gas

2010

PTQ supplement

gas cov copy.indd 1 18/3/10 11:55:58

 Creating excellence from raw materials
– you need a specialist
W W W. T O P S O E . C O M

Add value to your natural gas

with Topsøe Technologies
Haldor Topsøe offers you the technology, catalyst and experience for profitable use of your natural gas feed
– whether you are seeking a new plant or a revamp of an existing plant.

- hydrogen - ammonia
- synthesis gas to chemicals - methanol
- synthesis gas to fuels - DME

Through research-driven technologies, Haldor Topsøe offers impartial counselling on the selection of the best technology
and catalyst for your plant.

www.ptqenquiry.com
for further information

haldor.indd 1 17/3/10 13:01:48

 ptqPETROLEUM TECHNOLOGY QUARTERLY

2010
gas
www.eptq.com

5 A blow to exports
Chris Cunningham

7 Enhanced sulphur recovery from lean acid gases containing

COS and mercaptans
Angela Slavens, Justin Lamar, Sara O’Dell and Laura Francoviglia
Black & Veatch

19 Detecting and dealing with hydrate formation

Bahubali Chandragupthan PL Engineering
Girish Babu Nounchi L&T - Gulf

27 An optimum line-up for sour gas processing

Saeid Mokhatab
Tehran Raymand Consulting Engineers

35 Tray hydraulic operating regimes and selectivity

Ralph Weiland, Nathan Hatcher and Jaime Nava
Optimized Gas Treating

45 Processing NH3 acid gas in a sulphur recovery unit

Michael Quinlan and Ashok Hati
KBR

Waste Management Inc and Linde North America’s plant at the Altamont Landfill near Livermore, California, is the
world’s largest landfill gas to liquefied natural gas plant. Photo: Linde Group

©2010. The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights
reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means
– electronic, mechanical, photocopying, recording or otherwise – without the prior permission of the copyright owner.
The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while
every care has been taken in the preparation of all material included in Petroleum Technology Quarterly the publisher cannot be held
responsible for any statements, opinions or views or for any inaccuracies.

contents/ed com copy 7.indt 1 19/3/10 09:57:15

 www.ptqenquiry.com
for further information

merichem.indd 2 17/3/10 13:05:03

merichem.indd 3 17/3/10 13:05:21
 Model for our technologies.
Lurgi is the worldwide leading partner when clean conversion is postulated. We command
sustainable processes which allow us to make better use of oil resources or biomass than
ever before.

With our technologies we can produce synthesis gases, hydrogen or carbon monoxide: for
downstream conversion to petrochemicals. Based on resources like natural gas, coal and tar
sand we produce synthesis gas which we convert into low-pollutant fuels.

Enhanced sustainability: from biomass which does not compete with the food chain, we
can recover ultra-pure fuels burning at a low pollutant emission rate which are excellently
suited for reducing the carbon footprint. You see, we are in our element when it comes to
sustainable technologies.

Build on our technologies.

Call us, we inform you: +49 (0) 69 58 08-0
www.lurgi.com
A member of the Air Liquide Group
1122_e

www.ptqenquiry.com
for further information

lurgi.indd 1 17/3/10 14:36:39

 T tq
p he European Union has arguably
been the global leader in biodiesel
production and use, with overall
biodiesel production increasing from 1.9
million tonnes in 2004 to nearly 10.3 million
PETROLin
tonnes M TECHBiodiesel
EU2007. NOLOGY QU ARTERLY
production in the
US has also increased dramatically in the
A blow to exports

past few years from 2 million gallons in 2000

to approximately 450 million gallons in 2007.
According to the National Biodiesel Board,

W
171 companies own biodiesel manufacturing
Editor hen Qatar’s huge LNG production industry and its associated super-
plants and are actively marketing biodiesel.1.
Chris Cunningham chilled tankers got under way, the balance of global gas supply looked
editor@petroleumtechnology.com
The global biodiesel market is estimated set to change for the foreseeable future. The greater proportion of
to reach 37 billion gallons by 2016, with an liquefied Arabian production was destined to make up a growing shortfall in US
Production Editor
average annual growth rate of 42%. Europe domestic gas supply, while Russia’s Gazprom prepared to unleash major new
Rachel Zamorski
will continue to be the major biodiesel
production@petroleumtechnology.com supplies of pipeline gas into European markets as well as substantial amounts of
market for the next decade, followed closely LNG for the US.
by the USEditor
Graphics market. Not any more: the focus of breakthrough technology has shifted upstream to
AlthoughSamiuddin
Mohammed high energy prices, the production of so-called unconventional gas, which offers the promise of
graphics@petroleumtechnology.com
increasing global demand, drought continued self-sufficiency in North America. Meanwhile, the recession has hit
and other factors are the primary driv- demand from another major regional outlet for Qatari LNG, Japan and Korea.
Editorial
ers for higher food prices, food Supplies of LNG will grow increasingly abundant as new projects due to come
tel +44 844 5888 773
competitive feedstocks
fax +44 844 5888 667 have long been on stream in Qatar this year add another 80 million tonnes to annual supply,
and will continue to be a major almost 50% more than in 2008.
concern forSales
Advertising theManager
development of biofu- In response to changes in the market, Gazprom has postponed its project to
els. To compete, the industry has
Paul Mason develop the giant Shtokman gas field. The gas major had intended to liquefy
sales@petroleumtechnology.com
responded by developing methods to part of the production from Shtokman to supply the US. However, its remote
increase process efficiency, utilise or location and associated transportation costs mean it would be sold at a
Advertising Sales
upgrade by-products and operate with premium price at a time when gas prices are falling. When the customer’s
Bob Aldridge
lower quality lipids as feedstocks.
sales@petroleumtechnology.com position is no longer “buy it or nothing”, there is no future, at least in the short
term, in spending vast amounts on field development.
Feedstocks
Advertising Sales Office Five years ago, shale gas was one of a long list of energy technologies and
Biodiesel
tel +44 844refers to a diesel-equivalent
5888 771 resources championed by the few but with little prospect of a fruitful outcome.
fax +44 844 5888 662
fuel consisting of short-chain alkyl Now, in the US at the very least, it is seen as a viable stopgap between heavy
(methyl or ethyl) esters, made by the reliance on fossil fuels and future dependence on reliable and renewable
Publisher
transesterification of triglycerides, resources. Shale gas’s maverick reputation changed during the past decade as
Nic Allen
commonly known as vegetable oils or
publisher@petroleumtechnology.com a result of work by independent drillers Mitchell Energy in the Barnett Shale
animal fats. The most common form measures of Fort Worth Basin in Texas. The company developed techniques to
uses methanol, the cheapest alcohol
Circulation ensure its wells remained within the productive measures and without loss of
available,
Jacki Watts to produce methyl esters. pressure or water ingress, and its own brand of rock fracturing.
circulation@petroleumtechnology.com
The molecules in biodiesel are primar- Rounded estimates say that the breakthrough in unconventional gas produc-
ily fatty acid methyl esters (FAME), tion has assured a supply that will last for 100 years, provided, of course, that
Crambeth Allen Publishing Ltd
usually created by transesterification the uptake in new supplies does not deplete the resource too enthusiastically.
Hopesay, Craven Arms SY7 8HD, UK
between
tel +44 844 fats
5888 and776methanol. Currently, Proponents of unconventional gas say that it can be the bridge between fossil
biodiesel
fax +44 844is5888 produced
667 from various fuels and a renewable energy future. In 2000, gas shale technology accounted
vegetable and plant oils. First-genera- for 1% of US natural gas production, but today represents nearer 20% and may
tion
ISSN food-based
1362-363X feedstocks are straight achieve 50% by 2030, according to industry researcher IHS CERA. Taking into
vegetable oils such as soybean oil and account other sources such as coal bed methane, the “unconventional” portfo-
animal fats such as tallow, lard, yellow
Petroleum Technology Quarterly (USPS 0014-781)
lio has already achieved the 50% mark in the US, which has surpassed Russia
grease,
is publishedchicken fatannual
quarterly plus andCatalysis
the by-products
edition as the world’s leading producer of natural gas. And, to make matters even
by Crambeth Allen Publishing Ltd and is distributed
of the
in the USA byproduction
SPP, 75 Aberdeenof Omega-3
Rd, Emigsville, PA fatty more uncomfortable for Russia’s gas industry, Middle Eastern supplies of LNG
acids from fish
17318. Periodicals postageoil. Soybean
paid at Emigsville PA.oil and into Europe are being stepped up and represent heavy competition for
Postmaster: send address changes to Petroleum
rapeseeds oil are
Technology Quarterly c/o POthe common source Gazprom’s pipeline gas.
for biodiesel
Box 437, Emigsville,production
PA 17318-0437 in the US and Texas’s propitious gas-bearing geology is not unique. While several oil
Back numbers available from the Publisher
Europe in inc
at $25 per copy quantities
postage. that can produce majors, including BP, Total and StatOil, are busily signing up joint venture
enough biodiesel to be used in a agreements in the US, drilling for unconventional supplies is starting up in
commercial market with currently Europe and China, where there is an eagerness to hold down levels of
applicable technologies. imported gas. It all amounts to a rejuvenated future for local supplies of gas.
First-generation feedstocks for
chris cunningham

GAS 2010 5

contents/ed com copy 7.indt 2 17/3/10 14:54:39

 Travel with us,
and you’re always
on the road
to innovation.

THE CLEAN FUELS TEAM

www.criterioncatalysts.com
© 2008 Criterion Catalysts & Technologies L.P. (cri815_0508)
www.ptqenquiry.com
for further information

cri815_A_roadInnovation_ptq.indd 1 5/29/08 12:07:17 PM

criterion 1 9/3/09 10:13:45
 Enhanced sulphur recovery from lean acid
gases containing COS and mercaptans
Claus configurations are available for more effective sulphur recovery from lean
acid gas containing significant concentrations of organic sulphur compounds
Angela Slavens, Justin Lamar, Sara O’Dell and Laura Francoviglia
Black & Veatch
C
laus sulphur recovery from
lean acid gas can be problem-
atic. Low H2S concentration
in acid gas creates a cool thermal
reactor temperature and a tendency
for flame instability, reducing the
reliability of conventional operation
of a modified Claus sulphur recov-
ery unit (SRU). The acid gas
enrichment (AGE) process is
commonly employed to provide
greater reliability, flexibility and
improved Claus unit operations.
AGE achieves these results by
increasing lean acid gas H2S concen-
tration, producing an enriched acid
gas feed stream with significantly
reduced volumetric flow, for
processing in the Claus SRU.
Lean acid gas containing signifi-
cant concentrations of mercaptans
and/or carbonyl sulphide (COS)
creates difficulty for AGE to achieve
high sulphur recovery. The selec-
tive treating solvents commonly
utilised for AGE absorb H2S from
the acid gas, but do not absorb
mercaptan or COS components.
These organic sulphur species
remain in the AGE absorber over-
head stream and flow to the
incinerator without being recov-
ered, thereby negatively affecting
the efficiency of sulphur recovery.
With acid gas having high organic
sulphur content, 1% or more of the
sulphur in lean acid gas can be
present as mercaptan or COS
sulphur. Recovery of this sulphur is
required to achieve high sulphur
recovery efficiency.
This article compares the perform-
ance of several SRU schemes
processing a lean acid gas contain-
ing COS and mercaptans. A
conventional SRU/tail gas unit
www.eptq.com
b&v.indd 1
H2S and having significant COS and
Acid gas feed stream used for
comparison of test cases
mercaptan concentrations.
Component Acid gas, Acid gas Lean acid gas test cases
lbmoles/hr mole fraction The lean acid gas stream used for
comparison of the various process
H2S 275.459 24.968
CO2 756.060 68.530
configuration test cases is shown in
CH4S 5.509 0.499 Table 1. Acid gas sulphur content is
COS 5.509 0.499 100 t/d. About 4% of the acid gas
H2O 55.107 4.995 sulphur is present as COS and/or
CH4 3.110 0.282
C2H6 1.164 0.106
mercaptan; therefore, recovery of
C3H8 0.803 0.073 these sulphur species is important
iC4H10 0.100 0.009 if high sulphur recovery is to be
nC4H10 0.341 0.031 achieved. Acid gas such as this
iC5H12 0.040 0.004
nC5H12 0.020 0.002
could be produced from sweetening
C6+ 0.040 0.004 natural gas with high COS/
Total 1103.262 100.002 mercaptan content, where removal
Temperature, °F 113 of these components is necessary to
Pressure, psia 28.3
achieve a low sales gas sulphur
specification. Organic sulphur
Table 1 removal from natural gas is becom-
ing increasingly important in
(TGU), a conventional SRU/TGU today’s gas market.
with acid gas enrichment, a SRU/ Table 2 describes the seven
TGU scheme bypassing acid gas to process configuration test cases
the TGU, and a SRU/TGU scheme compared here, each of which is
that entails routing the AGE illustrated in Figures 1 to 7.
absorber overhead to the TGU The Case 1 process configuration
(based on a concept originally is shown in Figure 1. Acid gas from
patented by Shell in 1982) are the upstream sour gas treatment
compared on the basis of process- unit is processed in a conventional
ing a lean acid gas containing 25% two-bed Claus SRU. SRU tail gas
Process configuration test cases for comparison
Test case Description
1 Conventional SRU/TGU
2 Conventional acid gas enrichment, with SRU/TGU
3A SRU/TGU, with acid gas routed directly to TGU
3B SRU/TGU, with acid gas routed directly to TGU, with COS hydrolysis in TGU
4A AGE with enrichment absorber overhead routed to TGU
4B AGE with enrichment absorber overhead routed to TGU; TGU absorber rich solvent
routed to midpoint of AGE absorber
4C AGE with enrichment absorber overhead routed to TGU; TGU absorber rich solvent
routed to AGE absorber; no lean solvent to AGE absorber
Table 2
GAS 2010 7
18/3/10 09:31:35
 Acid gas

Claus converters

Incinerator

Thermal
reactor Fuel
Air

Sulphur Sulphur Sulphur

Air

Hydrogenation
reactor
Quench TGU Regenerator
column absorber
Fuel

Inline
burner/mixer
Air
WHB

Figure 1 Case 1 PFD: conventional SRU/TGU

flows to a methyl diethanolamine
(MDEA)-based TGU, where resid-
ual sulphur species in the SRU tail
gas are converted to H2S, which
is subsequently removed in the
downstream TGU absorber. The sulphur in this stream. The
TGU regenerator regenerates the combined acid gas H2S concentra-
rich MDEA solvent, and the over- tion is higher than in the inlet acid
head from this tower is routed to gas because the TGU regenerator
the front of the SRU for recovery of produces a recycle acid gas stream

Claus converters

Incinerator

Thermal
Air
reactor Fuel

Sulphur Sulphur Sulphur

Air
TGU absorber

AGE absorber

Hydrogenation
reactor
Quench Regenerator
Fuel column

Inline
burner/mixer
Air
WHB

Acid gas

Figure 2 Case 2 PFD: conventional AGE with SRU/TGU

8 GAS 2010 www.eptq.com

b&v.indd 2 18/3/10 09:31:47


Thermal
reactor
Air
Acid gas
Fuel
Inline
burner/mixer
Air
Figure 3 Case 3a PFD: SRU/TGU with acid gas routed to the TGU
that is more concentrated than the
inlet acid gas.
Case 1 is a conventional SRU/
TGU arrangement commonly
applied to achieve good sulphur
recovery efficiency, typically above
99.7% of the inlet sulphur. Most
COS and mercaptan sulphur can be
recovered with this approach.
Hence, it would be the preferred
process configuration for most
plants, except when acid gas H2S
concentration falls below a level
that enables sustainable Claus unit
performance. Below approximately
25% H2S, low thermal reactor
temperature and flame instability
make operation difficult, and
another flow scheme should be
considered to improve operation.
A split-flow configuration, where
some acid gas bypasses the thermal
stage, can improve operation of the
acid gas burner and thermal reac-
tor. However, split-flow exposes the
catalytic section to acid gas hydro-
carbons, which can foul the Claus
catalyst. Acid gas enrichment,
where acid gas is first fed to an
amine absorber to absorb H2S and
slip some CO2, produces an acid
gas feed stream with a higher H2S
concentration that is a more suitable
feed stream for a Claus SRU. In an
www.eptq.com
b&v.indd 3
Claus converters
Sulphur Sulphur Sulphur
Hydrogenation
reactor
Quench TGU
column absorber
WHB
effort to improve Claus plant opera-
tion with lean acid gas, an
additional six cases have been
developed for comparison, all of
which employ AGE schemes.
The Case 2 process configuration
is shown in Figure 2. Acid gas from
the upstream sour gas treatment
unit is first processed in an AGE
absorber, which absorbs H2S from
the acid gas using MDEA solvent.
The AGE absorber’s overhead gas,
containing most of the CO2 in the
acid gas and other components not
absorbed by the amine, flows
directly to the tail gas incinerator.
Concentrated acid gas from the
combined AGE/TGU regenerator,
containing all of the sulphur to be
recovered, plus the TGU’s recycled
sulphur, flows to the SRU/TGU for
sulphur recovery. The regenerator
acid gas has a significantly higher
H2S concentration than the inlet
acid gas feed stream to the AGE
absorber.
AGE re-absorbs the acid gas H2S
in MDEA solvent, which must then
be regenerated. The increased
solvent regeneration energy and
larger solvent system are significant
cost penalties when compared with
the conventional Case 1 approach.
However, greatly improved Claus
Incinerator
Fuel
Air
Regenerator
SRU operation with enriched acid
gas makes smooth, reliable opera-
tion possible for lean acid gases that
cannot be effectively processed
using Case 1. Additionally, AGE
significantly reduces the acid gas
volumetric flow rate to the SRU,
leading to significant reductions in
equipment sizes and corresponding
capital costs. Another advantage to
using AGE is that acid gas hydro-
carbons are, for the most part,
slipped to the incinerator by the
enrichment absorber. Removal of
these hydrocarbon components
from the Claus feed stream reduces
Claus air control problems, which
are often caused by fluctuations in
hydrocarbon concentrations.
AGE in Case 2 does not recover
sulphur from acid gas COS or
mercaptans. These components are
not absorbed significantly by
amine-based selective treating
solvents; therefore, they flow in
the enrichment absorber overhead
to the incinerator and are emitted
to the atmosphere as SO2, nega-
tively affecting sulphur recovery
efficiency.
The Case 3a process configuration
is shown in Figure 3. Case 3a
provides a method for combining
AGE with sulphur recovery from
GAS 2010 9
18/3/10 09:31:57

Thermal
reactor
Air
Hydrogenation
reactor
Acid gas
Fuel
Inline
burner/mixer
Air
Figure 4 Case 3b PFD: SRU/TGU with acid gas routed to the TGU, including COS hydrolysis in the TGU
COS and mercaptan components of
acid gas. This unconventional
approach involves preheating the
acid gas, mixing it with the
preheated Claus tail gas, and rout-
ing the combined stream to the tail
gas hydrogenation reactor. Here,
sulphur in the acid gas COS and
mercaptan components is converted
to H2S across the reactor. The tail
gas is subsequently cooled by water
quench, followed by H2S absorption
in the downstream tail gas absorber,
as in a conventional TGU configu-
ration. Concentrated acid gas from
the regenerator, containing all of
the sulphur to be recovered, plus
the TGU recycled sulphur, flows to
the SRU/TGU for sulphur recovery.
The tail gas absorber serves as the
enrichment absorber, slipping most
acid gas CO2 to the incinerator.
Case 3a improves upon Case 2 by
providing an opportunity for
removal of mercaptans and COS
via conversion to H2S and subse-
quent absorption in the TGU
absorber. However, in Case 3a, COS
hydrolysis conversion to H2S is
limited by equilibrium at the
high temperature condition
required in the hydrogenation
reactor. COS that is not converted
to H2S is not absorbed by the tail
gas absorber’s selective amine
10 GAS 2010
b&v.indd 4
Claus converters
Sulphur Sulphur Sulphur
COS
hydrolysis
reactor
TGU absorber
Quench
column
WHB
solution and therefore not
recovered.
The Case 3b process configuration
is shown in Figure 4. Case 3b is
similar to Case 3a and also provides
a method for combining AGE with
sulphur recovery from COS and
mercaptan components of acid gas.
The Case 3b flow sheet is identical
to that of Case 3a, except for the
addition of a COS hydrolysis reac-
tor downstream of the TGU
hydrogenation reactor, to improve
COS conversion to H2S and there-
fore improve overall sulphur
recovery efficiency. Following tail
gas hydrogenation, a COS hydroly-
sis reactor is added just upstream
of the quench column, where the
gas has been cooled to 350°F
(177°C) by a low-pressure waste
heat boiler. The lower temperature
greatly improves COS conversion
equilibrium, as the COS hydrolysis
reactor contains a catalyst bed
active for COS conversion at these
conditions. Using COS hydrolysis
in this manner improves sulphur
recovery, because COS slip to the
incinerator is minimised.
Case 3b improves upon Case 3a
by providing enhanced COS
hydrolysis in the additional catalyst
bed in the TGU, allowing for
greater overall sulphur recovery
Incinerator
Fuel
Air
Regenerator
efficiency. Cases 3a and 3b provide
a benefit similar to that achieved in
Case 2; that is, reducing Claus oper-
ating problems caused by acid gas
hydrocarbon components. These
components are, for the most part,
slipped to the incinerator by the tail
gas absorber. However, it should be
noted that any hydrocarbons
present in the acid gas feed are
routed directly to the hydrogena-
tion reactor in Cases 3a and 3b,
which present the potential for foul-
ing the hydrogenation catalyst.
The Case 4a process configuration
is shown in Figure 5. Case 4a is
similar to Case 2 except that, in
Case 4a, the AGE absorber over-
head stream flows to the TGU
hydrogenation reactor rather than
to the incinerator. Routing the AGE
overhead to the hydrogenation
reactor allows conversion of the
COS and mercaptans to H2S, for
absorption in the downstream TGU
absorber. In a similar fashion to
Case 2, concentrated acid gas from
the combined AGE/TGU regenera-
tor, containing all of the sulphur to
be recovered, plus the TGU’s recy-
cled sulphur, flows to the SRU/TGU
for sulphur recovery. The Case 4a
process configuration was patented
by Shell in 1982.
Case 4a improves upon the Case
www.eptq.com
18/3/10 09:32:09
 Claus converters

Incinerator

Thermal
Air
reactor Fuel

Sulphur Sulphur Sulphur

Air

AGE absorber
TGU absorber
Hydrogenation
reactor
Regenerator
Quench
Fuel column

Inline
burner/mixer
Air
WHB

Acid gas

Figure 5 Case 4A PFD: AGE with enrichment absorber overhead routed to the TGU

3 configurations by reducing the
hydrogenation reactor H2S concen-
tration, which improves COS
conversion to H2S. Case 4a also has
the advantage of increased absorber
H2S concentration, which allows 4a than it is for the Case 3 configu-
higher solvent loading and reduced rations, which reduces equipment
MDEA circulation. Finally, the sizes and consequently reduces
process gas volumetric flow rate capital cost. One disadvantage of
through the TGU is lower for Case the Case 4a process configuration is

Claus converters

Incinerator

Thermal
Air
reactor Fuel

Sulphur Sulphur Sulphur

Air

Quench
column
TGU absorber

AGE absorber

Hydrogenation
reactor
Regenerator

Fuel

Inline
burner/mixer
Air
WHB

Acid gas

Figure 6 Case 4B PFD: AGE with enrichment absorber overhead routed to the TGU, TGU absorber rich solvent routed to the midpoint of
the AGE absorber

www.eptq.com GAS 2010 11

b&v.indd 5 18/3/10 09:32:23

 Claus converters

Incinerator

Thermal
Air
reactor Fuel

Sulphur Sulphur Sulphur

Air

Quench
column

TGU absorber

AGE absorber
Hydrogenation
reactor
Regenerator

Fuel

Inline
burner/mixer
Air
WHB

Acid gas

Figure 7 Case 4C PFD: AGE with enrichment absorber overhead routed to the TGU, TGU absorber rich solvent routed to the AGE absorber

that acid gas hydrocarbons slip absorber, the total solvent circula- identical to Case 4b, except that, in
through the enrichment absorber tion rate is considerably decreased, an attempt to further reduce the
and flow to the hydrogenation reac- providing a significant reduction in total solvent flow and required
tor, where they present the potential regenerator size and regeneration regeneration duty, no lean solvent
for hydrogenation catalyst fouling. heat duty. is pumped to the enrichment
The Case 4b process configuration Case 4b has the same advantages absorber. The TGU absorber
is shown in Figure 6. Case 4b is over the Case 3 configurations that bottoms semi-rich solvent is
identical to Case 4a except that, in Case 4a has. In addition, Case 4b pumped to the top of the AGE
an attempt to reduce the total achieves lower solvent circulation absorber. The enrichment absorber
solvent flow and required regenera- rates compared with Case 4a as a bottoms rich solvent is more highly
tion duty, rich MDEA from the result of the use of semi-rich solvent loaded with H2S than in Case 4b.
TGU absorber is pumped to the in the AGE absorber. Case 4b An increased amount of H2S
midpoint of the AGE absorber. The suffers the same disadvantage as slips through the enrichment
tail gas absorber’s rich solvent is Case 4a, associated with the poten- absorber in its overhead stream
not highly loaded with H2S; it is tial for hydrogenation catalyst flowing to the hydrogenation reac-
“semi-rich” and has significant H2S fouling as a result of hydrocarbons tor, but this H2S is absorbed in the
absorption capacity at AGE in the acid gas feed. tail gas absorber downstream, so it
absorber conditions. By using the The Case 4c process configuration does not negatively affect recovery
semi-rich solvent in the AGE is shown in Figure 7. Case 4c is efficiency.
Case 4c has the same advantages
over the Case 3 configurations as
Sulphur recovery efficiency comparison for test cases
do Cases 4a and 4b. Additionally,
Case 4c achieves lower solvent
Case 1 2 3A 3B 4A 4B 4C circulation rates than those of Case
TGU H2S to incinerator, lbmol/h 0.32 0.13 0.32 0.31 0.30 0.31 0.24
TGU COS to incinerator, lbmol/h 0.16 0.05 0.99 0.18 0.12 0.14 0.16
4a and Case 4b, as a result of the
AGE H2S to incinerator, lbmol/h - 0.11 - - - - - use of semi-rich solvent in the AGE
AGE COS to incinerator, lbmol/h - 5.47 - - - - - absorber, and no lean solvent feed
AGE mercaptan sulphur to incinerator, lbmol/h - 5.25 - - - - - to the AGE absorber. Case 4c suffers
Pit sweep sulphur to incinerator, lbmol/h 0.02 0.03 0.03 0.03 0.03 0.03 0.03
Total sulphur to incinerator, lbmol/h 0.50 11.04 1.34 0.52 0.45 0.48 0.43
the same disadvantages as Cases 4a
Incinerator stack SO2, ppmv dry and air-free 200 5,021 563 221 196 206 187 and 4b, related to the potential for
Sulphur recovery, % 99.83 96.15 99.53 99.82 99.84 99.83 99.85 hydrogenation catalyst fouling as a
result of hydrocarbons in the acid
Table 3 gas feed.

12 GAS 2010 www.eptq.com

b&v.indd 6 18/3/10 09:32:35

 Sulphur recovery efficiency
comparison Recycle acid gas Stack gas
The overall sulphur recovery effi- 40.18 lbmol/h S 0.50 lbmol/h S
ciency for each of the test cases is
given in Table 3. When preparing Acid gas SRU tail gas TGU tail gas
286.48 lbmol/h S 40.66 lbmol/h S 0.48 lbmol/h S
these cases, an attempt was made SRU TGU Incinerator
to achieve the highest practical
sulphur recovery for each case, on
a consistent basis. The objective was Sulphur product Sulphur pit vent
not, for example, to attempt to treat 286.00 lbmol/h S 0.02 lbmol/h S
tail gas H2S with amine more
aggressively in one case compared Figure 8 Sulphur balance for Case 1: conventional SRU/TGU
with another. For any of the cases,
sulphur recovery could be increased
Stack gas
or decreased by using a different AGE absorber overhead 11.04 lbmol/h S
10.83 lbmol/h S
amine, changing circulation rates,
changing regeneration heat duty or Acid gas SRU tail gas TGU tail gas
some other adjustment. The authors 286.48 28.38 0.18
lbmol/h S lbmol/h S lbmol/h S
believe that the representations AGE SRU TGU Incinerator
presented herein reflect an accurate

Semi-rich amine
30.48 lbmol/h S
comparison of the characteristics

Acid gas 303.85

Rich amine 284.01

lbmol/h S

Lean amine 2.28

for each case, on a consistent basis. Sulphur
Lean amine 8.36

lbmol/h S
Sulphur pit vent
product

lbmol/h S
0.03 lbmol/h S
lbmol/h S

Case 2, the AGE/SRU/TGU 275.47

lbmol/h S
configuration, achieves the lowest
sulphur recovery. Case 2 does not
Amine
recover sulphur from acid gas COS regenerator
or mercaptans because these
components are not absorbed
significantly by amine-based selec- Figure 9 Sulphur balance for Case 2: conventional AGE with SRU/TGU
tive treating solvents; therefore,
they flow in the enrichment
absorber overhead to the Recycle acid gas Stack gas
306.25 lbmol/h S 1.34 lbmol/h S
incinerator.
Case 3a achieves higher recovery
efficiency than Case 2, but lower SRU tail gas TGU tail gas
21.08 lbmol/h S 1.31 lbmol/h S
than the other cases. Residual tail SRU TGU Incinerator
gas COS is much higher for Case
3a, because the large quantity of
H2S feeding the TGU hydrogena- Sulphur product Sulphur pit vent
285.17 lbmol/h S 0.03 lbmol/h S
tion reactor inhibits the equilibrium Acid gas
of COS conversion. Overall sulphur 286.48 lbmol/h S

recovery suffers from the high

residual COS level. Sulphur recov- Figure 10 Sulphur balance for Case 3a: SRU/TGU with acid gas routed to the TGU
ery can be improved for Case 3a by
adding a tail gas COS hydrolysis
step to convert the high residual
COS to H2S, such that it can be Recycle acid gas
subsequently absorbed in the tail 307.11 lbmol/h S
Stack gas
gas absorber. 0.52 lbmol/h S
Table 3 shows that the addition SRU tail gas TGU tail gas
of this COS hydrolysis step, through 21.12 TGU with 0.49
lbmol/h S lbmol/h S
the application of Case 3b, raises SRU COS Incinerator
sulphur recovery efficiency to a hydrolysis
reactor
level comparable with Case 1, albeit
at the expense of an additional Sulphur product Sulphur pit vent
reactor. 285.99 lbmol/h S 0.03 lbmol/h S
Acid gas
Cases 4a, 4b and 4c provide 286.48 lbmol/h S
significant increases in sulphur
recovery efficiency over Case 2,
because these options allow a Figure 11 Sulphur balance for Case 3b: SRU/TGU with acid gas routed to the TGU, with
significant portion of the COS and COS hydrolysis in the TGU

www.eptq.com GAS 2010 13

b&v.indd 7 18/3/10 09:32:47

 H2S content (less than approxi-
3TACKGAS
!'%ABSORBEROVERHEAD 
mately 25% H2S), Case 1 should not
LBMOLH3 LBMOLH3 be considered viable.
The sulphur balance for each of
!CIDGAS 325TAILGAS 4'5TAILGAS
   the test cases is illustrated in
LBMOLH3
AGE SRU
LBMOLH3
TGU
LBMOLH3
Incinerator Figures 8 to 14.

COS conversion comparison

2 I C HAM I N E
LBMOLH3

3 E M I R I C H A M I N E
LBMOLH3
LBMOLH3
!CIDGAS
Concentration and conversion data

,EANAMINE 3ULPHUR 3ULPHURPITVENT
PRODUCT

 ,EANAMINE
LBMOLH3 for the hydrogenation reactor in
LBMOLH3
LBMOLH3  each case, and for the hydrolysis
LBMOLH3
reactor in Case 3b, are shown in
Amine Table 4. For these conversion
regenerator
results, close to equilibrium COS
conversion was assumed in the
Figure 12 Sulphur balance for Case 4a: AGE with enrichment absorber overhead routed to hydrogenation reactor, and about
the TGU 95% of equilibrium conversion was
specified for the hydrolysis reactor.
Stack gas The Case 3 configurations, which
AGE absorber overhead
10.72 lbmol/h S
0.48 involve routing acid gas directly to
lbmol/h S
the TGU, can achieve a similar tail
Acid gas SRU tail gas TGU tail gas gas COS level as the other cases if a
286.48 28.98 0.45
lbmol/h S lbmol/h S lbmol/h S hydrolysis reactor is included, as
AGE SRU TGU Incinerator illustrated by the overall COS
conversion results for Case 3b.
Lean amine 4.95

Acid gas315.01

Without hydrolysis, COS concentra-

Semi-rich amine
44.26 lbmol/h S
lbmol/h S
lbmol/h S

Sulphur Sulphur pit vent

product 0.03 lbmol/h S
tion in Case 3 is an order of
286.03
lbmol/h S
Lean amine magnitude higher than in the other
5.01
lbmol/h S cases, as illustrated by the results
Amine
for Case 3a.
Rich amine
324.97 regenerator
lbmol/h S Amine system comparison
Amine circulation rates, corre-
sponding to rich/lean amine
Figure 13 Sulphur balance for Case 4b: AGE with enrichment absorber overhead routed loadings and other amine system
to the TGU, TGU absorber rich solvent routed to the midpoint of the AGE absorber design data for each of the test
cases are shown in Table 5
Case 1 does not perform acid gas
enrichment. Therefore, Case 1
Stack gas
AGE absorber overhead 0.43 lbmol/h S requires the lowest MDEA solvent
17.64 lbmol/h S
rate, much lower than the other
Acid gas SRU tail gas TGU tail gas cases that do employ acid gas
286.48 27.64 0.40 enrichment. Total amine circulation
lbmol/h S lbmol/h S lbmol/h S
AGE SRU TGU Incinerator rates range from 70–185% greater
for the AGE cases than for Case 1.
Acid gas 313.72

The Case 1 process configuration

Semi-rich amine
lbmol/h S

52.07 lbmol/h S

Lean amine 7.19

Sulphur Sulphur pit vent results in 30 mole% H2S recycle acid

product
lbmol/h S

0.03 lbmol/h S
286.08 gas leaving the regenerator and 25.5
lbmol/h S
Rich amine mole% H2S acid gas feeding the
320.91 Claus unit, for the test case acid gas
lbmol/h S Amine
regenerator feed described in Table 1. Difficulty
in Claus operation caused by a low
acid gas H2S concentration in Case
1 makes the AGE achieved in the
Figure 14 Sulphur balance for Case 4c: AGE with enrichment absorber overhead routed to other cases desirable, in spite of
the TGU, TGU absorber rich solvent routed to the AGE absorber their larger MDEA units with
higher energy consumptions.
mercaptans in the AGE overhead to 1, the conventional SRU/TGU Case 2 is the conventional AGE
be recovered in the TGU. The configuration. However, when reli- flow sheet, with enrichment
recovery efficiencies for Cases 4a, able Claus SRU operation cannot be absorber overheads flowing to the
4b and 4c are similar to that of Case guaranteed because of low acid gas incinerator. Case 2 requires a lower

14 GAS 2010 www.eptq.com

b&v.indd 8 18/3/10 09:32:59

 tail gas absorber solvent feed rate
than Case 1, because AGE lowers
the tail gas CO2 content and reduces
CO2 co-absorbed in the tail gas
absorber. Case 2 rejects most acid
gas CO2 directly to the incinerator
such that this CO2 does not flow
through the tail gas absorber, and
this is the only instance among the
cases where this is achieved. In the
other cases, all acid gas CO2 flows
through the tail gas absorber.
However, the Case 2 enrichment
absorber requires a large MDEA
flow to absorb all acid gas H2S. The
total lean MDEA flow for Case 2 is
more than double the MDEA
requirement for Case 1.
Case 3 performs an H2S enrich-
ment step similar to that in Case 2,
but has the advantage of recovering
sulphur from acid gas COS and
mercaptans using the TGU. Case 3
requires a 15–20% greater total
MDEA solvent circulation rate than
Case 2 because of the additional
H2S absorption resulting from the
conversion of COS and mercaptans.
Also, the Case 3 absorber H2S
concentration is lower than in Case
2; therefore, H2S partial pressure is
lower and the rich solvent cannot
be as highly loaded with acid gas.
The Case 4 alternatives also have
the advantage of recovering sulphur
from acid gas COS and mercaptans,
similar to Case 3. Case 4a has the
same enrichment absorber MDEA
requirement as Case 2 because, in
both process flow schemes, the acid
gas feed is routed directly to the
enrichment absorber. Case 4a has
nearly the same tail gas absorber
MDEA requirement as Case 1,
because in Case 4a all acid gas CO2
flows through the tail gas absorber,
similar to Case 1. With an enrich-
ment absorber MDEA requirement
similar to Case 2 and a tail gas
absorber MDEA requirement simi-
lar to Case 1, Case 4a has the
highest total lean MDEA require-
ment of any of the cases.
In Case 4b, the tail gas absorber’s
semi-rich MDEA is fed to the
bottom section of the enrichment
absorber, reducing the amount of
lean MDEA required at the top of
the enrichment absorber. As a
result, total MDEA circulation
required in Case 4b is reduced by
www.eptq.com
b&v.indd 9
Tail gas COS conversion comparison for test cases
Case 1 2 3a 3b 4a 4b 4c
Hydrogenation reactor
H2S inlet, mole % 1.03 1.45 11.0 11.0 0.75 0.86 1.09
H2S outlet, mole % 1.77 2.29 11.7 11.7 1.68 1.83 2.01
CO2 onlet, mole % 37.4 14.7 34.6 34.6 38.8 38.5 38.8
CO2 outlet, mole % 38.8 16.4 35.8 35.8 40.3 40.1 40.3
COS onlet, mole % 0.102 0.064 0.245 0.245 0.297 0.292 0.284
COS outlet, mole % 0.0070 0.0034 0.0378 0.0378 0.0056 0.0064 .0072
COS conversion, % 93.2 94.2 84.5 84.5 98.0 97.8 97.5
COS hydrolysis reactor
COS inlet, mole % - - - 0.0378 - - -
COS outlet, mole % - - - 0.0068 - - -
COS conversion, % - - - 82.0 - - -
Overall results
COS conversion, % 93.2 94.2 84.5 97.2 98.0 97.8 97.5
Table 4
Amine system comparison for the test cases
Case 1 2 3a 3b 4a 4b 4c
TGU absorber
Lean MDEA, gpm 325 150 800 820 375 350 550
Inlet H2S, mole % 2.19 3.19 14.1 14.1 2.07 2.26 2.48
Inlet CO2, mole % 47.9 22.8 43.2 43.3 49.6 49.4 49.6
Overhead H2S, ppmv 185 151 188 185 185 188 148
CO2 slip, % 90.5 90.6 85.3 85.2 91.5 91.6 88.3
AGE absorber
Lean MDEA, gpm - 550 - - 550 345 -
Overhead H2S, ppmv - 175 - - 139 180 10,000
CO2 slip, % - 77.4 - - 80.2 80.3 84.2
Regenerator
Lean MDEA temp., ° F 104 104 104 104 104 104 104
Acid gas H2S, mole % 30.0 57.6 64.6 64.5 54.5 55.0 54.6
Lean H2S, mol/mol 0.0046 0.0078 0.0094 0.0094 0.0077 0.0074 0.0067
Lean CO2, mol/mol 0.0024 0.0016 0.0011 0.0011 0.0016 0.0017 0.0019
Rich H2S, mol/mol 0.068 0.23 0.21 0.20 0.18 0.24 0.30
Rich CO2, mol/mol 0.14 0.14 0.09 0.09 0.13 0.16 0.21
Total lean MDEA, gpm 325 700 800 820 925 695 550
Reboiler steam, lb/hr 19 500 42 000 48 000 49 200 55 500 41 700 33 000
Table 5
approximately 25% compared with occurring in the enrichment
Case 4a. absorber to occurring in the tail gas
Case 4c attempts to further opti- absorber, essentially converting
mise the MDEA system by utilising Case 4c into Case 3a. Case 4c tail
only the tail gas absorber bottoms gas absorber over-circulation
semi-rich solvent in the enrichment produces a lower tail gas absorber
absorber, using no lean MDEA for overhead H2S concentration, which
enrichment. Total lean solvent improves sulphur recovery slightly
circulation required for Case 4c is above Cases 4a and 4b. Case 4c uses
the same as that required for the lean solvent effectively, where it
enrichment absorber in Case 2, contributes most to improving
which represents a reduction in sulphur recovery efficiency, and
total circulation of about 20% from uses rich solvent effectively, where
Case 4b, and a reduction of approx- it can achieve the highest acid gas
imately 40% from Case 4a. loading, resulting in the best
Case 4c over-circulates the tail gas performance with the lowest total
absorber to provide enough solvent solvent circulation rate.
for absorption of H2S in the enrich-
ment absorber. If solvent flow is SRU/TGU train size comparison
reduced below this value, H2S Most SRU and TGU equipment
absorption abruptly shifts from sizes are primarily determined by
GAS 2010 15
18/3/10 11:42:38
 SRU / TGU train size comparison for the test cases
metering and loading facilities, plus
construction indirects, spares, engi-
neering and commissioning. Items
Case 1 2 3a 3b 4a 4b 4c
such as owner’s costs, land costs,
SRU tail gas, lbmol/hr 1969 1073 1022 1026 1136 1141 1142
TGTU feed gas, lbmol/hr 1969 1073 2125 2129 1784 1790 1831 escalation and contingency are
Incinerator process gas feed, lbmol/hr 1729 1466 1698 1700 1631 1637 1664 excluded. Operating costs are
discounted for each year of plant
Table 6 life and are added to capital cost to
form each case’s net present cost.
process gas volumetric flow rate. ment absorber flows directly to the The conventional SRU/TGU
Volumetric gas flow rates for the incinerator and does not flow arrangement of Case 1 and the
SRU, TGU and incinerator for each through the TGU. Such is not the conventional enrichment arrange-
of the test cases are shown in case for Case 3, which routes the ment of Case 2 serve as benchmarks
Table 6. total acid gas stream through the for comparisons between the other
Case 1 requires the largest SRU TGU before CO2 is slipped to the cases. Case 1, with no acid gas
train size due to the fact that it incinerator. enrichment, has the largest Claus
processes SRU feed gas with the The Case 4 alternatives allow for section, adding significantly to its
lowest H2S concentration. Case 1 a relatively small SRU train size capital cost. Case 2 requires a larger
also requires a large TGU, only (approximately 60% of the Case 1 equipment count than Case 1, with
slightly smaller than for the Case 3 size) due to the employment of its additional requirement of an
alternatives, which require the AGE upstream of the SRU. Case 4 enrichment absorber and corre-
largest TGU as a result of routing results in an intermediate TGU size, sponding pumps, yet the increased
the entire lean acid gas feed smaller than Case 1 and 3 but larger cost of the additional equipment
stream directly to the TGU along than Case 2, as a result of routing plus the larger MDEA unit is more
with the tail gas stream from the the enrichment absorber overhead than offset by the lower cost of the
SRU. to the TGU. smaller SRU/TGU achieved with
Cases 2 and 3 minimise the size enrichment in Case 2. Despite the
of the SRU train because these alter- Capital and operating cost fact that Case 2 has the lowest capi-
natives minimise CO2 flow through comparison tal cost of any of the cases, it suffers
the SRU. Case 2 achieves this by Table 7 shows estimated capital and from low efficiency in sulphur
slipping CO2 through the enrich- operating costs for each case. Rough recovery because it recovers no
ment absorber and Case 3 achieves order of magnitude capital costs are sulphur from the COS and
this by slipping CO2 through the provided on a US Gulf Coast basis mercaptan in the acid gas.
tail gas absorber. Case 2 also mini- and include engineering, procure- The Case 3 alternatives have a
mises TGU train size because the ment and installation of all process Claus section comparable in size
CO2 that is slipped from the enrich- and utility equipment, storage, and capital cost to the Case 2 Claus
section. However, the Case 3 TGUs,
with all of the acid gas CO2 and H2S
Cost comparison for test cases
flowing through them, are larger
than the TGUs for any of the other
Case 1 2 3a 3b 4a 4b 4c cases; thus, the Case 3 TGUs are the
Fuel
most costly of any of the cases.
Consumption, MMBtu/hr 44.2 29.7 39.5 39.0 37.0 37.5 37.7
Yearly cost, $ 2.26 1.52 2.02 2.00 1.89 1.92 1.92 Additionally, Case 3b adds a COS
Power hydrolysis reactor to the capital cost
Consumption, kW 502 513 630 637 646 590 556 of Case 3a, making it the most
Yearly cost, $ 0.27 0.28 0.34 0.35 0.35 0.32 0.30
costly option of all.
600 psig saturated steam
Export, lb/hr 18800 17 300 18300 18 300 18100 18 000 18 000 Compared with Case 3b, Case 4a
Yearly credit, $ -1.33 -1.23 -1.30 -1.30 -1.28 -1.27 -1.27 has slightly larger Claus and MDEA
50 psig saturated steam sections, and an additional absorber
Import, lb/hr 5300 34 800 38 800 39 900 46 800 32 600 23 700
for enrichment. However, Case 4a
Yearly cost, $ 0.31 2.06 2.30 2.37 2.77 1.93 1.40
Maint. and non-utility operating costs has a significantly smaller TGU
Yearly cost (10% TIC), $ 5.31 4.65 5.65 6.04 5.75 5.15 4.85 than Case 3b, because all of the acid
Cost summary gas H2S does not flow through the
Total yearly operation and maint. cost, $ 6.82 7.28 9.01 9.46 9.48 8.05 7.20
Case 4a TGU, and Case 4a does not
Total installed cost, $ 53.1 46.5 56.5 60.4 57.5 51.5 48.5
Net present cost, $ 97 94 115 122 119 104 95 require the COS hydrolysis reactor
that is present in Case 3b. Savings
1. Costs are reported in $ millions. with the smaller TGU offset slightly
2. Costs were prepared using the following information: $6.00/MMBtu LHV fuel; $0.064/kWh power;
$6.96/1000 lb 50 psig sat steam; $8.31/1000 lb 600 psig sat. steam; 355 operating days per year; increased Claus and MDEA costs,
15% annual rate of return; 25-year plant life making Case 4a’s capital cost
slightly lower than Case 3b’s.
Table 7 Therefore, Case 4a offers

16 GAS 2010 www.eptq.com

b&v.indd 10 18/3/10 09:33:19

 efficiency of sulphur recovery comparable to Case 1’s,
at a lower capital cost than for Case 3b.
Cases 4b and 4c are attempts to optimise Case 4a by
reducing the size of the MDEA system. Cases 4b and
4c have Claus and tail gas section capital costs similar
to those of Case 4a. The reduced MDEA system sizes
in Cases 4b and 4c reduce their capital costs. With
Case 4c, capital cost is reduced to less than 10% greater
than the cost of Case 2. Therefore, with Case 4c, good ,
cExpo Europe
sulphur recovery and reliable Claus operation are Visit us at Sto 172P
achieved from lean acid gas containing COS and Booth 169P &
mercaptans, with only slightly increased capital cost.
The lowest net present costs are for Cases 2, 1 and
4c. Of these three cases, only Case 1 and Case 4c
achieve good sulphur recovery when COS and
mercaptans are present in the acid gas feed. As was
mentioned previously, Case 1 possesses difficulties in
Claus operation caused by low acid gas concentration.
Therefore, Case 4c offers the best choice for good
performance with low cost. Cases 3a, 3b and 4a have
higher capital and operating costs and are not cost
competitive with Case 4c. Case 4b has a capital cost
competitive with Case 4c’s, but has a higher operating
cost because of its larger amine system.

Conclusions Membrane
Technology
AGE units operate well and, provided that the energy
consumption penalty for solvent regeneration can be
tolerated, offer a good sulphur recovery solution for
lean acid gases that can cause problems for conven-
tional sulphur recovery. When acid gas COS/mercaptan for Processes and Environment
levels are high enough to reduce sulphur recovery
with conventional AGE below an acceptable level, one BORSIG Membrane Technology GmbH, a mem-
of the Case 4 design configurations should be ber of the BORSIG Group, is an internationally
considered. successful company that offers its customers
Cases 4a, 4b and 4c accomplish AGE with recycle leading innovative membrane technology solu-
of unabsorbed enrichment absorber overhead COS/ tions for a wide range of industrial applications.
mercaptans to the TGU for conversion to H2S, and
effectively achieve high sulphur recovery with good The BORSIG Membrane Technology product
Claus operation. The solvent optimisation schemes in range:
Cases 4b and 4c, using tail gas absorber rich solvent Emission Control Systems such as
for further absorption in the enrichment absorber, Vapour Recovery Units
reduce capital and operating cost. Case 4c approaches Carbon Retrofit Units
the capital and operating cost of the conventional Vent Recovery Systems
SRU/TGU in Case 1, maintaining good sulphur recov- Product Recovery Systems
ery, but with improved Claus unit operation provided Gas Conditioning
by enriched acid gas. Liquid Separation
Services

Other products of the BORSIG Group:

Compressors and blowers
Pressure vessels and heat exchangers
Angela Slavens is Vice President and Sulphur Technology Manager with Boilers and power plant technology
Black & Veatch, Overland Park, Kansas. Industrial and power plant services
Email: slavensaf@bv.com
Justin Lamar is Sulphur Section Lead with Black & Veatch. BORSIG Membrane Technology GmbH
Email: lamarja@bv.com Bottroper Str. 279
Sara O’Dell is a Process Engineer with Black & Veatch. D-45964 Gladbeck / Germany
Email: odellsk@bv.com
Laura Francoviglia is a Process Engineer with Black & Veatch. Phone: +49 (0) 2043 / 4006-01
Email: francoviglial@bv.com Fax: +49 (0) 3764 / 4006-6299
e-mail: info@borsig-mt.com
www.borsig.de
www.ptqenquiry.com
for further information
www.eptq.com

b&v.indd 11 18/3/10 09:33:33

axens_Sail_PTQ.qxd 21/11/07 16:37 Page 1

Refining is
like sailing

Leading-edge technology, an experienced team

and operational efficiency keep you
ahead of the competition
• Refining technology
• Catalysts
• Services

The performance improvement specialists

www.axens.net
www.ptqenquiry.com For more information
for further information Paris +33 1 47 14 25 14 Houston +1 713 840 1133 Email information@axens.net

axens.indd 1 9/3/10 14:53:21

 An optimum line-up for sour gas
processing
Analysis of technical, economic and licensing considerations in the selection of
an optimum gas treating line-up recommends a choice of technologies

Saeid Mokhatab
Tehran Raymand Consulting Engineers

S
our gas fields brought into
production in the near future Sales
Gas Gas Mercaptans HC dew point gas
will contain considerable sweetening dehydration removal controlling
amounts of sulphur components. unit unit unit unit

Such gas poses challenges in adapt-

HP Enriched acid gas
ing processing technologies to meet separator C5 + cut from HC
environmental standards, improve dew point
controlling unit
reliability, and achieve higher Raw
Sulphur
Tail gas
natural gas treating unit
margins. from well recovery
unit
This article discusses five Off-gas to atm.
approaches to the integration of gas Off gas Incineration
treating processes in order to estab- Sulphur
degassing
lish the optimum treating line-up
for the design of sour gas process- Solid
Condensate
ing plants, taking all the process stabilisation Sulphur sulphur
Slug unit solidification
limitations into account within a catcher unit
Liquid sulphur
flexible, operable and economically storage
justifiable window. The conclusion Condensate
Condensate to export
is that the best line-up, with respect storage
to maximisation of treating train
operability and flexibility, is the Sour water Water
stripping unit treating unit
route employing molecular sieve
technology for gas dehydration and
mercaptan removal, with hydrocar- Figure 1 Setup of a gas processing plant
bon dew point control using a
propane refrigeration system. and facilities at a rate at which it Water, mercaptan (RSH) and the
A range of gas processing systems can be handled properly. heavy hydrocarbon components of
are available to provide products to Gas from the outlet of the slug the sweetened gas from the GSU
the specifications defined by plant catcher is directed to a high- should be reduced to meet sales gas
owners. This variety places a huge pressure (HP) separator, where final specifications. The gas pressure also
burden on the owner to select the separation of the liquid from the needs to be boosted by HP compres-
right technologies to fulfill an gas takes place. The HP raw gas sors before passing through the
optimised scheme that meets tech- flows through to the gas sweeten- sales gas meter and into the export
nological and economic targets. ing unit (GSU), in which acidic line.
Given the magnitude of the invest- components including H2S and CO2 Liquids from the reception facili-
ment in a gas processing plant, it is are removed by means of chemical ties are separated again into
appropriate to carry out a rigorous solvents. The enriched acid gas produced water and condensate.
selection study to identify the most from the GSU is processed to Produced water is treated and used,
cost-effective and appropriate treat- produce elemental sulphur in a while the condensate is combined
ment package. A general scheme of sulphur recovery unit (SRU), with hydrocarbon liquids knocked
a typical gas processing plant is consisting of a Claus unit, and an out in the dew point control unit
shown in Figure 1. Field produc- associated tail gas treating unit (DPCU). It is then processed in the
tion, on arrival at the processing (TGTU) if higher recovery rates are condensate stabilisation unit (CSU)
plant, is processed in a slug catcher, specified for the SRU itself. The to reduce the Reid vapour pressure
which captures liquid and allows it final residual gas from the TGTU is (RVP) and enable storage in atmos-
to flow into downstream equipment incinerated. pheric storage tanks. For sour

www.eptq.com GAS 2010 19

mokhatab.indd 1 18/3/10 09:42:06

 these blocks (or combinations of
blocks) can be achieved in a variety
2EFRIGERATION
UNIT
of ways; some technology options
may result in one or more of these
3ALES functions being achieved simultane-
2AWGAS '$5-25 0ROPANE #OMPRESSOR
GAS ously (thus reducing the complexity
'35 REFRIGERATION
MOLECULARSIEVE
SYSTEM
STATION of the design). As an example,
(EAVY
HYDROCARBON
molecular sieves can be used for
both dehydration and mercaptan
-ERCAPTANS removal, reducing the two steps
&UEL
325
GAS RICHGAS presented in the scheme to one.
REGENERATION
The water and mercaptans
desorbed from the bed during
3ULPHUR
regeneration can be captured by a
physical absorption process to
concentrate the mercaptans in a
Figure 2 Treating line-up 1 stream being sent to the SRU. The
bulk of the mercaptans might be
reduced in this way and will limit
the sulphur species being condensed
2EFRIGERATION in the different natural gas liquid
UNIT
(NGL) cuts. Instead of having to
3ALES use expensive caustic-based proc-
2AWGAS '$5-25 $0#5 #OMPRESSOR
GAS esses, molecular sieves or other
'35 MOLECULARSIEVE SILICAGEL STATION adsorbents could be used for the
(EAVY fine sweetening of the liquids.1
HYDROCARBON
However, operational problems and
-ERCAPTANS
upsets in the molecular sieve unit
&UEL
325
GAS RICHGAS can make that unit the bottleneck of
REGENERATION
the plant. Where a molecular sieve
unit is a bottleneck, extensive know-
how may be needed to improve its
3ULPHUR performance.2

Process treating line-ups

Figure 3 Treating line-up 2 Natural gas processing is often
thought of as a mature technology
with little opportunity for improve-
ment or innovation. However,
0ROPANE
REFRIGERATION changes in the requirements of
UNIT customers continue to drive
improvements in technology. Five
3ALES
GAS aspects of integration of the treating
2AWGAS '$5-25 *OULE 4HOMSON #OMPRESSOR
'35 MOLECULARSIEVE EXPANSION STATION processes involved in a sour gas
(EAVY processing plant are proposed here
HYDROCARBON in order to establish the best treat-
ing line-up. A typical scheme for
325
&UEL -ERCAPTANS
RICHGAS
each line-up is shown in Figures 2
GAS
REGENERATION to 6.
Line-ups 1 to 3 (see Figures 2 to
3ULPHUR
4) have the same gas treating
sequence but different technologies
for controlling the hydrocarbon
Figure 4 Treating line-up 3 dew point. In fact, in line-ups 1 to
3, the sweetened gas from the GSU
condensate coming in, if any treat- ing units shown in Figure 1 is is first routed to the gas dehydra-
ment is required for the heavier typical and illustrates the functional tion and mercaptan removal unit
sulphur components, complex blocks required to achieve the utilising molecular sieve technol-
treatment involving hydrogen is objectives of a plant designed to ogy, then to the hydrocarbon DPCU
necessary. produce pipeline gas from a sour utilising propane refrigeration,
The arrangement of gas process- gas feed. The function of each of silica gel technology and Joule-

20 GAS 2010 www.eptq.com

mokhatab.indd 2 18/3/10 09:42:15

 Feed gas composition
(EAVY
HYDROCARBON
Temperature, C 60 3ALES
Pressure, bara 70 2AWGAS GAS
'$5$0#5 -25 #OMPRESSOR
Flow rate, MMscfd 500 '35 PROPANEREFRIGERATION MOLECULARSIEVE STATION
H2O 1.8699
N2 5.1157
CO2 5.8770
H2S 5.8461
Methane 78.0451 -%' -ERCAPTANS
325 RICHGAS &UELGAS
Ethane 0.9596 REGENERATION
REGENERATION
Propane 0.3610
n-butane 0.1849 7ATER
i-butane 0.0948 3ULPHUR
n-pentane 0.0908
i-pentane 0.0905
C6 cut 0.1431
C7 cut 0.1960 Figure 5 Treating line-up 4
C8 cut 0.1718
C9 cut 0.1529
C10 cut 0.1333 cable for definitive design purposes. gas specifications and environmen-
C11 cut 0.0992 More details of undefined sulphur tal requirements.
C12 cut 0.5085 species were not available, hence, a
COS 0.0062
Methyl mercaptan 0.0272
few properties, such as MW = 120, BTX issue
Ethyl mercaptan 0.0032 boiling point (BP) = 182°C and ideal In accordance with the given feed
Isopropyl mercaptan 0.0010 liquid density = 837.7 kg/m3, have gas composition, there will be some
n-propyl mercaptan 0.0003 been assumed for the simulation. technical issues involving benzene,
sec-butyl mercaptan 0.0004
Undefined sulphur 0.0216
The process line-ups shown in toluene and xylene (BTX) in the gas
Figures 2 to 6 were modelled and dehydration and mercaptan
simulated so as to produce a treated removal unit utilising molecular
Table 1 gas stream to meet sales gas specifi- sieve technology (BTX co-adsorp-
cations (CO2 <1% mole, H2S <4 tion on the sieve and presence in
Thomson expansion technology, ppmv, mercaptan sulphur <10 the regeneration gas) in proposed
respectively. ppmv, total sulphur <100 mg/scm, line-ups 1 to 3, but this can be
Line-ups 4 and 5 (see Figures 5 water <1 ppmv and -10°C HC dew handled. Considering this, two
and 6) have the same gas treating point at 55 barg). options would be possible for the
sequence with different technolo- design of a gas dehydration and
gies for gas dehydration and Technical analysis mercaptan removal unit with
hydrocarbon dew point control. In Selecting the correct technology and molecular sieves:
line-up 4, the sweetened gas from tailoring the right process for a • Option I: complete water and
the GSU is first routed to the gas given application first requires tech- mercaptan removal with 4A, 5A
dehydration and hydrocarbon nical analysis that clearly indicates and 13X molecular sieves
DPCU using propane refrigeration which alternative would be techni- • Option II: complete water
accompanying methyethylene cally acceptable. Determining the removal with 4A, but bulk removal
glycol (MEG) injection, and then to best process selection depends on of light mercaptans (C1/C2 and
the mercaptan removal unit utilis- initial feed gas conditions, treated n-C3SH) with 5A molecular sieves.
ing molecular sieve technology.
Line-up 5 (see Figure 6) is the same
as line-up 4, wherein silica gel tech- 7ATER 2EFRIGERATION
nology has been applied for both (EAVY UNIT
gas dehydration and hydrocarbon HYDROCARBON
dew point control. 3ALES
2AWGAS GAS
'$5$0#5 -25 #OMPRESSOR
'35 SILICAGEL STATION
MOLECULARSIEVE
Input parameters used for the study
In order to make a practical
comparison of an integrated and
optimised treating configuration for &UEL -ERCAPTANS
325 RICHGAS
the design of a sour gas processing GAS
REGENERATION
plant, typical gas composition and
operating parameters for a sour
gas field have been used in this 3ULPHUR
study (see Table 1). The data given
here is for illustrative purposes and
should not be construed to be appli- Figure 6 Treating line-up 5

www.eptq.com GAS 2010 21

mokhatab.indd 3 18/3/10 09:42:25

 Option I would enable complete
sweetening of the gas, including
sweetening of the C5+ cut recovered
from the DPCU. On the other hand,
BTX competition with heavy
mercaptans on the 13X sieve should
be taken into account. The overall
design (bed size) would be higher
than for Option II, resulting in a
higher regeneration gas flow rate.
Assuming correct design, BTX will
not render RSH removal by
molecular sieve impossible. But co-
adsorbed BTX will be present in the
regeneration gas. As the most
common treatment of the regenera-
tion gas is a physisorption, it will be
captured too and will have to be
separated from the mercaptans before
their treatment in the Claus unit.
Option II will have the advantage
of no BTX co-adsorption with, addi-
tionally, a lower regeneration gas
flow rate. Most of the mercaptans
will be removed from the gas; the
heavier mercaptans will be knocked
down with the C5+ cut, which is
commonly treated with the liquid
from the condensate stabiliser
containing other sulphur species,
which should be addressed by
further treatment.
Liquid condensation on
molecular sieves
Depending on the given feed gas
composition, heavy hydrocarbon
liquids will not be formed in the
gas dehydration and mercaptan
removal unit (utilising molecular
sieve technology) to reduce the effi-
ciency of the molecular sieves beds.
In fact, hydrocarbon liquid conden-
sation on the sieves during the
regeneration procedure is a result
of poor design and not due to the
use of molecular sieve technology.
Considering this, there is no essen-
tial need to replace the mercaptan
removal unit with a hydrocarbon
DPCU.
Liquid carry-over to molecular
sieve unit
The major operational upset that
can befall the molecular sieve units
is carry-over of liquid from an
upstream unit. The cause of this
could be an upset in the gas supply
system or the result of excessive
foaming in the operation of the
22 GAS 2010
mokhatab.indd 4
GSU. Such carry-over of liquid
water and amine solution causes
rapid degradation of the molecular
sieve. During normal operation it
should be prevented by proper
operation of a well-designed
upstream separation vessel.
However, precautions can be
taken to protect the bed against
liquid carry-over during an upset
by installing a guard layer on top
of the molecular sieve bed. The
guard layer is intended to adsorb
most of the liquid water and amine
solution that could be deposited on
the molecular sieve bed during an
upset resulting in liquid carry-over.
Frequently alumina or silica gel-
based material is used, because they
are stable in liquid water and have
The major
operational upset
that can befall the
molecular sieve units
is carry-over of
liquid from an
upstream unit
good capacity for water uptake at
high water concentrations. Although
operational upsets are not expected
on a regular basis, the guard layer
must be prepared to handle them
during the lifetime of the molecular
sieve inventory.
Design flexibilities and issues
Drying cannot be separated from
mercaptan removal. If drying and
mercaptan removal are necessary,
the use of a drying option other
than molecular sieve technology
very often needs another processing
unit for the removal of mercaptans.
When the beds are regenerated,
water and mercaptans desorb into
the regeneration gas, which can
then be treated using a physical
solvent process to sweeten it and
provide a concentrated mercaptan
stream that can be sent to the SRU.
Removing mercaptans in this way
reduces the quantity of sulphur
species feeding forward into the
hydrocarbon recovery unit and thus
into the liquid products. This results
in a significant reduction in liquid
treating requirements and can open
up opportunities for novel technol-
ogy; for instance, the use of
molecular sieves instead of caustic-
based processes. Considering this,
process line-ups 1 to 3, which utilise
molecular sieve technology for
gas dehydration and mercaptan
removal simultaneously, would be
in appropriate designs, taking into
account the operating flexibility
over the feed gas range. However,
line-up 3 would not be attractive
unless the outlet pressure require-
ment was much lower compared to
an inlet that does not require gas
compression, or the inlet pressure
was too high, with an operating
range in the critical region.
The use of propane refrigeration
systems in line-up 1 for controlling
the hydrocarbon dew point requires
more operator attention and main-
tenance than the silica gel system
used in line-up 2. However, there is
greater flexibility in line-up 1,
where a branch of cold propane can
be used to keep the feed gas
temperature of both the gas dehy-
dration and mercaptan removal
units at 40°C or below, which is
required by most solid bed vendors.
The disadvantages of a silica gel
adsorption system for a hydrocar-
bon DPCU arise from the need to
regenerate, by heating not only the
silica gel bed to a suitable tempera-
ture (approximately 250°C) to drive
off contaminants, but also vessels
and pipework. This requires a rela-
tively larger heater, usually a fired
heater, which incurs large capital
and operating costs. Internal insula-
tion of the adsorber vessels may be
justified to reduce heater and fuel
gas costs. The need for at least two
and possibly more large, HP
adsorber vessels also makes for a
relatively high capital cost.
Therefore, while silica gel provides
a simple, reliable and flexible proc-
ess facility, which requires little
operator attention, it can carry a
relatively high cost.3
The molecular sieve regeneration
gas, rich in mercaptans, in line-up 1
is more than its value in line-up 4,
which increases operation difficul-
www.eptq.com
18/3/10 09:42:33
 ties as well as unit cost; however,
there are some issues in process
GASTO
line-up 4:
• The MEG and condensed water/
hydrocarbon mixture has a high
viscosity and needs to be heated to
ensure good separation. Even then,
uncertainty in composition and the
GAS
extent of hydrocarbon absorption in
RICHGAS ,URGI0URISOL
the MEG may cause operational
difficulties
• Glycol co-adsorbs some BTX that
ends up in the regeneration vapour
stream. In fact, because glycol
shows some affinity for hydrocar-
bons, the off-gas from the glycol
stripper often contains BTX
• The line-up is completely
dependent on propane refrigera-
tion. When the propane system
goes down, there is no dehydration.
In fact, the purpose of injected MEG
is not to dehydrate the gas but to
prevent the formation of hydrates
• There is considerable loss of
glycol caused by its solubility in
hydrocarbon condensate
• Sending rich MEG to the MEG
regeneration system adds to the
system’s cost
• Determining the appropriate
glycol injection rate and successive
injection points is a difficult task
that requires more operator atten-
tion to prevent hydrate formation
in the hydrocarbon DPCU.
Silica gel technology, as shown in
line-up 5, enables the single step
removal of both water and heavy
hydrocarbons from natural gas in
order to reach the dew point pipe-
line specifications for water and
hydrocarbons, because it requires a
lower regeneration gas compared to
molecular sieves due to its short
cycle operation. Untreated gas is
used for heating and cooling, and is
recycled to the inlet of the unit after
partially condensing the water and
heavy hydrocarbons. However,
adsorption times usually range
from 12 minutes up to 2–3 hours,
which on the other side reduces
lifetime expectations. This could
also increase the required instru-
ment air as well as the power
consumption by the air compressor,
which results in relatively high
operating costs. Other disadvan-
tages of silica gel technology are
fuel gas consumption, resulting in
www.eptq.com
mokhatab.indd 5
,IQUID
SULPHUR 3ULPHURDEGASSING
,URGI!QUISULF ,URGI#LAUS TREATINGUNIT
!CIDGAS
2AW
GAS '35
"!3&ACTIVATED #HILLER
-$%!
Figure 7 Overall arrangement of the Lurgi OmniSulf package
emissions of CO2 and NOx, and the
need for exhausted adsorbent to be
disposed of by an approved
method.
Process guarantees
Considering the natural gas process-
ing setup designed to produce
pipeline gas from a sour gas feed
(see Figure 1), there are four main
blocks, which usually means four
licensors. Therefore, up to four
licence contracts are need for each
of the four treating steps, including
guarantees and well-defined border
limits for the different interfaces.
Such a large number of package
vendors may give good reason for
failure to meet product specifica-
tions. If a specification is not
met in one unit, the other licensors
cannot necessarily meet their
liabilities.
Unit optimisation is another
concern. Since each licensor is
responsible for their own contribu-
tion, a certain security margin is
assumed. This takes into account
the combined strengths and weak-
nesses of each unit.4 To eliminate
these constraints, leading technol-
ogy licensors commonly take the
approach shown in line-ups 1 to 3
for optimal design of gas process-
ing plants (see Figure 7). With this
integrated concept, single-line
project responsibility (one licence
contract, one overall guarantee and
liability, one transparent licence
fee), optimised tailor-made design
and customer-oriented supply from
a single source result in considera-
/FF
%NRICHEDACIDGAS
ATMOSPHERE
325 ,URGITAILGAS )NCINERATION
-ERCAPTANS
RECOVERYUNIT
0URISOLOFF
23(
'ASDEHYDRATION
$RY
-ERCURY SWEETGAS
MERCAPTAN REMOVALUNIT
REMOVALUNIT
MOLECULARSIEVES
OPTIONAL
2EGENERATIONGAS
ble savings in investment and
operating cost, with enhanced flexi-
bility across the whole plant.
Technical conclusion
In configuring an optimum flow
scheme, the designer must under-
stand the technology options
available, their opportunities for
integration and their limitations.
Technical risk, licensor experience,
degree of commercialisation, safety,
and health and environmental
factors all need to be weighed up
along with the performance of the
technologies concerned. Considering
these points, process line-ups 1 to 3
should be selected for designing a
sour gas processing plant; other-
wise, the overall concept will not be
optimised. However, to select the
optimum treating line-up, a
technical-economic analysis of the
proposed approaches to the design
of a hydrocarbon DPCU should be
carried out.
The most important technical
characteristics of any hydrocarbon
dew point controlling process are
the feed gas pressure and permissi-
ble unit pressure drop.5 Considering
this, and in accordance with the
given feed-gas pressure as well as
the high pressure drop resulting
from the refrigeration (Joule-
Thomson expansion) process
applied in treating line-up 3, this
method would not be attractive
if additional compression was
required. However, this technology
should also be compared, from an
economic point of view, with other
GAS 2010 23
18/3/10 09:42:42
 of technical options and arrive at
Summary of performed economic analyses for five proposed treating line-ups the best solution for the project.

Total capital investment, $ Electricity cost, $ Fuel gas cost, $ Present value, $
Line-up 1 85 465 391 3 103 664 1 083 852 121 165,145
Line-up 2 98 860 415 2 135 549 1 652 778 134 554 551
Line-up 3 92 973 363 4 022 459 1 083 852 134 821 975
Line-up 4 72 137 901 3 014 683 541 926 102 159 991
Line-up 5 98 860 415 2 135 549 1 625 778 134 489 366
Table 2
hydrocarbon dew point controlling
methods.
It should also be noted that the
proposed propane refrigeration
system can be designed with two
and/or three levels of refrigeration
within the refrigeration cycle.
Determination of the refrigeration
level is based on the minimum
temperature required in the proc-
ess. In this study, the final
temperature required in the process
(to meet the sales gas dew point
specification) and the inlet feed
temperature to the DPCU should be
-21°C and 40°C, respectively.
Assuming these temperature
requirements and providing an
adequate approach temperature
(about 7ºC) between propane refrig-
erant and the inlet feed temperature,
two cases could be realised.
It should be noted that in the case
of two levels of refrigeration (Case
1), cooled process gas is used to
keep the feed gas temperature of
the gas dehydration and mercaptan
removal unit at 40ºC, while in the
case of three levels of refrigeration
(Case 2) propane is used for this
purpose. However, in Case 1, the
cooled process gas is not available
in start-up mode, where relevant
feed enters the molecular sieve beds
above 40°C. This can seriously
damage the existing beds. Hence,
Case 2 is recommended as the
preferred propane refrigeration
cycle.
Economic analysis
Five options for the integration of
treating units in a sour gas process-
ing plant have been proposed.
However, in view of the complexi-
ties involved, identification of the
optimum scheme requires careful
economic evaluation of the alterna-
tives, on a case-by-case basis.
Based on the main equipment list
and the projected treated gas speci-
fications, Aspen Icarus was used for
economic analysis of the five
proposed alternatives for a sour gas
processing plant. It should be noted
that gas sweetening, sulphur recov-
ery and compressor station units
are common and in the same
sequence in all of the proposed
treating line-ups; therefore, they
have not been considered in this
evaluation. Table 2 shows a
summary of economic analyses for
the proposed treating line-ups 1 to
5. This economic analysis was
budgetary and only used for the
purpose of comparing the relative
Identification of the
optimum scheme
requires careful
economic evaluation
of the alternatives,
on a case-by-case
basis
costs of the alternatives. Lower
present value indicates lower
investment cost.
Considering a margin of error of
±15%, the results of the economical
evaluations, which are in the same
order of magnitude, should not
alone be considered as good criteria
for selecting the right line-up of
technology for the project’s circum-
stances. Attention should also be
given to the technical evaluations
as well as to some of the important
factors (technical risk, licensor expe-
rience, degree of commercialisation,
and so on) that enable the designer
to navigate through the minefield
Conclusion
Five process treating line-ups have
been investigated for the design of
gas processing units in a sour gas
processing plant. With the vast
array of technical opportunities
available, the designer must care-
fully navigate to an optimal process
flow scheme by meeting technologi-
cal and economical targets. Among
the alternatives, integrated solu-
tions presented in line-ups 1 to 3,
which are commonly used by lead-
ing licensors, would be the best
line-ups for designing gas process-
ing trains in a sour gas processing
plant, considering technical require-
ments alone. However, considering
the results of economic analysis and
in accordance with the design flexi-
bilities of the propane refrigeration
system, line-up 1 is recommended.
Selecting the best alternative is a
complex affair, sensitive to a range
of parameters each of which merits
careful attention. For instance, the
plant owner’s preferences may
come into play, which may deter-
mine the project’s final outcome or
impose restrictions on the treating
line-up.
References
1 McMahon D, Navigating the technical
minefield of gas processing options, GPA
Europe Meeting, London, 26–27 Feb 2004.
2 Bruijn J N H, Huffmaster M A, van de Graaf
J, Grinsven P F A, Grootjans H, Maximizing
molecular sieve performance in natural gas
processing, 81st Annual GPA Convention,
Dallas, TX, 11–13 Mar 2002.
3 Finn A, Tomlinson T, A case for dehydration,
Hydrocarbon Engineering, Dec 2007.
4 Mokhatab S, Meyer P, Selecting best
technology lne-up for designing gas processing
inits, GPA Europe Sour Gas Processing
Conference, Barcelona, 13–15 May 2009.
5 Mokhatab S, Poe W A, Speight J G, Handbook
of Natural Gas Transmission & Processing,
1st Edition, Gulf Professional Publishing,
Burlington, MA, 2006.
Saeid Mokhatab is Process Technology
Manager for Tehran Raymand Consulting
Engineers, Iran. He has been involved as a
technical consultant in several international
gas engineering projects and has published
more than 150 academic and industry papers
on related topics.
Email: saeid_mokhatab@hotmail.com

24 GAS 2010 www.eptq.com

mokhatab.indd 6 18/3/10 09:42:53

 PROCESS INSIGHT
Optimizing CO2 Capture, Dehydration and Compression Facilities
The removal of CO2 by liquid absorbents is widely implemented
in the field of gas processing, chemical production, and coal gasification.
Many power plants are looking at post-combustion CO2 recovery to meet
environmental regulations and to produce CO2 for enhanced oil recovery
applications. The figure below illustrates actual data of fuel consumption in
2005 and an estimate of energy demand for various fuels from 2010 to 2030.
The world energy demand will likely increase at rates of 10–15% every 10
years. This increase could raise the CO2 emissions by about 50% by 2030
as compared with the current level of CO2 emissions. The industrial countries
(North America, Western Europe and OECD Pacific) contribute to this jump in
emissions by 70% compared to the rest of the world, and more than 60% of
these emissions will come from power generation and industrial sectors.
formulated solvent without implementing any split flow configurations. This is
much less than the reported steam usage for the MEA solvent.
The design of a facility to capture 90% of the CO2 from the flue
gas of a coal fired power plant is based on the specified flue gas conditions,
CO2 product specifications, and constraints. Using the ProMax® process
simulation software from Bryan Research & Engineering, CO2 capture units
can be designed and optimized for the required CO2 recovery using a variety
of amine solvents. The following figure represents a simplified process flow
diagram for the proposed CO2 Capture Plant.

Despite the strong recommendations from certain governments, The table below presents the main findings for CO2 capture from the
there are very few actual investments in CO2 capture facilities geared toward coal fired power plant and the NGCC power plant, each designed to produce
reducing greenhouse gas emissions mainly because of the high cost of CO2 about 3307 ton per day (3,000 TPD metric). To produce the same capacity of
recovery from flue gas. CO2 capture costs can be minimized, however, by CO2, only one train with smaller column diameters is required in the case of
designing an energy efficient gas absorption process. Based on the findings the coal power plant and two trains with larger column diameters are required
of recent conceptual engineering studies, HTC Purenergy estimated the in the NGCC Power Plant case. This is mainly due to processing a larger
production cost to be US$ 49/ton CO2 (US$ 54/ tonne CO2) for 90% CO2 flue gas with lower CO2 content in the NGCC power plant. Consequently, a
recovery of 4 mole% CO2 content in the flue gas of NGCC power plants. A substantial reduction in the capital and production cost was reported for the
separate study showed the cost for 90% CO2 recovery of 12 mole% CO2 from coal fired power plant CO2 recovery facility.
a coal fired power plant to be US$ 30/ton CO2 (US$ 33/tonne CO2). The cost
of CO2 recovery from coal power plant flue gas is substantially less than that
of NGCC power plant flue gas due to the higher CO2 content in the feed.
The energy efficiency of a CO2 capture plant depends primarily on
the performance of the solvent and optimization of the plant. In traditional flue
gas plant designs, MEA was the primary solvent and was limited to 20 wt% to
minimize equipment corrosion. Recent developments in controlling corrosion
and degradation has allowed an increase in the solvent concentration to
about 30 wt% thus decreasing the required circulation and subsequent steam
demand. A recent DOE study shows the steam consumption for an existing
CO2 plant using 18 wt% MEA (Kerr McGee Process) is 3.45 lb of steam per
lb of CO2 for amine regeneration. A modern process that uses 30 wt% MEA
is expected to use 1.67 lb of steam per lb of CO2 for amine regeneration. The
HTC formulated solvent is a proprietary blend of amines and has a lower
steam usage than the conventional MEA solvent. Based on the material and
energy balances for the plant designed in the recent study, the reboiler steam For more information about this study, see the full article at
consumption is estimated at about 1.47 lb steam/lb CO2 using the proposed www.bre.com/support/technical-articles/gas-treating.aspx.

Bryan Research & Engineering, Inc.

P.O. Box 4747 • Bryan, Texas USA • 77805
979-776-5220 • www.bre.com • sales@bre.com
www.ptqenquiry.com
for further information

ProcessInsight 2010MAR9.A4.rev6.indd 1 3/11/2010 9:09:12 AM

bre.indd 1 17/3/10 14:20:22
 www.ptqenquiry.com
for further information

albemarle.indd 1 17/3/10 14:22:36

 Detecting and dealing with hydrate
formation
A review of cause, detection and treatment of hydrate plug formation in
gas pipelines
Bahubali Chandragupthan PL Engineering
Girish Babu Nounchi L&T - Gulf
G
as hydrates are solid crystal-
line compounds, which have
a structure wherein guest
molecules are entrapped in a cage-
like framework of the host
molecules without forming a chemi-
cal bond. It is a result of the
hydrogen bond that water can form
hydrates. The hydrogen bond
causes water molecules to align in
regular orientations. The presence
of certain compounds causes the
aligned molecules to stabilise, and a
solid mixture precipitates. The
water molecules are referred to as
the host molecules, and the other
compounds, which stabilise the
crystal, are called the guest mole-
cules. The hydrate crystals have
complex, three-dimensional struc-
tures in which the water molecules
form a cage, and the guest mole-
cules are entrapped in the cages.
The stabilisation resulting from
the guest molecule is postulated to
be caused by Vander Waal’s forces,
an attraction between molecules that
is not a result of electrostatic attrac-
tion. The hydrogen bond is different
from the vander Waal’s force
because it is due to strong electro-
static attraction, although some
classify the hydrogen bond as a
Vander Waal’s force. Another inter-
esting thing about gas hydrates is
that no bonding exits between the
guest and host molecules. The guest
molecules are free to rotate inside
the cages built up from the host
molecules. This rotation has been
measured by spectroscopic means.
No hydrate without guest molecules
has been found in nature. Thus,
clathrates are stabilised by the weak
attractive interactions between guest
and water molecules. However, the
www.eptq.com
l&T.indd 1
guest species has some restrictions
on its size. This arises from the fact
that there are a limited number of
cage types that encapsulate guest
molecules without deviation of the
hydrogen bond lengths and angle
from ideal ones. All of the cages are
not necessarily dependent on the
temperature and the pressure of the
guest compound in equilibrium with
clathrate hydrate.
There is no single
indicator that gives
the best warning of
hydrate formation,
but pressure drop is
the most common
indicator
The formation of a hydrate
requires three conditions: low
temperature and high pressure; the
presence of hydrate formers such as
CH4, C2H4, CO2 and H2S; and suffi-
cient quantities of water and
formation time.
Early warning signals
There is no single indicator that
gives the best warning of hydrate
formation, but pressure drop is the
most common indicator. Hydrates
in a well are announced most often
by abrupt flow blockages accompa-
nied by a high pressure drop. In
normal operation, however, the
well’s temperature is high enough
to prevent hydrate formation:
• Pigging returns should be exam-
ined carefully for evidence of
hydrate particles. Hydrate masses
are stable even at atmospheric pres-
sure (metastable equilibrium) in a
pig receiver discharges.
• If water arrival decreases appre-
ciably at the separator, hydrates
may be forming inline. Several field
trials have indicated that the earli-
est sign for hydrate formation in a
pipeline is a decline or stoppage of
the production of water. An
increase of water hold-up in the
pipeline is due to the formation
of non-transportable hydrates.
Unfortunately, the water produc-
tion rate is not accurately and
continuously monitored enough,
and therefore this early warning
sign is often not noticed.
Hydrates denude H2S from natu-
ral gases owing to the near optimal
fit of H2S in hydrate cavities. The
same is not true of other acid gases
and carbon dioxide.
The pressure drop (∆P) increases
and the flow rate decreases if the
pipe diameter is decreased by
hydrate formation at the wall in a
gas line. Since ∆P in pipes is
proportional to the square of turbu-
lent flow rates, the change in flow
with hydrates present can be
substantial; however, a large restric-
tion may be necessary over a long
length before a substantial pressure
drop occurs.
• In some cases, the pressure drop
over the line increases gradually
because of a build-up of hydrate
layer at the pipeline wall. This
behaviour has been observed in
gas/condensate lines that operate
in the annular flow regime at a
relatively low water cut. Usually
GAS 2010 27
18/3/10 11:43:54
 the increase in pressure drop is
accompanied by a decreasing gas
production rate.
Austvik suggests that while a
gradual pressure increase in hydrate
formation occurs for gas systems, a
gradual pressure increase is not
typical for gas/oil/condensate
systems. The flow regime of a
system may change due to the
formation of viscous hydrate slur-
ries. The flow in a system that
normally operates in the annular
flow regime may change to slug
flow because of hydrate formation.
This may result in a severely
fluctuating pressure drop over
the line after an initial period
of gradually increasing pressure
drop.
• If a pipeline operates in the strati-
fied or slug flow regime, hydrate
formation may cause large fluctua-
tions in the pressure drop. This
behaviour is due to the formation
of an increasingly viscous slurry of
hydrates that intermingles with
water, gas and oil. The continuous
formation and subsequent disper-
sions of slugs of this viscous slurry
initially cause variable and tempo-
rary restrictions in the pipeline
(although the bare line pressure
drop often increases continuously).
Once these fluctuations are
observed, a sudden jamming of the
flow line by a slug of hydrate slush
may be imminent. In gas/oil/
condensate systems, Statoil’s expe-
riences are that, without advance
warning, the line pressure drop
shows sharp spikes just before
blockages occur.
• Dead ends, tees and narrow
tubing are easily blocked by
hydrate. Pressure transducers
connected to a pipeline via narrow
tubing may give erratic readings
due to hydrate blockage in the
instrument lines before notable
hydrate restrictions are formed in
the pipelines itself.
• Patches of hydrate may be heard
to move through flow line or
production facilities. If these
hydrates dissociate in heat-traced
or low-pressure facilities (for
instance, heated separators) the
temperature may notably decrease
because of heat consumption by
dissociating hydrates.
28 GAS 2010
l&T.indd 2
Detection of hydrate blockages
Two methods are suitable for detec-
tion of hydrates in onshore systems:
thermal imaging and gamma ray
detection. A thermal imaging
camera is a handheld device that
measures infrared spectral trans-
mission as an indicator of system
temperature. It is applicable only to
onshore or offshore topsides.
The camera detects temperature
variations in the system and is very
sensitive to pipe coatings, variation
in wall thickness and pipe rough-
ness. A gamma ray densitometer
with temperature sensing employs
an emitter and a sensor on external
pipe walls. The transmission of
gamma rays to the sensor is a func-
tion of the density of the pipe’s
contents.
The differences in density of the
water and hydrates are very small
so the gamma ray densitometer
Two methods are
suitable for detection
of hydrates in
onshore systems:
thermal imaging
and gamma ray
detection
alone cannot differentiate between
water and hydrate. However, if the
equipment is used in combination
with temperature measurement
downstream of the densitometer,
hydrate formation can be identified.
Hydrates are indicated by a low
temperature (the Joule-Thomson
effect) and an increase in density,
whereas the water temperature is
similar to that of gas. A high-
density, low-temperature mass in
the pipeline is likely to be hydrate,
whereas a high-density plug with-
out a temperature drop is probably
water.
Thermal imaging and gamma ray
techniques should first be consid-
ered for offshore systems, but there
are additional methods to deter-
mine hydrate blockage locations in
offshore:
• In the event of a hydrate block-
age in offshore flow lines, it is
common to fill the line with an
inhibitor, particularly when the
blockage is close to the platform.
The inhibitor’s injection volume
enables the determination of a
blockage’s location relative to the
platform. However, given the size
of the line and the effects of liquid
retention within the pipeline, in
most cases this method is
ineffective.
• Reducing or cycling the line
pressure.
• Measuring internal pressure
through external sensors.
• Pressure location techniques:
■ Pressure reduction This simple
technique takes advantage of
hydrate’s porosity and permeability
to gas by decreasing the down-
stream pressure and monitoring the
rate of downstream pressure recov-
ery, as well as the rate of pressure
decrease of the upstream pressure
of the plug.
■ Back pressurisation One method
to locate a complete pipeline block-
age is to measure the pressure
increase as metered amounts of gas
are injected at the platform. The
rate of pressure increase is propor-
tional to the rate of gas input,
enabling determination of the
piping length between the platform
injection point and the blockage.
■ Pressure fluctuations Pressure
pulse travel time and pressure
frequency response methods can be
used to locate a hydrate blockage.
Both methods involve the measure-
ment of sound waves’ travel time
or frequency changes from the plat-
form to the blockage. However,
these analyses have not been
successful to date because of two
factors: acoustic response is a func-
tion of the relative amounts of gas
and liquid present, which are
usually unknown; and reflected
pulses are dampened by walls,
valves and bends.
Locating a hydrate plug
There is no perfect and precise
method for locating a hydrate plug,
so a combination of methods gives
the best results. In an above-ground
line onshore, a thermo camera is
recommended for locating a hydrate
www.eptq.com
18/3/10 11:44:04
 plug. For buried or subsea lines, the
following procedure applies:
• Estimate the hydrate formation
temperature and pressure of the
blockage relative to the conditions
of the pipeline. Use a simulation to
determine where the contents enter
the hydrate formation pressure and
temperature regime during the
normal operations.
• Depressurise the pipe down-
stream of the plug to approximately
two-thirds of the pressure between
the normal operating pressure and
the hydrate formation pressure. Do
not decrease the pressure on one
side of the plug below the hydrate
formation pressure. Monitor the
pressure increase downstream and
the pressure decrease upstream for
24 hours or until a significant pres-
sure change is obtained at each
point. Use the rate of pressure
decrease upstream and increase
downstream to determine the centre
point of the plug.
• In a subsea system, fill the riser
with glycol to determine the
distance between the platform and
the plug. This may be inaccurate
www.eptq.com
l&T.indd 3
because of the distance to the plug,
pipeline elevation changes and the
partial line filling with condensate.
• Back pressure the pipeline and
monitor the pressure increase for a
measured volume of gas input.
Estimate the downstream distance
from platform to plug by the rate of
pressure change relative to gas
input for a given compressibility
and simulated liquid retention
volume. Use this technique with
step two to determine pipe volume
before and after the plug.
• If possible, use a mechanical
device to determine the plug’s
location.
Safety
There are generally three safety
rules to be considered while
handling hydrate blockages. They
are:
• Always assume the presence of
multiple hydrate plugs; there may
be high pressure between the plugs.
• While attempting to move
hydrate plugs, their velocity and
mass can cause two types of failure:
damage to orifices, flanges, valves,
www.ptqenquiry.com
for further information
bends and elbows; and pipe rupture
downstream of the hydrate caused
by extreme gas compression.
• Thermal treatment is not applica-
ble for underground pipelines and
subsea pipelines. Any heating
should be done from the end of the
plug instead of the middle.
Gradually depressurising a plug
from both sides is normally
recommended.
Methods of hydrate prevention and
removal
Chemical inhibition
There are two types of inhibition:
thermodynamic and kinetic.
Thermodynamic hydrate inhibitors
are added to a two-component
system (water and gas) that changes
the energy of intermolecular inter-
action and the thermodynamic
equilibrium between molecules of
water and gas. There is a relation-
ship between the concentration of
an inhibitor in water and the
temperature of hydrate formation.
These chemicals work by altering
the chemical potential of the aque-
ous phase such that the equilibrium
GAS 2010 29
18/3/10 13:23:54
 dissociation curve is displaced to
lower temperature and higher pres-
sure. Electrolytic solutions and
alcohols are used as thermody-
namic inhibitors.
The chemical potential of a
constituent in a mixture is the
increase in the free energy that
takes place at constant temperature
and pressure when one mole of that
constituent is added to the system,
keeping the amounts of all other
constituents’ constant; one mole of
the constituents is added to such a
large quantity of the system that its
composition remains almost
unchanged. The chemical potential
of a constituent in a mixture is its
contribution per mole to the total
free energy of the system of a
constant composition at constant
temperature and pressure.
Kinetic inhibitors do not shift
the hydrate’s equilibrium condi-
tions. Rather, they decrease the rate
at which hydrates form, prevent-
ing plugs for a period longer than
the free water residence time in
a gas line. They are predomi-
nantly water soluble and dosed at
concentrations ranging between
2000 and 5000 ppm in water. They
adsorb on the surface of hydrate
microcrystals and dispersed
droplets of water, and sharply
change the diffusive-sorptional
exchange at the inhibitor-water
interface. This decreases the rate
of microcrystal growth, their
coagulation, sedimentation and
adhesiveness, thus preventing the
formation of large gas hydrate
plugs in wells/pipelines.
Kinetic inhibitors do not prevent
hydrate formation, but shift the
time and space involved in the
formation of large hydrate plugs.
They must be very soluble in water,
should not hydrolyse to form insol-
uble compounds and adsorb well
on a polar surface of hydrate micro-
crystals to form an external surface
that prevents association of hydrate
crystals. Fatty acids, mixtures of
fatty alcohols and amines can be
used as kinetic inhibitors; for
instance, poly-n-vinyl amides, poly-
n-vinyl-n-methyllaptamide and
poly-n-vinylpyrolidone (PVP). PVP
is injected in an alcohol or aqueous
alcohol solution into an oil or bonding among the water mole-
30 GAS 2010
l&T.indd 4
gas flow to form a 0.5% solution.
Used in multiphase, gas and
condensate flows, kinetic inhibitors
have the advantages of low cost of
application, low toxicity and are
proven in gas systems.
Alcohols
Methanol is widely used as a
hydrate inhibitor, but it has some
disadvantages; for instance, in LPG
pipelines. LPG is made up largely
of propane and mixed butanes and
the addition of methanol can form
propane/methanol and n-butane/
methanol azeotropes at unaccepta-
ble levels. Some corrosion inhibitors
added to pipelines are alcohol-
based; when methanol is added as
a hydrate preventer it may dissolve
the corrosion inhibitor, leading to
corrosion problems.
Methanol is usually stored in
tanks that are open to the atmos-
Kinetic inhibitors
have the advantages
of low cost of
application, low
toxicity and are
proven in gas
systems
phere. This allows some air to
dissolve in the methanol. Typically,
the amount of dissolved oxygen is
small, but in the long term this
could cause problems:
• Methanol is a volatile substance;
we have to pay attention to vapour
loss. In practical terms, this means
that more inhibitor must be injected
than the amount required, mitigat-
ing the hydrate formation
theoretically.
• Hydrate inhibition abilities are
less for larger alcohols; ie, methanol
> ethanol > iso-propanol.
• Safety issues (poison, volatility)
and inventory management.
Methanol, which is small mole-
cule, does not form a hydrate
because it is hydrogen bonded and
hence interferes with the hydrogen
cules. In addition, methanol is
highly soluble in water. Gases that
are highly soluble in water, such as
NH3 and HCl, do not form hydrates
even though their size might lead
one to believe that they should.
Methanol is seldom used to dissoci-
ate a hydrate plug unless the
injection point is vertically above a
hydrate plug (as in riser or a well).
Methanol is normally used for flow
line plugs. Most frequently than
methanol injected in to the pipe-
lines as a loss, although in gas
processing methanol recovery is
viable in turbo expander plant.
Methanol is fully miscible with
water, while the solubility of meth-
anol in hydrocarbons is very small.
Therefore, water can be used to
extract methanol from the hydro-
carbon condensate efficiently. Since
water is the solvent, this extraction
process is called water wash. More
than 96% of methanol injected can
be recovered, if good engineering
judgement and experience are
applied.
Methanol is less expensive per
gallon than monoethylene glycol
(MEG), although its price varies
significantly. The lower molecular
weight of methanol (32 vs 62 for
MEG) allows more dilute methanol
solutions to be used since hydrate
temperature. Depression is a molar
phenomenon. However, the amount
of methanol in the gas phase must
also be considered when calculating
the methanol injection rate. No
solid waste is generated in metha-
nol recovery processes. In some
overseas locations, such as the
North Sea, the methanol concentra-
tion of water discharged is limited
to 1000 ppm maximum.
Depending on operating condi-
tions, the solubility loss of methanol
into sales gas can be very high,
typically 1lb of methanol/1 MMSCF
for every percentage of methanol in
the water phase. Losses to the
liquid hydrocarbon are higher than
for MEG, but usually less than
1–2% of the hydrocarbon volume.
Depending on solubility losses,
chemical makeup requirements for
methanol can be very large and
expensive for both once-through
systems and methanol recovery
units.
www.eptq.com
18/3/10 13:24:20
 Ammonia
Ammonia is more than twice as
effective as methanol on a weight
basis, but is seldom used due to the
reaction of NH3 with CO2 in the gas
to form solid plugs of ammonium
carbonates. The handling of ammo-
nia also carries safety considerations.
Glycols
Glycols offer more opportunities for
hydrogen bonding with water than
alcohols because of their additional
hydroxyl group. Glycols generally
have higher molecular weights,
which inhibit volatility. The
commonly used glycols are ethyl-
ene glycol, triethylene glycol,
propylene glycol and polyalkylene
glycol. Alcohols and glycols form
hydrogen bonds with water mole-
cules and make it difficult for the
water molecules to participate in
hydrate formation. The higher
molecular weights of glycols cause
them to remain in the liquid phase
so they are more recoverable than,
say, methanol. In view of vapouri-
sation losses of glycols, Toensend
and Ride recommend that ethylene
glycol be used where system
temperatures are lower than 18°C
and that diethylene glycol be used
above that temperature. They also
suggest that five other factors
should be considered for the use of
a glycol injection system:
• The presence of salt is a serious
obstacle because of the risk of foul-
ing and corrosion.
• The pH of solutions should be
controlled at 7 to avoid corrosion.
• Glycol coats all surfaces where
hydrates form because it is more
difficult to atomise.
• Glycol-water solutions freeze
when they are too dilute or too
concentrated.
• Nelson indicates that at 40°C the
viscosity of 80 wt% of methanol
was only three times that of water
at ambient conditions, while the
viscosity of ethylene glycol is 480
times that value for the same condi-
tion. To ensure flow through valves,
maintain the viscosity at 100–150
centipoise.
• Dew point reductions of 35–40°C
without gas stripping and 50–60°C
with gas stripping are obtainable.
At contactor temperatures above
www.eptq.com
l&T.indd 5
60°C, tetraethylene glycol should be MEG, if present in sufficient quanti-
considered. ties relative to the volume of
The price of MEG is fairly stable. condensate or oil, inhibits paraffin
The controlling factor is the product deposition by forming an aqueous
of chemical cost times gallons continuous emulsion phase. For a
consumed, thus chemical losses glycol system, there is virtually no
rather than price per gallon have loss to the gas phase, and the solu-
the greatest bearing on chemical bility of glycol in hydrocarbon
costs. Solubility losses of MEG into liquids is typically 1 to 2 orders of
the sales gas are negligible, and loss magnitude less than methanol.
to the liquid hydrocarbon phase is Since salt is non-volatile, it will
very low. MEG has paraffin inhibi- remain in the lean glycol during
tion/dispersion properties, while regeneration as the water is boiled
methanol exhibits little inhibition. off. Unless glycol is reclaimed by
Increase Performance and Profits
with MPR Services’ World Wide
Amine Management!
Today, there’s a better, more cost efficient and environmentally friendly
way to clean amine systems and control heat stable salts (HSS). Around
the world, leading refineries, gas plants, ammonia plants and synthesis
gas plants, depend on our patented HSSX® process to clean millions of
gallons of amine solutions. HSSX is our patented proprietary technology
that provides the operational advantages you are looking for:
t
Non-hazardous, biodegradable effluent is suitable for discharge to your Waste
Water Treatment Plant
t
Continuously cleaning amine systems on line with our Amine Shield™ Slip Stream
Unit (permanent unit installation). Efficiently removes suspended solids (SSX™),
hydrocarbons (HCX™) and heat stable salts (HSSX) providing the clean amine your
system was designed to operate.
t
Periodic amine system cleanings off-line or on-line using our mobile services
t
Low amine losses save on replacement amine processes
t
HSSX technology can reduce water and chemical requirements by 50% or more
For more information, call 281-377-7424
or visit our website at www.mprservices.com
MPR Services, Inc., A Tessenderlo Kerley, Inc. Company.
1201 FM 646, Dickinson, TX 77539-3014 U.S.A.
5FMFQIPOF

t
'"9

E-mail: info-mprservices@tkinet.com
NQSFVSPQF!ULJOFUDPN
t
NQSNJEEMFFBTU!ULJOFUDPN
Contact us now to schedule The HSSX process at your location.
HSSX is a registered trademark and SSX and HCX are trademarks of MPR Services.
©2010 Tessenderlo Kerley, Inc.
www.ptqenquiry.com
for further information
GAS 2010 31
18/3/10 13:24:46
 removing the salt, serious plugging
of equipment and flow lines can
result. Depending on the magnitude
of the contamination problem,
solvent reclamation can be accom-
plished either by continuous or
intermittent removal from the total
circulation stream or a partial
slipstream.
Salts
Concentrated brines are known to
be very good inhibitors of gas
hydrate formation. They are also
called inert inhibitors; they do not
enter the gas hydrate phase, but
influence the thermodynamic equi-
librium through their effect on
water activity. The brine ions inter-
act with the dipoles of water
molecules with a bond that is much
stronger than Vander Waal’s forces,
leading to clustering around the
polar solute molecules. This cluster-
ing also causes a decrease in the
solubility of potential hydrate mole-
cules in the water (a phenomenon
known as salting out) as a second-
ary effect. These effects combine to
require substantially more sub-
cooling to cause hydrate to form.
Antiagglomerants
These are surface-active chemicals
(surfactants) that do not act to
prevent hydrate crystals from form-
ing, but rather prevent them from
agglomerating to form hydrate
plugs. Two types are available:
water-soluble antiagglomerants are
used only if the produced water
contains some salt; and oil-soluble
antiagglomerants also work if
fresh (condensed only) water is
produced.
The antiagglomerant works by
emulsifying hydrates in the hydro-
carbon liquid. Hydrates are carried
as a non-agglomerated slurry, with-
out a viscosity increase up to 50%
water cut. Thus, they do not display
the pressure-temperature limita-
tions of kinetic inhibitors. However,
antiagglomerants used to date only
work in systems with a continuous
hydrocarbon liquid phase, and their
effectiveness depends on the type
of hydrocarbon flow, the salinity of
the formation water and the water
cut. In addition, the operation of
the pipeline can be important, since
32 GAS 2010
l&T.indd 6
the dispersion of small hydrate
crystals is favoured by higher
velocities, whereas, at low flow
rates, crystals may settle out and
agglomerate. This is analogous to
the behaviour of water oil disper-
sions in wet crude oil pipelines.
Antiagglomerants provide protec-
tion for up to 40:60 water-oil ratios.
Hydrate slurry method
Researchers have previously
observed that emulsified conditions
(before hydrate formation) resulted
in a slush-like mixture, and segre-
gating the fluids resulted in hard
hydrate spheres. The observations
were attributed to velocity effects
rather than shear stress conditions
that might have promoted the
phenomenon. Later investigations
approached these observations from
a rheological point of view rather
than a flow aspect. It was proposed
that chemical classes such as
asphaltenes were adsorbed on the
hydrate surface, promoting a
reversible aggregation process. This
hypothesis helps to understand the
shear thinning behaviour observed
in some laboratory and flow loop
experiments.
As rheology theory continues to
give insight into the hydrate plug-
ging concept, other studies of
droplet size distribution, particle-
particle micromechanical forces and
phase inversion point allow the
application of pre-existing flow
concepts. These concepts are used
to quantify the force required to
overcome the yield stress generated
by viscous plugs during restart
operations.
Preliminary results confirm that
the pre-existence of an emulsified
phase before hydrate formation
improves hydrate transportability
as a slurry flow. However, the
enhancement of hydrate slurry flow
seems to be limited by the initial
physio-chemical properties of the
fluids. Oil with high emulsifying
characteristics and high viscosity
exhibited poor hydrate transporta-
bility when compared to less
viscous oils with a similar
asphaltene content. From these
observations, it could be inferred
that failures resulting from increas-
ing water cuts in oil continuous
systems are preset by the increase
of mixture density and viscosity.
For suspensions of hydrate parti-
cles in water up to 32%, with a size
distribution of 5–50 µm, in the
turbulent region, the pressure drop
in the pipeline carrying hydrate-
water slurries is no higher than the
pressure drop in a pipeline carrying
liquid water. It is postulated here
that the hydrate particles are small,
perhaps smaller in size than the
thickness of the laminar sub-layer,
and move along from the pipe wall
due to the shear gradient. It is a
possible way of overcoming hydrate
plug problems during deepwater
production of hydrocarbons. It does
not require any insulation or heat-
ing system, and it requires less use
of speciality or bulk chemicals.
Coiled tubing
Tubing is inserted through a lubri-
cator, usually on a platform or
floating work-over vessel, in an
effort to get an inhibitor, such as
glycol, to the face of the plug. The
tubing walls are porous to allow
air/gas to lubricate the tube’s travel
for greater penetration. If the block-
age can be accessed with coiled
tubing, methanol can be pumped
down the coiled tubing to the
blockage. In field applications,
coiled tubing has reached as far as
14 800 ft in remediation operations,
and industry is currently targeting
lengths of 10 miles.
This article is based on extracts from a
presentation by the authors at LRGCC 2009.
Bahubali Chandragupthan is Deputy Manager,
Process, with PL Engineering, Gurgaon, India.
He has a BTech in chemical engineering from
the University of Madras, India, and an MTech
in refining and petrochemical engineering from
the University of Petroleum and Energy Studies,
Dehradun. He has published and presented
more than 20 papers on gas hydrates and the
energy economy.
Email: b.chandragupthan@gmail.com
Girish Babu Nounchi is Assistant Manager,
Piping Engineering, with L&T - Gulf, a joint
venture of Larsen & Toubro and Gulf Interstate
Engineering, USA, in Faridabad, India. He has
a BE in mechanical engineering from Andhra
University, India, and an MTech in pipeline
engineering from the University of Petroleum
and Energy Studies, Dehradun.
Email: girishnounchi@gmail.com
www.eptq.com
18/3/10 13:25:03
 You Get More Than Just a Process Gas Compressor
Lubricated up to 1’000 bar, non-lubricated up to 300 bar

For longest running time: We recommend our own designed,

in-house engineered compressor valves and key components

Designed for easy maintenance

We are the competent partner with the full range

of services – worldwide

Your Benefit:
Lowest Life Cycle Costs

www.ptqenquiry.com
for further information

Like this image? Get more: www.recip.com/screensaver

$GB$3,B$LQGG 
burkhardt.indd 1 9/3/10 14:23:44
M$17XUER_SWT_LVVXHPDUFK_210x297mm_ 11.02.10

Robust, Reliable
and Versatile

Compressors Turbines Expanders Reactor Systems PrimeServ www.manturbo.com

The THM gas turbine is recognized in the industry as a proven
workhorse for many on- and off-shore applications. Designed and
continuously improved to meet ever more stringent requirements,
THM gas turbines do their job around the world, even under the
most adverse environmental conditions.
With the THM family of gas turbines, MAN Turbo furnishes straight-
forward solutions for many of today’s environmental challenges
and fuel flexibility requirements.
Ask us for a recommendation.

Engineering the Future – since 1758.

MAN Turbo www.ptqenquiry.com
for further information

m turbo.indd 1 17/3/10 14:30:15

 Tray hydraulic operating regimes
and selectivity
Tray hydraulics is discussed in terms of froth and spray operating regimes, and
possible impacts on selectivity are pointed out by reference to performance data
Ralph Weiland, Nathan Hatcher and Jaime Nava
Optimized Gas Treating
M
ost gas treating contactors
are highly liquid loaded,
with good designs operat-
ing at vapour loads approaching
the jet-flood limit. An optimal
design will also have downcomers
sized to approach the choke flood
limit at the design point, as this
allows for the maximum tray active
area, hence the smallest column
diameter for a given throughput.
At high weir liquid loads, trays
operate predominantly in the froth
regime, where the gas dispersion
rises through continuous liquid,
interfacial areas are large and the
liquid is highly agitated. This
results in high absorption efficiency
for both CO2 and H2S. Low weir
loads correspond to the spray
regime, where the liquid is
dispersed as large droplets that are
projected through a continuous gas
phase as a spray. The droplets
behave like nearly stagnant,
rigid spheres and have extremely
small liquid-side mass transfer
coefficients. Consequently, CO2
absorption is severely retarded
because its rate is controlled by the
liquid-side resistance to mass trans-
fer. This leads to a concomitant
improvement in H2S removal.
A number of cases have come to
light in which seemingly impossi-
bly low H2S leak rates have been
realised and much higher than
normal CO2 rejection rates observed
from trayed columns. In every case,
liquid rates were low and the weirs
were long enough to have weir
loads of 10 gpm/ft or less, putting
the trays well into the spray regime.
Simulations could be matched to
plant performance data only by
using greatly reduced liquid-film
www.eptq.com
gas weiland.indd 1
coefficients for mass transfer (stag-
nant liquid), and by increasing the
gas-side coefficient as expected for
turbulent gas flow past drops.
The spray regime is quite effective
in achieving ultra-low H2S leaks and
greatly improved CO2 rejection rates
and, if conditions cause an existing
plant to operate there, selectivity
will probably greatly exceed what
might be expected. Therefore, under
certain circumstances, consideration
might be given to trying to operate
in this region.
Usually, contactors containing
trays are initially designed to oper-
ate at 70–80% of jet and downcomer
flooding rates using very conven-
tional valve-type trays with few or
no hydraulic performance enhanc-
ers, such as tapered downcomers
and swept-back weirs. These are
safe designs because they use older,
broadly established technology and
leave plenty of room for easy capac-
ity improvement should they fall
short in the field and to accommo-
date the inevitable future increases
in plant throughput. Most contac-
tors in gas treating applications
have relatively high liquid-to-gas
(L/G) ratios, usually quite a bit
higher than in distillation, because
the large amounts of acid gases to
be removed require large solvent
flows for absorption. A well-
designed tray has a turndown ratio
of at least 4:1; in other words, it can
operate satisfactorily from
hydraulic standpoint down to about
one-quarter or less of design rates.
Below three- or four-to-one turn-
down, liquid starts to weep through
the trays to an extent that mass
transfer performance begins to
suffer.
At the L/G ratios common in gas
treating, trays normally operate in
the froth regime. In this flow
regime, the biphase on the tray has
the gas dispersed in a more or less
continuous liquid, and the mixture
flows across the tray as an intensely
agitated froth.
Sometimes a raw gas contains
either very small amounts of H2S
and CO2, or perhaps a small concen-
tration of H2S and a larger fraction
CO2, but the H2S must still be
removed to meet a parts-per-million
specification. In such cases, only
modest solvent flow is needed to
treat the gas with MDEA so the
L/G ratio in the contactor is low.
As the liquid flow is reduced, the
froth gives way to a spray because
there is insufficient liquid flow to
maintain a reasonable volume of
liquid on the tray. The still large
vapour flow shatters the liquid,
converts it to droplets and ejects it
into the downcomer, not as a
stream of continuous liquid flowing
over a weir, but as a swarm of
drops (a spray) thrown over the
weir and into the downcomer.
Interfacial areas, and gas- and
liquid-side mass transfer coeffi-
cients can be radically different in
the froth and spray regimes, with
the consequent likelihood that
selectivity may be affected
significantly.
a Tray hydraulics
Maximum tray hydraulic capacity
can be limited by any one of three
types of flooding mechanism: jet or
entrainment flood, downcomer
backup flood and downcomer
choke flood. The ability of a tray to
operate at the lower end of the
GAS 2010 35
18/3/10 09:55:24

Jet floo
t
en
tra ay
inm
en Spr
Vapor rate
Incipient weep
Acceptable weep
Liquid rate
Figure 1 Tray operating diagram
loading range is limited by weeping
and also by entrainment that occurs
even in the spray regime of opera-
tion as the liquid flow on the tray is
decreased while maintaining a high
vapour flow rate.
The other low liquid rate limita-
tion is vapour bypassing up
downcomers as lowered liquid
depths unseal the downcomers. In
a perfectly balanced design, jet
flood and choke flood occur simul-
taneously and the column’s
diameter is sized to be just adequate
for the proposed loads; certainly
pressure drop will not be high
enough nor downcomer clearance
tight enough to induce backup
flood problems, nor will vapour
bypass up downcomers.
Entrainment flood
Figure 1 shows a typical tray oper-
ating diagram illustrating stable
limits. The portion of the upper
operating line labelled spray
entrainment is in the froth regime
and corresponds to the upper limit
of vapour flow where the tray
begins to flood, the so-called jet
flood line. This should not be inter-
preted as a finely drawn, sharp line
because the definition of jet flood
varies from vendor to vendor. Some
vendors call 85% jet flood the
vapour velocity at which 10% of the
liquid is entrained from the tray
and flows with the vapour to the
tray above. Others call it the maxi-
mum useful capacity. Whether the
definition is 80 or 90%, jet flood
36 GAS 2010
gas weiland.indd 2
d
Do wnc
Do
wn om
co er
me ba
r c cku
ho p
ke
Dump point
corresponding to 10 or 20% entrain-
ment is arbitrary and the
pinpointing of maximum useful
capacity is certainly subjective.
However, near the jet flood limit,
the entrainment rate is exponen-
tially sensitive to small changes in
vapour rate and only a small
increase in vapour flow will cause
the column to flood. This is not a
condition at which one would like
to operate because the column
approaches instability and becomes
harder to control.
Vapour load is best described by
the C-factor, usually based on the
superficial vapour velocity, uS,
through the empty column but
sometimes based on the gas
velocity through the tray active area
or even through the area of the
actual openings (valves and so on).
The superficial velocity definition
is:
ρv
√ ρ –ρ
Cs = us
L v
A conventional valve tray with
13% straight downcomers operating
at normal weir loads (50–150 gpm/
ft of weir) will flood when CS is
much above 0.38 ft/s. A more
modern, high-capacity tray uses
special downcomer treatments to
recover much (and, in the case of
truncated downcomers, all) of
the area under the inlet down-
comer, so it enjoys between a 10
and 20% higher capacity. This
higher jet flood capacity is achieved
simply by converting more of the
tower area into tray active area.
One jet flood mechanism is froth
entrainment by gas, which carries
liquid to the tray above in large
enough quantities that downcomer
capacity is exceeded on the top
tray. Except for the bottom-most
and top-most trays, each tray in the
column receives as much entrain-
ment from the tray below as it loses
to the tray above, so it experiences
very little change in liquid load
caused by entrainment. The bottom
tray actually has a lower liquid load
than any other tray in the column
because it receives no entrainment
from below but loses liquid to the
tray above. However, the top tray
has most of its entrainment returned
to it by the mist pads commonly
installed in the top of the tower, so
the liquid load on the top tray is
incremented by the entrainment
rate from the tray below. Thus, the
top tray has the highest liquid load
and would be expected to flood
first.
However, froths on operating
trays are not well-behaved flows of
uniform depth — massive eruptions
and geysers occur with high
frequency, with the liquid being
thrown against the bottom of the
overlying tray and forced through
its perforations. At close tray spac-
ing, the froth may occupy the whole
space between the trays and flood-
ing occurs when the froth tries to
expand even further. This may not
be limited to close tray spacing — it
is hard to make more than a slight
distinction between froth occupying
the whole volume between trays
and numerous geysers throwing
liquid at the upper tray. Of course,
when foaming occurs, the whole jet
(and downcomer) flood situation is
exacerbated in the extreme.
Downcomer flood
Downcomers can flood by two
possible mechanisms: excessive
liquid backup inside the down-
comer (backup flood) and choke at
the downcomer mouth. The right-
hand side of Figure 1 does not
distinguish between the two mecha-
nisms and, indeed, it is often hard
to distinguish between them even
visually in a large-scale, well-
instrumented pilot plant.
www.eptq.com
18/3/10 09:55:50
 Backup flood
As shown in Figure 2a, the depth of
liquid (or froth) in a downcomer is
determined by the liquid depth on
the tray, the pressure drop across
the tray and the head loss for flow
of aerated liquid under the down-
comer skirt. The third factor can be
controlled by using the right down-
comer escape area, but the second
depends on tray pressure drop,
which is a function of the tray
bubbling area, the type of tray
opening being used and the frac-
tional hole area. Backup flood is the
easiest to avoid.
Choke flood
The cross-sectional area (and to a
lesser extent the shape) of the
downcomer mouth is the primary
factor that determines its capacity
for accepting a given flow of liquid
or froth. Referring to Figure 2b, at
choke, liquid does not crest over
the weir. Instead, the downcomer is
simply presented with a higher
volume flow of froth than can be
driven into it when hydrostatic
head is at its greatest, equal to the
tray spacing (the maximum head
possible). Once the downcomer
mouth becomes covered with froth,
liquid height over the weir is no
longer proportional to volume flow
to the power 2/3 as in the Francis
weir formula. Instead, liquid height
over the weir becomes proportional
to the liquid rate raised to the 1.5
power; thus, any further increase in
liquid flow requires greatly
increased head to drive the flow
into the mouth. Once the down-
comer mouth becomes buried in
froth, only small increases in liquid
load are enough to throw the
column into choke flood — the
column becomes increasingly unsta-
ble and eventually impossible to
operate.
During a foaming event, the
downcomer choke flood might be
considered the culprit because the
froth volume is much increased,
with the downcomer being asked to
admit a greatly increased volume
flow of even lower density material.
Under choke conditions, even a
large increase in froth depth drives
only a modestly increased flow. But
this is speculation because, once
www.eptq.com
gas weiland.indd 3
flooding begins, it is extremely hard
to ascribe one mechanism or
another as the cause — they are all
connected.
Weeping through tray openings
Referring again to Figure 1, as the
gas rate is decreased at a fixed
liquid rate, a point is reached at
which the tray starts to weep. This
is the incipient weep point, where
the first droplets of liquid find their
way downwards through a few
perforations. Further decreases to
vapour rate result in increased
weep rates until the weep rate
becomes, in some sense, excessive.
There have been instances where,
during a turnaround, all the trays
have been found in a heap in the
tower sump, yet the column was
making on-specification gas at
design rates just before the turna-
round. Obviously, what defines a
weep rate as excessive depends
very much on the operation.
However, when slipping CO2 is
important or meeting both an H2S
and a CO2 specification is the objec-
tive, the column must be more
tightly designed, and too much
weeping will compromise the sepa-
ration because to some extent it
spoils the split.
Spray regime
Kister and Haas (1987) have shown
that the entrainment rate goes
through a minimum with weir
liquid load (see Figure 3). The mini-
mum might be considered to be
somewhere in the transition
$0
$#FULL
$#LOSS
&ROTH
a Backup flood
Figure 2 Two types of downcomer flooding
between the froth and spray
regimes, the spray regime lying to
the left. However, the transition
itself is very gradual, so there is not
a single point associated with it.
The spray regime is entered at
higher liquid rates in hydrocarbon
systems (cyclohexane-n-heptane in
the figure) than in air-water (much
closer to amine treating fluids). The
higher the gas rate, the higher the
weir load at transition. However,
unless the tray is operating at an
extremely low vapour rate, full
transition to the spray regime
occurs at a weir liquid load (volume
flow per length of weir) of between
11 and 22 m3/m·hr, and the entrain-
ment rate climbs super-exponentially
as the liquid load is further
reduced.
As the column transitions from
the froth into the spray regime, the
gas goes from being the dispersed
phase to becoming continuous, and
the liquid switches from continuous
to dispersed. Spray is generally in
the form of fairly large droplets,
typically exceeding 1000 µm (1 mm)
diameter, which move at high
velocity. When well into the spray
regime, most of the liquid is in the
form of droplets and little or none
actually lies on the tray. The liquid
no longer flows over the weir, but
instead is projected through the
vapour space and into the down-
comer. The weir might as well not
be there at all, and it has been
observed in distillation that in this
region neither pressure drop nor
efficiency depend to any great
#HOKED$#MOUTH
MOUTHTOOSMALL
b Choke flood
GAS 2010 37
18/3/10 09:56:06
 sparging into and rising through
 liquid. Unconfined gas flow past
large drops probably is much more
0ORTERAND*ENKINSBASEDON&2)DATA
#YCLOHEXANEn.
(EPTANEK0A
turbulent than gas in froth, espe-
!D(MM cially near the interface, so higher
+ISTERETALAIRWATERK0A gas-side mass transfer coefficients
!D(MMHNMM might be anticipated. (At atmos-
pheric pressure, a 1 mm drop has a
terminal velocity of about 7 m/s
US – 1.25 m/s and falls at a Reynolds number of
about 500; in other words, in the
fully turbulent region.)
 The point is that the spray and
%NTRAINMENTRATE+'ENTRAINEDLIQUID+'GAS

froth regimes present completely

different types of fluid flow and,
US – 1.04 m/s
because hydraulics is one of the
major determining factors for mass
transfer, the mass transfer perform-
ance in these two regimes ought to
be different, with the difference
even being magnified in selective
gas treating. Does this really lead to
differences in CO2 slip and, there-
 fore, H2S absorption? The two sets
of plant performance data presented
US – 1.40 m/s next suggest that indeed it does.
It should be noted that operating
US – 1.17 m/s in the spray regime can result in
US – 0.94 m/s
reduced tray vapour handling
US – 0.94 m/s capacity unless the tray is carefully
designed. There are ways to regain
froth regime operation if this is
desired. One is to reduce the
number of tray passes. One of the
 main reasons for using multiple
     
tray passes in the first place is to
,IQUIDFLOWRATE,-(-WEIRLENGTH
reduce excessive weir loads but, if
the weir loads are already too small,
Figure 3 Dependence of entrainment rate on weir load at various gas rates for using multiple passes is counter-
cyclohexane - n-heptane and air-water (Kister & Haas, 1987) productive and fewer passes should
be used. The other approach is to
extent on the presence or absence resistance in the liquid phase, the use so-called weir blocks, or
of a weir. CO2 absorption rate should be picketed weirs, which effectively
In terms of mass transfer, the gas- reduced; in other words, more CO2 shortens the weir length by placing
liquid interfacial area per unit should be rejected and CO2 slip a series of vertical barriers on
volume of spray can be expected to increased. It is less easy to speculate the weirs to force part of the
be smaller than the same area in on the effect of spray regime opera- liquid spray back onto the tray and
good-quality, deep froth. The drop- tion on gas-side mass transfer make the effective weir length
lets, being small and probably resistance. In surface tension nega- shorter.
containing some amount of contam- tive systems (where surface tension
ination, are essentially rigid spheres. decreases as liquid flows down the Simulation and field data
Transfer of acid gases into the inte- column), higher distillation efficien- Before using the ProTreat simula-
rior of drops must take place by cies are observed in the spray tor’s mass transfer rate-based
pure molecular diffusion, a very regime (distillation is generally column model to explain why
slow process compared with trans- controlled by vapour-phase resist- certain treating plants with absorb-
fer into the highly agitated, ance) because smaller drops are ers operating in the spray regime
turbulent liquid in froth. Thus, the produced. However, gas treating is have been observed to perform
liquid-side mass transfer coefficient a surface tension neutral system. much better than they should, it is
might be expected to be very much Nevertheless, highly turbulent gas important to be convinced that
smaller than in froth. Since CO2 flow past drops is markedly differ- simulation is capable of truly
absorption is controlled largely by ent from gas bubbles and jets predicting plant performance at all.

38 GAS 2010 www.eptq.com

gas weiland.indd 4 18/3/10 09:56:20

 Unique Solution for Your
Petrochemical Application
Best for polyethylene, polypropylene,
ethylene glycol and ethylene oxide production
The ultimate contact-less labyrinth sealing
technology
No piston rings - no wear
Insensitive to particles and contaminated gases
Gas-tight design for zero emissions
Highly reliable components for highest
availability of your compressor
Count on our experience:
Thousands of compressors in operation,
hundreds of long-term customers

Your Benefit:
Lowest Life Cycle Costs

www.ptqenquiry.com
for further information

Like this image? Get more: www.recip.com/screensaver

$GB/DE\$LQGG 
burckhardt2.indd 1 9/3/10 14:25:09
 Predicting plant performance
A mass transfer rate-based column
model uses the fundamental mass
transfer characteristics of the
specific tower internals (trays,
random packing and structured
packing) together with phase equi-
librium and reaction kinetics to
calculate the actual mass transfer
rates occurring between the liquid
and vapour phases flowing on real
trays and through packed beds. The
approach is completely analogous
to the way heat transfer calculations
have been done and heat exchang-
ers have been designed for nearly a
century. The real difference is the
transferring entities: heat in one
case and chemical species in the
other, so temperature differences
are used in one case and concentra-
tion differences in the other. Where
heat exchangers’ heat transfer
properties are correlated, for exam-
ple, in terms of shell- and tube-side
heat transfer coefficients as func-
tions of exchanger hydraulics and
the transport properties of the
heated and cooled streams, the
mass transfer properties of trays
and packing are correlated as
liquid-side and vapour-side mass
transfer coefficients, again as func-
tions of hydraulics and the transport
properties of the two phases. The
analogy is a very strong one. Phase
equilibrium and the number of
species make mass transfer calcula-
tions a lot more challenging than
heat transfer, but this is only a
Treated
3 13
Wash water
Recycle
30
2 18
Raw gas
0.94% CO2
60 ppm H2S Absorber
1 15
4 5
Flash
drum
Figure 4 PFD for case study 1
40 GAS 2010
gas weiland.indd 5
computational technicality. If the
underlying correlations themselves
are reliable, the heat transfer rate
and mass transfer rate models
predict equipment performance
with equal trustworthiness and
fidelity. Engineers do not use effi-
ciency-corrected equilibrium heat
exchanger models, and there is no
longer need to use their mass trans-
fer equivalent in tower design.
The mass transfer rate-based
approach to column simulation as
implemented for example in Aspen
Plus and Pro/II under the name
Ratefrac (a general-purpose module)
by some solvent vendors and refin-
ers, and within ProTreat (specifically
for gas treating), has been used to
simulate hundreds of columns
around the world. The approach
has been in continuous commercial
use for some 20 years, so it is
evidently a sound one. The advan-
tage of mass transfer rate-based
simulation is that it is truly and
reliably predictive, without leaning
on any knowledge of plant perform-
ance to tweak the model to get the
right results.
Case studies
Both case studies involve plants
that are treating raw gas having
parts per million H2S, in the first
case 60 ppmv and in the second
240 ppmv. Both plants produce
treated gas having about 1 ppmv
H2S and, because neither is treating
gas containing a lot of CO2, only
Cooler
CB 17 16
Controller
Pump
Flash gas
HTXR
7
6
Transfer
pump
low solvent rates are necessary to
process the gas. In other words, the
contactors in these plants are oper-
ating with unusually low L/G
ratios.
Case study 1
Ineos Gas/Spec SS-3 solvent at 45
wt% strength is used to treat gas
containing 60 ppmv H2S and 0.94
mol% CO2 at about 460 psig. Figure
4 shows the flowsheet.
Straightforward simulation of this
process predicts that the treated gas
should contain 27 ppmv H2S and
the CO2 level should have been
reduced to 0.636 mol%, with the
rich amine loaded to 0.634 moles of
CO2 per mole of SS-3 solvent.
However, a series of carefully
planned and executed plant trials
showed unequivocally that the raw
gas, varying between 50 and 70
ppmv H2S, was actually treated to
between 1.0 and 1.2 ppmv H2S and
the treated gas was consistently at
0.865 mol% CO2 (93% CO2 slip!)
with a rich solution CO2 loading of
only around 0.13 mol/mol. The
simulation results were not even
close to reality.
These are one-pass valve-type
trays and, because of the low L/G
ratio, they have only 5% downcom-
ers; that is, 90% active area. The
trays run at about 50% of jet flood,
with the downcomers at only 11%
of choke flood. Jet flood and choke
flood are not issues. The striking
hydraulic parameter, however, is
Acid gas
11
12
200 °F
Regenerator
8
10
9
14
www.eptq.com
18/3/10 09:56:35
 SOx NOx

ACID MIST PARTICULATES

MANY SOLUTIONS, ONE SOURCE. IT’S A BELCO® WORLD.
Reducing refinery emissions has never been so easy.
BELCO® is a leader in complete air pollution From consulting to design, fabrication,
control solutions, ready to serve your facility installation, training and maintenance, the
anywhere in the world. Thousands of refinery, world turns to Belco Technologies Corporation.
industrial and municipal clients trust us for For more information and online access to
proven technology leadership and customized, our extensive Technical Library of white papers
turnkey services. and brochures, visit www.belcotech.com.

%NTIN2OAD0ARSIPPANY.*53!s0HONE  s&AX  sE MAILINFO BELCOTECHCOMsWWWBELCOTECHCOM
Copyright © 2008 DuPont. The DuPont Oval Logo, DuPont™, and BELCO® are registered trademarks or trademarks of E.I. du Pont de Nemours and Company or its affiliates. All rights reserved.

www.ptqenquiry.com
for further information

belco.indd 1 17/3/10 14:33:00

 the weir load at only 8 m3/m·hr (11
gpm/ft). At a superficial gas veloc-
ity of 0.39 m/s (1.27 ft/s), this puts
the trays well into the spray regime
of operation; on the other hand, the
simulation is based on mass trans-
fer characteristics that pertain to the
froth regime. Much work has been
reported on the froth-to-spray tran-
sition. However, no work appears
ever to have been done, and none
has been reported, which measures
such fundamental mass transfer
parameters as film coefficients and
interfacial areas for mass transfer in
the spray regime, possibly because
the general recommendation is not
to operate there. Therefore, all mass
transfer rate-based simulators do
their calculations using mass trans-
fer characteristics that pertain to
froths, and this takes care of the
majority of operating distillation
and absorption columns.
The measured plant data are
accurately reproduced by the simu-
lation if the liquid- and gas-side
coefficients are decreased and
increased, respectively, by a factor
of roughly 10. The simulated results
in this particular case were not very
sensitive to reasonable changes in
interfacial area. It is interesting to
note that decreasing the liquid-side
coefficient alone produced the
observed CO2 slip and also resulted
in a significant (albeit insufficient)
improvement in H2S pickup because
less solvent was used up by over-
absorption of CO2. A further
significant increase in the gas-side
mass transfer coefficient was
needed to reach the observed H2S
treat.
Case study 2
For this case, only data for the
contactor were available. A lower
pressure gas (11.7 bar, or 155 psig)
was being treated than in Case 1,
using a solvent mixture comprising
19.15 wt% MDEA and 10.85 wt%
DEA (30 wt% total amine) flowing
at only 1 m3/hr (5 gpm) to the
contactor. The raw gas contained
2.1 mol% CO2 and 240 ppmv H2S.
Lean amine CO2 and H2S mole
loadings were 0.0136 and 0.0005,
respectively. The H2S leak from the
absorber was measured to be
1 ppmv and the treated gas
42 GAS 2010 www.eptq.com
gas weiland.indd 6
contained about 1 mol% CO2.
However, simulation based on froth
regime, mass transfer coefficients
predicted 23 ppmv H2S and 0.72
mol% CO2. Again, too much CO2
pickup is predicted and, partly as a
consequence, simulated H2S
removal is too low by a substantial
margin. The hydraulic calculations
gave a weir liquid load equal to the
amazingly low value of 2 m3/m·hr
(2.8 gpm/ft), making the spray
regime an absolute certainty. But,
just as in Case 1, if the liquid- and
gas-side coefficients are decreased
and increased, respectively, by
roughly the same factor (10), the
plant data are perfectly matched by
the simulation. It should be
mentioned that the trays in this
example were turned down to only
11% of jet flood so serious weeping
was possibly another issue.
Nevertheless, regardless of whether
there was weeping or not, the liquid
was definitely in the form of a
spray and spray regime mass trans-
fer corrections should apply. The
results indicate that indeed they
do.
These are not small changes in
individual phase mass transfer coef-
ficients. However, it should be
realised that the flows of the phases
change from turbulent to stagnant
in the case of the liquid, and from
relatively quiet for gas bubbles to
very turbulent near drop surfaces
in the case of the gas. The size of
the corrections emphasises the radi-
cal difference in the flows.
Discussion
For the most part, gas treating engi-
neers seem unaware of the important
distinction between spray and froth
regimes of tray operation, and the
effect of weir load in determining
where a given set of trays actually
operates. More importantly, the
impact of froth versus spray flow
regime on selective performance
appears never before to have been
addressed quantitatively. However,
the idea that a laminar or stagnant
liquid might promote CO2 rejection
is certainly not new. Sigmund and
Butwell (1981), for example, devel-
oped a special tray intended to
produce a laminar liquid flow with
the specific objective of increasing
CO2 slip and, therefore, improving
H2S removal rates. Their idea was to
produce laminar liquid jets on trays,
with the liquid flowing into trun-
cated downcomers with perforated
bottoms. Although a version of their
tray design was installed in a New
Mexico gas plant owned by El Paso
Natural Gas, this patented tray has
never been truly commercialised.
The normal weir loading range
for conventional trays is from
roughly 30–40 m3/m·hr (40 or 50
gpm/ft) to about 120 m3/m·hr (150
gpm/ft). Modern high-performance
trays such as Sulzer’s VGPlus tray
and Koch-Glitsch’s Superfrac tray
have superior capacity across the
board, but the absolute upper limit
on weir load is independent of the
type of tray. In gas treating applica-
tions, L/G ratios are generally high
because most gases contain high
concentrations of acidic components
and the solvents have limited
capacity, so high solvent rates are
required relative to the gas flows
being treated. Furthermore, except
in tail gas treating and more
recently CO2 capture from power
plant flue gases, treating pressures
are usually in the 100s of psi range.
These factors make spray regime
operation somewhat exceptional
(although not rare) in gas treating,
perhaps explaining the general lack
of appreciation for the very exist-
ence of this regime.
The two cases discussed here
presented quite a puzzle when they
first came to light. It just seemed
physically impossible to produce
1 ppmv H2S gases under the stated
conditions and slip as much CO2 as
was claimed. No amount of suppo-
sition about such things as possible
effects of solvent contamination and
tray damage made any sense.
Perhaps a gamma scan might have
led to an earlier resolution;
however, in both cases, the contac-
tors were treating the gases
perfectly satisfactorily so there was
certainly no commercial incentive
to spend money diagnosing a non-
problem. The inquiry was driven
by the failure of the simulations to
come even close to measured
performance. It was intriguing that
simulation of these two plants,
operated by different companies
18/3/10 09:56:47
 and using quite different solvents,
showed similar deviations from
measured performance.
When more detailed tray hydrau-
lic calculations were done, it became
blatantly evident that, in both cases,
the contactors were operating well
into the spray regime. (The trays
used in these studies were not
designed specifically for extremely
low weir loads and spray regime
operation. However, they used inlet
sumps, so vapour bypassing up the
downcomers was likely not an
issue.) Therefore, the basic mass
transfer coefficients used in the
simulations were almost certainly
at fault. Confirmation of this think-
ing was that the simulations
required similar adjustments of
basic mass transfer coefficients to
bring the results into conformity
with measurements. One of the
outcomes is that much more
detailed tray hydraulic calculations
are now done as a routine part of
ProTreat simulations.
Very low L/G ratios are not the
norm in gas treating. However, in
instances where they exist, much
increased CO2 rejection and far
better H2S removal seem to result.
So, should spray regime operation
be encouraged (for example, by
using multi-pass trays with long
weir lengths when a one-pass tray
is sufficient) as a means to promote
better selectivity in gas treating?
Conventional wisdom is to stay
away from this region,* supposedly
because entrainment rates are
higher, so superficial gas velocity
would have to be reduced (tower
capacity sacrificed), necessitating a
more expensive, larger-diameter
tower. However, this is far from the
truth, as a properly designed tray
(no vapour bypassing up down-
comers) actually has higher
vapour-handling capacity in the
spray regime than in the froth
regime. There certainly are circum-
stances, such as the ones in the case
studies reported here, where a
significant advantage can be had by
purposefully operating in the spray
regime. Indeed, two of the world’s
highest capacity (and most expen-
sive) trays, Shell’s ConSep and the
Koch-Glitsch Ultrafrac, both operate
primarily in the spray regime over
www.eptq.com
gas weiland.indd 7
their entire operating range.
Nevertheless, using conventional
trays, some selective contactors
have been enjoying the great selec-
tivity benefits of the spray operating
regime for years. What is new is
knowing the reason for their Our new clean-fuels
success. Understanding certainly
plant is straining the
does not increase risk, it decreases
it, so where possible there is a real
auxiliary units!
incentive to operate in the spray
regime because it may tremen-
dously improve selectivity. What is
needed to determine quantitatively
the magnitude of this effect is a
truly predictive mass transfer rate
model, as featured within the
ProTreat simulator. How can we boost their capacity
Although trays are often not the without major construction?
best choice of tower internals for
TO MEET CLEAN FUELS require-
very low L/G operations, in the ments for gasoline, a medium-sized
present case simulation and experi- 0LGZHVWHUQUH¿QHU\DGGHGDORZ
ence indicate that when it comes to VXOIXUIXHOVWHFKQRORJ\SODQW*DVROLQH
selectivity neither random nor WKURXJKSXWZDVXQFKDQJHG+RZHYHU
structured packing can come even WKHVKDUSLQFUHDVHLQVXOIXUUHPRYDO
UHTXLUHGPRUHK\GURJHQIURPWKH
close to well-designed trays operat-
K\GURJHQXQLWDQGVHQWPRUHVXOIXU
ing in the spray regime. gases to the amine treaters and
GRZQVWUHDPVXOIXUXQLWV7KHVH
DX[LOLDU\XQLWVEHFDPHERWWOHQHFNV
* The authors thank Henry Z Kister for pointing RYHUGULYHQDWWKHFRVWRISURGXFW
out just how ill-founded this conventional SXULW\DQGDPLQHFRQVXPSWLRQ
wisdom really is. In fact, trays can have even
higher capacity in the spray than in the froth
regime if they are properly designed. Obviously,
downcomers must be designed to remain
sealed without reliance on the outlet weir; an
obvious way to do this is to use a sufficiently
deep inlet sump.
ProTreat is a mark of Optimized Gas Treating
Inc. Other marks are the property of their
owners.
2QVWXG\LQJWKHK\GURJHQDPLQH
Literature cited
DQGVXOIXUXQLWV$PLVWFRIRXQGPDQ\
Kister H Z, Haas J R , I. Chem. E, Symp. Ser., 104,
RSSRUWXQLWLHVIRULPSURYLQJVHSDUDWLRQ
p. A483, 1987.
HI¿FLHQF\DQGFDSDFLW\7KHSUREOHPV
Sigmund P W, Butwell K F, US Patent 4 278 621, ZHUHVROYHGZLWKRXWPDMRUFRQVWUXFWLRQ
14 July 1981. E\DSSO\LQJPRGHUQWHFKQRORJ\WRPLVW
HOLPLQDWRUVOLTXLGOLTXLGVHSDUDWRUV
DQGWRZHUWUD\VDQGSDFNLQJ5HVXOWV
Ralph Weiland is a founder of Optimized Gas LQFOXGHGKD]HIUHHSURGXFWDQG
Treating, Clarita, Oklahoma. He has BS, MA and UHGXFHGDPLQHFRQVXPSWLRQ1RZD
PhD degrees in chemical engineering from the GLHVHOFOHDQIXHOVSODQWLVEHLQJDGGHG
University of Toronto.
Email: ralph.weiland@ogtrt.com Read more on this topic at
Nathan Hatcher joined Optimized Gas Treating www.amistco.com
as Vice-President, Technology Development,
in 2009. He has a BS in chemical engineering
from the University of Kansas and is currently a
member of the Amine Best Practices Group.
Jaime Nava is Manager, Software Sales and 3KRQH‡)D[
Licensing, with Optimized Gas Treating. He has amistco@amistco.com
a BS in chemical engineering from Texas A&M KU(PHUJHQF\6HUYLFH
University.
www.ptqenquiry.com
for further information
GAS 2010 43
18/3/10 09:57:01
 ‡0HVKQXPHURXVDOOR\V SODVWLFV :LGHYDULHW\RIUDQGRPSDFNLQJW\SHV ‡&DWDO\VWEHGVXSSRUWV‡%DVNHWVWUDLQHUV
VL]HVDQGPDWHULDOVLQVWRFN$VNXVKRZ
‡9DQH&KHYURQ‡'RXEOH3RFNHW9DQH ‡1R]]OHV‡2XWOHW,QOHWEDVNHWV
RXUSDWHQWHG6XSHU%OHQGŒ3DFFDQLQFUHDVH
‡&DQGOHRU)LEHUEHG‡,QVHUWLRQ0LVWIL[®
\RXUFDSDFLW\DQGHIILFLHQF\ ‡'LVWULEXWRUV‡+XEDQGKHDGHUODWHUDOV

MIST ELIMINATORS RANDOM PACKING PROFILE WIRE SCREEN

When You Need Trays....NOW!

Valve (numerous options) Why wait 10, 15 or 20 weeks for your delivery! Ask AMISTCO Separation
Sieve or perforated
Products about our Fast Track delivery for trays and a complete range of
Bubble cap trays
Cartridge trays tower internals. With complete in-house engineering and fabrication, we
'XDOÀRZ can use your existing drawings or modify them to improve your process.
%DIÀH

www.ptqenquiry.com
!-EMBEROF&RACTIONATION2ESEARCH)NC for further information

WWWAMISTCOCOMsHR%-%2'%.#93%26)#%s   

DISTRIBUTORS & SUPPORTS COALESCERS STRUCTURED PACKING

0DQXIDFWXUHGWRFXVWRPHU ‡2LOZDWHUVHSDUDWLRQV‡+D]HUHPRYDO :RYHQVKHHWPHWDODQGNQLWWHG
VSHFLILFDWLRQVRUHQJLQHHUHGWR IURPIXHOV‡5HPRYDORIWRZHUZHWUHIOX[ VWUXFWXUHGSDFNLQJ%XLOWWRVSHF
PHHWSHUIRUPDQFHUHTXLUHPHQWV ‡&DXVWLFWUHDWHUDSSOLFDWLRQV RUSHUIRUPDQFHUHTXLUHPHQW

Trays_PTQ.indd
amistco.indd 1 1 7/27/09
12/8/09 6:26 PM
10:18:30
 Processing NH3 acid gas in a sulphur
recovery unit
The feasability and economics of a two-stage sour water stripper with an SRU for
NH3 contents of 25% and higher are discussed
Michael Quinlan and Ashok Hati
KBR
T
oday’s refineries are process-
ing crude slates with higher
sulphur and nitrogen contents.
Some of these crude slates have
such high nitrogen contents that the
feed to the sulphur recovery unit
(SRU) contains a far higher sour
water stripper (SWS) acid gas to
amine acid gas ratio than has been
typical for existing ammonia (NH3)-
burning SRUs. In the industry’s
experience with NH3-burning SRUs,
the SWS acid gas is processed in
the SRU along with the amine acid
gas for recovery of elemental
sulphur from hydrogen sulphide
(H2S), and the NH3 content of the
aggregate acid gas has been low
enough that the increase in SRU
equipment sizes to process the NH3
and the loss in Claus sulphur recov-
ery can be tolerated.
For these low NH3 content acid
gas feeds, the furnace temperature
is, or can be designed, high enough
to ensure complete NH3 destruc-
tion, and it is less expensive to
process the NH3 in the SRU than to
employ a two-stage SWS that
would strip the H2S and NH3 in
separate towers, enabling the NH3
to bypass the SRU.
With higher NH3 content in the
SRU feed, the refiner must decide if
a NH3-burning SRU is still the opti-
mum choice. The refiner will
recognise that even higher tempera-
tures are needed for NH3 destruction,
that the incremental mass flow
through the SRU to process the NH3
will be proportionately larger, that
the loss in sulphur recovery from
the additional N2 and H2O will be
more severe, and that there is little
operating experience with high NH3
content SRU feeds.
www.eptq.com
kbr.indd 1
Fuel gas/
LPG
Crude
Crude
distillation
unit
Stripped
sour water
for refuse
Figure 1 Feed to sulphur recovery unit in a refinery
Despite these disadvantages, it
might still be economical to select a
NH3-burning SRU, leading to
assessment of technical risk vs
economics. This article discusses
the issues with high content NH3-
burning SRUs, and looks at the
technical and economic issues that
the refiner must address in choos-
ing between a single-stage SWS
with a NH3-burning SRU compared
with a two-stage SWS with a SRU
that does not process the NH3.
SRU feed in a refinery
The crude feed to a refinery
contains both sulphur (S) and
nitrogen (N) compounds. These
compounds are converted to H2S
and NH3 as the crude is refined into
finished products, such as fuel gas/
LPG, gasoline, diesel and coke. As
Gasoline Coke
Diesel
Refinery units Amine
Lean amine
S to H2S regeneration
N to NH3 unit
ARU
Rich amine acid gas
Sour water
SWS acid gas Sulphur
recovery
unit
H2S (for 2-stage SWS) Sulhur
Sour water
stripper unit
(SWS) NH3 (for 2-stage SWS)
Stripped sour water to WWT
shown in Figure 1, amine and sour
water remove the H2S and NH3 to
meet finished product specifica-
tions. The amine regeneration units
(ARU) produce an acid gas contain-
ing H2S with traces of NH3. The
SWS may employ single- or two-
stage strippers. In a single-stage
SWS, H2S and NH3 are stripped
from sour water in a single column,
whereas a two-stage or double
stripper has the H2S and NH3 strip-
pers as two columns in series. All
of the amine acid gas is processed
in the SRU together with SWS acid
gas (containing both NH3 and H2S
if a single-stage SWS, or H2S only if
the SWS is a two-stage stripper).
As the SWS acid gas increases as
a proportion of the total acid gas
feed, or as the NH3 content of the
acid gas feed increases as a result of
GAS 2010 45
18/3/10 10:14:22

ARU Combustion ARU
acid gas chamber acid gas
Burner
SWS SWS
acid gas acid gas
Air
Figure 2a Single-zone furnace Figure 2b Two-zone furnace
processing crudes with higher N/S
content ratios, there may be techni-
cal and/or economical reasons why
the NH3 component of the SWS
acid gas should bypass the SRU.
The technical reasons why higher
NH3 content feeds may not be a
good fit for the SRU are incomplete
NH3 destruction, increased NOx
formation and inadequate burner/
furnace designs. The economical
reasons are the increased size of
SRU equipment and the additional
SRU equipment that may be needed
to compensate for the loss in
sulphur recovery.
NH3 destruction in a SRU
In a Claus SRU, about a third of the
H2S is burned to SO2 using air. The
produced SO2 then reacts with
uncombusted H2S to form elemen-
tal sulphur. The reactions are
shown as Equations 1 and 2 below.
The overall reaction is shown as
Equation 3:
H2S + 3/2 O2 → SO2 + H2O (1)
2 H2S + SO2 → 3 S + 2 H2O (2)
3 H2S + 3/2 O2 → 3 S + 3 H2O (3)
While the H2S is only partially
oxidised in Equation 1, the NH3 is
completely combusted to nitrogen
and water, as shown by Equation 4:
2 NH3 + 3/2 O2 → N2 + 3 H2O (4)
It is vital to destroy the NH3,
meaning the residual NH3 leaving
the furnace should be 30 ppmv
or lower. If the NH3 is not suffi-
ciently destroyed, NH3 -H2S salts
can form at the cold spots of the
Claus (for instance, in the final
sulphur condenser) and can plug
46 GAS 2010
kbr.indd 2
Zone 1 Zone 2
Combustion Combustion
Burner chamber chamber
Air
the SRU. The reactions that occur
in the Claus furnace are complex
and not fully understood, and the
destruction of NH3 is governed by
kinetics rather than equilibrium.
For NH3-burning SRUs, the three
Ts — turbulence, temperature and
time — are the key to ensuring that
the NH3 is sufficiently destroyed.
A number of different approaches
are available to destroy the NH3 in
NH3-burning SRUs. To achieve the
required turbulence, a high-inten-
sity burner is recommended. If the
burner is a high-intensity type (for
instance, Duiker or HEC), all of the
amine acid gas may be combined
with the SWS acid gas and the
combined stream sent to the burner
and a single-zone combustion
chamber, as shown in Figure 2a.
For low content NH3 feeds, a good
mixing of the high-intensity burner
plus a minimum combustion cham-
ber temperature of 2250°F (1230°C)
has been deemed adequate by the
industry to destroy the NH3. If the
combined acid gas stream is not
sufficiently rich in combustibles,
such that 2250°F is not attainable,
air and/or acid gas preheat may be
used to achieve the minimum
temperature for NH3 destruction.
Even when the combustibles are
rich enough in the combined acid
gas stream, it is often considered a
good idea to preheat the air and/or
acid gas anyway, as this ensures
better NH3 destruction.
An alternate approach to obtain
an NH3 destruction temperature is
to provide a two-zone combustion
chamber, with the amine acid gas
component of the SRU feed split
between the two zones (see Figure
2b). All of the SWS acid gas and
part of the amine (ARU — amine
regeneration unit) acid gas is
sent to the burner. The remaining
amine acid gas goes to the rear
zone of the combustion chamber.
This results in the front zone (where
all the NH3 is burned) having a
higher temperature than the second
zone. The higher temperature in the
front zone ensures better NH3
destruction.
The furnace temperature is
affected by a number of factors. The
combustibles (NH3, H2S and trace
hydrocarbons) and inerts (CO2, H2O
vapour) in the SRU feed determine
the adiabatic flame temperature.
While the hydrocarbons will elevate
the flame temperature and CO2
decrease it, the biggest factor may
be the water content of the SWS
acid gas, which can be 30 mole%
H2O or higher, depending on the
SWS acid gas feed temperature.
Air is normally the oxidising
medium in a SRU but, if oxygen is
available, the use of enriched air
will increase the furnace tempera-
ture. Kinetics and reaction pathways
are the other important factors in
determining the Claus furnace
temperature. There are multiple
reactions (decomposition of H2S
and NH3, formation of CO, H2, COS
and CS2) that occur in the Claus
furnace in addition to those shown
above, and all of the reactions and
their reaction pathways are not
fully understood. Furthermore,
there is now increasing evidence
that SO2 and not O2 is the predomi-
nant oxidising medium for NH3
destruction:
2 NH3 + SO2 → N2 + H2S + 2 H2O (5)
The result is that the refiner or
SRU designer has some uncertainty
about what precisely is the
SRU Claus furnace temperature.
However, it is known (based on
testing by SRU burner manufactur-
ers or by actual SRU performance
test runs) that as the NH3 content of
the SRU feed increases, higher
temperatures are needed to destroy
the NH3, but there is some uncer-
tainty about what that temperature
is for a given NH3 content. Usually
a design furnace temperature is
sought where there is confidence
that all the NH3 will be destroyed
and, if that temperature cannot be
www.eptq.com
18/3/10 11:46:59
 attained, the NH3 will not be proc-
essed in the SRU.
In addition to good burner mixing
and adequate temperature, ample
residence time is necessary to
ensure complete NH3 destruction.
While residence time is always
needed for the Claus reaction (see
Equation 2 above), it is also needed
to destroy the NH3 (Equation 4
and/or 5). When SO2 is the predom-
inant oxidising medium for NH3,
enough residence time must be
provided for the conversion of H2S
to SO2. Furthermore, as the NH3
content of the SRU acid gas feed
increases, there is increasing compe-
tition between the NH3 and the
uncombusted H2S for the produced
SO2. Therefore, more residence time
is usually needed for the SRU
furnace as the SRU feed’s NH3
content increases.
For a 200 tpd SRU, Table 1 shows
the oxygen (in air) requirements
and simulated flame temperature
when the NH3 content of the acid
gas is varied from 5–45%. The basis
is an amine acid gas with 98% H2S,
1% CO2 and 1% hydrocarbons satu-
rated with water at 120°F (49°C).
The SWS acid gas is 50% H2S, 50%
NH3 and saturated with water at
185°F (85°C). No air or acid gas
preheat is assumed and the furnace
has a single zone.
Table 1 shows that the furnace
temperature increases as the SRU
feed’s NH3 content increases. The
O2 for the H2S is based on burning
a third of the H2S to SO2. The O2 for
burning the NH3 is based on
combusting all of the NH3. The
actual O2 is less than the sum of the
O2 required for H2S and NH3, since
some thermal dissociation of H2S
and NH3 occurs. The table shows
that in the sub-stoichiometric SRU
furnace, the NH3 has to compete far
harder with H2S for combustion O2
(or SO2) and that, as the NH3
content increases, the NH3 is
demanding an increasing propor-
tion of the O2 supplied to the
furnace. It can be postulated that
the increase in temperature
increases the kinetics of NH3
destruction and causes more NH3
to thermally dissociate, and this
compensates for the lower O2/NH3
ratio.
www.eptq.com
kbr.indd 3
High NH3 content SRU feeds Licensor A then compares this
Over the last few years, KBR has temperature with the simulated
been involved in refineries that temperature. The simulated furnace
wished to process high-nitrogen temperature is based on kinetic
crudes, including cases where the models and reaction pathways that
NH3 content to the SRU was 25% or Licensor A believes best represent
higher. To assess technical risk, a the SRU furnace’s chemistry. The
number of SRU licensors, burner water content of the SWS acid gas
vendors and operating plants were is often crucial in determining if the
surveyed to determine if it is possi- simulated furnace temperature is
ble to process this high NH3 content high enough to ensure complete
feed in the SRU. We also looked at NH3 destruction. If the simulated
the NH3 destination options if a two- temperature is sufficiently above
stage stripper SWS is employed. the calculated NH3 destruction
After evaluating the technical and temperature, Licensor A will guar-
operating risks of a NH3-burning antee the NH3-burning SRU.
SRU operating with a high content Licensor B indicated that NH3
NH3 feed, we compared the econom- contents above 25% are too high for
ics of going for the “safer” two-stage processing in a SRU. Even with air
stripper SWS. and acid gas preheat and a two-
zone reaction furnace, Licensor B
SRU licensor responses does not feel comfortable that the
Licensor A indicated that it is often simulated furnace temperature is
possible to process the high content high enough to destroy all the NH3.
NH3 feed in the SRU, depending on Licensor B therefore recommends a
furnace temperature. Licensor A two-stage stripper SWS, unless
calculates the required NH3 destruc- oxygen is available at the SRU to
tion temperature using an in-house elevate the furnace temperature.
formula that linearly increases the Licensor C also indicated that a
temperature based on the NH3 two-stage stripper SWS is needed
content. If T is the required NH3 above 25% NH3. Neither Licensor B
destruction temperature, the or C was specific about what they
formula is of the type: believed is the temperature needed
for NH3 destruction at a given NH3
T = A + b*(y-c) content, but it appears the tempera-
ture is higher than required by
where Licensor A.
A = minimum temperature (°F) for
NH3 destruction at a SRU NH3 SRU burner vendors’ responses
content feed of c%. A typically is The two burner vendors contacted
~2250°F were Duiker and HEC. In personal
b = increase in temperature for communication1 and on its website,2
every 1% rise in NH3 content; b is Duiker states that SRU licensors
~ 5 to 6 allow 2–25 vol% of NH3 in the
y = SRU feed % NH3 content combined acid gas stream, but that
c = SRU NH3 feed content at which the company has experience with
a temperature greater than A is NH3 content up to 30%. Duiker
needed; c is ~2–5% suggests that higher temperatures
O2 competition trends and predicted furnace temperature
NH3/ (NH3 H2S, NH3, O2 for H2S, O2 for NH3, Actual O2, Furnace
+ H2S), mole% lbmols/h lbmols/h lbmols/h lbmols/h lbmols/h temp, °F
5 573 30 287 23 297 2294
10 573 64 287 48 318 2319
15 573 101 287 76 343 2344
20 573 143 287 107 370 2369
30 573 246 287 184 437 2418
40 573 382 287 287 510 2438
45 573 469 287 352 582 2486
Table 1
GAS 2010 47
18/3/10 11:47:21

¾19 psig
Sour water
from SW tank
SW
feed pump
Stripped
Sw cooler
Figure 3 Single-stage sour water stripper
and longer residence times are
needed at higher NH3 contents
because these conditions promote
higher NH3 thermal decomposition
or dissociation. Table 1 indicates
that the amount of air required for
the furnace changes from 95.8% of
stoichiometric requirements at 5%
NH3 content to 91% at 45% NH3
content. Duiker believes that
adequate NH3 destruction may be
possible for up to 45% NH3, and
that the dominant mechanism will
be thermal decomposition, but is
uncertain what that temperature
will need to be for a given NH3
content.
Duiker provided a reference list
up to 2007 for 50 of its NH3 SRU
burners. Some 22 of these burners
were operating on NH3 contents of
less than 10%, and 27 had NH3
contents between 10 and 20%. One
burner has a NH3 content of 22.1%.
Only one burner (at 30% NH3) was
for NH3 content greater than 25%,
and NH3 slip downstream of WHB
was <30 ppmv for this application.
Duiker also alluded to two other
48 GAS 2010
kbr.indd 4
SWS – AG
to SRU
CW
Reflux
Condenser drum
P=
Reflux
pump
SW
stripper
Steam
CW
Stripped sour water
commercial installations. Total
Vlissingen, according to local opera-
tors, has operated the SRU
occasionally on SWS acid gas
(containing 28% NH3) alone with a
Duiker burner 300. This suggests
that it may even be possible to
operate a SRU on SWS acid gas
alone as long as there is sufficient
H2S in the SWS acid gas to provide
the SO2 needed for NH3 destruction
and for the Claus reaction. NNPC
Kaduna has reportedly operated
with up to 40% NH3 using a Duiker
burner 45.
In a personal communication,3
HEC indicated that it seeks higher
temperatures (>2450°F) and resi-
dence times greater than one second
when the SRU feed’s NH3 content
is greater than 25%. However, HEC
has no experience with NH3
contents above 25%, but does not
believe it would pose problems for
the SRU in adequately designed
furnaces.
SRU operating plants feedback
A European refinery4 has been
intermittently processing a SRU
feed with a NH3 content of 28%.
The refinery frequently changes its
crude slate. One of the crude slates
has a high nitrogen content, result-
ing in a high NH3 content to the
SRU. The SRU is a Superclaus unit
with two Claus stages.
When processing an SRU feed
with a high NH3 content, the refin-
ery ensures the Claus furnace
temperature is 2300°F (1260°C) or
higher, although its SRU licensor
had suggested a minimum of
2400°F (1315°C). The refinery occa-
sionally uses oxygen or fuel gas
assist if this furnace temperature
cannot be achieved. The refinery
reports no problems with plugging
in the SRU, suggesting that NH3 is
being totally destroyed at the main-
tained furnace temperature.
However, the refinery does report
that it takes some time after a crude
switchover to get the required 99%
sulphur recovery.
It appears that the H2S content to
the Superclaus reactor varies after a
crude slate switch. The refinery did
not reveal the size of the sour water
tank nor whether the NH3/H2S
content in the SWS acid gas
changed with a different crude
slate. Assuming the SWS acid gas
composition is constant, it appears
that only NH3 dissociation differ-
ences can explain the change in H2S
content to the Superclaus reactor.
At higher NH3 contents, the
destruction of NH3 through dissoci-
ation or decomposition becomes
more important. As the NH3 decom-
position increases, less air or SO2 is
needed to destroy NH3. This means
that the feed-forward air demand
for the SRU is slightly in error
owing to differences in NH3
dissociation at different furnace
temperatures.
For high NH3 content feeds, the
required SRU furnace temperature
is probably in the 2400–2500°F
(1315–1370°C) range. Similar to an
enriched air SRU, it can be expected
that both H2S and NH3 will dissoci-
ate at these temperatures. The
evidence from enriched air SRU
plants is that the H2 formed during
these dissociation reactions is
higher at the furnace outlet than at
the waste heat boiler (WHB) outlet,
www.eptq.com
18/3/10 10:15:02
 ProjeCt #: NPRA-016
ContaCt: Steve Higley (shigley@npra.org)
DFH7
9ED<;H;D9;I
?\
OekÊh[
Dej
>[h["
OekÊh[
@kij
Ekj
J^[h[$
DFH7
H[b_WX_b_jo

CW_dj[dWdY[
9ed\[h[dY[
WdZ
;n^_X_j_ed
CWo
(+Å(."
(&'&
¼
>[dho
8$
=edpWb[p
9edl[dj_ed
9[dj[h
¼
IWd
7djed_e"
JN

Client: NPRA
<_dZ
oekh
mWo
je
DFH7Êi
H[b_WX_b_jo

CW_dj[dWdY[
9ed\[h[dY[$
?d
jeZWoÊi
Z[cWdZ_d]
[Yedeco"
oek
mWdj
ceh[
lWbk[
\eh
oekh
ZebbWh"
WdZ
DFH7Êi
H[b_WX_b_jo



CW_dj[dWdY[
9ed\[h[dY[
WdZ
;n^_X_j_ed
Z[b_l[hi
_j
m_j^
Wd
[nfWdZ[Z
fhe]hWc
j^Wj
_dYbkZ[i


d[Whbo
*&
mehai^efi
_d
[_]^j
ikX`[Yj
jhWYai"
jme
a[odej[
if[Wa[hi"
WdZ
Z_iYkii_ed
i[ii_edi


m^[h[
oek
YWd
Zh_bb
Zemd
_dje
j^[
jef_Yi
j^Wj
Wh[
[ii[dj_Wb
je
oekh
ikYY[ii$
J^[
;n^_X_j_ed
m_bb
i^emYWi[
j^[
bWj[ij
j[Y^debe]_[i
WdZ
i[hl_Y[i
\eh
oek
je
i[[
WdZ
[lWbkWj[$


Ceh[
j^Wd
(&&
YecfWd_[i
m_bb
[n^_X_j
j^[_h
[gk_fc[dj"
cWj[h_Wbi"
WdZ
if[Y_Wb
[nf[hj_i[
_d


iebl_d]
fheXb[ci
j^Wj
Wh[
Yh_j_YWb
je
oekh
Yecf[j_j_l[d[ii$
J^[
(&'&
HC9
_i
oekh
X[ij
effehjkd_jo
je
a[[f
kf
m_j^
_dZkijho
Z[l[befc[dji
Xo
c[[j_d]


m_j^
[nf[hji
WdZ
\[bbem
fhWYj_j_ed[hi
m^e
m_bb
i^Wh[
j^[_h
fhel[d
fhWYj_Y[i
_d
h[b_WX_b_jo"


cW_dj[dWdY["
fhe`[Yji"
fheYkh[c[dj"
WdZ
jkhdWhekdZi$

H[]_ij[h
jeZWo
Wj
dfhW$eh]%hcY
eh


YWbb
ki
Wj
(&($*+-$&*.&
\eh
ceh[
_d\ehcWj_ed$

NATIONAL PETROCHEMICAL JeZWoÊi
h[Ód_d]
WdZ
f[jheY^[c_YWb
Xki_d[ii[i$

& REFINERS ASSOCIATION
J^[
ceh[
oek
adem
WXekj
ki"
j^[
X[jj[h
m[
beea$

npra.indd 1 9/3/10 14:48:40

 suggesting that some of the H2
formed recombines with S2 vapour
in the WHB to form H2S. Thus, for
a high NH3 content SRU, the SRU
designer may need to look at the
responsiveness of the air control
system and allow for H2S re-
association when calculating the
WHB duty and surface.
Despite the burner vendors’ confi-
dence that NH3 contents greater
than 25% could be processed in a
SRU using air as the oxidising
medium, there are only a handful
of commercial applications, and
only one SRU licensor is prepared
to guarantee the SRU’s perform-
ance. There are risks, therefore, in
looking to the SRU to handle these
high content NH3 streams. These
risks may not be worthwhile if the
cost of the two-stage stripper is not
much more than the increased cost
for the SRU. The differences
between single- and two-stage SWS
strippers and between fuel gas- and
NH3-burning incinerators need to
H2S to SRU
claus train
CW
Stripped
SW cooler P=
¾123 psig
H2S
stripper
Off gas
H2S
stripper
boiler
2nd
feed/bottom
exchanger
1st
feed/bottom
exchanger
Sour water
from SW tank
SW
feed pump
Figure 4 Two-stage sour water stripper
50 GAS 2010
kbr.indd 5
be understood before the optimum with a reboiler, but does not have
configuration can be selected. an overhead condenser. For a
trayed column, the number of
Single-stage SWS actual trays usually varies from 30
Figure 3 shows a schematic of a to 48. Stripped sour water (or wash
single-stage SWS. Sour water from water) is used at the top of the
the SW tank is preheated and sent column to wash out NH3 from the
to the stripper. The stripper consists overhead stream. The overhead
of a column, a reboiler and an over- vapour is mainly H2S with traces of
head condenser (pumparound NH3 (~20–100 ppmv), and is sent to
condensing is also possible). The the SRU’s Claus unit.
stripper has usually 36 to 56 actual The NH3 stripper operates at low
trays, depending on product speci- pressure. This column is equipped
fication and steam usage. The with a reboiler and an overhead
stripped sour water has a low H2S condenser for reflux. The actual
content (<10 ppmw) and a small number of trays is 40 to 44 or
amount of NH3 (<50 ppmw). equivalent packed bed. The NH3
stripper overhead vapour, mostly
Two-stage SWS NH3 and residual H2S (~1500–3000
Figure 4 shows a schematic of a ppmv), is sent to the incinerator
two-stage SWS. The two-stage strip- and burned, assuming there is no
ping concept uses the boiling point end use for the NH3. The NH3
difference between NH3 and H2S to stream can be incinerated directly
remove them as individual streams. or after H2S scrubbing. The need to
The H2S stripper operates at ~125 scrub H2S from the NH3 stream is
psig to remove H2S from the sour dictated by the overall sulphur
water. The column is equipped recovery requirement, and/or
whether the local or environmental
authority judges the NH3 stream to
be a fuel gas source and thus limits
NH3 to
incinerator its allowable H2S content. A small
CW part of the stripped sour water is
NH3 Stripper sent to the H2S stripper top as
condenser NH3 scrubbing water and the rest is sent
stripper outside the unit for reuse by the
reflux
drum refinery users or for wastewater
P=
¾19 psig
treatment. The stripped sour water
has low H2S (<10 ppmw) and about
NH3 stripper
reflux pum
25–50 ppmw NH3.
Recycle to SW tank
Comparison of single- and
two-stage strippers
The two-stage SWS configuration
MP LP
requires an additional stripper,
steam steam reboiler and feed-bottom exchanger
and, therefore, has a higher capital
cost and requires a larger plot
NH3
stripper space. For an equimolar H2S and
NH3
stripper
boiler NH3 sour water feed, the H2S strip-
per diameter is ~15% smaller than
the single-stage SWS. The design
pressure of the H2S stripper is
higher and may require 300# class
flanges. The NH3 stripper has a
Stripped
SW pump diameter similar to the single-stage
SWS. The two-stage stripper
consumes more utilities. For exam-
ple, the reboiler duty for the
combined H2S and NH3 stripper is
~100% higher than for the single-
stage stripper. The overhead
www.eptq.com
18/3/10 10:15:17
 condenser in an NH3 stripper has a
higher duty than the single-stage
stripper’s overhead condenser.
NH3 destination options
The NH3 stream from a two-stage
stripper can be used in several
ways. The choice is usually:
• Liquid NH3 manufacture
• NH3 dissociator
• NH3 incinerator.
Liquid NH3 manufacture is
usually not considered unless the
NH3 quantity is above 30 tpd and
there is a market for the liquid NH3.
In most cases, these conditions do
not exist but, if they do, the NH3
overhead from the two-stage strip-
per must be treated to remove
contaminants (trace H2S etc) in the
NH3 stream. Chevron’s WWT proc-
ess has an optional add-on that uses
a two-stage scrubbing system to
remove H2S from the NH3 stream,
which is then liquefied to produce
anhydrous NH3.
The NH3 gas is dehydrated,
preheated by steam and cracked in
the NH3 dissociator. The dissocia-
tion of NH3 occurs at an elevated
temperature in the presence of a
catalyst. The cracked gases are then
cooled to generate steam and/or
preheat the feed gas. The formed
gas contains little dissociated NH3.
This NH3 may be removed in an
optional molecular sieve unit. The
formed gas is 75% hydrogen and
25% nitrogen. The hydrogen or
formed gas is used for the bright
hardening of metals. It is not
known if this hydrogen-rich gas
could be used to advantage in a
refinery, since the gas is low pres-
sure and has a significant nitrogen
volume.
While some older plants flared
the NH3, this is no longer environ-
mentally acceptable (due to NOx
formation), and because the heat
content of the NH3 gas is not recov-
ered. Today, NH3 is mostly sent to
a NH3 incinerator that is designed
for low NOx and for steam genera-
tion from NH3 combustion.
NH3 incinerator
The incinerator is used at the end
of the SRU to destroy residual H2S
in the SRU tail gas. For a NH3-
burning SRU, fuel gas is used. A
www.eptq.com
kbr.indd 6
typical SRU incinerator that
includes optional waste heat recov-
ery is shown in Figure 5.
Fuel gas is combusted with excess
air in the burner. The flue gas goes
to the incinerator furnace where
SRU tail gas is introduced. The
waste heat boiler (and steam super-
heater) is optional. The cooled
incinerated gas is released to the
atmosphere through a stack.
When a two-stage SWS is used,
the overhead stream from the NH3
stripper is used as incinerator fuel.
This stream is mostly NH3 (~96
mole%) with trace H2S (~1500–3000
ppmv). Depending on the refinery’s
sulphur recovery requirement, or
whether the NH3 is considered a
fuel gas, the NH3 may need
SRU tail gas
Combustion
air boiler
Fuel gas
Burner
Furnace
1450 °F
Figure 5 Typical SRU incinerator
pretreatment to remove H2S. If the
incinerator stack SO2 is not to
increase, a Superclaus unit must
achieve higher sulphur recovery
when the NH3 stream is not
scrubbed of H2S. A NH3-burning
incinerator is specially designed for
low NOx emissions. There are a
number of low NOx-producing
incinerator designs available, such
as the John Zink Noxidizer and
Duiker low NOx incinerator.
The Noxidizer5,6 uses phased
combustion and is a three-stage
process consisting of reducing,
quench and re-oxidation sections
(see Figure 6). There are about ten
Noxidizers operating as SRU
incinerators.
The NH3 stream and fuel gas (if
required) are burned at 70–90% of
stoichiometry. Part of the SRU tail
gas is admitted at the burner and
reduction chamber to reduce the
temperature so that the refractory is
not thermally stressed. The temper-
ature in the reduction chamber is
controlled at ~2300°F (1260°C) with
a residence time of about two
seconds. The high temperature
causes the NH3 to dissociate,
producing nitrogen. The nitrogen
and the combustibles, competing
for the limited quantity of oxygen
available, keep the nitrogen from
being oxidised. The products of
combustion from the reduction
stage, which contain uncombusted
NH3 and uncombusted CO, H2 and
H2S in the SRU tail gas, are cooled
in the quench section by the remain-
ing SRU tail gas to ~1400°F (760°C).
To atm.
Superheated
HP steam
HP steam
Waste heat Steam
boiler superheater
700 °F
BFW
In certain cases, cooled gases from
downstream of the waste heat
recovery are recycled to the quench
section.
In the re-oxidation furnace,
uncombusted NH3, H2 and CO are
burned with excess air. Enough
secondary air is injected to give an
excess oxygen level of 2–3% by
volume on a dry basis in the final
flue gases. The temperature is main-
tained at ~1900°F (1038°C) to limit
the formation of thermal NOx at
about one second residence time.
The temperature difference between
the reduction and re-oxidation
chambers is controlled and depends
on the sub-stoichiometry used at
the reduction stage.
High-pressure steam is generated
from the gases leaving the re-
oxidation chamber. The cooled flue
GAS 2010 51
19/3/10 11:01:28
 To atm.

Re-oxidation
SRU tail gas air blower Superheated
HP steam
Combustion
air boiler
HP steam

Reduction Quench Re-oxidation

Fuel gas furnace furnace Waste Steam
Burner section
2300 °F 1900 °F heat boiler superheater
1400 °F
2.0 sec 1 sec 700 °F

NH3 rich gas BFW

Figure 6 NH3-burning incinerator

gas from the waste heat boiler is Claus, SRU TGTU and NH3 throughput in Superclaus or SCOT,
released to atmosphere through a treatment. assuming that these processes need
stack and the NOx emission is <150 When the refinery SWS is an to be added to Claus, depending on
ppmv. integrated SWS that treats both whether 99% (Superclaus) or 99.9%
The fuel gas incinerator burns phenolic and non-phenolic sour (SCOT) sulphur recovery is needed.
enough fuel gas to meet the require- water, increases in capital cost are This is illustrated in Table 2 for a
ment for incinerator temperature, primarily due to the addition of an 200 tpd SRU, assuming the same
whereas the NH3 incinerator burns extra stripper and reboiler, and the basis as in Table 1. This shows that
all NH3 from the NH3 stripper even operating costs due to the higher a 45% NH3 SRU feed has almost
though that exceeds the tempera- steam consumption. If the refinery double the mass throughput of a
ture to incinerate the SRU tail gas. SWS uses segregated strippers for 5% NH3 SRU feed. In fact, every
While waste heat recovery is non-phenolic and phenolic sour tonne of NH3 in the SRU feed is
optional for a fuel gas incinerator, it water, it is often best to keep the equivalent to ~2.5 tpd of sulphur in
is mandatory if the heat of combus- phenolic stripper as a single strip- terms of SRU mass throughput. As
tion from burning NH3 is to be per, because a far greater shown in Figure 7, the SRU Claus
recovered. The Noxidizer requires proportion of the total NH3 in the will use two catalytic stages as a
bigger equipment (greater residence sour water is associated with the minimum.
time) and higher temperature non-phenolic sour water. The level of two-stage Claus
refractory. Economically, it is best to provide sulphur recovery will change,
an integrated SWS or, if segregated depending on whether a single- or
Overall economics with a two-stage SWS systems are in place or two-stage stripper is employed.
SWS option preferred for water reuse and Table 2 shows that the sulphur
In the preceding paragraphs, the management, to only use a two- recovery level decreases when the
differences in equipment and utili- stage stripper on the non-phenolic NH3 content increases, because of
ties of a single- or two-stage sour water. the water vapour in the SWS acid
stripper SWS, and in fuel gas or With a two-stage stripper SWS, gas and the greater concentration of
NH3-burning incinerators, were the NH3 from the sour water no N2 and H2O in the Claus process
discussed. However, the economics longer goes to Claus. This reduces gas when the NH3 is combusted. If
of a two-stage SWS must also considerably the mass throughput the 200 tpd two-stage Claus SRU
consider differences in the SRU in Claus. It also reduces the mass (at 5% NH3) costs $40 million, the
cost would increase to $58 million
SRU mass throughput and recovery % vs NH3 content if the NH3 content increased to 45%.
The process options (add-ons to a
two-stage Claus) for either 99%
NH3/(NH3 Amine AG, SWS AG, Air, SRU feeds, % sulphur recovery in Claus stages
or 99.9% recovery are shown in
+ H2S), % lb/h lb/h lb/h lb/h 2-stage 3-stage
5 19 511 1961 41 253 62 725 94.4 98.1 Figure 8.
10 18 307 4140 44 263 66 710 94.1 97.9 With a two-stage stripper, only
15 16 961 6575 47 637 77 173 93.7 97.8 two Claus catalytic stages may be
20 15 447 9314 51 442 76 203 93.2 97.6
needed upstream of Superclaus,
30 11 769 15 967 60 704 88 440 92.0 97.1
40 6865 24 652 70 927 102 444 90.7 96.6 whereas a single stripper will
45 3671 30 482 80 800 114 953 89.8 96.1 require three Claus stages.
Typically, the additional Claus
Table 2 stage will add 15–20% to the unit’s

52 GAS 2010 www.eptq.com

kbr.indd 7 18/3/10 10:15:41

 “organised
I was lucky to attend this well
and varied summit.
It certainly added to my
knowledge and broadened my

17th - 19th May 2010

knowledge spectrum.
”
Specialist Planning Manager, KNPC
Beurs - WTC Congress Center, Rotterdam, The Netherlands

www.refiningsummit.com 2010 SPeaKerS INClude: 2010 SPONSOrS INClude:

IMPROVING Dariusz Jacek Krawiec, President and CEO,

PKN Orlen S.A.
OPERATIONAL Dariusz focuses upon recovery from the economic
downturn and the future outlook of the refinery
PERFORMANCE industry.

TO SECURE Jean Jacques Mosconi, Senior Vice

FUTURE President of Strategy, Total.
Jean discovers the latest trends in global refining
PROFITABILITY as well as comparing how competitive mature
markets such as Europe and North America are
against emerging dynamic markets.

In association GLOBAL

with: Steve Welch, Chief Operating Officer, US

R&M and Global Petrochemicals, BP.
Steve identifies the latest developments in refining
operations and strategy.

Jarmo Honkamaa, Deputy CEO &

Executive Vice President of Renewable
Fuels, Neste Oil.
Jarmo investigates the current and future role of
biofuels.

Mutlaq Al Azmi, Operations Manager,

Shuaiba refinery, Kuwait National
Petroleum Company.
Mutlaq reveals operational excellence at the
Shuaiba Refinery.

2010 PlaTINIuM MedIa

Dr Frigyes Lestak, General Manager, CO2 ParTNerS INClude:
Efficiency Shell.
Dr Frigyes Lestak examines carbon efficiency and
management strategies.

TO regISTer, contact richard Jones:

Email: richard.jones@wtgevents.com
Call: +44 (0) 20 7202 7690 - Fax: +44 (0) 20 7202 7600
Visit: www. refiningsummit.com - Quote booking code: PTQ Researched and Produced by:

global summit.indd 1 9/3/10 14:50:12

 HP HP
HP steam RH 1 steam RH 2 steam

Converter 1 Converter 2

ARU acid gas RF WHB

LP steam LP steam LP steam Claus tail gas

Condenser Condenser Condenser

Air BFW

SW H2S + NH3 gas (IF 1–stage SWS)

SW H2S gas (IF 2–stage SWS) BFW S BFW S BFW S

Figure 7 Two-stage Claus sulphur recovery unit

costs. Even if Claus is designed to
accept the small amount of NH3 in
the phenolic SWS acid gas, the NH3
quantity may be small enough that
a two-stage Claus upstream of
Superclaus may be sufficient. For
99.9% sulphur recovery using
SCOT, the Claus will generally only
require two catalytic stages regard-
less of the phenolic SWS
configuration, but more hydrogen
and amine will be needed in
SCOT, and the recycle flow to the
Claus inlet will increase as the SRU
feed’s NH3 content increases.
Additionally, the amount of water
removed by the SCOT quench
section will increase with a rise
in the NH3 content. These will
2nd stage
claus tail gas
3rd stage
claus
Hydrogenation
Option
Figure 8 SRU options for 99.0/99.9% recovery
increase the SCOT and SWS costs.
The amount of steam generated
in Claus is less with a two-stage
stripper SWS. With a single-stage
SWS, the heat from the combustion
of NH3 is largely recovered in the
SRU. For operating costs, the value
of the additional steam is usually
worth more than the additional
power that is needed for the SRU’s
air blower to deliver the incremen-
tal air to combust the NH3.
Finally, the fuel gas-burning SRU
incinerator for a single-stage SWS
will become a NH3 fuel incinerator
for a two-stage stripper SWS. It can
be seen that the NH3 incinerator
will be more expensive than the
fuel gas incinerator. For a two-stage
Superclaus
99% recovery
To
incinerator
S assumed. Even with the most
SCOT
99.9% recovery
Amine
+
quench
treating
SW Acid gas
to SRU
stripper SWS, the heat from the
combustion of NH3 is recovered at
the incinerator rather than at the
SRU, as is the case for a single-stage
SWS.
The NH3 that goes to the incinera-
tor contains about 2500 ppmv H2S.
This can represent as much as 0.2
wt% of the inlet sulphur. If 99%
sulphur recovery is needed, this
may require the Superclaus unit to
achieve 99.2% recovery, which may
be beyond its capability and would
require that Superclaus be upgraded
to Euroclaus. If 99.9% sulphur
recovery is needed, it will be neces-
sary to treat the NH3 to remove
the H2S.
The differences in capital and
operating costs are shown in the
following example for a 200 tpd
SRU processing a SRU feed with a
45% NH3 content, when the
required sulphur recovery is 99%
and 99.9%, respectively.
In Table 3, only one Claus and
Superclaus or SCOT train is
favourable conditions (NH3 content
at 45%), the two-stage stripper’s
capital costs for 99% recovery are
the same or slightly more than for
the single-stage SWS. For 99.9%
recovery, the unloading of the NH3
from the SRU helps reduce both
Claus and SCOT costs, and, at very
high NH3 contents, the two-stage
stripper can be slightly cheaper.
Table 4 shows that the additional
stripping steam for the two-stage
stripper is not fully offset by

54 GAS 2010 www.eptq.com

kbr.indd 8 18/3/10 10:15:55

 savings in fuel gas and by the lower
Claus blower power. Overall,
Tables 3 and 4 show that a two-
stage SWS can be economic with a
single-stage stripper. There are
several variants that affect the cost
comparison, but a two-stage SWS
will usually look most attractive
when the SWS capacity is high, the
NH3 content is near maximum and
when 99.9% sulphur recovery is
required. Based on our experience,
the cost of a two-stage stripper is a
minimum of 5% greater, and some-
times can be as high as 25% greater
than a single-stage SWS when NH3
content is 25% or less and when
only 99% sulphur recovery is
needed.
Conclusions
• For NH3-burning SRUs, the
industry standard — that a mini-
mum temperature of 2250°F
(1230°C) is needed to guarantee
NH3 destruction — needs to be
modified and correlated with NH3
content. In fact, minimum tempera-
tures of 2400°F (1315°C) or higher
may be needed for NH3 contents of
25% and above.
• High-intensity burner manufac-
turers such as Duiker and HEC
believe that their burners can
achieve these higher temperatures
if water vapour content is not exces-
sive, and that a SRU feed NH3
content of greater than 25% can be
processed in a SRU.
• There are few NH3-burning SRUs
operating with a NH3 feed content
above 25%.
• If oxygen is available, enriched
air can be used at the SRU and
ensure that temperatures of NH3
destruction at the SRU are achieved.
This will nearly always be cheaper
than using a two-stage stripper at
the SWS.
• A two-stage stripper at the SWS
is usually more expensive than a
single stripper and requires a lot
more steam. However, these
increased capital and operating
costs are offset by a smaller SRU.
• The economics of a single strip-
per with a larger SRU against a
two-stage stripper with a smaller
SRU depend on several factors,
such as SWS flow, NH3 quantity,
SRU recovery and NH3 destination.
www.eptq.com
kbr.indd 9
Capital costs comparison
Single-stage SWS plus SRU 2-stage SWS plus SRU
SWS1 35 60
SRU Claus (2-stage) 58 40
Add 3rd-stage Claus 9 66
Superclaus 11 8
SCOT 48 33
Incinerator2 15 18
Total for 99% sulphur recovery3 128 132
Total for 99.9% sulphur recovery4 165 151
Notes:
1. Cost depends on SWS capacity
2. Both fuel gas and NH3 incinerators generate HP steam
3. Sum of SWS, SRU Claus, third stage, Superclaus and Incinerator
4. Sum of SWS, SRU Claus, SCOT and Incinerator
5. Cost of treating NH3 for H2S removal not included
6. 3rd Claus stage often not needed for 2-stage stage SWS
Table 3
Under the most favourable circum- International Conference on Environmental
stances, the capital cost for a Control of Combustion Processes, Honolulu, HI,
two-stage stripper is the same, but Oct 1991.
usually the capital cost is anywhere
between 5% and 25% more. Mike Quinlan is a Senior Process Manager
with KBR, Houston, where he is responsible for
• In cases where the cost increase
gas/liquid treating and sulphur. He has over 30
is no more than 15%, it may be best
years’ experience in acid gas removal and has
to use a two-stage stripper and designed and started up numerous amine, sour
avoid the potential operating diffi- water stripper and sulphur recovery units. He
culties of a NH3-burning SRU. has a bachelor’s degree in chemical engineering
from University College, Dublin.
Email: michael.quinlan@kbr.com
References Ashok Hati is a Senior Technical Advisor with
1 Email and phone conversation with Dennis KBR, Houston. He has 26 years’ technical
van de Giessen at Duiker, Netherlands. experience ranging from conceptual process
2 Statement concerning NH3 decomposition development to design engineering and
in Claus reaction furnaces, www.duiker.com. plant operation, the last ten years in gas/
3 Email and phone conversation with Nick liquid treating and sulphur recovery. He is a
registered Professional Engineer in Texas and
Roussakis at HEC, Calgary, Canada.
4 Phone conversation with Cees Koopman. holds a BTech in chemical engineering from IIT,
5 Noxidizer brochure, www.johnzink.com. Kharagpur, India.
6 Incineration of nitrogen bearing wastes, Email: ashok.hati@kbr.com
Operating costs comparison1
Unit Single-stage SWS 2-stage SWS
plus SRU plus SRU
SWS reboiler steam2 MMBtu/h 83 160
SRU 2-stage Claus steam MMBtu/h -97 -46
SCOT steam MMBtu/h 14 7
Incinerator fuel gas MMBtu/h 58 2
Incinerator steam MMBtu/h -48 -53
Total for 99% sulphur recovery3 MMBtu/h -4 62
Total for 99.9% sulphur recovery4 MMBtu/h 10 69
Electrical (blower) powers
SRU Claus blower power hp 1140 550
Incinerator blower power hp 160 150
Notes:
1. All positive values are consumption; negative values represent production
2. Depends on SWS capacity
3. Sum of SWS, SRU Claus and incinerator
4. Sum of SWS, SRU Claus, SCOT and incinerator
5. Cost of treating NH3 for H2S removal not included
Table 4
GAS 2010 55
19/3/10 11:02:16
 Alphabetical list of advertisers

Albemarle Catalysts Company 26 Euro Petroleum Consultants IBC

www.albemarle.com www.europetro.com

Amistco Separation Products 43, 44 Global Refining Summit 53

www.amistco.com www.refiningsummit.com

Axens 18 Haldor Topsøe IFC

www.axens.net www.topsoe.com

Borsig 17 Lurgi 4
www.borsig.de www.lurgi.com

Bryan Research & Engineering 25 MAN Turbo 34

www.bre.com www.manturbo.com

Burckhardt Compression 33, 39 Merichem - Gas Technology Division 2-3

www.burckhardtcompression.com www.gtp-merichem.com

Criterion Catalyst & Technologies 6 MPR Services 31, OBC

www.criterioncatalysts.com www.mprserve.com

DuPont Belco Clean Air Technologies 41 Worley Parsons 29

www.belcotech.com www.worleyparsons.com

For more information on these advertisers, go to www.ptqenquiry.com

56 GAS 2010 www.eptq.com

ad index copy 6.indd 1 18/3/10 12:21:25

 4th Asia Bottom of the Barrel
Euro Petroleum Consultants
Technology Conference BBTC 2010

18 & 19 May, Parkroyal Hotel, Kuala Lumpur

Key SPeAKerS inClude:
 M Radzif A Kadir,
Lube Base Oil Specialist,
MELAKA REFINERY
PETRONAS
 Dr Maizar Rahman,
Commissioner,
PERTAMINA
 Marcos Benicio Pompa Antunes,
Executive Director,
NANSEI SEKIYU K.K. /
PETROBRAS
Now entering its fourth successful year, Asia BBTC is established as the
only event capable of bringing you up to date with the latest economic
and technology developments for upgrading crude oil residues including
the modernisation of existing facilities, new equipment and catalyst
innovations.
With its expert speaker panel and unrivalled networking opportunities
with key regional refinery management, Asia BBTC has become the
industry standard for excellence and is the must-attend resid upgrading
event for your calendar.
Key highlighTS:
 New Decade - New Challenges? Outlook for the Asian Refining
Industry
 How to Select the Most Beneficial Technology for Residue Upgrading
 Latest Catalysts and Technologies for Resid Upgrading
 Cost-Saving Hydroprocessing Technologies & Equipment Innovations
 Operator Presentations by Petronas’ Melaka Refinery, Pertamina,
BP Kwinana and Petrobras

8th Asia Petrochemicals Strategy

Euro Petroleum Consultants
& Technology Conference APTC 2010

20 & 21 May, Parkroyal Hotel, Kuala Lumpur

Refinery-Petrochemical integration will be high on the agenda at APTC:

benefits and opportunities will be discussed and case studies from recent
projects in Korea, Middle East and Europe will be presented.
Leading solution providers and technology companies will present on their
latest cost-effective strategies and products, and we will hear positive
updates on new and ongoing petrochemical projects.
Join us in Kuala Lumpur to meet with the leading players in the Asian
petrochemical industry, be updated on the latest market issues and gain
Key SPeAKerS inClude: an insight into how to maintain and improve the future of your business.
 Byeong-Jae Park,
General Manager, Key highlighTS:
SK ENERGY
 Competitiveness in Petrochemicals in Uncertain Economic Times
 Ahmed Al-Emadi,
General Manager,  Focus on Refinery-Petrochemical Integration - Practical Examples
QATAR CHEMICAL  Case Studies from SK Energy, RAFO, PTT Phenol, Shell MEG Plant at
COMPANY SEPC, Q-Chem
 Thevarak Rochanapruk,  Outlook and Technologies for Aromatics, Olefins and Polyolefins
Vice President & Project Director,
PTT PHENOL  Maximising Petrochemical Assets. Optimisation Strategies

SPeCiAl rATeS for All refinery & PeTroChemiCAl PlAnT PerSonnel!

regiSTer at www.euroPetro.com, Tel: +44(0)20 7357 83964, email: conferences@europetro.com

epc.indd 1 9/3/10 14:46:44

 There is a company that can help you monitor and optimize the operation of your amine/
glycol treating units! MPR Services has been doing this successfully since 1989 for their
customers around the world!
What Does MPR Technology Do?
’ Remove contaminants
’ Reduce solvent losses
’ Eliminate foaming
’ Prevent corrosion
’ Stop forced shut-downs
Since 1989, MPR Services, Inc. has used their technologies to
clean amine and glycol solvents. Utilizing special ion exchange
resins with unique control and regeneration methods, the HSSX®
Process continues to be used successfully to remove Heat Stable
Salts (HSS), (organic acids and inorganic acid anions), as well as
chlorides, sodium and other contaminants.
The HSSX Process removes HSS from; MEA, DEA, MDEA, DIPA, Sulfinol M, Sulfinol D, Flexorb® SE, TEG,
and many other proprietary and generic amine and glycol solvents.
Upon completion of the HSSX Process, your amine or glycol system will exhibit improved performance with:
’ Less foaming,
’ Less filter plugging,
’ Lower amine losses,
’ Lower corrosion rates,
’ Reduced solvent circulation,
’ Reduced stripping steam requirements,
’ Lower NH3 levels at the waste water treatment plant.
The HSSX Process can take a slipstream from your system and clean the solution on-line, or it can clean the
solution off-line while the system is down. The HSSX Process may also be used to clean spent amine or glycol
from storage tanks.
MPR Services has cleaned over 4 million gallons of amine and glycols for many of the major oil companies,
gas plants, ammonia plants, and coal gasification plants. MPR provides all necessary personnel and specialized
equipment to perform their services at the plant.
MPR also offers full analytical laboratory services to support job planning, as well as support on-site to jobs in
progress.
MPR Services can help you monitor and optimize the operation of your amine/glycol treating units!
Contact us today!
MPR Services, Inc, a member of the Tessenderlo Group,
1201 FM 646, Dickinson, TX 77539-3014 USA
Telephone 281-337-7424 Fax 281-337-6534
FNBJM

JOGPNQSTFSWJDFT!ULJOFUDPN
t
Npreurope@tkinet.com
t
NQSNJEEMFFBTU!ULJOFUDPN
www.mprservices.com
HSSX is a registered trademark of MPR Services, Inc., U.S.A. SSX, HCX, and Oxex are trademarks of MPR Services, Inc., U.S.A.
SigmaPure is a trademark of D-Foam, Inc., U.S.A. Cartoon ©Steve Harris
www.ptqenquiry.com
© Tessenderlo Kerley, Inc. 2010
for further information

mpr.indd 1 9/3/10 14:06:53