COMBUSTION

This is a set of lectures on Combustion Basics (there is little here on practical systems), organised in the
following topics:
Combustion characteristics. A descriptive presentation of what is combustion, what it is for
(applications), how it is done (practical combustors), and what was historically known.
Fuels, Fuel properties, Fuel consumption, and Pyrotechnics. An extensive descriptive presentation,
with an historical development review.
Combustor characteristics. A short description, following a block diagram approach, of what to
study in a generic combustor: intake, internals, heat and work flows and exhaust.
Environmental effects and hazards in combustion. A descriptive presentation of potential source
of damege in combustion applications, with a review of fire safety, pollutant emissions, and
generic safety management.
Combustion thermodynamics. A mathematical formulation of equilibrium conditions, based on
the extent of reaction and the affinity of reaction, and with emphasis on the enthalpy of reaction
(the maximum heat, or heating value), the exergy of reaction (the maximum work) and the exhaust
equilibrium composition.
Combustion kinetics. A descriptive presentation of the detailed mechanism of reaction rates,
activation energy and its modification by catalysts.
Combustion models . A mathematical formulation of some key combustion problems:
combustion at rest, premixed combustion and non-premixed combustion, with emphasis on flame
geometry.
Combustion instrumentation. A descriptive presentation of devices and procedures for the setting,
control and diagnosis of combustion processes.
Back to index

COMBUSTION CHARACTERISTICS
Combustion characteristics (Introduction) .................................................................................................... 1
Combustion fundamentals (What it is) ..................................................................................................... 1
What it is not ......................................................................................................................................... 3
Thermal free-flame gaseous combustion .............................................................................................. 4
Thermal trapped-flame combustion in porous media ........................................................................... 4
Smouldering (Thermal non-flame combustion of porous media) ......................................................... 5
Catalytic combustion............................................................................................................................. 5
Detonating combustion ......................................................................................................................... 6
Combustion applications (What it is for) .................................................................................................. 6
Heating .................................................................................................................................................. 7
Propulsion and electricity...................................................................................................................... 7
Absorption refrigeration ........................................................................................................................ 7
Chemical transformations ..................................................................................................................... 8
Combustion systems types (How it is done) ............................................................................................. 8
Steady combustion chambers ................................................................................................................ 8
Unsteady combustion chambers............................................................................................................ 9
Catalytic combustors ........................................................................................................................... 10
Porous burners..................................................................................................................................... 10
Fluidised bed combustion ................................................................................................................... 11
Open fires ............................................................................................................................................ 11
Combustion history (What was known).................................................................................................. 11
Fuel history ......................................................................................................................................... 11
History of Combustion theories .......................................................................................................... 12

COMBUSTION CHARACTERISTICS (INTRODUCTION)

There is much more to combustion than a fuel in air and an ignition source. To better appreciate the wide
range of involved phenomena, a description of combustion basics (combustion types and processes), and
combustion applications (combustor types and systems), is presented here, before a more rigorous
treatment of the thermodynamics, kinetics and metrology of combustion.

COMBUSTION FUNDAMENTALS (WHAT IT IS)

Everyone knows from infancy what a fire is; humans have always felt a mixture of fear and magical
appeal for fire. Combustion is burning, a self-propagating oxidative chemical reaction producing light,
heat, smoke and gases in a flame front. Combustion is a process and fire is the actual outcome. What does
it mean in more detail?
Combustion means burning (lat. cum urere-ustus = burn), e.g. burning wood in air, natural gas in
air (or CH4/O2/N2 mixtures in general), hydrogen with oxygen (H2/O2 in gaseous or liquid forms,
and not only H2 in O2, but O2 in H2), and more bizarre burnings, such as sodium with chlorine
(Na(s)/Cl2(g)), aluminium powder with water, magnesium powder with carbon dioxide,
nitrocellulose (cellulose is -(C6H10O5)n- with n=300..2000) within any medium, etc. But the
meaning of combustion is usually restricted to easily flammable substances (typical fuels) in
ambient air. Fuels and oxidisers are presented aside.

Self-propagating, means that, once it ignites, it goes on, sustained by the high temperatures and
radicals (active species) produced, until either the fuel or the oxidiser practically runs out, or an
extinguishing agent is applied that prevents fuel-and-air mixing, or cools the system well below
autoignition, or scavenges active species. Notice the two sequential steps in combustion: first there
is an endothermic process of ignition, followed by a much more powerful exothermic process of
runaway oxidation that propagates the process. Any exothermic adiabatic system will show a
thermal runaway at some high enough temperature (autoignition), but for real non-adiabatic
systems, ignition criteria are governed by an interplay between heat-release rate and heat-loss rate.
Materials are termed non-combustible if they cannot be ignited below 1000 K.
Combustion propagation is usually a slow process; very slow indeed for solid and liquid fuels:
<0.1 mm/s for a candle flame (that were used as clocks to measure hours), 1 cm/s for flames
spreading on solid fuels, usually <0.5 m/s for premixed fuel-and-air gases burning at rest
(although it may be raised to 100 m/s in high-turbulent flows, and may reach >1000 m/s in
detonations). However, a common feeling and fear is that combustion is explosive, because the
pressure-rise in confined spaces may give rise to violent mechanical explosions (brick walls are
not pressure vessels).
According to the mechanism for propagation, several types of combustion processes may be
distinguished:
Thermal free-flame gaseous combustion: the usual case for combustion, e.g. in a candle, a
lighter, a bunsen, an internal combustion engine (reciprocating or turbine), a furnace, a boiler,
etc.
Thermal trapped-flame combustion in porous refractory media: a new combustion procedure
derived from the common one above-mentioned, to substantially increase combustion intensity
and stability, but problems of refractory materials exist.
Smouldering (or Thermal non-flame combustion of porous media): the slow burning without
flame of porous combustible matter, e.g. cigarettes, wood or coal embers, etc. It has little
engineering use, but great safety interest because uncontrolled fires usually start by
smouldering.
Catalytic combustion: active species, instead of temperature, may propagate the combustion
process, usually without flaming: e.g. room-temperature reaction of hydrogen and oxygen in a
platinum surface (it may flame if a large contact area exists, as in a porous catalyst that
releases so much heat as to ignite the rest).
Detonating combustion: when the combustion process is coupled to a high-pressure shock
wave, travelling at supersonic speed.

Oxidative chemical reaction. Combustion is an electron-exchange reaction (a redox one), not a

simple electron-cloud distortion as in proton-exchange (acid-base) reaction. The fuel atoms supply
electrons and get oxidised, whereas the oxidiser atoms get the electrons (get reduced).
Light, heat, smoke and gases in a flame front. Combustion results in a large temperature increase
in the products, typically from 300 K to 2500 K, causing them to be in the gas state (except for
soot and more rare refractory particles) and establishing a radiation imbalance in the infrared and
visible ranges. Light emission in flames only approaches blackbody radiation if there are solid
particles, such as soot found in non-premixed flames (e.g. yellow bunsen). For premixed flames,
light is by chemiluminescence in special spectral bands, and very dim in intensity (e.g. blue
bunsen).

It is the visible light of non-premixed flames that has been traditionally identified with combustion
(it is the standard symbol for fire). In fact, it might help to think of the flame as an invisible, very
hot, burning interface, made visible by non-burning incandescent substances passing by or being
created, for instance soot particles in non-premixed flames (their sublimation temperature is
around Tsubl=3900 K), sodium ions in salt-seeded flames (above the salt boiling point Tb=1690 K),
or calcium oxide in limelight (Tb=3100 K). The latter was used in the 19th c. in theatres as the
brightest, most natural-colour artificial light available, being produced by placing a block of lime
against a hydrogen/oxygen jet flame (practically invisible in spite of its 4000 K temperature; lime
melts at 2850 K). The Sun also gives light and heat (at a temperature of 5800 K in its surface) but
by nuclear fusion reactions in the interior (where the temperature may reach 107 K) and not by
chemical combustion.
What it is not
Combustion vs. explosions
Combustion means burning and explosion means bursting, i.e. combustion is a relatively slow chemical
process yielding light and heat, whereas explosion is a sudden mechanical process causing rupture and
noise, due to great pressure forces that may be originated chemically (e.g. from a confined combustion),
thermally (as in boilers, even electrically heated), mechanically (as in a balloon or any other gaspressurised vessel), nuclearly, etc. Detonation, the supersonic combustion taking place under some
circumstances in premixed fuel/oxidiser gaseous mixtures and many explosives, is studied aside.
Combustion vs. fuel cells
Fuel cells are electrochemical generators, like batteries but with continuous fuel-and-oxidiser supply.
Reactions inside a fuel cell, although globally equivalent to combustion, are not properly combustion
because they do not self-propagate (reaction in a fuel cell stops as soon as the electrical load is switched
off, it shows no thermal-runaway). A non-premixed burner (e.g. a lighter) may be thought of as
controllable as a fuel cell (as soon as the fuel injection stops, combustion ceases), but it does not simply
starts over if reopened. A controllable-area catalytic combustor, however, more closely resembles a fuel
cell: no need of igniter, simple reaction control, and for small active areas there is no runaway (a big
difference is that fuel cell directly generates electricity and the catalytic combustor just heat).
The igniter in a combustor (a spark or a hot wire) and the electrical connector in a fuel cell, act as
catalysts that provide a gateway for the reaction; in both cases there is an electron-transfer reaction (redox
reaction), the main difference being that the transfer of electrons from fuel to oxidiser is restricted in a
fuel cell by electrode-interface-area and electrolyte-ion-diffusion, with the external electrical connector
required all the time, whereas in normal combustion the electrons transfer is only limited by diffusion in
the bulk, and the igniter is only needed to start the process. Entropy generation, positive in both cases,
tends to zero in a fuel cell at very low intensities, but it is always above a certain finite value in
combustion.
Combustion vs. oxidation
Combustion is a self-propagating oxidative chemical reaction characterised by a thermal runaway; i.e. it
is a quick exothermal oxidation. The same system may undergo slow oxidation or combustion, with the
same initial and final states, but with different paths (e.g. paper turns yellow (and brittle) with the years
because of slow oxidation, but may burn in seconds). Notice also that oxidation may be exothermic or
endothermic, whereas combustion is always very exothermic.

THERMAL FREE-FLAME GASEOUS COMBUSTION

This is the usual case for combustion, that self-propagates as a result of the high temperature (1500..3500
K) developed after initial ignition (e.g. by a spark), due to a more-or-less adiabatic conversion of
chemical-bond energy to internal-thermal energy within a reacting gas mixture (for condensed fuels, the
latent heats for vaporisation and possibly decomposition has to be subtracted).
According to the initial state of mixing of fuel and air, combustion process can be classified in the limit as
premixed and non-premixed (real processes are in between), with a corresponding premixed and nonpremixed flame. In premixed combustion the unburnt gas is already a perfect mixture of fuel and air, and
the burning or flame-propagation speed is only limited by the chemical kinetics of the reactions involved
and heat diffusion forward, whereas in non-premixed combustion there is not a characteristic burning or
flame-propagation speed, the speed being fixed by the flow-rates of fuel and oxidiser that must approach
the flame by diffusion from each side.
Steady free flames demand an astonishing fine balance for heat and mass flows (e.g. think why a candle
flame sits at a precise distance up the wick). If adiabaticity of the initial ignition region is prevented by
nearby heat sinks, as a cold solid wall, this kind of 'thermal' combustion cannot propagate (safety lamps
and quenching grids are based on this fact); e.g. for premixed methane/air stoichiometric mixtures,
thermal free-flame combustion cannot propagate inside a metal tube of less than 2 mm.
Thermal flames are almost always established in a gas phase: in a fuel/air gas mixture or in the fuel
vapours diffusing in air, from liquid or solid fuels. Only refractory fuels like coal to some extent, tantalum
or zirconium, burn at the solid surface; iron and titanium, having intermediate melting points for both the
metals and the oxides, burn at the surface of a molten mixture of the metal and its oxide, whereas
aluminium and magnesium, which have low boiling points, vaporize and then burn in the gas phase.
The presence of condense matter is always a handicap (all condensed fuels burn worse than their
vapours), and this fact is used in fire-fighting. However, flames may be sustained inside liquids in a
suitable gas envelop. In order to maintain a steady underwater flame (e.g. a hydrogen or acetylene
welding torch), it is necessary to form a stable bubble, usually achieved with an additional compressed-air
jet-stream introduced around the tip of the torch, since the exhaust gases cannot maintain it by themselves
(a great deal of skill is required of divers who perform this kind of work).
Thermal free-flame combustion cannot propagate if the air/fuel ratio lies outside of the lower and upper
flammability limits at ambient conditions: e.g. 5% and 15% of fuel by volume of mixture for methane/air
flames, respectively, although increasing temperature widens this range. Neither can flames propagate
also at very low pressures (a fire safety rescue in spacecraft).
THERMAL TRAPPED-FLAME COMBUSTION IN POROUS MEDIA
Flames cannot propagate through small holes in a solid (this is how Davy's safety lamp work), unless the
solid is hot enough (say >1000 K), what can be achieved by holding a lit free flame close to the solid for
some time. For instance, if a premixed methane/air stream is forced through a finite porous medium
(usually solid, but also fluidised), and ignited at the exit, the free-flame formed may travel upstream or
downstream according to the flow speed. If the injected gas speed matches the deflagration speed, the
flame sits steadily at the mouth and, after the porous end gets hot, slowly decreasing gas injection speed
allows the flame to go backwards and penetrate the porous media, which is being heated by the slowly
moving flame front. The process is also known as filtration combustion.

Notice that the flame-front temperature is lower than the adiabatic value when the front moves
backwards, but, if the gas injection speed is increased to force the flame to travel downstream, then its
temperature is above the adiabatic value (it is moving into an already hot solid).
Presently only Al2O3 and ZrO2 can work above 2000 K, SiC and FeCrAl-alloys being used below 2000 K
with the advantage that they have larger thermal conductivities and mechanical resistance. In spite of this
basic materials difficulty, porous-medium burners have several advantages:
Less NOx emissions because of lower temperatures.
More compact because the deflagration speed increases from 0.5 m/s to 4 m/s (they reach 3000
kW/m2 instead of the 300 kW/m2 of normal burners).
Wider range of ignitable compositions (lower limit decreases from 5% to 4% in methane/air
flames, increasing the air ratio from =1.9 to =2.2, allowing for leaner mixtures to be burnt,
yielding lower emissions).
Wider power-modulation range (0.1 to 1 times full load, against 0.5 to 1 for normal burners, so
that start/stop cycles and accumulators are avoided). Porous media combustion was developed
aiming to stabilise premixed flames near their lower stability limit.
SMOULDERING (THERMAL NON-FLAME COMBUSTION OF POROUS MEDIA)
Before, combustion 'in' a porous-media was considered; now combustion 'of' a porous-media-fuel is
analysed. Some porous (solid) fuels may sustain a self-propagating combustion inside their matrix, i.e. an
heterogeneous reaction, at a very low rate and low temperature, taking the oxidiser from the ambient
through its pores, with little or no visible flame, but with change of external texture (which chars) and
smoke emission (sometimes very toxic).
A typical smouldering process takes place in the tip of a lighted cigarette. The porous fuel is inside a
porous cylindrical envelop (of paper in a cigarette, or a full tobacco leaf in a cigar) containing crashed
dried tobacco leaves (they were smoked or chewed by American Indians since ancient times because of
the euphoric action of nicotine). Once the cigarette lighted, if left in still air without drawing, a dim
burning happens with maximum temperatures of some 85050 K at the centre and some 65050 K at the
periphery. During drawing, however, those values go up to 100050 K and 85050 K, respectively.
A big fire-safety problem is that, the burning taking place in the insides, smouldering can progress
undetected for long periods of time (smoke detectors are used, but diffusion is very inefficient. Thus,
under certain circumstances, smouldering may undergo a sudden transition to flaming; that is why some
forced convection of cabin-air and avionics-air is always procured in spacecrafts, where buoyancy forces
are absent, to help an early detection.
CATALYTIC COMBUSTION
Even in the close proximity of cold solids, combustion may be sustained if there is some catalytic
substance that lowers the activation energy and sustains the combustion process at low temperature at the
porous surface; e.g. H2/air react over Pt at room temperature, CH4/air over Pt requires 350 C. The most
developed catalytic combustor uses C4H10/air over Pt-coated ceramic matrix at 500 C. Low-temperature
catalytic combustors work in the 300..600 K range, and high-temperature catalytic combustors work in
the 700..1400 K. They may work with premixed flows and with a non-premixed fuel-flow and ambient
air.
The advantage of power modulation, widening of ignition limits and less emissions, are maintained or
increases here (<1 ppm for NOx and CO, due to very low temperatures involved), but not the higher
burning rate (that was due to thermal conduction along the hot solid matrix, and thus only 600 kW/m 2 are

achieved instead of the 3000 kW/m2 of thermal porous-medium burners. At intermediate temperatures, an
hybrid regime of catalytic-assisted thermal combustion may be developed, where both heterogeneous and
homogeneous reactions take place, an interesting trade-off solution when the catalyst is too expensive and
very little can be used (in full catalytic mode, if there is insufficient catalyst, part of the fuel slips to the
exhaust (increasing pollution and expense), or even the whole fuel if the catalyst cools down and
deactivates. The influence of the flow rates is important; for instance, if a thin porous solid is doped with
a catalyst and a premixed methane/air stream is forced through, an exothermic oxidation of the fuel takes
place if the temperature is >600 K, releasing much of the lower heating value, what causes the matrix to
reach some 1500 K at the outer surface (the hotter) in a flameless regime with a power of up to 500
kW/m2; but if more gases are fed, small flames detach from the outer surface and at about 1000 kW/m2 a
blue flame at some 1900 K forms immediately close to the outer surface, that only reaches now some
600 K.
DETONATING COMBUSTION
Supersonic combustion, called detonation, is realised when combustion gets coupled to a shock wave
travelling at supersonic speed. Detonating combustion is considered under Explosives in Pyrotechnics,
and under Supersonic combustion in /Combustion kinetics.

COMBUSTION APPLICATIONS (WHAT IT IS FOR)

Combustion is an old technology to humankind, only second to simple mechanical tools made of wood,
bone and stone. Is it just then a subject of historical interest? For instance, from pre-historic times (some
500 000 years ago, at least) to the end of XIX c. the only lighting method was combustion, but today only
electrical lamps are used, so, why spending effort on candescent lighting?
We want to devote an effort to understand combustion processes for several reasons:
1. To understand it (i.e. for its own sake, the general interest to broad our knowledge of the world).
2. To minimise the risks and damages of controlled combustion and uncontrolled fires (safe fire
handling).
3. To advance the efficient use of fossil fuels (quickly exhausted) to satisfy our needs (e.g.
transportation propulsion, electricity generation).
4. To alleviate the problems associated to combustion emissions.
Combustion always gives off heat and gases, most of the times light, and sometimes smoke (a nuisance in
non-premixed flames, and the major personal danger in fires). Combustion is the major energy release
mechanism in the Earth (it is gravitational dissipation in jovian moons). Why? Because the Earth
atmosphere is plenty of free-molecular oxygen (50%wt of the whole ecosphere is oxygen, but mostly in
compounds) and Earth ecosphere is plenty of fuel (all living matter plus fossil fuels). Why? It is nearly a
miracle because at the start of the Earth ages, 4.5109 yr ago, there was no O2 but some H2O that
generated a small amount of O2 by solar-radiation hydrolysis, part of which was transformed to UVshield O3 by solar radiation and thus allowed the start and survival of primitive plants. Plant
photosynthesis has generated all O2 in the atmosphere and all living (and fossil) fuels on Earth.
Combustion (controlled and uncontrolled) releases an average of 101012 W in the globe, as compared to
0.51012 W due to nuclear disintegration (controlled and uncontrolled), 31012 W due to geothermal heat
flux and 120 0001012 W due to the solar absorption.
Combustion is key to humankind today as in the past and the foreseeable future; it is worldwide the major
energy-release process (nearly 90% of the world primary energy is to be burnt), it is the major
environmental-pollution process, and fire is one of the most feared natural disasters. Roman Emperor
Augustus established a corps of fire-fighting watchmen already in 24 b.C.; their fire-fighting techniques

are still applied: removing heat by throwing water with a bucket passed from hand to hand, removing the
fuel by cutting it aside with an axe (or, at a much later times, by providing firebreaks), and opening doors
and windows in enclosures to get rid of smoke and let fresh-air in for people sake (unfortunately feeding
the combustion process, too).
Even nowadays 88% of world commercial energy consumption (i.e. excluding photosynthesis and
uncontrolled fires) is by combustion (plus a 5% hydraulic and a 7% nuclear fission, roughly). This
combustion-released energy (as any energy) is used today for:
Heating. Internal chemical energy is first released to internal thermal energy (high temperature),
and then heat transfer to the load takes place.
Propulsion and electricity generation. Mainly by means of a heat engine with a working fluid that
may be an independent one (e.g. steam), or directly the combustion products.
Refrigeration. By an absorption refrigerator driven directly by the combustion products or through
an intermediate fluid (e.g. steam).
Chemical transformations. Mainly as a heating application, but sometimes directly for reduction,
oxidation, incineration, reforming, etc.
Light emission was a major application of combustion up to the XIX c., but nowadays it is a
rarity: candles, rescue signals, fireworks.
There are other unusual applications of combustion, as smoke generation for visualization, gas generation
for fuel-tank inertisation, aromatherapy, etc. Explosives applications are considered aside.
HEATING
Heating for space conditioning or for materials processing (domestic and industrial) has always been the
basic combustion application. Energy efficiency in heating is usually measured as the fraction of the
lower heating value of the fuel that is fed to the system to be heated (the rest is lost by the flue), and it is
close to 1 (that is not a limit: heat pumps can give more).
A special heating application is the cutting of materials. Thick metal plates can easily be cut with a
normal gasoline, propane or oxy-acetylene torch, with additional oxygen to burn the metal (that is a fuel
itself; hint: think of M+O2=MO2, as for carbon). For cutting or piercing thick ceramic objects, a thermal
lance is used, in which, an iron tube act as the fuel that burns (once ignited) with oxygen being supplied
through it (sometimes in the cryogenic liquid state). A small 5 mm-diameter tube, 0.5 m long, may burn
for one minute consuming some 5010-3 kg of oxygen, producing a >5000 K flame that would pierce a
150 mm-thick steel slab in less than 10 s (a 0.5 m deep hole can be pierced in stone or concrete in a few
minutes, anywhere: in the air, under water, or within mud and slurries).
PROPULSION AND ELECTRICITY
Motion and electricity generation, by means of heat engines (vapour and gas cycles), constitutes the
foundations of Thermodynamics. For electricity generation, energy efficiency is measured by the ratio of
electrical energy delivered divided by the lower heating value of the fuel (it typically ranges from 0.2 to
0.5), or by the specific fuel consumption (of the order of 60 g/MJ or 200 g/kWh), and it could be similarly
done for mechanical power, although for cars it is usually done in litres of fuel consumed per 100 km
travelled (or in miles per gallon in the USA; the change is x [mph]=235/x [L/100 km]).
ABSORPTION REFRIGERATION
Absorption-refrigeration machines use heat (usually from steam or from a fuel burning) to separate the
working vapour from a high-pressure liquid mixture. The working vapour condenses in a heat exchanger
with the ambient, and is flashed through a restriction to a low-pressure heat exchanger where the cooling

effect is produced, exiting into a liquid absorber that is pumped and closes the loop. Energy efficiency is
usually measured as the amount of cooling divided by the amount of fuel or steam consumed (in vapourcompression refrigeration, energy efficiency is measured by the ratio of cooling energy to electrical or
mechanical energy input).
CHEMICAL TRANSFORMATIONS
Some materials processing, e.g. mineral reduction, cooking, distillation, require an exergy supply (other
materials processes are used as an exergy source, as combustion); incineration may have positive or
negative heat balance, depending on the materials. In the typical case of energy consumption, energy
efficiency is usually measured as the amount of material processed divided by the amount of fuel
consumed.

COMBUSTION SYSTEM TYPES (HOW IT IS DONE)

According to the steadiness of the control frontier (real or imaginary), and to how the propagation is
maintained, one may distinguish several type of combustors, traditionally grouped in steady combustion
chambers and unsteady combustors.
STEADY COMBUSTION CHAMBERS
Steady state combustion takes place in industrial heaters, furnaces, boilers, steam turbine boilers, gas
turbines, and domestic appliances: from the cooking range to the gas lighter. Steady refers to the average
process, but it does not mean that the process is locally steady; most practical systems operate in a highly
turbulent regime. The combustion process may approach the non-premixed flame model or the premixed
flame model, according to how much fuel-and-air mixing exists before burning. Walls must be cooled to
avoid materials problems at the high temperatures involved (>2000 K), and, most of the times, excess air
is used to cool the main burnt flow, that is nearly stoichiometric (see Fig. 1a).
Recently, partial recirculation of exhaust gases is being applied (instead of the secondary air in Fig. 1a)
for emission reduction, heat recovery, autoignition and flame stabilisation. Exhaust gas recirculation
(EGR) helps to keep a high uniform temperature inside the combustor, but lower than the hottest region in
normal flames; e.g. a maximum T1800 K with maximum spatial deviations of T100 K, entering the
air/EGR mixture at some 1500 K with as low as 2% O2 and producing a wide dim green flame, whereas,
in normal combustion, air enters at 300 K with 21% O2, producing a bright and long yellow flame with
maximum temperature T2500 K with maximum spatial deviations of T1000 K.
A modern home boiler (Fig. 1b) is a friendly, efficient and environmentally-clean combustion appliance;
it is permanently ready (safe, reliable, no surveillance, no servicing, 20 years life), provides some 20 kW
of space heating or hot water for sanitary use, is fed by natural gas that burns with 3% excess air,
extracting 94% of the maximum heating value (105% in terms of its LHV, due to exhaust condensation),
is small and silent, and only produces 80 mg/m3 of NOx as major pollution (besides the CO2 inherent to
the fuel carbon content).

Fig. 1. Steady combustion chamber layouts: a) industrial burner, b) domestic water heater.
Solid fuels are no longer burnt in chunks over a grate, but gasified (pyrolysed), pulverised (fine dust) or
fluidised (coarse granulate made to levitate by air entrainment). Dry (<15% water) coal powder (1 mm
size) can be burnt in cyclonic or screw burners, although it is expensive and only worth for >3 MW.
UNSTEADY COMBUSTION CHAMBERS
Unsteady combustion chambers are used in reciprocating internal combustion engines (ICE); also in the
(non-reciprocating) Wankel engine, not further mentioned here because of lack of use. The combustion
process may approach the non-premixed flame model (for the major part of the fuel burning in a diesel
engine), or the premixed flame model (for the spark-ignition engine and the initial stage for compressionignition engines). The period of burning and fluids renovation is so short, that the thermal inertia of the
wall materials makes higher peak temperatures allowable in the gas, contrary to steady combustors.
The two basic ICE realisations are the spark ignition engine (SI, gasoline, or Otto) and the compression
ignition engine (CI, diesel oil, or Diesel). The normal working for a SI-engine is with a premixture of fuel
and air at the stoichiometric ratio, tuning the output power by essentially chocking the entrance duct,
whereas for a CI-engine non-premixed fuel is added to highly-compressed air. SI-fuels must have high
volatilities (or be gaseous) for quick mixing, and have a high resistance to autoignition, to avoid
knocking. CI-fuels must have a low autoignition temperature and short delay.
The best ICE are direct injection (DI) exhaust gas recirculation (EGR) engines, both of Diesel and of Otto
type, the latter with stratified mixture (by injection control) and exhaust catalytic converter. Typical
air/fuel relative ratios are =0.9..1.05 for Otto engines, =1.3..1.8 for normal Diesel engines, and
=1.6..2.1 for large (>500 kW) slow (<200 rpm) marine Diesel engines.

Fig. 2. Unsteady combustion chamber in an internal combustion engine.

A modern direct-injection turbo-charging diesel-engine for a small family-car (like in Audi A2 1.2 TDI,
or VW Lupo 3L TDI) is a friendly commodity: it is safe, easy to start, reliable (2 years typical
maintenance-free, 10 years typical mean-time between failure, 20 years life), providing some 50 kW of
mechanical power with its 1200 cm3 total piston displacement and 18:1 compression ratio, it injects the
fuel directly to its three cylinders in line at more than 100 MPa when the compressed air is at some 6 MPa
and 700 C, extracting some 50% of the maximum work-value of the fuel (some 30% energy efficiency in
terms of its LHV), it is small and not too noisy and shaky, producing less than 0.3 g/km of NO x as major
pollution (and <0.1 g/km of CO, and <0.02 g/km particulate matter, besides the CO2 inherent to the fuel,
yielding some 100 gCO2/km), and is able to power a 5-seat car at 170 km/h consuming 3 litres every 100
km (at 90 km/h on good roads). Presently, however, EU and Japan cars are typically 2-litre, 4-cylinder,
80-kW engines, whereas USA cars are typically 4-litre, 6-cylinder, >100 kW-engines.

It seems that both Otto and Diesel engines are converging towards an hybrid, the new Homogeneous
Charge Compression Ignition engine (HCCI), where fuel injection during the compression stroke creates
a very lean mixture that auto-ignites by premixed compression near the top dead centre, producing a
better combustion for a steady regime (it is difficult to control the time of ignition if unsteady): less soot
and lower NOx emissions (lower temperatures) than a diesel engine, and higher efficiency that an Otto
engine (but unburnt emissions in between).
CATALYTIC COMBUSTORS
Catalytic combustors are porous catalytic solids inside which a low-temperature combustion takes place.
They are not so much developed as the simple thermal combustors above-mentioned (steady and
unsteady). They have the advantage of very complete oxidation (low emissions) and low-temperature
work, but they are expensive and bulky. They may be used for:
Catalytic burners, where heat is generated by burning a main fuel (C4H10, CH4, H2) in a catalytic
porous media. They are mainly used as radiant catalytic heaters for stoves and open-air industrial
heating (e.g. to cure the epoxy used to join pipes in gasoducts and pipelines, to dry wood and other
biomass).
Catalytic-assisted burners, where very lean mixture may be burnt at moderate temperatures,
avoiding the formation of NOx. They are being introduced in gas-turbine engines and plants.
Catalytic converters, where emissions are remediated without interest in heat generation (e.g. CO
and VOC in car exhausts are completely oxidised). This is the most developed application; most
modern cars have a catalytic converter in the exhaust pipe (see Three-way catalytic converter).
Although simple stagnation flows over a catalytic plate are studied in the lab, practical systems consist of
monolithic honeycombs or granular material (porous media of 1 mm typical porous size), because of the
surface effect on heterogeneous reactions. The combustion catalyst per excellence is a platinum doped
refractory matrix.
Catalysts may work at room temperature (as in Pt for H2/O2), but most usually they require some initial
heating (as in Pt for CH4/air), achieved electrically (from the mains, or with a battery), by a conventional
pilot flame, or even by a room-temperature catalytic combustor (as in hydrogen-assisted natural-gas
catalytic combustor). The element preheats the catalyst bed to some 400 K, the temperature at which the
catalytic pad will sustain a chemical reaction with the fuel and ambient air, and the fuel supply is
regulated to operate steadily at some 700 K.
POROUS BURNERS
Porous burners are porous non-catalytic solids inside which a high-temperature combustion takes place
with small flames trapped in the pores. They are in the development stage, showing promise of very high
combustion intensities (3000 kW/m2 of cross-section) and low emissions, because of increase heat
transfer and flame stability. They may work with a premixed flow (fuel and air been forced through one
end), or, more rarely, with a non-premixed fuel flow that meets a back-diffusing air flow at the exit.
Premixed porous burners consist of two sequential stages: the premix fuel/air stream first enters a finepore hot solid matrix (below the flame quenching size, for safety), where it is heated until it enters the
second larger-pore hottest solid matrix, where alveolar flames stabilise themselves for a wide range of
flow rates (e.g. from v=0.2..4 m/s for CH4/air) and air/fuel relative ratios (e.g. from =1..2 for CH4/air).
Depending on the application, a third stage may be directly added to the burner, with a compact heat
exchanger.

The wide power modulation capability of porous burners make them ideal for applications such as home
water heaters, where the power needed to heat sanitary water cannot be less than say 15 kW and is
increasing, whereas the power for space heating is decreasing from typically 20 kW to 5 kW for betterinsulated apartments. Besides, being more compact, porous burners could be installed in wall niches
instead of protruding significantly. This compactness is also an advantage for industrial heaters and
driers, that must be 1 m to 3 m long if conventional (to protect from the flame), but can be shorten to less
than 0.5 m using a porous burner.
FLUIDISED BED COMBUSTION
Fluidised bed combustion is conceptually similar to porous burner combustion, but, instead of a rigid
solid matrix, a high-temperature combustion takes place within a particulate system (the fuel particles
themselves, or sand or other granular refractory), held over a porous surface. When an upward air-stream
(or a premixed fuel/air stream) is established, for small upward speeds the particulates act just as a filter,
but when the drag on the particles overcome its weight, a fluidised regime is establish (at the fluidising
speed), with more or less bubbling or boiling when speed is increased until at a certain value is
reached (the terminal velocity) beyond which the particles are ejected from the bed and carried away with
the stream.
The main advantage of this modern type of combustors is that additives can be easily added and the
porous material easily refurbished. The disadvantages are the difficult start-up, difficult ignition (it is
ignited above the bed), noise, and erosion of the heat-exchanger walls (that is partially or totally
submerged to limit the maximum temperature within the bed). Steady burning is achieved at 1100 K to
1300 K (uniform red-hot bed), but below 1100 K the burning is noisy with violent small bursts, and above
1300 K the particulate may start fusing and clogging. Fluidised bed combustion is currently used in some
coal-fired power plants, where coal-desulfuration is achieved during combustion by adding lime to the
bed. It is also been used to burn high-moisture fuels like agriculture and industrial residues and sludge.
OPEN FIRES
Most practical combustion systems are enclosed inside easily-identifiable walls (combustion chambers,
empty or filled with porous media). By 'open fires' we here just mean unenclosed combustion setups,
from the traditional fireplace to the wild uncontrolled forest fire, passing by the humble candle flame.
Because the information here compiled is basically historical, it can be found under Torches, oil lamps
and candles, in Fuels.
Smouldering is not intentionally used in engineering, but it is very important in uncontrolled fires. Porous
materials, either solid like wood or particulate like dust, may sustain a low-temperature combustion inside
with air penetrating by diffusion. A smoulder temperature may be identified to describe the flammability
behaviour of a flat dust layer on a hot surface (e.g. the lowest temperature of a heated surface capable of
igniting a 5 mm thick dust)

COMBUSTION HISTORY (WHAT WAS KNOWN)

Although the history of combustion is related to the history of fuels, we treated them here separately
because under fuels we focused on empirical findings and under combustion we centre the attention on
theories.
FUEL HISTORY
This theme can be found in Fuels. Briefly, Fuel history has the subheadings there: Biological fuels before
the XX c., Mineral fuels (coal, crude oil and natural gas), and Biological fuels in the XXI c.

HISTORY OF COMBUSTION THEORIES

Fire has always been a cause of panic, but also of immense power. It had a magical appeal: Prometheus
stole it from the gods, according to Greek mythology. Around 500 b.C., Heraclitus of Ephesus held that
fire is the primordial substance of the universe and that all things are in perpetual change (like a vital
flame). Around 450 b.C., Empedocles of Sicily set the theory of the four universal elements: fire, air,
water, and earth, further developed and spread by Aristotle around 350 b.C. (he added aether as the
quintessence, in his Metaphysics).
Coming down to our physics, it was appreciated from the beginning that combustion required a fuel (e.g.
wood, but not stone) and an igniter (by rubbing woods, sparking stones or compressing air, but not just
hitting the wood). However, it was only learnt in the 17th c. that air was also needed; around 1663, Boyle
realised that, under vacuum, sulfur could not be ignited by concentrated light, as usually achieved, but it
was Lavoisier in 1777 the first to realise that oxygen was the key, both to combustion and to animal
respiration. Much earlier than the need of air for combustion was realised, it was common practice to
preserve fire overnight by putting a fire cover (curfew) over it, to avoid air convection and heat loss.
The concept of energy (chemical, thermal, electrical), as an extension of the primitive mechanical idea of
combination of 'head' (potential energy) and 'vis viva' (kinetic energy), was only generally accepted after
mid 19th century. Before that, imaginary fluids were thought to explain the facts. Around 1660, J.J.
Becher and G.E.Stahl introduced the theory of phlogiston as a fluid that flowed away in combustion
processes. Lavoisier, in his paper "Rflexions sur le phlogistique" of 1783, rejected the phlogiston theory,
but set forth the caloric theory, a fluid that flowed away with heat (some authors even advocated for
another fluid for cooling processes, the frigoric). Although already in 1798, Count Rumford rejected the
caloric theory in his essay 'An experimental enquiry concerning the source of the heat which is excited by
friction', it was not until the 1840s that it was finally discarded and replaced by the principle of energy
conservation, by Mayer and Joule.
A good appreciation of the level acquired in combustion theory can be grasped from the titles of Michael
Faraday's famous series of Christmas Lectures for Children, delivered in the mid 19th c. (starting in 1826)
at the Royal Institution. The most important (actually six of them) was published as "The Chemical
History of a Candle" in 1860:
1. Construction of a candle. The flame.
2. Necessity of air for combustion.
3. Products of combustion; water from combustion.
4. The components of water are oxygen and hydrogen.
5. Components of air: oxygen and nitrogen. Product from a candle: carbon dioxide.
6. Human respiration and its analogy to a candle.
A summary of the chronological development of combustion theory is presented in Table 2.

The need of air

Phlogiston

Table 2. Chronology of combustion theories.

Robert Boyle-1663 found that combustion requires air, besides the fuel.
Proposed by the German physician and chemist Georg Ernst Stahl-1697 and
discarded by Antoine Lavoisier-1783-Rflexions sur le phlogistique, (lat.
phlegma=fire spirit): a substance (with mass) stored in fuels and released in
chemical reactions (the flammability component), aside of the caloric substance (the
heating component). Fuels released phlogiston quickly, whereas metals did it slowly
(rust). Afterwards, Joseph Priestley-1796 tried to come back to phlogiston.
Phlogiston principles:

Phlogiston is a material fluid present in fuels (wood, coal, metals and living
beings), that escapes from them when bodies are heated to burn.
Air is a phlogiston absorber, needed to let it escape from fuels. If air is
trapped inside a bell, fuels cannot yield all their phlogiston. Plants slowly
take phlogiston from the air on sunshine, and may suddenly release it if
burned.
Phlogiston is conserved; e.g. animals exhalate phlogiston by the mouth, is
stored in the air and plants, and returns to animals when they eat plants, or to
air if plants are burnt. The phlogiston released by metals when rusting may be
fed-back from phlogiston-containing substances (metals are obtained from
their ores by calcination with wood or coal). Water dissolves phlogiston, but
does not destroy it.
Phlogiston can be weighted by difference before and after burning the fuel.
But metals release so little phlogiston that it floats in air, making the metal
heavier when rusted.
Question. What was phlogiston: H2 (released by metals when pouring an acid over)
CO2 (exhaled by animals), or O2? Answer: the closest idea is to -O2, i.e. oxygen
affinity.
Caloric
Proposed by Lavoisier-1777 and discarded by Mayer-1842, a substance (with mass)
released in thermal processes. Similarly, electricity was a substance (with mass)
released in electrical processes.
Mass
Proposed by Lavoisier-1777: the overall mass is preserved in any confined chemical
conservation
reaction. He also established the taxonomy of chemical compounds, and identified
oxygen-in-the-air as the main oxidiser, and carbon and hydrogen as the main fuelconstituents.
Fixed
Proposed by Proust-1799: chemical reactions take a fixed amount of substances
proportions
relative to another, and the excess appears un-reacted (quite different to a mixture).
Multiple
Proposed by Dalton-1803: different chemical reactions between same reactives take
proportions
a fixed amounts of substances relative to another, the ratio between which are small
whole numbers. Berzelius-1826 accurately measured the relative masses in chemical
reactions and established chemical-formula notation (the first letter of the Latin
name with the relative factor as a superscript, later changed to subscript, e.g. H2O).
Flame
Sir Humphry Davy published in 1800 his personal trials on the physiological effects
quenching
of some gases of importance to combustion (laughing gas, NO, and water gas, i.e.
Emissions
CO+H2, which nearly poisoned him). Around 1815 he discovered that flames cannot
go through wire meshes and developed the miners safety lamp.
Premixed flame Robert Bunsen, Prof. At Heidelberg, developed his premixed burner in 1855,
Flame spectra
measuring flame temperatures and flame speeds. Bunsen-1860 repeated Newton's
experiment of light dispersion by a prism, but with the light of flame, and discovered
that it was not a continuum but separated bands, particularly with premixed flames,
and the bands changed with addition of different substances to the flame (the
spectrum was characteristic of the substance). Bohr-1913 proposed the orbital theory
of electrons in the atom, the quantum numbers, and explained the emission spectra
of simple atoms in absence of magnetic fields (quantum mechanics explains
everything up to now).
Thermal model Mallard and Le Chtelier, Profs. at cole de Mines de Paris, studied flame
propagation in 1868 and proposed the first flame structure theory in 1883.
Chapman and Jouguet in 1900 distinguished deflagrations from detonations and
computed detonation speeds.

Thermal Ignition: Semenov's theory of 1928 (uniform temperature), FrankKamenetskii's theory of 1945 (with thermal gradients).
Thermal explosions. Zeldovich theory of 1943. He also proposed in 1943 his thermal
model for the formation mechanism of NO.
Diffusion model Burke and Shumann in 1928 made the first theoretical computation of non-premixed
flame height and shape.
ThermalZeldovich in the 1940s and Sivashiusky in the 1970s analysed thermal-diffusive
diffusive model instabilities in two-dimensional flames.
Thermo-fluidTheodore von Krman, the founder of the U.S. Institute of Aeronautical Sciences
chemistry
(1933), of the Jet Propulsion Laboratory (1944), and a co-founder of the Combustion
theory
Institute, organised and led (under NATO and UN auspices) an international team,
around 1950, to compile and spread multidisciplinary knowledge on combustion
science, producing his famous Aerothermochemistry notes and lectures (T. von
Krmn and G. Milln, Proc. Combust. Instit. 4, 173 (1953)).
Finite rate
Asymptotic methods for different chemical and flow time scales: activation energy
chemistry
and reaction-rate ratio asymptotics. Steady-state and partial-equilibrium
approximations.
CFD codes
Simulation of multidimensional reacting fluid flow, including turbulence,
unsteadiness, multiphase flow, acoustic-coupled instabilities, etc.
(Back to Combustion)

FUELS
Fuels ..........................................................................................................................................................1
What is a fuel ........................................................................................................................................1
The oxidiser...........................................................................................................................................2
What is a fuel used for ..........................................................................................................................3
What is the problem with fuels? ...........................................................................................................3
Hydrocarbon nomenclature ...................................................................................................................4
Fuel types ..................................................................................................................................................5
By physical state ...................................................................................................................................5
Solid ..................................................................................................................................................5
Liquid ................................................................................................................................................5
Gas ....................................................................................................................................................5
By period of natural renovation ............................................................................................................5
Fossil fuels ........................................................................................................................................5
Renewable fuels ................................................................................................................................6
By production stage (resources type) ....................................................................................................6
Natural or primary fuels ....................................................................................................................6
Artificial or secondary fuels ..............................................................................................................6
By marketing .........................................................................................................................................7
Non-commercial................................................................................................................................7
Commercial .......................................................................................................................................7
Petroleum fuels .................................................................................................................................7
By application .......................................................................................................................................8
For spark ignition engines .................................................................................................................8
For compression ignition engines .....................................................................................................9
For gas turbine engines .....................................................................................................................9
For boilers .........................................................................................................................................9
For small portable applications .......................................................................................................10
For fuel cells....................................................................................................................................10
For pyrotechnics, propellants and explosives .................................................................................10
Fuel history .............................................................................................................................................11
Biomass fuels before the XX c. ..........................................................................................................11
The fireplace and the chimney ........................................................................................................11
Torches, oil lamps and candles .......................................................................................................12
Engines on biological fuels .............................................................................................................14
Mineral fuels .......................................................................................................................................14
Biological fuels in the XXI c. .............................................................................................................16

FUELS
WHAT IS A FUEL
A fuel (from Old French feuaile, from feu fire, ultimately from Latin focus fireplace, hearth) is a
substance that may be burned in air (or any other oxidant-containing substance), i.e. that so quickly
reacts with oxygen that heat and light is emitted in the form of a sustained flame.

Although the chemical reactions of wood in air, animal-fat in air, coal in air, natural gas in air
(approximated by CH4/O2/N2), hydrogen in oxygen (H2/O2), silicon in oxygen (Si/O2), sodium in
chlorine (Na(s)/Cl2(g)), zirconium powder in carbon dioxide (Zr/CO2), nitrocellulose (-(C6H10O5)nwith n=300..2000) in any medium, etc., are all of the same chemical type (self-propagating highlyexothermic re-dox reactions, i.e. combustion processes), usually 'fuels' and 'combustion' only refer to
easily flammable substances in air (the air is the oxidiser needed by a fuel to burn, and it is needed in
larger quantities than fuels, so, a first glance on it seems appropriate).

THE OXIDISER
Oxygen in the air is the basic oxidant for fuels: nitrogen is basically inert, although it combines
endothermically with oxygen at high temperatures to get the unwanted NOx pollutants. The most
abundant element on Earth's crust is oxygen, with 47%wt. It is also the most abundant in the
hydrosphere, 86%wt. In the atmosphere, it is second to nitrogen, with 23%wt. In the whole ecosphere
it is the first with 50%wt, and in the whole planet Earth it has 30%wt, only after Fe, 40%wt). Oxygen
is then readily available from Earth's atmosphere; that is why it is the main oxidiser. Silicon follows as
most abundant in the crust (28 %wt), but most natural Si-compounds are already fully oxidised, as
sand (SiO2) and silicate rocks (like CaO3Si); there is no silane gas (SiH4) free in Nature. Aluminium
follows (8 %wt), but again the main Al-compounds are already fully oxidised, as bauxite (Al2O3) and
silicates like Al2O3Si.
As for the fuels, hydrogen, the most abundant element of the Universe, is scarcely available in the
ecosphere, mainly fully oxidised as water in the hydrosphere, where it comprises only a 3%wt, and
only 0.9%wt in the whole Earth crust; the ninth most abundant element). Carbon has a poor
abundance: <0.1%wt in Earth crust, and it is found basically oxidised, both in the air as CO2, and in
the rocks as CaCO3, but there is some part un-oxidised in living and fossil matter, forming CnHmcompounds (hydrocarbons), which constitutes the conventional fuels.
Air is a constant-composition gas mixture, except for great variations in a minority component, water
vapour, which may range from nearly 0% up to 4%, with typically 2% in volume at the Earth surface.
Most combustion data for fuels refer to its burning with normal air, i.e. with 20% to 21% oxygen.
Oxygen-enriched air (up to pure oxygen) accelerates combustion processes and give way to higher
temperatures, whereas oxygen-depleted air may prevent combustion (usual fuels cannot be ignited if
xO2<10%). Lean-oxygen air is sometimes used as oxidiser (to diminish emissions), obtained by adding
to ambient air a part of the burnt gases, usually from the exhaust stream (exhaust gas recirculation,
EGR), but also by letting a sizeable amount of burnt gases within the cylinder in piston engines
(internal gas recirculation, IGR); in a 4-stroke engine, IGR may be just 4% but in 2-stroke engines may
reach 60%.
Before Lavoisier in 1776 gave name to most chemical compounds, all gases were named as air
variants (the word gas was introduced by Paracelsus around 1520 to name intangible substances).
Besides normal air (21% O2), flammable air (H2, obtained by dripping an acid over a metal) was
discovered by Priestley in 1766, fixed air (CO2, it fixed to quicklime, CaO, to form CaCO3, from
which it was obtained with sulfuric acid) was discovered by Priestley in 1771 and found to be unable
to support life or combustion, mephitic air (N2, 'poisonous' to respiration, the residue of combustion
inside a bell jar) was discovered by Rutherford in 1772 and found to be 4/5 parts of normal air, and
dephlogisted air (O2, obtained by heating HgO2 or KClO3) was discovered by Priestley in 1774.

WHAT IS A FUEL USED FOR

Fuels are mostly used as convenient energy stores because of their high specific energy release when
burnt with omnipresent ambient air (or other specific oxidiser); the same fuel substance may be also
used as a feedstock in chemical synthesis (e.g. polymers from petroleum), lubricants, paints (who has
never used a coal chunk to draw), and so on, but these uses are minority. Primary fuels (natural fuels)
may be difficult to find, and secondary fuels (artificial fuels) may be difficult to manufacture, but, once
at hand, fuels are very easy to store, transport, and use, with the only nuisance of safety (uncontrolled
combustion) and pollution (toxic emissions during storage and when burnt, dirtiness...).
Energy is a basic need to humans. Besides metabolic energy needed by any living being and supplied
by the catalytic reaction of food and oxygen (some 100 W for an adult person), descendants of homo
habilis need energy to change Nature to better suit their needs: to heat their home, to cook their meals,
to bring potable water in (e.g. from low river courses up to their dwellings) and waste waters out, to
remove the ground for aeration, to mill the grain, to turn the potter-mill, to throw weapons, for
industry, for transportation of goods and people, for telecommunications, etc. Fuels, and energy in
general, are used (see Fuel consumption, for more details) for heat generation, for work generation, or
for chemical transformations. A common problem to all human needs (except air, in most cases) is that
energy is not available at the location and time we desired, and sources must be found (for energy,
water, food, minerals), transportation to a better place must be arranged, as well as storage and end-use
details. Storage is sometimes the most cumbersome stage, e.g. for food (all food is perishable,
particularly meat and fish), and for electrical energy.
Specific energy storage values for fuels are presented aside, in comparison with other energy store
systems.

WHAT IS THE PROBLEM WITH FUELS?

Several:
Fuels are dangerous, because they accumulate a lot of chemical energy that may be accidentally
released, causing deathly thermal and chemical effects.
Fuels are pollutant when burnt (and even before; most liquid fuels are cancerous); they are
presently the major contribution to environmental pollution, both locally and at a global scale.
Fuels are scarce (fossil sources are being depleted) and the sources are unevenly spread (most
petroleum reserves are in the Middle East, causing economic and political instabilities).
Fuels are difficult to handle: coal is very dirty, crude-oil is too viscous, natural gas has very low
density, to say the less.
But, as just discussed, fuels are so convenient energy storage systems, that their associated problems
are but to be solved. Elaborating on the above problems:
Danger in fuel handling can be controlled and reduced to a relatively very low risk (in
comparison with other accepted risks, as in transportation or sports).
Pollution by fuels can be negligible if clean renewable fuels are used; basically a fuel is a C-Hcompound that combines with oxygen to yield CO2 (a natural compound of no harm if it does
not accumulate, as for biofuels) and H2O (the most life-compatible compound). A short-term
palliative to vehicle-engine pollution is to force a dual-fuel system (with dual fuel-reservoirs
and engine controls, or even two different engines), using a non-pollutant fuel inside cities,
ports and urban areas, and leaving the present more-convenient but more-pollutant fuels to
highway and cruising.
Scarcity of fuels is like scarcity of water: what we mean is scarcity of cheap good-quality
sources, i.e. that we have to devote a sizeable effort within our limited capabilities. The atoms

that make the fuel and oxidiser are preserved after combustion, and with the addition of some
external energy (freely available from the Sun) those atoms can be arranged to form the initial
fuel and oxidiser molecules. It may be difficult to think of such a new gadget added to our cars
that would regenerate the gasoline and air from the exhaust gases, but think on a fuel-cell car
that runs on H2+(1/2)O2=H2O and uses off-line-produced solar electricity (in the garage or at
the station) to recharge the tank by electrolysis H2O=H2+(1/2)O2. Thus, the problem is an
ancient one: it is hard to become a farmer if you can find suitable wild plants, animals and fuels.
Water-like fuels seem the best to handle, so coal should be liquefied (i.e., converted to liquid,
what is presently done by first gasifying it), crude-oil is distilled, and natural gas should be
liquefied (as done with coal gases by Fischer-Tropsch process, since the 1920s). Liquid fuels
are a must, particularly for vehicle propulsion (in land, water, air and space).

HYDROCARBON NOMENCLATURE
Most practical fuels are hydrocarbon mixtures (i.e. organic chemical compounds of carbon and
hydrogen atoms, and maybe some additional ones); the main exception are pure carbon (C, graphite,
since coal is already a mixture), and pure hydrogen (H2). Some chemistry refreshing is appropriate
here to better understand fuels:
Hydrocarbon nomenclature:
By type of additional atoms. Besides the ubiquitous carbon, hydrocarbons may be:
hydrogenated, oxygenated, nitrogenated, etc.
By type of molecular shape. chain (linear or branched), cyclic (homo- or heterocyclic).
By type of carbon bonding (single, double, or triple bond).
By mixture fraction specification (all natural fuels are mixtures):
From fractional distillation of raw materials (wood, coal, crude oil).
From chemical reforming (e.g. pyrolysis, cracking of heavy molecules, reforming
by synthesis with vapour or air, like syngas).
Hydrogenated compounds may be:
Aliphatic (oils and fats, little odour): alkanes (saturated hydrocarbons: paraffins (linear chain),
isoparaffins (branched chain), and cycloparaffins or naphtenes), alkenes (ethylenes or olefins),
alkynes (acetylenes).
Aromatic (strong smell): 1-ring: benzene, phenol (-OH), toluene (-CH3), aniline (-NH2),
nitrobenzene (-NO2), styrene (--CH=CH2)); 2-rings: naphtalenes; 3-rings: anthracenes; 4-rings:
tetracenes; etc...
Oxygenated compounds may be:
Carbohydrates: Cn(H2O)m, like sugars, starch and cellulose.
Alcohols: R-OH (phenols if R is aromatic; glycols if double OH-group, glycerols if three OH)).
Methanol comes from wood distillation or natural gas reforming. Ethanol comes from biomass
fermentation (directly from sugars, or after hydrolysis from starch).
Ethers: R-O-R'. Alcohol derivatives (obtained from alcohols and isobutene, and oil derivative;
in the lab, mixing alcohol and sulfuric acid); used as solvents.
Esters: R-COO-R'. From acid+alcoholester+water: R-COOH+R'-OHR-COO-R'+H2O.
Volatile fragrant fats (also found on fruits and flowers).
Aldehydes: R-CHO. Dehydrogenated alcohols. They have the divalent carbonyl group =CO.
Used in the manufacture of chemicals.
Cetones (or ketones): R-CO-R'. Volatile flammable pungent liquids, miscible with water, used
in the manufacture of chemicals and as solvents.

Nitrogenated compounds may be amines (primary amines R-NH2, secondary R-NH-R'...), amides (RCO-NH2), imines (R- (R'-) C=NH), imides ((R-CO)2-N-R'), azides (RN3), cyanates (ROCN), nitriles
(R-CN), nitro-compounds (R-NO2), etc.

FUEL TYPES
BY PHYSICAL STATE
Solid
As coal (mineral), charcoal (from wood) and biomass (wood, dung), but also waxes, metals and nonmetals (e.g. sulfur ignites easily, producing a pungent blue flame; aluminium particles are used in the
rocket boosters for heavy-lift launchers such as the Space Shuttle and Ariane 5).
Liquid
As crude-oil derivatives (gasoline, diesel, fueloil), alcohols, ethers, esters, but also LPG at low
temperatures.
Notice that the usual U.S., Canadian, and New Zealand word for gasoline is simply gas, and that the
usual British word is petrol.
Gas
As natural gas, oil derivatives (LPG), acetylene, manufacture gas (from coal or oil residue) and biogas
(from manure or sewage).

BY PERIOD OF NATURAL RENOVATION

Fossil fuels
Fossil fuels (coal, crude-oil and natural gas) were formed slowly (during millions of years, mainly at
certain remote epochs, not uniformly; e.g. American oil was formed some 90 million years ago,
whereas the rest dates from 150 million years) by high-pressure-decomposition of trapped vegetable
matter during extreme global warming. Fossil fuels are found trapped in Earths crust, up to 10 km
depth, although large pressure might stabilise them also at higher depths and temperatures (at 300 km
it might be 10 GPa and 1000 C). They are then non-renewable at humankind periods, and will
eventually be commercially depleted. Notice that 'resources' refers to the total amount in Nature,
whereas reserves refers to that portion of resources that can be economically recovered at today's
selling prices, using today's technologies and under today's legislation.
Table 2. Estimated reserves and availability of fossil fuels (oil-discovery peaked in 1960s and
oil-production is expected to peak around 2007; gas, some 20 years latter in both cases).
Commercial reserve-2000 Reserve/Consumption-2000
Coal
250 yr
10001012 kg
12
Crude oil
40 yr
10010 kg
12
Natural gas
70 yr
15010 kg.
Notice, by the way, that nuclear fuel reserves are also short-term: the 2109 kg present commercial
reserves would last some 50 years, if only 0.5% of natural uranium is profitable used as in present
nuclear plants (0.71% U-235 in natural uranium, times some 70% of burning depth of the enriched
stuff, typically from 3% down to 2% U-235 in nuclear fuel rods). Two differences, however, apply to
commercial reserves of nuclear fuel: first, that new breeder reactors can burn not only the U-235

fraction but the main U-238 fraction and other fissionable ores, and second, that nuclear ores have
been only marginally prospected and price increments of several orders of magnitude are tolerable due
to the small share of fuel price in nuclear power.
Renewable fuels
Renewable fuels (biomass) are formed in a year or a few years basis (synthetic fuels may come from
fossil or from renewable sources):
Gaseous: biogas from anaerobic fermentation or gasogen gas from pyrolysis of biomass.
Liquid: alcohols, ethers (biopetrol), esters (biodiesel).
Solid: wood, charcoal, fuel pellets (from wood or vegetable residues), agriculture residues,
cattle manure, urban waste.
In comparison with fossil fuels, particularly with oil and gas, renewable fuels are more disperse, have
less energy content, more moisture and ash content, and require more handling effort (but they are
renewable).

BY PRODUCTION STAGE (RESOURCES TYPE)

Natural or primary fuels
Any commodity can be artificially produced, but it may cost a lot; humankind progress has always
been based on finding raw materials that with no cost or little cost could satisfy their needs. The need
for energy, to make machines work, to transport people and goods, and so on, has been met in the past
and in the present by primary fuels (biofuels in the past and fossil fuels during the last two centuries).
Fossil fuels: coal, crude oil (not used unprocessed), natural gas and biomass. They are obtained by
mining (coal) or welling (oil and gas). Some pumping is usually needed. Actually, crude oil is
never used as a primary fuel because there is no economy (residual crude-oil products are cheaper)
and because it is difficult to handle (being a mixture of very light and very heavy substances, its
handling causes cavitation, vapour traps and sticky clogs).
Biofuels (from biomass). They can be directly taken from nature (e.g. wood and fuel crops), or
from human activity waste (agriculture residues, industrial residues, animal residues, or domestic
waste).
Artificial or secondary fuels
Distillates from natural fuels (fossil or biomass, but without chemical reaction): all petroleum
derivatives, plus alcohols used as additives and mixtures, may be obtained by distilling the raw
material. However, modern oil-refinery products really come from a combination of physical
methods (distillation) and chemical methods (reforming and cracking).
Reformed from natural fuels (fossil or biomass, by chemical reactions with heat, steam or partial
air). They are also called synthetic fuels, and may be gaseous, liquid, or solid: hydrogen, acetylene,
synthetic gasoline, synthetic oils, synthetic gases (syngas), charcoal, and coke. Synthetic liquid
fuels are most promising because of their high energy density; first results date from 1897 when
formaldehyde (HCHO, Tb=98 C) was obtained by electrical discharge on a CO/H2 mixture
(syngas), but the main milestone is the Fischer-Tropsch process of 1923, where, desulfurated
syngas (generated by passing water vapour over hot coal), was made to react in presence of Fe, Ru
and Co catalysts, to yield a liquid hydrocarbon blend (containing from methane to heavy waxes)
from which gasoline-like and diesel-like fuels were obtained (e.g. 24H2+12CO=C12H24+12H2O);
most of South Africa's diesel fuel is currently produced this way (they also use a 50/50 Jet A-1
substitute, half from oil, half from coal). Syngas preparation is the most expensive stage in the
process due to materials handling: purification of input coal, removal of sulfur, nitrogen, and ash.

Methanol synthesis is another important process, as a final fuel (CO+2H2=CH3OH), or as an

intermediate step to gasoline and diesel (e.g. through dehydration to dimethylether
2CH3OH=CH3OCH3+H2O).
Exotic fuels (not obtained from fossil or biomass fuels): hydrogen from water electrolysis, and
non-hydrogen non-carbon fuels, like metals (Si, Al, Mg, Fe), used as intermediate energy stores,
since they are first produced from their oxides (with cheap natural energy) and afterwards burnt to
form their oxides (providing valuable artificial energy).

BY MARKETING
Non-commercial
Some biomass materials (municipal solid waste (MSW), local industrial wastes, dung), rocket
propellants (e.g. black powder, hydrazine: N2H4), metals (Fe, Al, Mg, Si), NH3, etc., are considered
non-commercial or special fuels. They may be directly used as fuels (burnt with air), as
monopropellants (e.g. 2N2H4=2NH3+N2+H2), or as intermediate products; e.g. 2Al+(3/2)O2=Al2O3 is
used for aluminothermy, whereas Si+2H2O=SiO2+2H2 is used to produce hydrogen fuel
(Al+3HCl=AlCl3+(3/2)H2 is only used in the lab). Ammonia is not used as a fuel because of its high
auto-ignition temperature (925 K) and the difficulty of complete burning (there is NH3 in the exhaust),
but it might have a future as a hydrogen carrier since liquid NH3 at 20 C (p>854 kPa), with
=610 kg/m3, might yield 109 kg/m3 of H2 whereas the most advanced hydrogen storage systems
(metal hydrides) only store 25 kg/m3 of H2.
Commercial
Coal. It was the main fuel for the Industrial Revolution in the XIX c., and still traded for domestic
use during the first half of the XX c., but today it is only traded to power stations and heavy
industries; mainly bituminous coal, but also lignite.
Crude-oil derivatives (see Table 3). They were developed in parallel with the corresponding
internal combustion engine during the XX c. Crude oil is not used directly as a fuel, as said before
(it only burns in uncontrolled fires). Most crude-oil is now traded in relation to the spot price of
certain market crudes, such as West Texas Intermediate, North Sea Brent, Dubai, or Alaskan North
Slope. The spot price is the price of an individual cargo of crude traded at a particular location.
Natural gas. It seems to be the main fuel for the immediate future. Old town gas, manufactured
from coal or crude-oil, is no longer on the market.
Special commercial fuels, like acetylene (used for cutting and welding).
Petroleum fuels
More than 50% of world's primary energy comes nowadays from petroleum, i.e. all vehicle fuels, and
small and medium stationary applications fuels are petroleum derivatives, obtained by fractional
distillation and reforming. Main commercial fuels and their physical data (that change according to the
market requirements) are presented in Table 3.
Table 3. Main commercial fuels derivatives from crude-oil, and their main averaged properties.
Boiling Boiling Carbon
Density
Viscosity
Flash
Main use
range
range
chain (liquid at 15 C)
at 40 C
point
Tb [K] Tb [C] range
[kg/m3]
106 [m2/s] Tflash [C]
Liquefied
<300
<30
1..4
580
0.5
100
domestic
petroleum
heating, cars
gases (LPG)
Gasoline
300..500 30..200 4..12
730..760
0.5
30
cars

Kerosene
Diesel

450..650 150..350 10..14

500..600 200..300 10..20

780..850
820..880

3
3

40
40

Fuel oil
distillate
Fuel oil
residue

600..800 300..500 15..30

840..930

10

60

930.1010

500

100

>800

>500

20..40

aircrafts
cars, lorries,
boats, heaters
industry, ships
industry, ships.
Must be heated

BY APPLICATION
For spark ignition engines
Spark ignition (SI) internal combustion engine (ICE) fuels (Otto fuels) require knock retardation
(oxygenated compounds increase the octane number (RON=research-octane-number) and decrease CO
and HC emissions). The reference fuel is gasoline (also named petrol, shortened to 'gas' in the USA),
but some other alternative fuels exist.
Gasoline (Eurograde-95 in UE, Premium-95 in USA).
Propane (or better LPG), at its vapour pressure (around 1 MPa). It has good octane number
(of order 100, increasing with propane content up to RON=112), and yields less emissions
than gasoline (less than a half, particularly at small loads). It is fed liquid to direct injection
engines, or gaseous for carburated engines. Most propane-fuelled-engines can work
indistinctly with propane or gasoline (dual-fuel engines). LPG is a mixture of mainly
propane, propylene, butane and butylenes, with composition varying widely from nearly
100% propane in cold countries, to only 20..30% propane in hot countries (e.g. 100% in
UK, 50% in the Netherlands, 35% in France, 30% in Spain, 20% in Greece).
Natural gas (mostly as a compressed gas, CNG, at 30 MPa, but sometimes in as a
cryogenic liquid, LNG, at some 500 kPa). Little used in cars because of the storage, but
more and more used in slave-fleet buses, power plants and cogeneration plants because it is
clean at entrance and exhaust, and it has RON=130, what allows for a compression rate of
12 instead of 9, increasing the efficiency up to 40% (the engine must be specially tuned).
Ethanol and bioethanol. Usually not pure (E100-fuel, i.e. pure ethanol, is used in Brazil)
but added up to 20% to gasoline (E20-fuel or gasohol) to avoid engine changes, or nearly
pure for new 'versatile fuel vehicles' (E80-fuel only has 20% gasoline, mainly as a
denaturaliser). Anhydrous ethanol (<0.6% water) is required for gasoline mixtures, whereas
for use-alone up to 10% water can be accepted. Bioethanol is preferentially made from
cellulosic biomass materials instead of from more expensive traditional feedstock (starch
crops). RON=108.
Methanol is rarely used. Methanol is produced by steam reforming of natural gas to create
a synthesis gas, which is then fed into a reactor vessel in the presence of a catalyst to
produce methanol and water vapour. RON=107.
Ethers. Usually not pure but added up to 20% to gasoline. They are better additives than
alcohols because they are not so volatile, not so corrosive, and less avid for water. MTBE
(methanol tertiary butyl ether, (CH3)3-CO-CH3 or C5H12O), is a petroleum derivate (65%
isobutene, 35% methanol), regularly added in a 10% to gasoline in EU (but there are
concerns about its cancerous properties); its solubility is only 4.3%vol for fuel-in-water and
1.4%vol for water-in-fuel. A better alternative is ETBE (ethanol tertiary butyl ether,
C6H14O), obtained by catalytic reaction of isobutene with bioethanol (with only 55%
isobutene, that comes from petroleum): CH3CH2OH+(CH3CH)2=(CH3)3-CO-CH2CH3.
ETBE has lower volatility, lower water-solubility (2.3 kg/m3 at 20 C), and higher octane
number (RON=116). ETBE forms an azeotrope with ethanol (63% mol ETBE).

Hydrogen is only used in research prototypes as a possible intermediate to future integrated

hydrogen systems. RON=130.

For compression ignition engines

Compression ignition (CI) internal combustion engine (ICE) fuels (Diesel fuels) require very high
injection pressure and low autoignition temperature and delay (as for cetane, C16H34, n-hexadecane,
that is why it is measured as 'cetane number', defined as a cetane / methyl-napthalene mixture which
has the same ignition delay-time as the test fuel). The reference fuel is named diesel (formerly gasoil),
but some other alternative fuels exist.
Diesel oil (gasoil) from crude-oil distillation.
Fuel oil (heavy fuel or residual oil) is only used in large marine engines.
Natural vegetable oils (sunflower, colza, soybean) are usually not directly used because of
their high viscosity (10..20 times that of diesel) and glycerine waste.
Biodiesel (a mono-alkyl-ester mixture, obtained by transesterification of natural oils) can
directly substitute diesel oil in CI engines (a mixture of 30% biodiesel and 70% fossil
diesel is on the market), decreasing pollutant emissions. Biodiesel is renewable, non-toxic
and biodegradable.
Natural gas can be used as main fuel in a diesel engine if a small amount of diesel fuel is
used for compression ignition (dual fuel engine). The gas is usually added by direct
injection in the cylinders before ignition (to avoid high injection pressures). Hybrid
solutions in which a small amount of natural gas injected in a small pre-chamber and
ignited in a spark plug, provide the high temperatures needed for subsequent autoignition
of the main natural gas loading, are being developed. The same approach may be used to
burn other high-autoignition-temperature fuels (e.g. hydrogen) in autoignition mode (diesel
mode).
Notice that a mixture of gasoline and kerosene makes no good for either the SI-engine (less octanenumber) or the CI-engine (less lubrication). In case of accidental mixing, the best is to discard the
mixture. The use of diesel/bioethanol mixtures in CI-engines is being investigated; up to 10% in
volume of anhydrous ethanol (E10-diesel) may be burnt on unmodified CI-engines, significantly
reducing particulate-matter emissions (at the expense of some reduction in cetane-number, ignitionspeed and lubrication).
For gas turbine engines
As they are internal combustion engines (ICE), they cannot work with solid or heavy fuels; besides,
although gas turbine engines are in principle more tolerant on fuel type than gasoline and diesel
engines, high-performance gas turbine demand high-performance fuels.
Kerosene is used for mobile applications. It has high heating value and low flash point. Aircraft
jet engines currently use Jet A-1 fuel, basically 99% kerosene with 1% additives to enhance
cold operations and thermal stability. The first jet engine, however, that flew in the Heinkel He
178 on 27 August 1939, used gasoline as fuel.
Natural gas is used for stationary applications.
For boilers
For external combustion engines (vapour turbines) and very large heaters.
Pulverised coal
Fluidised-bed coal
Fuel oil
For small heaters (like domestic water heaters for space heating or for hot water)

Natural gas (if a network is available)

LPG
Diesel oil

For small portable applications

Fuel, which are odourless, smokeless and non-hazardous are required:
For lighters (portable fire sources)
Matches (see Pyrotechnics)
GLP (see SI-engines). The butane gas lighter dates back to 1933 and it is the most used
today. Gas lighters are refilled by first letting the remaining gas escape (enhanced by
warming up), and then connecting it to an upside-down GLP reservoir (enhanced by
heating and shaking the reservoir and cooling the lighter). Keep gas lighters below 50 C.
Gasoline (see SI-engines)
Waxed-wick fire-lighters
For illumination
GLP (see SI-engines)
Kerosene (see SI-engines)
Waxes (candles)
Acetylene (C2H2, ethyne). It is used in chemical synthesis (80%) and welding (20%), and it
is produced from CaC2 and water, or by hydrocarbon cracking, or by partial oxidation of
natural gas. CaC2 was first obtained in 1892 while searching for aluminium synthesis, by
passing an electric arc through calcinated limestone (CaO) and coal tar. Acetylene can be
produced in situ, but it is usually traded in high-pressure bottles, in liquid form, dissolved in
acetone (and stabilised within a solid porous material) since pure acetylene may decompose
explosively if p>205 kPa.
For heating (cooking, plumbing, cutting)
GLP (see SI-engines, above). Most used at movable cooking ranges and stoves.
Barbecue pellets or briquettes (charcoal, pressed sawdust). BBQ-hints: (BBQ=bar-byqueue): place the charcoal in the BBQ-pan and some firelighters in the middle cavities;
light it and left it for half-an-hour until the coal is ready to cook (it gets to embers covered
by a fine grey ash). After cooking, the fire may be extinguished by air sophocation, and the
unburnt fuel left in place ready for a new fire, or get rid of by soaking with water.
Acetylene (see Illumination, above). The oxy-acetylene torch is the common tool for
manual cutting and welding, because of its high heating power and high combustion
temperature, 3500 K, the maximum of any fuel. A typical workshop bottle of 40 litres, at
1.5 MPa, yields some 6 m3 of acetylene at room conditions (the flow rate should not be
higher than 1 m3/h to avoid acetone carry-over).
For fuel cells
Hydrogen is the nominal fuel for low temperature fuel cells, but the storage problem is not yet
solved satisfactorily. Since 2004, there are city buses powered by fuel cells operating in several
cities (three, with two 75 kW PEM fuel cell stacks each, in Madrid, costing some 1 200 000
each, instead of the 200 000 for a normal bus, roughly).
Reformed hydrogen from other (fossil or natural) fuels (e.g. natural gas, methanol, etc.).
For pyrotechnics, propellants and explosives
(See Pyrotechnics.pdf)

FUEL HISTORY
Humans must have mastered fire some 500 000 years ago (from the time of Homo Heidelbergensis).
For instance, excavations at Torralba (Spain) suggest fire-hunting for elephants, wild cattle, horses,
deer and woolly rhinoceroses 400 000 years ago, and the same in Anatolia 200 000 years ago. Of
course, wild fires must date from the beginning of terrestrial vegetation (evidence of fire has been
found in coal deposits formed 350 million years ago).

BIOMASS FUELS BEFORE THE XX C.

Most fuels used nowadays are fossil (remnant of plants that existed in the distant past), but the first
fuels used by humans were from biomass (living matter used as a source of energy, notably firewood).
Primitive humans must have used fire for lighting, heating, cooking, fighting, communication, religion,
etc. But only non-premixed flames with condensed fuels were used up to 1850, where Bunsen
mastered the premixed combustion with gaseous fuels in his famous burner.
Elementary carbon is not abundant on Earth (<0.1%wt in the crust), but it is the basic constituent of
life and it is found unoxidised on all living and fossil matter. The term biomass (and biological fuels)
is usually restricted to living matter (not fossilised). Since 500 000 years ago, humankind used natural
carbon-compounds as their main fuel. Would it continue like that in future?, or some new kind of fuels
like artificial C-compounds, H-compounds or Si-compounds will take over?
The first biofuels used were: firewood, charcoal, animal fat, animal dung, and vegetable oils. Charcoal,
i.e., partially burnt wood to yield a more energetic and biological-inert fuel, was used as a paint
pigment in the Palaeolithic, and as a strong fuel in metallurgy, from its earliest developments in 5000
B.C., to its take-over by coke in late XVIII c. The simple fire place was an earthy hearth surrounded by
stones, later a large flat stone and since the XIX c. a flat metal plate.
Fire was a sacred element to most ancient cultures, up to Classical Greece (myth of Prometheus), and
perpetual fires were maintained in front of principal temples. They used to light the flame by the suns
rays captured at the centre of a recipient called a skaphia (the ancestor of the parabolic mirror used
today for lighting the Olympic flame).
The fireplace and the chimney
Hearths were used for lighting, heating and cooking. The first human-controlled fires might have been
done out in the open where smoke presented no problem, but soon the benefit of having the fireplace
protected by a shelter wall was discovered and the hearth enter the cave, with the smoke burden,
perhaps the first severe anthropogenic environment pollutant (metabolic waste is not so asphyxiant).
The process of smoking meats and fish might have been naturally discovered that way. Perhaps the
next step was to have fireplaces in the open but surrounded by large stones that protect against gusts
(and radiate heat), but when the roof was added to protect from the rain, the problem of smoke
pollution and fire safety was aggravated. An opening in the roof was left to get rid of smoke and,
already in recent centuries, the importance of the wall surrounding the fireplace was appreciated to the
point of making a dedicated smoke tunnel: the chimney.
Chimney fireplaces have been in use at every home for heating and cooking until the middle of the XX
c., and, although no longer needed because of oil and gas heaters and cooking ranges, they are still
appreciated for their aesthetic and natural flavour. It must be realised, however, how embarrassing a
fireplace in a modern house may be; an enclosed wood stove with glass doors on the fireplace, may be
the best solution to heat with wood today; and remember that burning wet or green wood generates

creosote vapours that may inflamate in the chimney (and never burn plastics, painted wood or glossy
paper that generate toxic gases).
A chimney is a very effective ventilation set-up (much more than a whole in a wall), being able to
remove more heat from already warm walls by air convection than the chemical heat supplied by the
fuel (that is why walls extended around fireplaces to have comfort within, and not just the front side
roasted and a frozen back). For the same reason, it is difficult to approach, say, 900 K, and enclosed
chambers (furnaces) are required to reach higher temperatures. A chimney works on the Archimedes'
principle: generating a pressure-imbalance draught, p, on a column of hot air of height L: p=gL.
Chimneys get coated with soot and, if not cleaned regularly, may catch fire. A chimney cap is always
used to keep out the rain, leaves, and bird droppings, to inhibit downdraughts and as spark arrestor.
The chimney is a rather recent invention; the first one dated comes from an earthquake in Venice in
1347. B. Franklin in 1745 wrote some guidelines to avoid smoky chimneys, suggesting to narrow the
air entrance by covering most of the front with a hanging cloth (or better a plaster wall), and to widen
the flue, to keep smoke from coming out into the room. B. Thompson (Count of Rumford) in 1796
suggested to make the fireplace shallower, with widely angled side-walls, so they would radiate better,
and with a streamlined flue throat with a sudden expansion to avoid smoke rebuff (it was shown by
P.O. Rosin in 1932 that the sudden expansion is counterproductive). Modern fireplaces have a pre-cast
iron box with a pre-cast streamlined chimney throat, which is covered with heavy refractory masonry.
Torches, oil lamps and candles
Special devices were invented to transport fire (portable burning appliances), mainly for lighting. The
first portable fire was the torch, a burning branch plucked from a fire, later enhanced to a reed or tow
soaked in molten fat or oil, or naturally impregnated with resin or pitch. Torches date back at least to
50 000 B.C.
Much later on, animal fat in a bowl (sea-shell or skull) with a grass wick, a kind of torch with liquid
fuel pumped up by capillarity, must have been developed; the first remains were found at Lascaux
famous painting cave, dating from around 20 000 B.C. Oil lamps were common in Egypt in 3000 B.C.,
made of clay and burning seed oil with a cotton wick. Greek lamps from 500 B.C. looked like saucers
(later on with a groove) and burned olive oil in a pottery or bronze container (a modern oil lamp made
of brass is shown in Fig. 1). Candles, although known before, only were in widespread since 400 B.C.
The next light source, the carbon arc lamp, was not developed until the early 1800s (by Sir Humphry
Davy), although it was not in widespread use until late in that century, when the incandescent carbonfilament lamp was also developed (by Thomas Edison in the 1880s). Although the carbon arc lamp is
assisted by the burning of graphite rods, electric light marked the decline of fuel lamps, which
practically disappear from 1900 onwards, taking over first the traditional tungsten-filament
incandescent bulb (developed in the 1920s, and being retired in the 2010s because of its low
efficiency), and later by the fluorescent gas-discharge lamp (in widespread use since the 1940s,
although introduced around 1900).

Fig. 1. A smoky kerosene oil-lamp.

The candle is as a kind of solid-oil lamp, tuned to feedback heat, by radiation to melt the solid fuel,
made of frozen oil, semisolid fat, or wax, enhancing safety and handling (a no-spill lamp). The solid
fuel melted at some 55 C to 65 C by radiated heat, migrating up the wick by capillary action, where it
is pyrolysed at some 900 K (dark zone in the flame), generating gases that diffuse through the inner
cold bulb and reach the thin combustion zone, blue at the well-ventilated bottom, yellow at the richer
upper part, where products, including soot particles, are formed at some 1500 K (the hottest zone, up
to 1700 K, is found off the centre on the edge of the brighter yellow portion), by reaction with outside
air, whose flow is smoothly coupled to burning rate through buoyancy convection (notice that if air is
artificially supplied, the candle flame flickers a lot and may be extinct, and that, in absence of a force
field, as under microgravity, the candle slowly burns with a very precarious spherical blue flame).
The candle was known in Egypt (clay candle holders are found dating 3000 B.C.), but candles were
not used because of the higher cost, dimmer light and higher pollution than oil lamps. Candles were
made by coating a wick (the fuel pump) with pitch or animal fat. Chinese, Japanese and preColombian Americans also used candles. Bee-wax candles were used by wealthy Romans as a rarity:
they produce a bright flame, do not smoke, and produce a fragrant odour while burning. Candle usage
widespread during the Middle Ages to light monasteries and churches; ordinary candles were made of
animal fat (tallow), usually from sheep or cattle, had a dark yellowish colour and gave off a nasty
smell. From the XVI c. onwards, living standards improved as evidenced by the increasing availability
of candlesticks and candleholders and their appearance in households. Spermaceti candles were made
of oil present in the head cavities of sperm whales; they burned with so a bright light that a spermaceti
candle flame was used as a standard light measure for photometry. Since the early XIX c., candles are
made of crude-oil bleached paraffins, or better of stearin, after the chemist M.E. Chevreul discovered
that tallow was a mixture of two fatty acids, stearic acid and oleic acid, combined with glycerine; by
removing glycerine, he achieved a harder, brighter and longer lasting wax. Candle making may be by a
process of sequential dipping, or by moulding, or rolling, or extrusion, or casting.
Improvements on the wick were also achieved in mid XIX c. by weaving the cotton fibres flat and
treating them with a mordant (e.g. boric acid), what causes the wick tip to stand, curl and burn, instead
of just twisting the cotton fibres, what causes the fluffy wick tip to bend within the cold zone of the
flame and fall into the molten wax, causing a mesh if not snuffing (cutting the charred part of the
wick). If the wick would not curl and lean out of the flame and burn, it would suck too much fuel, form

too much soot, and smoke. The white smoke seen when a candle is blown out, is the condensation of
vaporised hydrocarbons not yet pyrolysed (it can be easily lighted).
A major development in the oil lamp took place in the 1780's when Argand, a Swiss chemist,
introduced the hollow-cylinder wick, that allowed air to reach the centre of the flame, yielding a
brighter flame; his assistant Quinquet added the glass chimney that bears his name (Da Vinci also
sketched it). Whale oil and colza oil were most used at this time, until kerosene came into scene in mid
XIX c. and finally Thomas Edison invented the incandescent light in 1879, pushing combustion
illumination to the corner, but not until Coolidge's ductile wolfram replaced Edison's wretched carbon
filaments (incandescent lamps have been improved by filling with gas and most recently by using a
halogen gas to recycle vaporised wolfram; the fluorescent lamp was invented in 1938). Olive oil lamps
made in brass have been in use in rural Spain up to mid 20th century.
Engines on biological fuels
Steam engines, the Otto engine and Diesels engine, all started their development burning biomass
fuel: wood in the boilers, alcohol in Ottos engine, and vegetable oil in Diesels, but wood has a small
heating value and was been exhausted, and liquid biofuels were too much expensive. Up to the modelT Ford, cars were designed to burn the coarse petrol distillate of the time, or alcohol (distilled from
fermented sugars), or any mixture of them. But during the XX c. the petroleum industry swept all the
fuel markets.

MINERAL FUELS
Mineral fuels were later discovered and only in the last three centuries massively used (up to
depletion!). They are basically coal, petroleum, and natural gas. They have several advantages over
biomass: fossil fuels are more concentrated, have higher energy density, lower moisture content, more
constant composition, and, except for coal, less ash content. Their major drawback is being not
renewable, with two direct consequences:
They are short-term commercially available (they will be exhausted at current trends in one
or two generations).
The increase pollution, and particularly global warming, without short-term natural
recycling, what means that fossil-fuel emissions must be artificially counterbalanced (e.g. by
getting back excess atmospheric CO2, which increases the greenhouse effect, and burying it
in a non-pollutant way).
History of coal
Coal was known since ancient times, but only used when available on site; otherwise charcoal from
wood was used.
Coal was used in China in 1100 BC, in Wales during the Bronze Age, and as a home heating fuel in
England since Roman times and during the Middle Ages, but mining really started in the XVIII c., also
in England, driven by the Industrial Revolution hunger to feed steam boilers. In the mid XIX c. coal
also replaced charcoal in making black-powder and explosives. A great amount of coal is also used in
the iron industry, its second consumer.
Around year 1880, primary world energy came from traditional biomass and coal 50%/50%, and up to
mid 20th c. the tonne-coal-equivalent (1 tce=30 GJ) unit was used, later substituted by the tonne-oilequivalent unit (1 toe=42 GJ).
History of natural condensed hydrocarbons

Natural condensed hydrocarbons (solid, sticky or liquid), generically called bitumen (Latin bitumen
sticky) were found in Mesopotamia 5 000 years ago, and used for lighting, heating, gluing bricks
together, sealing boats (Noah's Ark) and paving streets. Egyptians coated mummies and sealed
pyramids with pitch. Two kind of bitumen were found: crude oil (or petroleum, Latin petroleum stone
oil), that flows, and asphalt (Greek binder) that does not flow.
Crude oil has always aroused interest as a substitute of vegetal and animal oils. In 1771 George
Washington (the 1st USA President) acquired a piece of ground in what is now West Virginia because
it contained an oil-and-gas spring. As demand for kerosene for illumination grew, oil well drilling
started. The first success was in Titusville, Pennsylvania, where E. Drake found oil in 1859 in a 23-m
deep well. Drake refined the crude oil into kerosene for lighting, and gasoline and other products made
during refining were simply thrown away because people had no use for them. John D. Rockefeller
enters the oil refining business in 1870 forming the Standard Oil Company (Cleveland, Ohio) to
produce kerosene, for lighting purposes.
Crude-oil today is usually measured in 'toe' (metric ton oil equivalent), but in USA, the basic unit
remains the 'barrel' (of 42 gallons of 231 cubic inches; there are many other barrel and gallon sizes),
though there are no more old-fashioned wooden crude-oil-barrel since Drake's time except in museums
(they run out of barrels, and time to fill them, just after a few years). Crude oil is always transported in
bulk by tankers, pipelines, barges, and trucks. It was in 1866, seven years after Drake drilled his well,
that Pennsylvania producers confirmed the 42-gallon barrel as their standard volume unit (not a real
container), as opposed to, say, the 31.5 gallon wine barrel, or the 32 gallon London ale barrel, or the 36
gallon London beer barrel.
Today only fluid crude-oil is put in the market, and all of it is distilled, asphalt being the residue (more
than 80% of asphalt production is used for road paving, handled at >100 C to overcome yield stress,
or lately as a water emulsion with 30%w asphalt particles about 1 m size). The generic name for
natural condensed hydrocarbons is bitumen, and, according to the fluidity, it may be crude oil
(petroleum, Tflash<65 C, Tpour<0 C) or asphalt (Tflash>65 C, Tpour>0 C). They are non-crystalline
highly viscous, black or dark brown pastes, adhesive and waterproof (boat, roof and road paving), with
>80%wt C and >15%wt H.
History of natural gas
Humans must have found many natural-gas sources seeping from the ground (by their hissing or
bubbling, because they usually have no smell; odorants must be artificially added for safety alarm), but
the ones they could not miss were the burning springs (ignited by a lightning) known as 'eternal fires'
referred to in most ancient traditions (India, Persia, Greece; the most famous might be one on Mount
Parnassus around 1000 B.C.).
This first well intentionally drilled to obtain natural gas was drilled in 1821 in Fredonia, New York, by
William A. Hart: a 9 m deep well to enhance a surface seepage of natural gas. In the XIX c. natural gas
was used locally in many USA states. Coal seams store CH4 inside micropores, but get mixed with air
in mining, releasing a 50/50 methane/air mix, unsuitable to gasoducts. It is only after 1950 that big gas
pipelines (gasoducts) were built (in USA and Russia). Natural gas for transmission companies must
generally meet certain pipeline quality specifications with respect to water content, hydrocarbon dewpoint, heating value, and hydrogen-sulphide content.

BIOLOGICAL FUELS IN THE XXI C.

Use of biomass fuels has never been abandoned, but their importance during the Industrial Revolution
fade away. A good example of this biomass back up in those days is that when Rudolph Diesel
developed in 1893 his compression-ignition engine, he intended to burn coal powder (fossil), but
problems with that choice (later solved with oil fuels), dictated that he demonstrated his engine at the
World Exhibition in Paris in 1900 using peanut oil as fuel (biomass).
Late in the XX c., fossil fuel depletion indicators, and local and global pollution associated to fossil
fuels, have being pressing to come back from 'black fossil power' to 'green renewable power', and, as
fuels continue to be the best solution for energy storage, specially for transport applications, biofuels
are at the stage again.
The terms: biofuels, biomass fuels and renewable fuels, may be used indistinctly if they refer to natural
or artificial fuels obtained from renewable sources, although other times distinctions are introduced
and then biofuels may refer to biomass derivatives directly substituting fossil fuels for the same
combustor, biomass may be restricted to unprocessed biomass (forest waste, crops and agriculture
waste, animal waste, domestic waste), and renewable may include fuels like hydrogen obtained by
electrolysis and not from biomass.
Biomass fuels have always presented several problems that might have been forgotten during the two
centuries when we have profited from massive fossil fuel sources:
Biomass-fuel sources are not found concentrated in Nature (contrary to fossil-fuel fields), and
there is an inherent inefficiency in collecting them (e.g. forest wood waste, urban solid waste),
although sometimes it is compensated by the need to get rid of that matter (e.g. to prevent forest
fires, to improve sanitation).
Biomass fuels are mostly solid, and some pre-processing is needed (gasification, liquefaction)
to produce fluid fuels, the kind of fuel best fitted to both engines and stationary combustors.
Biomass fuels are less energetic than fossil fuels, because living matter is roughly a water
suspension of oxygenated hydrocarbons, and fossil fuels were slowly 'cooked' over the aeons to
separate water and most of the oxygen. Besides, some biomass fuels have non-fuel components
that must be separated (e.g. soil in forest-waste, metals in industrial waste).
Biomass fuels are also contaminant, not contributing to global warming (because the CO2
produced compensates with that synthesised from the air during the biomass growth), but
producing much more particulates and new chemical emissions (e.g. dioxins) if not properly
treated.
But the move to biofuels is not based on their short-term advantage over fossil fuels but on the longterm need to have fuels of any kind. And for the time being, living matter and their residues are a
handy alternative to dying fossil fuels.
Some biofuel production methods considered are:
Ethanol by fermentation of biomass sugars, starch or cellulose by yeast or bacteria. In Japan, a
bacteria has been bred which produces ethanol from paper or rice-straw without any pretreatment.
Methane (actually a biogas mixture) by anaerobic digestion of biomass waste (manure, straw,
sewage, municipal solid waste (MSW)).
Oils (biodiesel) by reforming oleaginoseous plant seeds (e.g. colza, sunflower, soya). The
marine microscopic algae Botryococcus Braunii has been shown to accumulate a quantity of
hydrocarbons amounting to 75% of their dry weight

Methanol from wood-waste distillation.

Hydrogen by reforming other biofuels (e.g. ethanol or methane), or from water electrolysis by
solar or wind energy (of course, this sometimes called 'solar fuel' is not bio in the sense that it
has nothing to do with living matter, but what is life: a self-organised system based on water
photolysis?

For mobile applications, because of higher energy density and simpler infrastructure, liquid biofuels
are preferred (ethanol and biodiesel), gaseous biofuels being restricted to stationary applications.
As an aid in transition from fossil fuels to biofuels, mixtures of both are being progressively put on the
market for old engines and combustors, and new engines and combustors are progressively developed
to run on 'biofuel prototypes' derived from current fossil fuels (e.g. producing methanol or hydrogen
from natural gas).
Another possible biofuel in the future may be hydrogen produced biologically, e.g. by a photosynthesis
of hydrogen instead of carbohydrates (there are some algae that do that). Producing hydrogen by other
sustainable means (direct solar energy or wind energy), however, seems closer in the future.
(Back to Combustion)

FUEL PROPERTIES
Fuel properties...........................................................................................................................................1
Crude oil ................................................................................................................................................2
Gasoline ................................................................................................................................................2
Bioethanol and ETBE .......................................................................................................................3
Diesel oil, Kerosene, and Jet fuel ..........................................................................................................4
Jet fuels .............................................................................................................................................5
Biodiesel............................................................................................................................................6
Fueloil ...................................................................................................................................................6
Heavy fueloil .....................................................................................................................................7
Natural gas, biogas, LPG and methane hydrates ..................................................................................7
Hydrogen ...............................................................................................................................................9
Production .........................................................................................................................................9
Storage & transport .........................................................................................................................11
Safety ..............................................................................................................................................12
Purity ...............................................................................................................................................12
Price comparison of hydrogen energy ............................................................................................13
Hint: Hydrogen balloon combustion & explosion ..........................................................................13
Coal .....................................................................................................................................................13
Origin ..............................................................................................................................................13
Types ...............................................................................................................................................13
Uses .................................................................................................................................................14
Composition: proximate analysis and ultimate analysis .................................................................14
Air requirement for theoretical combustion ....................................................................................15
Heating values .................................................................................................................................16
Wood ...................................................................................................................................................16
Composition ....................................................................................................................................16
Biomass ...........................................................................................................................................18
Fuel pyrolysis ......................................................................................................................................19

FUEL PROPERTIES
Fuels, as for any other type of substance, can be assigned some physical and chemical properties (e.g.
density, thermal capacity, vapour pressure, chemical formula, etc. However, most of the times,
combustion properties are also assigned to fuels, in spite of the fact that these properties depend on the
oxidiser (e.g. air, pure oxygen) and the actual process (e.g. the explosion limits depend on the
boundary conditions for a given fuel/oxidiser pair). Fuel price, availability, risk, and so on, could also
be considered fuel properties (attributes).
An introduction on fuels and fuel types, including some relevant properties, can be found apart in
Fuels.pdf. Fuel consumption and Pyrotechnics are also covered separately. A summary table of fuel
properties for normal combustion in air can also be found there. What follows is just a collection of
additional notes, mainly physicochemical data, on particular fuels.

CRUDE OIL
Crude oil is not used directly as a fuel but as a feedstuff for the petrochemical factories to produce
commercial fuels, synthetic rubbers, plastics, and additional chemicals. Oil refineries were originally
placed near the oil fields, in part because natural gas, which could not then be economically
transported long distances, was available to fuel the highly energy-intensive refining process, but since
1950, for strategic reasons crude oil was transported by tankers and oleoducts to local refineries.
Most data are highly variable with crude-oil field; typical ranges are given.
Density. Typically 900 kg/m3 (from 700 kg/m3 to 1000 kg/m3 at 20 C; floats on water). Linear
temperature variation fit. The density of spilled oil will also increase with time, as the more
volatile (and less dense) components are lost, so that, after considerable evaporation, the
density of some crude oils may increase enough for the oils to sink below the water surface.
Freezing and boiling points. When heating at 100 kPa a frozen crude-oil sample (from below 210 K),
solid-liquid equilibrium may exist in the range 210 K to 280 K, and liquid-vapour above 280 K;
vapours start to decompose at about 900 K.
Viscosity=510-6..2010-6 m2/s at 20 C. Exponential temperature variation fit. Pour point= 5..15 C.
Vapour pressure. 5..20 kPa at 20 C (40..80 kPa at 38 C). Vapours are heavier than air (2 to 3 times).
The characteristic time for evaporation of crude-oil spills at sea is 1 day (25% in volume
evaporated).
Composition. Each crude-oil field has a different composition, that can be established by a
combination of gas-chromatography, fluorescence-spectroscopy and infrared-spectroscopy
techniques, and that may be used, for instance, in forensic analysis of oil spills at sea (even
after refining, crude-oil derivatives may be associated to their source field). Saturated
hydrocarbons content is around 60%wt, aromatics 30%wt, resins 5%wt. Sulfur content is
0.5..2%wt. Heavy metals <100 ppm. Crude-oil vapours are mainly short-chain hydrocarbons
(only about 10% in volume have more than 4 carbons).
Flash-point and autoignition temperature: some 230 K and 700 K approximately.
Ignition limits: lower 0.5..1%, upper 7..15%.
Organoleptic: black, brown or dark-green colour, aromatic or sulphide odour.
Solubility. <0.4%wt, due mainly to volatile compounds.
Surface tension: 0.029 N/m with its vapours, 0.023 N/m with water.
Price: for a 100 $/barrel (very variable), with 159 L/barrel it is 0.63 $/L, and with 900 kg/m3 and 42
GJ/toe, it is 16.6 $/GJ.

GASOLINE
Types. In EU: Eurosuper-95, Eurosuper-98 (both lead-free). In the USA: Regular (97 RON) and
Premium (95 RON).
Density=750 kg/m3 (from 720 kg/m3 to 760 kg/m3 at 20 C). Thermal expansion coefficient=90010-6
K-1 (automatic temperature compensation for volume metered fuels is mandatory in some
countries).
Boiling and solidification points. Not well defined because they are mixtures. (e.g. when heating a
previously subcooled sample at constant standard pressure, some 10% in weight of gasoline is
in the vapour state at 300 K, and some 90% when at 440 K).
Viscosity=0.510-6 m2/s at 20 C.
Vapour pressure. 50..90 kPa at 20 C, typically 70 kPa at 20 C.
Heating value. Average Eurosuper values are: HHV=45.7 MJ/kg, LHV=42.9 MJ/kg.
Theoretical air/fuel ratio: A=14.5 kg air by kg fuel.

Octane number (RON)=92..98. This is a measure of autoignition resistance in a spark-ignition engine,

being the volume percentage of iso-octane in a iso-octane / n-heptane mixture having the same
anti-knocking characteristic when tested in a variable-compression-ratio engine.
Cetane number=5..20, meaning that gasoline has a relative large time-lag between injection in hot air
and autoignition, although this is irrelevant in typical gasoline applications (spark ignition).
Composition. Gasoline composition has changed in parallel with SI-engine development. Lead
tetraethyl, Pb(C2H5)4, a colourless oily insoluble liquid, was used as an additive from 1950 to
1995, in some 0.1 grams of lead per litre, to prevent knocking; sulfur was removed at that time
because it inhibited the octane-enhancing effect of the tetraethyl lead. Its typical hydrocarbon
composition is presented in Table 1. Average molar mass is M=0.099 kg/mol, and ultimate
analysis (by weight; see coal analysis below for more details): 87%C and 13%H (corresponds
roughly to C7.2H12.6).
Table 1. Gasoline composition*.
60% saturated (4..8 -C-)
15% lineal (n-)
best combustion, low RON
(is increasing)
30% branched (iso-)
high RON
15% cycle
40% unsaturated (5..9 -C-)
5% olefins (alkenes)
bad smell
(is decreasing)
35% aromatics (benzenes) toxic, yield soot, high RON
<500 ppm Sulfur in 2000
<100 ppm Sulfur for 2005
*A sample showed 21% cycle-hexane, 17% iso-octane, 16% iso-pentane, 16% ethyl-bencene, 15%
toluene, 12% n-decane, 3% naphthalene, and all other <1%.
Solubility in water depends on the actual compounds: hydrocarbons are very insoluble in water, but
alcohols readily mix. Table 2 presents some data.
Table 2. Solubility data at 25 C of some gasoline compounds.
Substance
Solubility of substance in water Water solubility in substance
ethanol (& methanol)
100%wt
100%wt
benzene
0.18%wt
0.06%wt
cyclohexane
0.006%wt
0.01%wt
iso-octane
0.0003%wt
0.006%wt
Price. In Europe in 2013, about 1.5 /L, with variations of 30% amongst countries (in USA some 1
/L). In Europe, in % of retail price, the price structure is roughly: refinery output 20, transport
1, station benefit 6, special fuel tax 60, value added tax 13.
Bioethanol and ETBE
Bioethanol is bio-fuel substitute of gasoline; i.e. it is ethanol obtained from biomass (not from fossil
fuels), and used as a gasoline blend.
Pure bioethanol (E100-fuel) is by far the most produced biofuel, mainly in Brazil and USA. More
widespread practice has been to add up to 20% to gasoline by volume (E20-fuel or gasohol) to avoid
the need of engine modifications. Nearly pure bioethanol is used for new 'versatile fuel vehicles' (E80fuel only has 20% gasoline, mainly as a denaturaliser). Anhydrous ethanol (<0.6% water) is required
for gasoline mixtures, whereas for use-alone up to 10% water can be accepted.

ETBE (ethanol tertiary butyl ether, C6H14O, =760 kg/m3, LHV=36 MJ/kg), is a better ingredient than
bioethanol because it is not so volatile, not so corrosive, and less avid for water. ETBE-15 fuel is a
blend of gasoline with 15% in volume of ETBE. ETBE is obtained by catalytic reaction of bioethanol
with isobutene (45%/55% in weight): CH3CH2OH+(CH3CH)2=(CH3)3-CO-CH2CH3. To note that
isobutene comes from petroleum. The other gasoline-substitute ether, MTBE (methanol tertiary butyl
ether, (CH3)3-CO-CH3), is a full petroleum derivate (65% isobutene, 35% methanol).
Bioethanol is preferentially made from cellulosic biomass materials instead of from more expensive
traditional feedstock such that starch crops (obtaining it from sugar-feedstocks is even more
expensive). In Japan, a bacteria has been bred which produces ethanol from paper or rice-straw
without any pre-treatment. Steps processes in ethanol production are:
Milling (the feedstock passes through hammer mills, which grind it into a fine meal).
Saccharification. The meal is mixed with water and an enzyme (alpha-amylase) and keept to 95
C to reduce bacteria levels and get a pulpy state. The mash is cooled and a secondary enzyme
(gluco-amylase) added to convert the liquefied starch to fermentable sugars (dextrose).
Fermentation. Yeast is added to the mash to ferment the sugars to ethanol and carbon dioxide
(CO2, a byproduct sold to the carbonate-beverage industry). Using a continuous process, the
fermenting mash is allowed to flow, or cascade, through several fermenters until the mash is
fully fermented and then leaves the final tank. In a batch fermentation process, the mash stays
in one fermenter for about 48 hours before the distillation process is started.
Distillation: The fermented mash contains about 10% ethanol, as well as all the nonfermentable
solids from the feedstock and the yeast cells. The mash is pumped to the continuous flow,
multicolumn distillation system where the alcohol is removed from the solids and the water.
The alcohol leaves the top of the final column at 96% strength, and the residue from the base of
the column is further processes into a high protein-content nutrient used for livestock feed.
Dehydration: To get rid of the water in the azeotrope, most ethanol plants use a molecular sieve
to capture the remaining water and get anhydrous ethanol (>99.8%wt pure).
Denaturing: Fuel ethanol is denatured with a small amount (2%-5%) of some product such as
gasoline, to make it unfit for human consumption.

DIESEL, KEROSENE, AND JET FUEL

Diesel fuel is any liquid fuel used in diesel engines, originally obtained from crude-oil distillation
(petrodiesel), but alternatives are increasingly being developed for partial or total substitution of
petrodiesel, such as biodiesel (from vegetal oils), and synthetic diesel (usually from a gas fuel coming
from coal reforming or biomass, also named gas to liquid fuels, GTL). In all cases, diesel nowadays
must be free of sulfur.
Kerosene is a crude-oil distillate similar to petrodiesel but with a wider-fraction distillation (see
Petroleum fuels). Jet fuel is kerosene-based, with special additives (<1%). Rocket propellant RP-1
(also named Refined Petroleum) is a refined jet fuel, free of sulfur and with shorter and branched
carbon-chains more resistant to thermal breakdown; it is used in rocketry usually with liquid oxygen as
the oxidiser (RP1/LOX bipropellant). The tendency to change to biofuels or GTL fuels is also
applicable here. Contrary to its etymology, present-day kerosene and derivatives are less waxy than
diesel (i.e. less lubricant). Diesel and kerosene should not be taken as fully interchangeable fuels at
present, because kerosene has no cetane-number specification and thus it may have large ignition
delays (producing lots of unburnt emissions and engine rough-running by high-pressure peaks);
besides, kerosene has less lubricity, and diesel-fuel less cold-start ability.

Diesel types. In EU: type A for road vehicles, B for industries (agriculture, fishing; same properties as
type A, but red-coloured for different taxation), C for heating (not for engines; blue-coloured).
In USA: No. 1 Distillate (Kerosene), and No. 2 Distillate (Diesel).
Density=830 kg/m3 (780..860 kg/m3 at 40 C). Thermal expansion coefficient=80010-6 K-1. 880 kg/m3
for biodiesel (860..900 kg/m3 at 40 C).
Boiling and freezing points. Not well defined because they are mixtures. In general, these fuels remain
liquid down to 30 C (some antifreeze additives may be added to guarantee that).
Viscosity=310-6 m2/s (2.010-6..4.010-6 m2/s at 40 C) for diesel; 4.010-6..6.010-6 m2/s for biodiesel.
Vapour pressure=1..10 kPa at 38 C for diesel and JP-4, 0.5..5 kPa at 38 C for kerosene.
Cetane number=45 (between 40..55); 60..65 for biodiesel. This is a measure of a fuel's ignition delay;
the time period between the start of injection and start of combustion (ignition) of the fuel, with
larger cetane numbers having lower ignition delays. This is only of interest in compressionignition engines, and only valid for light distillate fuels (because of the test engine; for heavy
fueloil, a different burning-quality index is used, calculated from the fuel density and
viscosity).
Flash-point=50 C typical (40 C minimum). In the range 310..340 K (370..430 K for biodiesel).
Heating value. HHV=47 MJ/kg, LHV=43 MJ/kg (HHV=40 MJ/kg for biodiesel).
Composition. All natural fuels are mixtures (and most synthetic fuels too). The analysis can be
ultimate (i.e. mass fraction of chemical elements), or structural (mass fraction of identified
molecules). The ultimate analysis of desulfurized kerosenes (<0.2% S), by weight, may yield
some 84..86% C, some 13..15% H, and 1% impurities and additives. The structural analysis
shows, by volume, some 66% of saturated hydrocarbons (linear and cycle chains), 30%
aromatics (benzene derivatives), and 4% olefins (unsaturated hydrocarbons). From the ultimate
analysis one may establish a reduced molecular formula (per unit carbon atom) of CH n with
n=1.8..2 (e.g. for 86% C and 14% H, n=(14/1)/(86/12=1.95). If the structural analysis is also
considered, a mean molecular formula can be found (i.e., with whole number of atoms and
typical carbon-chain-length, as C11H21, or C12H23, or C12H26, or C13H26, or C14H30; dodecene
and tridecene are the most usual surrogates). Composition of biodiesel, by weight, may be:
77% C, 12% H, 11% O, 0.01% S.
Price. In Europe in 2013, diesel costs about 1.4 /L, with variations of 20% amongst countries (in
USA some 1 /L). In Europe, in % of retail price, the price structure is roughly: refinery output
20, transport 1, station benefit 6, special fuel tax 60, value added tax 13.
Jet fuels
Jet fuel is used for commercial (Jet A-1, Jet A, and Jet B) and military (JP-4, JP-5, JP-8...) jet
propulsion; aviation gasoline (avgas) is used to power piston-engine aircraft. They are basically
mixtures of kerosene and gasoline (half-&-half for JP-4, 99.5% kerosene for JP-5 and JP-8, 100%
kerosene for Jet A-1), plus special additives (1..2%): corrosion inhibitor, anti-icing, anti-fouling, and
anti-static compounds. Jet A-1 comprises hydrocarbon chains with 9 to 15 carbon atoms. Jet B (also
named JP-4, with composition distribution from 5 to 15 carbon chains), is used in very cold weather,
and in military aircraft.
Jet A-1 is the international standard jet fuel, with a freezing temperature of Tf=50 C (47 C as a
limit); Jet A (with Tf=40 C) is a low-grade Jet A-1 only and mostly used in USA; and Jet B (Tf<50
C), the commercial name of JP-4, is only used in very cold climates. They all have a lower heating
value of 42.8..43.6 MJ/kg. Minimum flash point is 60 C for JP-5, 38 C for Jet A-1 and JP-8 (typical
value for Jet A-1 is Tflash=50 C, with a vapour pressure at this point of 1.5 kPa; 1 kPa at 38 C), and
Tflash=20 C for JP-4. Typical density at 15 C is 810 kg/m3 for Jet A-1, and 760 kg/m3 for Jet B. Jet
fuel must withstand 150 C without fouling (dissolved oxygen in fuel exposed to air reacts with the

hydrocarbons to form peroxides and eventually deposits after few hours); further heating leads to
thermal cracking.
Jet A-1 specification is Tflash=494 C at 100 kPa (but it might decrease to Tflash=15 C at cruise altitude
with 25 kPa). Fuel tank ullage can be inertized with nitrogen-enriched air with xO2<12%. JP-4 has
Tflash=20 C. Jet A-1 surrogate is 1-dodecene (C12H24), whereas Jet B (also named JP-4) surrogate is ndecane C10H22. Jet A-1 viscosity at 20 C is about 8.010-6 m2/s.
Price: Jet A-1 sells at some 0.8 $/L (about 20 /GJ in terms of LHV).
Rocket propellant RP-1 fuel properties may be assumed to be the same as jet fuel properties.
Biodiesel
Biodiesel is a biomass-derived fuel, safer, cleaner, renewable, non-toxic and biodegradable direct
substitute of petroleum diesel in compression-ignition engines, but more expensive. Biodiesel is a
mono-alkyl-ester mixture obtained from natural oils, currently produced by a process called
transesterification, where a new or used oil (sunflower, colza, soybean, or even animal fat) is first
filtered, then pre-processed with alkali to remove free fatty acids, then mixed with an alcohol (usually
methanol) and a catalyst (usually sodium or potassium hydroxide); the oil's triglycerides react to form
esters and glycerol, Fig. 1, which are then separated from each other and purified. Usually 10%
methanol (non-renewable) is added, and some 10% glycerol forms. Colza is also known as rape
(RME=rape methyl ester, and REE=rape ethyl ester). Biodiesel surrogates are longer-chain
hydrocarbons than petrodiesel: C13H28, C14H30, or C15H32.

Fig. 1. Transesterification of vegetable oil to biodiesel (R is typically a 16 to 18 C-atoms hydrocarbon

with 1 to 3 double bounds.

FUELOIL
Types. There are two basic types of fueloil: Distillate fueloil (lighter, thinner, better for cold-start) and
Residual fueloil (heavier, thicker, more powerful, better lubrication). Often, some distillate is
added to residual fueloil to get a desired viscosity. They are only used for industrial and marine
applications because, although fueloil is cheaper than diesel oil, it is more difficult to handle
(must be settled, pre-heated and filtered, and leave a sludge at the bottom of the tanks). Notice
that, sometimes, particularly in the USA, the term 'fuel oil' also includes diesel and kerosene.
Density. Some 900..1010 kg/m3. Varies with composition and temperature.
Viscosity. Widely variable with composition; some 100010-6 m2/s at 20 C (400010-6 m2/s at 10 C,
(10..30)10-6 m2/s at 100 C). Varies a lot with composition and temperature. Must be heated for
handling (it is usually required to have <50010-6 m2/s for pumping and <1510-6 m2/s for
injectors). Pour point in the range 5..10 C. Fueloils are usually graded by their viscosity at 50
C (ISO-8217).
Vapour pressure. 0.1..1 kPa at 20 C.
Composition. Distillate fueloils are similar to diesel oil.

Price. Typically half of crude-oil price.

Heavy fueloil
Heavy fueloil (HFO) is the residue of crude oil distillation that still flows (the quasi-solid residue is
asphalt); waste oil from other industries are often added. It is the fuel used in large marine vessels
because of price (about half the price of distillates). A typical HFO is IF-300 (Intermediate Fuel),
which has a viscosity of 30010-6 m2/s at 50 C (300 cSt), 2510-6 m2/s at 100 C, =990 kg/m3 at 15 C,
HHV=43 MJ/kg, and the flash-point at 60..80 C.
HFO (also named Bunker-C, or Residual fuel) may have a composition of 88%wt C, 10%wt H, 1%wt
S, 0.5%wt H2O, 0.1%wt ash, and may contain dispersed solid or semi-solid particles (asphaltenes,
minerals and other leftovers from the oil source, metallic particles from the refinery equipment, and
some dumped chemical wastes), plus some 0.5% water. HFO leaves a carbonaceous residue in the
tanks, and may have up to 5% of sulfur (MARPOL directive is to limit it to 3.5% by 2012 and to 0.5%
by 2020).

NATURAL GAS, BIOGAS, LPG AND METHANE HYDRATES

Natural gas is a flammable gaseous mixture, composed mainly of methane: 70..99% CH4 (e.g. 70% in
Libya, 99% in Alaska), 1..13% C2H6, 0..2% C3H8, and minor concentrations of H2O, CO, CO2, N2, He,
etc. It is found on many underground cavities, either as free deposits (e.g. Indonesia, Algeria, New
Zealand) or linked to petroleum fields (e.g. Saudi Arabia, Nigeria, Alaska). Since the mid-20th century
it is traded by large continental gasoducts (up to 2 m in diameter, with sensors and control valves every
25 km and pumping stations every 100 km) and LNG-ships (Liquefied Natural Gas carriers typically
of 140 000 m3 in capacity, 250 000 m3 for new ones). The liquefaction of natural gas requires the
removal of the non-hydrocarbon components of natural gas such as water, carbon dioxide and
hydrogen sulfide to prevent solid plugs and corrosion.
Liquefied natural gas tankers (LNG ships) were developed in 1960s. The Algeria-Italy submarinegasoduct started operation in 1983, and the Algeria-Morocco-Spain one in 1997. The latter gasoduct,
under the Gibraltar Strait, consists of two 0.5 m in diameter welded steel pipes (tested at 16 MPa) with
a concrete overcoat to protect it from anchors, 50 m apart, with 45 km undersea length up to 400 m
deep (the one under Sicily Strait reaches 600 m depth). Large LNG tanks of up to 50 000 m 3 and
gasholders of up to 100 000 m3 are used as accumulators, but underground cavities (natural or
artificial) seem a better solution. Before putting the dry natural gas on the market, it is sweetened (H 2S
and CO2 are removed by amine absorption), dehydrated (by glycol absorption, to avoid water freezing
and hydrate formation), and some liquefying fractions are extracted (to produce LPG, by isentropic
expansion); natural gas associated to oil fields may contain appreciable fractions of butane and heavier
hydrocarbons, and it is then called 'wet gas'. Differences in natural gas composition have a sizeable
impact on heating value, particularly on volume basis; e.g. HHV=36 MJ/m3 for NG from Russia, 38
MJ/m3 for NG from USA, 33 MJ/m3 for NG from Netherlands, 39 MJ/m3 for NG from UK, 42 MJ/m3
for NG from Algeria, etc. Mind also that HHV of LNG is some 2 MJ/kg lower than natural gas, due to
its low temperature.
Price of natural gas varies not only with time but among world regions (because, contrary to crude oil,
transport costs are significant, and particularly if cryogenic transport is involved); in 2013, based on
LHV, price is around 12 $/GJ in Europe and 4 $/GJ in USA.
Biogas is a flammable gaseous mixture, composed mainly of methane and carbon dioxide, obtained by
anaerobic fermentation of condensed biomass (manure or sewage). The production may range from

20..70 m3 of biogas per cubic metre of manure, lasting 10..30 days within a digestor (depending on the
temperature, that is 20..40 C), where biomass is first hydrolysed by some bacteria in absence of
oxygen, yielding monomers that are made to ferment by other bacteria, yielding alcohol that later turns
to acetic acid and finally decomposes to methane plus carbon dioxide, the later step being the
controlling stage.
LPG (liquefied petroleum gas) are petroleum derivative mixtures (gaseous at ambient temperature, but
handled as liquids at their vapour pressure, 200..900 kPa), mainly constituted by propane, n-butane,
iso-butane, propylene, and butylenes, with composition varying widely from nearly 100% propane in
cold countries, to only 20..30% propane in hot countries (e.g. 100% in UK, 50% in the Netherlands,
35% in France, 30% in Spain, 20% in Greece. In Spain, the traditional bottle for domestic use (UD125) holds 12.5 kg of commercial butane (56% n-butane, 25% propane, 17% iso-butane, 2% pentane,
0,1 g/kg H2O and 1 mg/kg mercaptans, with Tb=0.5 C and L=580 kg/m3 at 20 C). The new
aluminium bottle holds 6 kg (13 kg total, 290 mm diameter and 376 mm height). For higher rates or
cold ambient, propane bottles works better. For vehicles EN-589-1993 applies. LPG is also marketed
in small expandable containers for laboratory use (containing some 50..300 g of LPG, 190 g is the
commonest), and portable 'camping gas' bottles (containing some 2..4 kg of LPG, 2.8 kg is
commonest), with a rough molar composition of 40% propane and 60% butanes (n-butane and isobutane).
All gaseous fuels are odourless (except those containing traces of H2S), and odour markers (sulfurcontaining chemicals, as thiols or mercaptans, e.g. ethanethiol, CH3CH2SH) are introduced for safety
because its detection threshold for human smell is 0.4 ppm in volume).
Table 3. Data for some gaseous fuels.
NG
NG
NG
Propane
Butane
Biogas
a
(Alaska) (Algeria ) (North Sea) (commercial) commercial) (typical)
at 15 C [kg/m3]b
0.74
0.74
0.74
2.0 gas,
2.4 gas,
1.1..1.2
c
520 liquid
560 liquid
HHV (LHV) [MJ/kg] 54.3 (49) 54.3 (49)
53 (48)
50 (46)
49 (45)
33 (30)
CH4 %vol
99
89
82
60
C2H6 %vol
8
9
C3H8 %vol
2
5
>80
<30
C4H10 %vol
1
2
<20
>70
olefins %vol
<20
<20
N2 %vol
1
1
2d
CO2 %vol
1
40
a
For natural gas delivered through LNG carriers, which follows a different treatment than the one
pumped through gasoducts (e.g. methane content may be as low as 83% in the latter case).
b
Standard conditions are usually defined as 15 C and 100 kPa, whereas normal conditions are usually
defined as 0 C and 101 kPa, thus one standard cubic metre equals 0.95 normal cubic metres
(sometimes written 1 Sm3=0.95 Nm3).
c
Thermal expansion of liquid propane =1.510-3 K-1.
d
Typical biogas composition: 55..65% CH4, 35..45% CO2, 1..4% N2, 1..2% H2 and <1% H2S and
H2SO4 before desulfuration.
Pure methane, propane and butane can be easily found from local chemicals suppliers, if the
commercial mixtures traded (natural gas, commercial propane and commercial butane) are not good
for some laboratory work. For small lab demonstrations they may also be obtained in situ; e.g.

methane may be easily produced by means of Al4C3(s)+6H2O(l)=3CH4(g)+2Al2O3(s), or by heating a

50/50
mix
of
anhydrous
sodium
acetate
and
sodium
hydroxide:
NaOH(s)+NaC2H3O2(s)=CH4(g)+Na2CO3(s), as did his discoverer, the American Mathews, in 1899.
Methane hydrates are solid icy-balls (of some centimetres in size) found trapped under high pressure
(>30 MPa) and chilling temperatures (0..5 C) in plant-covered moist places like the continental
sediments on the sea floor and permafrost soil on high-latitude lands. They might be the major source
of natural gas in the future; presently they are a nuisance in high-pressure gasoducts, where they may
block valves. Strictly speaking, they are not hydrates (chemical compounds of a definite formula), but
clathrates, i.e. an unstable network (they tend to the liquid state) of host polar molecules like water,
characterized by H-bonds and regular open cavities, stabilised to a solid state by incorporating small
guest non-polar molecules of appropriate size (to which they are not bonded; only van-der-Waals
forces act to stabilise the network). Besides methane, carbon dioxide, hydrogen sulphide, and larger
hydrocarbons such as ethane and propane, can stabilize the water lattices and form hydrates"; smaller
molecules like nitrogen, oxygen or hydrogen are much more difficult to stabilise in water.
Methane hydrates (approx. CH46H2O) fizzle and evaporate quickly when depressurised, yielding
some 150 times its volume of methane. Since this methane comes from very large-time biomass
decomposition, the problem of global warming remains: it yields CO2 on burning, and released CH4
losses are worse: 20 times more relative greenhouse effect that CO2. Hydrates soils are prone to
accidental landslides, particularly during exploitation, what constitutes a high risk to extraction
platforms.

HYDROGEN
In the long term, hydrogen-energy appears as the final solution to face the energy-environment
dilemma of scarcity and pollution, not only for the much-pursued nuclear-fusion power stations (using
hydrogen isotopes), but for the using of hydrogen as an intermediate energy carrier (like electricity),
cleanly produced from water and solar energy, and cleanly converted back to water, to drive fuel cells
engines and clean combustors.
Production
Pure hydrogen (H2) is an artificial product on Earth (1 ppm in the atmosphere), but nearly 100%
of Jupiter atmosphere and 90% of all atoms in the Universe (nearly 3/4 of its total mass). On
Earth, it is found combined in water, living matter and fossil matter. Discovered in 1766 by
Cavendish (used in 1520 by Paracelsus as inflammable air), and named by Lavoisier in 1781
First massive production in 1782 by Jacques Charles (Fe(s)+2HCl(aq)=FCl2(aq)+H2(g)) to
inflate a balloon (he flew 25 km from Paris on the same year of Montgolfier's brothers fly with
a hot-air balloon). Present use is mainly for chemical synthesis (e.g. ammonia), metallurgy,
ceramics, for the hydrogenation of fats, as a cryogenic fuel in rockets, in cryogenic research,
and as a fuel-cell fuel.
Production at large (world production in 2010 was 40109 kg), is based on fossil feedstock:
Some 50% of world H2 production is from natural gas reforming. The process,
CH4+aH2O=(2+a)H2+bCO2+cCO is carried out at 1150 K with Co-Ni catalysts, but >2000
ppm-CO is left and PEFC-type fuel cells required <20 ppm-CO. Instead of fully purifying
the H2, it is easier to purify until <1000 ppm-CO and add oxygen to get rid of the CO at the
catalyst (but if more O2 is used, it reacts with H2 at the catalyst producing just heat). In
MCFC & SOFC the CO is an additional fuel. Natural gas reforming is presently the best
method to produce hydrogen while renewable sources are being developed, but availability

of natural gas is in question if new major sources, as seabed clathrates, are not made
available (and in that case with CO2 sequestration).
Some 30% of world production is based on naphtha reforming in crude-oil refineries.
Another 20% of world production was based on coal reforming (declining rapidly): C+H2O=
H2+CO at 1300 K and CO+H2O = H2 + CO2 with Fe0-CrO2-ThO2 catalyst.
A small percentile of world production (<4%) is based on water electrolysis from cheap
hydroelectric energy in Canada and Scandinavia: H2O = H2+(1/2)O2, with e=0.65. It is the
purest H2. As fossil fuels are being exhausted, water electrolysers seem to be the most
popular hydrogen sources in the future. Electrolysers with liquid potash lye produce
hydrogen cheaper than other kinds of electrolysers (e.g. proton-exchange-membrane, PEM,
electrolysers).
Production at intermediate locations (for transportation or for stationary applications) by
reforming (see 'Reforming' details below)
Methanol reforming (methanol has 12.5%wt of hydrogen).
Ethanol reforming (ethanol has 13%wt of hydrogen). This may be the best H2 production
method in the long term, and not electrolysis with renewable-energy.
Gasoline reforming (gasoline has some 16%wt of hydrogen; diesel has a little less and is not
used). Gasoline may be thermally decomposed at >800 C (or best at 300 C with Nicatalyst), CuHv+uH2O = (u+v/2)H2+uCO, syngas=synthesis gas, endothermic but good
energy rate (78%), but contaminates a lot (and desulfurisation is required). With more
H2O(g) may yield H2+CO2 (40%/60%). The more aromatics in gasoline, the worst reforming
(that is why diesel is bad).
NG (natural gas, methane) or LPG (propane+butane) reforming. Methane has 25%wt of
hydrogen; water only 11%, however, gaseous fuels are no good for storage and
transportation. Natural gas is most used today in low-temperature fuel cells (PEFC and
PAFC); it is first desulfurised (from previously added safety odorants!), then steamreformed (yields 10% CO), afterwards shift-converted (reduces CO to 1%), and (only for
PEFC) finally selectively oxidised (to <10 ppm CO).
Coal reforming. No good because of pollution and high temperature work.
From other intermediate storage compounds (too expensive at present): NH3 (liquid at 1
MPa & 298 K, but it is a poison), N2H4 (flammable liquid).
Production in the lab
Zn(s)+2HCl(aq)= H2(g)+ZnCl2(aq), i.e. dripping a strong acid over a metal.
Na(s)+H2O(l)= (1/2)H2(g)+NaOH(aq).
Si(s)+2H2O(l)=SiO2(s)+2H2(g)+339 kJ/mol. In practice there is an intermediate steps (if not,
it is far too slow): Si(s)+2NaOH(aq)+H2O(l)=Na2SiO3(aq)+2H2(g)+424 kJ/mol at room
temperature, and Na2SiO3(aq)+H2O(l)=2NaOH(aq)+SiO2(s)-85 kJ/mol at 220 C. This
technique may become commercial (Si is non-toxic non-CO2 and non-CO producing,
compared to methanol). Presently Si is produced by SiO2-reduction with charcoal in an
electric-arc furnace, producing 109 kg/year at 1 /kg (consuming 29 MJ/kgSi plus 2.7
kgcoal/kgSi). Direct electrolysis of molten SiO2 is not developed. Si is not good for direct fuel
(in a SOFC) due to the very small anode voltage. Order of magnitude world production and
cost for metals is: Fe 0.2 /kg, Al 25109 kg/year at 0.4 /kg. Pb 40106 kg/year at 0.4 /kg,
Pt 15000 /kg.
Reforming a fuel is producing another fuel from it. Several reforming processes exist:
Partial oxidation (PO), sometimes catalytic (PCO). It is the exothermic reaction with deficient
oxygen; it is the simplest reforming process, but gives rise to very high temperatures (2000 K)
and pollutants (NOx, NH3, HCN) without appropriate catalysts. For methanol reforming,
CH3OH(vap)+O2=2H2+CO2+667 kJ/mol (in reality yields 40% H2(g) instead of 67%), a Pd

catalyst is used. For natural-gas reforming, CH4(g)+O2=2H2+CO+36 kJ/mol (in reality yielding
1.3 molH2/molCH4 instead of 2), a Pt or Ni catalysts on alumina are used, working at 1200 K (a
big problem is the formation of soot, 2CO=C+CO2, that clogs the catalyst).
Steam reforming (SR). It is the endothermic reaction with water vapour; it is the most widely
used reforming process and the one that yields more hydrogen, but a very complex one because
of the required external heating (only used for production >500 kg/day). Methanol reforming,
CH3OH(vap)+H2O(vap)=3H2+CO2-49 kJ/mol (in reality yields 70% H2(g) instead of 75%, and
consumes up to 20% of HHV, at 250 C, with Ni plus a final pass through Pt to further oxidise
CO to CO2, CO+H2O=CO2+H2+41 kJ/mol; or with Cu/ZnO); although it works for PEFC, it is
best suited to PAFC because of the CO2, and PAFC have the advantage that the vapour is
produced with the by-product heat. Natural-gas reforming, CH4(g)+H2O=3H2+CO-206 kJ/mol is
carried out at 900 K to 1200 K in a gas furnace, with Ni-catalyst on alumina plus a final pass
through Pt to further oxidise CO to CO2; although low pressure favours the reaction, it is not
used in practice.
Autothermic reforming (AR). It is just a combined PO+SR process that it is adiabatic overall;
this is presently the most economic method of H2 production.
Thermal decomposition (TD). It is the endothermic cracking at high temperature. For methanol,
CH3OH(vap)=2H2+CO-95 kJ/mol (it is not used alone but adding water, i.e. SR). For methane,
CH4(g)=2H2(g)+C(s)-75 kJ/mol, using a Ni-catalyst on silica, that must be regenerated with
oxygen from time to time to get rid of the carbon deposited; notice that no CO is involved.
Carbon-dioxide reforming (not much used). For methane, CH4(g)+CO2(g)=2H2+2CO-248
kJ/mol.
Membrane reactor. It is not reforming itself but a post-processing stage to reforming; after SR
(or PO or AR) the gas flow is exposed to a selective membrane that yields 90% H2.

Storage & transport

History of hydrogen storage and transport is always associated to the Hindenburg-1937 and
Challenger-1986 catastrophes (as nuclear energy to the Hiroshima bomb).
Several hydrogen storage systems may be used:
Compressed gas at 30 MPa and Tamb (e.g. 3 kg for 500 km in a car, 0.18 m3 and 25 kg tank).
Requires container inspection every few years. It is the most used storage method for small
applications.
Cryogenic liquid at 20 K and pamb (up to 0.3 MPa really; e.g. 3 kg, 0.10 m3 and 45 kg tank). With
present dewar-tanks it evaporates in two weeks, and it presently costs 35 MJ/kg just to liquefy
H2(g) from ambient conditions (30% of its lower heating value).
Chemically absorbed in metal hydrides (up to 2%; e.g. 3 kg, 0.05 m3 of LaNi5H6), stored by
moderate overpressure and/or cooling, and given-off by depression and/or heating. Hydrogen, as
most other gases, hardly dissolves in liquids (e.g. 1.6 ppm by weight in water at normal
conditions).
Chemically combined with alkaline metals. A non-flammable non-volatile alkaline aqueous
solution of sodium borohydride (NaBH4) at 20%wt, gives off pure H2(g) by contact with a
catalyst (ruthenium, cobalt) at room temperature, by exothermic hydrolysis:
NaBH4(aq)+2H2O=4H2+NaBO2(aq). Power is controlled by insertion/removal of the catalyst.
Chemically stabilised in clathrates (i.e. metastable crystal networks of host polar molecules like
water). Hydrogen clathrates are more difficult to get than methane clathrates; they are realised by
slow compression of water and H2(g) at room temperature up to 200 MPa followed by cooling to
some 250 K. Although this requires a refrigerated storage at 250 K, this could be easily managed
for instance with a small venting of innocuous liquid nitrogen vapours.

 Gas trapped in glass micro-spheres or nano-fibres. Very high pressure is needed to put the gas
inside, but afterwards the pressure can be lowered without leakage (safe to transport and store),
and released by heating.
It is curious that 'hydrogen is roughly as light in air as in water': 7% (of course meaning that the gas
has 7% of the density of air, and the liquid has 7% of the density of water, although the last couple can
never be in thermal equilibrium).
Safety
Hydrogen is a dangerous flammable gas, with the same self-ignition temperature as methane (850 K),
but much wider flammability limits (in air, 4..75% instead of 5..15%), smaller energy for ignition (15
times less), smaller quenching distance (0.6 mm instead of 2 mm), nearly invisible non-premixed
flame, and more prone to detonation (in air, detonability limits are 18..59% instead of 6..14%). But, in
relative terms to other fuels, hydrogen is not so much dangerous (some claim it is safer), its main
advantage being its extreme lightness, which, in ventilated spaces, makes H2 leaks and flames to
vertically escape quickly, minimising possible horizontal spreads (most ignition sources and valuables
accumulate horizontally). Hydrogen is not toxic itself, and burns with not toxic fumes (most deaths
caused by fire are actually due to deadly fumes and gases).
As further reassurance on H2 safety, it must be acknowledged that, before natural gas took over, town
gas (50% hydrogen) was pipelined to most dwellings in most large cities for nearly a century, with the
same or even less damage to people and goods. The often-used stereotypical example of the
Hindenburg catastrophe was not due to its lifting hydrogen filling, but to its badly-designed canvass
coating, made of a flammable combination of dark iron-oxide and reflective aluminium paints to avoid
solar heating. It is concluded now (at the time it was attributed to a H2-leakage) that, while mooring to
a 50 m high metallic docking tower in stormy weather, an electrostatic spark ignited the coating and
set fire to everything combustible: the canvas, hydrogen inside the bags, and engine diesel fuel (35 out
of the 97 passengers lost their lives, only 8 people by burning, all of them by diesel flames in the
cabin; the other victims fall or jumped to the void in despair; a helium-filled air-ship would have had
the same type accident and caused the same casualties, it is said).
Hydrogen is most dangerous in poorly-ventilated and closed spaces, due to risk of explosion by
deflagration or even detonation. Compressed hydrogen storage posses the additional problem of
pressure-vessel explosions, entirely similar to any other compressed gas vessel, but great advances
have been recently made on safe-breakage of very high pressure containers made of laminated fibrereinforced polymers.
Purity
Required hydrogen purity depends on intended use (it is easy to understand that requirements for the
ammonia industry are different than those for food hydrogenation). For fuel cell applications, standard
purity (e.g. the one offered at the few hydrogen-supply-stations) is compiled in Table 4. A safety
problem related to purity is that presently hydrogen cannot be marked with any suitable odorant to ease
early leak-detection by humans, as done with all other odourless flammable gases.
Table 4. Admissible impurities in hydrogen supply for fuel cell vehicles.
Component
Tolerance
Inert gases (He+Ar+N2,)
<1%
Agua lquida (H2O(l))
<0.5%
Oxygen (O2)
<500 ppm

Carbon oxides (CO+CO2)

Hydrocarbons (total, including lubricants)
Sulfur
Ammonia
Inorganics

<2 ppm
<1 ppm
<1 ppm
<0.01 ppm
<0.01%

Price comparison of hydrogen energy

Price of an electricity generator, by kW-installed. As for year 2000, electrical utility stations cost
500 /kW (20-yr life), internal combustion engines 30 /kW (20-yr life), fuel cells 3500 /kW
(<10-yr life), solar cells 5000 /kW (<10-yr life).
Price of energy, by MJ-produced (before taxes). As for year 2000, gasoline 0.006 /MJ,
methanol 0.012 /MJ, hydrogen 0.020 /MJ, electricity 0.025 /MJ (notice that fuel energy
refers to its higher heating value).
As for year 2000, hydrogen costs four times the price of natural gas for the same energy content.
Hint: Hydrogen balloon combustion & explosion
A balloon filled with hydrogen gas burns when ignited but does not produce a large bang. However, if
a balloon filled with hydrogen is allowed to remain untouched for a while, oxygen gas from the air
begins to effuse through the tiny pores in the balloon material. After a while, the balloon contains a
mixture of H2, O2 and some other non-reactive gases. When ignited with a lighted candle, the gases
inside the balloon quickly react to produce yellow flames and a large bang.

COAL
Coal is the most plentiful fuel available on Earth; proven commercial reserve in 2000 were >10001012
kg (some 30% in USA, 20% in Russia, 15% in China, 15% in EU-15), but it is a finite, non-renewable
and very-polluting source. Estimations in 2000 gave a reserve/consumption ratio of some 250 years.
Coal was known since ancient times, but only used when available on site; most often, a charcoal, a
similar stuff obtained from wood, was used (charcoal leaves less ashes, but yields less heat also).
Origin
Coal is a compact black or dark-brown sedimentary rock (a mixture, not a mineral) formed some 300
million years ago (Carboniferous Period, although smaller deposits exists from 200 Myr, dinosaurs era,
and 100 Myr ago), by high pressure and temperature anaerobic decomposition of dead plants (mainly
ferns), but the degree of metamorphosis varies a lot and several types of coal can be found today;:
from more to less cooking: anthracite, bituminous coal, lignite, and peat.
Types
Coals are classified in a maturity rank according to age and fixed carbon content:
Anthracite is the hardest, purest, more brittle and scarce coal; too precious for a fuel (it is used
for chemicals). It is a dense black solid, with brilliant lustre and very low moisture. (<5% of
trapped water).
Bituminous coal is a dense black solid that frequently contains bright bands with a brilliant
lustre. It has some 2..10% of trapped water. Volatile matter range is 10..30%, and the typical
heating value is 30 MJ/kg.
Lignite, or brown coal, is the most abundant form of coal. It has some 40..60% of trapped
water. Sometimes, the remaining texture of the original vegetation can be discerned. Volatiles
may reach up to 30% in some lignite, and the typical heating value is 15 MJ/kg.

Peat has some 70..90% of trapped water, and it is not used as commercial fuel (it must be dried
to at least 30% in water to be burned, and was used in open fires). Plant decomposition has
progressed only partially, and it is possible sometimes to identify the remains of individual
leaves in peat, cellulose being still the main component.

Uses
Coal was used in China in 1100 BC, in Wales during the Bronze Age, and as a home heating fuel in
England since Roman times and during the Middle Ages (Newcastle coast was plentiful of so-called
sea-coal), but mining really took off in the 18th c., also in England, driven by the Industrial Revolution
hunger to feed steam boilers. It was after a wood-shortage in West Europe in late 16th c. that coal
started to be used in the metallurgical industry, with the coke discovery in 1603 by H. Platt, by heating
coal. In the mid 19th c. coal also replaced charcoal in making black-powder and explosives. A great
amount of coal is used to make iron (reducing iron ore). Coke, used in blast furnaces to yield clean
concentrated heat, is almost pure carbon but amorphous, not as graphite) obtained by anaerobic heating
of coal, to separate volatile matter, that is distilled aside.
Coal is not so valued as to worth it deep underground mining; even today, more than half of the world
coal is from surface mining (open cut or strip mining), although mines some 500 m deep have been
worked (following a surface coal layer deeper and deeper).
Coal had the leading share in world energy production from 1800 to 1950, when oil took over, and still
has a large share: 24% of primary energy consumption, mainly restricted to electricity generation in
large combustion power plants (38% of world electricity generation).
Since coal consumption in its raw solid form is cumbersome, several fluidification processes have
been developed:
Pulverisation by mechanical grinding.
Gasification by pyrolysis. Heating coal with air (and preferably also water vapour) produces
coal gas (town gas) and leaves a solid residue (coke). In 1792, Murdock, a Scottish engineer, lit
his home with gas lamps that burned coal gas. Early in the XIX c, gas lamps had come into use
as street lamps in London. The FischerTropsch process (1920s) convert a mixture of carbon
monoxide and hydrogen into liquid hydrocarbons (as done by Germany in II World War;
synthetic gasoline was transparent).
Distillation by pyrolysis. Heating coal in absence of air at >1000 K, produces a mixture of
volatiles (CH4, H2, CO2, HCN), a liquid pours out (coal tar), and a solid remains (coke: nearly
pure carbon, used in blast furnaces). Coal tar is further distilled: a first fraction yields benzene
and toluene, a second fraction naphthalene's, a third fraction creosotes, a fourth anthracenes,
and the residue is called coal-tar pitch.
Composition: proximate analysis and ultimate analysis
Coal is a natural composite material, not a single chemical compound, and its composition, being a
solid substance, is measured by weight, and referred to either to as-received coal (wet coal), or on a
dry basis, or on a moisture-free-ash-free basis. Two types of coal analysis are commonly used:
Proximate analysis, specifying: fixed coal + volatile coal + ash + water.
Ultimate analysis, specifying: water + percentage of C, H, O, N, and S.
In any case, coal is first drained, and moisture contents is computed by differential weighting after
heating the sample to 103 C for several hours. Sulfur content is computed either by chemically
converting all S-ions to sulfate-ions and precipitation to BaSO4, or by chemically fixing the sulfur

oxides formed on combustion of the sample. Ash is measured as the solid residue on combustion;
although there is a difference in composition and weight between initial mineral matter and final ash,
the latter is most often used in analysis.
In the proximate analysis, once water is eliminated, volatile coal is computed by differential weighting
after heating the dry sample to 950 C for several minutes (volatiles consist mainly of H 2, CO, CH4,
C2H6, O2, H2O and maybe other volatile organic compounds). Fixed coal is computed by subtracting
from sample weight moisture, ash, sulfur and volatile coal.
In ultimate analysis, once water is eliminated and sulfur and ash measured aside, the percentage in
carbon is measured from CO2 content of complete burning the sample, the amount of hydrogen is
computed as 1/9 of the weight of water produced in combustion, the amount of nitrogen (usually less
than 1% in weight) is measured aside or neglected, and the amount of oxygen is computed by
subtracting from the total.
A typical dry coal composition is 85% C, 5% H, 5% O and 5% ash by weight). The main impurity in
coal combustion has been sulfur: 2..10% in weight. Some sulfur can be removed by just grinding and
washing, because it is associated to iron in pyrite chunks that settle, but all modern power plants have
flue gas desulfuration units (scrubbers) to get rid of it by absorption in a lime spray.
In summary, coal composition, as received, is as follows:
5..15% of moisture, i.e. water mass divided by dry mass (what is left at 103 C). Lignites can
have 30..60% moisture. Wood moisture is 20..80%.
10..20% of ash, i.e. the solid residue after burning. Wood ash is 1..3%.
The remaining 70..90% is dry ash-free coal, further divided in volatile matter (20..40% by
weight in dry ash-free basis) and fixed coal (the solid residue). Overall dry ash-free coal has a
composition of:
Carbon content in dry ash-free coal is in the range 80..90% by weight. Notice, however,
than coal from the mine has less %C by weight (some 70%) than oil (85%) or natural gas
(75%).
Hydrogen content in dry ash-free coal is in the range 5..6% by weight.
Oxygen content in dry ash-free coal is in the range 1..10% by weight (it may range from
1% in dry ash-free anthracite to some 45% in wood just cut). This oxygen decreases both
the theoretical air required for complete combustion of the fuel and its heating value, since
it is chemically combined with hydrogen and carbon atoms (the HHV reduction per
kilogramme of oxygen would be 17.9 MJ/kg if combined with H, or 12.3 MJ/kg if
combined with C; a empirical value of 15 MJ/kg is often used, what gives for the higher
heating value, HHV=0.33%C+1.43%H0.15%O in MJ/kg, for the mass-percentages
composition).
Sulfur is typically 1..5% in dry ash-free coal.
Nitrogen is typically 0..2% in dry ash-free coal.
Air requirement for theoretical combustion
Pure carbon demands 11.5 kg of air per kg of fuel, corresponding to 1 mol of oxygen per mol of
carbon (C+O2=CO2) and 23.2% by weight of oxygen in the air. In terms of the dry ultimate analysis of
a coal, u%C+v%H+w%O+x%N+y%S+z%Ash, the air demand is the sum of the C-, H-, and S- airdemand, minus the contribution of its oxygen content:

A0

2.67 %C 8 %H 1 %O 1 %S
(e.g. A0=11.5 kg/kg for %C=100)
23.2

(1)

with typical coal values of A0=10 kg of air per kg of coal.

Heating values
Pure carbon has a higher heating value of HHV=32.8 MJ/kg, corresponding to the enthalpy of reaction
C+O2=CO2+393.5 kJ/mol. Actual coals would have a heating value made up of the contribution of C-,
H-, and S- oxidation, assumed to be bond-free, minus the contribution of the existing bonds,
particularly those with oxygen, which could not be further oxidised; i.e. oxygen decreases the heating
value for complete combustion, since it is already bonded to hydrogen and carbon atoms.
The HHV reduction per kilogramme of oxygen would be 17.9 MJ/kg if combined with H (286 kJ per
half-a-mole of oxygen, H2+(1/2)O2=H2O+286 kJ/mol), or 12.3 MJ/kg if combined with C (393.5
kJ/mol per mole of oxygen, C+O2=CO2+393.5 kJ/mol); an empirical value of 15 MJ/kg is often used,
what gives for the higher heating value of a coal in terms of its dry ultimate analysis,
u%C+v%H+w%O+x%N+y%S+z%Ash:

HHV
0.328 %C 1.43 %H 0.15 %O 0.09 %S
[MJ/kg]

(2)

with typical values in the range HHV=20..30 MJ/kg for soft coal (bituminous coal), HHV=10..20
MJ/kg for brown coal (lignite), and HHV=8..13 MJ/kg for peat and schist (but peat briquettes may
reach 17 MJ/kg).
Equations (1-2) can be used for any kind of complex fuel of unknown molecular formula (for which
the ultimate analysis per weight is available), either solid (coal, wood), or liquid (gasoline, diesel,
heavy fuel or crude-oil).
Price of coal for thermal power plants in 2013 is around 60 $/t (at 30 MJ/kg it means 2 $/GJ, much
lower than any other fossil fuel, e.g. natural gas may be 5 times more expensive).

WOOD
Wood is the hard, porous, fibrous substance found beneath the bark of trees and shrubs. Wood is used
for timber (construction), for paper making, and as a fuel (up to 30% of wood production in
industrialised countries, mostly the debris, but up to 90% in developing countries). Hard wood
(resistant to sawing) comes from deciduous broad-leafed trees: oak, elm and fruit trees. Soft woods
come from pine, cedar, fir. Paper is made from chemically and mechanically processed wood fibres
(typically 30 mm in diameter and 2 mm long) which are self-binding when dried from a wet state.
Composition
Wood is a natural composite material consisting of hollow polymer fibres in a polymer matrix. The
hollow fibres are tubular cells (most of them dead), with cellulosic walls (70% cellulose and 30%
lignin) holding aqueous solutions in the inside space (water content varies a lot, from 60%wt in freshly
cut trees, to 5%wt in artificially-dried furniture-wood); the matrix is made of hemicellulose and lignin.
Trees are cut in winter to minimise initial water content.
Chemically, dry wood is an aggregate of 40..45%% cellulose (a long homopolysaccharide, (C 6H10O5)n
with n=5000..10000 and M=500..10000 kg/mol), 20..30% hemicellulose (a short-chain polymer with

n=150..250 and M=20..30 kg/mol) and 20..30% lignin (a reticular polymer with a 3-D structure that
glues all together), with some 7% ash (the main components being SiO2 and CaO); the elementary
chemical composition of wood is presented in Table 5. Cellulose, Fig. 2) is the most abundant
naturally occurring organic substance; it does not dissolve water, but it forms very porous structure
that makes it very hygroscopic (e.g. wood, paper, or cotton, that is practically pure cellulose; pure
cellulose is a white crystalline powder).

Fig. 2. Molecular structure of cellulose ((C6H10O5)n with n=5000..10000).

Table 5. Wood composition (%wt) and lower heating values.
%wt water
0
20
40
60
C
50.30 40.24 30.18
20
H
6.20
4.96
3.72
2.7
O
43.08 34.46 25.85
17
N
0.04
0.03
0.02
0.01
S
0.00
0.00
0.00
0.00
ash
0.37
0.31
0.23
0.15
Total
100
100
100
100
LHV kJ/kg
19900 15400 10950 6500
The biological structure of wood is basically a hard hollowed tissue in a cylindrical arrangement. The
inner part is mainly xylem (mostly dead conduits of dxylem=2510-6 m, say between 10 m and 100 m;
the outermost are alive and allow for water and nutrients supply to the leaves from the roots), whereas
the outer part is mainly cambium and phloem (living conduits that return elaborated sap from the
leaves to everywhere). Cambium divides every spring, yielding xylem to the interior and phloem to the
outside (cambium divides only transversally, increasing the diameter some 2 mm (conifers) up to 7
mm soft woods); apical tissue divides axially, increasing the length). Old xylem vessels (the core or
duramen) die and become clogged with dry metabolic wastes, such as tannins, dyes, and resins, just
providing structural support to the plant. There is also a tissue that transports water transversally from
xylem to cambium (wood rays)
Density. It depends on water and air content (porosity). Maximum is cellulose=1550 kg/m3, but very
soft wood (balsa wood) only has 50 kg/m3.
Thermal conductivity. It depends on water and air content (porosity), and on direction (anisotropy).
Dry wood is a very good insulator due to the air spaces, with k=0.3 W/(mK) along the fibres
and k=0.1 W/(mK) transversally.
Heating value. It depends on water content (moisture). Maximum is HHV=20 MJ/kg for dry wood;
wood pellets may have HHV=17..18 MJ/kg, log wood HHV=15..16 MJ/kg, and wood chips
HHV=13..15 MJ/kg.

Why using wood as a fuel, if a gas or oil burner works smoothly and effortless?. First, there is the case
of developing countries where wood is the only fuel, aside of domestic waste. Second, there is the case
of wood residues from forest cleaning or industry, which may be used as a commercial fuel instead of
just burning it to get rid of. But, besides any practical justification, there is the aesthetic reason of the
intrinsic beauty of wood fire and its cultural heritage (our ancestors developed around the fireplace
heart). It takes time, effort and thought to burn firewood, but it is creative, inspiring and really warm,
bringing you closer to Nature.
Biomass
Here, biomass is synonymous of vegetable matter used as fuel (biofuel), either grown for that purpose,
or recovered from other industries waste (forestry, farming, food industry); urban and animal waste
might be included too, but its importance is marginal. Municipal solid waste (MSW) has great organic
content and can be used as a fuel in incineration power plants, with HHV=7..12 MJ/kg, but dioxin
emission is a problem. It excludes organic material which has been transformed by geological
processes into coal, petroleum, or natural gas (fossil fuels).
Biomass is a renewable fuel, and, to a first approximation, carbon neutral, in the sense that the CO 2
released in biofuel combustion was previously captured from the environment during biomass growth,
although in October 2007, Nobel Laureate Paul Crutzen published findings that the release of Nitrous
Oxide (N2O) from rapeseed oil, and corn (maize), contribute more to global warming than the fossil
fuels they replace.
The traditional biomass through the ages has been wood. Besides the biofuel production here
discussed, biomass is also used as a fertiliser (compost), paper industry and other chemical stuff,
building (e.g. straw in adobe and roofs, timber), etc.
Biomass can be directly burned in furnaces and boilers, but the preferred way to easy handling and
transportation, and to minimise pollution, is by transforming raw biomass into gas (known as biogas or
syngas), liquid (which may range from alcohols to tars), and solid (char, pellets). At present, liquid
biofuels (bioethanol and biodiesel) are mixed with oil derivatives (gasoline and diesel) in a 5%..20%
biofuel fraction. Table 6 presents some examples of biomass composition:
Table 6. Examples of biomass composition (%wt).
Sugar cane bagasse Pine sawdust Almond shells Grape stalks
moisture
7
9
12
8
C
46
45
41
41
H
5
5
5
6
O
40
39
39
40
N
0
0
1
0
S
0
0
0
0
ash
1
1
3
5
PCI kJ/kg
16 200
16 400
16 000
16 700
A problem facing energetic crops is the impact on food crop price (the food vs. fuel debate), since, at
present, some crops and some fields are of double use, as when cereals are grown for either food or
bioethanol, depending on market price, instead of producing the ethanol from cellulose in the straw.
An additional problem is the impact on water resources. Presently, most common energy crops are:
corn, soybeans, rapeseed, sugar cane, barley, sorghum, palm oil In the future, biofuels from nonfood crops and from algae, should be used.

FUEL PYROLYSIS
Pyrolysis is the chemical decomposition of compounds caused by high temperatures with no access to
air or oxygen (i.e. it is not a partial oxidation or combustion process). Most of the times, the
decomposition products are separated by distillation (the whole process is called dry or destructive
distillation). Pyrolysis implies chemical reactions; the first stages in heating a fuel, up to 400 K,
usually imply just a moisture loss (although de-polymerisation may occur after some time). When
water is added during the pyrolysis of a fuel, the process is best known as reforming. Recently,
catalytic pyrolysis is being developed to drive the process to the most demanding yields. We focus
here on pyrolysis of some raw fuels (wood, coal, crude oil, biomass) to get more suitable fuels.
Pyrolysis is the major step for crude-oil cracking in the petroleum industry to get lighter fuels (and to
get ethylene and hydrogen for synthetic chemicals).
The earliest use of pyrolysis was to produce coke (a smokeless form of coal without volatiles, with
more than 80%wt carbon, used in metallurgy and as a fuel), by heating coal to more than 1200 K. The
vapours can be cooled to get a liquid condensate (known as coal tar) and then distilled to yield, as a
first batch up to 500 K, a mixture of benzene, toluene and xylene; from 500 K to 550 K mainly
naphthalene, from 550 K to 600 K mainly creosotes, and from 600 K to 650 K a mixture of quinoline,
anthracene and phenanthrine, leaving a solid residue called coal-tar pitch, used as a heavy fuel or as a
binder in making coal or coke briquettes. Similar organic products are obtained from the pyrolysis of
any natural organic material: wood, coal or crude oil.
Wood and wood residues are basically cellulose, like paper. When heated, up to 100 C (<400 K),
wood tends to dry, i.e. to loose moisture by evaporation, and some internal depolymerisation starts,
slowly decreasing wood strength. Between 100 C to 200 C (400..500 K), non-combustible products
are released, as water vapour, carbon dioxide and other trapped gases. Above 200 C (>500 K) the
cellulosic polymers decompose yielding volatile organic compounds, with a weight loss of some 10%
from 100 C to 300 C (400 K to 600 K). Above 400 C (>700 K) all volatile material is gone, with a
sudden 80% weight loss from 300 C to 400 C (600 K to 700 K), and a carbonaceous char (charcoal)
remains (charcoal is often approximated as pure carbon, but it is really partially pyrolysed cellulose
best approximated by the empirical formula C7H4O). In the presence of air, the flash point is around
250 C (500 K), giving a slow-burning glow or a prominent flame according to the heat losses, and the
autoignition is around 400 C (700 K). A now-obsolete method for making methyl alcohol (wood
alcohol) was to pyrolyse wood chips and separate the volatile alcohol by fractional distillation.
Pyrolysis of biomass (straw and other agricultural feedstock, or organic wastes from the cellulose,
paper and sugar industries) generates three different energy products in different quantities: coke
(char), oils (tar) and gases (syngas). In low-pressure fast pyrolysis the biomass is fast heated and
allowed to decompose into vaporised oils that are fast cooled and condensed, in order to eliminate any
polymerisation. In catalytic high-pressure pyrolysis the biomass is heated at a high pressure with a
reducing gas and a catalyst, and a synthetic gasoline is produced.
Pyrolysis of plastics and rubbers, taking place in most uncontrolled fires, produce very toxic gases.
Some plastics decompose to their monomers (e.g. (PE, PMMA), whereas others generate new
substances, as when PVC decomposes to HCl). High-temperature resisting plastics decompose to
almost-exclusive carbonaceous residue.
Table 7. Thermal softening, decomposition, and heating value of some materials.
Material
Density Softening Decompo- Minimum Autoignition Combustion
[kg/m3]
Tg
sition Td ignition Tflas
Tself-ign
HHV [MJ/kg]

Bakelite (Phenol1300
NA
formaldehyde, PF)
Cellulose (90% in cotton)
1600
NA
280 C
300 C
400 C
17
Coal
1400
NA
200 C
28
Methacrylate (PMMA)
1180
85 C
180 C
300 C
450 C
26
Nylon (Polyamide, PA)
1140
220 C
300 C
450 C
500 C
32
Polyacrylonitrile (PAN)
250 C
450 C
550 C
Polyester
1380
260 C
18
Polyethylene (PE)
930
100 C
350 C
350 C
370 C
46.5
Polypropylene (PP)
910
170 C
350 C
350 C
370 C
46
Polystyrene (PS)
1040
100 C
300 C
350 C
500 C
42
Polyurethane (PU)
1100
NA
300 C
400 C
PVC (polyvinylchloride)
1400
75 C
200 C
350 C
450 C
20
Teflon (PTFE)
2250
330 C
500 C
550 C
600 C
Wood
500
NA
200 C
300 C
400 C
25
NA. Not Applicable (e.g. glass-transition temperature is not applicable to thermosetting polymers)
Cellulose is the most abundant polymer on earth. Like most polymers it is not a well-defined molecule
but a family of closely related chemical structures known as polysaccharides, consisting of long
unbranched chains of H-bond-linked glucose units, -(C6H10O5)n- with n=300..2000, with size of order
10-7 m and M50..300 kg/mol. The linear chain is very coiled, masking the H-bonds and making
cellulose highly insoluble in water, but it forms very porous structure that make it very hygroscopic
(e.g. paper). Cellulose is the main constituent of plant cell walls, but not digestible by humans; if its
hydrolysis, saccharification and fermentation were mastered (like in ruminants and some insects like
termites), it would be a less expensive source of glucose than starch.
(Back to Combustion)

FUEL CONSUMPTION
Fuels and energy utilisation .......................................................................................................................... 1
Fuel consumption ...................................................................................................................................... 2
Fuels in world energy production ......................................................................................................... 3
Fuels in electricity production............................................................................................................... 4
Fuels in end-use energy consumption ................................................................................................... 5
Fuel-to-energy conversion factors ........................................................................................................ 6
Energy management .................................................................................................................................. 8
Energy future......................................................................................................................................... 8
References ................................................................................................................................................. 9

FUELS AND ENERGY UTILISATION

Besides air, water, and food, humankind needs energy services: lighting, heating, cooling, pulling and
pushing goods (actuators), transporting goods and people, and powering information devices
(semaphores, radio and TV transponders, computing machines...).
We might question how much energy we need to satisfy our wanted services, but who knows the answer.
If we look into the past, trying to extrapolate into the future, humankind energy expenditure has grown
differently on different energy services:
Energy used to procure food and water is now (say year 2000, per capita) five times larger than
106 years ago. This is the consequence of most people living in large service cities, and only few
people devoted to provide food and water to all.
Energy used for transportation is now sixty times larger than 500 years ago (the start of ocean
travels). Is that surge in transportation-energy consumption really needed? Human mobility a
basic human need, but to what extent? Is it not really a burden sometime, wasting nowadays
several hours a day to go to work and back home?
Telecommunication technology, on the other hand, seems no so energy-eager (compare a
videoconference-meeting with a presence-meeting arranged via individual car transportation).
However, every conceivable non-inert system (from biological organisms to just mechanical
clockwork mechanisms) generates entropy, which must be evacuated as heat, and must be
compensated by an exergy input to keep the process steady.
At present we only use two final-user commercial-energy carriers: fuels (piped or batch-delivered) and
electricity (wired through the grid, or stored in batteries). Human metabolism needs some 100 W/cap (100
watts per capita), and humankind consumes some additional 1800 W/cap of final energy, coming from
2400 W/cap of primary energy: some 300 W of electricity (produced from some 900 W of primary
energy, mostly from fossil fuels), plus some 1500 W of end fuels (refined from raw fossil fuels, and used
for transportation, heating, and so on).
Most of the energy trade involves fuels, presently, in the past, and in the foreseeable future, as
summarised in Table 1.
Table 1. Short summary of fuel share in world energy utilization.
Year 2000
Year 2020 prediction

2
Primary energya Energy carriers (end use)
90% Fuels
84% Fuels
7% Nuclear
16% Electricity
3% Hydro
Electricity production:
66% Fuels
17% Nuclear
17% Hydro

Primary energyb Energy carriers (end use)

90% Fuels
82% Fuels
5% Nuclear
18% Electricity
3% Hydro
Electricity production:
2% Wind
65% Fuels
10% Nuclear
15% Hydro
10% Wind, solar...
a
18
Gross values: 2400 W/cap (46010 J/yr=10900 Mtoe/yr); population, 6.1109 cap; GDP, 6700 /cap.
b
Gross values: 3200 W/cap (6201018 J/yr=14600 Mtoe/yr); population, 7.6109 cap; GDP, 10000 /cap.
When dealing with world-wide-average energy usage, we must recall how uneven (unfair) the distribution
can be, with a third of mankind presently lacking electricity.

Fuel consumption
Fuel consumption, as fuel price and fuel availability, may be considered as market fuel-properties, and be
jointly dealt with physico-chemical properties of fuels, treated aside, but we have preferred to deal with
separately here.
The substances collectively known as fuels (basically coal, oil, gas, biofuels and synthetic fuels) are
mainly used as convenient energy stores, because of their high specific energy-release when burnt with
ambient air, a most fortunate situation, because a 15-fold (for hydrocarbons; a 34-fold for hydrogen) mass
of air is required to burn a given mass of fuel, and air is freely available everywhere anytime (has not to
be carried on). The burning process, however, is not essential for the release of fuel-and-oxidiser energy;
the same global process takes place in fuel cells without combustion. Fuels, as energy source, are used for
heat generation, for work generation, for cold generation, or for chemical transformations (see details in
What fuels are used for). Fuels are also used for non-burning purposes, as for the chemical synthesis of
materials, mainly polymers (fibres, plastics, cosmetics, pharmaceuticals, mineral oils, etc.), not
considered here furthermore. In a summary, fuels may be used (chronological or difficulty ordering):
1. To produce heat in a burner (thermo-chemical converter). This heat may be used for direct
heating, indirect heating (heat exchangers), for candescent lighting, for feeding a thermal machine
(heat engine, refrigerator, or heat pump) to produce power, cold, or more heat, or for materials
processing.
2. To produce work (and heat) in a heat engine (mechano-chemical converter). This work may be
used to produce propulsion, or electricity, or cold, or more heat.
3. To produce electricity (and heat) in a fuel cell (electro-chemical converter). This electricity may
be used to produce propulsion, cold, more heat, or for materials processing.
4. To produce materials (and heat) in a reactor (chemo-chemical converter); e.g. polymer synthesis,
oils, perfumes...
Fuels may be considered as primary energy (i.e. directly extracted from natural sources and put on the
market), as energy carriers or secondary-energy source (i.e. manufactured fuels such as crude-oil
distillates and synthetic fuels), or as final energy (bought by the end-user for final consumption).
Fuel consumption, both as primary energy (i.e. as found in Nature) and final energy source (i.e. as input
to the end user), is today the major contributor (near 90%) to energy use, both at source and at destination
(up to the Middle Ages, animal power, water-mills and wind-mills were large contributors; in the far
future, nuclear fusion might take over). The analyses of the utilization of: energy as a commodity
(sources, transportation, storage and consumption) is sometimes called Energetics.

Fuels major share in world energy market (80% to 90%) means that the two terms, fuels and energy, can
be used indistinctly both for primary and for final consumption. Beware, however, that some people used
indistinctly 'electricity' and 'energy', without such a rational as above. On the other hand, it is worth
considering that all terrestrial energy (except the minor contribution of gravitational tidal energy) is
ultimately of nuclear origin: nuclear fission inside the Earth generates geothermal energy (also a minor
share of the overall Earth energy budget), and nuclear fusion at the Sun providing the major energy input,
that is partially converted in the short term (weeks) to hydraulic energy and wind energy, in the mid term
(a year) to biomass energy, and in the very long term (million years) to fossil fuels, that is the dominant
commercial source nowadays.
FUELS IN WORLD ENERGY PRODUCTION
Primary energy production (i.e. resource consumption) is computed from the budgets and estimates of
industrial producers:
Coal mines and coal importers
Crude-oil extractors and importers
Natural gas extractors and importers
Nuclear energy generators
Renewable energies: hydroelectric plants, wind-mill fields, solar energy fields, biomass industries,
etc. Most of them are accounted basically by the subsidies they ask for.
The total primary energy consumption in the world (year 2000) was 4601018 J/yr (i.e. 11 000 Mtoe/yr, or
an average of 15 TW in the world, or 2.4 kW average per person). Table 2 presents the distribution by
type of energy source and its time evolution. Traditional energy balances are presented in toe-units
(tonne-oil-equivalent) per year) or other odd units, but the average per unit time (e.g. in 1012 W=1 TW)
seems a more rational rate measure and allows easier comparison with single power devices (e.g. with a
typical nuclear power station of 1 GW). We agree that a year period is a more natural unit of time than a
second, for human activities (who measures salaries in /s?), but conversion errors and encumbrance are
minimised if only SI-units are used, and one may use in most cases the simple approximation 1 yr=30106
s, which is less than 5% below the exact figure, since the data may have not higher accuracy (1 yr=107 s
gives less than 0.5% error). Besides, world averages have less dispersion than local ones (at a given
instant, some places have daylight and others night, some have summertime and others winter). Some
energy unit conversions are presented in Table 3 (e.g. 4601018 J/yr=151012 W=1301012
kWh/yr=11 000106 toe/yr, that divided by 6.1109 people corresponds to 2.4 kW/cap).

Fig. 1. Time evolution of world annual primary-energy consumption: a) in Mtoe, b) in TW. From IEA
http://www.iea.org/.
Table 2. Primary energy consumption in % (year 2000).

Fuels
coal
crude oil
natural gas
biomass (not traded)
Nuclear
Hydroelectric

World
First world
Third world
18
18
(46010 J/yr) (25010 J/yr) (1901018 J/yr)
90
86
95
24
20
29
36
41
26
21
22
20
9
3
20
7
11
2
3
3
3

EU-15
Spain
18
(6510 J/yr) (5.11018 J/yr)
81
85
18
18
43
53
19
13
1
<1
15
13
4
2

Table 3. Some energy unit conversions.

Unit
Equivalence
Tonne oil equivalent (toe)
1 toe
=
42109 J
Tonne coal equivalent
1 tce
=
30109 J
3
Cubic metre of natural gas (STP) 1 m NG =
40106 J
Kilowatthour
1 kWh = 3.6106 J
Thermie (106 cal)
1 thermie = 4.2106 J
Therm (105 BTU)
1 therm = 0.11109 J
3
Barrel of crude (0.159 m )
1 barrel = 6.1109 J
British Thermal Unit
1 BTU = 1.1103 J
Kilocalorie (Calorie)
1 kcal
= 4.2103 J
Quad (1015 BTU)
1 Quad = 1.11018 J
Exajoule
1 EJ
= 1.01018 J
Electronvolt
1 eV
= 0.1610-18 J
Erg
1 erg
= 0.1010-6 J
Notice that renewable energy sources (RES) in 2000, basically hydroelectric and biomass, only amount to
a 10% of world energy coverage (6% in the UE) and all the rest come from exhaustible sources; there is a
firm will however, to come back to a more sustainable exploitation of energy resources, and the objective
is of covering by renewable sources up to 30% of the world energy production in 2020 and up to 60% in
2100 (UE target to 2010 is 12% of RES). Notice that 'resources' refers to the total amount in Nature,
whereas reserves refers to that portion of resources that can be economically recovered at today's
selling prices, using today's technologies and under today's legislation.
Per capita consumption of energy is oddly distributed (more than food, but less than water): the 2.2
kW/cap average comes from 4 kW/cap in EU, 8 kW/cap in USA, and less than 1 kW/cap in the Third
World). It might be compared with the metabolic consumption of 0.1 kW/cap and the averaged Sun input
on Earth of 30 000 kW/cap. Spain primary energy consumption is 5.11018 J/yr=120106 toe/yr=4 kW/cap.
It may be interesting to compare fuel consumption (basically energy) to other basic human needs: world
annual per-capita consumption is some 1000 kg of drinking water, 300 kg of oxygen from the air, 200 kg
of solid food, and 600 kg of coal, 500 kg of crude-oil and 300 kg of natural gas (i.e. 1400 kg of traded
energy-products; more if biomass from developing countries were added).
FUELS IN ELECTRICITY PRODUCTION
An intermediate step in energy utilisation (between primary consumption and final consumption) is
energy transformation into a more useful form of energy, basically electricity production (but also town
gas and coque production) from primary energy sources. In 2000, some 30% of all primary energy was
used in the production of electricity, split by type of primary energy in Fig. 2 and Table 4. Notice that the

5
major source for electricity is the combustion of fuels, in spite that many first-grade books say that
electricity comes from water).

Fig. 2. Time evolution of world annual electricity production (in TWh. From IEA http://www.iea.org/.
Table 4. Electricity production: total and percentage by source type (year 2000).
% in the world % in USA % in EU-15 % in Spain
(1.7 TW=
(440 GW= (300 GW=
(20 GW=
15 000 TWh/yr= 3900 TWh= 2500 TWh= 170 TWh=
541018 J/yr) 141018 J/yr) 91018 J/yr) 0.51018 J/yr)
Fuels
65
73
51
46
coal
37
52
27
39
crude oil
9
3
6
3
natural gas
17
16
18
4
biomass (not traded)
2
2
<1
Nuclear
20
20
34
35
Hydroelectric
15
7
15
19
FUELS IN END-USE ENERGY CONSUMPTION
Final energy consumption is the energy finally consumed in the transport, industrial, commercial,
agricultural, public and household sectors (it excludes deliveries to the energy transformation sector and
to the energy industries themselves). It can be measured by type of final energy (Fig. 3 and Table 5) or by
type of end use (Table 6).

Fig. 3. Time evolution of world annual final-energy consumption (in Mtoe). From IEA
http://www.iea.org/.
Table 5. Final energy use by energy type (year 2000).

Fuels
coal
crude oil
natural gas
biomass (not traded)
Electricity

% in the world
(3151018 J/yr)
84
15
41
17
13
16

% in EU-15
% in Spain
18
(4410 J/yr) (3.81018 J/yr)
79
81
3
3
56
64
26
14
4
21
19

The total final energy consumption in the world is 3151018 J/yr=7000106 toe/yr. The largest share in
electricity generation is by coal (50% world-wide, 40% in EU). Final energy consumption in Spain is
3.81018 J/yr=86106 toe/yr.
Table 6. Final energy use by sector (year 2000).
% in the world % in EU-15 % in Spain
(3151018 J/yr) (441018 J/yr) (3.81018 J/yr)
Industry (construction, manufacturing, extractive)
30
32
35
Transportation
25
31
40
Commerce and services (offices, hospitals)
15
11
10
Residential (home)
20
23
15
Non-energy consumption
10
3
About 20% of world primary energy (30% of the final-energy consumption) is used to power
transportation (1% coal, 90% oil-derivatives, 6% gas, plus 3% electricity). Some 17% of anthropogenic
CO2 emissions also come from transport, being also responsible for some 20% of the projected increase
in both global energy demand and greenhouse gas emissions until 2030. Fossil fuels will continue to
provide the largest share in vehicle power consumption; EU forecast for year 2020 stills base >80% of
that power from fossil fuels, with a rising on natural gas to 10%, supplemented by some 8% renewable
biofuels and some 5% hydrogen (from fossil and renewable sources). Rough average energy consumption
in transportation is:
Per passenger: 3.0 MJ/km by plane, 1.8 MJ/km by car, and 0.9 MJ/km by bus or train. In
equivalent fuel litres per 100 km, the figures are: 10 L by plane, 6 L by car, 3L by bus, and 2 L
by train.
Per tonne of freight: 3.0 MJ/km by truck, 0.7 MJ/km by ship and 0.5 MJ/km by train.
It is appropriate here to quote the CO2 emissions of different transportation means: some 250 g/(kmpax)
for plains (down to 100 g/(kmpax) for the most efficient), some 200 g/km for cars (down to 130 g/km for
new cars in EU from 2012), some 200 g/km for motorcycles, some 80 g/(kmpax) for buses, and some 60
g/(kmpax) for trains.
Nearly 40% of final-energy consumption in UE takes place inside buildings (heating, lighting, cooling
and other appliances).
FUEL-TO-ENERGY CONVERSION FACTORS
Notice that the conversion from material budgets (coal, oil, gas, uranium) to energy budgets is an agreed
standard and not the thermodynamic exergy (the difference is not very large however, except for nuclear
fuels); the lower heating value is chosen for coal and oil, the higher heating valued for natural gas.

7
For nuclear power plants, a standard value of 33% in energy efficiency is assumed, taking no account of
the amount of uranium used, i.e., for an amount Ee of electricity generated, a raw energy of 3Ee is
accounted for (1 MWh0.086/0.33=0.2606 toe). The actual amount of nuclear raw-material used
depends a lot on the technology used; e.g. a given uranium-ore, would yield some 50 times more
electricity if processed in a breeder reactor (where most of the fertile U-238 atoms transform in fissile Pu239 atoms) than if processed in a normal reactor.
Similarly, a standard value of 10% in energy efficiency is assumed for geothermal plants.
Table 7. Recommended conversion factors.
Fuel
Ascribed energy
Coal bit.
0.58 toe/tonne
Coal black lignite
0.32 toe/tonne
Coal brown lignite
0.18 toe/tonne
Crude oil
1.02 toe/tonne
LPG
1.13 toe/tonne
Gasoline
1.07 toe/tonne
Kerosene
1.06 toe/tonne
Diesel
1.03 toe/tonne
Fuel-oil
0.96 toe/tonne
Natural gas
0.090 toe/thermie
Hydraulic energy
0.086 toe/MWh
Nuclear energy
0.2606 toe/MWh
Geothermal energy
0.86 toe/MWh
From IEA (International Energy Agency) data.
Talking about mass-to-energy conversion factors, it is worth mentioning that the main mass-percentage in
fossil fuels is carbon, which burns with oxygen in the air to yield nearly four-times more mass of carbon
dioxide (44 g every 12 g, from stoichiometry C+O2=CO2), so that in crude words, to release the average
energy we trade in the world, per person and year, we are shovelling 1 tonne of carbon from below
ground to the troposphere above us (and that tonne was bonded to some 0.2 tonnes of hydrogen, and is
released bonded to 3.7 tonnes of oxygen); world CO2 emissions in 2005 were 241012 kg (6500 MtC/yr,
from the 11 000 Mtoe/yr of primary energy).
Notice that final-energy consumption must be less than the primary-energy production, because of the
'unavoidable energy losses' in the production and transportation processes (e.g. world production 440
EJ/yr and world consumption 315 EJ/yr). Thermodynamics, however, just says that:
Energy is conserved, in an isolated system, no matter the processes taking place.
Exergy, i.e. energy available for work, in an otherwise isolated system in a given environment,
can only decrease with time in every process.
From that, if the only energy need of humans were comfort heat, cooking power and even sanitary hot
water, this final energy consumption could be met by extracting a much smaller primary energy from
fossil fuels and forcing the rest of the energy to come from the ambient, like in a common heat pump, that
produces three or four times the energy it consumes.
To measure energy-consumption efficiency, two ratios are used:
Per capita energy consumption

Per gross-national-internal-product energy consumption. Third-world countries consume three

times more energy per GNP than developed countries.

Energy management
Energy is a first-need good, as food and water, and its supply has been traditionally managed by public
administrations. The trend, however, is towards a free-market management with political restrictions (e.g.
taxes on fossil energy) and incentives (e.g. subsidies on renewable energies) to procure the following
social guaranties (safety, security, affordability and sustainability):
Reasonable energy safety. Some risks always exist, and society and individuals must establish the
level of acceptable risks (should you carry a gas-lighter in your pocket?, in an airplane?, should
the domestic grid voltage be low or high?).
Reliable energy supply. Some unreliability always exists, and society and individuals must
establish appropriate levels of reliability, knowing that the costs grow exponentially (should a
one-minute electricity-dropout in a commercial store be considered an admissible minor nuisance,
or a great costly disturbance to be protected from?; how wide should the margin in supply voltage
or frequency be acceptable?).
Reasonable energy pricing. Should disperse occasional users (e.g. weekend second-residences)
pay energy (and water, telephone, etc.) at the same price as central city dwellers? Should large
energy consumers pay more or less per unit energy consumed?
Reasonable energy impact on the environment. Some environment impact always exists, and
society and individuals must establish the level of acceptable impact that energy utilisation
(production, transportation and end use) may cause.
There are other aspects related to fuel consumption that have not been considered here: strategic reserves,
strategy to fulfil demand variations, marketing policies, waste management, etc.
On the strategic dependence side, for instance, Europe imports more than 50% of the primary energy it
consumes (in 2000, and it is increasing; in the case of Spain this external dependence is >70%).
On another side, to adapt electricity generation to demand, in view that electricity can hardly be
accumulated, an order of power-plant activation priority is established, with non-storable hydroelectric,
windmill and nuclear plants being always enabled, then cheap-coal and storable-hydroelectric power
stations, and then combined-cycle natural-gas plants, that are more expensive to run. Besides, due to this
changing-load effect, and particularly to the changing-input conditions (low hydraulic year, low winds,
pre-programmed maintenance, unexpected shut-downs, and so on, the design capacity of available power
plants must be larger than the expected average production.
ENERGY FUTURE
The future of energy (as a human commodity) looks dark nowadays, even darker than the future of clean
water and food. The key problem is that energy consumption is growing not proportionally to population
growth (as food may be), but at a much higher rate (because of the 'developed' way-of-life, and new
energy demands from a crowed world, like massive water desalination), with two associated
consequences:
Environmental impact, because the largest share in energy production comes from fuel
combustion, which generates global-warming gases and chemical pollution (global and local),
and other energy sources do not show a clear alternative: nuclear fission has the unsolved
problem of waste fuel and proliferation, and renewable energy sources are not so powerful
neither free of environmental impact (e.g. effects of wind mills on fauna and landscape).

Scarcity of cheap resources, because readily-available oil, gas, and coal deposits, are being
exhausted at a quicker pace than new reserves are found.

As a clear solution to this energy problem is presently not at hand, the most rational approach might be to
push along several fronts, looking forward to solving some of the inconveniences (being alert for new
possibilities), and weighting more on those showing better promise at the time being. In particular:
New fossil fuel plants seem to be unavoidable for decades to come, at least. Cleaner and more
energy-efficient combustion processes must be develop for the traditional fuels, e.g. using
natural-gas combined-cycle plants with a thermal efficiency nearly double than old coal-fired
plants, capturing CO2 emissions from traditional exhaust gases (e.g. using the carbonatationcalcination process), or helped by the oxy-combustion process, or directly from the fuel by
reformation of the fossil fuel to less-contaminant fuels before combustion (what drives towards
the hydrogen economy), etc.
New nuclear fission plants can alleviate in the short term the energy problem, their problem
with nuclear waste perhaps being solved in the future, but their remote risk of massive life
destruction renders them too risky for wide-world proliferation (energy consumption in the
future will increase the most amongst presently underdeveloped societies). Power plants
intrinsically safe to runaway, intrinsically non-proliferating, and making best use of fissionable
material, should be developed. Nuclear fusion research must be further encouraged, as being the
only panacea in the horizon.
New renewable plants must be promoted, even subsidized if one takes account of the social
costs implied in traditional power plants (from human health to world politics), but not as a
present panacea: nowadays, they cannot provide a complete substitute to fossil-fuel plants, nor
in decades to come. Among renewables, the two approaches with wider future are, first, biomass
cultures for biofuels (from non-alimentary plants), and second, thermal solar energy plants,
although wind energy is developing faster, at present.
Perhaps the best summing up is:
Make people aware of this gigantic energy-problem by fostering scientific and social education.
Public acceptance is a pre-requisite in developed societies, without which, economic criteria and
technology availability are powerless.
Make energy economy more explicit (including waste management and health-care costs), for
consumers to minimize the real cost/benefit ratio. Energy seems to be presently too cheap for
people to care about (e.g. when buying powered appliances, or when using vehicles), or too
rigid to allow for sensible choices (e.g. biofuels are not yet alternatives vehicle fossil-fuels).
Invest in basic and applied research on energy management and related environmental impact.
There is no proportion between the ratio of R&D expenses and consumption expenses, between
energy technologies and other technologies.
Meanwhile, diversify the effort according to actual achievements (facts) and reasonable
expectations (expert prospective; without preconceptions, fears and utopias). Avoid being too
enthusiast on a single goal; the best energy diet may be, as for a food diet, variety and
temperance (with green matter being preferable to meat, and paying attention to 'oysters and
lobsters').

References

www.worldenergy.org. World Energy Council (WEC); a UN-accredited non-government, nonprofit organisation.

www.iea.org. International Energy Agency (IEA); a 26-member-states policy-advice cooperative
agency.

10

Back

http://europa.eu.int/comm/energy/index_en.html. European Union's Directorate-General for

Energy and Transport.
http://www.bp.com/statisticalreview. BP is a global energy company.

PYROTECHNICS,

PROPELLANTS

AND

EXPLOSIVES
Pyrotechnics, propellants and explosives..................................................................................................1
Classification .........................................................................................................................................1
Applications: blasters, propellants, airbags, sparklers, matches, fireworks... .......................................2
Explosives .............................................................................................................................................3
Greek fire ..........................................................................................................................................4
Black powder ....................................................................................................................................4
The match ..........................................................................................................................................4
Nitroglycerine and dynamite .............................................................................................................6
Trinitrotoluene ..................................................................................................................................7
Ammonium nitrate ............................................................................................................................7
Propellants .............................................................................................................................................8
Shuttle fuels.......................................................................................................................................8
Ariane fuel.........................................................................................................................................9
References .................................................................................................................................................9

PYROTECHNICS, PROPELLANTS AND EXPLOSIVES

CLASSIFICATION
Pyrotechnics (from Greek , fire) refers to making fire by chemical reaction, with the goal to
produce light, heat, noise, or gases. It is always done by combustion of a fuel and an oxidiser (a red-ox
reaction), but distinguishes from normal combustion in the speed: combustion refers to slow processes
whereas pyrotechnics is associated to almost instantaneous combustion (solid-rocket propellants being
in between).
For pyrotechnics to be effective, fuel and oxidiser must be premixed (double-base pyrotechnics) or,
even better, they should be part of the same molecule (single-base pyrotechnics) with zero or slightly
positive oxygen balance, they should be highly exothermic, and they should be in condense form and
generate a lot of gas. Nitrogen atoms are found in most explosives, because they yield nitrogen
molecules that release great energy and expanding gases (the bond energy of NN is 941 kJ/mol).
In double-base pyrotechnics, the oxidising agents may be nitrates, chlorates, peroxides, oxides,
chromates and perchlorates (the best), all providing oxygen. The reducing agents may be charcoal
(carbon), sulfur, or metal powders. Notice that all practical double-base pyrotechnics are powder solids
mixed-up, with some gluing agent to keep them bounded, because liquid mixtures are too unstable.
By physical state pyrotechnics may be grouped as:
Solids. The majority of cases, because they are more stable and easier to handle.
Liquids. Very unstable even if single-base, as nitrocellulose and nitroglycerine. Separate
liquids, like LH2 and LOX used in cryogenic rockets, are treated as combustion processes.
Gases. There are no single-base pyrotechnic gases (they would decompose), and premixed
explosive gases are considered under normal combustion.
An explosion is a mechanical process generating a destructive high-pressure wave in a fluid; this shock
wave (the blast caused by rapidly expanding gases), and the associate projection of entrained solid

debris, cause mechanical damage by impact (besides other possible associated risks, as fires, toxic
fumes, radioactive waste...). A container with pressurised gas, a confined mixture of premixed
flammable gases, a mist of combustible particles in air, a high-explosive, a high-power electrical
discharge in a solid, a nuclear reaction... all these systems may explode.
By use, pyrotechnics are grouped as:
Explosives. Substances that, by chemical decomposition, generate a supersonic reaction wave,
propagating at several km/s within the material, generating a lot of hot expanding gases. They
are also called high-explosives, and the process is known as detonation; e.g. dynamite.
Sensitive materials that can be exploded by a relatively small amount of heat or pressure are
called primary explosives (e.g. nitroglycerine, lead azide), and more stable materials secondary
explosives (e.g. TNT, ANFO).
Propellants. Substances that, by chemical decomposition, generate a subsonic reaction wave,
propagating at a few cm/s or m/s within the material, generating a lot of hot expanding gases.
They are also called low-explosives, and the process is known as deflagration, as in
combustion; e.g. black powder.

APPLICATIONS:
FIREWORKS...

BLASTERS,

PROPELLANTS,

AIRBAGS,

SPARKLERS,

MATCHES,

According to their purpose, pyrotechnics may be classified as:

Blasters, for mining, tunnelling, demolition, quick-release devices, and weaponry
(warhead). They are high-explosives that undergo supersonic combustion when detonated
by a low-explosive or shock-wave (they slowly burn if just approached by a flame). If the
blast is just to cause an abrupt noise, with insignificant blasting, the device is called a
firecracker (see below).
Propellants, for rockets and weaponry. They generate a large gas stream (like all other
pyrotechnics) that is channelled with one free end to give propulsive thrust to a projectile
or to the combustor body. The main difference between rocket propellants and gun
propellants is the working pressure reached, which in rockets is around 10 MPa, and in
guns more than 100 MPa, with the consequent change in burning rate (recession speed vr is
modelled by Vielle's law, vrpn, with 0.4<n<0.7).
Gas generators, for airbag inflators. Airbags are car-safety-devices that use a pyrotechnic
inflator. They commercially started in the late 1970s in USA and Germany, and widespread
to all new cars in late 1990s. They were based on the combustion of sodium azide (NaN 3)
with an oxidiser, rapidly producing a great quantity of nitrogen gas that inflates a nylon or
polyester bag in some 50 milliseconds (it inflates at 100 m/s); the bag is micro-perforated
to allow progressive cushioning by deflation when the passenger-body hits it. A low
combustion temperature (2000..2300 K) is preferable to avoid massive formation of toxic
CO and NO. Airbag deployment is harsh: it generates some toxic substances (NaOH, a
strong alkali that cause eyes irritation), it generates a bang (some 170 dB, but too short to
break eardrums) and white smoke of talcum powder (used to lubricate the deployment),
and it usually causes burns to passengers, either by friction, chemical attack or high
temperature. In case of accident, beware of un-deployed airbags. If NaN3 is exposed to
water in a landfill, it is converted into hydrazoic acid, which is an extremely toxic, volatile
liquid. Since most airbags will never be deployed and since each airbag contains between
50 and 150 grams of NaN3, concerns have been raised regarding landfill pollution. Less
dangerous pyrotechnics are taking over NaN3, like triazole (C2H3N3), tetrazole (CH2N4)
and derivatives.

Light generators (sparklers), for fireworks or for rescue signals. Some metal powders (Al,
Fe, Zn, Mg, Na) are added to the black powder in order to create bright light (yellow-white
by hot emission at >1500 K) and coloured shimmering sparks (by actual particle burning
and gas line-emission: yellow Na-line, orange CaCl-band, red SrCl-band, green BaCl-band,
blue CuCl-band). Underwater torches use a mixture of high-gassing solid reactives that,
when ignited, creates enough pressure gases to keep water away (they only work at small
depths; a few meters).
Fire generators, for domestic use (matches, fuel pellets for field stoves) or military purpose
(incendiary grenades; thermite, a powder mixture of iron oxide and aluminium or
magnesium dust, was used to spread fire by the splash of liquid iron generated:
2Al(s)+Fe2O3(s)=Al2O3(s)+2Fe(l))).
Heat generators. Aluminium has already been mentioned as an incendiary metal. Other
incendiary metals include zirconium, magnesium, titanium, and depleted uranium. They all
burn at very high temperatures. A particularly useful metallic incendiary is "thermite",
which is a mix of ferrous oxide (Fe2O3, essentially rust) and aluminium. The thermite
reaction is Fe2O3+2Al=Al2O3+2Fe. The reaction burns very hot and releases a tremendous
amount of energy. Thermite is often used in demolition grenades to burn or melt down
military gear that has to be abandoned to an enemy.
Smoke generators, for rescue signals or military purpose (smoke grenades). Smoke from
combustion is an aerosol formed by a suspension of microscopic solid particles from the
poor combustion of carbonaceous fuels. Theatre 'smoke' is just a mist formed by
condensation in the ambient of boiling glycol entrained by an air jet, or the condensation of
water-vapour forced over dry ice or liquid nitrogen.
Noise generators (firecrackers), for fireworks or for rescue signals. Black-powder slowly
burns if in the open, but confined within the paper wrapping of a firecracker, it explodes
(yields a high pressure pulse, but not supersonic combustion).

EXPLOSIVES
CAUTION. The author reminds the reader that the information collected below is intended to satisfy
human curiosity; as for children, it is better to educate on dangers, than to let them explore on their
own. The author is wise enough to keep away of foreseeable dangers, and advises other people to be so
prudent (simple things may kill if put to bad use, like the donkey bone in Cain and Abel story, but
explosives are risky even if put to good use).
Fire was used by humans since 500 000 years ago, and it is known that little explosions may occur in
the fireplace that cause a loud noise and throw sparks away, but control of 'sudden fires' is a very
recent happening. One may find precedents in the incendiary substances developed in the Middle
Ages.
Notice that an explosion is a sudden mechanical process causing rupture and noise, due to great
pressure forces that may be originated chemically (e.g. from a confined combustion), thermally (as in
boilers, even electrically heated), mechanically (as in a balloon or any other gas pressurised vessel),
nuclearly, etc. An explosion is a travelling wave with a sizeable pressure jump across; in air, a very
loud noise (e.g. a close-up turbine) produce an acoustic wave with p<0.1 kPa, nearly hurting the ear;
a 3 kPa jump may break window panes, a 10 kPa jump may throw down people, a 20 kPa jump may
throw down thin walls and a 50 kPa jump thick walls; beyond p=200 kPa the wave becomes
supersonic in ambient air, above p=200 kPa there may be some casualties, and at p=400 kPa more
than 90% people die.

Greek fire
Greek fire is a water-resistant fuel-mix used by the Byzantines. With it, they manage to destroy the
Arab's wooden ships during the siege of Constantinople in 670 a.D. Greek fire was a mixture of pitch,
sulfur, petroleum and quicklime, that burned vigorously and could not be extinguished with water. Its
present-time successor, Napalm (an acronym from naphtha and palmic acid, or from sodium
palmitrate), is a highly incendiary jelly, usually consisting of a naphtha liquid made viscous and sticky
with a thickener: a sodium soap, an aluminium soap, or polystyrene plastic beads). Another incendiary
is FAE (acronym for fuel-air explosives) that sprays out an aerosol cloud of a hydrocarbon liquid, and
then ignites it to create a flaming explosion over a wide area.
Black powder
Black powder may be considered the first pyrotechnic. It was known in China more than 1000 years
ago, and used to make firecrackers and rockets for public entertainment and to frighten enemies in
combat. Black powder knowledge spreaded to the West in the Middle Ages. The English monk Roger
Bacon described a formula for it in 1242; he wrote (in code because of the lethal nature of the
material) that when heating a finely ground mixture of 6 weights of saltpetre with 5 weights of
charcoal and 5 weights of sulphur, a vigorous flame suddenly appears. In the 14th century, black
powder led to the development of fire weapons. Saltpetre is potassium nitrate, a shiny white crystalline
material that could be found on the walls of caves or in well-aged manure piles. Charcoal is pyrolysed
wood, often approximated as carbon, but it is really partially pyrolysed cellulose best approximated by
the empirical formula C7H4O. Sulfur, found in volcanic deposits, decrease the ignition temperature
(and provided additional fuel). Eventually, the formula for black powder was refined to a mix of
saltpetre, charcoal, and sulfur, in the proportions 75:15:10 by weight. A simple approximate
stoichiometry for its reaction is 4KNO3+7C+S=K2CO3(s)+K2S(s)+3CO2(g)+3CO(g)+2N2(g). The
large amount of solids formed (>50 % in mass) makes powder combustion very sooty; this fact, and
the fact that moisture turned some of the soot into a caustic corroding solution (with KOH), demanded
a thorough cleaning of old fire arms for maintenance.
Black powder is an excellent pyrotechnic in many respects. Its raw materials are cheap, abundant, and
reasonably safe: non-toxic ingredients easily shaped, non-detonating (it burns readily, but by
deflagration, though it may explodes under confinement), it can be stored indefinitely if kept very dry,
and it can be easily ignited with a spark; but, on burning, it gives off a dirty smoke with a characteristic
smell. A binder (e.g. a moistened starch or sugar slurry) is used to give shape to the mixture, and the
compound slurry can easily coat a support wire or fill a tube. Black powder, also known as gunpowder,
was used to propel ammunition until late 19th century, when cellulose nitrate (nitrocellulose,
C6H7O2(OH)3+3HNO3=C6H7O2(ONO2)3+3H2O, in concentrated sulfuric acid to get rid of water) was
developed by F.A. Abel in 1865, giving way to smokeless powders (of which guncotton was the first,
followed by cordite) that are used almost exclusively since then.
Magicians commonly use different forms of nitrocellulose (flash paper, flash cotton, flash string) to
produce bright flashes of fire; less dangerous is lycopodium powder, custard powder and even
powdered milk, used by fire-breathing magicians.
The match
Up to the beginning of the XIX c., the usual way to start a fire was by striking a flint-stone with an iron
to get a spark, close to an easily burning material (tinder). In the friction match (from Old French
meiche), the piezoelectric spark is substituted by a low-temperature combustion process initiated by
rub-heating of phosphorus or one of its compounds (pure white phosphorus catches fire spontaneously
in air).

The first trial to make matches is due to R. Boyle, that in 1680 tried to enhance the old friction method
of making fire, by using small wood-sticks impregnated with sulfur (used in black powder) and
phosphorus (just discovered in 1669 by Hennig Brand, who extracted it by evaporating urine to
dryness and distilling the residue with sand, looking for the philosopher's stone), rubbed against
another wood, but it was not reliable. There were other chemically-ignited matches developed, as the
one in Paris, in1805, where a thin strip of wood or cardboard, tipped with a mixture of potassium
chlorate and sugar, spontaneously ignited when brought into contact with sulfuric acid (soaked in
asbestos inside a bottle), but the common friction match was invented by the English chemist John
Walker in 1826 (it is said that he accidentally scraped the stick he was using to mix phosphorus with
antimony sulphide and potassium chlorate over a rough surface and caught fire). Matches were
produced on a commercial scale first in 1833 at Darmstadt (D), and the match box developed in late
XIX c. in USA. The basic design of the match may be split in three parts: igniter (phosphorus in air),
booster (potassium chlorate oxidiser mixed with sulfur fuel) and sustainer, the slow-burning splint to
which the head is glued with a binder (such as gum arabic or wax, also preventing moisture). The
ordinary yellow phosphorous initially used was highly poisonous to match-makers, producing necrosis
of the jaw-bone and mental disorder. The match box widespread with smoker fashion in XX c., until
development of the disposable gas-lighter in 1969. Burning a match releases around 1 kJ of heat.
Phosphorus (Gr. phos, light, and phoros, bearer; ancient name for the planet Venus when appearing
before sunrise) can have several allotropic forms: white, red, violet ... The equilibrium phase at
standard conditions is a transparent (whitish-yellow by impurities) soft crystalline solid named white
phosphorus (M=0.031 kg/mol, =1820 kg/m3, Tm=44 C, Tb=277 C, arranged as tetrahedral units of
four atoms). Phosphorescence is the emission of light (the glow can only be seen in darkness) by slow
oxidation of white phosphorus in air (red phosphorus does not phosphoresce). Transition from white to
red phosphorous may occur by boiling or by exposure to sunlight. Red phosphorus is relatively stable
and easy to handle (it is an amorphous solid wit =2340 kg/m3, that sublimates at 417 C), but white
phosphorus is a deadly poison that spontaneously ignites at room temperature (it is kept under water).
Black phosphorus is produced by heating white phosphorus in the presence of a mercury catalyst, and
it is the least reactive and of least commercial value. Phosphorus is the most abundant mineral element
in plants (0.7%wt), and only second (after calcium) in animals (0.7%; present in every cell, but 85
percent of the phosphorus is found in the bones and teeth). Phosphorus is commercially obtained from
phosphate ores.
Safety matches, first patented by Pash in 1844 in Sweden, do not have the igniter (the phosphorus) in
the same place at the booster (the head); i.e., the match-head must be rubbed against a special surface
(glued on the package of matches) to get ignited. The striking surface is coated with powdered glass
and red phosphorus mixed in a binder. When the match is scratched over it, the heat from friction
causes some red phosphorus to become white phosphorus, which burns spontaneously in air,
decomposing the potassium chlorate, liberating oxygen that quickly reacts with sulfur and lights the
wood of the match, which has been dipped in a fireproofing agent to keep it from burning too easily.
Matches and their boxes are tested to avoid burning problems as exploding heads, spitting heads, splint
afterglow, heat resistance, friction resistance, etc., as well as non-combustion-related problems as
splint strength, tray retention of the box, etc. Safety matches were slowly introduced into the market
because technical difficulties in producing red phosphorus, and the ease-of-use of strike-anywhere
matches.
A few basic safety-rules on how to light a match should be pursued (at least for child education): 1)
Verify there is no flammable substances nearby, 2) Take only one match, 3) Close the cover of the

match box before striking the match, 4) Strike the match away from the direction of the body, holding
the match an arm's length away, 5) Throw a match away only after the flame is extinguished and cool
to the touch (a waste basket is not an ash tray), 6) Matches are fire-making tools, not toys (do not allow
children to play with match-boxes or lighters, even when empty).
Nitroglycerine and dynamite
Explosives may be grouped in low explosives (like black powder) and high explosives (like dynamite).
In low or deflagrating explosives the explosion propagates through the material at subsonic speed
through a sustained combustion process, whereas in high explosives the explosion propagates by a
supersonic detonation.
The first commercial high explosive was found in 1846 by an Italian chemist named Ascanio Sobrero
adding glycerol to a mixture of nitric and sulfuric acids, setting off an explosion that nearly killed him.
Glycerol, or glycerine, C3H8O3 (Fig. 1a) is a syrupy liquid, a by-product of soap manufacture. Sobrero
decided that the liquid he called "nitroglycerine" was dangerous and tried to keep it a secret, but he
failed: a Swedish chemical manufacturer, Alfred Nobel, began to produce nitroglycerine for rock
blasting in 1863.
Nitroglycerine, the nitric acid triester of glycerol, CHONO2(CH2ONO2)2, M=0.227 kg/mol (Fig. 1b), is
a dense oily liquid with a density of 1600 kg/m3 that melts at 80 C and detonates if heated to 218 C
or if subjected to mechanical shock, releasing a lot of gas and energy:
C3H5(NO3)3(s) = 3CO2(g)+(5/2)H2O(g)+(3/2)N2(g)+(1/4)O2(g) + 1,5 MJ.
Nitroglycerine tablets and sprays are used as heart medication (diluted by inert matter and completely
non-explosive). A tiny amount of nitroglycerine (<1mg) placed under the tongue causes blood vessels
to dilate, instantly lowering blood pressure.
Alfred Nobel, began to produce nitroglycerine for rock blasting in 1863. Liquid nitroglycerine is very
unsafe to handle (Nobel's brother Emile was killed while working with it). But he made two major
findings:
He found a stabiliser: nitroglycerine absorbed in diatomaceous earth (a porous clay that
consisted of the deposits of the skeletons of tiny sea creatures laid down aeons before) is stable
to chocks. This material could be packed into cardboard tubes and reliably transported, handled,
and detonated. Nobel called it dynamite. Dynamite was particularly useful in coal-mine
blasting, where the quick, cool flame was not likely to ignite the explosive methane-coal dust
mixture in the tunnel.
He found a detonator: dynamite could not be set off by a spark or a flame (it was even safer
than black powder). In 1865 Nobel devised the first detonator, a blasting cap consisting of a
small black power charge with a cord fuse, to set it off. Nobel's detonator was a significant step
forward in the development of modern explosives technology. Detonators have traditionally
been made from fulminate of mercury, Hg(CNO)2, or lead azide, Pb(N3)2. These are salts of
hydrazoic acid (HN3), which is a dangerous and unstable liquid explosive, and have the
respective formulas of Hg(N3)2, and Pb(N3)2.
Besides dynamite in 1867, Nobel developed in 1875 blasting gelatine (gelatinised nitroglycerine and
nitrocellulose), that can be detonated underwater, and in 1887 ballistite, a smokeless powder (a
mixture of nitrocellulose and nitroglycerine).

Trinitrotoluene (TNT)
The first military high explosive to be put into service was picric acid, C6H2OH(NO2)3, a toxic yellow
crystalline substance used by the French in 1885. However, picric acid has a high melting point,
making the process of filling shells with it difficult; reacts with heavy metals to form very toxic
compounds; and tends to be sensitive.
Another explosive, trinitrotoluene (TNT, trilite, Fig. 1d), was first discovered in the 1860s, but was not
used until the German military adopted it in 1902. It was widely used in World War I. It is relatively
insensitive to chock and can be melted at low temperature to allow it to be poured into bombs and
shells.

Fig. 1. a) glycerol (glycerine), b) trinitroglycerine, c) toluene, d) trinitrotoluene (trilite, TNT).

Ammonium nitrate
Ammonium nitrate (NH4NO3,=1750 kg/m3, M=0.080 kg/mol, Tf=445 K, Tdecomp=470 K) is a
colourless odourless highly soluble crystalline solid, obtained by the reaction between gaseous
ammonia and aqueous nitric acid, and used mainly as a fertiliser and in explosives. In the 1950's
ammonium nitrate derivative explosives were commercialised. Ammonium nitrate is stable if pure,
even in the molten state or in aqueous solutions, yielding NH4NO3=2H2O+N2O when slowly heated in
the open, but violently exploding as NH4NO3=2H2O+N2+O2 when subjected to hot confinement or a
strong shock. Contaminants increase the explosion hazard, particularly if acid. If pure (<0.2%wt
combustible matter within) it is catalogued as a strong oxidiser by the U.S. Department of
Transportation, but for >0.2% of combustible substances as an explosive.
One of the worst blasts ever happened took place on 16 April 1947 on a French freighter docked at
Texas City and loaded with ammonium nitrate fertiliser. The ship caught on fire in the morning and
attracted a crowd of spectators along the shoreline, who believed they were a safe distance away. In
mid-morning the ship exploded, killing hundreds with the tremendous blast, sending a tidal wave
surging over the shoreline, and setting refineries on the waterfront on fire. A second ship loaded with
fertiliser exploded after midnight, but emergency workers were given warning, they evacuated the
vicinity of the vessel, and only two people were killed. Fires burned for six days after the disaster.
Official casualty estimates came to a total of 567. An explosion in Toulouse on 21-09-2001 of some
250 tons of ammonium nitrate in a large fertiliser factory caused 31 deaths and more than 2000
injuries. A truck with 25 tonnes exploded in Spain on 09-03-2004 in a car accident, leaving 2 deaths
and a 3 m diameter crater. ANFO, ammonium nitrate and fuel oil, is the most common explosive
nowadays for surface-mine blasting, having practically replaced dynamite, due to lower cost and safer
handling
In the early 1980s, US Army researchers experimented with "cubane" a cubical C-network
hydrocarbon synthesised in 1960s of double the density of gasoline; its high density means faster
propagation of breakdown reaction, leading to a more powerful and compact explosive.
Octaninitrocubane consists of a cubic core of eight carbon atoms, with an N2O group attached to each
corner of the cube; it seems to be very stable, twice as powerful as TNT and its breakdown products
are non-toxic carbon and nitrogen compounds.

PROPELLANTS
Propellants are used to force an object to move forwards, from a firework-rocket, to a rocket to put
people on the Moon. As propellants must be carried aboard, condense phase propellants (solid or
liquid) are much preferable.
Liquid propellants may be grouped as:
Single base liquids, such as hydrazine (N2H4(l)) and derivatives (as mono-methyl-hydrazine
MMH N2H3CH3(l)) and unsymmetrical-dimethyl-hydrazine UDMH N2H2(CH3)2(l))), or
hydrogen peroxide (H2O2 more than 70%wt concentrated), that are stable at ordinary
temperatures, but decompose into hot gases when exposed to suitable catalysts or high
temperatures (e.g. 2N2H4=2NH3+N2+H2).
Double base liquids, stored apart. As oxidiser, liquid oxygen (LOX) or nitrogen tetroxide (NO4)
may be used. As fuels, liquid hydrogen or kerosene may be used.
Solid propellants (they do not allow intermittent operation) are categorised as "single-base", "doublebase", and "multi-base" or "composite" powders or pastes:
Single-base powders, such as cordite, are mostly nitrocellulose (guncotton with some
nitroglycerine). Black powder was used for solid-propellant rockets in the first third of the 20th
c. They burn cool and cause little barrel wear in firearms. Double-base composite powders are
now generally used as a propellant for solid-fuel rockets.
Double-base powders are mixtures of nitroglycerine and nitrocellulose such as ballistite, and
burn hotter.
Composite powders are modern formulations that do not contain nitrocellulose or
nitroglycerine, instead using more modern propellants that burn cool but are as powerful as
double-base powder. Ammonium perchlorate composite propellant (APCP), one of the best
solid rocket fuels, can be cast into shape (instead of pressed as a powder), making its setup
more simple and reliable.
Hypergolic propellant are substance pairs that ignite upon contact with each other without a spark, heat
or other external aid, such as N2H4/N2O4, N2H4/NO4, aniline and red fuming nitric acid (concentrated
nitric acid in which nitrogen dioxide is dissolved), etc.
Shuttle fuels
The Shuttle (the Space Transportation System) has, for take-off, two solid rocket boosters (2 minutes
burn) and three liquid-fuelled main engines, providing in total up to 30106 N thrust (80% solid
rockets, 20% liquid rockets) to lift the 2106 kg total mass. Besides, the Orbiter (what most people call
the Space Shuttle) has two propulsion systems: OMS (Orbital Manoeuvring System) used to change
orbit and to return to earth, and the RCS (Reaction Control System) used for station-keeping and
attitude control; both systems burn liquid hydrazine with cryogenic oxygen.
Each solid rocket is 45 m high and 3.7 m in diameter, has 500 000 kg of fuel (590 000 kg total mass)
and yields 13106 N thrust during the first few seconds, slightly decreasing during the 2 minutes burn.
Their solid propellant consists of a mixture of aluminium powder (fuel, 16%wt), ammonium
perchlorate powder (NH4ClO4, oxidant, 70%wt), and a dash of iron oxide powder (catalyst, 0.4%wt),
held together with a polymer binder (12%wt) and epoxy resin (curing, 2%wt). The burning reaction is
basically NH4ClO4=2H2O+0.5Cl2+0.975O2+0.25N2+0.25N2O+167.4 kJ (lower heating value). The
white cloud formed at launch is due to alumina particles. Russia uses liquid-fuel rockets also as
boosters.

The liquid rocket engines use liquid hydrogen and liquid oxygen contained in the external tank, the
largest element of the Shuttle (8.4 m in diameter, 2000 m3 of cryogenic propellants with a multilayer
thermal protection 25 mm thick). Hydrogen is supplied from a 1450 m3 bottom tank (106 t of LH2 at
20.6 K and 300 kPa, =73 kg/m3), at a rate of 210 kg/s (3 m3/s), while oxygen is supplied from a 540
m3 top tank (630 t of LOX at 90.6 K and 250 kPa, =1150 kg/m3), at a rate of 1300 kg/s (1.1 m3/s),
both through 0.43 m diameter feed lines, during the 8.5 minutes burning time, positioning the Shuttle
at 115 km height. Notice that for 1300 kg/s of LOX, the stoichiometric LH2 is 1300/8=163 kg/s, but
210 kg/s is supplied, this excess hydrogen being used as coolant to protect the engine. Each of the
three rockets yield up to 2106 N thrust.
Ariane fuel
Ariane 5 has two solid boosters and two liquid-propellant stages, with a total thrust of 11106 N and a
total take-off mass of 750103 kg. Each booster is 31 m long and 3 m in diameter, uses 240103 kg of a
PCA + HTPB + aluminium solid mixture, yielding 5106 N thrust during 130 s (up to 55 km). Stage 1
is 31 m long and 5.4 m in diameter, using 160103 kg of LH2/LOX and yielding 1106 N thrust during
580 s; the combustion chamber is at 10 MPa. Stage 2 is 3.3 m long and 4 m in diameter, using 10103
kg of monomethyl hydrazine / nitrogen tetroxide and yielding 27103 N thrust during 1100 s.

REFERENCES
http://en.wikipedia.org/wiki/Pyrotechnics.
R.E. Reish, 1996, "Airbags", Scientific American, 116.
J.A. Conkling, 1996, "Pyrotechnics", Scientific American, 102.
(Back to Combustion)

COMBUSTOR CHARACTERISTICS
Combustor characteristics ............................................................................................................................. 1
Inlet flow (intake)...................................................................................................................................... 2
Dual fuel combustors ............................................................................................................................ 3
Internal flow .............................................................................................................................................. 4
Outlet flow (exhaust) ................................................................................................................................ 4
Heat and work flow ................................................................................................................................... 4
Power intensity...................................................................................................................................... 5
Efficiency .............................................................................................................................................. 5
Power modulation ................................................................................................................................. 5
Combustor internals .................................................................................................................................. 6
Type of flames ...................................................................................................................................... 6
Type of geometry .................................................................................................................................. 6

COMBUSTOR CHARACTERISTICS
A combustor is the equipment where combustion takes place, usually a combustion chamber, and it can
be a complex system that is best studied by considering its main parts and processes separately. The
different types of combustors were reviewed in Combustion characteristics; we here intend to give a
general idea of the basic parts of all combustors.
The analysis of a combustor may be, in order of increasing generality:
Specific combustor types: steady and unsteady combustion chambers, and practical burners.
Generic combustor subsystems and processes: intake, injectors, igniters, energy flows, exhaust,
safety and controls.
Basic combustion processes: combustion at rest, non-premixed flame, premixed flame
(deflagration wave, detonation), pollutant generation.
Fundamentals: Chemistry (in particular Chemical Kinetics), Thermodynamics (in particular
Chemical Thermodynamics), Heat and Mass Transfer, Fluid Mechanics (in particular gas flow and
turbulence)...
Ancillary sciences (needed for the modelling and validation of theories and practical design):
Mathematics (analytical and numerical modelling) and experimentation (test rig preparation,
instrumentation, data acquisition and analysis).
The practical goal in the analysis of combustion devices is the prediction of its performance (for present
and future combustors), in terms of the multiple physico-chemical phenomena involved; it is thence a
prerequisite to analysis the latter.
We try to focus here on combustion basics, i.e. consider a generic combustion system as used in
applications (Fig. 1), but analyse mainly its fundamentals. The objective is to understand the influence of
the main parameters in combustion, in terms of predictive models (not just descriptions of particular
cases).
Q
Inlet
Combustor
Environment

Outlet

F
O

P
W

Fig. 1. a) Generic steady-state combustion chamber and its interactions. b) unsteady chamber in
reciprocating engines (F: fuel, O: oxidiser, P: products, Q: heat, W: work).
It is good to always keep in mind the steady-state combustor with one inlet and one outlet (a simple
control volume approach), and consider other cases as variations of this simple scheme: the enclosure
may be imaginary as in a simple candle flame, the inlet duct may be actually split in two (one for the fuel
and another one for the oxidiser), there may be neither inlet not outlet (in reciprocating engines with
valves closed), the outlet may (theoretically) be considered split in as many ducts as chemical species are
(to provide a common thermochemical reference to each species), etc.

Inlet flow (intake)

What to study:
Fuels. Types, availability, price, convenience (better if fluids), additives (water, octane index
promoters, lubricants, stabilisers). A detailed description of fuels can be found in Fuels.pdf.
Oxidisers. Types: ambient air is the most used because it is free, but preheated air, enriched air (up
to pure oxygen), depleted air (mixed with exhaust recirculation), and other oxidisers (as for rocket
propulsion) may be used.
Mixture ratio specification. There are several ways to specify the mixture ratio: fuel molar-fraction
in the mixture, xF; fuel mass-fraction, yF; fuel/air ratio (molar or mass), f; air/fuel ratio (molar or
mass), A; air ratio (to stoichiometric air for constant fuel flow; also known as aeration or simply
'lambda'; it the same in molar and mass basis), =A/A0; equivalence ratio (the fuel/air ratio relative
to stoichiometry; it is the same in molar and mass basis), =1/; excess air, e=-1; mixture (massflow-rate) fraction, etc. (Note that sometimes ; is named excess air ratio or even excess air,
instead of air ratio). Most industrial burners and SI-engines use near stoichiometric mixtures,
whereas CI-engines and gas turbines use very lean mixtures. SI-engines using three-way afterburning catalysers must use stoichiometric mixtures (=1) for proper operation, but homogeneouscharge SI-engines may work up to <1.6, a range that is extended in stratified-charge SI-engines
(direct gasoline injection may work up to <3, and direct natural-gas injection up to <10 of
overall mixture ratio).
Mixture preparation. It depends on the type of combustion wanted: premixed, or non-premixed.
Fuel and oxidiser must be pumped for normal pressurised combustion (they are aspirated in
depression combustors). Liquid fuels must be quickly vaporised within the oxidiser (only vapours
burn). The liquid fuel may be supplied at high pressure into the combustion chamber in order to
quickly form a burning spray, or vaporised into the oxidiser stream, either far upstream in the
carburettor of a classical premixed-combustion engines, or at the intake manifold as in most
current Otto engines, or within the cylinders (direct injection). In Otto engines, indirect fuel
injection at the manifold may be at 300..400 kPa, whereas direct injection in the cylinder at the end
of the compression phase may be at 3..10 MPa for direct injection spark ignition Otto-engines
(DISI) and 30..200 MPa for direct injection compression ignition Diesel-engines (DICI). Most
modern injection systems are electronically controlled. Lean pre-vaporised premixed combustion
(LPP) is under study to reduce the temperature and NOx emissions in gas turbines. For coal chunks
and other solid fuels burning over a grate, the best in mixture preparation is to split the air intake in
two streams, one for under-fire air, and the other for over-fire air.

3
Chemical notation (particularly hydrocarbon chemistry) and stoichiometry (the proportions in which
atoms are combined in molecules) must be considered, and distinction made between stoichiometric
equations (e.g. C+O2=CO2 or C+(1/2)O2=CO) and actual mixture equations (e.g.
C+aO2=bCO+cCO2+dO2 or aCuHvOwNxSy+aA(c21O2+c79N2)+cH2O+dO2=(xiMi), depending on the
choice of unit-fuel-amount or unit-exhaust-amount of substance, for a given oxidiser/fuel ratio, a in the
first instance or A in the second. Actual mixing, by non-premixed feeding, convection, evaporation and so
on, may be considered also here (e.g. for spark engines: carburetion, intake-port injection or directcylinder injection).
DUAL FUEL COMBUSTORS
Combustors have been traditionally classified according to the type of fuel they were designed for, and
for their size, of course. For instance, a gasoline spark-ignition engine cannot be fed with any other fuel;
really, some amount of a compatible fuel may be added, as when up to 20% bio-ethanol is added to
normal gasoline. And similarly for a kerosene-fuelled gas turbine or any other internal combustion
application. Of course, external combustion applications, like a boiler and some type of furnaces, may run
with heat produced by the combustion of any kind of fuel, but the hearth where the combustion takes
place is also fuel-specific, within small variations.
But there has always been a push to provide a dual-fuel capability, initially to be able to burn new fuels in
old systems (coal in wood grates, oil in coal hearths, natural gas in oil burners, etc.), later to cope with
fuel restrictions (gasogens during the wars), and nowadays to limit pollutant emissions (e.g. city fleets of
taxis and buses on LPG and NG, clean diesel oil for port and coastal operations of ships, and cheap dirty
fuel-oil for deep-sea operation, etc.). We do not consider under the dual-fuel heading the capability to use
a fuel and electricity (as a second energy source).
Dual fuel combustors can switch to a different fuel by just manually or automatically turning a control
valve; of course, any combustor prepared for any fuel might be fully modified to run on any other fuel,
but this is of little interest, except when an existing combustor is to be definitively transformed, as when a
fleet of city buses is transformed from diesel-oil to natural gas (to reduce pollution) and the modified
engines cannot work with diesel-oil any more.
The common dual-fuel systems in use are:
Otto dual-fuel engines, where a gasoline engine system is modified to run also on propane (LPG
really) or methane (really compressed natural gas CNG typically at 20 MPa, or liquefied natural
gas LNG from a cryogenic reservoir). The changes required are: the additional fuel tank (to be
refilled in situ or exchanged and refilled off-line), the fuel lock valve, a liquid-to-vapour
converter (if any), a gas mixer or a gas injector device, and the switching valve for the driver
(most times an additional fuel storage meter is added). The engine itself is not modified, or very
little, since the combustion process is similar or even better (e.g. LPG and NG have larger octanenumber than gasolines). Otto dual-fuel engines switch from one fuel to the other by a simple
feeding switch, manually or automatically operated.
Diesel dual-fuel engines, where a diesel engine system is modified to run also on CNG or LNG.
The problem here is that the high-pressure gas-fuel-injection would consume a lot of power
(some 30% of the heating value of the gas), what might be reduced by compressing LNG if
available, or, for most practical systems, by feeding the gas at the modified intake manifold and
using some diesel oil in the normal way to ignite the mixture (always), and (in most cases) to
complement the required fuel load (up to 20% diesel oil plus more than 80% natural gas is the
common approach). Dual fuel diesels presently run only in the 4-stroke mode. Notice that
sometimes the term dual-fuel diesel engine refers to the capability to run on light and heavy fuel

oils. Notice also that most diesel-oil engines would run on biodiesel with none or minimal
change, and even on straight vegetable oil with some modifications. A control electronic unit is
required to adjust the dual-fuel supply, and some liquid fuel is always required to quickly ignite
the mixture at each cycle (engine start-up is done with 100% diesel oil).
Brayton dual-fuel engines, i.e. gas turbines running on kerosene or natural gas. Most dual-fuel
systems work in the normal non-premixed mode, but some lean premixed combustion systems
(vaporising the liquid fuel before combustion) are also been developed.

External combustion applications running on dual fuel are basically a dual burner within the same
combustor. A typical dual-fuel external-combustion applications is the steam turbine in a classical LNG
carrier ship, where the natural-gas boil-off is burnt in the boilers, besides the normal heavy-fuel oil.

Internal flow
Internal flow field depends a lot on the type of combustor. In modern combustors, recirculation of exhaust
gases (EGR) is a common practice. In all practical cases, the internal flow is highly turbulent, and many
times with forced swirling at the injectors. Typical mean flow velocity inside combustion chambers of gas
turbines are in the range 15..25 m/s for industrial gas turbines and 25..50 m/s for aeronautical turbines.

Outlet flow (exhaust)

What to study here: types of exhausts conditions according to heat transfer (from isothermal to adiabatic),
exhaust gas composition (complete-oxidation model, chemical-equilibrium model with
(xip/p0)i=exp(K(T,p0)), possible condensation, emission of contaminants) and exhaust temperature.
Environmental effects are usually identified with pollutant emission through the tail-pipe because this is
the major source, but there are others (see below). The use of exhaust scrubbers and exhaust catalysts are
crucial for pollutant control. Emissions from cars, the major pollution source in cities, have realised a
unitary reduction (per car) of 50% from 1970 to 1990, and a drastic reduction around 1992 (another 40%
from 1990 to 2000). A detailed description of pollutant emissions and fire safety can be found in
Environmental effects and hazards in combustion.pdf.
Closely related to the exhaust ducting is the chimney, the vertical part of the exhaust piping intended to
create a natural draught and a far disposal of pollutant emissions, used in stationary and large mobile
combustors.

Heat and work flow

What to study: energy balance for the combustor: dE=dQ+dW+Hi*dni, amount of heat transfer (from
zero to the calorific value), amount of work transfer (from zero to the maximum value: the exergy of
reaction). Heat transfer mechanisms in combustion, the main one being convection but radiation effects
sometimes crucial. Heat sinks in combustion: walls, heat exchangers, heating of a furnace load.
Combustor walls usually have to withstand high temperatures, like within furnaces, and thence an inner
lining of refractory material (fire bricks or monolithic) is applied.
Notice that heat and work are path variables, being different for constant-volume combustion than for
constant-pressure combustion, for the same fuel/oxidiser mixture. Furthermore, it is sometimes wrongly
assumed that a given chemical reaction always gives more heat than work; it might yield even more work
than the maximum heating value, as for instance with C+O2=CO2, that has a maximum heat potential at
standard conditions of 393.6 kJ/mol (-hr) and a maximum work potential of 394.5 kJ/mol (-gr).

5
POWER INTENSITY
As for other engineering processes, the main interest is not on integral heat and work, but on heat and
work flow-rates, i.e., on combustion intensity, either per volume of combustor (in kW/m 3), or per crosssectional area perpendicular to the main flow in the combustor (in kW/m2). Combustion intensity is the
product of mean reaction rate times fuel heating value; to increase the former, the area being burnt must
be increased (e.g. by turbulence), the temperature must be increased (e.g. inside a porous matrix burner),
or a catalyst must be used (catalytic combustion may be used to stabilise low temperature flames instead
of increasing their intensity).
EFFICIENCY
Different kinds of efficiencies can be defined:
System efficiency. It is usually defined as the ratio of useful energy extracted from the system (a
boiler, an engine) to the available heating value of the fuel supplied (care must be paid to find if
the HHV or the LHV has been chosen). Typical values are 90% for a boiler and 40% for an
engine; the rest of the heating potential goes as enthalpy of exhaust gases, or as heat loses to the
environment through the walls. These are just energy ratios that do not take into account
thermodynamic limits, an exergy ratio being sounder (the exergy extracted, per unit of fuel exergy
input, is some 30% for a boiler and 40% for an engine).
Combustion efficiency. It is usually defined as the ratio of actual heat released with intake at
standard conditions and exhaust transported to standard conditions at constant composition, to the
maximum heating value of the fuel (roughly the HHV); typical vales in well-operated combustors
is >98%. Real combustion processes do not yield their maximum heating value because of unburnt
emissions (soot, CO, volatile organic compounds) and condensables. Combustion efficiency in
hydrocarbons practically coincides with the carbon conversion fraction from fuel to CO2, since
hydrogen oxidises more readily, but nitrogen oxides may be present too. Sometimes, a boiler
efficiency is called combustion efficiency, but the numerical values usually clarify the distinction.
Combustion efficiency in 'adiabatic combustors'. In a fix wall (i.e. without work exchange)
combustion chamber intended to be adiabatic (like in a gas-turbine combustor; i.e. without a
secondary on-purpose heat-extracting fluid), combustion efficiency is defined as the fraction of the
LHV that exits as thermal enthalpy (sensible and latent) of the exhaust stream, with the 'lost
fractions' being the residual chemical energy of unburned components in the exhaust, and the heat
loses through the wall. Notice that this definition corresponds to a system efficiency if one
considers the combustion gases themselves as the energy-extracting medium (like the heated water
in a boiler).
In any case, the usual assumption in most combustors is that the heat lost through the external combustor
walls is negligible against the enthalpy flows involved.
POWER MODULATION
Combustors are usually optimised to work at a nominal load, but are usually required to work at different
loads. Non-premixed flames are stable in a very wide range of flow rates, and thus of power, but the
more-powerful normal premixed flame gets unstable below 50% of full load (flash-back extinction
occurs, and blow-off extinction in the upper limit). Outside of this flame-stability range, either a nonpremixed flame is allowed (with size increase, and CO and NOx increase), or some burners must be
switched off (causing uneven temperatures and start-stop increased emissions), or new porous combustors
developed...
For instance, consider the case of home boilers. They usually provide space heating and hot water. Hot
water (in line) requires a minimum of 15 kW for small periods (less than 15 minutes), whereas a 100..150

6
m2, 4-person family home in mid-latitudes may require from 2..15 kW of heating, depending on wall
insulation and external temperature. The normal premixed flow, the one used in traditional water heaters,
yields an unstable flame below 50% of full load, thus current home boilers may be modulated between
7..15 kW (heating is temporary stopped when hot water is needed), matching the demand by an on/off
control that increases pollutant emissions at every start. On the other hand, for better comfort on hot water
demands, i.e. if more than 15 kW were needed (e.g. two showers at a time), if intermittent user-demands
(e.g. at the sink tap) are to be smoothed, and if startup-times of the boiler are to be skipped, the best is to
provide a hot-water storage tank of some 50..100 litres, at the expense of its continuous heat loss.

Combustor internals
What to study: description of types of burners (see Combustor types), types of flames, mechanisms for
ignition, stabilization and extinction; and analytical models for internal processes and kinetics (flow and
chemistry), to evaluate the combustion rates and required sizes and times. Chemical kinetics and physical
transport processes may become too complex if the flow field is not drastically simplified. Sometimes
measurement and control systems are here considered (e.g. flame detectors), but instrumentation is also
applied to the intake to regulate the flow and to the exhaust to monitor emissions.
TYPE OF FLAMES
According to the type of flame, combustion processes may be grouped as:
Non-flame combustion. Usually identified with low-temperature catalytic combustion.
Non-visible-flame combustion. Usually identified with porous media combustion, although there
are also nearly-invisible flames, as premixed hydrogen/air or methanol/air...
Non-premixed combustion, usually divided in:
Homogeneous combustion of gaseous fuels, as a butane jet in air.
Homogeneous combustion of condensed fuels, as gasoline vapours in air.
Heterogeneous combustion of condensed fuels, as for coal burning and solid charring.
Premixed combustion, usually divided in:
Unsteady process: ignition (local or global light-off), extinction (effect of air/fuel ratio,
quenching).
Steady process: deflagration speed, anchored and travelling flames.
Chemical kinetics is perhaps the hardest theme to study, due to the intricate mathematical models (very
stiff problems because of the widely different orders of magnitude involved) and the scarcity of reliable
experimental data. Topics to be studied include: speed of chemical reactions, law of mass action,
elementary reaction mechanism, first-, second- and third-order reactions, partial-, steady- and frozenequilibrium reactions, initiation-, propagation-, branching- and termination-reactions, Arrhenius law, etc.
TYPE OF GEOMETRY
Concerning the geometry of the combustion process, the traditional division in combustor modelling is:
Uniform combustion model, as for still-reactors and well-stirred reactors.
One-dimensional combustion models, as for plug-flow reactors and spherical combustion.
Two-dimensional combustion models, as for planar and axial jets.
Three-dimensional combustion models, basically generic CFD codes.
We have focused here only on the characterisation of combustors from the combustion point-of-view, not
considering other important aspects of combustors as their construction, operation and maintenance
details, which depend a lot on the particular application envisaged: furnace, boiler, turbine, etc.

7
For instance, in jet-engine combustor, an outer case holds the pressurised gases against the low external
ambient pressure, and an inner perforated liner holds the hot burning gases while providing primary air
sustain the flame, and secondary air to dilute the burnt gases and lower the temperature some 500 K
before entry to the turbine; in spite of being cooled by secondary air on both sides, special thermalresistance alloys are still required for the liner. Fuel is injected behind the leeward dome of the liner
through a swirler to accelerate mixing (liquid fuel gets atomised, and air flow gets more turbulent;
however, the higher the turbulence, the higher the pressure loss in the combustor).

Fig. 2. Diagrams of components and flow paths in a gas turbine combustor (from Wikipedia).
Back to Combustion

ENVIRONMENTAL EFFECTS AND HAZARDS IN

COMBUSTION
Environmental effects and hazards in combustion ....................................................................................... 1
Fire safety. Protection and remediation .................................................................................................... 2
Fire safety education ............................................................................................................................. 2
Fuel detection ........................................................................................................................................ 4
Fire detection......................................................................................................................................... 5
Smoke detectors .................................................................................................................................... 5
Water sprinklers .................................................................................................................................... 5
Inert atmospheres .................................................................................................................................. 6
Fire-breaks and fire extinguishers ......................................................................................................... 6
Pollutant emissions ................................................................................................................................... 7
Emission quantification......................................................................................................................... 8
Fuel tank and crank-case ventilation ..................................................................................................... 9
Exhaust emissions and pollution ........................................................................................................... 9
Vibrations and noise............................................................................................................................ 14
Electromagnetic interferences ............................................................................................................. 14
Safety management ..................................................................................................................................... 14
Risk: a combination of hazard and damage ............................................................................................ 14
Emergency response ............................................................................................................................... 15
Analysis of accidents .............................................................................................................................. 15
Prevention of accidents ........................................................................................................................... 15
Glossary of terms: accident, exposure, fear... ......................................................................................... 16

ENVIRONMENTAL EFFECTS AND HAZARDS IN COMBUSTION

Combustion is a hazard, and, besides the many services it provides to humankind, it may cause nuisance
(e.g. noise, smoke), damage to property (deformations, loss of strength, burnings and explosions), and
damage to people (injury and loss of life). Damage may be ranked in top-down severity order as: a) loss
of life (mortality), b) loss of health (morbidity), c) loss of property and d) loss of activity.
Besides damage caused by the combustion process itself, there is also damage associated to the
management of fuels (and oxidisers, if special). Coal handling produces respiratory hazards, and crude-oil
derivatives are carcinogenic. Liquefied petroleum gases, LPG, and particularly cryogenic fluids like
LNG, may cause severe frostbite and structural damage (carbon- and low-alloy steels show a marked
ductile to brittle transition at freezing temperatures).
Damage may be caused to individuals or to the environment in general. A general summary on safety
management is included below; here, some techniques to minimise risks associated to combustion are
studied.
Combustion is a physico-chemical hazard, and, to minimise its impact, one has to be aware of it, rely on
safer fuels (e.g. diesel instead of gasoline, natural gas instead of butane), reduce unnecessary fuel stores,
avoid fuel leakages and provide fuel detectors, reduce uncontrolled ignition-sources (sparks and hot
spots), decrease the impact of the controlled combustion processes (emissions), and plan for the best
rescue actions in case of uncontrolled combustion (fire detection and fighting). As usual, pyrotechnics
hazards (from amusement firecrackers to weaponry) are not considered under the combustion heading.

A quick-summary of hazard types associated to combustion may be:

Physical hazards: mechanical (explosion), thermal (excessive heat, out-range temperatures),
radiation (blinding flare).
Chemical hazards: oxygen depletion, gas poison, aerosols, liquid poisons, solid poisons.
The effects of combustion on the environmental may be grouped in two categories:
Sudden uncontrolled events (combustion accidents). This impact is fought with proper prevention
(minimising risk and educating people) and proper fire fighting (smoke detectors, automatic firesprinklers, portable fire-extinguishers, fire-fighting brigades).
Continuous degradation, due to pollutant emissions during normal combustion. This impact is
fight by developing better fuels (sulfur-free diesel, unleaded gasoline), better combustion
processes (fluidised bed, porous media) and post-combustion treatment (catalytic conversion).
Only a descriptive view of the subject is here presented (some theoretical insight can be found on
Combustion Kinetics). Thermal effects on materials in general, which may be originated from a
combustion process, are presented aside. We have tried to follow a top-down approach in the analysis of
the environmental effects of combustion, i.e. from deadly explosions to inconvenient electromagnetic
interferences.

Fire safety. Protection and remediation

Fear of the fire has always frightened humans. Risks must be prevented first (minimising risk), their
consequences foreseen (minimising damage), their occurrence appropriately handled (early warning,
gateways, fighting), and the consequent damage allocated in justice.
Contrary to common sense, there are many fires that are not fortuitous accidents but directly caused by
irresponsible people, from children left playing with matches, to arson. People are the primary cause of all
accidents. Again, education on fundamental principles, and not only instructions about particular items, is
the best preventive measure.
Fire protection can be achieved by passive means (using fire-resistant materials), or by active means (fireextinguishers). The most effective protection is based on proper fire safety education.
Minimising risk: fire safety education, inert atmospheres, smoke detectors and fire-doors.
Minimising damage: fire-breaks, fire sprinklers and fire extinguishing.
FIRE SAFETY EDUCATION
Education is the best prevention. For every identified risk, plan a fire safety procedure (e.g. get a hose or a
water bucket nearby when making an outdoor fire), and rehearse it (e.g. evacuate multi-store buildings
once a year, even at a pre-established time; it teaches people and safety personnel a lot). Clean habits
already help to prevent fires, since accumulation of waste (packing material, oily rags, rubbish) and even
dust may be a fire hazard (old stores must be kept tidy, too!). Smoking is reported to be a major cause of
fire accidents.
Hazard types to humans (for goods there are only two concerns: either destruction by fire, or fire
promotion):
Toxicity of emissions, i.e. poisoning. The main toxic gas in combustion is CO; xCO>1% for 1 minute
is mortal (maximum permissible 50 ppm); other toxic gases may be SH2 (xSH2>0.1% for 1 minute is
mortal), SO2, CNH, HCl, NO2, NO, NH3. As a rule, hydrocarbon vapours in concentration >0.1% in
volume (1000 ppm) cause eye irritation after some minutes, whereas >1% ppm produces

unconsciousness after a few minutes. Smoke, i.e. particles and droplets aerosols, cause eyes and nose
irritation and orientation lost (panic).
Anoxia, i.e. insufficient oxygen (asphyxia means lack of pulse). Normal air has xO2=21%; above
xO2=18% normal breathing can be sustained, but below xO2<15% muscle fatigue is noticeable, at
xO2<10% reasoning stops, at xO2<5% consciousness stops. Breathable oxygen concentration range is
19.5%..22%; if <19.5% work with oxy-mask, if >22% risk of fire. Monitored by electrochemistry
(ZrO2). CO2 is not toxic, just deprives of oxygen: adding >10% CO2 to air is mortal after 1 minute.
Burning, i.e, dehydration and ablation of tissues, may range from mild skin burns to incineration,
usually graded as first-degree burns (skin surface painful and red, a small blister), second-degree
burns (large blisters on the skin), third-degree burns (destruction of both epidermis and fatty dermis;
dry and painless), and calcination. Burning may be due to hot objects, a very hot environment, or
direct flame contact: superficial skin burn (1st and 2nd degree) takes 6 hours at 45 C but 1 second at
70 C. Monitored by explosimeters (oxidation in a hot Pt-wire): ok if <10% LEL (), if >10%&<20%
be on alert, if >20% leave the room.
Suffocation and hyperthermia, i.e. inability to evacuate metabolic heat to the environment, resulting in
heat exhaustion and heat stroke. At 40 C and 100% RH death follows after a few hours (but a steamsauna, Turkish bath, at 40 C for a few minutes may be healthy; at 50 C and 100% RH death follows
after a few minutes). At 120 C, even in the driest, death follows after a few minutes (but a dry sauna,
Nordic sauna, at 90 C and less than 10% RH for some minutes may be healthy).

Each fire is unique, but most domestic and industrial fires (explosions not included) have the following
common characteristics:
tbuildup=105 minutes, i.e. a build-up period of some 10 minutes after ignition, while sufficient
heat is released for a thermal runaway; suddenly, oxygen is locally depleted and CO and HCN
abound. The spreading of fire over a quiescent liquid-fuel pool (e.g. oil spillage) at room
temperature is at some 0.1 m/s, and some 0.01 m/s over solid fuels.
tmax heat=3010 minutes, i.e. the maximum heating taking place 3010 minutes after ignition
Tbuildup=1100100 K, i.e. a maximum heating (maximum temperature relative to the ambient) of
1100100 K, what may be explained in terms of the fuel/oxidiser characteristics.
Tresidual=700100 K, i.e. a residual slowly-decreasing heating of Tamb+700100 K, after one and a
half or two hours after ignition.
The ISO standard fire-event is just an exponential heating, T=Tamb+Alog(1+Bt), with A=345 K and B=8
min-1.
There are some basic differences between open and enclosed fires: flame tip temperature, burning rate,
pressure jumps, and, most important, the ceiling layer.
Consider the same fuel in air geometry, e.g. a circular oil-pool fire of a 1 m2, in calm air, either outdoors,
or indoors in the middle of say a 333 m3 room; if the enclosed fire is close to a wall, significant changes
take place. In both cases there is a turbulent flame that completely fills the space above the fuel to some
height (nearly independent of pool size for >1 m2), plus some flame tongues or flame tips going further
up (some 2..3 m in total). In both cases the bulk region is at a nearly uniform temperature of 1100100 K,
but the temperature at the flame tips is 65050 K in open fires and 80050 K in enclosed fires, due to the
different heat radiation loss in each case.
Open fires show larger burning rates because of better air supply; the values for a 1 m 2 kerosene pool
being 0.02..0.05 kg/(m2s) for enclosed fires and some 0.07..0.08 kg/(m2s) for open fires.

Enclosed fires give way to significant pressure changes in the interior. Depending on openings and
ventilation, of course, inside pressure suddenly rises some 0.2..0.8 kPa for a few minutes after fire builds
up, due to the initial expansion of hot gases; then the pressure fluctuates around the original ambient
pressure. If the fire is extinguished, there is a small underpressure of order 0.1 kPa for a few minutes
(depending on ventilation), due to contraction on cooling of the hot gases inside. If the fire goes on and a
flashover in the ceiling layer takes place, a sudden pressure increase occurs that may break windows or
walls.
The most important difference, however, is that, although in both cases the temperature slightly decreases
with height further up the flame tips, and the vertical speed of gases also decrease due to air entrainment
(from a maximum of up to 10 m/s), after a few minutes from ignition, enclosed fires build a hot ceiling
layer that slowly increase temperature with time, due to the hindered heat losses, until the flashover
temperature of 90050 K (typical autoignition temperature) is reached in the ceiling layer with the sudden
burning of all combustible gases generated by pyrolysis and vaporisation of other combustible matter
existing in the room.
A two-zone thermal model, a uniform ceiling layer of hot gases floating over a floor layer of fresh gases
and a narrow hot plume, exchanging mass and energy, is often used as a first stage in the analysis of
enclosed fire development, before incurring on the expense of CFD codes.
FUEL DETECTION
A fire needs a fuel, an oxidiser (that is always available, ambient air), and an ignition source. Because
ignition sources are commonly present (cigarettes, heaters, electrostatic sparks), fire safety starts by
identifying and detecting fuel sources.
Besides pyrotechnics and explosives, that must be specially treated, fuels may be classified as flammable
substances (gaseous fuels and high-vapour-pressure condensed fuels) and combustible substances (lowvapour-pressure condensed fuels).
Non-flammable combustible substances are of secondary concern to fire safety, since they will not be
ignited by accidental sparks or distant flames; amongst them, only liquid fuels (particularly gas-oil and
kerosene) are of special risk, and the emergency measures should be proportional to spillage and
inversely proportional to flash-point temperature. The usual fuel-leakage detection method for
combustible substances is visual inspection of liquid-fuel tanks.
Flammable substances must be so labelled, and any uncontrolled treated as an immediate fire hazard with
ample ventilation and evacuation. Flammable liquids (gasoline, alcohols, liquefied petroleum gases) and
gaseous fuels are an obvious hazard since on a leakage occasion, they form ignitable mixtures with air
during the process of mixing (and perhaps after full mixing, in closed spaces). Accidental ignition sources
are so common (sharp knocks, electrical contacts, cigarettes, stoves and other hot points), that the risk of
ignition is very high.
Fuel-gas leakage detectors are also called explosimetres. They are usually based on the electrical
resistance variation of a platinum wire, due to the temperature increase caused by rapid catalytic
oxidation of the fuel-air mixture at room temperature in the surface of the catalyst.
Portable system with field replaceable measuring cells are in the market capable of sensing minute
concentrations of natural gas, butane and propane, either along pipes and combustor (gas leak), or in the
ambient (explosimetres).

Sometimes, to avoid explosions, it is not enough to have a quantity less than the lower ignition limit in a
closed room, since very light fuels like H2 or very heavy gas fuels as diethyl ether (C4H10O) will stratify a
lot.
Mercaptan odorants are always added to gaseous fuels so that they can be detected by scent before
reaching explosive levels.
FIRE DETECTION
Once a fire breaks out, it can be detected by different methods:
Smoke detectors
Flame detectors (infrared or visible)
Heat detectors or, more generally, thermal detectors.
The detector can be used to ring an alarm or directly fight the fire by active means, usually by water
sprinklers, using low-melting-point or thermoelastic switchers (e.g. an aluminium rod that expands and
breaks a glass stopper), or any other fire-fighting fluid. Some flame detectors and thermal detectors can
be found in Combustion instrumentation or in Thermal effects on materials, aside. For thermal detection,
care should be taken not to confuse fire with other inoffensive heating; fire alarms usually require a
temperature above 60 C plus a heating rate above 1 C/min.
SMOKE DETECTORS
Smoke detectors are used to provide a quick alarm (acoustic at least) once a fire breaks out. They are
usually combined with automatic water sprinklers (after a warning period of alert, since fire extinguishers
always cause damage, and the smoke detector trigger might be due to a false alarm). Smoke detectors
should be located in exposed areas of the room, and not in dead-air corners (a forced air convection is
applied in spacecraft racks, for that purpose, and to enhance heat removal).
There are two basic type of smoke detectors:
Ionization detectors. They contain a very small amount of radioactive material (e.g. americium, a
radioactive metallic element produced by bombardment of plutonium with high energy neutrons) that
ionizes the air, making an electrical path, where an electric current is established. When smoke enters
within the path, the smoke molecules attach themselves to the ions and the electric current changes,
triggering the alarm.
Photo-electric detectors. They contain a light source (usually a small bulb) and a photocell. When
smoke enters within the light path, the transmissivity changes, triggering the alarm.
WATER SPRINKLERS
Automatic water sprinklers, connected to a wet-pipe system (only suited for areas not subject to freezing),
provide both fire detection and suppression. They were introduced as back as the XIX c., and they are
based on an operating element, a low-melting-point or thermoelastic device, that clears the water exit;
water pressure pushes the sprinkler cap aside, and water is discharged only through sprinklers that have
opened due to exposure to heat, all the others remaining intact. Water sprinklers are very reliable and safe
(the chances of a sprinkler accidentally going off are remote) and minimise the water damage associated
to fire fighting.
In large systems, the flow of water out of the sprinkler system trips a monitoring device, typically a water
flow detector, which activates a central fire alarm.

Dry-pipe systems are essentially the same as wet-pipe systems except that the pipes in the protected area
contain pressurized air, allowing for freezing temperatures. When a sprinkler head is activated by a fire,
the pressurized air is released and water flow occurs.
A new fire-sprinkler technology is based on delivering water at very high pressure (requires special
pumping), producing a fine and high-efficiency water-vapour mist that maximize cooling, extinguishing
fires with minimal amounts of water (a few litres for a typical room).
INERT ATMOSPHERES
Many industrial fires take place when handling flammable materials or when welding near combustible
materials. Procuring a local inert atmosphere (i.e. depleting oxygen from the air) may be the best fireprevention action, since it is impossible to get rid of combustible materials, and very difficult to get rid of
ignition sources. In most circumstances, an inert atmosphere is one having xO2<5% (not to be confused
with an under-lean atmosphere, which is one with less flammable vapours than the lower ignition limit for
the fuel).
The two basic inertisation procedures are:
Nitrogen atmosphere, where the oxygen molar fraction in air xO2=0.21 is reduced below xO2=0.08
by adding nitrogen (nitrogen with 5% impurities, as obtained with selective permeation
membranes, is good enough). Inert atmospheres are routinely used onboard crude-oil tankers, low
flash-point chemical cargo ships, and fuel tanks; the largest nitrogen-generator producing 1.7 m3/s
from air by membrane technology. Besides fire prevention, nitrogen systems are used as super dry
(Tdew<200 K), non-freezing, non-oxidative, non-corrosive atmospheres.
Burnt atmosphere, where the oxygen molar fraction in air xO2=0.21 is reduced below xO2=0.08 by
adding exhaust gases from a combustor (the air/fuel relative ratio cannot be too large; up to =2 is
good enough).
For a tank to be really protected with an inert atmosphere, it must always have some overpressure,
relative to the surrounding air.
FIRE-BREAKS AND FIRE EXTINGUISHERS
Once a fire is noticed, an early alarm to trained personnel is most of the times better than a personal
attempt to extinguish a fire. Small fires may be fought with fire extinguishers, but for large fires the first
action must be to isolate the damage with fire-breaks. A fire-break is anything that stops a fire from
spreading: a strip of cleared land in a forest, a masonry wall, a fire-retardant door, a hydraulic seal, etc.
Materials are usually termed non-combustible if they cannot combine with oxygen at all or if their flashpoint is higher than 1000 K, i.e., if when heated to 1000 K in a furnace for 3 hours, they neither selfinflame, nor can be ignited by a near spark (ISO 1182). But there are several other characteristics that
make a material fire-resistant to different degrees. Fire-resistance tests are performed on building
construction elements (ISO-384), vehicles, wires, industrial installations, etc.
Extinguishers must be located near escape routes. Fires and extinguisher are classified according to the
material that is on fire: solid, liquid, electrical or metal, although many extinguishers are good for several
types of fire (and so are labelled; e.g. CO2, halons, their substitutes (e.g. HFC-227ea, HFC-236fa, FM-200
and Inertgen), and dry-powder (NH4H2PO4, mono-ammonium phosphate, or ammonium dihydrogen
phosphate, ADP), are all good for types A, B and C):
Type-A fires involve ordinary solid combustible substances (coal, wood, cloth, rubber and many
plastics), and can be extinguished with a cooling agent (water, foam, earth) to arrest the generation

of fuel vapours. Beware of arrested solid fires; their embers remain hot inside for long, and may
inflame many hours later by a blow of air. Typical hazards at home are cloth fires by cigarettes,
stoves and unsafe playing with lighters and matches. These fires are smoky and cause
disorientation and toxic fumes (remember: the air in a fire is cleaner at the floor).
Type-B fires involve flammable liquids and gases (oil, gasoline, natural gas, GLP, paints, lacquers,
grease, solvents), and are extinguished with an air-barrier dry-chemical (CO2, Inertgen, FM-200)
and never with water (the fire would spread), with the advantage that they do not leave a mesh of
foam. Typical hazards at home are fires at the kitchen by overheated oil or grease when cooking
(covering with a lid is usually enough to stop the fire; do not carry the pan outside or throw water
over). Sometimes, liquids are called flammable only if their flash-point is below 48 C (that of
diesel), the rest being called combustible liquids.
Type-C fires involve electrical potentials (wiring, fuse boxes, energized electrical equipment), and
must be extinguished with an electrically insulating air-barrier (e.g. CO2; water may cause
electrocution).
Type D fires involve metals such as magnesium, aluminium, iron, titanium or sodium; they are
very dangerous and should not be handled by untrained personnel.

As for refrigerant fluids, the standard fire-extinguisher fluids during the last half of the 20th century were
halogenated hydrocarbons (halons), which had to be substituted in the 1990s by non-ozone-depletion
fluids. The most used were Halon 1211 (CF2ClBr, bromochlorodifluoromethane, a liquefied gas), and
Halon 1301 (CF3Br, bromotrifluoromethane, a compressed gas), which have been substituted by
heptafluoropropane (R227, C3HF7, also named FM-200).
CO2 fire extinguishers are loaded with liquid CO2 that, when the valve is triggered, is expelled from the
bottom through a riser pipe and finally expanded at the end of the high-pressure hose, where the liquid
flashes to a mist of dry-ice particles and CO2-gas, focused along a long funnel (that may become very
cold and should not be touched). Inergen (IG-541) is a mixture of 52 % N2, 40% Ar and 8% CO2, which
has substituted CO2, to avoid the problem of blocking blood decarboxylation, i.e. when mixed with air
more or less half and half (they are almost isodense), the resulting x O2<10% is already low to sustain fire
propagation (but able to keep human conscience during several minutes), whereas the resulting x CO2<4%
is low enough to not blocking blood release of CO2 at the lungs.
Powder fire extinguishers are loaded with NH4H2PO4 or NaHCO3 powders and a pressurised CO 2
cartridge that, when triggered, pushes down and expels the powder as a fine mist that dissociates to CO2
and H2O in contact with the fire, and which penetrates everywhere (the mesh they create may outbalance
their use in small fires).
Fire engine trucks carry an extensive assortment of tools and equipment: pumps, hoses, water tanks,
portable extinguishers, ladders, axles, breathing apparatuses, communication equipment, electric lights
and generators, etc. They are used to fight domestic, industrial and forest fires, but the latter usually
requires straightforward measures as firebreaks (removing further fuel ingest), spreading chemicals that
decrease the flammability of wood, flooding with water, or spreading smothering materials such as dirt or
sand.

Pollutant emissions
Life is a polluting process, because life must live at the expense of the environment. The problem is that
the amount and concentration of pollutants emitted by human activities has gone too far and seriously
menace life, both locally and globally.

The first and traditional approach to fight pollution is to go away or throw it away, e.g. to make a
chimney for venting the fireplace, to have out-side-air sealed combustion boilers, to build power stations
aside, etc.
The effects on the environment are usually identified with pollutant emission through the tail-pipe of
combustors, but handling of fuels, fuel losses at the inlet, and product losses from the combustor shell, are
other sources (up to 20% of the hydrocarbon emissions in a car do not go out along the tail-pipe). Mass
losses and energy losses may be a danger to humans, animals, plants and goods: explosion danger (in
confined places), open flame danger, toxicity from CO (and other toxic gases in chemical fires),
suffocation or anoxia from CO2, hyperthermia by heat, respiratory and visual irritation by smoke and
noxious gases, etc. Mechanical pollution (noise) and electromagnetic pollution (interferences, EMI), are
dealt apart.
Main contaminants, besides the unavoidable CO2 in carbon-containing fuels that contribute to the
greenhouse warming (the legacy left by these emissions would be felt mainly by our offspring), are:
VOCs (Volatile Organic Compounds), CO, NOx, SOx, and particulate matter (PM, or PM10 to explicitly
restrict to sizes <10 m). Soot is formed in non-premixed flames and on premixed flames for equivalence
ratios >1.5. Diesel engines produce more pollutants in the stated order (more PM and less VOC),
whereas Otto engines do just in the reverse order. Some 40% of man-produced VOCs come from
transport (not only through the tail-pipe but from reservoirs and at the stations). Sometimes, instead of
VOC, the term HC (hydrocarbons) is used, even splitting between methane and NMHC (non-methane
hydrocarbons).
EMISSION QUANTIFICATION
Quantifying emissions is not a trivial issue and several methods can be followed:
Mass-to-mass emission indices, EIi, defined as the mass of i-contaminant emitted divided by the
mass of fuel burnt. Although a dimensionless ratio, units of g/kg are common. This measure is
independent of the application at hand and the dilution applied, thus it does not enter into
considerations of the actual effects of the emissions or the real need of the process, but it is the
best method from the combustion point-of-view to quantify emissions.
Mass-to-energy emission ratio, EERi, defined as the mass of i-contaminant emitted divided by the
energy supplied by the fuel burnt. It is intended to weight the burden against the benefit of the
combustion process, but great care must be paid to ascertain if the energy supplied is just the
heating value of the fuel burnt, or the final energy delivered to the payload (shaft work for
combustion engines, or water heated energy for boilers). This ratio serves to compare pollution
from other energy sources different to combustion, but with little relevance since the emissions are
so different (e.g. nuclear vs. fossil power stations).
Mass-to-purpose, e.g. mass-to-distance travelled emission ratio, DERi, defined as the mass of icontaminant emitted divided by the distance travelled. It is intended to weight the burden against
the real benefit of the combustion process, but great care must be paid to measure the benefit; of
course, the goodness of transportation is not the power of the engine but the distance the payload
travels; but time is also important, and the payload is not fixed beforehand (e.g. distanceemissions ratios are given for a car in a given trajectory, city cycle or highway, irrespective of the
number of actual occupants).
To give an idea of the pollution in a major city, the average figures for measured values in Madrid are:
1000 g/m3 of CO, 70 g/m3 of NO2, 1000 g/m3 of CO, 40 g/m3 of PM10, 30 g/m3 of O3 and 20
g/m3 of SO2.

FUEL TANK AND CRANK-CASE VENTILATION

Up to 20% of the hydrocarbon emissions in a car may not go out along the tail-pipe but directly from the
fuel tank, the crank-case and other fuel leakages.
Fuel leakage can be detected by several methods. For condensed fuels the key leakage diagnostic is
spillage tracking, whereas for gases, bubbling (usually by applying a soap solution), UV-fluorescence,
ultrasonic, and catalytic methods are used.
Fuel tanks are pressurised only for gaseous fuels; liquid-fuel tanks are not pressurised but ventilated to
prevent build-up of pressure and facilitate refuelling. Nowadays, liquid-fuel tanks are no longer directly
ventilated to the ambient but through an active-carbon filter (canister) that retains the fuel vapours; during
normal engine operation, some fresh air is passed through the canister to the admission pipe to burn the
vapours and regenerate the filter.
The crank-case (oil box) is also ventilated to the admission pipe, because there is an important fuel loss
from the combustion chamber to the oil box through the slipping rings, particularly during cold operation
(at starts).

EXHAUST EMISSIONS AND POLLUTION

The complete combustion of CuHvOwNx fuels would only yield CO2 and H2O as new compounds. Water
is thought to be processed globally by the hydrological cycle (although local condensations may be
corrosive and be the best culture for moulds and bacteria), but CO2 is already considered a pollutant due
to the overall green-house-effect contribution if non-renewable fuels are used. We intend here to analyse
pollutants going out the exhaust pipe of mobile and stationary combustors that, although in concentration
much lower than CO2 and H2O (typically less than 1%), have a great impact on the environment.
Combustion processes where non-commercial fuels are burn (e.g. wild fires, waste incineration, rocket
propulsion) give off many more hazardous contaminants not considered here: halogenated (dioxines,
furanes, hydrogen chloride), organometallic and others..
Just in EU, the annual emission of CO2 is some 31012 kgCO2 (40% due to electricity generation, 30% due
to transport engines, and the rest to industrial and other activities), and is stabilised (other emissions are
slowly declining: 15109 kgVOC, 15109 kgNOx, 10109 kgSO2, etc.). A major milestone in world pollution
control was achieved in the Kiotto-1997 protocol, where a target on anthropogenic greenhouse gases
emissions of less than 5% the 1990 world figures was established for 2008.
Environmental pollution locally depends on the amount and type of industrial activities, density and age
of vehicle fleet, efficiency in combustion, weather conditions, climatology, etc. In most developed
countries, the transport sector is a major source of air pollution (amongst it, road transport takes more
than 80%), and the dominant source in urban areas. Continuous emissions monitoring (CEM) is
mandatory in most major combustors. Exposure to air pollution can cause adverse health effects, most
acute in children, asthmatics, and the elderly, and can damage vegetation and materials (notably, the
cultural heritage). ICE are responsible for nearly half of the world contamination load. About 1 % of the
exhaust gas stream is harmful, and consists of carbon monoxide (CO), oxides of nitrogen (NOx), volatile
organic compounds (VOC, also named hydrocarbons: HC) and particulate matter (PM).
In premixed combustion, emissions depend upon the air-fuel ratio, but unfortunately when the
concentration of CO and HC decreases the concentration of NOx increases, and vice versa. Non-premixed
combustion may be much more pollutant: the flame sits near the stoichiometric diffusion-rates and thus

its temperature is very high, causing the formation of NOx, and the pure fuel approaching it gets
pyrolysed with large production of soot (that give the characteristic yellow colour to non-premixed
flames) and other volatiles.
CO2
Carbon dioxide (CO2) is an unavoidable emission in the combustion of carbon-containing fuels. There is
little concern with local contamination by CO2 emissions, being an inert gas like nitrogen. Like nitrogen,
it may cause local suffocation; a 10%CO2 is fatal to a person after a few minutes, by anoxia, and a 5%
already produces troubles after 1 hour. The main concern of CO2 emissions is at global scale, with the
positive correlation found between anthropogenic CO2 generation and the increase of CO2 fraction in air
(e.g. from 310 ppm molar in 1950, to 380 ppm molar in 2007), and the foreseeable consequences on
climate change due to the associated increase in the global greenhouse effect, which is feared to be highly
non-linear, with a small global warming but larger regional changes (desertification, floods, hurricanes,
and so on.
There are two ways to decrease CO2 emissions: to decrease overall emissions by increasing the energy
efficiency of the combustor, engine and associated transmissions, and to use low-C fuels (shift from coal
with C/H=1.3 and liquid fuels with C/H=0.5, to GLP and, better, to natural gas with C/H=0.25, and
eventually to H2 with C/H=0).
However, the foreseeable reduction in CO2 emissions seem not enough, and technologies to capture the
CO2 before it goes to the atmosphere must be applied in the near future. Use of CO2 as by-product to
other industries is too small: as a chemical stuff for methanol and urea synthesis, as an inert gas in
welding and inertization of large fuel installations (dry stoichiometric exhaust gas from a combustor is the
cheapest choice), CO2 is used to help recover residual crude-oil at some wells, as a refrigerant (including
dry ice applications), as a solvent in the cleaning industry, in the preparation of carbonated beverages, etc.
The only way out is then CO2 capture (or sequestration), and permanent disposal in deep soil reservoirs
(deep-sea disposal seems too risky, environmentally). CO2 capture may be based on different approaches:
Capture the emitted CO2 from conventional sources. This is impractical once the CO2 is diluted
in the atmosphere (x=0.04% molar), and seems impractical too in transport, but is being tried in
large coal-fuelled power stations (with x=0.15..20% molar). Physical separation (by membrane
technology, cryogenic distillation, or sorption) is presently too expensive, but chemical
separation has made good progress in the carbonatation-calcination process (CO(gas
mix)+CaO=CaCO, followed by CaCO+heat=CaO+CO).
Capture the emitted CO2 in a better way by modifying the combustion process, to be performed
with pure oxygen what is known as oxy-combustion, yielding a pure CO2 exhaust, or a CO2 /
H2O gas mixture easily separated by water condensation. To avoid too high temperatures, CO 2
is always partially recirculated. To get pure oxygen from air, traditional cryogenic distillation or
membrane separation may be used, although new approaches are been tried, as using an
intermediate metal that gets oxidised, separated, and then reduced to yield pore O2 and the
regenerated metal
Fuel decarbonisation before combustion, i.e. capture the carbon (in the form of C, CO, or CO).
This has no sense for coal, which is some 85% by weight C, but a possible solution for natural
gas (e.g. a reforming stage can be used to decompose CH4=C+2H2, get rid of the solid C, and
change to a hydrogen economy.
An indirect way out to CO2 capture is to reforest, for the trees to fix the atmospheric CO2 into
biomass (which can be later used as fuel source or not). Some other large scale enhanced
natural ways to get rid of CO2 have been suggested, as promoting carbonatation of marine biota

by seeding the sea-surface with iron particles, and so on (it is a fact that only a third of the CO 2
emissions accumulate in the atmosphere, the rest being captured by the oceans).

CO
Carbon monoxide (CO) is found in exhaust emissions due to a poor combustion process (i.e. too rich a
mixture, unburnt fuel pyrolises at crevices, or not enough residence time for equilibrium), particularly in
the Otto engine at cold starts, idle, and full power conditions. Unburnt pyrolysed fuel would be in
negligible amounts if sufficient time for equilibrium at the low exit temperatures were allowed, as in large
combustion chambers and large marine engines, where the residence time is near one second (combustion
is similar to eating; it takes time for a good digestion, starting by proper food preparation, chewing and so
on).
CO is a deadly poison that reduces the ability of the blood to absorb oxygen and, as a result, lowers the
blood oxygen content by producing carboxyhemoglobin. Even as low a proportion as 0.5 percent by
volume of CO in the air can prove fatal within 1 hour (>50% carboxyhemoglobin in blood), 0.05%
produces headache after 10 hours (10% carboxyhemoglobin in blood). In uncontrolled fires, like in a
hotel room, typical concentrations of up to 5% are achieved after the fire runs away of control. Measured
in the blood. For clean air standard EU sets a limit of 10 mgCO/m3 in a 8 hours average.
Two types of approaches may be followed to fight CO emissions (besides minimising fuel use): to avoid
CO formation inside (e.g. using stratified charging), and to eliminate the CO inside the tailpipe, usually
by a catalytic oxidiser (e.g. the tree-way catalyser in Otto engines). CO and CO2 are usually measured by
infrared absorption.
NOx
NOx stands for all nitrogen oxides, mainly NO and NO2, but also N2O, N2O2, N2O3, N2O4 and N2O5, and
all appear from atmospheric nitrogen during combustion with air (coals and heavy fuel-oils have some
intrinsic nitrogen also, up to a few percent by weight). All nitrogen oxides are unstable at ambient
conditions when pure, i.e. they dissociate, but their decomposition may be very slow. They are formed at
very high temperatures in the presence of air (there is a high peak in the range =1..1.1), and two
approaches are followed to avoid their emission: avoidance of high temperature formation (by using very
lean mixtures, exhaust gas recirculation, porous burners, catalytic burners, water injection), and catalytic
reduction at the exhaust. Notice that the peak in NOx production practically coincides with the range of
maximum combustion efficiency (minimum entropy production), so that it might be said the NO x
emission is a sign of good combustion, contrary to unburnt emissions.
The most polluting of NOx-components is nitrogen dioxide, NO2, a brown gas at normal conditions (but
readily condensable, Tb=11 C). When heating an ampoule containing NO2 from above, some dinitrogen
tetroxide N2O4 (2NO2=N2O4(g)+57 kJ/mol) is exothermically formed; N2O4 is a colourless heavier gas
that appears at the bottom (because of buoyancy), although the reaction would be more displaced to the
left in equilibrium at high temperature (but kinetics dominates). A third gas appears when heating at 600
C, nitrogen oxide, NO (N2O4=2NO+O2, also transparent. Nitrogen oxide, NO, is a colourless gas that in
the presence of atmospheric oxygen, rapidly converts to yellow NO2; NO concentration can be measured
by chemiluminescence with ozone: NO+O3=NO2+O2+h. NO2 smells pungently and causes pronounced
irritation of the respiratory system if > 10 ppm, and is fatal if >100 ppm after minutes; due to the fact that
it destroys the lung tissue, for clean air standard EU sets a limit of 40 gNO2/m3. When NO2 or N2O4 are
cooled, a blue liquid condensate first develops (a strong mixture of N2O4 in NO2), and after further
cooling, a blue solid appears, mainly consisting of N2O3. N2O is a powerful greenhouse gas. Atmospheric

ozone, O3, is another pollutant (contrary to stratospheric ozone), and, although not directly emitted in
combustors, it is formed by reaction with air of NOx emissions. Nitrogen oxides also combine with water
vapour to form acid mists (pH<5.6 at 288 K) that give way to acid rain, damaging forest, lakes and rivers
ecosystems, one of the key reactions being NO(g) + (3/4)O2(g) +(1/2)H2O = H+(aq) + NO3-(aq).
Two types of approaches may be followed to fight NOx emissions (besides minimising fuel use): to avoid
NOx formation inside by avoiding high temperatures (e.g. using lean mixtures, using exhaust
recirculation, using water-emulsified fuel), and to eliminate the NOx inside the tailpipe, usually by a
catalytic reductor (e.g. the tree-way catalyser in Otto engines, the urea catalyser in Diesel engines). NO x
concentration may be measured by chemiluminescence or by infrared absorption. Some amount of
exhaust gas recirculation (EGR), typically in the range 10..20% (but up to 50% for small loads) is
currently applied in all kind of new engines (Diesel and Otto).
VOC
Volatile organic compounds, coming from unburnt fuel and pyrolysed fuel, are a group of chemicals with
Tb<250 C that includes important air pollutants like benzene, 1,3 butadiene, and acrylic aldehyde
(CH2CHCHO, deadly if >10 ppm), that is the cause of the bad smell from tail-pipes. From uncontrolled
fires, with typical solid substances as wood, wool, plastics and flesh, very toxic substances are released,
as hydrogen cyanide (HCN, deadly if >0.3%, and found up to 0.1% in typical home fires), ammonia
(NH3, deadly if >0.3% in half an hour), hydrogen sulphide in rubber and flesh burning (H2S, deadly if
>0.1%), and phosgene from PVC burning (COCl2, deadly if >0.1 ppm). For clean air standard EU sets a
limit of 5 gC6H6/m3.
VOC is a cul de sac, comprising all chemical emissions except the singled-out H2O, CO2, CO and NOx;
and the expected trend is to go on with the singularisation of emitted substances, since there have widely
different effects on the environment. In that move, separate analyses have been already applied to natural
gas combustion, classifying its VOC (or gaseous HC) as methane and non-methane (MHC and NMHC).
Other approaches split further the bunch of substances identifying e.g. polycyclic aromatic hydrocarbons
(PAH), BTEX (benzene, toluene, ethylbenzene and xylene), and others, as the more dangerous to health.
VOC (and CO) emissions should be very low for premixed combustion with excess air (even with
stoichiometric air), but Otto engines are the major source of them because of the small residence time
(some milliseconds for combustion, against near one second in premixed industrial burners), and the
associated small size of the combustion chamber (limited to say half a litre per cylinder for this fact). The
most pollutant are the small two-stroke Otto engines used in motorcycles and gardening, because the
fresh mixture is directly thrown to the exhaust to sweep the burnt gases in the cylinder. By the way, for
reciprocating engines, VOC are not only due to the fuel but to the lubrication oil that seep through the
segments and gets burnt (in the small two-stroke engines oil is add directly to the fuel).
Two types of approaches may be followed to fight VOC emissions (besides minimising fuel use): to
avoid VOC formation inside (e.g. using stratified charging), and to eliminate the VOC inside the tailpipe,
usually by a catalytic oxidiser (e.g. the tree-way catalyser in Otto engines). VOC may be measured by
flame ionisation or by infrared absorption.
PM
Particulate matter (PM, or PM10 to explicitly restrict to sizes <10 m) is harmful to the respiratory
system for sizes smaller than say 10 m (larger particles do not follow the air stream and get stuck at the
nose and trachea), but the worst are sizes <2 m. Particulate matter consists of soot from all kind of
hydrocarbon combustion (mainly in non-premixed flames), and fly ash from coal and waste combustion

and incineration. Premixed combustion starts producing soot for air-to-fuel relative ratios <0.5 or 0.6,
depending on the fuel.
Particulate matter was also characteristic of diesel engines at low loads (e.g. during acceleration), due to
very inefficient burning of fuel drops on cold surfaces, with a dense dark smoke at the exhaust. Remedies
have been the elevation in fuel-injection pressure (up to 200 MPa), preheating of air, and exhaust
filtering. Nowadays, a non-visible-smoke exhaust is mandatory in practically all types of engines (even
for ships). Tobacco smoke, incense burning, charcoal grills and the like, are well-known sources of
particulate matter associated to combustion (as well as some related processes as deep frying), but
emphasis here is on engineering applications.
Finer particles, of sizes smaller than 2.5 m (named PM2.5), are even worse than PM10; they are issued
from combustor exhaust, but also form by atmospheric reactions of NOx and SO2 emissions (forming
nitrates and sulphates).
Two types of approaches may be followed to fight PM emissions (besides minimising fuel use): to avoid
PM formation inside (e.g. using homogeneous mixture compression and suitable injection rates), and to
eliminate the PM at the exhaust by an appropriate filter, that should be periodically regenerated (e.g. by
passing hot gases rich in oxygen, as done in Diesel engines). PM and smoke are distinguished by the
measuring method: PM is measured by weighting a filter, and smoke by light absorption. For clean air
standard EU sets a limit of 20 g/m3 for sizes <10 m.
SO2
Sulfur dioxide (SO2) emissions mainly depends on type of fuel and not on combustion details, and the
trend has been to get off the market sulfur-containing fuels (by desulfurising those that need it), or
implementing desulfurising agents in fluidised-bed combustion (e.g. adding lime for
CaO+SO2+2H2=CaSO32H2O, or limestone), or in the exhaust (deSOx dry or wet scrubbers). Nowadays
only very large marine engines and large power stations still burn sulfur-containing fuels (residual fuel oil
and coal, respectively), causing severe local and global pollution (acid rain); even in 2005 the IMOMARPOL limit on marine fuel is 4.5% in sulfur (down to 1.5% in special areas like in EU seas). Power
plants alone contribute to 2/3 of SO2 global generation in 2000 (and 1/4 of NOx and 1/3 of mercury, from
the fuel). SO2 has been shown to have detrimental effects on the selective reduction of NOx with
ammonia (SCR and SNCR), by formation of the highly corrosive ammonium bisulphate, although the
dirty exhaust associated to the combustion of sulfur-containing fuels usually prevents the use of
catalysers.
Emission regulations
In the European Union, exhaust emission limits and their testing are regulated: 1992 Euro I (adoption of
catalytic converters), 1997 Euro II, 2000 Euro III and 2005 Euro IV. Table 1 presents the typical emission
rates of new (2000) cars, as per-distance travelled. Table 2 compares these internal combustion engines
with other power-producing devices, in a per-shaft-energy basis. Notice that a typical car running at 100
km/h demands a propulsive power of some 20 kW, consuming thence 1 kWh of work every 5 km or so.
In the USA, the Environmental Protection Agency (EPA) is the organism in charge of protecting human
health and safeguarding the natural environment, and thus responsible for emission control.
Table 1. Typical emission rates of new (2000) cars a.
Otto
Otto (tree-way cat.)
Diesel
Exhaust manifold
Emissions
Exhaust manifold

Diesel
Emissions

CO2
CO
VOC
NOx
PM
a

(wet basis)*
8..10%
0.1..1%
0.5..1%
NO: 100..1000 ppm
NO2: 10..30 ppm
<0.001 g/m3

>120 g/km
>0.2 g/km
>0.03 g/km
>0.01 g/km
<0.001 g/km

(wet basis)b
4%
0.1%
0.5%
NO: 100..1000 ppm
NO2: 10..30 ppm
>0.01 g/m3

>100 g/km
>0.05 g/km
>0.03 g/km
>0.2 g/km
>0.01 g/km

Stationary engines have tighter emission regulations (e.g. CO<700 ppm, NO x<250 ppm, or NOx<400 ppm for biogas engines).
Water molar fraction in Otto engine exhaust is some 15%, and in Diesel engines some 7%.

Table 2. Typical emission rates and fuel consumption of different power engines, in g/kWh a.
Coal Steam Diesel ICE
Diesel ICE Gas Turbine Gas Turbine Phosphoric Acid Fuel Cell
b
Power Plant Euro III
with catalyst
Oil-fired
NG-firedc
NG-fuelled
CO2
800
500
600
800
600
450
CO
5
10
VOC
0.8
2
NOx
1
5
0.4
0.7
1
0.004
PM
0.1
0.1
0
SO2
3
0
0.05
0.08
0
Fuel-in
>400
>170
>170
>200
>200
>120
a

Notice that specific fuel consumption is the inverse of specific energy output (e.g. 200 g/kWh = 18 MJ/kg).
Marine diesel (>150 kW engines) NOx 2000-regulation <10 g/kWh (up to 17 g/kWh for very low engines, below 200 rpm).
c
Gas Turbine NOx 2000-regulation <5 ppm in the exhaust.
b

VIBRATIONS AND NOISE

Combustion generates acoustic waves due to shear flows and turbulent fluctuations. This aerodynamic
noise, besides the possible mechanical noise associated to oscillating or vibrating solid parts in
combustors, is a source of environmental pollution.
In reciprocating engines in particular, exhaust gases leave the cylinder under high pressure that, if allowed
to escape to the atmosphere directly, would create an unbearable loud noise. That is why muffler silencers
are installed in all exhaust pipes for sound attenuation.
ELECTROMAGNETIC INTERFERENCES
In spite of the undeniable fact that combustion pollution is mainly due to tail-pipe chemical emission
(CO, NOx, VOC and PM), the electromagnetic contamination due to operation of electrical and electronic
devices in reciprocating engines keeps increasing since the old days (1897) when the magneto substituted
the open-flame hot-tube ignition system in Otto engines.

SAFETY MANAGEMENT
Risk: a combination of hazard and damage
Humankind has always evolved in a risky environment, plenty of possible sources of injury (hazard), and
different levels of possible impact (damage). We cannot envisage life without risks (it might turn that risk
is a native ingredient in life evolution), but we must be aware of risks and try its optimum management,
accepting minor risks and injuries as natural events (e.g. freely using glassware, cutlery, matches..., but
restricting the use of powerful machinery, or flammable materials and explosives).

The basic rule in risk management is that "the higher the possible damage, the lower the acceptable
hazards", always realizing that we cannot devote our life to risk prevention (e.g. most minor accidents and
unsafe exposures are recoverable, and bald remedies requiring a century are useless). Galen, in the 2nd
century, already said that doses alone make the difference between medicine and poison.

Emergency response
A quick-list of actions to do after an accident occurs may be (intervention protocol:
1. What happened? Keep calm, to best identify the accident for proper acting and for reporting (identify
the type of damage and its extent).
2. Ask for help? Do not panic but hurry up; alert the closest knowledgeable people. An early alert to
professional teams is key to control large emergencies.
3. Isolate damage! Has it ended? Avoid escalation in damage; isolate causes and damage extension:
close valves, put further risks away (including children).
4. Injuries? Keep priorities: is life threatening? Do not think about property until life and health is safe.
5. Rescue. Keep priorities (benefit/cost) on what can be rescue. Organise actions to minimise overall
cost, accident and countermeasures; sometimes fighting accidents cause increased damage.
6. Remediation. Many times, damage costs increase with time due to lack of benefit of use, so that quick
temporal repairs are often applied (they should not contribute to enlarge risks).
7. Be better prepared for emergences. Individuals cannot be on permanent watch, but society must keep
watch teams alert all the time (e.g. policemen, firemen). Both, society and individuals must keep
proportional emergency stores (e.g. first aid box), and rehearse emergency procedures (e.g. evacuation
rules).

Analysis of accidents
1.
2.
3.
4.
5.
6.

What happened. Identification of the accident: what happened, when and where.
How it happened. Immediate cause of the accident (first explanation).
Why it happened. Immediate responsibility: first explanation of why it happened (had proper
prevention measures been taken?).
How it was fought. Damage control, rescue and remediation.
Who pays. Responsibilities.
Rules updating. Risk acceptance is a subjective measure, and people may change priorities
with experience.

Prevention of accidents
Individuals and society accept risks in accordance to the benefit the risky activity brings or may bring, as
when settling by river banks that might flood, under roofs that might fall over, when travelling in
sophisticated vehicles, or when using powerful tools.
There is, however, a demand on risk management, usually delegated on social authorities, to control that
risks have been evaluated, and procedures have been adopted in developed societies to guarantee a safe,
efficient and respectful development of human activities at reasonable risk.
Risk management can be split in several stages:
1. Hazard identification and labelling (the first part of risk management). Hazards must be categorised
and regulated during production, transport and use.
2. Consequences identification (i.e. possible damage prediction, the second part of risk management).
3. Risk assessment (e.g. evaluating risk by comparison with other better-known risks). Known risks are
less feared of (fear might increase risk).

4.
5.
6.

7.

Acceptability identification (avoiding risks disproportionate to benefits, and minimising acceptable

risks).
Distribution of risk according to profiteers.
Prepare intervention protocols to fight in case of accidents (if they can, they will). It is very
important to have an early detection, early alert, early containment and early fight, but above all to
avoid escalation by improper action (quick-response systems may become unstable).
Distribute damage. Prepare intervention protocols to repair possible damage. Insurance companies
aim at reducing the financial risk of individuals by spreading financial loss to a community.

Glossary of terms: accident, exposure, fear...

Accident

Unexpected undesirable sudden event that cause direct damage.

If damage is not undesirable, but intentional, it is an incident (minor offence), aggression or
attack.
If damage is not sudden, but progressive, it is wear or illness.
If damage is not direct, but indirect, it is side-effects or after-effects.
Severe accidents (that affects a lot of people and cause great damage) are termed disasters or
catastrophes.
Damage
Actual harm caused to people, property or the environment.
Damage usually refers to a sudden event; if more progressive, it is said spoil.
Sudden damage to the human body is specifically termed injury.
Defence
Protection against threat or attack from enemy circumstances (e.g. weather) or people (from
thieves to sabotage and military attack).
Defence and protection against accidents is usually split in prevention (before) and response
(after).
Danger
Imminent or probable risk, i.e. threatening circumstances.
Explosion Sudden conversion of potential energy (mechanical or chemical) into kinetic energy of a
high-pressure gas released (pre-existing or being produced), with the production of a bang
noise, and some mechanical work (forcing to move or shattering nearby objects).
Exposure Cumulative amount of hazardous substance or electromagnetic radiation, received by living
or non-living systems.
Fear
Subjective feeling of insecurity or risk (real or imaginary). Perceived risk may be more
frightening than real danger (a USA poll amongst risk experts rated the three most-risky
activities from a given set of 30 as: driving, smoking and alcohol drinking; the same poll
amongst college students yielded: nuclear power, handguns and smoking).
Injury
Physical damage or hurt. Also moral harm (from injustice).
Hazard
Potential source of damage (threat). Hazards may be physical, chemical, nuclear or
biological, and they are inherent to the materials, conditions, processes and activities
involved. Spoil.
Hazard types:
Physical hazards: mechanical (chock, explosion, vibration and noise, particulates),
electrical (chock), thermal (fire, chock, out-range), radiation (ionising, UV, flare, IR,
MW, RW).
Chemical hazards: oxygen depletion, gas poison, aerosols, liquid poisons, solid poisons.
Biological hazards: infection by bacteria, by viruses, by prions.
Ergonomic and psychological hazards: bad posture, bad planning, bad timing, changing
interfaces (e.g. multitasking, dealing with the public), etc.
Risk
Product of probability by consequences of a hazardous event.
Risk likelihood may be imminent or remote, events per time (e.g. 10 deaths per year),
fractions of event cases (e.g. 1 per million).

Spoil

Risk possible consequences may be on health (deaths, injuries, exposure), property (ruin,
spillage), or the environment (pollution, evacuation).
Degradation of usefulness by exposure to harmful environments.
Environmental effects may be physical (wear, temperature, weather), chemical (reactions),
biological (viruses, bacteria) or radiological (ultraviolet, nuclear).
Sudden spoil and accumulated spoil are named damage.

(Back to Combustion)

COMBUSTION THERMODYNAMICS
Combustion Thermodynamics ...................................................................................................................... 1
Combustion ............................................................................................................................................... 1
Thermodynamics fundamentals ................................................................................................................ 2
Thermodynamics of mixtures ................................................................................................................... 4
Equilibrium ........................................................................................................................................... 4
Chemical potential expressions ............................................................................................................. 4
Thermal capacity averaging .................................................................................................................. 5
Water vapour condensation ................................................................................................................... 6
Thermochemistry ...................................................................................................................................... 6
Stoichiometry, extent and affinity ......................................................................................................... 7
Enthalpy of formation and absolute entropy ......................................................................................... 8
Heating value ........................................................................................................................................ 8
Maximum work ................................................................................................................................... 10
Adiabatic combustion temperature ..................................................................................................... 11
Equilibrium composition..................................................................................................................... 14

COMBUSTION THERMODYNAMICS
COMBUSTION
Combustion is a self-propagating exothermic oxidative chemical reaction, mostly in gas phase, producing
light, heat and smoke in a nearly-adiabatic flame front. The overall process in combustion is analogous to
those taking place in fuel cells and living-matter respiration, so that the same overall results apply in all
three cases, in spite of their details being so different; combustion is characterised by the very high
temperatures reached.
The practical goal in combustion study is the prediction of its performance, for a safe, efficient and clean
design and operation of fire-making devices, in terms of the multiple physico-chemical phenomena
involved; it is thence a prerequisite to analysis the latter. Those phenomena may be split in two groups:
equilibrium behaviour (what we need and what we get), and kinetics (how we get it; at what rate).
Combustion Thermodynamics focuses on the former physico-chemical phenomena: fuel/air ratios,
heating values, maximum work obtainable, exhaust composition, etc., whereas Combustion kinetics
focuses on mixing process, flame geometry, ignition, extinction, propagation, stability, etc.
The study of combustion is based on the more general subject of Thermodynamics of Chemical
Reactions, usually called Thermochemistry. We shall deal here only with the peculiarities of the
combustion reaction, i.e. with focus on the thermodynamics of a fuel-and-air gas-phase reaction.
It may be helpful to have in mind a concrete instance of a combustion process, and we proposed the
idealised burner sketched in Fig. 1, a steady-state combustor burning natural gas (here idealised as pure
methane) in air, with some heat output (as in a domestic water heater). This model already emphasises the
black-box approach typical of thermodynamic analysis (it does not look into the internal details of how
the fuel and air mix, the geometry of the flame, or any other gradient or discontinuity within), and
assumes the combustor is large enough for the exhaust to be in equilibrium (no longer reacting, no
gradients).

2
CH4

air

combustor

products
heat

Fig. 1. A simplified model of a natural-gas combustor.

A detailed description of fuels, oxidisers, their mixing, the ignition process and the kinetics of its
propagation, can be found elsewhere. A succinct description of the process in Fig. 1 may be as follows.
As it will be shown below, at least 9.5 m3 of air are required for the complete combustion of 1 m3 of
methane at same p-T conditions (the molar stoichiometric air/fuel ratio is A0=9.5), with a maximum heat
output of 55 MJ/kgCH4 (or 37 MJ/m3CH4, the higher heating value of methane) that would decrease as the
exhaust temperature increase until a maximum when no heat is exchanged, some 2200 K (the adiabatic
combustion temperature for the stoichiometric mixture). The exhaust composition for stoichiometric
mixture consists of 71% by volume of N2, 19% H2O, 9% CO2, and much less than 1% of undesirable
gases called emissions, that are noxious to the health: CO, NO, NO2, aromatic-hydrocarbon vapours and
maybe soot. Those figures already show that thermodynamic properties of exhaust gases can be
approximated by those of air for a crude analysis (as air properties can be approximated by those of
nitrogen); similarly, they show the importance of water as the only condensable gas in the exhaust, and
the small proportion of contaminant emissions, although the massive use of combustion renders their
effects very obvious. Nowadays, even the inert gas CO2 is considered an undesirable emission, since it
contributes to the menacing global warming and its associated climate changes.
Exercise 1. Combustion emissions.

THERMODYNAMICS FUNDAMENTALS
Perhaps a rough summary of Thermodynamics Fundamentals seems appropriate at the beginning of a new
application of the general theory, instead of the recourse to as you may know. We should only consider
systems such that their equilibrium states are just characterised by its energy E, volume V and amount of
each chemical species ni. For such a system in such states, the traditional formulation of Thermodynamics
is based on these principles:
Zero Law: there exists an state-function named temperature, T(E,V,ni), indicating the thermal level
of the system, such that if two systems having different temperatures are put in contact, their
energy varies, flowing from the hotter to the colder one until both reached the same value
(equilibrium).
First Law: there exists a path-function named heat, Q, measuring the thermal energy exchanged by
the system, such that when two control-mass systems having different temperatures exchange
energy, the heat flow at the frontier is just the change in stored energy minus the work received by
the system; i.e. Q=EW.
Second Law: there exists a state-function named entropy, S(E,V,ni), measuring the distribution of
thermal energy within the system, such that for any process in a control-mass system, its variation
is lower-bounded to SdQ/T, with the equality holding only for the limit case of a nondissipating process.
Third Law: there exists a singular value for the state-function entropy, such that all entropy
variations tend to die for any processes at that limit T=0 K, i.e. ST0 K0.
These basic general principles and other particular assumptions on the behaviour of some type of
substances, give way to a formulation that is actually applied when solving problems, and that is here
briefly refreshed, following our Thermodynamics Lectures.

3
Details on the concept of energy, an additive and conservative function of kinetic and potential terms, can
be found aside, but the energy balance of a control mass system (one that cannot exchange mass with the
surroundings), and the perfect substance model for stored thermal energy, should be kept in mind:
E=Q+W (energy balance of a control mass)

(1)

E=mcvT (stored thermal energy model for a perfect substance)

(2)

Details on the concept of entropy, an additive non-conservative function measuring the internal
distribution of energy and other conservative properties, can be found aside, but its basic expression as a
function of other variables, and its particularisation for a perfect gas, should be retained:

dQ dEmdf
dU pdV

T
T

(general expressions for entropy change)

(3)

T2
p
mR ln 2
T1
p1

(entropy change in a perfect gas model)

(4)

PGM

S mc p ln

The combination of energy and entropy called exergy, measuring the maximum work obtainable from a
system, or the minimum work required to reach a global non-equilibrium state, is, for a control mass:
Wumin Wu

Suniv 0

E p0 V T0 S

(5)

Many thermodynamic variables (some extensive, others intensive) are used to simplify the analysis of
systems, amongst which enthalpy, thermal capacities, dilation and compressibility can be distinguished:
HU+pV, c p T

s
T

h
1 V
s
u
1 V

, cv T
,
,
T p
V T p
V p
T v T v

(6)
T

But it must be remembered that, for pure substances, there are only two independent state variables, and
all others can be obtained by a combination of this two. In particular, the ideal gas equation of state is
omnipresent in all combustion studies:
pV=nRT with R=8.314 J/(molK), or in the form

p
RT

or v

RT
p

(7)

where the same symbol is used for the universal gas constant, Ru=8.3 J/(molK), and the particular gas
constant, R Ru/M, for a gas of molar mass M. Most combustion processes take place within a control
volume, the combustor, and thus the equations developed in Control Volume analysis must be known,
particularly the omni-present steady-state mass and energy balances:
0

openings

vA
me with m

0 Q W

(8)

openings

me hte

(9)

4
W being the shaft work input, Q the heat received, and hte the total specific enthalpy for each entrance
(or exit) with mass flow rate me . There are few combustion problems were equations (7-9) are not
involved.

Phase changes in pure substances are needed in combustion only to deal with pure liquid fuels, and it may
be enough to recall the equation for the vapour pressure curve, i.e. Clapeyrons equation, more usually
used in the form of Antoines fitting:

dp
dT

sat

hlv
Tvlv

ln

p hlv 1 1
p

A
ln
T
p0 RT T T0
p0

B
Tunit

(10)

The Thermodynamics of mixtures is so important to combustion, that a more detailed summary is here
included.

THERMODYNAMICS OF MIXTURES
Combustion always involve a mixture, at least of a fuel and an oxidiser (usually air, another mixture
itself!), and the exhaust is always a mixture of burnt gases (except in the two ideal cases C+O 2=CO2 and
H2+(1/2)O2=H2O). The general theory of Thermodynamics of Mixtures is developed aside, but the
important points to combustion are summarised here.
EQUILIBRIUM
An isolated mixture with conservative amounts of substance ni tends to reach an equilibrium state defined
by its entropy S(U,V,ni) being a maximum, what implies, in absence of external fields, that the
temperature T=U/S, pressure p=TS/V, and chemical potential for each conservative species
i=TS/ni, are uniform at equilibrium (in absence of external forces). Multi-phasic mixtures, and
mixture segregation due to external force fields, can be found aside.
CHEMICAL POTENTIAL EXPRESSIONS
The chemical potential, i, measures the tendency for an species to migrate, i.e. the escaping tendency of
chemical energy, in a similar way as temperature measures the escaping tendency of thermal energy, and
pressure the escaping tendency of compression energy. For a mixture in contact with an infinite
environment at T=constant and p=constant, as the ambient atmosphere, it is better to work with the Gibbs
function for the system G(T,p,ni)U+pVTS=HTS, instead of with the entropy of an isolated system. It
can be deduced, from Gibbs equation and Euler theorem for homogeneous functions, that G=ini; i.e.:
C

G U pV TS H TS ni i

(11)

i 1

The differential form of the Gibbs function is used a lot in thermochemistry:

dG=SdT+Vdp+idni

(12)

Several useful relations can be derived from it. First, the Gibbs-Duhem equation, subtracting (12) to the
total differential of (11), i.e.:
0=SdTVdp+nidi

(13)

Second, the general dependence of i(T) and i(p); the former comes from the equality of the crossed
second derivatives 2G/(Tni)=si=2G/(niT)=i/T and from gi=hiTsi, what yields:

T
1

hi

(14)

p , ni

known as vant Hoff equation; on the other hand, i(p) comes directly from the equality of the crossed
second derivatives 2G/(pni)=vi=2G/(nip)=i/p; i.e.:
i
p

vi

(15)

T , ni

For an ideal gaseous mixture, IGM, the chemical potential for an species i takes the form:

i(T,p,xi)IGM=i(T,p,1)+RTln(p/p)+RTlnxi

(16)

indicating that the chemical potential varies with temperature according to the enthalpy function (13), and
varies with pressure logarithmically according to (12) and vi=RT/p, and varies with the molar fraction xi
also logarithmically.
THERMAL CAPACITY AVERAGING
The perfect gas model, besides the ideal gas equation of state (7), assumes constant thermal capacities cp,
simplifying the computations a lot; but a good averaged value of cp must be taken in combustion studies,
since cp varies considerably: e.g. for air at 300 K cp=1000 J/(kgK)=29 J/(molK), but at 3000 K cp=1240
J/(kgK)=36 J/(molK), and for water-vapour at 300 K cp=1900 J/(kgK)=34 J/(molK), but at 3000 K
cp=3100 J/(kgK)=56 J/(molK). Figure 2 gives a plot of cp(T) for the most important combustion gases;
from that, averaged values that may be used for preliminary computations are presented in Table 1.

Fig. 2. Variation of thermal capacity with temperature for gases of interest in combustion. For
preliminary computations, the usual averaged values are presented in Table 1.
Table 1. Mean values used for constant-thermal-capacity models in combustion.
Gas
cp [J/(molK)]
Diatomic molecules: N2, O2, CO, NO, H2, OH
34
Water vapour: H2O
47
Carbon dioxide: CO2
54
Monoatomic molecules: Ar, He, H, O, N
21
For a more crude manual analysis, the drastic simplification of assuming an averaged thermal capacity for
the exhaust mixture of say cp=36 J/(molK) may be cost effective. On the other hand, for high-precision
computations, temperature correlations for thermal capacities must be used, as the traditional JANAF
tables, or other polynomial fitting.

6
WATER VAPOUR CONDENSATION
Except in a few theoretical cases (C+O2=CO2, CO+(1/2)O2=CO2, S+O2=SO2, etc., that really take place
also with some water vapour), water is a genuine combustion product. Leaving aside the important
problem of acid rain formation in the combustion of sulfur-containing fuels, water is the only condensable
component in the products, being produced in such sizeable amounts (typically xH2O10%) that its
detailed account on mass and energy balances is of paramount importance).
Gaseous mixtures with a condensable component are dealt with in detail for the case of humid air aside.
Here we are concerned mostly with humid flue gases, but the approximation of non-condensable products
by air may be good enough; if not, the only modification is in the thermal capacity of non-condensable
gases, that should be changed from cp=1000 J/(kgK)=29 J/(molK) for air at about 15 C to cp=1100
J/(kgK)=32 J/(molK) for a typical exhaust mixture at about 50 C; the molar mass changes even less
because the increase due to the formation of carbon dioxide practically compensates with the decrease
due to water formation.
Recalling that liquid-vapour equilibrium, for ideal two-phase mixtures, implies uniform chemical
potential for each species I, one arrives at Raoult's law (see Mixtures):
xi,vap/xi,liq=pi*(T)/p

(17)

i.e. the molar fraction of species i in the vapour phase, divided by the molar fraction in the liquid phase, is
equal to the vapour pressure for the pure i-component at that temperature, pi* (obtained from Clapeyrons
equation or Antoines correlation), divided by the pressure of the mixture, p. It often helps to think of the
pressure of the mixture as the summation of partial pressures attributed to each i-component, pixi,vapp, so
that at liquid-vapour equilibrium pi=xi,liqp*(T).
For the case of one condensing species (water), the important equations are:
Maximum water-vapour fraction, xvap (notice that vap now refers to the substance and not only
to the phase) dissolved in a given flue gas at given pressure p, and temperature T:
xvap=pi*(T)/p

Dew point, i.e. condensing temperature, Tdew, for a given pressure p, and water molar fraction
xi,vap, it is solved from:
xi,vap=pi*(Tdew)/p

(18)

(19)

Total pressure, p, for the two-phase system at equilibrium (humid exhaust plus condensate) at
temperature T:
p=pi,non-cond+pwater*(T)

(20)

THERMOCHEMISTRY
A detailed thermodynamic treatment of generic reacting systems can be found aside. We will focus here
just in combustion reactions of simple fuels: hydrogen, carbon, and pure hydrocarbons (commercial fuels
are complex mixtures of hydrocarbons; see Fuel properties).

7
STOICHIOMETRY, EXTENT AND AFFINITY
A combustion process involves a set {Mi, i=1..C} of chemical species i, identified by their molecular
formula Mi (Mi is also used for their molar mass), undergoing a simultaneous set of chemical reactions
{Ri, i=1..R}, each one specified by its so-called stoichiometric equation:
C

i 1

i 1

i 1

ir' M i ir'' M i or 0 ir M i for r=1..R (e.g. H2 21 O2 H2O )

(21)

where the first form is preferred for kinetic studies when a direction in the process is implicit (it is said to
occur from reactants (left) converting into products (right), and sometimes an arrow is used instead of the
equal sign), whereas the last form is more simple for equilibrium studies where no direction is privileged.
In any case, (21) serves as the mass conservation equation if Mi is the molar mass of species i, and serves
to build the set of elementary conservation equations when the molecular form of Mi is considered. The i
are called stoichiometric coefficients for that reaction. Notice that the stoichiometric coefficients change
if one writes 0=2H2+O2-2H2O, or 0=H2+(1/2)O2-H2O, or 0=H2O-H2-(1/2)O2.
Most commercial fuels are hydrocarbons (chemical notation should be briefly refreshed, e.g. to
distinguish between n-octane and iso-octane). According to the stoichiometric ratio for full oxidation of a
fuel, air/fuel mixtures fed to a combustor are classified as:
Lean mixtures (little fuel content, excess of air).
Stoichiometric mixtures (with the precise or theoretical amount of fuel as established in a given
reaction as (21)).
Rich mixtures (more fuel than needed, but excess fuel will pyrolise to small-molecule fuels, and
only small molecules appear at the exhaust).
The ratio air-to-fuel in molar base is Annair/nfuel, and in massic base Annair/nfuel, although it is often
stated simply as A (but notice the values are different, though non-dimensional; e.g. theoretical air ratio
for
methane,
CH4+2(O2+3.76N2)=CO2+2H2O+7.52N2,
is
An=2(1+3.76)/1=9.5,
and
Am=AnMa/Mf=9.50.029/0.016=17, sometimes written as A=9.5 mola/molf=17 kga/kgf). The relative air-tofuel ratio is with respect to the stoichiometric air-to-fuel ratio A0; i.e. A/A0 (now independent of the
molar or massic base). In the USA, the inverse of these functions are often used, namely the fuel-to-air
ratio, f=1/A (either in molar or massic base, as above), and the equivalence ratio =1/.
As for the products of combustion, the molar fractions are always used (also said 'volume fractions' since
they are the same with the ideal-gas mixture model predominant in combustion exhaust). Notice,
however, that, to avoid condensation of water inside the instruments, measurements of exhaust gases are
taken on a dry mixture that is obtained by passing the exhaust gases through desiccants (an ice bath is
often good enough, since most water is condensed and left behind).
When a control-mass mixture with initial composition ni0 reacts, the extent of the reaction at any later
time, (Greek letter xi, also called progress or degree of advancement of a reaction, a convenient
normalised mole-account system) is defined as:

ni ni 0

for a given 0 = i Mi

(22)

The extent of a reaction marks the state of progress; to measure the tendency to progress, the chemical
affinity, A, is defined:

8
C

A i i

(23)

i 1

such that Gibbs function variations are:

C

dG SdT Vdp i dni SdT Vdp Ad

(24)

i 1

the minus sign being introduced in the definition of A to ensure that natural evolution (entropy increase in
an isolated system, or Gibbs-function decrease in a system at T and p constant) corresponds to a positive
affinity, i.e. dG=Ad0 progress, i.e. d/dt>0, only if A>0. It is important to notice that reactions with
negative affinities can naturally progress only at the expense of other reactions with positive and larger
affinities, since the real limit, at T and p constant, is dG=Ad0 for the whole set of reactions.
Although both, affinity and extent, are thermodynamic state functions, the reaction rate d/dt is not,
similarly to heat transfer. Reaction rates depend a lot on the presence of catalysts, that have no influence
on reaction equilibrium.
ENTHALPY OF FORMATION AND ABSOLUTE ENTROPY
Compounds are not conserved-entities in combustion, but atoms are. Thus, energy and entropy reference
states for each atom must be agreed in reacting systems, instead of the energy and entropy reference state
for each compound used in non-reacting systems. References are free to choose by the observer, but they
must necessarily be consistently related. The best references are:
Enthalpy reference: zero enthalpy is assigned to the most stable natural form of each chemical
elements at standard temperature and pressure (T=298 K and p=100 kPa). Enthalpies for nonelementary species at standard temperature and pressure, called standard enthalpies of formation,
hf, are experimentally measured (usually by calorimetry, most of the times indirectly) and
tabulated. Enthalpy of an species at other temperature and pressure are computed from the general
relation dh=cpdT+(1T)vdp, according to the model used (e.g. for perfect gases h=hf+cp(TT)).
Entropy reference: zero entropy is assigned to each species (not just to each elementary species but
compounds too) at 0 K and any pressure, since it is an experimental fact explained by information
theory that entropy changes in the limit T0 also tend to zero. Entropy values at the standard state
(T=298 K and p=100 kPa), known as absolute standard entropies, s, are computed by
integration of experimentally measured (usually by spectrometry) thermal capacity data, according
to ds=(cp/T)dTvdp, and tabulated.
It is customary to include in the thermochemical tabulation not only hf and s, but also the standard
Gibbs function of formation of compounds from their elements, gf, although it is redundant since:
C

g f h f T i si for the reaction of formation of the compound: 0= i Mi . (25)

i 1

HEATING VALUE
The standard practical heating value (PHV, also called practical calorific value, PCV) of a fuel is defined
as the heat transfer to cooling water in a Junkers-type calorimeter, i.e. a flow calorimeter with inputs (fuel
and air mixture, and cooling water) and output approaching standard conditions (298 K and 100 kPa), i.e.
the specific combustion enthalpy, hPHV (or its molar value). The fuel is burnt with excess air for complete
combustion, since the heating value is independent of excess air, while the cooling flow-rate is traded off
for best resolution (least uncertainty in Q mw cw Tw mf hPHV ).

9
However, in order to provide a common reference to allow direct combination of heating values for
several reactions, a theoretical standard is defined by assuming that inlet and outlet flows run through
separated pipes for each pure chemical species. Thus, a so called standard higher heating value (HHV or
hHHV, sometimes known as high calorific value HCV) is defined as the heat release to the ambient when
fuel, oxidiser and products flow at 298 K, 100 kPa and species-separated), and computed by adding the
latent heat of the water-vapour (exiting the calorimeter above-mentioned), to the PHV. For instance, the
HHV for methane is (from the Combustion Data Table) hHHV=890 kJ/mol (hHHV=55 MJ/kg), meaning that
if a steady flow of methane is made to burn with air (or oxygen) at 100 kPa, both inlets (methane and air)
being at 298 K, and the outlet is cooled to 298 K, and only CO2 and H2O were produced and all the water
in the exhaust were condensed, 890 kJ by mole of methane will be released to the ambient (55 MJ by kg
of methane). In practice, there will be just one tail-pipe and at 25 C the water will be some 95%
condensed (forming a mist of micrometric droplets) and some 5% dissolved in the gas stream, and some
traces of unburnt hydrocarbons and possibly soot will also show up in the exhaust, and the measured
heating value would be slightly smaller.
However, for applications where the exhaust gases are above their dew point (a very common situation
since the exhaust dew point is typically below 60 C), it is advantageous for the calculations to establish a
different standard, called lower heating value (LHV or hLHV), that assumes all water exits at 100 kPa and
25 C but in the gaseous state (a virtual state not possible in practice for pure water); thus, the LHV is
equal to the HHV minus the vaporisation enthalpy of the exhaust water at 25 C (2.44 MJ/kg or 44 kJ/mol
of water produced). For instance, for methane, its combustion being CH4+2O2=CO2+2H2O, the LHV is
hLHV=890244=802 kJ/mol. Care is needed when using tabulated heating values, to know if they are
HHV or LHV.
Notice that the heating value of a fuel depends on input and output conditions and the path followed; heat
is a path variable, not a state function; HHV and LHV are defined as state functions, the negative of the
reaction enthalpy (hLHV=hr,comb), being for a given constant-pressure process. Fortunately for combustion
reactions, the heat release is practically independent on pressure, and depends very little with
temperature; even the difference between burning at constant pressure or at constant volume is not
significant. Notice also that the heating value of a fuel must be understood as the heat released in its
combustion with air (or any inert mixture with oxygen). Finally, recall that it has no sense to talk about
the 'energy content' of a fuel or of any other system, since energy is only defined between two states of a
closed system, by EW|Q=0; similarly, heat is only defined as energy transfer through an impermeable
boundary, by QEW.
For combustion at constant pressure, be it in a flow system (the most usual case) or in a control mass (e.g.
in an ideal Diesel cycle), the energy balance is:
Q=H=nfuelhHHV

(26)

assuming all water is in condensed form. For combustion at constant volume of a control mass (e.g. in an
ideal Otto cycle), the energy balance is Q=E=U, but since the available data are enthalpies,
Q=(HpV)=HVp. If there is no condensed matter, i.e. for initial and final gaseous states, pV=nRT
and Q=HVp=H(nfinalninitial)RT; for instance, if 1 mol of CO is burnt at constant pressure in air
at 298 K and 100 kPa, the heat release is hHHV=hLHV=283 kJ/mol, whereas if the initial state is the same
but the burning is inside a rigid vessel at constant volume, the heat release may range from 283 kJ/mol for
a very diluted mixture to 283103+(11.5)8.3298=282 kJ/mol for an stoichiometric CO/O2 mixture. If
there is some condensed matter, either initially or at the end, a more involved analysis is required,
although, if the saturation pressure of the condense matter at the standard temperature is small, the same
equation as for gaseous mixtures may be applied just considering the non-condensed species; for instance,

10
if 1 mol of H2 is burnt at constant pressure in air at 298 K and 100 kPa, the heat release would be
hHHV=286 kJ/mol if the exhaust water were fully condensed, hLHV=242 kJ/mol if all the exhaust water
were in its vapour state, or a value close to the HHV in the practical case, whereas if the initial state is the
same but the burning is inside a rigid vessel at constant volume, the heat release may range from 242
kJ/mol for a very diluted gaseous mixture, to 286103+(01.5)8.3298=284 kJ/mol for an stoichiometric
H2/O2 mixture, that would produce near-zero gaseous moles. In summary, the heat release slightly varies
from constant pressure to constant volume combustion, the larger variation being with the physical state
of water at the end: for long-chain hydrocarbons the LHV is 6.5% lower than the HHV, the difference
increasing with decreasing carbon content (6.9% for C8H18, 9.1% for CH4 and 15.4% for H2).
All the above-defined heating values are global values for a non-equilibrium combustion process between
the initial and the final states quoted, but no difference is found if the instantaneous equilibrium changes
are introduced and the heating value is defined as the enthalpy of reaction (changed of sign because the
assumed direction for the heat release), i.e.

hHHV hr

i 1

i 1

i hi
p ,T

ni hi
nfuel,in

and hLHV hHHV vH2O hlv

(27)

i.e. the ratio of the change of the enthalpy of a reacting system at equilibrium to the change in the extent
of reaction when an infinitesimal advance in the extent of the combustion reaction (scaled per unit of
amount of fuel) takes place. Similarly, the heat release at constant volume can be identified with the
internal energy of reaction at constant volume, or with the internal energy of reaction at constant pressure
for the case: U/|V,T=U/|p,T+U/p|,Tp/|V,T. The value for the molar enthalpy of water, for the
liquid-to-vapour phase change at T=298 K is hlv=44.0 kJ/mol=2442 kJ/kg.
MAXIMUM WORK
A combustion process may exchange both heat and work with the surroundings, only heat, only work, or
nothing of the two. The energy balance of the combustor relates the heat and work flows with the
entrance and exit temperatures.
The work obtainable from the combustion of a fuel depends on the actual process (as does the heat
release); if the combustion is performed in a rigid open chamber (as in a gas-turbine), no work will be
produced (at the combustion chamber), but if it is performed in a chamber with moving parts (as in a
reciprocating engine), some work will be produced. The maximum obtainable work, or the minimum
required work to reverse the process, for a steady process in the presence of an environment at
temperature T0 is the change in flow-exergy from inlet to outlet, =htT0s, that for standard
conditions is:
wmin==gr=hrTsr=ihf,iTisi

for a given 0= i Mi

(28)

where hf,i are standard enthalpies of formation of the participating compounds i=1..C, and si their
absolute standard entropies. For instance, for the theoretical combustion of methane with oxygen,
CH4+2O2=CO2+2H2O, =gr=hrTisi=890103298(214+2701862205)=817103 J/mol,
i.e. 817 kJ/mol of work might be produced (the remainder to the heating value being evacuated as heat, on
account of the energy balance q+w=ht). Similarly, for H2+(1/2)O2=H2O a maximum of
=286103298(701310.5205)=237103 J/mol, i.e. as much as 237 kJ/mol of work might be
generated. If an energy efficiency is defined as the quotient between work produced and heating value,
the maximum efficiency would be 0.92 for methane and 0.83 for hydrogen (but mind that it would be
1.002 for carbon).

11
Recall here that there is a small difference between the maximum work obtainable from combustion and
the maximum work obtainable from the fuel (in spite of ambient-air having no exergy), because the
streams (to be considered separate components, since the thermochemical values refer to separate
streams) have some chemical exergy relative to the atmosphere, even when at T and p. For instance, the
maximum work obtainable from methane is CH4=831 kJ/mol (against the 817 kJ/mol above), computed
by adding the exergy of one mole of carbon dioxide, CO2=20 kJ/mol, and two moles of water,
H2O=1.3 kJ/mol, and subtracting the exergy of two moles of oxygen, O2=3.9 kJ/mol, to the exergy of
the CH4+2O2=CO2+2H2O reaction, |r|=817 J/mol. Similarly, the exergy of hydrogen at the standard
state is H2=236 kJ/mol, to be compared with the exergy from H2+(1/2)O2=H2O, |r|=237 J/mol.
Although an indirect realisation of the same global reaction can approach that maximum work, as in fuel
cells, direct combustion processes cannot approach those values. To begin with, if a constant-pressure
combustion process (the difference with a constant-volume process being small) were very slow, the heat
release would just dissipate to the environment without any possible work, i.e. all the exergy of the
reaction would be lost by entropy generation inside. If the process is rapid, as most combustions really
are, it may be assumed adiabatic to a first approximation, the heating value being spent in heating the
exhaust gases themselves instead of in heat transfer to another medium; no work is actually produced, but
some work may be obtained afterwards, since the exhaust is hotter than the environment.
To evaluate the maximum work obtainable from the adiabatically burnt gases, its exergy, one must
computed the flow exergy, =htT0s, of the hot exhaust stream relative to the atmosphere. It is good
enough to retain just the thermomechanical exergy, since the chemical exergy of dilution into the
atmosphere is much smaller. For hydrogen at standard conditions, for instance, the maximum obtainable
work (236 kJ/mol in an ideal fuel cell) is reduced to some 170 kJ/mol in the adiabatic exhaust at 100 kPa
and 2480 K after isobaric combustion of hydrogen with stoichiometric air, or to some 180 kJ/mol in the
adiabatic products at 870 kPa and 2950 K after isochoric combustion with stoichiometric air, due to the
unavoidable entropy generation inherent to the combustion process (entropy increase without entropy
flow).
Burning with pure oxygen increases the temperature of the adiabatic exhaust and thus its exergy. For
instance, in the case of H2/O2, some 190 kJ/mol of exergy per mole of hydrogen may be obtained from the
very hot adiabatic exhaust at some 4000 K and 100 kPa (difficult to compute because of dissociations). In
spite of this advantage of higher temperatures (already discovered by Carnot in the most general case),
practical combustion is most of the times with common air, even in some excess to further decrease the
temperature achieved, since common materials cannot withstand such high temperatures.
ADIABATIC COMBUSTION TEMPERATURE
The adiabatic model very accurately predicts homogeneous combustion temperatures in gas phase,
because of the very poor thermal conductivity of gases and the low speed (<0.5 m/s laminar burning
speed for ordinary hydrocarbon/air mixtures). The adiabatic model, however, overestimates the
temperature in the case of high radiation emission in non-premixed flames, and for high speeds flows
relative to solids (as in afterburner heat-exchangers).
The adiabatic temperature is a thermodynamic state variable that results from the conversion of internal
chemical energy to internal thermal energy after the combustion process takes place, and depends on the
actual fuel and oxidiser (e.g. CH4/air), their ratio (e.g. lean, stoichiometric or rich mixtures), their initial
thermodynamic state (e.g. premixed gases at 298 K and 100 kPa, although energies associated to mixing
and pressure are negligible), and the type and proportions of the compounds formed. It can be measured
at the exhaust of an adiabatic combustor, or just after a flame, but the practical difficulties in ensuring

12
minimal heat losses, particularly from the thermometer probe, render a theoretical computation more
precise that the actual measurement.
The problem with the computation is that the type and proportions of the compounds formed must be
known, what is partially solved assuming a set of compounds dictated by experience, and assuming that
they are in thermodynamic equilibrium at that pressure and temperature. But, even for the simplest case
of a constant-pressure combustion process (very good for open combustors since the pressure only rises
some 2% after a deflagration), the temperature is the unknown result and thus the equilibrium
composition cannot be directly computed, yielding an implicit and stiff algebraic mathematical problem.
The energy balance for a generic reactor is:
openings
openings
C
dE
C

W Q n j h j W Q n j xij hf ,i xij (hi hf ,i )

dt
j 1
j 1
i 1
i 1

(29)

where the enthalpy contributions at each opening j is split in a chemical part (proportional to the molar
fraction i of each chemical component of standard enthalpies of formation hf,i, and a thermal part (from
the standard state to their actual state at the opening) of the participating compounds i=1..C. Notice that
the components are assumed known at entry and exit openings, but their molar fractions may be functions
of the temperature and pressure (e.g. when there is chemical equilibrium at an exit).
A drastic simplification of the energy balance occurs if one assumes complete combustion, because then
one can resolve the exhaust composition independently of temperature. This approach is useful for
combustion of lean mixtures in air, but not for combustion in pure oxygen where temperatures are higher.
With this simplification the adiabatic temperature for a stoichiometric H2/air mixture, for instance, is
computed to be 235050 K, in good agreement with a measure of 2300100 K, whereas the adiabatic
temperature for a stoichiometric H2/O2 mixture is computed to be 5400200 K, in clear disagreement
with a measure of 3000100 K.
With this complete-oxidation model, separating the entry flows and the exit flows, assuming W 0 (no
work input/output through the impermeable walls), and assuming there is no phase change from the
standard state to the actual input or output states (see below), the energy balance for a steady state
combustor is
C

0 Q nentry xi hi xi c pi (T T ) nexit xi hi xi c pi (T T )
i 1

i 1

(30)

Notice that water must be treated as an ideal gas to apply (30); otherwise, the enthalpy of phase change
must be added.
The actual mass balance or, better, molar balance for a combustor may be established in different ways,
according to the scale of amount of substance chosen. The main choices to write the so called mixture
equation are: per unit amount-of-substance of input stream, per unit amount-of-substance of output
stream, and per unit amount-of-fuel in the input stream. The last two usually take the form:
C

i 1

i 1

aM fuel bM air cM other xi M i and M fuel AM air BM other i M i

(31)

13
where a, b, c, A and B are numeric factors (A=b/a is the air/fuel ratio), Mi the molar masses of each
component (associated to the molecular formula stated; e.g. aC+bAir=xCOCO+xCO2CO2+xO2O2+xN2N2), xi
are the molar fractions in the exhaust (xi=1), and i are the amounts of exhaust species per unit amount
of fuel (not the stoichiometric coefficients).
Choosing the first option in (31), i.e. per unit amount-of-substance exiting, the energy balance (30) with
the definition of LHV in (27), yields:

0qa h

fuel

air

other

c p fuel (T T ) b h c pair (T T ) c h

q ahLHV (ac p fuel bc pair cc pother )(Tin T )

x h

c pother (T T )

Cexhaust

i 1

xi c pi (Tout T )

Cexhaust
i 1

c pi (T T )

(32)

which can be read as follows: the thermal energy in the exhaust stream (the last term) is the contribution
of the external heat added through the walls per unit of exhaust q (zero for an adiabatic process), plus the
chemical heating value of the fuel input (the lower heating value, hLHV, since water is taken as a gas), plus
the thermal energy in the intake stream.
The adiabatic combustion temperature is obtained from (32) with q=0:

Tad T

ahLHV (ac p fuel bc pair cc pother )(Tin T )

Cexhaust

i 1

xi c pi

ahLHV
c pair

(33)

where the last simplification (a most crude approach) neglects the thermal enthalpy input, and
approximates the thermal properties of the exhaust gas mixture by those of air, i.e. cp=1000 J/(kgK)=29
J/(molK) for cold air, or better cp=34 J/(molK) to account for the growth of thermal capacity with
temperature. Of course, the best precision is obtained with the integration cp/T)dT instead of the
approximation cpT in (30), but the first form of (33) with cp=34 J/(molK) for any diatomic molecule,
cp=47 J/(molK) for water vapour, and cp=54 J/(molK) for carbon dioxide (Table 2), is thought to be the
best compromise on precision/effort for non-programmed computations.
One may wonder why all vales of the maximum adiabatic combustion temperature in air are so close
(2300200 K, from combustion data Table), in spite of the widely different heating values (e.g. hLHV=10
MJ/kg for CO and hLHV=(1422.4)=140 MJ/kg for H2). A simple explanation can be found for
hydrocarbons, CnHm; their LHV may be approximated by the LHV of nC+(m/2)H2, which is
(n394+m242/2) kJ/mol. The adiabatic jump is TadT=hLHV/nicp,air, and it happens that the amount of
substance of the products varies almost proportional to the molar heating value. For the complete
stoichiometric combustion in air, CnHm+(n+m/4)(O2+3.76N2)=nCO2+(m/2)H2O+3.76(n+m/4)N2, and
taking a mean value cp,air=40 J/(molK) for the exhaust gas, the total mole number in the products is
n(1+3.76)+m(1/2+3.6/4), which makes TadT=(n394+m121)/(n0.190+m0.058) K, a biparametric
function almost constant, of value TadT2000 K.
When dissociation is important due to the high temperatures associated to near-stoichiometric
combustion, or due to lack of oxygen in rich flames, computation of adiabatic temperatures is coupled to
computation of equilibrium composition. The effect of pressure on adiabatic temperature is through its
effect on equilibrium composition: at high pressure dissociation is negligible and the completecombustion adiabatic temperature gets its highest value, whereas reducing the pressure increases
dissociation in accordance with Le Chtelier principle, lowering adiabatic temperature.

14
EQUILIBRIUM COMPOSITION
The intake to a combustor is not at equilibrium (it would not react if it were), although it may be
considered at a metastable equilibrium (it would not react if not ignited), or even at separate fuel and air
inlets at their own equilibrium conditions before mixing. The exhaust from a combustor may be at
equilibrium (not with the surroundings but within itself, i.e. in chemical equilibrium at a given T and p), if
the combustion process has had time enough to develop completely. In practice, there are always some
partial reactions so slow that they do not reach equilibrium in the times considered, as for some trace
contaminants, but the assumption of perfect chemical equilibrium is a very good approximation to the
overall real combustion process, and, on top of that, serves to explain why a further simplification, the
complete combustion model, is an acceptable approximation to many combustion processes.
Chemical equilibrium in absence of external force fields implies no gradient of chemical potential, i=0,
for each component i=1..C, and no affinity to go on, Ar=0, for each reaction r=1..R. The latter establishes
a relation between molar fractions in the equilibrium composition xi, and T and p, for a given reaction
0=iMi, that for a reacting ideal gas mixture is (see Chemical reactions):

p
xi

i 1
p
C

p
K (T , p )
p

g
h T
exp r r 1

RT
T
RT

(34)

where the approximation lnK=A+B/T for the reaction constant K has been applied, and the constants gr
(standard Gibbs function of reaction) and gr (standard enthalpy of reaction) are computed from the
standard enthalpies and Gibbs functions of formation by:
C

i 1

i 1

hr i h fi , g r i g fi

(35)

For systems that are not at equilibrium yet, the ratio calculated from the mass-action law is called a
reaction quotient Q. The Q-values of a closed system have a tendency to reach a limiting value over time,
the equilibrium constant K.
Back to Combustion

COMBUSTION KINETICS
Combustion kinetics ................................................................................................................................. 2
Physical mixing and its effects on ignition, propagation and extinction .................................................. 2
Mixing................................................................................................................................................... 3
Mixture ratio specification ................................................................................................................ 3
Diffusion ........................................................................................................................................... 3
Convection ........................................................................................................................................ 4
Evaporation ........................................................................................................................................... 5
Evaporation from a planar surface .................................................................................................... 5
Droplet evaporation .......................................................................................................................... 7
Ignition and extinction ........................................................................................................................ 10
Igniter types .................................................................................................................................... 12
Propagation of non-premixed flames...................................................................................................... 13
Gas fuel jet .......................................................................................................................................... 13
Flame length ................................................................................................................................... 13
Condensed fuel ................................................................................................................................... 14
Flash point (ignition) ...................................................................................................................... 14
Flame stabilisation by porous feeding: oil lamps and the candle flame ......................................... 15
Fuel sprays ...................................................................................................................................... 16
Droplet combustion ........................................................................................................................ 18
Particle combustion ........................................................................................................................ 24
Propagation of premixed flames ............................................................................................................. 25
Laminar combustion ........................................................................................................................... 26
Deflagration speed .......................................................................................................................... 27
Flame thickness .............................................................................................................................. 29
Flame stabilisation .......................................................................................................................... 30
Flame quenching............................................................................................................................. 31
Flammability limits ......................................................................................................................... 31
Autoignition temperature ................................................................................................................ 33
Supersonic combustion ....................................................................................................................... 34
Turbulent combustion ......................................................................................................................... 36
Axisymmetric turbulent jet flames ................................................................................................. 38
Turbulent premixed flames ............................................................................................................. 38
Chemical kinetics.................................................................................................................................... 39
Reaction mechanism and reaction rate ............................................................................................... 39
Mass action law .............................................................................................................................. 40
Types of elementary reactions ........................................................................................................ 41
Arrhenius law.................................................................................................................................. 42
Collision theory .............................................................................................................................. 43
Relation between rate coefficients and equilibrium constants ....................................................... 44
Kinetics of NOx formation .................................................................................................................. 45
Catalysis .............................................................................................................................................. 46
The three-way catalytic converter .................................................................................................. 46
Selective catalytic reduction (SCR) ................................................................................................ 47
Catalytic combustion ...................................................................................................................... 47

COMBUSTION KINETICS
Thermodynamic laws, establish limits to natural and artificial processes, i.e. bounds to the possible paths,
but the path actually followed, and the pace (the process rate), depends on other circumstances. For
instance, Thermodynamics does not say that a piece of paper will burn in air, not even after being ignited,
and does not deal with the burning rate; it just says that the system paper/air might reach a more stable
equilibrium state (more entropy) by burning, and determines that end state (which might be reached also
by secularly-slow oxidation).
It is Kinetics science which deals with how fast things happen: instantly (i.e. more quickly than
monitored, as in explosions), evolving at a sizeable pace (i.e. in the monitoring time-span, as in
combustion), or at a negligible rate (i.e. more slowly than monitored, as in slow oxidation). For instance,
a piece of paper enclosed in a transparent container with more air than the theoretical one, may not burn
completely if ignited (e.g. by a concentrated light), because, as oxygen concentration gets reduced,
convection and diffusion might not supply enough oxygen to maintain the minimum heat release needed
for propagation. Hydrocarbon fuels cannot burn in N2/O2 mixtures if xO2<12% (<5% for H2 fuel), and
hydrocarbon fuels cannot burn in CO2/air mixtures if xO2<15% (<6% for H2 fuel). At room temperature,
without ignition, the piece of paper in air will oxidize very slowly (unnoticeable to the eye).
Topics covered below are: rate of physical mixing and its effects on ignition, propagation and extinction,
and rate of chemical reaction once mixed (including a review of reaction mechanisms). Some basic
models of global combustors and of flame structure are dealt with apart.

PHYSICAL MIXING AND ITS EFFECTS ON IGNITION, PROPAGATION AND

EXTINCTION
It cannot be stressed enough that kinetics is what finally controls combustion (or any other reaction);
Thermodynamics indicates if the reaction is natural (i.e. may proceed in an isolated system) or artificial
(i.e. requires some exergy input from outside). Thermodynamics says that a fuel and air may naturally
react, but if the kinetics is too slow, an observer concludes that there is no reaction. Some illustrative
examples are: the burning of a piece of paper or a candle inside a closed container; the fuel will not burn
completely if there is not enough room inside for a good air-convection to the flame, even if with more
than stoichiometric air is enclosed.
A vivid example of the controlling effect of mixing is presented in Fig. 1, showing two similar candle
flames, one on the ground and the other on a space platform (no buoyancy effects); candle flames on earth
are long, slender and yellow, whereas under weightlessness they are nearly spherical, blue, and burn
much slower (sometimes get extinguished), due to the lack of air draught by buoyant convection.

Fig. 1. Candle flames on earth (left) and on a space station under weightlessness (right).
One of the basic features of combustion is its self-spreading power: under usual conditions, a fuel/air
mixture, once ignited, generates a flame-front that sustains itself, i.e. that transmits the activation to the

fresh mixture through heat and mass transfer. Two limit cases can be considered for this propagation
according to the state of the mixture: combustion propagation when fuel and air are at each other side of
the flame (i.e. for a non-premixed mixture), and combustion propagation when both fuel and air are at the
same side of the flame (i.e. for a premixed flame).
Mass transfer is essential to combustion, which is a special case of combined heat and mass transfer
reacting system; and not only inside the combustor itself, but for other combustion-related reactors as in
the after-burning catalysts, fuel reformers, etc.

MIXING
Mixing is a pre-requisite for combustion. Mixing (i.e. decreasing bulk differences), is a natural process
(i.e. it does not require an energy expenditure), so that, if fuel and oxidizer gases are brought to contact
and enough time allowed, a perfect mixing would take place in their energy level (temperature), relative
speeds and chemical composition (with the natural stratification in the presence of gravity or another
force field).
But mixing is a slow physical process if not forced by convection (large-scale transport) and turbulence
(large-scale to small-scale transport). Turbulent mixing is the rule in all practical fluid flows at scales
larger than the millimetre, from the piping of water, fuels, gases..., to the wakes behind vehicles of any
sort, to all atmosphere, ocean and stellar motions (there are some exceptions, as the laminar diffusing
contrails left by jet aircraft)..
Two extreme cases of mixing are considered in combustion: combustion in a premixed system (prepared
well-before-hand, or well-stirred), and combustion in the common-interface layer where non-premixed
fuel and air come into contact. For premixed combustion, the mixture ratio specification is established
before-hand, whereas for non-premixed combustion, the mixture ratio specification depends on the actual
feeding flow-rates of fuel and air.
Mixture ratio specification
Mixture ratio specification may use different units: molar fraction of fuel in the mixture xF, mass fraction
yF, fuel-to-air ratio f (molar or mass), air-to-fuel ratio A (molar or mass), equivalence ratio (the actual
fuel/air ratio relative to the stoichiometric one, with the same value in molar and mass basis), air relative
ratio (the air/fuel ratio relative to stoichiometry, with the same value in molar and mass basis), mixture
fraction (mass-flow rate ratio of injected fuel to mass flow rate of products), etc.
Diffusion
Actual mixing of chemical species is governed by mass transfer laws, very similar to heat transfer laws
for conduction (diffusion) and convection. In a homogeneous media, without phase changes or chemical
reactions, the basic kinetic law for mass diffusion is Ficks law:

ji = Di i (Ficks law, similar to Fouriers law for heat transfer q =k T ) (1)

ji m i A being the diffusion-mass-flow-rate of species i per unit area, Di the mass-diffusivity for species
i in the given mixture, and i=yi the mass-density of species i in the given mixture. Notice that only the
flux associated to the main driving force is considered in Eq. (1), i.e. mass-diffusion due to a speciesconcentration gradient; there are also secondary fluxes associated to other possible gradients (e.g. massdiffusion due to a temperature gradient, known as Soret effect, and mass-diffusion due to a pressure
gradient; alternatively, there may be heat-diffusion due to a species-concentration gradient, known as
Dufour effect, and heat-diffusion due to a pressure gradient), but most of the times those cross-coupling
fluxes are negligible. Besides, selective force fields may yield diffusion (e.g. ions in an electric field).

Notice that only molecular diffusion is considered here, i.e. for particle sizes <10-8 m; particles in the
range 10-8.. 10-6 m (soot, mist, smoke), are studied with Brownian-motion mechanics, and particles >10-6
m with Newton mechanics.
To better grasp the similarity between species diffusion and heat diffusion, the balance equations for
mass-transfer and heat-transfer, applied to a unit-volume system, are here presented jointly:
Magnitude
Chemical species i
Thermal energy

Accumulation
Production
wi
yi
=

T
t

c p

Diffusive flux Convective flux

+ Di 2 yi
( yi v )
(2)
+ a 2T

(Tv )

(3)

with wi being the mass-production rate by chemical reaction, and a=k/(cp) the thermal diffusivity. Notice
that there is only one driving heat-transfer-function, T, but many mass-transfer functions, yi (one for each
species), although most problems are modelled as a binary system of one species of interest, i, diffusing
within a background mixture of averaged properties. Typical values for Di and a are given in Mass
diffusivity data. Because of the nearly-equal values of Di and a for gases (Dia10-5 m2/s), the thermal
and solutal relaxation times in absence of convection, are nearly equal (trelL2/DiL2/a), in spite of the
fact that the coefficients in (1) are widely different (Di10-5 m2/s and k10-2 W/(mK)), T is not bound,
and i is bound to i<<1 kg/m3 for diffusion in air under normal conditions. In liquids and solids, mass
diffusion relaxation times are much smaller than their thermal counterparts.
Convection
A quicker mixing process than diffusion is convection, where bulk fluid-flow transports species as if
encapsulated, instead of having to migrate by its own random fluctuations.
As in the study of convective heat transfer, the fluid flow should be solved in conjunction to diffusion,
but, as in heat transfer, one usually resorts to empirical correlations to compute mass-transfer nondimensional parameters (Sherwood Sh, or mass-Nusselt number) in terms of non-dimensional stimuli
(e.g. Reynolds number of the imposed flow Re, Rayleigh number of the imposed thermal gradient Ra,
etc.). Thus, instead of solving the whole fluid-dynamic problem with (2-3) and momentum equation,
empirical correlations applicable just to the boundary values are often used as:
hm L

=
Sh ( ReL , Sc ) for mass convection
Di

hL
Nu
Nu ( ReL , Pr ) for heat convection
=

Sh

(4)

where both, the solutal and thermal convection correlations are sketched (Sc/Di, Pr/a). Notice,
however, that the boundary conditions in practical solutal-convection problems can be very different to
the classical heat-convection problems where a single fluid sweeps a hot or cold rigid boundary. The case
of a submerged jet is a good case of similarity between solutal convection (e.g. a jet of fuel gas emerging
to ambient air) and thermal convection (e.g. a jet of hot air emerging to ambient air).
Phase changing systems are more conspicuous, although most of the times the process of phase change
(e.g. the evaporation of liquid fuels studied below) can be separated from the more complex gas-phase
combustion. But, in the burning of solid fuels the process are entangled because there may be

decomposition reactions within the solid (as in wood burning), or heterogeneous combustion at the
interface (as in coal burning and metal burning). For instance, when aluminium particles burn in a carbondioxide atmosphere (2Al+3CO2Al2O3+3CO), the initial and final phases take place with a detached
vapour flame (at some 2600 K), consuming two thirds of the mass, whereas an intermediate stage takes
place at the surface, controlled by a carbon layer (perhaps Al4C3) formed by heterogeneous reaction of
carbon monoxide there (and not by an alumina layer as thought; Tm(Al2O3)=2320 K)..

EVAPORATION
Evaporation (sometimes called vaporisation) is the net flux of some species, at the interface between a
condensed phase and a gas mixture, due to a normal concentration-gradient of that species in the gas close
to the interface. Examples: the evaporation of water from a glass of water in air (the level decreases some
1 mm/day); the evaporation of ethanol from a glass of wine in air (water evaporates too); the evaporation
of ammonia from an open bottle of water-ammonia solution in air (water evaporates too); the evaporation
of naphthalene in air (sometimes called sublimation). Volatiles liquids and volatiles solids are smelly. It is
often simpler and more efficient to transport fuels in condensed form, and safer to burn them with nonpremixed flames, in which case, fuel droplet evaporation from the injector spray constitutes a first stage
to the combustion of the generated vapours within the oxidiser stream (droplet burning is dealt with
below; we follow on here just with the evaporation process).
Evaporation should not be confused with boiling (which may also be properly called vaporisation), which
is the change of phase within the liquid phase due to an increase in temperature or a decrease in pressure.
Perhaps the most clarifying difference is that boiling is a bulk process (bubbles form at hot points, usually
the walls of a heated container), and may take place in pure substances, whereas evaporation is a freesurface process (no bubbles form, and the interface region cools) that only happens in mixtures.
Evaporation is a basic topic in combustion of condensed fuels, as well as a more general mass-transfer
topic in mechanical engineering (humidification, drying, cooling towers) and chemical engineering
(reactors, materials processing, oxidation, electrochemistry, scrubbing, desalination by reverse osmosis
and other membrane processes, etc.). Evaporation and condensation in a mixture are always combined
mass-and-heat transfer problems (boiling and condensation in a pure substance and the Stefan problem of
melting or solidification, are just heat transfer problems).
We only consider here evaporation of a pure liquid, typically water, in air (a mixture), controlled by
diffusion of both, species and heat. When considering the vaporisation of practical fuel droplets (like
diesel oil), the variation with time of the composition and vaporisation temperature may be very
important due to multi-component equilibrium; however, it is common practice in theoretical analysis to
assimilate commercial fuels to pure-component reference-fuels, usually n-octane for gasolines, and ndodecane or n-tetradecane for diesel oils.
Evaporation from a planar surface
Let us start by the simplest one-dimensional planar diffusion-controlled evaporation problem. Consider a
test-tube with water in open air. Assuming that the air in the tube is quiescent, but the air outside is stirred
enough as to maintain constant conditions at the mouth (T0, p0, 0), and assuming a steady state (the water
level is thought to be kept steady by some slow liquid supply from the bottom; in the real test-tube case,
the liquid level would slowly decrease), Eq. (2) with its initial and boundary conditions would solve the
problem, although the first integration can be skipped directly establishing the series of mass conservation
relations:

m
A

mass-flow-rate
per unit area

d
=
=i v + ji =
v =
i v + vdiffi +
i v Di i =liq vliq (5)
0

dz

mass-flow-rate
mass of i
mass of i

mass-flow-rate
of i per unit area

of other species
per unit area

convected

diffused

mass-flow-rate
of liquid to keep
the level steady

where is the density of the mixture (assumed constant in a first approximation), i=yi, and vliq the
speed of liquid injection (really, the liquid-level descent with time). Notice that, with the constant-density
approximation in the gas phase, the global velocity is also constant along the tube. Dividing by and
integrating, the value of the global velocity is obtained:
1 yi1
ln
1 yi0 yi <<1
yi yi0
d ln (1 yi )
dy
D dyi
v =Di
v =Di
= Di 1
v =yi v Di i v = i
(6)
1 yi dz
dz
dz
z1 z0
z1 z0
yi1 being the mass-fraction of species i at the mouth (assumed to be that of the ambient), yi0 the massfraction at the bottom of the gas column (i.e. close to the liquid surface), assumed to be that of two-phase
equilibrium, i.e. Raoults law (see Mixtures), and z1z0 being the depth (the diffusion length). This result
might have been anticipated by a simple dimensional analysis of the function v=v(Di,yi,z).
For water evaporation in ambient air, the mass-fraction of water-vapour, yi, at the ambient and at liquidlevel equilibrium are related to the pure vapour pressure at that temperature by Raoults' law:

pi* (Tamb ) M i
=
y
at the ambient
i1
p
Mm
xi M i
Mi
=
yi = xi
(7)

xi M i M m y = pi* (Tliq ) M i at the liquid surface

i0
p Mm

where Mi and Mm are the molar mass of species i (the volatile liquid, e.g. water) and the mixture
(practically that of ambient air, for small xs), respectively, and is the relative humidity of the ambient.
The evaporation speed, (6), with the relation v=liqvliq from (5), and the linearization in (6) yields:

pi* (Tamb )
v
Di M i pi (Tliq )
vliq =
=

p
liq liq z M m p
*

(8)

e.g. the evaporation rate for water at 20 C in ambient air at 20 C, 100 kPa and 60% humidity ratio, for a
1 cm diffusion layer, is:

vliq

1.2 kg/m3 2.2 105 m 2 /s 0.018 kg/mol 2.3 kPa

2.3 kPa
0.6
=1.510-8 m/s (1.3 mm/day)

3
2
1000 kg/m
1 10 m 0.029 kg/mol 100 kPa
100 kPa

where Di was found from Mass diffusivity data, and the vapour-pressure value obtained from steam tables
(or from Clapeyron equation, or Antoine fitting). Notice that the evaporation rate increases with vapour
pressure (e.g. ethanol evaporates some 10 times faster than water at 20 C vliq=12 mm/day), whereas for a
less volatile fuel like n-decane it is <0.1 mm/day), and it falls with deepness (e.g. 0.13 mm/day for water
down 10 cm in a test-tube).
It was J. Dalton in 1801, just after he introduced the partial-pressure concept, who first said that the
evaporation rate is proportional to the difference in partial pressure of water vapour from the saturated
boundary layer, to that of the air far aside, and that it increases with free air velocity. The first analytical

model of evaporation is due to J. Stefan in 1872. The effect of wind speed is to decrease the boundary
layer thickness, directly related to the diffusion depth, z, we have assumed known, consequently
increasing the evaporation rate.
Evaporation not only implies a mass transfer but also a heat transfer, since the vaporization enthalpy at
the liquid surface must come from either the liquid side or the air side, at the steady state. This heattransfer implication was not apparent in the numerical application of (8) just done, because the two
temperatures were assumed known, but in reality, the temperature at the liquid interface will depend on
the heat transfer and energy balance (the effect of a possible radiation trapping there, e.g. from sun rays,
will enter here also); for a small amount of liquid the liquid quasi-steady temperature would be the wetbulb temperature (see Humid air), a little below ambient temperature, making the approximation above
acceptable.
The analysis of evaporation in very hot environments (e.g. inside a furnace), gets more involved. To start
with, Raoult's law is no longer applicable to the far field (r) because the liquid could not be in twophase equilibrium with such a hot gas (i.e. T>Tcr, or at least p*(T)>p), although they are still applicable
at the liquid surface because the droplet tends to reach a quasi-steady-state temperature close to its boiling
point. Besides, the linearisation in (6) may no longer be valid, and the large temperature and
concentration variation makes the constant-property assumption less accurate.
Droplet evaporation
Evaporation of a droplet in a quiescent atmosphere is a more practical application of this diffusioncontrolled model, since small drops (say of 1..10 m), as found in mists and spray combustion, quickly
accommodate their speed to that of the surrounding gas flow, suppressing forced-convection effects
afterwards; besides, natural convection effects, due to buoyancy of the lighter humid mixture nearby, is
neglected in this analysis. Convection would increase the evaporation rate, of course.
Consider the system defined by the control volume enclosed by a generic sphere of radius r>r0, r0 being
the radius of the droplet, kept steady by an imaginary supply of liquid at the origin (in reality, the droplet
radius will decrease with time since there is not such an imaginary supply). The expected radial profiles
for the mass fraction and temperature fields are presented in Fig. 2.

Fig. 2. Radial profiles of temperature and mass fraction of the volatile component, in a droplet
evaporation process.
The mass and energy balance for this strictly-steady system (with a liquid source at r=0 and a vapour sink
at r=), can be set as follows:
Magnitude

Accumulation

Mass balance (global)

Production Diffusive flux

=

m L

+0

Convective flux
m ( r )

(9)

Mass balance (volatile species)

m L

+ ADi

di
dr

m ( r ) yi

(10)

Energy balance

m L hLV0

+ Ak

dT
dr

m ( r ) c pi (T T0 )

(11)

where the enthalpy reference has been chosen h=0 for the vapour, i, at the liquid temperature T0 (that is
why a hLV,0 input at the origin is accounted for in (11). Notice that the thermal capacity in the convective
flux must be the vapour one, because the sum ivicpiA reduces to just the vapour term ivicpiA (because
dry air cannot move in the steady state), and iviA=vA= m .
In (9), m L stands for the mass-flow-rate of liquid consumed (the parameter sought), and one concludes
that the global mass-flow-rate of gases across the spherical surface, m ( r ) , is independent of r, and, from
m = vA , the bulk radial velocity v, assuming constant gas density (pressure is constant and
temperature does not vary too much), and with A=4r2, finally yields:

=
m =
vA constant
=
v (r )

v0 r02
r2

(12)

Dividing (10) by m L = vA and upon integration, yields:

Di dyi
r 2 Di dyi
=
yi
=
yi
0 1+
0 1+
v dr
v0 r02 dr

liq dr0
D 1 yi ,0
=
v0 = i ln
dt
r0 1 yi ,

Di dyi
1
= d

2
v0 r0 1 yi
r

dr02
2 Di 1 yi ,0
=
K
ln
liq
dt
1 yi ,

tevap =

2
r0,ini

(13)

i.e. the well-known Langmuirs law (1918), or simply the r2-law, stating that the droplet life-time, tevap, is
proportional to the square of the initial radius; often, the diameter, d, is used instead of radius (with our K,
d 2 (t=
) d 02 4 Kt ).
There is, however, an internal unknown in (13), yi,0, which is controlled by the energy balance (11): the
heat for the liquid-to-vapour transition must come from the gas phase (droplets always attain a quasisteady temperature below that of the environment, even if injected hotter); evaporation always implies a
coupled heat and mass transfer.
Dividing now the energy balance (11) by m L = vA and upon integration yields:

0=
hLV0 +

ln

k dT
c pi (T T0 )
v dr

hLV0 + c pi (T T0 )
hLV0

=
tevap

hLV0

v r 2
v
dT
dr =
=

0 2 0 dr
k
kr
+ c pi (T T0 )

dr
c p liq 0 r0
c pi v0 r0
dt
= =

k
k

hLV + c pi (T T0 )
dr02
2k
=

ln 0
liq c pi
hLV0
dt

2
2
r0,ini
r0,ini
=
c p (T T0 ) 2 Di 1 yi ,0 (T0 )
2k
ln
ln 1 + i

1 yi ,
liq
hLV0
liq c pi

(14)

where the latter equality comes from (13), and is the closure for the evaporation problem, relating mass
diffusion and heat diffusion, and having only one unknown T0 after imposing Raoult's law (7):
yi ,0 (T0 ) = ( pi* (T0 ) p ) ( M i M m ) .
For instance, if a 0.1 mm in diameter water-droplet is injected in air at 1000 C, the droplet-temperature
quickly (in less than 1 ms) adapts to a temperature of 63 C, evaporating in some 0.17 s. The initial
relative motion decays also quickly (in some 10 ms). The vapour-mass-fraction close to the droplet is not
'near 1' as one might guess from such a 'burning' environment, but a mere 16% (notice that the air
surrounding the drop is just about 63 C and not 1000 C). As another example, the evaporation time for a
10 m n-decane droplet issuing at 400 K (50 C below its boiling point) into air preheated to 800 K in a
furnace, is some 1.210-3 s. Fuel droplets are usually assumed to evaporate in a zero-fuel environment, i.e.
into 'dry air' but, sometimes, the surrounding atmosphere may already containing some fuel vapours from
the evaporation of surrounding droplets in a spray.
As in the planar-geometry evaporation case treated above, convection effects in droplet evaporation can
be neglected in the species balance (10) and in the energy balance (11), but not in the overall mass
balance (9), if the gas is not too hot (let say for T<100 C), what is equivalent to linearising the
logarithms in (14), reducing to:
2
2
r0,ini
r0,ini
=
tevap
=
2k (T T0 ) 2 Di ( yi ,0 (T0 ) yi , )
liq hLV0
liq

(15)

which, together with yi ,0 (T0 ) = ( pi* (T0 ) p ) ( M i M m ) , allow the finding of T0, yi,0, and tevap.
It can be shown that, for room-temperature water droplets, T0 is the wet-bulb temperature, practically
coinciding with the adiabatic-saturation temperature analysed in Humid air and easily found with the
psychrometric chart, or analytically by an enthalpy matching. For instance, a water droplet 0.1 mm in
diameter in ambient air at 25 C, 100 kPa and 50%RH, attains a temperature of Twet bulb=17.6 C, and
evaporates in some 16 s.
Notice that the choice of system for the analysis can be changed, and, instead of the spherical volume of
radius r>r0, used above to establish the mass and energy balances (9-11), a unit area in a spherical surface
of radius r>r0 could have been used, and typical Fluid Mechanics formulation applied:
D ( i ei )
Dt

k d 2 dT
dT

+ ( i hi vi ) = q
0 + v c pv
vv = 2
r

r dr dr
dr

(16)

which, after performing a first integration from r0 to r, knowing from the mass balances (9) that

ivi=v=v0r02/r2, would yield a similar equation to (11):

r2
r

dT
dr

cp
v0 02=
v

k d 2 dT
r
r 2 dr dr

dT

v0 r02 c pv =
(T T0 ) kr 2 v0 r02 hlv0

dr

(17)

lv0 = 4 r02 k dT dr r = r . A second

where the last term comes from the energy balance at the surface, mh
0
integration would yield:

 hlv + c pv (T T0 ) v0 r0 c pv liq vliq r0 c pv

v0 r02 dr
dT
(18)
=

ln 0
=
=
2

hlv0 + c pv (T T0 )
k r
h
k
k
lv
0

and so on as from (14).

IGNITION AND EXTINCTION

We shall deal basically with steady-state kinetics and propagation, but sooner or later, as for most
engineering problems, besides analysing the steady running, one has to face the more complex
phenomena of starting up and stopping down the process (ignition and extinction), not only to better
understand combustion, but to provide a practical control of fire.
Ignition of a reactive system may be homogeneous (only for perfectly premixed systems uniformly
heated, as in compression ignition), or heterogeneous, as when locally initiated by a spark or a catalyst in
either a premixed or a non-premixed system (a subsequent deflagration front, i.e. a flame, develops).
All combustion reactions have positive affinities and thus could proceed naturally in an isolated system,
but most common fuels and oxidisers combine too slowly at ambient conditions, if an activation energy is
not externally supplied, i.e. an ignition source, like an electric spark or a hot object, to cause a selfacceleration of the reaction. After ignition, the reaction self-propagates because a flame is an ignition
source, as any other hot object.
One may consider different ways for the fuel/oxidiser reaction to run away:
Contact ignition, i.e. by just coming into contact some special fuels with special oxidisers
(ignition by contact with a hot object is covered under autoignition). There are substances (called
pyrophoric substances) which have autoignition temperatures below that of the environment (say
<15 C), so that the combustion takes place as soon as they are put in contact (they cannot exist
premixed). At room conditions, the mixture of hydrogen and fluorine, H2(g)/F2(g), burns without
an ignition source; the mixture H2(g)/Cl2(g) does not burn by contact in the dark, but intense light,
or the presence of Pt, Pd or Ni ignites it. Dimethyl-hydrazyne (N2H2(CH3)2(l)) and dinitrogen
tetroxide (N2O4(g)) also burn just by contact, and the mixture H2(g)/O2(g) also burns at room
conditions by contact in the presence of platinum catalyst. Even in normal air there may be
spontaneous combustion of some fuels, as when phosfine bubbles, obtained by immersion of
phosphorous
in
a
sodium
hydroxide
solution,
P4(s)+3NaOH(aq)+3H2O(l)=
PH3(g)+3NaH2PO2(aq), escapes at the free surface and react with air, PH3(g)+2O2(g)=
(1/4)P4O10(s)+(3/2)H2O(g). Also, when a solution of white phosphorus in carbon disulfide is used
to write on paper, the carbon disulfide evaporates within a few minutes and the paper chars. Pure
iron in minute amounts also burn by contact with air (this is how sparks are produced when hitting
a piece of iron on a flint). Although not properly a combustion process, the rapid exothermic
oxidation of metallic sodium in contact with water, Na(s)+H2O(l)=Na(OH(aq)+(1/2)H2(g), may
release so much energy as to ignite the hydrogen produced. There have been accidents where oilimpregnated rags have got fire by themselves because of the rapid oxidation of oil in air with such
a large contact area and low heat transfer.
Spark ignition, i.e. by locally supplying the activation energy required to triggers the reaction,
based on the production of a tiny burning particle (e.g. by striking a piece of flint and steel
together) as in prehistoric times, or on the electric arc produced between two closed electrodes
when the dielectric strength of the medium is surpassed (3 kV/mm for air). The latter is a very
convenient ignition method, used in Otto engines, gas turbines, furnaces and boilers, but requires a
premixture at least locally near the spark plug (see Flash point, below). An order-of-magnitude
estimation of the energy required for ignition is the energy required to heat a cubic volume of size

Lth to the autoignition temperature, i.e., Eign~L3thcpTign~(103)31103103~103 J, as mentioned

above. It is estimated that the spark produces a plasma filament at some 20 000 K in the discharge,
quickly turning into a hot gas-ball of one or two millimetres in diameter at some 5000 K, from
which the normal deflagration theory applies. The spark plug was invented by N. Tesla in 1890
(also attributed to R. Bosh and to K. Benz).
Autoignition, usually by compression ignition, but also by hot-surface ignition, i.e. by raising the
temperature of the gases by rapid compression or by heat transfer, either previously mixed, or,
more commonly, by raising the temperature of the oxidiser in a sudden compression and then
injecting the fuel when desired and at the desired rate. This is a very convenient ignition method,
which may be applied to the non-premixed mixture as in the usual Diesel engine (i.e. to ignite a
cold liquid jet being injected in hot surrounding air), or to a fully premixed mixture (as in some
new stationary Diesel engines where all the fuel is injected before ignition takes place by
compression), or to a dual-fuel Diesel engines (as when a lean natural gas / air mixture is
premixed and made to ignite by compression-ignition of a low-autoignition secondary fuel pilot
injection). Autoignition by hot spot is usually an undesirable event, only used purposely in minute
model engines.
Catalytic ignition. By using a catalyst to lower the activation energy. For instance, a H2(g)/O2(g)
mixture within a glass or a steel bottle at room temperature and pressure is not seen to react
(hardly any molecule yield fertile chocks, releasing negligible heat); the effect of a sizeable spark
(>1 mJ) in a sizeable container (larger than a few millimetres by all sides), a hard knock, or the
presence of a sizeable amount of platinum, is required to ignite the system at ambient conditions
(it may also be ignited by heating the mixture to 850 K at 100 kPa whatever the container).

Combustion ignition in reciprocating engines must be stimulated at each cycle (either by a spark or by
quick compression), but in steady combustors like gas turbines, furnaces and boilers, the system only
needs to be ignited at start of the duty (usually by spark plugs); although sometimes a pilot flame is
maintained if operation is discontinuous (e.g. in some domestic hot-water heaters).
Once ignited, for a flame to propagate, a certain balance between the heat released (depending on
chemical kinetics) and heat transfer towards the fresh mixture, must be warranted, what explains the
existence of flammability limits and quenching distance (explained below), as well as the process of
ignition and extinction themselves.
Combustion may be difficult to start, but it may be more difficult to stop. Besides the obvious extinction
procedure of letting all the fuel to burn (the amount of oxidiser is assumed infinite in terrestrial
applications), and avoiding any more fuel to be near the flame, a combustion process may be arrested by
providing unfavourable conditions to its propagation, namely, by cooling and/or by dilution of its reactive
species (fuel, oxidiser, and the many radicals needed for flame propagation).
The traditional fire-fighting method of massively adding water, works mainly as a cooling agent to lower
the temperature below the autoignition value, to prevent flame propagation, whereas the typical CO2 fireextinguisher works mainly as a dilutant to prevent air coming into the flame (see Fire safety).
Cold solid walls, with their massive thermal sinks, also act as cooling agents that prevent flame
propagation near them; safety lamps and quenching grids are based on that fact. For instance, premixed
methane/air combustion cannot propagate inside a metal tube of less than 2 mm bore. This value
corresponds to the quenching distance for stoichiometric mixtures; for lean and rich mixtures the
quenching distance grows parabolically towards infinite at the ignition limits (e.g. for methane(air, it goes
from 1.8 mm at =1, to 4 mm at =0.8 and =1.8).

Many types of igniters have been developed. A summary of ignition devices, classified by the ignition
method, follows.
Igniter types
Thermal
Rapid-compression heating
Air-compression igniter, as in a diesel engine; some primitive people knocked in a crude
piston-cylinder device to ignite some dry wood-shavings inside.
Spark heating
The flint-and steel spark. Usually, a gripping steel wheel is force to rotate a little against a
small flint stone, and iron sparks are projected over some easily flammable substance:
LPG (since 1933), gasoline vapour (since 1932) or waxy wool (tinder, since prehistoric
times). Little sensitive to moisture.
Two-electrode sparks. Electric discharge through a 1 mm air-gap (requires a coil
autotransformer to overcome rupture voltage in gases, i.e. >3 kV/mm), on use since 1850.
Sensitive to moisture.
Piezoelectric sparks. Created when some natural crystalline substances (quartz, barium
titanate) are hammered (a mechanical trigger is used), causing opposite electrical charges
to concentrate on each side of the crystal. Non-sensitive to moisture.
Electronic sparks. Since 1980. All lighters until 1990 where of non-premixed type, but
since then, induction-type premixed flames, nearly invisible and much more powerful, are
produced in 'jet lighters' (miniature bunsens).
Dissipative heating
Mechanical energy dissipation. The bow and drill is the most-efficient simple-friction firemaking procedure since prehistoric times. Very sensitive to moisture.
Electrical energy dissipation. A battery and a wire or a steel-wool load that gets red-hot by
Joule heating. Sometimes the wire is a catalyst (see below).
Pyrotechnic heating
The match. It is the most handy igniter since XIX c. With a little friction, a specially
prepared chemical substance quickly reacts and forms a flame. Sensitive to moisture.
Radiation heating
Solar radiation and a glass magnifier (water-filled bottles, or even a clear ice ball)
focussing light onto tinder. Non-sensitive to moisture.
Pulse laser can be focused on the most convenient spot; the combination of pulse-mode
semi-conductor lasers and optical fibres seem most promising. Non-sensitive to moisture.
Pure chemical (without heating)
Simple contact of special chemicals (e.g. hydrogen with fluoride).
Catalytic igniter. Platinum wire gets red-hot, reaching >1000 K, when in contact with an
hydrogen/air mixture. sustaining combustion even of other combustible mixtures. For propane
and butane flames, the catalytic wire is usually preheated electrically. This device is used in
modern, wind-proof lighters, producing a steady blow-torch type of flame nearly-invisible in
the bright (sometimes a ceramic flame-window that glows when hot is used to indicate the
lighter is on).
In spite of all the above possibilities, to be successful in using an igniter to start a camp fire, as important
as the lighter, is to get ready dry flammable tinder (fine wood shavings, dried grass, straw, down from
corn or from birds or seeds, etc.).

Ignition is a requisite for combustion, but fire-keeping demands also self-sustained conditions for its
propagation (very easily obtained with flammable material, but sometimes not so easy with other
combustible material).

PROPAGATION OF NON-PREMIXED FLAMES

Non-premixed flames (also diffusion flames, but all flame requires diffusion) are established by some
ignition process (usually by a spark or by high temperature of the oxidiser), near the boundary between
fuel and oxidiser, not having a defined propagation speed and flammability limits as premixed flames.
The main characteristic of a non-premixed flame is its size and shape.
Non-premixed flames may be laminar (e.g. candle flame, droplet burning) or turbulent (e.g. open fires),
and they are quite luminous (yellow, because of blackbody emission by soot) except for hydrogen flames,
that are nearly invisible (alcohol flames are difficult to see, too).
Burke and Schumann established in 1928 the first model of non-premixed flames (coaxial flows with the
same speed), based on the assumption that the characteristic chemical time for reaction is much shorter
than the time required for the reactives to reach the flame (mixing time), what implies very fast kinetics
and very thin flame thickness. The fluid field is split by the flame in a fuel-region with no oxidiser (yO=0),
and an oxidiser-region with no fuel (yF=0; i.e. yFyO=0 in the whole field), the flame sitting just where the
two reactive flows meet with stoichiometric rates (i.e. for a reaction F+OOPP, where

O vO M O + O FvF M F =
0 ); besides, if chemical kinetics is so fast that the global process is diffusioncontrolled, both yF=0 and yO=0 at the flame front. This Burke and Schumann's model of fast chemical
kinetics is the first step in most analysis of non-premixed combustion.
Non-premixed flames are most used in practical system (coal burning, gas-fired or oil-fired industrial
burners, diesel engines, gas turbines, and solid-fuel rockets), because of their higher safety and ease of
control, in spite of being less power-intensive and more pollutant than premixed combustion. They are
easier to establish than premixed flames, as everybody lighting a bunsen realises, because they naturally
hold attached to the injector rim over a very wide regime range, and they are more radiant, what is an
advantage for old lighting-methods and long-range heating. Unsteady non-premixed flames occur in
diesel engines.
To ignite a non-premixed fuel/oxidiser mixture, some ignition agent must be applied at the mixing layer
(outside this fluid interface, the fluid is not flammable because only fuel or oxidiser exist). The usual way
to ignite this local premixture is by an electrical spark or a pilot flame, but auto-ignition at high
temperature is used in diesel engines. Detailed studies in simple configurations, like the autoignition of a
small fuel droplet in hot air (well above the autoignition temperature), shows that ignition takes place
(after some delay time) in the lean and hotter zone sited some 10 droplet-radius afar (with relative air/fuel
ratios =2.5..3). The square-law for droplet-diameter evolution, d2(t)=d02Kt, applies very well to both
the non-reacting evaporation and to the combustion processes, with typical values of K heavily dependent
on temperature (K=10-7..10-6 m2/s in droplet combustion in cold or hot air).

GAS FUEL JET

Flame length
Flame length is the most characteristic parameter of a non-premixed flame established at the mouth of an
injector; the most common example may be the flame of a cigarette lighter. Flame length depends on the
measuring principle: visible light, other radiations, temperature field, concentration fields, and the
averaging principle for fluctuating flames. It also depends on buoyancy; e.g. propane/air jet flames
issuing from a 0.8 mm tube, linearly increase their height with exit velocity up to 12 m/s (Re1000), with

heights of 0.25 m for buoyant flames (on the ground) and 0.33 m for non-buoyant flames (under
microgravity). Further increasing of exit velocity causes a transition to turbulent flames with heights
receding to 0.18 m for buoyant flames, until they blow out at 33 m/s, whereas for non-buoyant flames
heights further increase to 0.39 m, until some 35 m/s, when they shorten to some 0.29 m before blow out
at 40 m/s. Notice that, at high speeds, buoyancy should not be as relevant as it appears in practice
(perhaps the behaviour would tend to match, if blow-out did not prevent it).
A quick guess for the laminar flame length, Lfl, may be given by the distance travelled by the jet at the
exit speed w0 during the period of fuel diffusion tdif, from yF=1 at the jet centre, to the stoichiometric value
yF=yF,stq at the flame front; i.e. Lfl=tdifw0, with tdif=(1yF,stq)(D0/2)2/Di and thence:
Lfl=(1yF,stq)w0D02/(4Di).

(19)

For instance, for the propane/air jet-flame above-mentioned, at w0=12 m/s, with D0=0.8 mm, Di=105
m2/s and yF,stq=1/(1+16)=0.06, we get Lfl=0.18 m, against the 0.25 m measured.
Notice that with this model (only applicable to laminar jet flames, Re<1000) the flame length is
proportional to the volumetric flow-rate, and increases with stoichiometric air/fuel ratio (e.g. carbonmonoxide flame are much shorter than hydrocarbon flames, LCO/LCH40.2, LH2/LCH40.2, LC4H10/LCH43).
Turbulent non-premixed flames flicker with a wide frequency spectra, but the main first natural
frequency, fn, decreases with size in the form f n f 0 / D D0 , with f0=1.5 Hz and D0=1 m, D being the
diameter of the fuel-injection tube or the mean horizontal size for open fires (it seems that this is due to
the Kelvin-Helmholtz vortex-ring instability in the shear layer of the hot plume, outside the flame).
Non-premixed flames may radiate a lot of energy. For gas-jet flames of a few kilowatts issuing from a
few-millimetre injector, from 10% (for methane/air) to 50% (for propane/air) of their heat release is lost
by radiation, mainly by blackbody emission at some 2500 K, from the in-flame soot (with a smooth
maximum at CWien/T=2900/2500=1.2 m, but with a more pronounced peak between 4 m and 5 m due
to molecular band emission by H2O and CO2.
Similarly to the temperature field, soot production in a gas-fuel jet flame, concentrates in the mixing shell
where the flame locates at small heights, and wide-spreads to the centre downstream (e.g. for a steady
methane jet of 10 mm diameter issuing in air at 80 mm/s, soot volume fractions profiles have a two-bell
shape up to some 50 mm height, and one central bell-shape further along the axis).
In most practical non-premixed burners, air is coaxially fed with the fuel, what may increase the flame
length and pollutant emissions if the co-flow pattern reduces the mixing. To prevent this, special swirlinducing nozzles are used to quickly mix fuel and air, greatly reducing flame length.

CONDENSED FUEL
Flash point (ignition)
Another aspect related to the combustion of non-premixed mixtures is the so called flash point of a
condensed fuel in the presence of ambient air. In the limit case of the initial liquid fuel before evaporation
(i.e. no fuel-vapours in the air), combustion is not possible, but, as soon as evaporation proceeds, a
gaseous fuel/air mixture develops near the liquid surface, with a fuel-vapour mole fraction that may be
below the lower flammability limit (see Flammability limits, below), in between the lower and upper
flammability limits, or above the upper limit. Of course, all this depends on temperature, since vapour
pressure exponentially grows with temperature.

The flash point is the lowest temperature at which a liquid has a sufficient vapour pressure to ignite when
a small pilot flame is brought close to the surface of the liquid. Table 1 shows some flash point data
(more can be found aside).
Table 1. Flash-point temperature for liquid fuels in air at 100 kPa.
Fuel
Tflash
Methane
188 C 85 K
n-Butane
63 C 210 K
Iso-Octane
12 C 261 K
n-Heptane
4 C 269 K
Methanol
12 C 285 K
Ethanol
12 C 285 K
n-Octane
13 C 286 K
n-Decane
44 C 317 K
Diesel
50 C 323 K
Kerosene
57 C 330 K
n-Dodecane
71 C 344 K
Lubrication oil 200 C 500 K
Liquids (and solids) with flash points below 40 C (or 50 C, the flash point for diesel; the choice is
arbitrary and depends on legislation) are labelled flammable, and above that are labelled combustible.
Many common organic-liquid solvents have a flash point below room temperature, posing a great fire
hazard: e.g. benzene (11 C), acetone (18 C), diethyl ether (45 C). Some of them will maintain their
flammability risk even at concentrations as low as 10% by weight in water; e.g. isopropanol and ethanol
have flash points below the threshold of 38 C at concentrations as low as 30% by weight in water. A
pool of light fuel oil with a flash point above 60 C cannot be ignited with a match.
Crude-oil is easy to ignite because of its many volatiles, requiring extensive safety measures for its
handling and transport, but crude-oil spills at sea cannot be ignited because volatiles swiftly go away and
heavies cannot be heated enough because of the large thermal sink of underlying water (oil spills have to
be cleaned mechanically and with chemical solvents).
The state of the fuel sample (particularly for solid fuels), its humidity, the composition of the oxidising
atmosphere, or finely dividing the fuel (sprays and powders), can significantly change the flash point.
Sprays of pure fuel show lower flash points, but sprays of diluted fuels (e.g. alcohol/water mixtures) show
higher flash points. Wood shaving ignites more readily than a massive piece of wood.
Flame stabilisation by porous feeding: oil lamps and the candle flame
Condensed fuels can be ignited by a spark if they are above their flash-point temperature (at least locally);
if not, they have to be heated before ignition starts, perhaps just holding close a match for a while,
providing the heating and the igniter at the same time. Once ignited, the non-premixed flame will travel
along the fuel surface, usually becoming unstable, and even extinguishing by poor ventilation (lack of
air). All these problems are solved when the flame can get stabilised around the tip of a porous solid that
pumps liquid-fuel by capillary suction towards a well-ventilated region, what is the foundation of oil
lamps and candles; a torch being a hybrid device between a candle and an oil lamp (a torch is a wooden or
tow shaft dipped in wax, tallow or pitch, and set alight by an igniter). Some oil lamps developed in the
19th century used gravity-pumping instead of capillary pumping. A detailed description of the historical
developments of oil lamps (where the fuel is already in the liquid state), and of candles (where the fuel is
in the solid state, but a local molten pool is maintained by heat radiation from the flame), can be found
aside, in History of Fuels. Torches, oil lamps and candles).

The length of this type of flames depends to some extent on the length of the wick, what can be controlled
in oil lamps (too long wicks will burn themselves if unable to pump enough liquid). The structure of a
typical candle flame is depicted in Fig. 3. Initially, the wick holds some solid wax in its pores (from
previous use or from manufacturing). The wax can be of animal origin (esters of fatty acids, like stearin
and beeswax) or of mineral origin (long-chain paraffins, beyond eicosane, C20H42, Tm=36.4 C). When a
match is brought nearby, first the wax melts (between 50 C and 60 C), then the vapours heat up to the
flash point (around 230 C, i.e. 600 K), ignition starts, and a quasi-steady-state regime is reached; a 20
mm in diameter candle burns some 5..10 g/h of fuel, yielding some 60..120 W (LHVwax=44 MJ/kg), with
a flame height of 40..45 mm in still air, and a maximum flame diameter of some 8..10 mm.

Fig. 3. Mixing processes and structure of a candle flame.

These flames radiate a lot of energy, and, although most of this low-temperature blackbody-radiation is in
the infrared spectrum, the small visible part more than justified their prominent use for artificial lighting
since prehistoric times, until the development of electric light at the end of the 19th century.
A typical candle flame has a colour temperature of some 1700 K, what is responsible for the typical
yellow colour of non-premixed flames. A volume fraction of soot less than 1 ppm already gives luminous
blackbody emission. This is similar to the incandescent radiation from a filament electrically heated to
near 3000 K as in the traditional incandescent bulb-lamp (white light corresponds to the solar spectrum,
nearly a perfect blackbody at 5800 K). Soot formation is a relatively slow chemical process of nucleation
and aggregation, and it is enhanced by increasing the residence time within the flame and by the presence
of polycyclic aromatic hydrocarbons (PAH), so that, when flickering occurs in turbulent flames, soot
production quickly rise.
Fuel sprays
Because of the advantage of easy storage and high energy intensity, liquid fuels are used in the majority
of vehicle-engines (cars, trucks, trains, ships, and aircraft), and in a sizeable part of stationary engines and
power plants. The combustion is performed by injection of the liquid fuel in an oxidiser gas stream
(usually air), and it is of utmost importance for a good burning to promote the interface area between
liquid and gas, where, in most cases, a non-premixed flame will be established; the exception is when the
injected liquid can be left to fully evaporate and a subsequent premixed flame produced, as in gasoline
engines.

The scattering of the injected liquid into fine droplets (spray, or atomisation; see Fig. 4), is usually
achieved mechanically by creating a high-speed jet (>100 m/s, what requires very high pressure jumps,
currently up to 200 MPa in some diesel injectors), which disintegrate into droplets (much smaller than the
jet diameter), by shear forces. Think about how fast the processes follow each other: in a few
milliseconds, a thin liquid stream in turbulent motion (which may cavitate and flash-boil when
approaching the tip of the fuel injector), forms a spray that breaks down to a myriad of droplets, which
evaporate, ignite, burn, and generate unwanted emissions. Wall impingement is undesirable in most
circumstances, increasing emissions by poor combustion near cold walls; even in gasoline injection at the
manifoild, wetted surfaces introduce an undesirable time-lag in the fuel control loop.

Fig. 4. Jet break-up, droplet formation and evaporation in a diesel injector.

The break-up length, i.e. the length where the first droplets appear (at the jet surface), tends to zero in
atomised jets (Fig. 5). Other methods of spraying, as impinging jets, air-entrained or water-entrained jets,
are less used in combustion.

Fig. 5. Liquid-jet break-up regimes in terms of Reynolds number ReuD/, and Ohnesorge number
Oh/(D)1/2. For a given injector diameter D, and given liquid (of density , viscosity , and
surface tension ), the only variable is injection speed u. In the Rayleigh regime, drops larger than
the jet diameter are cast by capillary instability, and the break-up length grows proportionally to
the speed (A to D). In the first wind-induced regime, droplet size decreases to the range of jet
diameter, and the break-up length decreases too /E to F), but still being much larger than the

diameter. In the second wind-induced regime, droplet size further decreases, and break-up length
tends to zero (F to G, at the surface; there is a longer liquid spike at the core that grows with
speed, F to H). The atomization regime stands beyond this zero-break-up-length point (beyond G).
Steady non-premixed flames occur in liquid-fired gas turbines, boilers and furnaces, whereas unsteady
non-premixed flames occur in diesel engines. Typically, a jet of liquid fuel is injected (at some 350 K
once warmed by the hot engine), inside hot air (heated to some 950 K by adiabatic compression). Due to
the fine atomisation achieved by the high pressure injector, a cloud of fuel vapour, well-mixed with air,
develops around the liquid jet already after a few millimetres from the injector; at a distance of a few
centimetres, this rich vapour mixture already attains its autoignition temperature (the minimum for
diesel/air is 480 K, but up to 800 K are measured before proper ignition), pyrolyses and burns, attaining
more than 2500 K and showing chemiluminiscent radiation; beyond that, soot starts to form at the tip of
the jet due to thermal pyrolysis of air-depleted vapours, and soot formation continues until there is liquid
remaining.
In the dual-fuel diesel engine, a lean mixture of air / natural gas is usually formed in the admission
manifold, and, once compressed to the normal compression ratio of the diesel engine, is ignited by a
injection of a small amount of diesel oil at the usual high pressure near the top dead centre; the
temperature increase during compression of the mixture is high enough to auto-ignite the oil (480 K for
diesel/air) but not enough to auto-ignite the gaseous mixture (850 K for methane/air), helped also by the
larger autoignition delay of methane than that of diesel oil.
Notice that LPG injection, and heated gasoline and diesel injection, can give rise to flash-boiling (i.e. if
the injection enthalpy is higher than the saturated-liquid enthalpy at the discharge pressure), with a very
effective atomisation. Presently, it is only of interest in gasoline injectors in the intake manifold.
Droplet combustion
A single droplet of fuel burning in air, is not of any practical use (except for research), but it is of
paramount importance for the understanding of spray combustion, a very practical arrangement used in all
liquid-fuel burners, all diesel engines, and kerosene gas-turbines. The key parameter to be found is the
droplet burning time, what governs the path length before burnout (and consequently the minimum size of
the combustion chamber required), and the required residence time for unsteady processes (and
consequently the minimum period in reciprocating engines).
We here assume that the combustion has already been started, either by a spark in cold air (above the
flash-point temperature), or by autoignition in hot air (above autoignition temperature). Pressure, assumed
constant during droplet combustion, has an influence on several aspects of the droplet-burning process,
including flame dynamics, combustion chemistry, evaporation rate, heat-up period, and ignition delay
time. For instance, droplet combustion experiments show that, after ignition, the flame increases in size
initially by thermal expansion and internal convection, but later decreases in size, following the shrinking
of the droplet (but not linearly proportional as deduced below with a simple model).
To further simplify the single-droplet problem, we assume it to be spherically-symmetric (a good
approximation under microgravity, because, on Earth, a vertical plume develops; by the way,
microgravity also allows easy droplet positioning by levitation). Additionally, we focus just on the quasisteady rate of burning, after the initial processes of ignition (heating, species diffusion, and reaction
activation) had been settled. As for droplet evaporation, we can imagine a liquid source at the centre of
the drop to keep a strict steady state. In this steady-rate regime, sketched in Fig. 6, the droplet temperature
keeps a constant value T0 (a heat-and-mass diffusion trade-off value to be found, lower than its boiling
point), with heat from the flame diffusing inwards and forcing evaporation, which supplies fuel vapours

with a mass fraction yF0 close to the liquid, diffusing outwards and feeding a non-premixed flame, located
at r1/r010, fed also from the outside by a diffusive flow of oxygen from the air, with both fuel and
oxidiser concentrations approaching zero at the flame (with slopes proportional to the stoichiometry).
Products being generated at the flame (we assume the single-step reaction model F+OOPP), and part
of the heat released, diffuse outwards. A possible inert gas, notably nitrogen foe combustion in air,
completes the set of species playing (one must have at every point yF+yO+yP+yN=1).

Fig. 6. Droplet combustion. a) Temperature and species-concentration profiles (combustion in air). b)

Evolution of drop radius, area, and volume, with time.
We start the analysis of droplet combustion by rewriting the equations for droplet evaporation (9-11), i.e.,
the global mass balance, species mass balance, and energy balance, for a spherical volume of radius r
(now it must be r0<r<r1), with yF+yP+yN=1 and only the fuel flowing (from the source at the origin, to the
sink at the flame; there is no sink for yP or yN at r0):
Magnitude

Accumulation

Production Diffusive flux

Mass balance (global)

m F

+0

Mass balance of fuel

(volatile species)

m F

+ ADF

Energy balance

m F hLV0

+ Ak

Convective flux

dy F
dr

dT
dr

m ( r )

(20)

m ( r ) yF

(21)

m ( r ) c pF (T T0 ) (22)

These equations must be now supplemented with the mass diffusion balances for the other species (all
minus one, if we make use of the global balance). To simplify nomenclature, we make use of the
stoichiometric relation, F+OOPP, from the start. The choice of system for analysis is always a control
volume of radius r in strict steady state (with appropriate sources and sinks at discontinuities). For the
oxidiser, we must choose a sphere with r>r1 (there is no oxidiser in r<r1). The oxidiser has a source at
r and a sink at the flame (at r=r1), of algebraic amount m O related to the fuel flow-rate by the
stoichiometry at the flame: m O M O = O m F M F ; notice that the stoichiometric coefficients i are in molar
basis (that is why the molar masses appear in the mass-equations).
Magnitude

Accumulation

Mass balance of oxidiser (r>r1)

Mass balance of products (r>r1)

Mass balance of products (r<r1)

Production Diffusive flux Convective flux

M
dy
= m F O O + ADO O
m ( r ) yO
(23)
MF
dr
M
dy
= m F P P
+ ADP P
m ( r ) yP
(24)
MF
dr
=

+ ADP

dyP
dr

m ( r ) yP

(25)

Mass balance of nitrogen (any r)

+ ADO

dyO
dr

m ( r ) yO

(26)

the latter being redundant since at any stage we have yF+yO+yP+yN=1.

The overall mass balance (20) is still valid for any r, on either side of the flame, since there are sources
and sinks of species at the flame front, but total mass is conserved. Thus, the global mass-flow-rate of
gases across any spherical surface, m ( r ) , is independent of r, and the bulk radial velocity v, assuming
constant gas density (no as good an approximation as for the droplet evaporation case studied before,
since, although pressure is also assumed constant, temperature varies a lot in droplet combustion), yields:

v0 r02
=
m =
vA constant
= m F
v=
(r ) 2
r

(27)

The convective speed in the gas phase close to the drop, v0, relates to the radius recession rate in the real
process through the mass conservation at the liquid surface:

liq dr0
dr
vA= v0 A0= liq 0 A0
v0=
dt
dt

(28)

The species balance of fuel in r0<r<r1, equation (21), divided by m F = vA and upon integration, yields:

DF dyF
r 2 DF dyF
1+
yF 0 =
yF
v dr
v0 r02 dr
DF
1 1

ln (1 yF,0 ) =
2
v0 r0
r0 r1

0=
1+

DF dyF
1
d
=

2
v0 r0 1 yF
r
(29)

The species balance of oxidiser in r1<r<, equation (23), divided by m F = vA and upon integration,
yields:
M
D dy
0=
O O + O O yO
MF
v dr

DO
ln
v0 r02

M O r 2 DO dyO
0=
O
+
yO
M F v0 r02 dr

MO
+ yO,
MF
1
=
M
r1
O O
MF

DO
dyO
1
=
d
2
M
v0 r0
r
O
+ yO
O
MF

(30)

The energy balance in r0<r<r1, equation (22), divided by m F = vA and upon integration yields:

hLV0
0=

ln

k dT
+
c p (T T0 )
v dr

hLV0 + c p (T1 T0 )
hLV0

hLV0

v0 r02
v0 r02 1
v
dT
=

dr =
dr = d
+ c p (T T0 )
k
kr 2
k
r

dr
c p liq 0 r0
c p v0 r02 1 1 c p v0 r0
dt
=
=

k
r
r
k
k
0 1

c (T T )
dr02
2k
ln 1 + p 1 0 K
=

hLV0
dt
liq c p

tburn =

2
r0,ini

(31)

i.e. getting the well-known r2-law (similar to Langmuirs law (1918) for droplet evaporation), stating that
the droplet life-time, tburn, is proportional to the square of the initial radius; often, the diameter, d, is used
instead of radius (with our K, d 2 (t=
) d 02 4 Kt ). The first derivation of this burning time was done by
Spalding and Godsave in 1953. Notice that this constant K for the burning problem will differ in value
from that of the simple evaporation problem.
If the steady-state temperature of the drop, T0, and its corresponding vaporisation enthalpy, hLV0, are
approximated by the boiling-point values (Tb and hLVb, or any other suitable value, because the influence
on (31) is not large), the energy balance gives directly the burning rate; afterwards, the species balances
can be used to get the actual droplet temperature and other internal unknowns, as fuel-mass-fraction near
the liquid surface, and products-mass-fractions.
The simplification introduced along (31), 1/r1<<1/r0, is justified by practical results (r1/r010). Later, the
flame position r1 will be found from (27), but the droplet life-time is directly given by the integrated
energy equation (28) because we know the flame temperature T1, which is the adiabatic flame
temperature, Tad, (we neglect radiation losses) given by:
=
T1 T=
T +
ad

LHVmolar
prod

(32)

i pi

where LHVmolar stands for the lower heating value of the fuel in molar basis, T=298 K is the standard
reference in Thermochemistry, ni are the molar coefficients of the products (including any possible inert
species), and cp,i their molar thermal capacities. One can make the approximation of considering just one
thermal-capacity value for the products mixture, cp, usually that of air at some appropriate conditions
intermediate between the extreme temperatures; the values cp=34 J/(molK)=1200 J/(kgK) may be used
for combustion in air (care is needed, however, to account for the mol-ratio or the mass-ratio of products,
including inerts, relative to the fuel. As a rule of thumb, an order-of-magnitude value for the adiabatic
temperature of Tad=2500 K may be used for a first approximation.
In summary, for a given fuel droplet in a given environment, data are: initial droplet radius (e.g. r0,ini=10-5
m), ambient air temperature, pressure and oxygen concentration (e.g. T=300 K, p=100 kPa, yO=0.23),
thermal properties of ambient air (for the given state, cp=1000 J/(kgK), =0.024 W/(mK), and
MO=0.029 kg/mol), properties of fuel (e.g. for n-dodecane, a pure-substance model for diesel oil (a
commercial distillate), liq=780 kg/m3, Tb=489 K, hLVb=257 kJ/kg, MF=0.170 kg/mol,
O=37/2=18.5, P=12+13=25, LHVmass=45 MJ/kg, and Tad=2600 K), the latter obtained from the
thermochemical analysis:
Stoichiometry: C12H26 + (37/2)O2 = 12CO2 +13H2O + 7575 kJ/mol
the latter obtained from LHVmolar=ihf,i=[12(393.52)13(241.82)+(291)]=7575 kJ/mol and
thence LHVmass=7.575/0.170=44.6 MJ/kg. The actual mixture used in the case of burning in air is:
Mixture used: C12H26 + 18.5O2 + 69.6N2= 12CO2 +13H2O + 69.6N2 + 7575 kJ/mol
and consequently:

LHV
7575 103
T1 =
Tad =
T + prod molar =
298 +
=
2650 K
(12 + 13 + 69.6) 34
icp i

or, in mass terms:

T1 =
Tad =
T +

LHVmass
44.6 103
298 +
2630 K
=
=
0.170 + 18.5 0.032 + 69.6 0.028
m prod
1200
cp
0.170
m fuel

From the energy balance (31) we get the burning time:

r2
tburn =0,ini , with
K

c (T T )
c (T T )
2k
2k
ln 1 + p 1 0
ln 1 + p 1 b =
liq c p
liq c p
hLV0
hLVb

K=

2
2 0.024 1200 ( 2650 489 )
6 m
ln 1 +
=
=0.12 10
780 1200
257 103
s

2
r0,ini
(10 ) =0.83 103 s
= =
K
0.12 106
5 2

tburn

i.e., a r0,ini=10 m n-dodecane droplet burns in 0.8 ms (typical uncertainties are 20%). Associated to the
burning rate in terms of radius (31) is the burning rate in terms of fuel mass:
dr0
2 r0 K =
2 K r0,2ini Kt m F ,ini =
2 r0,ini K =
2 0.003 0.12 106 =
2.3 109 kg/s
=
dt
The flame radius, r1, (relative to the drop radius, r0) can be obtained from the oxidiser mass balance (23):
m F =
4 r02

c p (T1 T0 )
2
k
r
d
1
ln
1 +

D
liq dr0
cp O
v0 =
h
r1 liq
LV
0

dt
2 dt

=
=

r0

y
y
DO ln 1 + O,
ln 1 + O,
MO
MO
O
O

MF
M F

the latter obtained by substituting the burning-rate result (31) and introducing the approximation of equal
thermal and solutal diffusivities in the gas phase (Schmidt number unity). With our example data:
M
O O + yO,
DO
MF
1
=
ln
2
M
v0 r0
r1
O O
MF

c (T T )
1200 ( 2650 489 )
ln 1 + p 1 0 ln 1 +

hLV0
257 103
ln (1 + 10 )
r1

=
37
=
=
=
r0
ln (1 + 0.066 )

0.23
y
ln 1 +
ln 1 + O,
15 0.032
MO

2 0.170

M
F

i.e. the flame stabilises itself at a distance of some 37 radii from the drop (of course, with an uncertainty
proportional to the simplifications assumed).
The mass fraction of fuel vapours close to the drop, can be computed from the fuel mass balance (21).
Proceeding as before, we get:

DF
1
ln (1 yF,0 ) =
2
v0 r0
r0

(1 y ) 1 +
F,0

1 dr02

ln (1 yF,0 ) = liq 2 dt
DF
liq dr0
v0 =
dt

c p (T1 T0 )
1
=

hLV0

yF,0 =
1+

1
hLV0

c p (T1 T0 )

=
1+

cp

DO

c (T1 T0 )
ln 1 + p

h
LV
0

1
1
=
=
0.91
3
1
257 10
1+
10
1200 ( 2650 489 )

i.e., there is some 91% in mass of fuel vapours close to the drop (the uncertainty may be large).
The mass fraction of products, yP, in the outer region can be computed from the products mass balance for
r>r1 (24), similarly to that of oxidiser, but now with yP,0, what yields:
MP
M
D dy
D
MF
1
=
0 P P + P P yP
P2 ln
=
M F v dr
v0 r0 M P y
r1
P
P,1
MF

1200 ( 2650 489 )

ln 1 +

257 103

=37
0.030

25 0.170
ln

0.030
25
yP,1
0.170

c (T T )
ln 1 + p 1 0

hLV0
r1

=
M
r0

P
P M

ln
M P y
P,1
PM

yP,1 =0.35

i.e., at the flame, yP,1=0.35, and, since there is neither fuel nor oxidiser at the flame radius (see BurkeSchuman's model, above), the rest must be inert component, yN,1=0.65. In the inner region the equation is
(25), what yields:

0 0+
=

y
DP dyO
D
1
yP
P2 ln P,1
=
v dr
v0 r0
yP,0 r1

1200 ( 2650 489 )

ln 1 +

257 103

=37
0.35
ln

yP,0

c (T T )
ln 1 + p 1 0

hLV0
r1

=
r0
y
ln P,1
yP,0

yP,0 =0.03

i.e, close to the drop, r=r0, the composition is {yF,0,yO,0,yP,0,yN,0}={0.91,0,0.03,0.06}, at the flame, r=r1,
the composition is {yF,0,yO,0,yP,0,yN,0}={0,0,0.35,0.65}, and at r {yF,0,yO,0,yP,0,yN,0}={0,0.23,0,0.77}.
The last development could have been applied to the inert component, what yields the relation
yN,1/yN,0=yP,1/yP,0. As for the composition profiles along r, equations (29-31) should not be integrated
between the limits, but up to a generic r, what yields exponential profiles for all the variations, as
sketched in Fig. 6.

Particle combustion
We consider here the combustion of solid particles, liquid particles being considered above under Droplet
combustion. The main differences to the latter are:
Particles shape can depart a lot from the spherical shape, because there are no surface tension
effects as in droplets. Fine solid particles, however, are usually the result of a milling process that
tends to produce not too-elongated particles.
In the case of metallic particles, the combustion products are usually solid oxides (even at the
high-temperatures involved).
Inert solid material (ash) may be a sizeable part of solid fuel (e.g. in some coals), leaving a porous
solid layer around the burning fuel core, significantly distorting mass and heat transfer.
Contrary to liquids, the combustion of solids can take place at the interface (heterogeneous
combustion). This is because liquids cannot get hot enough to sustain a rapid surface-oxidation
process with the adsorbed oxidiser; they vaporise before. But solids can indeed get very hot and
yield a sizeable rate of chemical recombination at their surface layer. For a reaction scheme

F+OOPP, fuel and oxidiser flow-rates still must verify O vO M O + O F vF M F =

0 at the
reaction front, but now there are discontinuities in the species concentration at the front, with yF=1
and yO=0 in the solid phase, and yF=0 and yO>0 in the gas phase, the oxidiser concentration
decreasing towards zero as temperature increase.
Under some circumstances, two combustion fronts may develop: the heterogeneous one at the
particle surface mentioned above, and a homogeneous detached flame as in droplet burning.
Solid particles are far more difficult to ignite than droplets (they have higher flash points).
Ignition and combustion of pulverized coal particles is of major interest to modern coal-fuelled power
plants. Combustion of metal powders and polymeric dust in air is a major safety concern in many powder
industries (e.g. flour in food processing, saw-dust in wood processing).
The nature of coal combustion depends a lot on composition and temperature. Leaving aside coal
composition (see Coal analysis, in Fuel properties), the effect of temperature on pure-carbon combustion
is as follows (temperature range is very sensitive to humidity):
For T0<1500 K, i.e. when the surface temperatures (after ignition) does not go over 1500 K, a
heterogeneous combustion process is established at the surface according to C+O2CO2.
For 1500 K<T0<2500 K, a heterogeneous endothermic reaction sets at the surface, according to
C+CO22CO, driven by a gas-phase homogeneous combustion with a blue free flame at some
distance from the surface, according to CO+(1/2)O2CO2.
For 2500 K<T0<3500 K, the external free-flame disappears, leaving just the heterogeneous
combustion at the surface according to C+(1/2)O2CO.
For T0>3500 K, evaporation is so effective (carbon sublimates at 3910 K at 100 kPa) that a
detached flame develops as in droplet combustion, according to C+(1/2)O2CO, although oxygen
dissociation becomes important too.
The surface temperature in carbon combustion depends on how quick heterogeneous reactions proceed
relative to the species-diffusion rates, what depends on geometry, trace contaminants, catalysts, etc. In the
simplest case of diffusion-limited combustion at low ambient temperatures, once the heterogeneous
reaction C+O2CO2 is established, a surface temperature is in the range 1000 K<T0<1500 K is obtained,
with a radial profiles as sketched in Fig. 7a, whereas for higher temperatures, 1500 K<T0<2500 K, a
detached flame appears, with a radial profiles as sketched in Fig. 7c.

Fig. 7. Combustion of a carbon particle: a) Radial temperature and species concentration profiles when
1000 K<T0<1500 K (the only product, P, is CO); b) Correlation of species concentration close to
the surface, with surface temperature, for 1000 K<T0<1500 K; c) Radial temperature and species
concentration profiles when a detached flame appears (1500 K<T0<2500 K).
The species and energy balances for the case without free flame, closely following the analysis detailed
above for droplet burning, yield:

MO

O
+ yO,
dr 2

2 DO
2 DO yO, yO,0
MF
0 =

ln
K =

sol
sol M O
dt
MO + y
O,0
O
OM

MF
F

P P
dr 2

2 DP
2 DP yP,0
MF
0

ln
=
K =

sol
sol M O
M P y
dt
P
P,0
O

MF

MF

HV + c p (T T )
dr02
2k
2k c p (T T0 )

ln
=
K =

c p sol HV + c p (T0 T )
c p sol
HV
dt

(33)

what relates all other variables with the surface temperature attained; e.g., assuming DO=DP=k/(cp), with
HV=33
MJ/kg
for
C+O2CO2,
and
cp=1200
J/(kg/K),
one
gets
(yO,yO,0)/(32/12)=yP,0/(44/12)=1200(T0T)/(33106), with T in kelvin, what is plotted in Fig. 7b.
Exercise 1. Partial oxidation of carbon particles at high temperature

PROPAGATION OF PREMIXED FLAMES

People learnt to make non-premixed flames on torches some 500 000 years ago, but it was only very
recently, after R.W. Bunsen conceived his famous burner in 1855, that premixed flames are used, not only
in most gas-fired burners (from cooking ranges to industrial furnaces) but also in spark-ignition engines
and some gas turbines, because premixed combustion is more energy-intense and less pollutant than nonpremixed flames. Premixed flames may be laminar (e.g. lab bunsen burner) or turbulent (industrial
burners).

Full mixing of a fuel in air takes a lot of time, even when the fuel is already a gas. For mixing to occur in
any reasonable engineering-time-span, it must always be made by turbulent convection. The quickest way
is by letting a fine dispersion of liquid fuel (obtained by high-pressure injection) to evaporate and mix
with air (as in gasoline engines); gaseous jets are more expensive to produce (require more power for
pressurisation) and are less ready to disperse (recall the long bunsen tube).
The degree of mixing in traditional Otto engines (both, for the new port-fuel-injection models and the old
carburettor models), is perfect, because there is a lot of time from mixture preparation outside the cylinder
to ignition after compression. However, the mixing is not so good in the new direct injection spark
ignition engines (DISI, also named GDI for gasoline direct injection), which work with a stratified
mixture, since there may be no time enough from injection to ignition; DISI engines (first developed by
Mitsubishi in 1995), only admit air to the cylinders (without a throttling valve), and inject fuel within the
cylinder during the compression stroke, with the injector (placed near the spar-plug) controlled in such a
way that as to have a stratified mixture, only rich near the spark-plug (i.e. easily flammable), which
causes the burning of the overall lean mixture (increasing a lot the efficiency on partial loads, one of the
weakness of Otto engines).

LAMINAR COMBUSTION
Perfectly premixed, quiescent flames are the easiest to analyse, the most instructive to study, and a
prerequisite to understand any combustion phenomena. Although the most common laminar premixed
flame is obtained at the tip of the bunsen burner, it is easier to consider a flame propagating inside a long
tube filled with a premixture (e.g. a 10% in volume of methane in air, at ambient temperature and
pressure, being ignited by a spark), assuming as a first approximation a planar one-dimensional
propagation.
The main feature of premixed flames is their characteristic propagation speed, but there are other
attributes that distinguish them from non-premixed flames, none the least their colour: contrary to the
yellow colour of non-premixed flames (due to blackbody emission by soot particles), premixed flames
show a green-bluish coloration and much lower luminous intensities. This light emission is not due to
blackbody radiation because, although the gases are at similar temperature in both types of flames, a huge
volume of gas is required to make it optically thick (several metres); otherwise, only some characteristic
spectral bands show up over the tenuous bulk emission, those corresponding to the species present. For
instance, rich premixed methane/air flames (e.g. as in a cooking range, when some primary air is used)
are blue/green due to CH and C2 radicals, whereas lean premixed methane/air flames are green due to C2
radicals. Although the predominant colour in premixed flames is blue, as for methane and natural gas,
propane and LPG flames are greener and brighter, and alcohol and hydrogen flames are purple and
tenuous.
The additional non-premixed flame surrounding the rich premixed methane/air flame is also blue (and not
yellow, as in normal non-premixed flames) because now there is little fuel remaining and consequently
soot formation is hindered.
The strong temperature gradient is one of the main characteristic of a flame. Mallard and Le Chtelier
proposed in 1883 a model for the propagation of a flame in quiescent premixed gases (laminar
deflagration), assuming that the heat-release zone was much smaller than the temperature-change zone
(Fig. 8), i.e. just a thermal model, equivalent to the steady-state in the adiabatic motion of a planar heatsource.

Fig. 8. Internal structure of a premixed laminar flame, and its equivalent thermal model.
In axes moving with the front, the energy balance (heat equation) for a generic unit volume yields:
T Tf
dh
2T
dT
d 2T
dh

xV
=a 2 ,
= k 2 , VL
= exp L

= q , VL
dx
dx
dx
x
Tb T f
dt
a

(34)

where for the fresh mixture entering from the left at Tf and the burnt products exiting to the right at Tb;
this is valid for <x0 (it is T=Tb for x0). The heat release per unit area is
=
Q / A f VL c p (Tb T f ) 1 MW/m 2 (i.e. some 100 W per square centimetre of flame), having substituted
typical values for density of fresh gases, =1 kg/m3, for the laminar deflagration speed VL=0.5 m/s, for
the thermal capacity cp=1000 J(kgK), and for the temperature increase TbTf=2000 K.
The most important feature of this thermal conduction model is obtained by a mere order of magnitude
analysis that shows that the thermal thickness, Lth, is related to the deflagration speed by just Lth=a/VL,
from the exponential argument in (34), i.e., for a mean temperature of 1500 K (a=210-5 m2/s at room
temperature, but at 1500 K a=0.3510-3 m2/s) one gets LthVL=0.3510-3 m2/s, what agrees very well with
typical values of Lth=10-3 m and VL=0.4 m/s. A more refined analysis yields LthVL=2a. A further analysis
of the laminar deflagration speed VL follows.
Deflagration speed
The laminar deflagration velocity VL is the flame-front speed relative to the quiescent, premixed, fresh
gases (i.e. the volumetric rate of fresh mixture being burnt, per unit area). Besides, VL, it is sometimes
named SL, for laminar speed, or Vb, for burning velocity, or Vf, for flame velocity, or VD, for deflagration
velocity, etc., but VL seems preferable because it reminds of the laminar character (practical combustion is

turbulent) and the apparent flame speed may vary with the configuration (e.g. the observable speed of a
flame in a tube only coincides with VL for the open-exit / closed-intake tube).
The laminar deflagration velocity VL varies with type of fuel, type of oxidiser, relative air ratio in the
mixture, =A/A0, temperature, Tf, and pressure, pf.
Effect of fuel. VL has a maximum value of near 0.5 m/s for most hydrocarbons: VL=0.45 m/s for
CH4, C3H8, C6H6, CO, but VL=0.75 m/s for C2H4, VL=0.85 m/s for CH3-CH2OH, VL=1.6 m/s for
C2H2 and VL=3.5 m/s for H2.
Effect of oxidiser. It depends on oxidiser composition (air, O2, H2O2). For air, VL increases one
order of magnitude from xO2=0.21 to xO2=1 (e.g. VL=3.9 m/s for CH4/O2, VL=3.6 m/s for C3H8/O2,
VL=3.3 m/s for C4H10/O2, VL=7.5 m/s for C2H2/O2 and VL=14 m/s for H2/O2). Notice that pure
oxygen may give rise to a detonation (see Supersonic combustion, below). Also, for a given xO2,
VL varies with the dilutant: with Ar and He it is double than with N2. However, although adding an
inert gas lowers the deflagration speed, in some special cases it may act as a positive catalyst and
increase it, as in the striking case of CO/O2 combustion, where a 0.2% H2O causes an eight-fold
increase of VL. Other striking case is the reduction of VL from 2.7 m/s to 0.15 m/s for H2/air
flames when adding 3.5% butane.
Effect of mixture ratio. Experiments show that VL has a maximum value for slightly over =1,
and decreases parabolically towards the ignition limits (to some 30%..10% of its maximum value,
not to zero). For premixed methane/air mixtures at ambient conditions, VL decreases from VL=0.45
m/s at =1.05, to VL=0.15 m/s at =0.65 and at =1.7; a minimum value of VL=0.036 m/s at
=2.08 has been found under microgravity conditions); Fig. 9 presents additional data.

Fig. 9. Laminar deflagration speed vs. equivalence ratio, =1/, for some fuels, at room
conditions.

Effect of temperature. VL increases exponentially with fresh gases temperature, Tf (e.g. from
VL=0.45 m/s at 300 K to VL=1.2 m/s at 600 K).
Effect of pressure. VL decreases with increasing pressure to a half from 100 kPa to 1 MPa (it
increases 50% from 100 kPa to 10 kPa). An empirical correlation from Andrews and Bradley for
stoichiometric CH4/air flames is V=
( c1 + c2T 2 ) p p0 , with c1=0.1 m/s, c2=3.710-6 (m/s)/K2
L
and p0=100 kPa.

Some prediction of all the above-described VL-dependence may be obtained following the thermal model
above. The burning rate density, /V , must be proportional to the overall concentration of fresh gases,
cf=p/RuTf=fMm, and inversely-proportional to the time taken to burn the gases inside the flame thickness,
i.e. to tth=Lth/VL, so that, upon substitution of LthVb=a in /V =cf/tth and elimination yields:

VL
=

a
=
cf V

M m dcF
f dt

(35)

or with a more refined analysis:

=
VL

a (1 + O )

M F dcF
f dt

(36)

where Mm and MF are the molar masses of the mixture and the fuel, O is the stoichiometric mass
coefficient for the oxidiser (O2) relative to unit mass of fuel (e.g. O=15.6 for methane in air), f the
unburnt mixture density, and dcF/dt the speed of reaction for the fuel. But the effect of the chemical
kinetics is difficult to guess beyond Arrheniuss law and the law of mass action.
There are many procedures to measure laminar deflagration speeds, usually grouped in stationary and
moving flames. For planar geometries, a flame stabilised over a porous burner is better than the travelling
flame along a tube, since the effect of gravity on the later, gives way to different speeds for upward and
downward propagation, and a non-symmetric parabolic flame in horizontal propagation, increasing the
burning area and the speed. For cylindrical geometries, the typical bunsen flame may be used, and better
if a planar velocity profile for the issuing jet is achieved by a contraction (in any case, the overall
injection speed, times the cross-section area, divided by the flame area, gives a 20% approximation to
the deflagration speed). For spherical geometries, central ignition inside a soap bubble filled with a
premixed fuel/air mixture may be used.
The laminar deflagration speeds here consider refer to open flames. Flames inside porous media have
higher values; e.g. for methane/air, maximum deflagration speed increases from 0.45 m/s in free-flames
to 4 m/s in porous-media flames, explained because of the 10-fold increases in thermal diffusivity a (100
times in k, but 10 times in c).
It is important to keep in mind that the products leave the flame at a quicker pace than the deflagration
speed, due to the effect of temperature on mass conservation. For instance, in the planar geometry,
fVL=bvb, i.e. vb/VL=(pf/pb)(Mf/Mb)(Tb/Tf), where the last factor in the right-hand-side product is the
dominant (of order 7 or 8, the other two being of order 1). Thus, for a typical 0.4 m/s flame speed towards
the fresh gases, the products are ejected at some 2.5 m/s to the other side.
It must be reminded that all this refers to the laminar speed; propagation speed greatly increases
(parabolically) with Reynolds number in the turbulent regime, what is profitably used in practical systems
(typical turbulent speeds are some 5 times larger than the laminar speed).
Flame thickness
The flame front is very thin, in either premixed or non-premixed flames; usually less than a millimetre.
Some comments made when analysing non-premixed flame lengths, apply also here: the uncertainty in
measuring the thickness of a solid sheet is well understood, even for an elastic material, but, what about
for the thickness of a gaseous layer as in a flame? One possible answer may be its thermal thickness (once
we agree on a certain norm, e.g. the mean distance between points with temperature say 1% greater than
the entering fresh mixture and the points with temperature say 1% lower than the exiting products, from a
temperature profile as in Fig. 8). But another possibility may be its light-intensity thickness (once we
agree also on a certain norm, e.g. using a certain photographic plate, or a given monochromatic band, and
the threshold levels); or some chemical thickness; etc. The visible thickness is always smaller than the
thermal thickness (say one half), and the visible layer stays at the downstream-end of the thermal layer
(see Fig. 8), there where radicals recombine.

As reasoned before, the thermal thickness, Lth, is related to the deflagration speed by just LthVL=a, with
typical values of 1 mm, that may grow to 1 cm at very low pressures (10 kPa). This applies to nearstoichiometric deflagration speeds; outside that range, flame thickness grows parabolically (e.g. for
premixed methane/air mixtures at ambient conditions, Lth grows from 1 mm at =1, to Lth=10 mm at
=0.6 and at =2.
To help develop appropriate kinetic models, it is very important to diagnose the species production rate
profiles within the flame. The main reactives (fuel and oxygen) show a simple consumption rate profile
(say a Gauss bell in the dci/dt profile), similar to the main products (H2O and CO2), that show a simple
production rate profile, whereas CO, NOx and radicals show first a neat production rate profile, followed
by a less pronounced consumption rate profile, that, if there is not enough time to equilibrium, causes a
neat production and consequent emission.
Flame stabilisation
Contrary to non-premixed flames, that naturally stabilise at the shear layer between fuel and oxidiser
because they cannot propagate to one side or the other, premixed flames have their own propagation
speed and, when fed with the fresh mixture, except in a very short range of matching the propagation
speed, they tend to flashback or blowoff. For instance, if a stable CH4/air premixed-flame is to be realised
at the mouth of a bunsen burner, the mixture (within its ignition limits) must be fed between twice and
five times the laminar deflagration speed.
Several methods are employed to stabilise a partially or totally premixed flame a desired location in a
combustor, against natural uncontrolled fluctuations and the intended operation range, all methods trying
to procure a region where natural flame speeds are matched:
Sudden flow expansion, giving rise to detached flows with a vast range of speeds.
Swirling by jet induction or bypass flows, giving rise to a recirculation with a vast range of speeds
Bluff-bodies, given way to a vast range of speeds at their wakes.
Local hot spots (e.g. ceramics with large thermal inertia) that permanently ignite the mixture.
The trade-off between flame stability (greater for non-premixed flames) and flame energy intensity
(greater for premixed flames) is usually solved by a sequential combination of a premixed flame and a
non-premixed flame; i.e. some primary air is added to the fuel (more than the lower flammability limit
and less than the stoichiometric ratio), a premixed flame is established, and the remaining fuel is burnt in
a second non-premixed flame with another supply of air (e.g. from the ambient). These partially-premixed
combined flames are the usual feature in bunsen burners and domestic gas-stoves and gas-heaters. Their
stability diagram, in terms of the primary air (that used in the premixed flame) is presented in Fig. 10.

Fig. 10. Stability of bunsen flames in terms of primary air. When both, fuel and air are supplied the
stability diagram is presented in terms of the equivalence ratio of the mixture supplied.

The easiest way to light a bunsen is with the primary-air openings closed (0% primary air, infinite
equivalence ratio of the feeding mixture) because this flame is most stable (it would detach only at an
order of magnitude larger exit speeds); but this flame is very long and yellow, and a better combustion
can be achieved by increasing the primary air to 60% or 70% of the stoichiometric (i.e. increasing the
relative air/fuel ratio to 0.6 or 0.7, or decreasing the equivalence ratio to 1.8 or 1.5). Lean mixtures are
more unstable, and, if more air than that corresponding to the lower flammability limit is supplied, the
mixture cannot be ignited at room temperature.
Flame quenching
Quenching is the extinction of a flame by an excessive heat-loss towards relatively-cold solid boundaries
nearby. Quenching prevents the propagation of combustion along narrow tubes or orifices and through
fine wire-meshes (see Fig. 11). Quenching distance is typically of 2 mm for normal hydrocarbon flames
(e.g. for stoichiometric fuel/air mixtures at 100 kPa and 20 C, dQ=1.9 mm for CH4, dQ=2.1 mm for C3H8,
dQ=2.0 mm for n-C8H18, dQ=1.8 mm for CH4O but dQ=0.6 mm for H2). These values increase with the
air/fuel ratio outside the minimum value (corresponding to 0.9), and decrease with pressure, with
corrugation (non-circular ducts have slightly lower values), with temperature (the quenching effect is lost
at high temperature, as in combustion within porous media), and with the presence of catalysts. Domestic
gas-fired appliances usually have multi-hole or multi-blade burners (slightly corrugated slots to increase
mechanical stiffness), with passage sizes below 3 mm to prevent flame ingestion.

Fig. 11. A fine wire mesh prevents flame propagation through, so that if the premixed jet is ignited at one
side of the mesh, the flame cannot traverse to the other side.
Flammability limits
Non-premixed flames are very stable in the sense that, once ignited, they tend to stay anchored to the
burner under most variable circumstances, but a premixed fuel/air mixture may not propagate a local
combustion (e.g. by a spark stimulation) if there is too-much air or too-much fuel, because a certain
energy balance must be accomplished between the chemical energy release and heat transfer.
The amounts of gaseous fuel (usually in volume terms in the mixture) between which a flame propagates
in a quiescent mixture at ambient conditions, excited by a spark of some 10 J, are called flammability
limits (also ignition or explosive limits), defining a flammability range. The lower (or lean) flammability
limit, LFL (also known as lower explosive limit LEL, or lower ignition limit LIL), is of greater interest
than the upper flammability limit, UFL (or rich flammability limit or upper explosion limit, or upper
ignition limit), since, if the fuel molar fraction is lower than xLFL there is no further risk of accidental
combustion, but if it is greater than xUFL, a subsequent escape of mixture will yield flammable mixtures
with omnipresent air.
Some typical flammability limits are presented in Table 2. Common units for LFL and UFL are molar
units, i.e. fuel molar fraction in the gaseous mixture, or, what is the same, in percent volume, %vol. Those
values can be converted to other units: e.g. for methane, {xLFL=5%, xstq=9.5%, xUFL=14%} from Mass

diffusivity data, may be expressed also in terms of equivalent ratios {LFL=0.5, stq=1, UFL=2}, relative
air/fuel ratios {UFL=0.5, stq=1, LFL=2}, molar air/fuel ratios {UFL=5.7, stq=9.52, LFL=19}, mass
air/fuel ratios {UFL=10, stq=17.2, LFL=34}, etc. Le Chtelier proposed in 1898 a rule to compute the
lower flammability limit for a mixture of gaseous fuels of molar fractions xi: xLFLm=1/(xi/xLFLi), where
only the xi of the actual fuels, not of the possible inert components, must be considered.
Table. 2. Flammability limits (in %vol) for premixed fuel/air gaseous mixtures at 25 C, minimum
quenching diameter at 25 C, and autoignition temperature, all at 100 kPa.
Fuel
Lower
Stoichiometric
Upper
Minimum
Autoignition
Flammability
mixture
Flammability
quenching
temperature
Limit, LFL
Limit, UFL distance [mm]
[K]
CO, carbon monoxide
14
29.5
70
900
H2, hydrogen
3.9
29.5
75
0.61
850
CH4, methane
4.9
9.47
14
2.0
850
C2H6, ethane
3.0
5.64
13
1.8
800
C3H8, propane
2.0
4.02
9.5
1.8
750
C4H10, n-butane
1.5
3.13
8.5
700
C7H16, n-heptane
1.1
1.87
6.7
560
C8H18, n-octane
1.0
1.65
6
500
C8H18, iso-octane
0.95
1.65
6
690
C10H22, n-decane
0.80
1.34
5.4
2.1
480
C12H26, n-dodecane
0.60
1.12
480
C2H2, acetylene
2.5
7.75
80
2.3
600
CH4O, methanol
6.7
12.3
36
1.5
680
C2H6O, ethanol
3.3
6.54
19
630
C4H10O, ether
1.8
3.13
37
440
Notice that for a liquid fuel tank at room conditions, three cases can be considered for the flammability
risk (e.g. in the event of an electrostatic spark inside the tank), according to its equilibrium vapour
pressure, pF*(T): fuels with vapour pressure lower than xLFLp pose no spark-risk (e.g. diesel oil, fuel oil),
fuels with vapour pressure xLFLp<pF*(T)<xUFLp pose a great spark-risk (e.g. methanol, ethanol), and fuels
with vapour pressure pF*(T)>xUFLp pose no spark-risk inside the tank (e.g. gasoline, ether, acetone),
although in the latter case the venting gases will be flammable when in contact with more air. Another
way to verify this tank-flammability behaviour is to put some liquid fuel in an Erlenmeyer flask (e.g. 5
cm3 in a 100 cm3 beaker), close it, agitate, open it, and approach a lighter flame; diesel oil will not burn,
gasoline burns with a non-premixed flame outside the beaker, and methanol burns violently inside the
flask (without explosion because of the wide-enough opening).
Flammability limits above refer to fuel and air at room temperature and pressure conditions. The effects
of changes in these parameters are:
With pure oxygen, the upper flammability limit increases a lot, not so the LFL (e.g. for H2/O2 the
change is from xFL=4..75%vol to xFL=4..95%vol, for CH4/O2 the change is from xFL=5..14%vol to
xFL=5..60%vol, for C2H6/O2 the change is to xFL=3..66%vol, for C3H8/O2 to xFL=2..45%vol, for
C4H10/O2 to xFL=1.5..40%vol and for C2H2/O2 xFL=2.5..95%vol).
With the addition of an inert gas, the flammable range gets smaller, up to a point where the
mixture is no longer flammable at room conditions; e.g. for CH4/air mixtures, this point is reached
with 21% of CO2 in volume, or 26% of water vapour, or 35% N2, or 48% He. Notice that for
CH4/O2 mixtures up to 80% N2 may be needed to have a non-ignitable mixture (Fig. 12).

With temperature, the flammable range gets slightly wider (particularly along the rich side) up to
an autoignition temperature (depending on pressure and composition) where any mixture ratio will
burn. For instance, for H2 at 300 C xFL=2.4..83%vol, or for CH4 at 300 C xFL=3.8..17%vol.
With pressure, the flammable range widens: e.g., for methane/air, xLFL=4.9% and xUFL=14% at 0.1
MPa, xLFL=3.8% and xUFL=55% at 10 MPa, and xLFL=3.1% and xUFL=60% at 20 MPa.

Fig. 12. Influence of an inert gas on the flammability limits of CH4/air and CH4/O2 mixtures at room
conditions.
Autoignition temperature
A premixed fuel/oxidiser mixture will spontaneously burn at elevated temperature irrespective of
pressure, composition, size, etc. The minimum temperature for this to occur at 100 kPa in air is called
autoignition temperature (also known as spontaneous ignition temperature, SIT, some values of which
can be found in Table 2). Autoignition is related to heat release and dissipation; according to van't Hoff,
Tautoign is just the temperature at which the rate of generation of heat becomes greater than its rate of
dissipation. Substitution of oxygen for air has little effect on autoignition temperature (e.g. 850 K for
methane/air and 825 K for methane/oxygen), implying the process is governed more by the Arrhenius
expression than by concentrations.
Combustion in this case is not by a flame spatially propagating but by homogeneous combustion of the
whole volume at once. A non-premixed fuel/oxidiser mixture will also start burning at high temperature,
but here the propagation is as for normal non-premixed flames: at the rate of supply. Practical use of that
is made in diesel engines, where intake air is quickly compressed to get it hot enough for the fuel to burn
by contact with the hot air, and in some steady combustors where high exhaust-gas-recirculation rates
yields high-enough intake temperatures. In reality, some premixed combustion takes place in diesel
engines too, since some fuel gets vaporised and mixed with air before autoignition actually occurs (that is
why a two-function Weibe model best fit actual pressure plots). Moreover, in the new dual-fuel diesel
engines, a lean premixture (prepared with a high-octane fuel at intake pressures) is fed to the cylinders
(with relative air/fuel ratio in the range =0.15..0.45, outside the flammability range), compressed below
its high autoignition temperature (e.g. <850 K for natural gas) and made to burn by compression ignition
of a minute injection at high pressure of low-autoignition-temperature fuel (e.g. >480 K for diesel oil).
Autoignition may be a hazard in some special cases at low temperature: phosphorus has an autoignition
temperature of only Tautoign=34 C, carbon disulphide has Tautoign=90 C, diethyl ether has Tautoign=160 C
(it can be ignited by a hot plate), sulfur has Tautoign=240 C, etc. Additionally, it must be recalled that the
quoted autoignition temperatures refer to sudden inflammation, but, if the system is well insulated,
autoignition may occur at much lower temperatures after some self-heating period (e.g. a H2/air mixture

will suddenly burn by itself at some 850 K, but it has been reported that in well insulated envelops it selfignites after a couple of hours at 750 K).

SUPERSONIC COMBUSTION
Combustion is a slow process (think of the maximum premixed laminar flame speed of 0.45 m/s for
methane/air, or the slow burning of wood in the fireplace). High burning rates required in engineering
applications, due to short residence time dictated by kinematics in reciprocating engines or size in other
combustors, are achieved by promoting turbulence (e.g. in a gas turbine combustor the average speed is
some 50 m/s, two orders of magnitude larger than typical laminar flame speeds, without the flame being
blow-off (aided by some flame-holder devices). But, could the flame speed be raised to say 2000 m/s?
The answer is yes, supersonic combustion sometimes takes place accidentally in coal mines and other
closed places, as well as within condensed explosives, but no controlled application has been developed
yet. H2/O2 and C2H2/O2 mixtures are particularly prone to detonation. Notice that the ignition limits for
supersonic combustion are usually narrower that for deflagration; e.g. for H2/air premixed combustion,
the flammability limits change from xFL=4..75%vol to detonability limits of xDL=18..59%vol.
Supersonic combustion, called detonation, is realised coupled to a shock wave travelling at supersonic
speed. Deflagration to detonation transition (DDT) is difficult to understand. If premixed gases inside a
tube closed at both ends are ignited at one end, a laminar flame first develops, travelling fast being pushed
by the expanding hot products behind. The inverted small pressure-jump across the flame generates local
pressure pulses that wrinkles the flame, creates turbulence, extends the burning area and increases the
burning rate, with a positive feedback that, if positively combined with pressure pulses reflected from the
other end, might compressed the fresh mixture to the autoignition temperature.
Although it is very difficult to have a planar geometry in practice (both in subsonic and supersonic
combustion), we analyse just the planar case because it yields the maximum benefit/cost for learning.
For a planar fluid front, be it a subsonic deflagration, a sonic acoustic wave, a supersonic shock wave, or
a supersonic combustion behind a shock wave, the mass, momentum and energy balance equations
through it at a steady state are:

=
=
m A
1v1
2 v2

(37)

p1 + 1v12 =p2 + 2 v22

(38)

h1 +

v12
v2
+ q = h2 + 2
2
2

(39)

where m A is the mass-flow-rate per unit area traversing the front, and q=yFPCI is the heat release
equivalent to the chemical energy of combustion (if there is such), that is taken apart from the thermomechanical enthalpy h.
The intention is to analyse the possible exiting conditions for a given entry conditions, what is
traditionally visualised in the pressure-volume diagram (p-v, or better p-1/, in terms of density, since we
keep v for velocities).
From (37) and (38) one gets the so called Rankine line (a straight line in the p-1/ diagram):
1
1
dp2
p2 p1 =
2 v22 1v12 =
m A2
= m A2
1
2 1
d

(40)

showing that, for a given entry conditions (point A in Fig. 13) the only possible exiting conditions lay in
the second and fourth quadrant from A, since m A is always positive.

Fig. 13. Possible exiting conditions (Rankine straight line) after a planar fluid wave, in the pressurevolume diagram, for a given entry conditions at point A. After a normal flame, point F is reached
(the Rankine line would join A and F). Acoustic waves take place as oscillations around point A
along q=0 (the Rankine line would be the local tangent). Normal chock waves correspond to
points like N (the Rankine line would join A and N). A detonation corresponds to point C-J
(Chapman-Jouguet), the tangent point for the Rankine line from A to the Hugoniot curve q>0
(positive heat release).
From (37), (38) and (39) one gets the so called Hugoniot curve (or Rankine-Hugoniot curve):

v2 v2
v +v
v + v p p1
p p1 1
1
h2 h1 =q 2 1 =q 2 1 ( v2 v1 ) =q + 2 1 2
=q + 2
+ (41)
2
2
2 2 1
1v1
2 2
that, with the perfect gas model (h=cpT, T=p/(R), cp/R=/( -1) yields:

p2 p1 1
1
p2 p1
q
=+
+
2 2 1
1 2 1

(42)

that, although not evident, corresponds to a kind of hyperbolic curve in the p-1/ diagram (Fig. 13),
displaced to the right of point A because of the q term (always positive).
In the case of a normal flame, the jump through the front is from A to F, with a small pressure drop (some
2%) and a large specific-volume increase (some 7 times).
The case of acoustic waves correspond to small departures to the left of A along the curve for q=0.
Assuming negligible energy dissipation, the process is isentropic, and from the differential forms of
equations (37) and (38), d(v)=0 and dp+d(v2)=0, one gets the general equation for the speed of sound:

2v 2
1
dp
= v2 =
vsound
dp + d
dp + 2 v 2 d =
0
=
d

p PGM
=
RT (43)
S

where the particularisation for the perfect gas model (PGM) can simply be obtained from differentiating
the logarithm of (p/)=constant.

The case of a normal shock wave correspond to large departures to the left of A along the curve for q=0.
Substituting =p/(RT) and v = M RT in equations (37), (38) and (39) for the perfect gas model, M
being the Mach number Mv/vsound), one gets:

pM
p
= constant
M RT
RT
T
p
p + v2 =
p+
M 2 RT =
p 1+ M 2 =
constant
RT
v2
M 2 RT
1 2
h + = c pT +
T 1 +
M =
constant
2
2
2

v =

(44)
(45)
(46)

from which one may get an explicit expressions for the exiting Mach number:
M2 =

2 + ( 1) M 12

2 M 12 ( 1)

(47)

and, upon substitution, the rest of the variables. Notice, by the way, that the Hugoniot curve must lay to
the right of the isentropic curve (p/)=constant passing by A, since there can only be entropy increases in
real adiabatic processes.
Finally, for the detonation case, the exiting conditions correspond precisely to the Chapman-Jouguet point
(C-J in Fig. 13); the rationale is that, if it were above C-J, the exiting velocity would be subsonic and the
combustion would not be able to sustain that strong shock wave, whereas if it were below C-J, the exiting
velocity would be supersonic and the combustion would tend to accelerate the shock wave; that is why it
is found in practice that detonation waves always approach the C-J point. The internal structure of the
detonation can be separated in a very thin leading region (a few molecular free-paths) of a strong shock
with a pressure jump of say 40-fold (and temperature and density jumps of say 7-fold and 5-fold,
respectively) leaving the stream subsonic with a local M0.4, followed by a much thicker combustion
region where the fluid accelerates towards the local sonic speed M=1 while increasing further the
temperature and decreasing both pressure and density. For a stoichiometric methane/air mixture at
ambient conditions, the detonation velocity is 1800 m/s (advancing speed relative to quiescent fresh
gases), varying little with pressure and composition and more with temperature, the corresponding Mach
number is M1=5.1, and the exiting temperature and pressure 2780 K and 1.72 MPa..

TURBULENT COMBUSTION
Combustion usually takes place in a turbulent flow, from combustion chambers in gas turbines, boilers
and furnaces, to reciprocating engines and uncontrolled fires. But, as in Fluid Mechanics, a larger share of
the learning effort is devoted to study laminar flows because they are more amenable, and more advanced
turbulent models are presently built upon them.
The basic characteristic of turbulence is that all local variables (velocities, pressure, temperature,
concentrations) show a random fluctuation superimposed to a running average changing more slowly (or
being constant if the flow is steady in the average; turbulence is always unsteady in the detail). This
randomness does not mean the flow is not deterministic; it means that the flow is too sensitive to
uncontrolled initial and boundary perturbations. In pipe flow, for instance, turbulent slugs may appear at
low Reynolds number (Re~200), decaying more slowly the higher this value, until at Re~2300 they no
longer decay in most setups; but in very smooth walls, smooth entry and quiet environment, the laminar
regime has been maintained up to Re~100 000.

Turbulence introduces a very effective exchange of mass, momentum and energy in cross-stream
directions, increasing physical transport by orders of magnitude, what is a blessing in engineering
problems limited by the mixing rate, as in combustion (obviously for non-premixed flames, but also for
extending the flame front in premixed flames, and for the dispersion of pollutant emissions).
For instance, for a velocity component u, along a given direction, x, a very quick scanning (e.g. laserdoppler, hot-wire or ultrasonic velocimetry) would yield a fluctuating variable u(t,x,y,z) at a given point
(x,y,z). But u(t) is aperiodic and its average would depend on the time-span considered, t. In practice,
one can find suitable t values that depend on the problem at hand; for typical engineering problems it
may be t<10-2 s, i.e. fluctuations below 102 Hz are retained, and above that are averaged), whereas for
meteorological problems it may be t103 s, i.e. fluctuations below 10-3 Hz are retained, and above that
are averaged). The theoretical value for t would be determined by the autocorrelation of the function
being averaged. Assuming t is well defined, the fluctuating variable could be decomposed in a running
mean, u , and a random fluctuation u:

=
u

u +
u'

mean

inst.

with u =

fluct.

t +t

t t

udt

(48)

Turbulence theories try to model this statistic problem by reducing the number of representative
parameters. Unfortunately, the non-linear convective term in Navier-Stokes equations make the statistic
analysis non-linear, introducing higher-order coupling terms at any stage in the decomposition problem,
giving rise to a closure problem. For instance, the momentum equation along x for incompressible flow,
when averaged, reduces to:

( ( u + u ') )

( p + p ')

+ ( ( u + u ')( v + v ') ) =
+ g x + 2 ( u + u ')
t
x
( u )
( p)

+ ( uv ) + u ' v ' =
+ g x + 2u

t
x

(49)

since the mean of single random variables is zero. The new term appearing in turbulent flows (the third in
the left hand side) may be transposed to the right-hand-side of (49) and thence it is called turbulent
Reynolds stress, defining an empirical turbulent viscosity, t, usually much larger than the viscous one):

u ' v ' t 2u

(50)

Turbulence is always a three-dimensional spatial-temporal problem, and thus spatial and temporal
correlations are sought. The parameters introduced to quantify turbulence are of two types:
Turbulent intensities of the variables, i.e. root-square-means in the temporal fluctuations of
velocities, temperature, concentrations), that may be of the same order as the mean values.
Turbulent scales in the coordinates, defining a region of fluid with correlated properties (an eddy
or vortex, since turbulent motion is highly rotational), that may be of the same order as the
geometrical size of the problem.
Several length scales may be defined in turbulent flows; from the largest to the smallest: the main length
of the geometry, L (e.g. the diameter in a pipe flow), the large-eddy scale, LLE=

0 )u 'rms ( r )) , the Taylor mesoscale, LTY= u 'rms /(u '/ y ) rms , and the Kolmogorof
r0 u '(r0 )u '(r )dr /(u 'rms3 (r1/4
microscale, LKG=( /) =[3(LLE/urm)/((3/2)u2rm )]1/4 (where is the kinematic viscosity and the
energy dissipation rate, =(ui'/xi)rms), beyond which there is no further convective-transport but
viscous-dissipation. Associated Reynolds numbers are defined as ReL=uL/, ReLE=uLLE/, ReTY=uLTY/

and ReKG=uLKG/, with the ratios LLE/LTY=ReLE3/4 and LLE/LKG=ReLE1/2. In reality, there is a continuum in
length scales (the turbulent cascade) and in time scales (the frequency spectrum of energy distribution is
continuum), all due to the non-linear character of turbulence.
Several turbulent intensities may be defined in turbulent flows: the eddy viscosity just introduced in (50),
t, the relative turbulent intensity, iuurm/ u , the specific turbulent kinetic energy, k u '2rms / 2 , the
turbulent mixing length, m (defined by t = 2m ( u ) / y ), the turbulent kinetic energy dissipation rate, ,
etc.); the traditional k- model is widely used. More elaborated models are being tried, as the Reynolds
stresses model in terms of ternary correlations, the large-eddy simulation (where only smaller scales are
averaged) and the direct numerical simulation of the whole statistical problem (only affordable for
academic configurations).
Trying to render the formulation of real turbulent flows more amenable, some of the following
simplifications may be imposed on the model:
Steady turbulence, when durm/dt=0 for all fluctuating components.
Homogeneous turbulence, when durm/dx=0 for all fluctuating components in all directions.
Isotropic turbulence, when durm/d=0 for all fluctuating components in all angular directions.
Sometimes only local lateral isotropy is imposed, i.e. two turbulent intensities are retained: the
axial one along the local mean velocity, and the lateral one (assumed isotropic).
Axisymmetric turbulent jet flames
Perhaps the more representative case of turbulent flow in combustion is the turbulent jet. When one
stream issues into an otherwise quiescent fluid (gas in gas, or liquid in liquid; the case of a liquid in a gas
being very different), the incoming stream detaches at the exit lip, becoming a predominantly axial flow
(jet) that decays along the axial length while entraining external fluid and broadening in a more-or-less
conical flow; i.e. for large distances compared to the exit diameter, D0, there is a conical similarity
solution in r/z. The radius where the axial speed is half of the central speed, r1/2, or either where it is 1%,
r99, is used to characterise the width of the jet.
For laminar flows (Re=w0D0/<1000 or even Re<10 for non-parabolic jets) de semi-angle is
small and given by r1/2/z=6/Re (e.g. 3.5 for Re=100), with the central speed decreasing
hyperbolically as w=(3/32)(w0D0/z)Re. This conical solution is only valid for z/D0>9.
For turbulent flows, Re>2000, the semi-angle is larger (typically 9), with the central speed
decreasing hyperbolically as w=6.6w0D0/z, both independent of the Reynolds number. This
conical solution is only valid from z/D0>8 to z/D0<100 with the origin at z/D0=0.6.
Non-premixed combustion of an axisymmetric gaseous fuel jet in air show a laminar behaviour up to
Re=w0D0/<1000, with a flame height increasing linearly with exit velocity (see Flame length). A
transition to turbulent jet combustion takes place at higher Re (up to Re>2000 for H2 but Re>9000 for
C3H8), with a fully-turbulent flame-length independent of speed for higher Reynolds numbers. For
instance, for a methane/air jet issuing from an orifice 1 mm in diameter, laminar flames are obtained for
injection speeds up to 20 m/s, and turbulent flames up to 40 m/s, beyond which the flame detaches from
the rim of the injector but remains stabilised in a lifted position (increasing with flow-rate to more than 10
diameters) up to 80 m/s, when it blows-off and extinguishes. Larger diameter flames do not blow off,
explaining why natural-gas and oil-well fires cannot be extinguished by blowing out.
Turbulent premixed flames
Laminar premixed flames were characterised by a propagation speed that only depended on fluid
properties (fuel and oxidiser type, concentrations, temperature and pressure). Turbulent premixed flame
speed, however, is dominated by turbulence intensity. For low turbulence-intensity levels, the flame
appears wrinkled in many connected flamelets along the main flame front. For larger turbulence-intensity

levels, the flamelets get disconnected, being transported by the eddies, until a generalised bulk reaction
zone develops.

CHEMICAL KINETICS
Thermochemistry considers how much heat, work and temperature can be obtained from a combustion
reaction, and how is the exhaust composition, but says nothing about the reaction rate, i.e. how quick the
process takes place, what is key to define the size of combustors.
The interest now is in the reaction rate, , for a given reaction (or better its density /V , to have an
intensive variable), instead of the reaction advance at equilibrium, eq, as in Thermochemistry. Recalling:

ni ni 0

dni
dci
d
=
, = =
,
, for a given reaction 0 = i M i
dt V V i dt i dt

(51)

where ci is the molar concentration of species i in the mixture (the common units used being mol per litre,
mol/L, and the SI unit mol/m3).
But before specific kinetic models can be developed to study basic combustion processes, a review of
general chemical kinetics is due.

REACTION MECHANISM AND REACTION RATE

Consider the reaction of hydrogen with oxygen to yield water, that we usually write as H2+(1/2)O2=H2O,
and let us think at the molecular level of detail. First, we should write 2H2+O2=2H2O because you can
have half-a-mole, but not half-a-molecule. Second, we should write 2H2+O22H2O because we focus the
attention on the forward reaction and not on the backward one (i.e. the possible decomposition of water to
hydrogen and oxygen). But third, it is clearly unrealistic to think that in a single step two H-H molecules
interact with one O=O molecule to yield two
molecules.
Instead of the above global reaction, a more realistic elementary mechanism must be thought of in terms
of uni-molecular, bimolecular and ter-molecular steps.
Uni-molecular reactions are those where one molecule of reactive breaks down by collision with
an unspecified molecule at a very high relative speed, or when an already excited molecule hits
another and decompose. The unspecified molecule is required to balance the momentum and
energy of the uni-molecular chemical change. Example: the combustion of hydrogen must start by
dissociation of a hydrogen molecule (it requires less energy than oxygen dissociation):
H2+MH+H+M, where M stands for an unspecified molecule (hydrogen, oxygen, a hot wire).
Uni-molecular reactions are very rare at low temperature (the fraction of molecules with very
large speeds according to the Maxwell-Boltzmann distribution law is not enough to flame the
system and propagate the reaction, except when locally excited (e.g. spark ignition) or globally
heated (autoignition).
Bimolecular reactions are those where two molecules of reactives recombine by collision because
of their very high relative speed or because they are already excited; example in the combustion of
hydrogen may be H2+O2HO2+H and H2+OHH2O+H, respectively.
Ter-molecular reactions are those where three molecules happen to collide together at the same
time. Ter-molecular collisions are very much improbable than bimolecular collisions, but they are
more fertile (a larger fraction of chocks yield chemical change) because there are more
possibilities for the redistribution of momentum and energy. An example in the combustion of
hydrogen may be H+OH+MH2O+M.

The many intermediate species appearing (e.g. H, HO2, OH,) are called radicals (e.g. atomic hydrogen
radical, hydroperoxide radical, hydroxyl radical,); they are electrically neutral but with unpaired
electrons. For instance, in the combustion of hydrogen in oxygen, a reaction mechanism with some 40
elementary reactions involving 8 species (H2, O2, H2O, OH, O, H, H2O and H2O2) has been found
appropriate to model the real reaction in an ample temperature and pressure range. For the combustion of
methane in air, current successful models use nearly 300 elementary reactions involving some 50 species.
Most of the times, a reaction mechanism is dependent on the range of temperatures and pressures
involved, and different mechanisms have to be developed for free-flame kinetics, for ignition kinetics, for
very-low-pressure kinetics (wall effects become predominant), for porous burning kinetics, etc.
But for either the global reaction or any elementary reaction in a reaction mechanism, the question
remains: what are the parameters that influence the rate of advance? It was well-known that increasing
temperature had a tremendous influence on the reaction rate; for the influence of pressure there was little
knowledge and little interest; but how about the influence of the reactives themselves? The answer is the
law of mass action.
Mass action law
The Norwegian chemistry professors Cato Guldberg and Peter Waage (brothers in law), first presented in
1864, in their Studies Concerning Affinity, what they called the action of masses in chemical
reactions, now called the law of mass action: at a constant temperature and pressure, the rate of
chemical reaction is directly proportional to the concentration of the reacting substances raised to
particular exponents. For instance, for the combustion of hydrogen in oxygen at some temperature and
pressure, it is found:

d [H 2 O] d [H 2 ] d [O 2 ]
= =
= = k (T , p )[H 2 ]1.5 [O2 ]0.7 for 2H2+O22H2O
V
2dt
2dt
dt

(52)

where the common notation for concentrations [M]cM has been used. The coefficient k(T,p) is called
rate coefficient. The sum of the exponents is called reaction order (order 2.2 (1.5+0.7) in the example),
and each exponent is called reaction order in that species (1.5 order in H2 and 0.7 order in O2).
Theoretical kinetic models have been developed for practical combustion processes, basically for
combustion of hydrogen and of methane, but the mechanism is very complex, with tens of elementary
species and hundreds of elementary reactions. For hydrocarbons, the two controlling steps seem to be the
formation of molecular hydrogen from hydrogen radicals (e.g. CH4+H=CH3+H2) and the consumption of
molecular oxygen by combination with hydrogen radicals (H+O2=HO2).
Empirical reaction rates for hydrocarbons have been developed based on the simplification that the fuel
(with the oxidiser) first oxidises to CO and H2O, and in a second step CO further oxidises to CO2 (for
lean mixtures), and even for the global complete combustion to CO2 and H2O:
Fuel+aO2 bCO2+cH2O with

d [ Fuel]
dt

a
b
T
= [ Fuel] [ O 2 ] A exp A
T

(53)

Where a, b and c are the stoichiometric coefficients, and a, b, A and TA are empirical coefficients whose
values are given in Table 3.
Table 3. Empirical coefficients in Eq. (53) for the global combustion reaction of some
hydrocarbons (units for A, a and b compatible with concentrations in mol/m3).

Fuel
Methane
Propane
Butane
n-Octane
Methanol

A
83109
27106
23106
15106
10106

a
b
TA
-0.3 1.3 15 000 K
0.1 1.65 15 000 K
0.15 1.6 15 000 K
0.25 1.5 15 000 K
0.25 1.5 15 000 K

But the real importance of the law of mass actions comes when applied to elementary reactions, since
then the reaction order is just the number of molecules participating. For instance, for H+O2OH+O the
reaction speed density is d[OH]/dt=k[H][O2]. For a generic elementary reaction aA+bB=Products, the law
of mass action is then:

dc A dcB
= =
= k (T , p ) c Aa cBb for aA+bBP
V
adt
bdt

(54)

Types of elementary reactions

According to their reaction order, the main type of reactions are first, second and third order reactions.
First order reactions serve to model decomposition reactions A+MB+M when the concentration
of the unspecified molecule M is much larger than the reactive species A. The concentration of the
active species varies exponentially with time, d[A]/dt=k[A], and the rate coefficient can be
computed from two concentration measurements (at two instants). It was Lindemann in 1922 who
first explained the reaction mechanism for uni-molecular decompositions.
Second order reactions serve to model the majority of elementary reactions (formation of activated
complex, isomerisation, double decomposition) since bimolecular reactions are the most common.
Third order reactions serve to model ter-molecular reactions. Ter-molecular collisions, although
not as probable as bimolecular collisions, are much more fertile and thus many third-order
reactions appear in reaction mechanisms. The common instance is the recombination of two active
radicals after collision with a third body that absorbs energy, e.g. H+H+MH2+M (without it,
simple collisions between two active radicals are not productive).
According to their effect, the main types of elementary reactions are:
Chain initiation and chain termination reactions. Initiation steps generate radicals from stable
molecules (but they are very rare), and termination steps consume radicals yielding the end
products (they are also very rare). Examples: H2+MH+H+M is an initiation reaction in
hydrogen decomposition, and OH+H+MH2O+M is a termination reaction in water formation.
Combustion reactions generally start with the generic initiation step RH+O2=R+HO2, where R is a
hydrocarbon radical (e.g. CH4+O2= CH3+HO2).
Chain propagation reactions are the majority in a reaction mechanism. Propagation steps produce
new radicals by consuming other radicals (in the same amount, see below). Example:
H2+OHH2O+H.
Chain branching occurs when one radical yields more than one radical upon collision with a stable
molecule. Example: O+H2OH+H. It is this branching effect that characterises what are known
shortly as chain reactions; reactions that become more and more productive because more and
more active radicals are generated, causing and exponential growth rate termed explosive, only
limited by the total amount of initial reactives. Combustion is the traditional example of chain
reaction (also nuclear fission).
In order to render amenable the solution of the mathematical problem for the reaction mechanism, several
simplifications are introduced, the most common being the steady state, the frozen equilibrium, the partial
equilibrium, and the high activation energy.

The steady state approximation is the assumption that very active radicals form and decompose at
the same (high) rate, with negligible accumulation in the system. This reduces a differential
equation to an algebraic one. For instance, in the Zeldovich mechanism of NO formation
(O+N2NO+N and N+O2NO+O) this approximation implies d[N]/dt=k1[O][N2]-k2[N][O2] =0.
The frozen equilibrium approximation is the assumption that slow rate reactions can be solved
after faster reactions are solved, the former being sensitive only to the integral effect of the
quicker processes. This eliminates a differential equation.
The partial equilibrium approximation is applied when there are two species linked by a single fast
reaction in both senses (forward and backward); then, the concentration of one of them can be
simply put as proportional to the other through the equilibrium constant for that equation (forward
and backward). This reduces a differential equation to an algebraic one.

Arrhenius law
Once the effect of the concentration of reactives has been singled out with the law of mass action, the rate
constant in (32) will depend on temperature, pressure and other side effects, as the presence of catalysts,
or the type of solvent for liquid-phase reactions. The effect of pressure is negligible in condense-phase
reactions, and for gaseous reactions is already taken into account through the concentration of species, at
least in the ideal gas model. But the effect of temperature is very great, with reaction rates doubling just
by increasing temperature a few tens degrees. Well, most chemical reactions accelerate when increasing
temperature, but there are some exceptions, notably in catalytically-controlled biological systems (i.e.
with enzymes), where first the reaction rate increases but then decreases with further temperature rise.
Svante Arrhenius proposed in 1889 a dependence of rate coefficients with temperature, based on the
integration of the theoretical deduction by vant Hoff in 1884 of his famous equation for the variation of
the equilibrium reaction constant, K, with temperature (see e.g. Thermodynamics of Chemical Reactions):

d ln K hr
=
1
R
d
T

(55)

where hr is the standard reaction enthalpy and R the gas constant. Arrhenius proposed:
Ea

k (T ) = Ae RT

(56)

and gave to EA the interpretation of an energy barrier that the molecules have to jump over, to actually
react; thence EA is termed activation energy, and it is much larger than the denominator in (56) for
combustion reactions. Coefficient A in (42) is known as the pre-exponential factor or the collision factor,
since it can be calculated from collision theory, as shown below. Another variable, named activation
temperature, TA, is introduced instead of the activation energy as TAEA/R, with values of order 104 K for
combustion reactions (Table 3).
The Arrhenius exponential temperature dependence, was deduced from vant Hoff equation for
equilibrium, but it can be deduced also from collision theory assuming that the probability of a fertile
chock is proportional to the number of molecules with kinetic energy above a threshold (the activation
energy). In effect, Maxwell-Boltzmann distribution of molecular kinetic energies shows that the fraction
of the total number of molecules, f, with normalised kinetic energy between and +d ( is the kinetic
energy divided by the mean kinetic energy), and its integral from to are (Fig. 14):

Fig. 14. Maxwell-Boltzmann distribution, f, of molecular kinetic energies (relative to the mean kinetic
energy), and its integral from to , F.

fd =

( )

e d F ( ) =
1 erf +
fd =

>>1

2
= e

with

1 2
mv
2
3
kT
2

(57)
where m is the mass of a molecule, v the modulus of its velocity, and k Boltzmanns constant (the gas
constant divided by Avogadros number, k=R/NA). That is, the fraction of molecules with kinetic energies
larger than a given value is proportional to an exponential factor in temperature of the form exp(1/T).
Notice that there is 1/T and not T in the exponential argument in (56), contrary to what might be thought
by saying that reaction rates grow exponentially with T (in fact, the Arrhenius exponential factor tends to
1 for large temperatures, not to infinity). One may recover the reasoning in terms of temperature
variations, T, relative to a given value, T0, and expanding exp(TA/(T0+T)) exp(T). The molecular
kinetic-energy distribution also shows a T-1/2 pre-exponential temperature-dependence (the term), but
it does not corresponds well to experimental data, and an empirical exponent, Ta, is sometimes used
instead.
Collision theory
Collision theory for bimolecular reactions gives a simple explanation to both the mass action law and
Arrhenius law. The argument is as follows. The reaction rate density should be proportional to the
number of collisions per unit volume and unit time, multiplied by the probability that a collision be fertile
(i.e. that produces new species; most collisions are sterile in the sense that only yield thermal
redistribution of energy, i.e. changes in translational, rotational or vibrational energies, without chemical
change, i.e. changes in electronic configuration).
The number of collisions per unit volume, for a molecule A with molecules B (assume there are type A
and type B molecules only), can be estimated roughly as the number of intersecting points (molecules
centres) in a cylindrical volume of length vdt and diameter double than the average molecule diameter
(double because only their centre points were considered before); i.e.:
number of collisions for one molecule N B ( 2 )
v

=
unit volume unit time
4

(58)

where NB/V is the number of B-molecules per unit volume (its molar concentration, cB, times Avogadros
number), and is the averaged diameter of the molecules. When all molecules are considered for the
number of bimolecular collisions, and substituting the characteristic speed, v , according to the MaxellBoltzmann distribution, the final result is:

8k B T 2
number of total collisions

= c A cB
AB
unit volume unit time

(59)

On the other hand, the probability that a collision be fertile must be proportional to the fraction of
molecules with a high-enough speed (the tail of the Maxell-Boltzmann distribution, Fig. 14), and to some
steric factor that accounts for the fact that, for the same high speed, the geometry of the multi-atom
molecules on the collision must be important to fertility (e.g. ceteris paribus, the fertility of the OH+H
collision to yield H2O must depend on the orientation of the linear OH-radical relative to the centre of the
H radical (chocks where the O-atom lies in the middle will be more productive). Considering just the tail
of the Maxell-Boltzmann distribution, it is proportional to the exponential of the chosen lower energy
bound (43), i.e.:

probability that a
EA

exp

RT
collision is fertile

(60)

where EA is the activation energy for the reaction (for the reaction A+BP; not just to species A). In
summary:

dc A dcB
E
T
= =
= c AcB A exp A
V
dt
dt
K
RT
b

(61)

where A is the pre-exponential factor as in Arrhenius law, and K stands for the temperature unit, the
kelvin, to have a non-dimensional exponent b (the K is omitted in most writings). For real work, the three
parameters A, b and EA are experimentally determined because the collision theory do not provide any
values for them, although a more refined activated-complex theory (with bonding energies and steric
factors) could be developed to supply estimations of those parameters in terms of more basic data.
Activation energy data are given in MJ/mol units (usually written as kJ/gmol), or as activation
temperatures, TA=EA/R in kelvin units. Exponent b is tabulated in non-dimensional form, as said above, in
spite of the fact that most of the times T b is written in (61) instead of (T/K)b. Finally, B is commonly
given in CGS-units of (cm3/gmol)n-1/s, where n is the reaction order, instead of the SI-units (m3/mol)n-1/s.
Table 4 gives a sample of experimental values.
Table 4. Empirical values in Eq. (61) for some elementary reactions in the combustion of hydrogen with
oxygen (from Warnatz-1984).
Reaction
A
b
EA
Temperature range
17
3
H+O2OH+O
1.210 (cm /gmol)/s
-0.91 69.1 MJ/mol
300..2500 K
7
3
O+H2OH+H
1.510 (cm /gmol)/s
2.0 31.6 MJ/mol
300..2500 K
H+OH+H2O2H2O 1.41023 (cm3/gmol)2/s
-2.0
0
1000..3000 K
H2O+H2OH+OH+H2O
1.61017 (cm3/gmol)/s
0
478 MJ/mol
2000..5000 K
Relation between rate coefficients and equilibrium constants
Global reactions indicate overall transformations, with a clear distinction between starting reactives and
end products, i.e. there is an assumed direction of advance. Elementary reactions, however, indicate
detailed interactions that, in principle, can be assumed to have no privileged direction of advance, i.e. they
should be at equilibrium (generating products at the same rate as they react back). For a reaction at
equilibrium, the quotient of rate coefficients (forward/backward) must equal the equilibrium constant. In
effect, applying Eq. (52) to both senses of a generic reaction aA+bB cC+dD, one gets:

= k f (T , p )c Aa cBb
c + d a b
k f (T , p ) cCc cDd

V
p
(62)
b
==
K c (T , p ) =
K (T , p )
f =

kb (T , p ) c Aa cBb
RT
b
c d

= kb (T , p )cC cD

Measurement of rate coefficient is so difficult that it is usual to get just one significant figure in accuracy,
so that even far from equilibrium, the relation between rate coefficients at equilibrium and the equilibrium
constant is used to evaluate either the backward or the foreword rate coefficient, based on the
measurement of the other.
Several kinetic mechanisms have been worked out of combustion interest. Some of them are very simple,
as the one step kinetics for ozone formation (O2/O3/O kinetics), other are relatively simple, as the burning
of hydrogen with halogens or the formation of nitrogen oxides, but typical combustion processes are
really entangled; a short-mechanism for CH4/air combustion already involves 14 species: CH4, O2, N2,
CO2, H2O, CO ,H2 , OH, H, O, OH2, CH3, HCHO and CHO. In the case of a lean combustion, the
oxidation pattern seems to be CH4CH3HCHOHCOCOCO2.
To simplify the kinetic mechanism, two basic assumptions are often introduced, the steady state and the
partial equilibrium hypothesis:
Some radicals are assumed to be produced and consumed at a steady rate (not valid at the start and
the end of the process)
Some elementary reactions are assumed to be so quick that they are at equilibrium (forward rate
equal to backward rate) at the time scale considered.

KINETICS OF NOX FORMATION

Chemical kinetics has not yet been able to provide accurate models for practical combustion reactions,
but has proved invaluable to understand some basic associated processes, particularly the formation of
emissions. The best known example is the Zeldovich-kinetics mechanism for NOx formation.
Normal combustion in air produces some 1000..4000 ppm of NO and some 10..40 ppm of NO2. Although
NOx formation may come from a N-containing fuel, from reaction of N2 in air within the flame front, or
from reaction in the hot-products region, for lean adiabatic combustion it is the latter, the thermal
formation after the flame, that is dominant. The extended Zeldovich mechanism is:

dcNO
m3
TA
8
O+N 2 NO+N
1.8 10
, TA =
38000 K (63)
=
cO cN2 A exp
with A =
mol s
dt
T
dcNO
m3
T
NO+O N+O 2
3.8 103
, TA =
21000 K (64)
=
cNO cO A exp A with A =
mol s
dt
T
dcNO
m3
T
NO+H N+OH
1.7 108
, TA =
25000 K (65)
=
cNO cH A exp A with A =
mol s
dt
T
the last reaction being added to the basic Zeldovich mechanism to better model near-stoichiometric
combustion cases. The rate of NO formation is computed from:
dcNO
k
k
k
= k1cO cN2 1 cNO cN k2 cNO cO + 2 cN cN2 k3cNO cH + 3 cN cOH
dt
K1
K2
K3

(66)

with k1, k2 and k3 being the rate coefficients for the three reactions above, K1, K1 and K1 being their
equilibrium constants, and cO, cN2, cN, cO2 and cOH corresponding to the equilibrium exhaust composition
in absence of NO.

CATALYSIS
A catalyst is a substance that changes the velocity of a chemical reaction while not being changed itself
overall, and may be highly selective in their activity, as shown for instance by their effect on ethyl
alcohol, which is stable at ambient temperature; in the presence of an alumina catalyst, ethyl alcohol will
react to form diethyl ether or ethylene, and, if the alumina is replaced by copper, the products become
acetaldehyde and hydrogen. It should be emphasized that a catalyst only speeds up the rate of a reaction
that is thermodynamically allowed, even if its rate is so small as to be negligible. A platinum catalyst will
initiate and accelerate the reaction of hydrogen and oxygen to form water, but no catalyst will facilitate
the reverse reaction, the decomposition of water into hydrogen and oxygen at room temperature. The
latter reaction is not allowed by thermodynamics to run alone.
Catalysts are usually classified as:
Homogeneous catalysts. Of little relevance in combustion. They are often coordination compounds
of transition metals soluble in the liquid medium in which the reaction takes place.
Heterogeneous catalysts. They are finely divided solids (metals, metal oxides, metal sulphides, or
acidic oxides) that catalyze reactions in a fluid media. Metal catalysts are usually transition metals
such as iron, which catalyzes the synthesis of ammonia; platinum, which catalyzes the oxidation of
hydrogen to water and of ammonia to nitric acid; platinum on alumina, which catalyzes the
reforming of petroleum to high-octane gasoline; molybdenum sulphide, which catalyzes removal
of sulfur from crude petroleum; etc. Gasoline is processed by a catalytic reaction carried out over
alumina silicates (zeolites).
Enzymes. They are biological catalysts, complex organic compounds that contain a protein entity,
and are the most active in Nature.
The combustion catalyst per excellence is platinum (Pt); the catalytic property of Pt were discovered by
the German chemist J.W. Dobereiner in 1820 in the hydrogen/oxygen reaction to form water at room
temperature. In 1887 the French pharmacist M. Berger patented the Berger lamp, a 'hygienic lamp and
smoke absorber', that consists on an alcohol lamp (with secret essential-oil odorants), with a wick
connected to a burner made up of ceramic and black platinum, that, when heated up for a couple of
minutes, has the ability to retain a smokeless incandescence at some 500 C (it was much used in
mortuaries and hospitals, and nowadays in aromatherapy).
The three-way catalytic converter
For unburnt emissions, the catalytic after-treatment of the exhaust gas is considerably more effective than
the purely thermal after-burning of the exhaust gases in a thermal reactor. Using a modern catalytic
converter, more than 90% of the amount of toxic substances can be converted to harmless substances.
The so called three-way catalytic converter (TWC, see Combustion Instrumentation), has come into
widespread use for gasoline engines (on all new cars since 1993). The term "three-way" means that all
three toxic substances CO, HC and NOx are eliminated at the same time, although there are really only
two different processes: the oxidation of the unburnt emissions (CO and HC), and the reduction of the
nitrogen oxides. The converter shell contains a ceramic "honeycomb" which is coated with a noble-metal
combination (platinum, palladium and rhodium), although some other catalysts are being tried, as
perovskite (a greyish-black mineral form of calcium titanate, with some rare-earth elements, which is
used in certain high-temperature ceramic superconductors). Only lead-free gasoline may be used with
noble-metal converters because the lead otherwise destroys the catalytic properties of the metals. The

catalytic converter also works on natural-gas burning engines, but not for sewage and landll biogases
because sulfur and heavy metals content rapidly deactivate the catalyst.
The three-way catalytic converter only reduce NOx in reach mixtures and only oxidise CO and HC in lean
mixtures, with a very thin overlapping region near the stoichiometric air-fuel ratio (a deviation of only 1
% has considerable adverse effects on either oxidation or reduction efficiency). That is why they are only
useful for gasoline engines, and only if running stoichiometric. Moreover, the best open-loop enginecontrol would be unable to maintain the air-fuel mixture within the close tolerances required for optimum
work of the three-way catalytic converter, and so an extremely accurate closed-loop electronic control
(featuring almost zero lag), with an oxygen sensor at the exhaust (the lambda probe), must be added to the
air-fuel-mixture management system.
Selective catalytic reduction (SCR)
As the successful three-way catalytic converter only works close to stoichiometric mixing, other catalysts
are being tried for lean-mixture gasoline and all diesel engines (always lean), specifically designed to get
rid of the NOx since there is little CO and HC in lean-combustion engines.
For lean-mixture gasoline engines, CO and HC are oxidised with a platinum catalysts in a first-order
reaction mechanism since only the CO or HC concentration is controlling, oxygen concentration being
much higher in lean combustion products (e.g. for CO+(1/2)O2CO2, / V = k (T ) xCO ).
For diesel engines, best results (>90% NOx reduction) are presently obtained by spraying urea in the flue
gases, to generate ammonia by thermo-hydrolysis (NH2)2CO+heat=HNCO+NH3 followed by hydrolysis
of the isocyanoic acid HNCO+H2O+heat=NH3+CO2, in a selective catalytic reduction (SCR) process
through a platinum-coated titanium-oxide matrix. The urea aqueous solution (40%wt) is carried in a
reservoir and fed by a dosing pump (the amount of urea depends on engine load) to a compressed-air
spraying system. The urea SCR only works at T>250 C, and it is found that near-equimolar NO/NO2
mixtures proceed much faster than pure NO or NO2. In the first case the reaction is directly
NO+NO2+2NH3=2N2+3H2O, whereas in the second case additional steps involving ammonium nitrate
take place.
If the fuel contains sulfur (as some heavy fuel oils), there is the problem of condensation of (NH4)2SO4
downstream (that is why the SCR is placed after the exhaust turbine). Ethanol and other hydrocarbons
have also been proved efficient for SCR, as well as other selective non-catalytic reducers (SNCR).
Catalytic combustion
The usual thermal free-flame combustion process cannot be sustained in the vicinity (say within 1 mm) of
normal cold solids (metals, ceramics or polymers), and thus it cannot propagate through normal cold
porous media. But flames can propagate through hot porous solids and through cold porous solidcatalysts.
The catalytic substance helps to maintain the combustion process at low temperature at the surface, as
when a mixture of H2/air reacts over Pt at room temperature (CH4/air over Pt requires 350 C), that in its
absence would require much higher temperatures: the minimum autoignition temperature for
homogeneous H2/air combustion is 850 K
Catalytic combustion works for premixed flows and for a non-premixed fuel flow in ambient air, with
steady temperatures in the catalytic matrix from 700 K to 1400 K. The advantages of power modulation,
widening of ignition limits and less emissions, are similar to high-temperature porous combustion, or
even better (<1 ppm for NOx and CO, due to very low temperatures involved), but not the higher burning

rate (that was due to thermal conduction along the hot solid matrix, and thus only 600 kW/m2 are
achieved instead of the 3000 kW/m2 of thermal porous-medium burners. At intermediate temperatures, an
hybrid regime of catalytic-assisted thermal combustion may be developed, where both heterogeneous and
homogeneous reactions take place, an interesting trade-off solution when the catalyst is too expensive and
very little can be used (in full catalytic mode, if there is insufficient catalyst, part of the fuel slips to the
exhaust (increasing pollution and expense), or even the whole fuel if the catalyst cools down and
deactivates. The influence of the flowrates is important; for instance, if a thin porous solid is doped with a
catalyst and a premixed methane/air stream is forced through, an exothermic oxidation of the fuel takes
place if the temperature is >600 K, releasing much of the lower heating value, what causes the matrix to
reach some 1500 K at the outer surface (the hotter) in a flameless regime with a power of up to 500
kW/m2; but if more gases are fed, small flames detach from the outer surface and, at about 1000 kW/m2, a
blue flame with some 1900 K forms immediately close to the outer surface, that only reaches now some
600 K.
Back to Combustion

COMBUSTION INSTRUMENTATION
Experimentation and instrumentation ....................................................................................................... 2
Experiment planning............................................................................................................................. 2
Experiment conditioning ...................................................................................................................... 3
Experiment stimulation and control...................................................................................................... 3
Experiment diagnostics ......................................................................................................................... 3
Experiment analysis .............................................................................................................................. 4
Experimental techniques ........................................................................................................................... 4
Controls and safety ............................................................................................................................... 6
Fuel and air supply............................................................................................................................ 6
Flame detectors ................................................................................................................................. 7
Emission detectors ............................................................................................................................ 7
Flame diagnostics ............................................................................................................................. 8
Exhaust control ................................................................................................................................. 8
Flow measurement ................................................................................................................................ 9
Hot-wire anemometry (HWA) .......................................................................................................... 9
Particle image velocimetry (PIV) ..................................................................................................... 9
Speckle velocimetry (SV) ................................................................................................................. 9
Laser doppler velocimetry (LDV) .................................................................................................. 10
Temperature measurement.................................................................................................................. 10
Analytical techniques.......................................................................................................................... 10
Gas chromatography ....................................................................................................................... 10
Mass spectrometry .......................................................................................................................... 11
(Radiation) Spectrometry................................................................................................................ 11
Optical diagnostic techniques ................................................................................................................. 11
Applications ........................................................................................................................................ 11
Fundamentals ...................................................................................................................................... 12
Classification ...................................................................................................................................... 13
Object types .................................................................................................................................... 13
Radiation sources ............................................................................................................................ 13
Detector types ................................................................................................................................. 13
Radiation emission and radiation absorption .................................................................................. 13
Radiation scattering ........................................................................................................................ 14
Holography ..................................................................................................................................... 15
Tomography.................................................................................................................................... 15
Interferometry ................................................................................................................................. 15
Speckle interferometry (SI) ............................................................................................................ 15
Moire deflectometry (MD) ............................................................................................................. 16
Raman scattering (RS) .................................................................................................................... 16
Coherent anti-Stokes Raman Scattering (CARS) ........................................................................... 16
Laser induced fluorescence (LIF) ................................................................................................... 16
Laser induced incandescence (LII) ................................................................................................. 16

EXPERIMENTATION AND INSTRUMENTATION

The aim here is to analyse the devices (and their implementation) used to measure and control
combustion process, i.e. fuel supply and control, air supply and control, ignition, flame detection, exhaust,
emissions detection, overall diagnosis, etc., for proper operation or, in the more demanding case, for
investigation of special details, as the internal structure of a flame (including radicals evolution).
Depending on the application of the combustion process, e.g. to heat engines, additional variables and
instruments may be of interest (e.g. engine tachometers and dynamometers).
Combustion always involves complex thermo-chemical and fluid-mechanical interactions, and thus, the
instrumentation to condition, control and diagnose all the variables of interest may become rather
complex. Instrumentation for research on combustion is much more sophisticated than instrumentation for
proper operation of a combustor, but they are not unconnected; on one hand, instruments for routine
operation were in the past research instruments, and, on the other hand, all experimental research make
use of common operation instruments (e.g. flow controls), besides special sophisticated instruments (e.g.
laser diagnostics). Aiming at research instrumentation, both aspects are covered.
Experiments are purposely tried events, to check some preconceptions. Instrumentation is the actual set of
devices used and their way of implementation to carry out the experiment (experimentation).
Experimental techniques for combustion (and in general) may be grouped according to its function of
conditioning, stimulation and diagnosing the event. We envelop these techniques, exposed in
chronological order of execution, with a previous planning phase and a final analysis phase.

EXPERIMENT PLANNING
Design of experiments (DOE) is a difficult creative endeavour, usually involving sophisticated tools for
their implementation. Clearly defined experiment objectives (known target), a mathematical model to use
for making predictions (checkable expectations), development of an ordered set of nominal trials to
perform (known procedures), a feasibility study to find the weak points and choose the most reliable tools
for actual implementation, redundancy in measurement, and planning for eventual malfunctions
(experimental robustness), are a guarantee to success, but the key rule may be that, the simpler the
experiment, the better (and try it at least twice!).
Experiments must be designed robust, i.e. with a planning insensible to uncontrolled environmental
perturbations (Taguchis experiment-quality approach). One aspect of robustness is redundancy;
validation of the results is greatly enhanced by the accumulation of data that demonstrate with a high
degree of confidence that the process is repetitive (without the need of lots of trials).
Instrumentation heavily depends on the objective of the test. According to its main objective,
instrumentation may be aimed at:
Normal operation of a working system (reliability and economy are the drivers)
Experimental research in a test-rig or prototype system (accuracy is the driver).
Instrumentation equipment serves the following purposes:
Controls, of the initial state and of the boundary conditions (sometimes depending on
measurements), that may be split in:
o Conditioning, i.e. generic controls, usually permanent and passive (not requiring power) to
establish a known configuration: structure, reservoirs, piping, test section, power supply,
heating, cooling, and other interfaces.
o Stimulation (or actuators), i.e. specific controls, eventful and active (requiring power or a
trigger) to force some stimuli or avoid it.

Diagnostics, i.e. sensors to measure the evolution of selected variables, including the initial state
and boundary conditions. Real time diagnosis is required for real-time control of the experiment,
but it is most desirable for analysis too (sending samples to a distant laboratory may still be
necessary for some special biological or chemical analysis, but fluid physics has always used flow
visualization for immediate diagnosis).

Combustion experiments depend on so many intermingled parameters that it has not been possible to
carry out scale-model experiments, as for other aerodynamic and hydrodynamic problems, and there is a
big jump from academic setups to industrial prototypes.

EXPERIMENT CONDITIONING
Besides instrumentation for diagnostics and control, combustion equipment always includes the basic
infrastructure to set up the process, i.e. to condition and control the configuration: combustor body (its
structure and supports), feeding ducts and exhaust, access doors and viewing windows, thermal
insulation, safety devices, etc.
Experiment conditioning is so application-dependent, that no further details are presented here.

EXPERIMENT STIMULATION AND CONTROL

The key stimuli in combustion is the ignition event (usually an electric spark), but, as for typical thermofluid-mechanical experiments, there are other controls for the fluid supply (e.g. fuel pump, air blower),
fluid flow-rate (valves), thermal state (heaters, coolers), etc.
Every stimulation requires some sensors to timeline, regulate or control its action, so that sensors should
be studied prior to actuators (and the number and variety of sensors in a test-rig is much greater than that
of actuators).
We do not extend here on combustion stimuli but on diagnostics, not without recalling that any sensing
involves some stimuli on the test-subject (there is no measuring without perturbation).

EXPERIMENT DIAGNOSTICS
Diagnostics means the taking of measurements and samples with the aim at identifying how the system
behaves in space and time. Our main aim in this survey of combustion instrumentation and experimental
techniques is really here, on measurement techniques.
Diagnostics instrumentation in combustion is based on three main types of variables: flow (velocities),
thermal (temperatures) and chemical (species). Most sensors (and actuators) act by direct contact with the
object, but non-contact coupling (electromagnetic) is increasingly being used for sensing (and sometimes
for actuators).
Besides the transducer itself, both, sensors and actuators, usually require some power and signal
conditioning (digital or analogue), nowadays interfacing to a digital computer for automated operation
(data acquisition and control, DAQ; digital signal processing, DSP; remote operation; etc.).
When research on combustion is implied, and not merely combustion control (with its inherent
diagnostics), very sophisticated techniques are usually applied, most of the time optical (or better
radiometric) in nature, since the high temperatures (>1000 K), small scales (<1 mm), and large gradients
(106 K/m), make intrusive techniques only applicable to intake and exhaust control. The most demanding
task is measuring active species within a flame (radicals and ions). Advanced diagnostics rely on
spectrometric methods using laser instrumentation and advanced computer modelling codes.

EXPERIMENT ANALYSIS
The analysis of an experiment depends a lot on its purpose, it is very different to check that a certain
magnitude is within allowable bounds (e.g. to check the concentration of CO in a vehicle exhaust), that to
investigate the formation of soot.
Redundancy is so important to guarantee data validation, that a big problem in data analysis is the
reduction of the overwhelming amount of data taken; statistical analysis helps a lot, but the most
important is to have a parametric model to fit the data to, i.e. data fitting to models and not just to generic
curves and surfaces.
However, as we said before, we deal here only with instrumentation, and leave actual data acquisition and
analysis aside.

EXPERIMENTAL TECHNIQUES
Experimental techniques may be classified as above in configuration setups, stimuli (actuators: flow
inducers, valves, spark plugs), and diagnostics (sensors: flow, temperature, concentration); it should be
recalled that besides the sensor or actuator itself (the transducer), some power supply and conditioning
circuit is usually involved.
For combustion instrumentation, experimental techniques could also be classified according to function
as: fuel-related, air-related, fuel/air-related, safety-related, emissions-related, and flame-structure-related
(only for combustion research), or according to the setup: operating installation, standard test-rig,
specially built research set-up. For combustion experimentation, special burners have been devised,
starting from the now-standard axisymmetric Bunsen premixed burner in 1855, Wolfhard flat-flame
premixed burner in 1939, Wolfhard-Parker twin-slot diffusion burner in 1949, Pandya-Weimberg
opposite-jets diffusion burner in 1963, etc.
The traditional classification of experimental techniques is, however, according to the type of variable
being measured or controlled: temperature measurement, flow rate or velocity, pressure, composition,
etc., sometimes grouped in physical and chemical techniques. But nowadays, and particularly in
combustion research, the same experimental technique is used to measure several magnitudes at once, as
when Rmn scattering is used to measure composition, concentration and temperature. Here we follow
the traditional approach, but a separate presentation of general non-intrusive diagnostics is included
below.
Time measurement
Many experiments on combustion deal basically with steady states, but timing is important in any case
(all steady states start and end). Timing in combustion is not as critical as in other fields, except when
analysing the initial phase of the ignition process.
Geometry, level, edges and particle measurement
Fixed geometry is measured before or after the experiment. Low varying geometries as liquid level, solidfuel borders, actuator position (for stimuli or for sensor location), and so on, are measured with
potentiometers, capacitive sensors, ultrasounds, etc. More subtle edges, as flame fronts, smoke plumes, or
the multiplicity of particle sizes (fuel sprays, tracers, soot) usually require optical techniques.
Thermometry
The contact thermometers most used in combustion are thermocouples, thermistors (negative temperature
coefficient resistors, NTC) and metallic resistance thermometer devices (RTD), in that order. More

advanced non-contact thermometers (sometimes named pyrometers) are dealt with below under Optical
techniques.
Velocimetry and flow rating
Flow meters may be global, to know the overall mass flow rate or volume flow rate, or local to know the
velocity field. In the first case, tank weighting or level change in condense fuels, or general flow meters
as rotameters, calibrated nozzles and diaphragms, turbine-meters and other more sophisticated methods as
thermal capacity and Coriolis effect devices, may be used.
Measuring the velocity field is much more cumbersome at least because of the amount of data implied,
and sophisticated methods are used, as explained below.
Piezometry
Pressure measurement (piezometry) is not a difficult problem in combustion except in reciprocating
engines, where rapid changes and large pressures are involved, and there piezoelectric sensors are used
(quartz-crystal transducers develop an electrical charge when compressed; they are bored into the head of
the cylinder or adapted within a modified spark-plug). Piezoresistive transducers are semiconductors
(doped silicon) that change their electrical resistence upon compression, and are used to follow the highfrequency pulsation at the intake and exhaust ports in reciprocating engines. For small pressure
differences the best are silicon-chip capacitance sensors.
Chemical analysis
Different stages in a combustion processes may demand chemical analysis, i.e. qualitative and
quantitative finding of the composition in a mixture: intake (fuel and air analysis), inside (radicals formed
for kinetic studies), and exhaust (emissions). The main components in combustion processes are: N2, O2,
fuel, CO2 and H2O; trace components are: CO, NOx, OH, O, H, CHO, OH2, e-, OH-, CHO+, etc. In
general, analytical techniques may be grouped as:
Chemical methods of analysis: characteristic reactions, selective absorption, electrochemical
techniques, etc.
Physical methods of chemical analysis, ranging from electrical, thermal or optical, to the most
sophisticated spectroscopic techniques.
Many times, a sample of the mixture is analysed off-line and discarded, often through a separation
process of chromatography, but most advanced analytical techniques are non-intrusive and on-line.
Chemical sampling may be intrusive, through a quartz tube of some 1 mm or less, followed by chemical
analysis (using Orsat selective absorbers, or gas chromatography, or mass spectrometry, or selective
solid-electrolyte conductometry), or non-intrusive sampling (radiometric) that may be passive or active,
classical or quantum. Sometimes a special chemical type, the electrochemical one, is singled out, since it
is widely used to directly measure concentrations of many different gases (CO, NO2, SO2, H2S, HCl, but
only in the ppm-range). Before gas chromatography took over in mid XX c., and later radiometry and
mass spectrometry, the standard method in exhaust analysis was developed by Orsat in late XIX c. After
filtering solid particles and dehumidifying the exhaust sample, the amount of CO2 was first measured by
volume subtraction after passing the gas through a NaOH solution; afterwards, non-saturated
hydrocarbons were removed by a KOH and pyrogalic-acid solution, oxygen was removed by a
NH4Cl/CuCl solution, and CO by H2SO2.
Note that composition ranges may be different not only for different species, main or trace, but for
different function; e.g. for CO-toxicity the maximum human exposure is 50 ppm, whereas for COcombustion the minimum concentration for ignition is 12.5% (the instruments are different).

Experimental techniques are described below; first those related to the overall control and safety of the
process, and later on the more sophisticated diagnostic techniques.

CONTROLS AND SAFETY

The general goal of combustion instrumentation is to procure a safe, energy-efficient and emission-free,
combustion process for the intended use: operation or research. To that purpose, electrical, pneumatic and
hydraulic actuators, automatically or manually operated, are used to control the process.
FUEL AND AIR SUPPLY
Fuel supply
Most fluid fuels are already available at a supply pressure (either bottled or piped). If not, some pumping
should be implemented, as in vehicle engines and coal-fires burners. The usual fuel flowrate control is a
solenoid valve (a needle electrovalve). Heating power control is based on fuel-supply control, either
on/off or modular. Related to fuel presence and the possibility of uncontrolled combustion is the
'explosimeter', a fuel-gas sensor, usually based on the rapid oxidation of the fuel at room temperature in
the surface of a catalyst (a Pt-wire that gets hot and changes its electrical resistance in the presence of a
reactive atmosphere).
Commercial liquid fuels require pumping and filtration, and sometimes also heating systems; solid fuels
usually require more handling and preparation; neither of those additional systems are dealt with here,
Air supply
The air supplier may be a variable speed fan (the speed is varied by changing the voltage, the frequency
or the wave profile), the engine suction itself, or a natural draught induced by the fuel supply.
Manometers (diaphragm, piezoelectric) may be part of the diagnostics. Air flow metering with integrated
temperature sensor is fed to the electronic control unit (ECU), if any. Air must be supplied not only for
combustion, but for purging purposes to bring the system to known safe conditions.
Air/fuel ratio
Good control of air/fuel ratio is important in premixed flames for efficiency and polution-avoidance, and
has become critical for operation of exhaust catalysts. A simple fix regulation may produce unwanted
air/fuel ratios due to fluctuations in fuel and/or air supply (pressure or composition), or changes in load
condition (e.g. secondary air in a burner does not follows fuel flow rate); for that reason, an O2-detector in
the exhaust is used to control the air/fuel ratio (an oxygen sensor is chosen because xO2 monotonically
grows. with A-A0, quasi-linearly (xO2=0..0.1 from =1..2), whereas for instance xCO2 decreases
parabolically with A-A0, and on both sides! of =1, besides the sensor being more expensive).
Several O2-detectors have been developed since the old -probe (Saab/Bosh-1977) that revolutionised
electronic ignition and injection control in Otto engines, where stoichiometry is now maintained to
=1.000.01 (domestic water-heaters work with 10..50% excess air). They are electrochemical cells
yielding a voltage depending on the difference in oxygen concentration (O2- really) between the exhaust
and the ambient air (highly non-linear emf, as seen in Fig. 1), through the electrolyte (a ceramic sheet of
ZrO2). The electrodes are gas-permeable platinum layers. They only work when hot (>300 C, that is why
they were placed at the exhaust manifold); besides, the output depends on the operating temperature.
Since 1990 all -probes (in the front and at the rear of the catalyser) are heated to work also when idle
and at part throttle. Other resistive semiconductor probes have been tried without too much success (TiO2,
SnO2).

Fig. 1. Functional details of a lambda probe sensor.

FLAME DETECTORS
Different types of fire alarms and flame detectors exist; some are good for close-proximity detection and
others for overall surveillance (indoors or outdoors):
Thermal. A probe that changes with temperature (bends a bimetallic strip, bends a burdom-type
vapour-pressure phial, etc.). They are pasive devices (no need of power), simple, and fail-safe.
Just for safety, a wire that melts may be used to break a contact.
Thermoelectric. An emf is generated that may power a solenoid to keep both, the main and the
pilot fuel supply (a manual start is needed, usually the user holding a push-button while the
thermocouple gets hot). It was the standard for small appliances as home water heaters.
Ionisation. Measures the change in electrical conductivity of air through a flame (a flame is a
plasma with some 1 ppm charged-particles). It is very quick and can be automated, being the
method presently used in home appliances, in spite of the fact that it is not passive (it requires
power).
Infra-red emission (IR with CO2=2.7 m or better CO2=4.1..4.6 m). Solar and lamp radiation
may cause false alarms. It is better to use several wavelengths to discard solar radiation
reflections. All radiometric methods are expensive, so they are only used in large equipment.
Ultra-violet emission (UV with =0.2..0.3 m). Lightning and arc-welding may cause false
alarms. A combination of UV/IR sensors is better.
Chemiluminescence. For known flame types, a photomultiplier tube with an spectral filter may
sense characteristic radiation emissions.
EMISSION DETECTORS
Gas leak detectors and explosimeters. Usually based on the electrical resistance variation of a
platinum wire, due to the temperature increase caused by catalytic oxidation of the fuel-air
mixture. Portable system with field replaceable measuring cells are in the market capable of
sensing minute concentrations of natural gas, butane and propane, either along pipes and
combustor (gas leak), or in the ambient (explosimeters). Sometimes, to avoid explosions, it is not
enough to have a quantity less than the LEL in a closed room, since very light fuels like H2 or very
heavy gas fuels as diethyl ether (C4H10O) will stratify a lot.
Flue gas emission analysers. Usually based on selective radiation emission or absorption, or on
electrochemical cells. Oxygen in the exhaust is measured to know the air/fuel ratio used, and the
most used method is the zirconium-oxide cell explained above. Other O2 and NOx sensors are
based on wet electrochemical cells, consisting of coated electrodes (sensing, reference, and
sometimes a counter electrode too) and a small volume of an acid or base solution (the

electrolyte); gases diffuse through orifices on the sensing face to the porous sensing electrode,
reaching the electrolyte, and generating a very small electrical current proportional to gas
concentration; their response time is low, and they have a consumable counter electrode. Portable
system with field replaceable measuring cells are in the market capable of measuring at once flue
velocity (0..50 m/s), H2O (0..30%), O2 (0..25%), CO (0..10000 ppm), NO (0..1000 ppm), NO2
(0..1000 ppm), SO2 (0..1000 ppm), differential temperature (0..1000 K) and CO2 (0..25%). Excess
air and energy efficiency can be easily computed from those measurements.
Smoke and particulate analysers. Usually based on light transmission, scattering or reflection, or
by -radiation absorption, or by the tribo-electric measuring principle (the tribological probe
measures the charge on the particles that strike a metallic rod, which depends on the flow velocity
and the concentration of the dust in the flue gas).

Detector must be calibrated from time to time (according to required standards), using certified
concentrations of test gases.
FLAME DIAGNOSTICS
The most conspicuous feature of combustion to analyse is the visible flame. Most flames flicker, and it is
difficult to have a steady flame to look at: it must be protected from minute air-drafts and fuel-supply
perturbations.
Flames tend to get anchored to solid edges because there are ample ranges in temperature and velocities
nearby, where stabilisation may take place. When flames cannot stabilise on the rim of the burner, flame
holders (flow gutters in I-, V- or H-shape) are placed downstream of the fuel injectors, as in gas turbine
combustion chambers where flames are stabilised in streams up to 100 m/s).
Although contact heat devices may be used for flame detection and analysis as said above, optical (nonintrusive) techniques are preferred, either based on classical optical effects (ray geometry and photometry,
as intensity absorption, particles, edges, shadowgraphy, interferometry, moir, speckle, polarisation, etc.),
or based on quantum effects (spectral intensity analysis: the frequency identifies the species, and the
intensity gives their concentration). A generic overview of optical diagnostic techniques is given below,
after flow and chemical instrumentation is presented.
Laminar premixed flame experiments are the standard way to get validation data for combustion
chemistry models (e.g. for pollutants and soot formation).
EXHAUST CONTROL
Flow rates are controlled at the intake by virtue of fuel and oxidiser supply systems; the aim here is how
to provide a safe exhaust for the operation of combustors. The usual procedure is to get rid of the exhaust
gases through a chimney to the atmosphere, far enough to minimise nuisance and danger (overheating,
asphyxia, intoxication, deflagration). Sometimes the exhaust is cleaned to minimise emissions, but we
focus here on fuel releases.
Normal practice to cope with unavoidable fuel releases through the exhaust, due to lack of appropriate
ignition or unexpected extinction, is to force fresh air for a while to dilute the mixture below its ignition
limit, before any other trial is performed.
A better procedure in experimental setups is to pass the exhaust through a pilot flame to guarantee that
explosive mixtures do not build up.

FLOW MEASUREMENT
Flow measurement is used for accounting and for control. Basic fluid-mechanical instrumentation, as
used in other applications to measure liquid tank level, flow velocity fields, turbulence level, pressure,
particle size and distribution, etc., are also used in combustion instrumentation: positive-displacement
counters (as the domestic gas meter), differential-pressure meters (diaphragm and venturi meters), turbine
flow-meters, hot-wire anemometers (HWA), laser doppler velocimetry (LDV), particle image velocimetry
(PIV), ultrasonic velocimetry, etc. The most precise flow-rate sensors are the heater-based for gases
(thermal capacity) and the Coriolis-based for liquids. Advanced optical techniques are covered below.
As said before, there is an emphasis on particle characterisation in combustion diagnostics, both because
of the importance of soot formation and particulate emissions, and because of the importance of spray
combustion, what is related to flow measurement (e.g. PIV may be used to measure single-fluid flow by
adding tracer particles, or to measure particle velocities in proper two-phase flows: fuel sprays and
exhaust soot). Modern high-speed digital cameras are used nowadays to better analyse two-phase flows.
HOT-WIRE ANEMOMETRY (HWA)
In this technique, a very small metal wire is heated and the power dissipated and temperature reached are
measured (the electrical resistance depends on temperature); as the amount of cooling depends on the
convective velocity of the fluid in which it is immersed, proper calibration provides a measure of fluid
velocity in terms of power supply for a constant temperature difference, or in terms of temperature
difference for a constant power supply.
The wire is made of platinum or wolfram, and is very delicate, with diameter in the range 0.5..5 m and
lengths from 0.1..2 mm, making HWA only usable for relatively clean gases like ambient air (that is why
they are usually named anemometers). Because of the geometry, a single hot wire only yields one
component of the velocity field, so that multi-sensor probes are used for three-dimensional fields. Notice
that by just measuring the electrical resistance, hot-wires are also thermometers.
PARTICLE IMAGE VELOCIMETRY (PIV)
In this technique the fluid is seeded with small particles (of order of 10 m or larger), of the same density
as the fluid, in an amount such that there are as many as possible without overlapping in the image.
In two-dimensional PIV, a thin powerful sheet of light is shed on the object (preferably by a pulsed laser,
to avoid heating by absorption), and the light scattered by the individual particles is focused in an image
plane, where their positions are tracked to compute velocity vectors from consecutive images.
Stereographic and photogrametric and holographic techniques are being used to provide direct threedimensional measurements.
This technique best applies to quasi-steady, quasi-planar flows. Refractive index gradients could distort
the image. Collimated light may be used to avoid parallax distortions.
SPECKLE VELOCIMETRY (SV)
Similar to PIV but with many more and smaller particles (of order of 1 m) whose individual images
overlap, forming speckles (a small granulation) without any meaning to direct observation; but if the
speckle pattern is illuminated with a coherent light beam, Young interference fringes appear, proportional
to the average particle separation, that, if subtracted from a reference speckle image, gives the apparent
velocity field.

LASER DOPPLER VELOCIMETRY (LDV)

In this technique, the time tracers take to cross along consecutive fringes formed at the intersection of two
nearly-parallel beams from a laser, is measured in a scattered-light detector. As this simple setup only
yields the speed modulus in one direction (along the in-plane counter-bisect), timing shifts and multiple
laser beam-pairs are used to measure the whole velocity vector at a point.
In spite of its fast and precise response and non-intrusive character (most of the times there is no need to
add the tracers, as common fluids as air and water always carry fine particles in suspension), LDV has the
strong handicap of being a one-point sampling technique (i.e. zero-dimensional, when most velocity
fields are three-dimensional).

TEMPERATURE MEASUREMENT
Combustion thermometry usually focuses onto the gas phase; measurement of surface temperatures in
conducts and combustor walls is also important, but not so demanding (except perhaps at moving surfaces
as gas-turbine rotor blades). Measuring high temperatures in a gas is a difficult subject since thermal
conductance from gas to probe is poor, heat conduction through the lead wires and the metallic sheath is
important, and radiative coupling between probe and walls is high. Moreover, in thin, fluctuating,
reacting, hot regions (as within a flame), there is response-time problem, probe micro-vibrations
problems, materials-resistance problems, and basic non-equilibrium ill-defined-temperature problems.
For high spatial and temporal resolutions, fine-gauge thermocouples are the best; platinum-resistance
probes (Pt-100) are too large, and thermistors (NTC) have a small temperature range. Accuracy and
response time are key issues. As for any contact thermometer, one assumes that thermal equilibrium of
the sensor and the object is established, but with low-conducting gases at high temperatures, losses
through the probe support and to the walls are large and only a steady state is reached. The response time,
of the order of 1 s for 1 mm size thermocouples, may be lowered down to 10-2 s for the smallest 10 m
wide thermocouples (at the expense of durability and accuracy). For high temperature measurement in
combustion, thermocouples must be sheath with SiO2 to avoid metal oxidation and catalytic effects.
But intrusive thermometers as the thermocouple cannot provide whole field temperatures in fluctuating
flows, as in reciprocating internal combustion engines; knowledge of the spatial distribution of gas
temperature prior to ignition is needed for modelling engine combustion with stratified load and/or
exhaust gas recirculation, since large temperature inhomogeneities are involved.
Non-intrusive radiometric temperature measurement is preferred for advanced multidimensional analysis,
the physical probes been used for calibration. Thermometry based on gas density measured by the
refractive index is easy, but resolution is low at high temperatures. Infrared thermometry based on CO2
and H2O bands, an ultraviolet radiometry at the 0.309 m band due to OH are also used.

ANALYTICAL TECHNIQUES
Analytical techniques refer here to chemical-composition measurement. A short description of some of
the traditional analytical techniques used in combustion follows, with the more advanced optical
diagnostics being covered under the Optical Diagnostic Techniques heading, below. Sometimes the main
products in the combustion of organic matter are separated before further exhaust analysis; water vapour
is traditionally absorbed and weighted in phosphorus-oxide, P4O10(s)+6H2O(v,l)=4H3PO4(s), and carbon
dioxide in sodium hydroxide, NaOH(s)+CO2(g)=NaHCO3(s).
GAS CHROMATOGRAPHY
Chromatographic techniques (gas or liquid) are based, as for other means of mixture separation, in the
natural segregation of species between two immiscible phases. In gas chromatography, a small sample is

diluted in an inert gas carrier (He, H2, Ar, N2...) and forced to flow through a porous media or a stationary
liquid, what introduces a selective speed-lag (dependent on size and affinity of species to plug material)
that allows selective elution for collection or discarding, i.e. a fractioning, as when a drop of mixed dyes
spread over a tissue (the origin of chromatography).
After separation, traditional chemical analysis may be performed, but, with proper calibration with known
samples at standard temperature and pressure, a catalogue of retention index can be prepared, and
qualitative (just by a look-at table) and quantitative (e.g. by light absorption) measures can be obtained.
Taking physical samples in combustion processes is not simple because the probe may induce catalytic
reactions at those high temperatures. The probe is typically a narrow SiO2 tube (down to 0.1 mm in
diameter have been achieved)
MASS SPECTROMETRY
In mass spectrometry, a small sample is put into vacuum and bombarded with an electron beam to
produce a stream of charged fragments in different proportions and mass-to-charge ratios (ion source
generation), that, when subjected to a magnetic or electric field, produces a proportional deviation, i.e. a
fractioning of different mass-to-charge ratios, with the intensity at the target being proportional to
concentration.
(RADIATION) SPECTROMETRY
Note. Spectrometry is understood to refer only to electromagnetic spectrometry, i.e. spectro-photometry
or spectro-radiometry; no mass spectrometry. Sometimes, spectrometry is also used as a generic name for
all quantum-optical techniques.
Three types of radiation effects are used in spectrometry: emission, absorption, and scattering. Either the
wavelength, , or the wavenumber k=1/, or the frequency, =c/ (with c the speed of light), is measured
(only the frequency is independent of the refractive index). Details are covered below.
Although high-resolution microwave and infrared spectroscopy are applied to optical diagnostics in
combustion, visual laser spectroscopy is the most common advanced technique for flame analysis: Raman
spectroscopy, Coherent Anti Raman Spectrometry (CARS), Laser Induced Fluorescence (LIF).

OPTICAL DIAGNOSTIC TECHNIQUES

APPLICATIONS
Optical diagnostic techniques are commonly used in all kinds of fluid diagnostics, not only in
combustion, but they are specially critical to flame diagnostics because of the hardship of the
environment and small size.
Optical diagnostic techniques may serve to study:
Geometry: fluid boundaries, moving objects, moving fluid-interface fronts, etc.
Particle size and distribution: nephelometry, Rayleigh scattering.
Velocimetry: particle image velocimetry (PIV), laser doppler velocimetry.
Thermometry: disappearing-filament pyrometry, sodium-line reversal.
Concentrations: refractometry (schlieren, moir).
The fact that most novel diagnostic techniques, both for physical analysis and for chemical analysis, are
based on electromagnetic radiations, is because of the following advantages of optical systems:
They are non-intrusive

Whole field sensing possibility

Immediate qualitative diagnosis with possibility of quantitative analysis
Digital video processing automation possibility

Optical diagnostic techniques naturally started by photographic recording of what the observers eye was
watching; it first developed along classical optics, but most applications nowadays are based on quantum
optical effects (spectrometry).
Optical techniques have some disadvantages, however. First of all, the radiation coming from the object
may go through a complex path (windows, intermediate fluids, environmental air, and so on), with optical
properties not well known or controllable. Second, the optical setups used to be most delicate and
expensive, but introduction of fibre-optics and afocal optics has alleviated some of these problems. Third,
optical techniques need to be calibrated in a similar configuration (a laminar premixed flame is usually
used as a standard).

FUNDAMENTALS
Electromagnetic radiation can best be understood at the microscopic level in the particle-sense (from the
wave/particle dualism), as a photon gas; a photon has zero mass at rest and an energy proportional to its
frequency, E=h. The interaction of electromagnetic radiation with a material substance may be small,
giving rise to thermal interactions (energy redistribution without molecular breakings; only translational,
rotational, vibrational and electronic energy level rearrangement), or the interaction may be stronger
(giving rise to loss of electrons in the molecule, i.e. ionisation and chemical reactions, or even causing
nuclear reactions).
Energy jumps in the interaction matter-radiation are widely separated in the spectra and are studied
independently (the experimental techniques and equipment are different): nuclear transitions yield -rays
independently of atomic and molecular structure, atomic (electronic) transitions yield X-ray and UV-ray
(fine spectral lines) independently of nuclear and molecular structure, and molecular transitions (vibration
and rotation) yield visible and IR bands (thick lines) independently of nuclear and atomic structure (e.g.
ions radiate the same as neutral atoms in the visible and IR).
Radiation is constantly emitted by matter at equilibrium due to the continuous energy transitions
associated to the prevailing temperature. A system out of equilibrium may emit additional radiation
depending on the reactions taking place inside. Most optical techniques, however, rely on the analysis of
stimulated emission due to an apply coherent radiation source (laser), although the stimuli may also be a
diffuse light, an electron beam, an electric or magnetic field, a spark, an arc, a flame, etc.).
Stimulated radiation due to a coherent radiation source may cause:
Optical resonance when the source frequency coincides with some natural frequency of the
system, and the molecules radiate at the same frequency and in all directions, without lag.
Chemiluminescence emission, when the source frequency coincides with some natural frequency
of the system, and the molecules radiate at a lower frequency and in all directions, with a lag that
may be >10-4 s (and then it is called phosphorescence), or <10-4 s (and it is called fluorescence).
Elastic scattering (Rayleigh scattering), when the system is excited with any frequency and emits
at the same frequency with a small amplitude (some 10-3 the intensity of the source, larger at
larger frequencies), and non-isotropic but lobular and polarised.
Inelastic scattering (Raman scattering, 1928), when the system is excited with any frequency and
emits at the several different frequencies with a very small amplitude (some 10-6 the intensity of
the source) and non-isotropic but lobular and polarised. The frequency differences with the

source depends on the type of molecules, and they may be below the exciting one (called Stokes
lines) or above it (called anti-Stokes lines).

CLASSIFICATION
Three main items must be considered in optical diagnostics: the object, the radiation source, and the
radiation sensor.
OBJECT TYPES
According to the object, optical diagnostic techniques may be classified in two main groups:
Heterogeneous systems, basically particles (isolated or forming a mist) in a fluid matrix, and fluid
boundaries.
Homogeneous systems, basically smoothly changing fluid volumes.
RADIATION SOURCES
According to the radiation source, optical diagnostic techniques may be classified as follows:
Own radiation emission by the object (thermal radiation and chemiluminescent radiation)
External radiation source
o Visible white light, usually from incandescent lamps, but sometimes from special
luminaries (Nernst, Globar, Hg, Na).
o Monochromatic light, mostly from lasers (monochromators with white light yield much
lower spectral power). Although it depends on temperature, typical He-Ne lasers have 10
mW and =0.633 m (the iodine stabilized Helium-Neon laser has =0.6329914 m),
cheap diode lasers may have 10 mW and ~0.8 m (AlGaAs) or =0.65 m (AlGaInP),
powerful NdYAG lasers have 100 W and =1.064 m (invisible near infrared), and the
most powerful, CO2 lasers have >1 kW and =10.600 m (invisible far infrared). Excimer
lasers are pulsed gas discharge lasers which produce optical output in the ultraviolet (the
actual wavelength can be changed by changing the gas mixture).
DETECTOR TYPES
According to what is detected (image), optical diagnostic techniques may be classified as follows:
Radiation origin: object own emission (by its temperature or internal processes), source own
emission outlined by the object, source reflection on the object and surroundings (typical eyesight), source refraction along the object, object emission stimulated by the external source
(dispersion, luminescence).
Image information: ray intensity (bright, contrast, colour and saturation of typical eye-sight
imaging), ray deflection refractometry (moir, schlieren, shadowgraphy), interferometry (radiation
phase measurement), and holography (intensity and phase measurement).
Spatial dimensionality: point sampling, line of sight, bidimensional, tridimensional.
Spectral dimensionality: white light, monochromatic.
Spectral frequency band: visible, infrared, ultraviolet, microwave, etc.
Image sensor: photochemical or photoelectric. The traditional chemical photogram (Gr. photo,
light, gram, message) was much used since its development by Daguerre in 1839 until the end of
the XX c., that was substituted by digital photography. Photoelectric sensors may be
onedimensional (photodiodes, photomultipliers, field-effect transistors) or twodimensional
(vidicon tubes, charge coupled devices, CCD, are the most used). Image sensors (eye,
photographic plate, CCD) are two-dimensional.
RADIATION EMISSION AND RADIATION ABSORPTION
The own emitted radiation from a sample can be used to diagnose hot or excited objects as flames. The
radiation absorbed by an object from a known wide-band light, can be used to diagnose cold and hot

objects. Molecules absorb light only at certain characteristic wavelengths called its spectrum. At infrared
wavelengths, the spectrum results from vibrations of the atoms in the molecule, while at visible and
ultraviolet wavelengths the spectrum is caused by the electrons orbiting the molecule. The spectrum can
be calculated from quantum mechanics or measured in the laboratory
Condense matter yields a continuous spectrum dependent on its temperature (it approaches blackbody
radiation for dielectric materials like soot), but gases yield clearly separated spectral bands (they would
require optical sizes of many metres to approach blackbody radiation).
Radiation emission may be used to measure temperature or to analyse species and concentrations. If the
object is cold, it can be raised to high temperatures by contact with a hot wire or by exposure to a flame
(e.g. by dipping an inert platinum wire in the solution or powder sample and putting it on a Bunsen
flame).
Light is often said to have a colour temperature (not the real temperature of the emitting body, but the one
of a blackbody that would give the same colour sensation); the colour temperature of some common light
sources are: 2000 K for a candle flame, 3000 for an incandescent lamp, 4000 K for a carbon-arc or a
magnesium-flash light, 6000 for direct sunlight and 10 000 K for open sky.
The colloquial usage of "red hot," "white hot," and so on, is part of the colour sequence black, red,
orange, yellow, white, and bluish white, seen as an object is heated to successively higher temperatures.
RADIATION SCATTERING
When electromagnetic radiation impinges on a material particle, some part is re-radiated at different
angles from incidence (what is known as scatter). Scattering radiation depends on the size of the particle
and radiation-source wavelength. The interaction may be described as follows:
For large particles (say >1 mm that is the typical minimum width of a collimated beam, very much
larger than the source wavelength, ), some part of it is absorbed, some part is reflected (both
specularly and diffusively), and the rest is refracted according to the transmissivity of the particle.
The geometrical optics approximation applies (i.e. ray tracing), without any influence of .
For medium-size particles (say from 1 m to 1 mm) the bending of outlining light-rays, i.e. edge
diffraction, has to be considered. Diffraction is the name given to any deviation from the laws of
geometrical optics; it was modelled by Fresnel in 1818. Diffraction in the far zone, the Fraunhofer
approximation, is good enough in this size range, instead of the full Mie scattering theory.
For small particles (say 1 m, comparable to the wavelength of the light used, ), some part of the
incident radiation is scattered in a non-isotropic, non-symmetric lobular pattern with intensity
independent of frequency, larger to the foreword, what is known as Mie scattering. Tyndall effect
is due to Mie scattering, as well as other whitish effects are: clouds, dispersion opalescence,
foams, mists, and dust haze (they scatter all wavelengths equally).
For very small particles (<1/10 m, or better /10), some part of the incident radiation is scattered
in a non-isotropic, symmetric lobular pattern with intensity proportional to the fourth power of the
frequency, what is known as Rayleigh scattering. The blue colour of sky and the reddish colours at
sunset and sunrise result from Rayleigh scattering. Scattered light intensity depends on species
concentration and on species cross-section, so that if one is known, the other may be found (this
simple technique is widely used in combustion, and named Rayleigh line scattering, RLS). Most
scattered radiation has the same frequency as the incoming radiation (Rayleigh scattering), but
some parts have different frequencies and are given the specific name of Raman scattering.
Radiation dispersed from a monochromatic source is the best mean to diagnose flames. Scattered
radiation is fractioned in a tuneable monochromator (a prism or better a grating) and focused onto a

detector (visual or infrared), for measuring the actual species (based on wavelength) and its concentration
(based on intensity detection).
HOLOGRAPHY
Holography, developed by Gabor in 1947, is the method for recording and reconstruction the whole
information of an optical scene (intensity and phase of the light waves). In traditional imaging, light
scattered by the object in the scene is focused by a lens onto a 2D-plate, building a planar projection of
the scene proportional to light intensity without any phase information. For reconstruction, i.e. for
viewing, we just look at the image with any light and we see a 2D-frozen image of the original scene,
irrespective of the eye position.
To build a hologram, the procedure is to split a coherent light beam, shine one part on the scene, and store
the interference of the reference beam with the beam scattered by the object. For reconstruction, i.e. for
viewing, the recorded hologram is places in a beam of the same coherent light used for recording, what
sheds a virtual replica of the initial scene as if the objects were there; the recording plate acts like a
window through which the observer can peep through.
Holography requires a large temporal coherence of the source light (a very monochromatic laser) to
produce sharp interferences, and large spatial coherence of the source light (a very planar beam) to enable
reconstruction by any similar laser and not necessarily the recording one. Holography can be used to
visualise objects and particles, or to produce interferograms.
TOMOGRAPHY
Tomography is an image synthesis technique based on building a three-dimensional visualisation by
juxtaposition of many two-dimensional adjacent images (sectional views). Tomography is used in flame
structure research to avoid the accumulated contribution from all the scene; flame sheets are diagnosed at
a time, and later the three-dimensional structure is built with computer aid.
INTERFEROMETRY
Interferometry is the measuring of the refractive index field of an object phase by means of the
interference pattern formed when a reference light beam is combined with the beam going through the
object. Both beams have to be spatial and time coherent, and that is why a single laser beam, split in two,
is used.
The refractive index field is directly related to the density of the fluid object, what depends on
temperature and composition, allowing to resolve one of them if the other is known.
Many different types of interferometers have been developed, most of them requiring an optical bench (a
very rigid setup, isolated from external noise) to work.
SPECKLE INTERFEROMETRY (SI)
A coherent light is split in two collimated beams; one beam passes through the object, then through a
scattering solid, then it is combined with the other beam, and the combination is focused on an image
recorder, generating a speckle pattern (a blurred image to the eye). When two such speckle images are
subtracted (optically or digitally), a clear image of the refractive index field appears. Usually a reference
speckle is shot ob the object before the flow starts, so that continuous subtraction to the subsequent
speckles yields a sequence of refractive maps.

MOIRE DEFLECTOMETRY (MD)

A collimated light beam (preferably coherent) goes through an object phase, then through two separated
parallel gratings and is finally focused on an image plane. The superposition of the two fine linear
gratings yields an X-pattern of intensity fringes (moir pattern) that is distorted by the ray deflection
within the object path. The sensitivity of the instrument can be tuned by adjusting the distance between
gratings and their relative rotation.
RAMAN SCATTERING (RS)
Raman scattering is the most used spectrometric technique of analysis. It measures the inelastic radiation
scattered by molecules when a strong laser light is shone on them (the signal is weak). The shifting in
wavelength from the source depends on the type of molecules, whose concentration depends on the line
intensities, whereas the relative intensities amongst the lines for a single species is related to the local
temperature (in reality, to the vibrational-rotational temperature of the molecules, that may be different to
the translational temperature if there is no local equilibrium).
Coherent anti-Stokes Raman Scattering (CARS)
In this technique, two lasers are focused on the sample (a point, sheet or volume) at the same time, with
their frequencies tuned to enhance the response of one type of molecule by the process of stimulated
emission. Coherent anti-Stokes Raman scattering is being used for accurate temperature measurements in
research on turbine combustors.
LASER INDUCED FLUORESCENCE (LIF)
In this technique, a laser of appropriate wavelength is focused on the sample so as to excite the electronic
energy levels in some type of molecules, notably OH-radicals and polycyclic aromatic hydrocarbons
(PAH), that upon subsequent relaxation yield a characteristic emission. LIF has been successfully used,
for instance, on optically accessible engines to measure two-dimensional temperature fields and fuel
concentration fields, before and after ignition, using excimer lasers at 248 nm and 308 nm tracer
molecules (3-pentanone, 10%wt) added to the fuel (iso-octane).
LASER INDUCED INCANDESCENCE (LII)
This technique is used for detailed analysis of soot formation. Visible radiation from a sooting flame
comes from all regions of the flame, giving only an overall picture of a three-dimensional and often
unsteady phenomena; with a powerful laser, however, a line or sheet is shed onto the flame and only soot
incandescence in that part is sampled, since the natural emission is much dimmer.
(back to Combustion)