Differential Scanning Calorimetry
SpringerVerlag Berlin Heidelberg GmbH
G. W. 
H. H6hne . W. F. Hemminger 
H. J. 
Flammersheim 
Differential Scanning Ca lori metry
2nd revised and enlarged edition
With 130 Figures and 19 Tables
"
Springer
Dr. G.W.H. H6hne
(Retired from Universitat Ulm) Morikeweg 30
88471 Laupheim, Germany
email: gwh.hoehne@tonline.de
Dr. W. F. Hemminger
PhysikalischTechnische Bundesanstalt Bundesallee 100
38116 Braunschweig, Germany
email: WolfgangHemminger@ptb.de
Dr. H.J. Flammersheim
Universitat Jena Institut fur Physikalische Chemie Lessingstraf5e 10
07743 Jena, Germany
email: C7flha@unijena.de
ISBN 9783642055935
Library of Congress CataloginginPublication Data Hohne, G. (Glinther) Differential scanning calorimetry: an introduction for practitioners / G. W. H. Hohne, W. Hemminger, H. J. Flammersheim.  2nd rev. and en!. ed. p. cm. Includes bibliographical references and index. ISBN 9783642055935 ISBN 9783662067109 (eBook) DOI 10.1007/9783662067109
1. Calorimetry. 1. Hemminger, W., 1941II. Flammersheim. H.J., 1942II1. Title. QC29I.H64 2003
536'.6dc21
2003050472
This work is subject to copyright. AII rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilrns or in any other ways, and storage in data banks. Duplication of this publication or parts thereof is only permitted under the provisions of the German Copyright 1.aw of September 9, 1965, in its current version, and permission for use must always be obtained from SpringerVerlag Berlin Heidelberg GmbH. Violations are liable for prosecution under the German Copyright Law.
http://www.springer.de
© SpringerVeriag Berlin Heidelberg 2003 Originally published by SpringerVerlag Berlin Heidelberg New York in 2003 Softcover reprint ofthe hardcover 2nd edition 2003
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.
Product liability: The publisher cannot guarantee the accuracy of any information about dosage and applica tion contained in this book. In every individual case the user must check such information by consulting the relevant literature.
The instructions given for the practical carryingout of HPLC steps and preparatory investigations do not absolve the reader from being responsible for safety precautions. Liability is not accepted by the author.
Cover Design: Klinkel & Lopka, Heidelberg Typesetting: Dataconversion by FotosatzService Kohler GmbH, Wlirzburg Production Editor: Christiane Messerschmidt, Rheinau
Printed on acidfree paper 52/3020  5 4 3 2 1
Preface
Differential Scanning Calorimetry (DSC) is a well established measuring method which is used on a large scale in different areas of research, development, and quality inspection and testing. Over a large temperature range, thermal effects can be quickly identified and the relevant temperature and the characteristic caloric values determined using substance quantities in the mg range. Measure ment values obtained by DSC allow heat capacity, heat of transition, kinetic data, purity and glass transition to be determined. DSC curves serve to identify sub stances, to set up phase diagrams and to determine degrees of crystallinity. This book provides, for the first time, an overall description of the most important DSC measuring systems and measuring programs, including the modulated temperature DSC. Furthermore many examples of typical and wide ly used applications of Differential Scanning Calorimetry are presented. Pre requisites for reliable measurement results, optimum evaluation of the mea surement curves and estimation of the uncertainties of measurement are, how ever, the knowledge of the theoretical bases of DSC, a precise calibration of the calorimeter and the correct analysis of the measurement curve. The largest part of this book deals with these basic aspects: The theory of DSC is discussed for both heat flux and power compensated instruments, and for the recently introduced modulated temperature mode of operation (TMDSC) as well. Temperature calibration and caloric calibration are described on the basis of thermodynamic principles. Desmearing of the measurement curve in differ ent ways is presented as a method for evaluating the curves of fast transitions. The instrumental data which are most important for the characterization of Differential Scanning Calorimeters are defined, and it is explained how they are determined experimentally. This enables every potential instrument buyer to ask the manufacturer for measured characteristic data which will allow him to compare the different instruments available. To make measurement results com parable, in addition to their traceability (via reference materials) a standardized evaluation of measurement uncertainty is indispensable. How this may be done is explained in some detail. We are indebted to S.M. Sarge for a critical examination of the manuscript and valuable suggestions for improvement, to S. Rudtsch for contributions to the chapter on calibration and to the expression concerning uncertainties and in particular to M. J. Richardson for helping us with the translation.
Braunschweig, Jena and Ulm March 2003
G. W. H. Hahne
w. F. Hemminger H.f. Flammersheim
Contents
1 
Introduction 
1 

2 
Types of Differential Scanning Calorimeters and Modes of Operation 
9 

2.1 
Heat Flux DSC 
10 

2.1.1 
Heat Flux DSC with DiskType Measuring System . 
10 

2.1.2 
Heat Flux DSC with TurretType Measuring System 
13 

2.1.3 
Heat Flux DSC with CylinderType Measuring System 
14 

2.2 
Power Compensation DSC 
. 
. 
. 
. 
. 
17 

2.2.1 
Function Principle 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
17 

2.2.2 
Special Power Compensating DSC 
22 

2.3 
DSC with Combined Heat Flux and Power Compensation 

Measuring System 
25 

2.4 
Modes of Operation 
25 

2.4.1 
Constant Heating Rate 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
26 

2.4.2 
Variable Heating Rate (Modulated Temperature) . 
27 

3 
Theoretical Fundamentals of Differential Scanning Calorimeters . 
31 

3.1 
Heat Flux DSC 
31 

3.2 
Power Compensation DSC 
48 

3.3 
TemperatureModulated DSC (TMDSC) 
50 

3.3.1 
The TemperatureModulated Method 
51 

3.3.2 
Influences of the Sample 
52 

3.3.3 
Influences of Heat Transport 
57 

3.3.4 
Conclusions 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
63 

4 
Calibration of Differential Scanning Calorimeters 
65 

4.1 
Aspects of Quality Assurance 
66 

4.2 
Basic Aspects of Calibration 
66 

4.3 
Temperature Calibration 
69 

4.3.1 
Temperature Calibration on Heating 
69 

4.3.2 
Temperature Calibration on Cooling 
84 

4.4 
Caloric Calibration 
86 

4.4.1 
Heat Flow Rate Calibration 
87 

4.4.2 
Heat (Peak Area) Calibration 
90 

4.4.3 
Examples of Caloric Calibration 
92 
4.4.4
Caloric Calibration in Cooling Mode
97
VIII
Contents
4.5 
Conclusions Regarding the Calibration of DSCs 
97 

4.6 
Reference Materials for DSC Calibration 
98 

4.6.1 
Reference Materials for Temperature Calibration 
99 

4.6.2 
Reference Materials for Heat Flow Rate Calibration 
101 

4.6.3 
Reference Materials for Heat (Peak Area) Calibration 
107 

4.7 
Additional Calibration in TemperatureModulated Mode of Operation 
107 

4.7.1 
Calibration of Magnitude 
113 

4.7.2 
Calibration of Phase 
113 

5 
DSC Curves and Further Evaluation 
. 
115 

5.1 
Characteristic Terms of DSC Curves 
. 
116 

5.2 
Parameters Influencing the DSC Curve 
118 

5.3 
Further Evaluation of DSC Curves 
119 

5.3.1 
Determination of the Real Sample Heat Flow Rate 
119 

5.3.2 
The Baseline and the Determination of Peak Areas (Enthalpy Differences) 
. 
121 

5.4 
Desmearing of the DSC Curve 
. 
126 

5.4.1 
Correction of the Temperature and Heat Flow Rate Indicated 
128 

5.4.2 
Subtraction of the Zeroline 
129 

5.4.3 
Calculation of the True Heat Flow Rate into the Sample 
129 

5.4.4 
Advanced Desmearing 
133 

5.4.5 
Further Calculations 
140 

5.5 
TMDSC Curves 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
. 
140 

5.5.1 
Reversing and NonReversing Heat Capacity 
141 

5.5.2 
Complex Heat Capacity 
143 

5.5.3 
Curves from StepScan Evaluation 
. 
145 

5.6 
Interpretation and Presentation of Results . 
145 

6 
Applications of Differential Scanning Calorimetry 
147 

6.1 
Measurement of the Heat Capacity 
147 

6.1.1 "Classical" ThreeStep Procedure 
148 

6.1.2 The ''Absolute'' Dual Step Method 
153 

6.1.3 General Precautions for the Minimization of Errors and their Estimation 
154 

6.1.4 
Procedure of Small Temperature Steps 
155 

6.1.5 
The TemperatureModulated Method 
159 

6.1.6 
Typical Applications of Heat Capacity Measurements 
160 

6.2 
Determination of Heats of Reaction 
162 

6.3 
Kinetic Investigations 
. 
168 

6.3.1 
Introduction and Definitions 
. 
168 

6.3.2 
Experimental Prerequisites for a Reliable Kinetic Analysis 
173 

6.3.3 
Selection of the Measuring Conditions  Isothermal or NonIsothermal Reaction Mode 
175 

6.3.4 
Activation of the Sample by UV Irradiation 
180 

6.3.5 
Different Strategies of Kinetic Evaluation . 
183 
6.3.6
Selected Examples and Possible Predictions
189
Contents
IX
6.4 
The Glass Transition Process 
200 

6.4.1 
The Phenomenology of the Glass Transition 
. 
200 

6.4.2 
The Nature of the Glass Transition and Consequences for DSC Measurements 
201 

6.4.3 
Definition and Determination of the Glass Transition Temperature Tg • • • • • • • • • • • • • • • • • • 
203 

6.4.4 
Applications of Glass Transition Measurements 
212 

6.4.5 
The Dynamic Glass Process, an Example 
217 

6.5 
Characterization of Substances, the Phase Behavior 
219 

6.5.1 
Applications in Biology and Food Science 
219 

6.5.2 
Applications in Pharmacy 
221 

6.5.3 
Other Applications 
224 

6.5.4 
Porosity Measurements 
228 

6.6 
Determination of Phase Diagrams 
230 

6.7 
Safety Aspects and Characterization of HighEnergetic Materials 
232 

6.8 
Characterization of Polymers 
233 

6.8.1 
Effects of Origin and Thermal History 
234 

6.8.2 
Determination of the Degree of Crystallinity 
236 

6.8.3 
Advanced Characterization with the TMDSC Method 
238 

6.9 
Purity Determination of Chemicals 
241 

7 
Evaluation of the Performance of a Differential Scanning Calorimeter 
245 

7.1 
Characterization of the Complete Instrument 
245 

7.2 
Characterization of the Measuring System 
245 

7.3 
Characterization of the Results of a Measurement: 

Uncertainty Determination 
251 

7.3.1 
Black Box Method . 
253 

7.3.2 
GUM Method 
253 

7.4 
Check List for DSCs 
257 

Appendix 1 
259 

Appendix 2 
263 

References . 
281 

Subject Index 
291 
List of Symbols
A 
area, preexponential factor 
C 
heat capacity, electric capacity 
D 
diffusivity 
E 
energy 
H 
enthalpy 
K 
factor (calibration), coefficient 
L 
thermal conductance 
P 
electric power, transfer function 
Q 
heat 
R 
resistance (thermal, electric), gas constant 
S 
entropy 
T 
temperature 
U 
internal energy, voltage 
V 
volume 
W 
work, electric energy 
a 
coefficient, apparatus function 
c 
specific heat capacity 
d 
distance 
i 
electric current 
k 
proportionality factor, calibration factor, rate constant, conductivity 
1 
length, distance 
m 
mass 
n 
reaction order 
p 
pressure 
r 
rate of reaction, radius 
t 
time 
w 
weight fraction 
x 
mole fraction 
a 
degree of reaction 
f3 
heating rate 
6 
phase angle 
y 
expansivity coefficient 
E 
emissivity 
A 
thermal conductivity 
v 
stoichiometric number 
tP 
heat flow rate 
p 
density 
a 
standard deviation 
XII
List of Symbols
r 
time constant 
w 
angular frequency 
X 
compressibility 
extent of reaction, composition, conversion 

Subscripts 

A 
activation, amplitude 
a 
amorphous 
bl 
baseline 
c 
extrapolated offset, crystalline 
cal 
calorimeter 
e 
extropolated onset 
el 
electric 
eq 
equilibrium 
exp 
experimental 
F 
furnace 
f 
final 
fix 
fixpoint 
fus 
fusion 
g 
glass 
h, i, k, n running numbers
initial 

iso 
isothermal 
I 
liquid 
lin 
linear 
lit 
literature 
M 
measurement point 
m 
measured 
mix 
mixing 
o 
onset 
p 
peak, constant pressure 
prop 
proportional 
4> 
related to heat flow rate 
Q 
related to heat 
R 
reference sample 
r 
reaction 
Ref 
reference material used for calibration (e. g. Certified Reference Material) 
S 
sample 
s 
solid 
st 
steady state 
th 
thermal 
tot 
total 
tr 
true 
trs 
transition 
u 
underlying 
V 
constant volume 
o 
zero, zero line 
1 Introduction
The objective of calorimetry is the measurement of heat. To measure heat means to exchange heat. The exchanged heat tends to effect a temperature change in a body that can be used as a measure of the heat exchanged, or the process of heat exchange creates a heat flow which leads to local temperature differences along its path which again serve as a measure of the .flowing heat. As chemical reactions and many physical transitions are connected with the generation or consumption of heat, calorimetry is a universal method for inves tigating such processes. Measuring devices in which an exactly known amount of heat is input into a sample, or abstracted from it, and the temperature change in the sample is measured (determination of the heat capacity, for example), are also referred to as calorimeters. Caloric measurements have been carried out since the middle of the 18th cen tury. Although modern Differential Scanning Calorimeters (DSC) are widely used today, the "classic" calorimeters cannot be dispensed with in precision measurements and for special applications. The most important classic calorime ters will be described only briefly in Appendix 2 to give the reader a more com prehensive survey of the field of calorimetry (for a more general presentation of calorimetry, cf. Hemminger, Hohne, 1984; Oscarson, Izatt, 1992). The topic of this book is Differential Scanning Calorimetry (DSC) using as a measuring instrument the Differential Scanning Calorimeter (DSC) available in various designs (see Chapter 2). In addition to the measurement of heat, DSCs are used to measure heat flow rates (power) and characteristic tempera tures of a reaction or a transition as well. The precise measurement of heat capacities, which is of an increasing importance, has distinctly been improved with modern DSCs. The measurement of heat not only includes integral (total) heats of reaction or transition but also the determination of "partial heats" developed within a selected temperature interval. Such values are of impor tance for kinetic evaluations, determination of crystallinity and purity (see Chapter 6). An accurate definition of Differential Scanning Calorimetry (DSC) is as follows:
Differential Scanning Calorimetry (DSC) means the measurement of the change of the difference in the heat flow rate to the sample and to a reference sample while they are subjected to a controlled temperature pro gram.
2
1 Introduction
It is important to understand that DSC measures the change of a property  namely of a heat flow rate difference  which normally is released due to an al teration of the sample temperature. When there is no alteration of the sample temperature, no change of a heat flow rate difference can be measured (except for possible chemical reaction heat flow rates). That means, a distinct tempera ture program, or in general a "mode of operation", is always part of a DSC mea surement (cf. Chapter 2). In the case of thermally activated reactions/transitions, e. g. by a controlled stepwise change from one constant temperature to another, the sample's temperature undergoes an alteration by internal processes and causes the DSC signal. It should even be mentioned, that heat only flows if there is a temperature difference present. In other words, a nonzero heat flow rate difference implies always a temperature difference between the sample and the reference and their surroundings and a change of the heat flow rate difference implies a change of the temperature as well.
Differential Thermal Analysis (DTA)
To distinguish a heat flux DSC from an apparatus for Differential Thermal Analysis (DTA), the latter will be characterized in the following. Differential Thermal Analysis is applied to measure the temperature differ ence between the sample to be investigated and a reference sample as a function of temperature (or time). This temperature difference indicates a heat exchange qualitatively. DTA is of more recent date than classic calorimetry. Its advantages as compared with conventional calorimetry are the dynamic mode of operation ("scanning") which allows reactions or processes to be investigated which can be thermally activated, and the high sensitivity to anomalies of the temperature time function. DTA allows characteristic temperatures to be determined and qualitative statements made on heat due to reaction. The further development of DTA has led to the construction of Differential Scanning Calorimeters (DSCs) with disktype measuring systems. The widely used term Differential Thermal Analysis (DTA) means a thermo analytical method, which is more than a thermoanalytical (measurement) tech nique (see the following definition), because the method (analysis) includes in principle the thermoanalytical investigation procedure, _{i}_{.} e. the evaluation and interpretation of the measured values. An accurate definition for the thermo analytical technique is:
Differential Thermometry (DT) means the measurement of the change of the difference in temperature between the sample and the refer ence sample while they are subjected to a controlled temperature pro gram.
The accurate term for the affiliated thermo analytical method is then Differen tial Thermal Analysis (DTA). For a short description of the principle of DTA measuring systems in comparison with DSC, see Appendix 1.
1 Introduction
_{3}
Calorimetry and OTA Today
Owing to the new materials and sensors used and the application of modern mechanical manufacturing (micro and nano technologies), advanced data pro cessing by modern electronic systems, all calorimeters available today  includ ing the socalled "classic" calorimeters  are instruments which allow precise measurements to be carried out with high sensitivity and repeatability (see Sect. 7.2), their operation being relatively simple or perhaps even automated. Caloric methods are used in many fields for quality assurance purposes _{(}_{c}_{f}_{.} series of standards ISO 9000). The standards to be applied in the field of quality assur ance (e.g., the standard ISO/IEC 17025, 1999: General requirements for the competence of testing and calibration laboratories) demand a large variety of measures to ensure reliable measurement results, for example, the application of wellproven methods for the preparation of samples, reliable measurement methods, calibration methods and calibration materials traceable to standards (see also Chapter 4). In the following, reference is made to some of the fields in which calorimeters are widely used today (see Hemminger, Hohne, 1984; Wunderlich, 1990; Oscar son, Izatt, 1992). A suggestion how to classify calorimeters is given in Appendix 2.
 With bomb calorimeters, combustion heats (gross calorific values) are mea sured on a large scale in industry (costs of fossil fuels) (Sunner, Mansson,
1979).
 Gas calorimeters (flow calorimeters) are used for the continuous or discon tinuous measurement of the calorific value of fuel gases (e.g., natural gas), both in frontiercrossing commercial transactions (between supplier and buyer) and for the calculation of the costs for the individual consumer (Hyde, Jones, 1960; Hemminger, 1988; Sarge, 1997; Ulbig, Hoburg, 2002; Jaeschke, 2002; Dale et al., 2002; Alexandrov, 2002).
 Drop calorimeters (usually selfmade) allow mean heat capacities or enthalpy differences to be quickly measured. Drop sample temperatures of up to 2000°C are realized (Chekhovskoi, 1984).
 Different types of isoperibol mixing calorimeters are used to investigate reac tions between two fluids or between a fluid and a solid (reaction heats, heats of solution, heats of mixing, adsorption heats) (see, e.g., Parrish, 1986).
 Reaction calorimeters and safety calorimeters allow model tests of procedures
applied in industrial chemistry to be carried out. These instruments provide valuable support in development and optimization tasks, as all test parame ters (temperature, time, addition of substances etc.) are completely docu mented and can be varied automatically (Landau, 1996). Questions of process technology which are also important for the dimen sioning of production facilities, may be well in the fore here (Regenass, 1985). Reaction heat released is connected with the degree of conversion and exploitation, for example, in biotechnology, where calorimetric studies are carried out on industrial production plants (largescale calorimetry, megacalorimetry, see, e.g., von Stockar, Marison, 1991 in Lamprecht et aI.,
1991).
_{4}
1 Introduction
Special calorimeters (commercially available) are also used to investigate aspects of safety technique, for example, to determine characteristic temper atures of decomposition or the kinetics of reactions (decompositions, run away reactions) under certain boundary conditions (see, for example, Grewer, 1987; Schwanebeck, 1991; Singh, 1993; for a survey,cf. Grewer,Steinbach, 1993).
 In highly sensitive flow calorimeters (usually of the isoperibol type, i. e., with surroundings at constant temperature), the heat generation of biological sys tems and their change at varying conditions of life are investigated (for ex ample, addition of pharmaceuticals to bacterial cultures). In different types of biocalorimeters, the metabolism of organisms and their change due to exter nal influences (optical, acoustical, mechanical, thermal, chemical) are studied (see, e.g., Spink, Wadso, 1976; Lamprecht, Schaarschmidt, 1977; Beezer, 1980; Lamprecht et aI., 1991).
 Heat capacities and heats of transition are directly measured with high accu racy using adiabatic calorimeters. The characteristic data of materials deter mined with their aid are an indispensable basis for the calibration of the DSCs (see, e.g., Gr0nvold, 1967; Kagan, 1984; Jakobi et aI., 1993; Gr0nvold, 1993). Properly calibrated DSCs allow specific heat capacities to be measured with an accuracy of 1 to 2% (see Sect. 6.1; cf. Richardson, 1992a).
New fields of application are constantly being opened up for the modern, highly automated DSCs. DSC is increasingly used in the field of quality assurance for many purposes: for the inspection of raw materials, as an accompanying mea sure in the manufacture and for the control of the finished products. The basic limitations to, and the problems of these instruments should not, however, be forgotten in view of the ease of operation and evaluation. DSCs allow reaction heats and heats of transition, or heat flow rates and their changes at characteristic temperatures, to be quickly measured on small sample masses (milligram range; in the case of classic calorimeters: gram range), in wide temperature ranges and with an accuracy which is usually sufficiently high for the respective purpose. DSCs are applied in the following fields (among others):
 characterization of materials (in particular polymers),
 comparison (relative) measurements (quality control, identification of sub stances or mixtures),
 stability investigations,
 evaluation of phase diagrams,
 purity determinations,
 kinetic investigations,
 safety investigations,
 determination of heat capacity and complex heat capacity (with TMDSC).
Now as before, instruments used for Differential Thermal Analysis (DTA) offer particular advantages when special problems are to be investigated:
 they can be used at very high temperatures (up to about 2400°C),
 they are highly sensitive,
 they are most flexible as regards the volume and form of the crucibles,
 their reasonablypriced measuring system can be easily exchanged.
1 Introduction
_{5}
Characteristic temperatures of transitions or reactions can be very well deter mined by DTA. Heats can still be estimated with an uncertainty between 20 and 50 %. DTA is applied in the following fields:
 comparison (relative) measurements (identification, control, comparison),
 safety research (stability investigations, also longtime investigations), see, e. g., Hentze, 1984; Hentze, Krien, 1986,
 investigation of transitions, decompositions, reactions with gases,
 evaluation of phase diagrams.
DSC and DTA are also used together with other methods of thermal analysis or other analytical techniques (simultaneous thermal analysis), most frequently in connection with thermogravimetry (TG), more rarely with Evolved Gas Ther moanalysis (EGA), Thermomicroscopy (TOA) or Thermosonimetry (TS). The coupling of DSC or DTA with Thermogravimetry (TG) is of particular importance. In addition to information on changes in the heat flow rate (due, for example, to changes of Cp) and heats of transition, the TG signal provides infor mation on whether volatile components are involved and which changes in mass are to be attributed to a transition. The different methods of gas analysis used together with DSC or DTA (usual ly together with TG/DSC or TG/DTA systems) are of increasing importance. Via a carrier gas stream, gaseous reaction products from the calorimeter or DTA device are transported to an apparatus (e. g., IR or mass spectrometer) in which these products are analyzed. This allows a correlation between the characteris tic temperature of a reaction, gravimetric and/or enthalpic information and the composition of the volatile reaction product to be made (see examples in
Mathot,1994a).
It is difficult to transfer the volatile products without adulteration from the hot sample to the analysis system (condensation must be avoided). Various sys tems for coupling quadrupole mass spectrometers are commercially available (up to temperatures of 2400 0c). Examples of the investigation of volatile com ponents applying methods of gas analysis can be found in Ohrbach et al., 1987; Matuschek et al., 1991; Matuschek et al., 1993; Kaisersberger, 1997. The advantage of simultaneous measurements is that the same sample is investigated under identical conditions and that diverse information is ob tained by one measurement run which is important for interpretation pur poses. The following potential disadvantages of simultaneous instruments should be mentioned: lower sensitivity, higher susceptibility to failure, in creased time and effort required for preparation and operation, higher instru ment costs. In Differential Scanning Calorimetry which is widely applied, there are still some actual problems of which one should be aware:
 The theoretically wellfounded complete understanding of DSC is not yet totally accepted. The average user does not yet know the limits to this method and the sources of the systematic errors by which it is affected. This is in par ticular true for the temperaturemodulated DSC, its theoretical background is still a matter of controversial discussion.
_{6}
1 Introduction
 There are not yet any practicable and experimentally tested recommenda tions for temperature and heat calibration and for measurement procedures which are internationally accepted (see Chapter 4; and some examples in Della Gatta et al., 2000).
 There is no international agreement on a single set of substances for the tem perature and heat calibration of DSC which have been measured with suffi cient accuracy, including a metrologically sound traceability to national/in ternational standards; instead there is a confusing variety of "certified refer ence materials" whose characteristic data are in part contradictory. Sub stances which are certified on a metrologically basis (i. e., traceable to the SI units) are offered by the German Metrology Institute (PTB) (cf. Sect. 4.6).
Due to the rapidly increasing use of DSCs in various fields of application, some negligence has gained ground  favored by the ease of operation and evaluation
 which would be inconceivable in "classic" calorimetry. Remedial measures
should be taken; for example, the national societies of the International Con
federation for Thermal Analysis and Calorimetry (ICTAC) should offer special training courses, and precise specifications for instruments and programs should be drawn up. The following can frequently be observed:
 A realistic estimate of the uncertainty of measurement is rarely made (cf. Sect. 7.3). The calibration capability (cf. Chapter 4), different influencing quantities (cf. Sect. 5.2) and known theoretical considerations (cf. Chapter 3) should be taken into account. In many cases, the repeatability of a DSC is, for example, simply, but wrongly, indicated as accuracy of the measured data (cf. Sect. 7.3).
 Interpretation of the DSC measurement results is often insufficient or faulty, when
 uncertainties of measurement are not taken into consideration;
 systematic error sources are disregarded;
 the measured curve is not "desmeared" (cf. Sect. 5.4);
 the laws of thermodynamics, kinetics, are not taken into account;
 uncritical confidence is placed in the evaluation programs provided by the manufacturer;
 the results are not confirmed by other measuring methods.
A great number of calorimetric methods has not been mentioned here as, in
comparison with DSC, they are used only in specific fields. To name a few ex
amples:
 lowtemperature calorimetry for measuring heat capacities (see, e. g., Gmelin,
1987),
 more
recent techniques of c _{p} measurement (see,
1981; Maglic et aI., 1984),
e.g., Lakshmikumar, Gopal,
 measurement of the energy of particle radiation (see, e. g., Domen, 1987),
 measurements on biological systems (see, e. g., Lamprecht et al., 1991; Wadso, 1993 and Kemp, 1993; Kemp et aI., 1998; Hansen, 2000),
1 Introduction
7
 deformation calorimetry on polymers (see, e. g., Kilian, Hahne, 1983; Godovsky, Hahne, 1994),
 hightemperature calorimetry in material science (see, e.g., Bruzzone, 1985; Bros, 1989).
 (high)pressure calorimetry (see,
 AC calorimetry (Sullivan, Seidel, 1968) and 3w calorimetry (Rosenthal, 1961).
e. g., in Mathot, 1994 b; Hahne, 1999 b),
Summary
During the last decades the development of Differential Scanning Calorimetry from a halfquantitative DTA method (producing "thermograms") toward a true calorimetric method (yielding quantitative caloric quantities) forged ahead. This becomes manifest from the following items which hopefully will be proved within this book.
 The theory of (heatflux) DSCs is wellknown and the function principle un derstood.
 Systematic (unavoidable) error sources and methods of correction are known.
 Modern data treatment and powerful computers make it possible for every one to make the necessary evaluations without problems.
 Metrological founded calibration and measurement procedures improved the certainty of the results which nowadays are comparable with those obtained with classical calorimeters.
 The difference in quality of the results from heatflux and powercompensat ed DSCs has diminished.
 Precise heat capacity measurements are possible and open the door for other welldefined thermodynamic quantities.
 Coming to terms with the problem of the thermal inertia (thermal lag) en ables kinetic investigations which otherwise need much more experimental efforts.
 DSC is a fast and reliable method in a wide field of different applications.
2 Types of Differential Scanning Calorimeters and Modes of Operation
Two basic types of Differential Scanning Calorimeters (DSCs) must be distin guished:
 The heat flux DSC,
 The power compensation DSC.
They differ in the design and measuring principle what will be dealt with in this chapter. The theoretical bases are presented later in Chapter 3. Common to all DSC's is a differential method of measurement which is defined as follows:
A method of measurement in which the measured quantity (the measurand)
is compared with a quantity of the same kind, of known value only slightly dif ferent from the value of the measurand, and in which the difference between the two values is measured (International Vocabulary of Basic and General Terms in Metrology, 1984). The characteristic feature of all DSC measuring systems is the twintype design and the direct indifference connection of the two measuring systems which are of the same kind. It is the decisive advantage of the difference princi ple that, in first approximation, disturbances such as temperature variations in the environment of the measuring system and the like, affect the two measuring systems in the same way and are compensated when the difference between the individual signals is formed. Moreover, the difference signal  which is the measurement signal actually of interest  can be strongly amplified, as the high basic signal (signal of the indi vidual measuring system) is almost compensated when the difference is formed. An extension to form multiple measuring systems (three or four) connected back to back does not mean a fundamental change of the difference principle. The differential signal is the essential characteristic of each Differential Scan ning Calorimeter. Another characteristic  by which it is distinguished from most classic calorimeters  is the dynamic mode of operation. In other words, the DSC can be operated by applying various "modes of operation": The tem perature can not only be held constant but also raised or lowered at a preset rate which might be superimposed with a temperature modulation (see below).
A characteristic common to both types of DSC is that the measured signal
is proportional to a heat flow rate iP and not to a heat as is the case with most
of the classic calorimeters. This allows time dependences of a transition to be observed on the basis of the iP(t) curve. This fact  directly measured heat flow rates  enables all DSCs to solve problems arising in many fields of application (see Chapter 6).
_{1}_{0}
2 Types of Differential Scanning Calorimeters and Modes of Operation
2.1
Heat Flux OSC
The heat flux DSC belongs to the class of heatexchanging calorimeters (for the classification, see Appendix 2). In heat flux DSCs a defined exchange of the heat to be measured with the environment takes place via a welldefined heat con duction path with given thermal resistance. The primary measurement signal is a temperature difference; it determines the intensity of the exchange and the resulting heat flow rate tP is proportional to it. In commercial heat flux DSCs, the heat exchange path is realized in different ways, but always with the measuring system being sufficiently dominating. The most important fundamental types are:
 The disktype measuring system, where the heat exchange takes place via a disk which serves as solid sample support. Features: Simple and easily realizable design with a high sensitivity, the sample volume is small, but the heat exchange between furnace and sample is limited which allows only medium heating and cooling rates.
 The turrettype measuring system, where the heat exchange takes place via small hollow cylinders which serve as elevated sample support. Features: More sophisticated design with high sensitivity and fast thermal response which allows large heating and cooling rates, the sample volume is small.
 The cylindertype measuring system, where the heat exchange between the (big) cylindrical sample cavities and the furnace takes place via a path with low thermal conductivity (often a thermopile). Features: Very sensitive with a large sample volume but with a large time constant which allows only low heating rates, the sensitivity per unit volume is, however, very high.
2.1.1
Heat Flux OSC with OiskType Measuring System
The characteristic feature of this measuring system is that the main heat flow from the furnace to the samples passes symmetrically through a disk of medi um thermal conductivity (Fig. 2.1 a). The samples (or the sample containers) are positioned on this disk symmetrical to the center. The temperature sensors are integrated in the disk. Each temperature sensor covers more or less the area of support of the respective container (crucible, pan) so that calibration can be car ried out independent of the sample position inside the container (cf. Sect. 4.3). To keep the uncertainties of measurement as small as possible, the arrangement of sample and reference sample (or of the containers) and temperature sensor in relation to one another and to the support must always be the same (center pin or the like). Metals, quartz glass or ceramics are used as disk materials. Type (and number) of the temperature sensors (e.g., thermocouples, resistance thermometers) dif fer. The use of modern sensors on the basis of semiconducting material leads to a significant increase in sensitivity.
2.1
Heat Flux DSC
11
3
rzzzzzzzz£l~zzzhzzZl
_{a}
S
R
6T
5
T( f) 17>',.
calibra ticn
KIT)
b
time, temperature
~
Fig.2.1. a Heat flux DSC with disktype measuring system. 1 disk, 2 furnace, 3 lid, 4 differ ential thermocouple(s), 5 programmer and controller, S crucible with sample substance, R crucible with reference sample substance, IPps heat flow rate from furnace to sample crucible, IPpR heat flow rate from furnace to reference sample crucible, <Pm measured heat flow rate, K calibration factor. b Measured heat flow rate <Pm (schematic curve) (according to Hem minger, 1994)
When the furnace is heated (in general linearly in time, more recently also in a modulated way), heat flows through the disk to the samples. When the arrange ment is ideally symmetrical (with samples of the same kind), the same heat flows into sample and reference sample. The differential temperature signal6.T (normally in form of an electrical potential difference) is then zero. If this steadystate equilibrium is disturbed by a sample transition, a differential signal
_{1}_{2}
2 Types of Differential Scanning Calorimeters and Modes of Operation
is generated which is proportional to the difference between the heat flow rates to the sample and to the reference sample:
c;llFS
c;llFR 

AT
with
AT= Ts TR
As neither ideal thermal symmetry of the measuring system at all operating temperatures nor thermal identity of the samples can be attained in practical application, not even outside the transition interval, there will always be a signal ATwhich depends on the temperature and the sample properties (cf. definitions of zero line in Sect. 5.1). Calculations in Chapter 3 are based on the assumption that this portion of the total signal is zero or has already been subtracted from the measurement signal proper. In the DSC the measurement signal AT is always obtained as electric voltage. The heat flow rate c;llm (m: measured) is internally (in the software) assigned to this signal AT by factoryinstalled provisional calibration:
c;llm = k'·AT
The measurement signal output by the DSC (Fig. 2.1 b) and accessible to the user is c;llm (in IIW or mW). The calibration of the DSC must be checked by the user:
to what extent does c;llm represent the true heat flow rate c;lltrue which is released or consumed by the sample? Such a test (verification of calibration) can be carried out
 by measuring the steadystate heat flow rate into a sample of known heat capacity C by "charging" this heat capacity in the quasisteadystate of scan ning with a constant heating rate f3 =dT/dt:
(2.1)
 by comparing the integral over a transition peak with the expected (known) heat of transition Qr (energy balance):
(2.2)
(for definition of baseline 4\" see Sect. 5.1) (When resistance thermometers are used as temperature sensors, the ther mometers can be used also as heaters which is advantageous in calibration pro cedures, see Wolf et aI., 1994). Questions of calibration capability and calibration are dealt with in Sect. 4.4; problems of "desmearing" arising in connection with it are discussed in Sect. 5.4. Heat flux DSCs with disktype measuring systems are available for tempera tures between 190°C and 1600°C. The maximum heating rates are about 100 K minI. Typical time constants of the ATsensor (empty systems, no samples) are between 3 and 10 seconds. The noise (definition see Sect. 7.2) of the measure ment signal lies between 0.511W and 20 llW (it also depends on the temperature and the heating rate). The total uncertainty of the heat measurement amounts to
2.1
Heat Flux DSC
_{1}_{3}
about 5 %, and it is to be expected that it cannot be reduced to less than 2 % even if more time and effort were spent (cf. Sect. 4.4).
Heat Flux DS( with Triple Measuring System
An extension of the disktype measuring system described above was presented by TA Instruments (formerly Du Pont). It is a "Dual Sample DSC" with three locations on the metallic (constantan) disk to receive sample crucibles. The three locations are equipped with temperature sensors (cf. Fig. 2.1 a for the normal DSC) which can measure the temperature differences between them. This configuration makes it possible to measure the heat capacity in a single run, using one measuring position for the sample, the second for an empty cru cible, and the third for the calibration (reference) material with well known heat capacity (Jin, Wunderlich, 1993; see even Sect. 6.1). The temperature range of this Dual Sample DSC is from 125 to 1000 K with heating rates up to 100 K minI. The development of an adiabatic high temperature triplecell DSC system (up to 1500 K) has been described by Takahashi, Asou, 1993.
Pressure Heat Flux DS(s
Special containers for heat flux DSCs with disktype measuring system allow high er pressure, in general up to 7 MPa (70 bar), to be applied. This makes it possible to determine, for example, vapour pressures and heats of evaporation by means of DSC (Wiedemann, 1991; Perrenot et al., 1992, see also in Mathot, 1994b).
2.1.2
Heat Flux DSC with TurretType Measuring System
The characteristic feature of this recently developed measuring system is that the essential heat flow passes from the bottom of the furnace through the jacket of two thinwalled cylinders to the top of them which serve as sample and ref erence sample support (cf. Fig. 2.2). Because of the short time since introduction of this system there are only few experiences yet, but it seems to offer some ad vantages against the common disktype DSC which are:
 a small system size combined with low mass,
 a very short heat conducting path,
 a strictly direct connection of the sample and reference sample to the furnace,
 no interference (cross talk) between sample and reference sample events.
This leads to a smaller furnace as well and the thermal response time becomes much lower, which allows faster heating and cooling rates and higher frequen cies in the temperaturemodulated mode. In addition to the temperatures of the sample and reference sample support the temperature in the middle of the bottom plate To (see Fig. 2.2) is measured in this measuring system. With these data it is possible to determine additional quantities which are essential for the performance of the DSC and correct for the influence of thermal inertia. The theoretical background of this socalled
14
2 Types of Differential Scanning Calorimeters and Modes of Operation
R
~::::::::1
Fig.2.2. Heat flux DSC with turrettype measuring system (TA Instruments). 1 elevated con stantan platform for sample and reference sample, 2 chromel area thermocouple, 3 constan tan body, 4 chromel  constantan thermocouple, 5 silver furnace, S sample substance, R reference sample substance,~T platform temperature difference, To body (furnace) temper ature
Tzero™ DSC technology will be outlined in Chapter 3, it is much more sophisti cated than the simple formulae presented in Sect. 2.1.1 for the simple disktype DSC. Modern computer techniques offer the possibility of expensive online cal culations and enable the output of the corrected heat flow rate and the calculat ed temperature of the sample pan even while the measurement runs. The tiny design necessitates a very accurate production of the components with very narrow tolerances to ensure a good symmetry which is indispensable for the construction of differential calorimeters. However, modern technology enables this new type of heat flux DSC which seems to be a large step in the right direction to overcome the wellknown drawbacks of heat flux DSC (see Chap ter 3) in near future.
2.1.3
Heat Flux DSC with CylinderType Measuring System
A blocktype cylindrical furnace is provided with two cylindrical cavities, each containing a cylindrical, fixed sample container which is connected with the fur nace or directly with the other container by means of several thermocouples (thermopiles), which are the characteristic features of this type of measuring system. In the original cylindertype measuring system (according to Calvet,1948, Fig. 2.3), the outer surfaces of each sample container are in contact with a great number of thermocouples connected in series between the container and the furnace. The thermocouple bands or wires are the dominating heat conduction path from the furnace to the samples. Heat conduction path and temperature difference sensors are identical. Both sample containers are thermally decou pled; heat exchange takes place only with parts of the massive furnace. The
2.1
Heat Flux DSC
15
2
4
2
3
'0 !1 T 0'
Fig.2.3. Heat flux DSC with cylindertype measuring system (Calvet, 1948; thermally decou pled sample containers) (according to Hemminger, 1994). 1 containers to take up sample and reference sample, 2 thermopiles, 3 furnace (with program mable temperature controller), 4 lid, S sample substance, R reference sample substance, 1'1T temperature difference between the containers
measurement signal proper is the temperature difference I1T of both sample containers averaged over the surfaces; it is generated by differential connection of both thermopiles:
By analogy with the disktype DSC, the following is valid for the (steadystate) heat flow rates exchanged between furnace and sample _{!}_{P}_{F}_{S}_{,} between furnace and reference sample !PFR, and for the measured heat (output) flow rate !Pm and the true heat flow rate !P _{t}_{r}_{u}_{e} into the sample:
!PFS 
!PFR 
I1T
!Pm = k' ·I1T
!Ptrue = K~· !Pm
An electric voltage proportional to I1T, or a heat flow rate signal internally cal culated from I1T (as in the case of the DSCs with disktype measuring system, cf. Sect. 2.1.1) is put out as a measurement signal. It must be checked in both cases to what extent the measurement signal corresponds to the (steadystate) heat flow rate actually exchanged or to what extent the integral over the measured peak corresponds to the known heat of transition (cf. Sects. 4.4 and 5.4). In modified cylindertype measuring systems (for example, according to Petit et al., 1961), only a small fraction of the heat usually flows from the furnace to the samples via thermocouples (Fig. 2.4). The greatest part of the heat exchange be
16
2 Types of Differential Scanning Calorimeters and Modes of Operation
4
3
5
Fig.2.4. Heat flux DSC with modified cylindertype measuring system (thermally coupled sample containers) (according to Hemminger, 1994). 1 containers to take up sample and reference sample, 2 thermopile(s), 3 furnace (with pro grammable temperature controller), 4 lid, 5 support of containers, S sample substance, R ref erence sample substance, I1T temperature difference between the containers
tween furnace and samples takes place via the holders of the sample containers, via leads and gas layers between furnace and sample containers. Thermocouples are preferably used to directly measure the temperature difference between sample and reference sample containers. Both containers are no longer thermally decoupled, however, the temperature difference measured between them is again proportional to the differential heat flow rate from the furnace to the samples, part of which is now also exchanged between the two containers. In all cylindertype measuring systems, the calibration factor may depend on the position and height of the sample in the container, the reason being that the efficiency of heat conduction paths, such as holders, leads and gas layers, depends on the position of the heat source (sample) in the container. When the sample comes close to the upper edge of the container, the calibration factor can change by more than 50 %, depending on the calorimeter type (cf. Sect. 4.4). Compared with disktype measuring systems, the classical cylindertype measuring system according to Calvet has the advantage of a much larger useful volume. In return, greater thermal inertia (time constants of up to 40 min) must
usually be put up with. The large cavity volume allows different special contain ers (for example, for electric calibration, gas flow, mixture, high pressure etc.) to
be used and makes intervention (electrical, mechanical, gas exchange,
direct observations (acoustical, optical) in the sample cavity possible. The great number of thermocouples (up to 1000) generates a high output signal at rela
tively low noise. Heat flux DSCs with cylindertype measuring systems with sample volumes of about 10 mL are available for a temperature range between 196°C and 15000c. Due to the relatively great time constants of the measuring systems, the
) and
2.2
Power Compensation DSC
17
maximum heating rates  depending on the cavity volume  lie at about 30 K
minI. Special instruments with large volumes
ample, in biology for the investigation of small animals at constant temperature (cf. Moratzky et al., 1993; Schmolz et al., 1993). High temperature systems can also be used in isoperibol (isoperibol: sur roundings at constant temperature) mode of operation, as a drop calorimeter (cf. Appendix 2) to measure, for example, enthalpies of mixture up to 1500°C (see, e.g., Fan et al., 1993). Other systems with smaller sample volume (around 850 }lL) and semicon ductor temperature sensors are used for highly sensitive measurements (detec tion limit around 50 nW). They are applicable in a temperature range from 45 to 120°C and can be equipped with special vessels according to various investi gations (batch, ampoule, fluidcirculation, mixing, electric calibration by means of a "Jouleheater", etc.). These DSCs can usually be heated maximally with 1 K minI, isothermal operation is possible as well (noise ca. 0.2 }lW). For heat flux DSCs with cylindertype measuring systems according to Calvet (decoupled sample containers), a rather simple theory has been developed which quantitatively describes the functional correlation between the instanta neous measurement signal IPm and the original event IPr in the calorimeter (cf. Sect. 3.1). According to this theory, "desmearing" (cf. Sect. 5.4) of the measured
curve  to obtain the IPr(t) function  can be performed rather easily, even on line, for example, by means of a commercially available electronic devices or via proper software on the computer. For disk and turrettype heat flux DSCs the desmearing procedure is not that simple but can be performed as well. For the latter DSC type the advanced Tzero™ technology makes it possible to do the necessary calculations during the measurement run (see Sects. 2.1.2 and 5.4.3).
(up to 100 cm 3 ) are used, for ex
2.2
Power Compensation DSC
The power compensation DSC belongs to the class of heatcompensating calorimeters (see Appendix 2). The heat to be measured is (almost totally) com pensated with electric energy, by increasing or decreasing an adjustable Joule's heat.
2.2.1
Function Principle
The commercial power compensation DSC I
ment with an isoperibol operation. The measuring system (Fig. 2.5) consists of
two identical microfurnaces which are mounted inside a thermostated alu minum block. The furnaces are made of a platinumiridium alloy, each of which contains a temperature sensor (platinum resistance thermometer) and a heating
most frequently used is an instru
1 The description of a power compensation DSC is based on the widely used DSC of Perkin Elmer Instruments.
_{1}_{8}
2 Types of Differential Scanning Calorimeters and Modes of Operation
Fig.2.S. Power compensation DSC (PerkinElmer Instruments). Setup of the measuring sys tem (according to Hemminger, 1994). S sample measuring system with sample crucible, microfurnace and lid, R reference sample system (analogous to S), 1 heating wire, 2 resistance thermometer. Both measuring systems  separated from each other  are positioned in a surrounding (block) at constant temperature
@
Fig.2.6. Power compensation DSC (PerkinElmer Instruments). Block diagram showing the function principle (according to Hemminger, 1994). Ts temperature of the sample furnace, TR temperature of the reference sample furnace, IlT = Ts  TR, P _{a}_{v} average heating power, IlP compensation heating power, cf>m measured heat flow rate (measurement signal)
2.2
Power Compensation DSC
_{1}_{9}
resistor (made of platinum wire). The microfurnace is about 9 mm in diameter, approx. 6 mm in height and has a mass of approx. 2 g. The time constant of the DSC without sample pans is 1.5 s, and the isothermal noise is about 2 II W. The maximum heating power of a microfurnace is about 15 W, the maximum heat ing rate is 500 K minI. Maximum cooling rate can reach 200 K minI, it depends on the temperature difference between block and sample (PerkinElmer Pyris Diamond DSC). The measuring range extends from 175°C (block cooled with liquid nitrogen) to 725°C, During heatingup, the same heating power is supplied to both microfurnaces via a control circuit (Fig. 2.6) in order to change their mean temperature in accordance with the preset heating rate (see Watson et aI., 1964 2 ). If there is ideal thermal symmetry, the temperature of both microfurnaces is always the same. When an asymmetry occurs, for example, as a result of a sample reaction, a tem perature difference results between the microfurnace accommodating the sam ple and the microfurnace containing the reference sample. The temperature dif ference is both the measurement signal and the input signal of a second control circuit. This second circuit compensates most of the reaction heat flow rate by proportional control by increasing or decreasing an additional heating power of the sample furnace. The compensating heating power I1P is proportional to the remaining temperature difference I1T (because of the proportional controller). The time integral over the compensating heating power is proportional to the heat Qr which was consumed or released in the sample (Fig. 2.7). Again, a heat flow rate <Pm is assigned to the real measurement signal I1T as a result of a factoryinstalled calibration, and fed in. The relations between I1T, <Pm and the compensating heating power I1P are as follows:
I1P=k _{l} ·I1T <Pm =k _{2} ·I1T
The factor kl is a factoryset fIxed quantity of the proportional controller, k2 can be changed at the instrument with the aid of a potentiometer or it is adjusted via the software (calibration). The factor k2 is almost independent of measurement parameters (e.g., temperature), as  via kl  a given compensating heating power always corresponds to a given 11T, independent of the temperature; k2 can there fore in principle be determined by one calibration measurement at one temper ature only. As regards the formal aspects (cf. Sect. 2.1 for the heat flux DSC), the output signal is also given as a heat flow rate signal _{<}_{P}_{m} (e. g. in mW), and the relation be tween <Pm and the true heat flow rate exchanged with the sample, <P _{t}_{r}_{u}_{e} = K<[>· <Pm, must also be determined by caloric calibration (see Sect. 4.4). At higher temperatures of the measuring system, the heat flows exchanged with the (isoperibol) surroundings (conduction, radiation, convection) are relatively
2 In the basic paper by Watson et al. the term Differential Scanning Calorimeter (DSC) is coined, but the title of the paper is: A Differential Scanning Calorimeter for Quantitative Dif ferential Thermal Analysis, i. e., the inventors of this DSC recognized clearly the systematic relationship between this type of DSC and classic DTA.
20
2 Types of Differential Scanning Calorimeters and Modes of Operation
t
t
tJp
t
tJT
t
timel>
Fig.2.7. Power compensation DSC. Diagrammatic view of the signals in question (according to Hemminger, 1994). tPr heat flow rate released in the sample (exothermic, therefore negative), qr heat released in the sample, ~T* temperature increase in the sample furnace which would build up due to the exothermic effect unless it was compensated (as in heat flux DSC), _{~}_{p} compensation heating power (negative, to compensate the exothermic effect), _{~}_{T} temperature increase actually occurring in the sample furnace (corresponds to the residual deviation from the theoretical
value, which cannot be completely compensated by the proportional control), tP _{m} output sig
nal (proportional to the (negative) amplified ~T signal)
2.2
Power Compensation DSC
21
large (several Watt) compared with the quantity to be measured which is three orders of magnitude lower. High requirements must, therefore, be met as far as the uniformity of the heat exchange between the two microfurnaces and the sur roundings is concerned in order to keep a good symmetry and the uncertainties of measurement small. Moreover, the shares of the various heat exchange mecha nisms and their respective amounts must depend only on the temperature, and strict repeatability must be ensured. This is a rather hard demand if one considers the huge heat exchange of the furnaces with the thermostated surroundings via radiation and convection. Consequently the conditions for this heat exchange should be kept as constant as possible. Conclusion: The microfurnaces must be covered with lids of the same kind in order to "cover up" possible inhomogeneity of sample and reference sample; the thermophysical properties of the crucibles and lids must depend only on the temperature and never on sample properties. As a result an asymmetry of heat exchange with the surroundings of, say, 10 4 causes a differential signal of about 1 mW, the zero line of the power compen sated DSC is never zero and often strongly curved. The manufacturer made allowance for these problems and took care of an electronically compensation of thermal asymmetries of the measuring system, which become apparent as a cur vature of the zero line. In this way, it is also possible to "straighten" the zero line and/or incline it as desired, and with it the measured curve outside the peaks. Compared with heat flux DSCs, the power compensation DSC offers the fol lowing advantages:
 The short heat conduction path between samples and heater and the relative ly small masses of the microfurnaces allow an almost instantaneous response to a sample reaction. Due to the small time constant, desmearing is required only in a few cases (cf. Sect. 5.4).
 Reaction heat flow rates are rapidly and to a large extent compensated by elec trical heating power. As a result, only small temperature differences I1T re main between the microfurnaces of sample and reference sample (approx. 1110 of those occurring in heat flux DSCs using the same sample and refer ence sample). This means that the calibration factor Kcp is practically inde pendent of the intensity and kinetics of the sample reaction.
 The total compensating energy (J I1Pdt) is equal to the reaction heat or heat of transition.
 The temperature dependence of the control circuit properties (above all of the temperature sensors and the heating elements) is known and strictly re peatable. It can be taken into account with the aid of special electronics or by the software. A single caloric calibration is then  in principle  sufficient to determine the correlation between CPm (or CPr) and I1P (cf. Sect. 3.2).
The user must, however, keep in mind that the temperature difference between the two microfurnaces is not compensated totally in commercial power com pensation DSCs. During the peak there is still a temperature difference 11T pro portional to the reaction heat flow rate. That is to say, the power compensation DSC can be considered a kind of DTA instrument with a 11T as the measurement signal, but with a I1T arising from a specific thermal event which is much small er than that developed in a heat flux DSC measuring system.
22
2 Types of Differential Scanning Calorimeters and Modes of Operation
In conclusion it may be said that all the attributes of heat flux DSC systems which depend on this temperature difference can also be found in the "real" power compensation DSC, but to a lesser degree. The calibration factor in par ticular is not a constant figure but depends in principle on temperature, heat flow rate, heating rate and peak area (cf. Sects. 3.2, 4.4). Though these effects are not very pronounced, they should be carefully tested and a thorough verification (with the relevant parameters varied) should be performed if the demands on the accuracy of the measurements are high (H6hne, G16ggler, 1989).
2.2.2
Special Power Compensating DSC
Because of the importance of this type of calorimeters several modifications have been done for essential special purposes:
High Pressure DSC
High pressure measurements are of great importance from the thermodynamic point of view. The change of pressure enables a better insight into the thermo dynamic behavior of materials. High pressure is often used during production and processing of materials and the change of properties with pressure is of great interest (H6hne, 1999b). While high pressure DTA already exists for
a longer time in several laboratories worldwide (see, e.g., Szabo et aI., 1969) at pressures up to 1 GPa, the number of high pressure DSCs is very limited (Schmidt et aI., 1994). To our knowledge there exists only one power compen sated DSC (based on the Perkin Elmer DSC7), which works up to a pressure of 500 MPa. Design, specifications as well as applications have been described by Blankenhorn, H6hne, 1991 and H6hne, Blankenhorn, 1994.
PhotoDSC
Another modification concerns the possibility to follow light induced reactions in a DSC. Slight changes in the design of a DSC make it possible to irradiate the sample with light. A DSC modified in this way is referred to as PhotoDSC. Irradiation of the sample with light, sufficiently rich in energy, leads to reaction in the PhotoDSC. The heat generated during these reactions is recorded. First applications of this type of DSC were described by Wight, Hicks, 1978; Tryson, Shultz, 1979; Flammersheim, 1981. Since 1987, the DPA7 of PerkinElmer is commercially available as a supplementary device of the DSC7. (Heat flux DSCs are also available with suitable additional devices.) The principle is shown in Fig. 2.8. Sample and reference substance are contained in the calorimeter at constant temperature. The sample crucibles are either open or covered by quartz disks. The lid of the microfurnace is provided with openings sealed with quartz disks.
If light high in energy is incident on the sample, it is either absorbed directly or
 as is usually the case  with the aid of a photoinitiator. A reaction usually takes
2.2
Power Compensation DSC
23
Fig. 2.S. Light transfer in a PhotoDSC, schematic representation 
light source 
optics 
LU
W
quartz cuvette with water as IRabsorber
monochromator neutral density fil ters
calorimeter
place only as long as activation energy is supplied by irradiation. Highpressure mercury vapor lamps with a large number of spectral lines are frequently used as a light source; however, hydrogen, deuterium or xenon lamps are also suitable. The undesired infrared portion of the spectrum can be quantitatively absorbed by quartz cells filled with water. The portion of the spectrum which is of inter est is selected using monochromatic filters and the desired intensity is adjusted by means of neutral density filters or metal sieves with different mesh width. It is a particular advantage that the DSC is also suitable for the direct measurement of the incoming radiation flow (light intensity). For this purpose, a graphite disk of known emissivity is substituted for the sample. This disk absorbs the greater portion of the incident light and transfers it to the calorimeter in the form of a heat flow. For most applications, it is of secondary importance whether sample and reference cells are irradiated or the sample alone. For precise measurements electronic stabilization of the light intensity of the lamp is important. Otherwise the total noise of the observed curve will be substantially higher than the noise of the DSC measuring system. Lightactivated reactions under conditions closely related to practice are usu ally very fast. The greatest part of the reaction takes place within a few seconds (cf. Figs. 6.13 and 6.14). There are two consequences:
 There is a considerable change in the measurement signal in periods compa rable with that of the DSC time constant. The measured and the true heat flow rates differ substantially. Desmearing is necessary prior to every evaluation in which the time variable is involved, for example, when calculating the con version versus time curve (see. Sect. 5.4).
 Heat flow rates are so large (see Fig. 6.13) that the sample temperature devi ates considerably from the temperature of the microfurnace, even in power compensated DSCs. If a thermal resistance of 40 K W 1 between sample and temperature sensor is assumed, temperature differences of 5 to 10K would result even for samples with very good thermal conductivity. For organic samples, these differences are even greater. This means that the measure ments are no longer isothermal, not even as an approximation. Kinetic analy sis must allow for this.
24
2 Types of Differential Scanning Calorimeters and Modes of Operation
Another PhotoDSC is described by Saruyama, 1999. In this case the light is used to transfer energy to the sample by absorption. A commercial DSC (Rigaku DSC 8230) has been modified to enable the light to reach the lids of the crucibles, which were covered with carbon black for better absorption. The light intensity was modulated in time by two polarizers, one fixed and one rotating with a certain frequency. This way a periodical changing (sinusoidal) heat flow rate is added to the sample which causes a temperature modulation. The lightmodulated DSC is used for similar investigations as the temperature modulated DSC (see Sects. 3.3, 5.5) but allows a wider range of frequencies up to 1 Hz.
Special DSCfor Fluids
For precise biochemical investigations or for the investigation of reactions in fluids, commercial power compensation DSCs are often used that are operat ed adiabatically or quasiadiabatically (Privalov, Plotnikov, 1989). In a former version of such a calorimeter the sample cell is heated with constant power and the (heated) jacket follows the cell temperature as closely as possible (adia batic or quasiadiabatic conditions). Consequences of this operation mode are that the heating rate is not constant but depends on the behavior of the sample (temperature change due to release of heat of reaction, change in heat capacity) and that cooling requires a jacket temperature that is significantly below the cell temperature to remove heat from the samples, which leads to nonadiabatic conditions. A power compensation calorimeter for liquids (MicroCal LLC, see Plotnikov et al., 1997) uses twin coinshaped measurement cells (nonremovable, volume 0.5 mL) in an adiabatically (for heating) operated jacket. The cells are connect ed to the outside via capillary tubes. Thermoelectric sensors measure the tem perature difference between the cells (power compensation control loop) and between reference cell and jacket (adiabatic control loop ). The time constant is approx.5 s. The temperature range is from 10 to 130°C (45 to 150°C for an ex tended temperature range version) with scan rates up to 1.5 K minI. Heating (adiabatic) and cooling (nonadiabatic) are performed by means of Peltier ele ments. To avoid boiling of the liquids, the system can be pressurized up to 3 bar. The repeatibility of the baseline is stated to be around ± 0.2 p.W, the short time noise (RMS, s. Sect. 7.2) around ± 70 nW. New calorimeters offered by Calorimetry Sciences Corp. (CSC) use a temper atureleading thermostat which is heated or cooled by means of Peltier elements according to the preset scanning rate. The cells are placed nonremovable inside a jacket and are thermally connected to the thermostat via capillary tubes which also serve to fill, to clean etc. the cells. The temperature difference sensors be tween sample cell and reference sample cell are semiconductor (bismuth tel luride) thermopiles. The calorimeter can be operated between 10 °C to 130°C (extended temper ature range version to 160°C) with scan rates (heating or cooling) up to 2 K min I and with a pressure up to 6 bar. Small sample masses of 10 p.g or 50 p.g are in vestigated in capillary type or cylindrical cells.
2.4
Modes of Operation
25
Power compensation heaters serve to minimize the temperature difference between sample and reference sample cells. The time constant is approx. 5 s, the baseline noise is stated to be ~ ± 15 nW and the repeatibility of the baseline  even after refIlling the cells  to be within 0.5 ]lW. These calorimeters have suc cessfully been used in biochemistry to investigate proteines, in particular their phase behavior in solution.
2.3
DSC with Combined Heat Flux and Power Compensation Measuring System
The "pure" power compensation DSC measuring system uses two thermally decoupled microfurnaces with an integrated temperature sensor and heater in each furnace (cf. Sect. 2.2). So called "hybrid systems" show a pair of sensor heater combinations on a disk. The temperature difference between the posi tions of the sample and the reference sample is measured by the temperature sensors and (almost) compensated for by means of controlling the integrated heaters. A good thermal coupling between temperature sensor and related heater is a prerequisite for short time constants and a negligible cross heat flow between the two sensorheater elements. This construction combines advan tages of both the pure heat flux and the pure power compensation measuring systems: stable baseline, short time constants, high resolution, low noise and small temperature differences between environment (furnace) and measuring system. Another version with a combined heat flux and power compensation system uses the disk with integrated thermopile and compensation heaters as bottom plate for two distinct cylindrical chambers which are machined in a small cylin drical metal block. The crucibles for sample and reference sample are put into these chambers and are positioned on the detector disk (temperature range 150 to 550°C, detection limit approx. 0.5 ]lW, crucibles with 30 or 100 ]lL). In principle the addition of heaters plus a power compensation control loop is not restricted to disktype DSCs. Whenever a heater is part of the measuring system  e. g., a calibration heater inside a cylinder type system  this heater may be used for compensation purposes.
2.4
Modes of Operation
DSCs are generally operated by a controlled program which changes the tem perature in time. We distinguish between those modes of operation which leave the heating rate constant (i.e., the classical DSC operation mode) and those with variable (periodical or nonperiodical) change of the heating rate.
26
2 Types of Differential Scanning Calorimeters and Modes of Operation
2.4.1
Constant Heating Rate
In this mode of operation the controlled program follows the time law:
(2.3)
where To is the starting temperature of the run and Po the heating or cooling (neg ative) rate. In other words the temperature changes linearly in time. With common DSCs heating rates up to 150 K min 1 (power compensated DSC: 500 K minI) can be reached. The real cooling rates are lower (< 150 K minI), because the transport of heat out of the sample needs time (a compensation is not possible).
Isothermal Mode
In this mode the heating rate [Po in Eq. (2.3)] is zero, i.e., the temperature To is kept constant. Consequently, if no transitions or reactions take place in the sample, there is no heat exchanged with the sample and the heat flow rate should read zero. This is in practice not the case, as all DSCs are not perfectly symmet ric and the heat exchange of the sample and reference sample with the sur roundings is somewhat different. This results in a nonzero heat flow rate even in the isothermal case. One reason to perform isothermal measurements in a DSC is to measure and check this asymmetry. Isotherms are used for calculation of heat flow rate corrections due to asymmetries, they are included before and after scanning sections in the case of precise heat capacity measurements, for example (see Sect. 6.1). Another reason to perform isothermal measurements is to determine the latent heat of reactions or transitions taking place in the sample at a certain temperature (examples can be found in Chapter 6). Of course, the baseline (the isotherm obtained without processes of the sample) must be sub tracted from the measured curve to get the true reaction heat flow rate.
Scanning Mode
In this mode the temperature changes linearly in time. Every material needs a certain heat to be warmed up and the heat flow rate is proportional to the heat ing rate:
dT
tPcp= Cp'
dt
(2.4)
with Cp as the proportionality factor. In a DSC the differential heat flow rate de pends on the differential heat capacity and heating rate. Generally the measured heat flow rate in scanning mode is never zero and is made up of three parts:
(2.5)
The first term on right hand side is caused by the (unavoidable) asymmetry of the DSC, the second term is caused by the difference in heat capacity of sample
2.4
Modes of Operation
_{2}_{7}
and reference sample and the third term is the heat flow contribution from a re action or transition (latent heat) occurring in the sample. The first two parts de fine the "baseline" and the third part the "peak" of the measured curve. In case of total symmetry and absence of processes, only the sample heat capacity caus es a signal (and can be determined from it). Scanning is the mostly used mode of operation with DSCs.
HighSpeed Scanning Mode
Recently Pijpers et al., 2002, published a method to perform DSC measurements at very high, controlled (including constant) cooling and heating rates of hun dreds of degrees per minute with submilligram amounts of substance. To over come the thermal inertia (thermal lag) of the heat transfer to the sample special measures have to be taken: the mass of the sample has to be very low and the ther mal contact between DSC and sample as good as possible. This is needed to re duce the time constant of the respective RCelement (cf. Sect. 3.3.3) which other wise can be much larger than the time constant of the DSC and the controllers causing a huge smearing (cf. Sect. 5.4) of the measured heat flow rate curve. The method works, however, well and allows, for example, one to investigate melting kinetics and to override recrystallization processes which are unavoid able when testing polymers and often indistinguishably superimposed at lower heating rates.
2.4.2
Variable Heating Rate (Modulated Temperature)
In this case a certain modulation term is added to the linear part of the temper aturetime function. The simplest and most used modulation type is the peri odic (harmonic, i. e., sinusoidal) one and Eq. (2.3) reads:
T(t) = To + Po· t + TA • sin (wt)
(2.6)
with TA the amplitude and w the (angular) frequency of the modulation. To gether with Po, the "underlying" heating rate, there are three parameters which can be chosen freely within certain limits. This influences the heating rate, which follows from Eq. (2.6) as:
dT
 = Po + TA _{•} W· cos (wt)
dt
(2.7)
Depending on the relation of Po to TAW, we have to distinguish four cases (Fig. 2.9).
QuasiIsothermal Mode
In this mode the underlying heating rate is zero and the temperature varies around a constant temperature (see Fig. 2.9 curve a). As the amplitude of the modulation is low (0.01 to 0.5 K normally) the temperature is almost constant
28
2 Types of Differential Scanning Calorimeters and Modes of Operation
Fig.2.9. Temperaturemod _{S} ulated modes of operation:
(a) 
quasiisothermal, 
(b) 
heatingcooling, (c) heat 
ingiso, (d) heatingonly
I
4
3
^{~}
ClJ
'
::J
d
'
ClJ
Cl.
E
ClJ
2
0
^{}^{1} 0
time
_{•}
10
(quasiisothermal). The heating/cooling rate varies between ± TAW. The advan tage of using this mode is the possibility to determine heat capacities even in the isothermal case (see Sect. 6.1.5) what otherwise is not possible.
HeatingCooling Mode
In this mode the amplitude is so large, that the heating rate (Eq. 2.7) changes its sign periodically (see Fig. 2.9 curve b),i.e., the sample is periodically heated and cooled. This is the case if the product TAW is larger than the underlying heating rate Po. The advantage of this mode is a rather large heating rate and therefore good signaltonoise ratio which is in particular useful if we have small sample masses. The disadvantage on the other hand is that the sample is heated and cooled during the run, this may cause problems in some cases, e. g., for polymers where different processes (with different time constants) could occur within the heating and cooling period. This would complicate the evaluation of the processes.
HeatingOnly Mode
In this mode the amplitude is so small that the heating rate (Eq. 2.7) is always positive (see Fig. 2.9 curve d), i.e. the sample is only heated and never cooled dur ing a run. This is the case if the product TAW is smaller than the underlying heat ing rate Po. The advantage of this mode is that all processes only occurring during cool ing of the sample are suppressed and the remaining processes are easier to eval uate. The disadvantage is clear from the rather low amplitude which gives a bad signaltonoise ratio for all quantities derived from it.
HeatingIso mode
This is the limiting case between the two modes above where TAW = Po: the heating rate varies periodically between TAW + Po and zero. This is the mode
2.4
Modes of Operation
29
Fig.2.1O. Special tempera turemodulated modes of operation: (a) stepscan, (b) sawtooth
t
<lJ
L
:::I
0
L
<lJ
CI
E
(lJ
_{5}
4
_{3}
2
time
10
with maximum of amplitude (and signaltonoise ratio) but without cooling the sample (see Fig. 2.9 curve c). Another advantage is that, at heating rate zero, the (even periodical changing) heat flow rate has a minimum with no contribution from the sample heat capacity [which follows from Eq. (2.4) and (2.7)]. This way the lower contour of the fluctuating heat flow rate curve can be used to detect endo or exothermic processes occurring in the sample. After subtracting of _{c}_{P}_{o} there is only _{c}_{P}_{r} re maining in the moment of minimum (i.e. zero) heat flow rate [ef. Eq. (2.5)].
Sawtooth Mode
In this mode of operation different linear scanning rates change periodically (Fig. 2.10 curve b). Again we can distinguish between heatingcooling, heating only, and heatingiso modes of operation, depending on the choice of the re spective rates of temperature change. Every DSC can be operated in this mode, as programming of several different linear scanning ramps in one run normally is possible without special addi tional electronics or software. This is one advantage of this mode of operation. Another advantage follows from the fact that every (nonsinusoidal) periodical function can be expanded as a Fourier series (see textbooks of mathematics). In the case of the sawtooth mode, the periodical function, superimposed on the linear temperature change, reads:
(2n
1
3 . 2n
1
5 . 2n
T(t) _ =c sin2"sin+2"sin
tp
3
tp
5
tp
)
with c as a proportionality factor containing the amplitude and period tp of the periodic modulation. This function contains not only the basic frequency f = lIt _{p} but also (odd) higher harmonics. The same is true for the measured heat flow rate. By Fourier analysis the modulated heat flow rate at different frequen cies can be determined from one measurement.
_{3}_{0}
2 Types of Differential Scanning Calorimeters and Modes of Operation
A disadvantage of this temperature modulation mode is the discontinuous (steplike) change of the heat flow rate, which may cause problems with the fur nace controller (overshoot) and produce additional (nonlinear) heat flow rate fluctuations.
StepScan Mode
The stepscan mode, recently introduced by PerkinElmer Instruments for powercompensated DSCs and by MettlerToledo for heat flux DSCs, is a special mode of operation: After changing the sample temperature, say, by 0.5 K at a high heating rate (steplike, see Fig. 2.10 curve a) the heat flow rate is allowed to equilibrate again. The respective response is evaluated. This can be done many times in a periodical or nonperiodical manner. In the periodical case the step scan is a special sawtooth mode, which is similar to the heatingiso mode in the sinusoidal case, but with an asymmetric triangle shaped form of the tempe raturetime curve. This results in more than only the odd harmonics which one can separate from the modulated part of the heat flow rate signal. Another advantage is that a series of temperature steps is equivalent to a series of heat ing rate pulses. The heat flow rate response on a pulselike event enables one to determine the apparatus (Green's) function of the DSC with and without a sample at every moment of the run. This offers an easy possibility to come to correct ("desmeared", cf. Sect. 5.4.3) reaction heat flow rates as well as to mea sure the complex heat capacity Cp(w) from one run (see Sects. 4.7, 5.5.3).
3 Theoretical Fundamentals of Differential Scanning Calorimeters
In all DSCs, a temperature difference AT  given as a voltage  is the original measurement signal. In almost all instruments a heat flow rate tl>m (differential heat flow rate) is internally assigned to AT (cf. Chapt. 2). Independent of whether the user obtains AT or tl>m from the respective DSC, knowledge of the functional relation between the measured signal (AT, tl>m) and the quantity searched for (the real heat flow rate tl>r consumed/produced by the sample) is important for
 the timerelated assignment of tl>r to AT or tl>m (investigation into the kinetics of a reaction),
 the determination of partial heats of reaction,
 the evaluation and assessment of the influences of operating parameters and properties of the measuring system with regard to this relation,
 the estimate of the overall uncertainty of measurement.
The relation between tl>r and AT or tl>m can be derived in varying degrees of approximation to real DSCs. Analytical solutions of the differential equations are possible only for simple boundary and initial conditions and for quasi steady states. Numerical procedures and solutions can approximate the actual conditions more exactly, however, without the clarity of the functional relations given by an analytical solution. Basic considerations in this field are given by Gray, 1968. To ensure better differentiation from tl>n in the following section, AT  instead of tl>m  is assumed to be the measurement signal, i. e., we search for the relation tl>r(AT). These two quantities, AT and tl>m, are under normal circumstances strictly proportional, with the exception of the opposite sign.
3.1
Heat Flux DSC
Three steps of an analytical description of the functional principle of a heat flux DSC will be presented in the following. The results of a numerical calculation will be discussed afterwards.
32
3 Theoretical Fundamentals of Differential Scanning Calorimeters
Fig.3.1. Heat flux DSC (disktype), model for zeroth approximation (linear model).
S sample, R reference
sample, F furnace, A cross section of the heat conductor
between furnace and Sand R,M distance between tem perature measurement point and furnace
Tr
_{F}
_{/}
_{/}
_{t}_{J}_{!}
l
_{S}
Ts
T.
_{/}
/
T,
_{F}
Zeroth Approximation
First, the heat flux DSC is represented by a very simple linear model (Fig. 3.1). The following simplifications have been made in this approximation:
 steadystate conditions (i.e., constant heat flow rates),
 only one thermal resistance, the apparent resistance between furnace and sample, is taken into account with no interaction (crosstalk) between sample and reference sample,
 only the heat capacities of the sample and reference sample (Cs, C _{R} _{)} are taken into account, other heat capacities are neglected,
 sample temperature and measured temperature are assumed equal,
 no heat exchange with the surroundings (i. e., no heat leak).
Figure 3.2 shows the equivalent electric circuit diagram for the zeroth approxi mation. This diagram serves to better understand the interrelations. From the physical point of view, electric charge transport and heat transport are equiva lent processes, and many people find it easier to read electric circuit diagrams than to visualize heat flows in real equipment.
Fig. 3.2. Equivalent electric circuit for the linear model
of the heat flux DSC
(see Fig. 3.1).
C capacitance, R resistance,
i current, U voltage, Sub scripts: S sample, R refer U, ence, F furnace
'fS
if.
^{R}^{,}^{s}
^{R}^{'}^{R}
[.
fjU
T
[s
T
OT~~
/77J777
3.1
Heat Flux DSC
_{3}_{3}
The BiotFourier equation of the (steadystate) conduction of heat, together with the formulation in absolute values, reads as follows:

A
~
=  A . grad T
or
I~I

A
=  A ·1 grad T 1
The amount of the heat flux tP/A is proportional to the gradient of the tempera ture; the thermal conductivity .\ is the proportionality factor. In the onedimensional model referred to above, this equation is reduced as follows for the lefthand and righthand subsystem shown in Fig.3.l (with Tp>Ts, TR in scanning mode):
In the case of thermal symmetry, tPps = tPPR is valid, with Ts = TR. If a constant (exothermic) heat flow rate (tP _{r} < 0) is produced in the sample, Ts increases by t:.Ts, the temperature difference Tp  Ts and thus the heat flow rate tPps decrease. When the steadystate (i.e., uniform heating rate) is reached again, for reasons of balance, the change of tPps, (t:. tPps), must be equal to tP _{r} _{:}
Nothing has changed on the side of the reference sample, hence:
Consequently
(Here, Ts > TR,hence tP _{r} is negative: exothermic effect; t:.T = t:.TSR: measurement signal)
In this simple steadystate model, K is given completely by the properties of the heat conduction path between the furnace and the samples. This means that
in the steady state, there is direct proportionality between the measurand tP _{r} and
can
be achieved when in scanning operation sample and reference sample have dif ferent, temperatureindependent "heat capacities". A greater amount of heat will always flow into the sample whose heat capacity is higher, in order that the steadystate heating rate is maintained. With Cs > C _{R} the following is then valid
the measurement signal t:. T. The conditions of constant heat consumption
for the difference between the heat flow rates to sample and reference sample:
34
3 Theoretical Fundamentals of Differential Scanning Calorimeters
There is no steady state during sample transitions or reactions; the above approximation does not apply in these cases. Furthermore, Cs and C _{R} (and thus I:! cPSR) change with temperature, but these changes are in many cases rather slow and do not affect the steadystate condition very much. When there is such a quasisteady state, the following is valid in approxima tion:
(3.1)
This relation describes the shape of the baseline and is the basic equation to de
termine the heat capacity Cs. For
the empty reference crucible (C _{R} = 0)
I:!T Cs = K'·
P
is valid. When the heat capacity C _{R} is known (i. e., if C _{R} = C _{r}_{e}_{f} "# 0, using a reference material), we get
I:!T
Cs=CRK' .
P
In practice, first a zero line I:!To will be recorded with both crucibles empty (to check the asymmetry of the apparatus) which is subtracted from the measured curves (for the measurement of heat capacities, see Sect. 6.1).
First Approximation
In the first approximation, nonsteadystate processes in the sample are also permitted which manifest themselves as "peaks" of the measured curve. This means that I:!T is not constant in time. For the rest, the simplified arrangement as for the zeroth approximation is used here. cPps is the heat flow rate from the furnace to the sample, cPr(t) the time dependent heat flow rate produced inside the sample (reaction, transition). The following balance for the heat flow rates is then valid for the sample of heat capacity Cs:
dTs
Cs 
dt
=
cPps 
cPr
(exothermic: cPr negative, endothermic: cPr positive)
Withl:!T= Ts  T _{R} ,
dT _{R}
Cs
dt
results.
dl:!T
+ Cs 
dt
= cPps 
cPr
3.1
Heat Flux DSC
_{3}_{5}
Accordingly, the following holds for the reference sample (cPr =0 by defini tion):
dT _{R} C _{R} = cP _{F}_{R}
dt
When the difference between the two balance equations is calculated, the fol lowing is obtained:
The following is valid for the heat flow rates cP _{F}_{S} and cP _{F}_{R} :
and
where R _{F}_{S} and R _{p}_{R} are global heat resistances between the furnace and the sam ples. In the case of thermal symmetry, R _{F}_{S} =RFR =R, thus:
flT
dTR
cPr =  (C _{s}  C _{R} )
R
dt
dfiT C _{s} 
dt
(3.2)
This equation links the reaction heat flow rate cPr searched with the measured signal fl T. The second term takes the asymmetry of the measuring system into account as regards heat capacities of sample and reference sample. The third term considers the contribution of the thermal inertia of the system when a measured signal flT(t) appears. In analogy to the charging or discharging of a capacitor of capacity C, a time constant T can similarly be defined for the heat flow rates:
T=Cs·R
When fl T is changed, R is the effective thermal resistance to the "charging or dis charging" of the "capacity" Cs. With this resistance and with dTR/dt = /3, the heating rate; as the reference sample is always in a steadystate heating mode, the following results from Eq. (3.2):
flT(t)
cPr (t) = R 
T
 (Cs  CR ) • /3  R . ~
dflT(t)
(3.3)
The measured signal flT is not (I) proportional to the heat flow rate cPr at a given moment but delayed with time and thus also distorted ("smeared"). Further more, when Cs "# C _{R} is valid, the measurement signal is not equal to zero  even if cPr is equal to zero and steadystate conditions prevail  but has the value
36
3 Theoretical Fundamentals of Differential Scanning Calorimeters
t
DoT
"
time

"
Fig.3.3. Measured curve of a heat flux DSC (schematic) with heat released inside the sample (exothermic effect). t1T measured signal, R thermal resistance between furnace and sample, f3 heating rate, t1C = Cs  C _{R} difference between the heat capacities of sample and reference sample (here negative, because C _{R} > Cs), Q) peak area (exothermic effect); it is a measure of the heat released between t, and tz,@ area below the baseline; it is a measure of the heat required to heat the sample be tween t, and tz
R . p. (C _{s}_{} C _{R} ) which is the initial deviation after the quasisteady state has been reached in scanning operation [cf. Eq. (3.I)]. This contribution is the measured curve before/behind a peak which is parallel to the abscissa if R, l1C = Cs  C _{R} and pare constant (Fig. 3.3). In reality, the term R (Cs  C _{R} _{)} reflects the temperature dependence of the thermal resistance R (in general: of the heat transfer conditions) and of the heat
causing a temperature dependence of the measured curve
capacities Cs and C _{R}
even without any thermal effect due to the sample. [NB: Eq. (3.3) is in principle the so called Tian equation, originally derived for cylindertype calorimeters.] Regarding Eq. (3.3), the conclusions are as follows:
1. When the signall1 T measured at a given moment is to be assigned to the heat flow rate IPr by which it is caused, the third term in Eq. (3.3) must be taken into account (cf. desmearing, Sect. S.4). R must be determined by calibration (cf. Sect. 4.4); the time constant T can also be obtained from calibration measurements (cf. below "Higherorder ap
proximations" and Sect. 7.2).
For t> to, the solution of the differential Eq. (3.3) for a heat pulse (/Jr at the moment to has the form (see textbooks of mathematics)
3.1
Heat Flux DSC
37
Fig. 3.4. Measured signal of a heat flux DSC for an exothermic heat pulse at the time to (model of the 1st approximation). r time constant; symbols ^{s}^{e}^{e} ^{F}^{i}^{g}^{.} ^{3}^{.}^{3}
~~r~~~
t
IlT
tTj
t,
time
with T = Cs ' R. Up to the moment t = to, the solution is the steadystate curve
For t ~ 00, the function ~T(t) returns to this curve (cf. Fig. 3.4).
2. For the total heat of reaction or transition Qr developed/consumed in the sample, the following balance equation is valid:
where t _{1} _{,} t2 are the beginning and end, respectively, of the peak. With Eq. (3.3) inserted, the following is obtained:
1 [t2
t2
t2 T d~T
Qr= f~T(t)dtf(R·~C'f3)dt f·dt
R
tl
tl
tl
R
dt
]
(3.4)
The content of the square brackets corresponds to area CD, i. e., the socalled peak area between the measured curve and the (interpolated) baseline (defi nition see Sect. S.1) in Fig. 3.3.
 When ~C and R are not temperature (or time )dependent, the measured
abscissa. In this case, d~T/dt is
zero before and behind the peak. The contribution of the 3rd term vanish es when integration is carried out over the whole peak.
 In the real case, the 2nd term of Eq. (3.4) is not constant, i. e., the curve be fore and behind the peak is not parallel to the abscissa. In this case, the 3rd
term will not vanish but represents a correction of the values obtained by peak integration.
curve outside the peak is a parallel to the
_{3}_{8}
3 Theoretical Fundamentals of Differential Scanning Calorimeters
Fig. 3.5. Partial integration of a peak (heat flux DSC, exothermic effect).
C _{R} > C _{s} , CD partial peak area
(between tl and t*),@ area below baseline (between tl and t*); symbols see Fig. 3.3
t
6.T
t,
time
.
3. For the partial integration (Fig. 3.5) of the peak between t, and t*, the contri bution of the 3rd term must be taken into acount at the point t*:
Qr(t*} = ~ [Y ilT(t} R
tl
\.
dt f (R· ilC·f3) dt]  f ~. dilT dt
tl
y
tl
R
dt
) 'vl
partial peak area
correction term
Partial integrations of the peak are, for instance, necessary for kinetic investiga tions and to determine purity (see Sects. 6.3 and 6.9).
HigherOrder Approximations
The temperatures of sample and reference sample (assumed here to be homo geneous) are not measured directly. There is a certain distance and a thermal
Fig. 3.6. Heat flux DSC,
(one half), model of the 2nd approximation.
S sample, F furnace, _{T}_{M}_{S} tem
perature at sample measure ment point
/
/
3.1
Heat Flux DSC
39
4
s
Fig.3.7. Equivalent electric circuit for the 2nd approximation. S sample side, R reference side, F furnace, Ms measurement point on sample side, MR mea surement point on reference side, ll.U corresponds to ll.T, UFcorresponds to T _{F} _{,} i corresponds to tP _{r} _{•} RFM thermal resistance between furnace and measurement points Ms, M _{R} , R _{M}_{S} thermal resistance between sample measurement point and sample, _{R}_{M}_{R} thermal resistance between reference sample measurement point and reference sample, Cs heat capacity of the sample, C _{R} heat capacity of the reference sample, C _{F}_{M} apparent heat capacity between furnace and mea surement points Ms , MR
resistance between the temperature measurement points and the respective sample (Fig. 3.6). Depending on the design of the measuring system, the resis tance is made up of several parts differing in quantity and originating in the transition layer between sample and bottom of the crucible on the one hand and crucible and support on the other hand as well as further resistances between support and temperature sensor. When the sample temperature changes, the temperature measurement point reacts only after some delay. The analogue electric circuit diagram of a DSC in this "2nd approximation" is shown in Fig. 3.7. The following is valid for this socalled "thermometer problem":
TMS = Ts 
dT Ms
r2 • 
dt
TMS temperature of the measurement point (e. g., junction of the thermocouple) for the sample
sample temperature (assumed homogeneous) characteristic time constant for the temperature relaxation between sample and measurement point.
r2
Ts
In analogy, the following is valid for the reference side:
T _{M}_{R} =T _{R} 
dT MR
r2·
dt
(r2 is assumed to be equal for both sides).
_{4}_{0}
3 Theoretical Fundamentals of Differential Scanning Calorimeters
With these equations, the following results for the difference:
flTsR = Ts T _{R} = T _{M}_{S}  TMR + T2
(
dTMS
dTMR)

dt
dt
fl Tm = TMS  TMR is the measured temperature difference. In other words, with known T2 it is possible to calculate the temperature difference between sample and reference sample from the temperature difference measured at the position of the temperature probe. This was made use of in the socalled advanced Tzero™ technique for the turrettype DSCs recently introduced by TAInstruments (see Sect. 2.1.2). However, the following relation results in analogy to the mathemat ical procedure of the 1st approximation if we use fl TSR instead of fl T:
1
iPr (t) =  Ii
[
dflTm
fl T m + R . (Cs  CR ) • f3 + T 1 dt + TIT 2 ~
d2flTm J
(3.5)
[Tl = T from the first approximation, see Eq. (3.3)1
In addition to the 1st derivative (slope) of the measured curve flTm(t), the 2nd derivative (curvature) must be used in the 2nd approximation to get the heat flow rate converted in the sample. The two time constants must be determined experimentally by proper calibration procedures. The first time constant _{T}_{l} is de termined by the apparent thermal resistance and heat capacity between furnace and temperature sensor. The second time constant is determined by the appar ent thermal resistance and the "effective" heat capacity between sample and sample temperature measurement point. Exact symmetry between the sample and the reference side is assumed in this calculation. The abovedescribed approximation can be refined as desired: Possible asym metries can be taken into account by introducing different thermal resistors and capacities on the sample and reference side into the network. This results in more complicated equations which, however, can be solved. This enables one to cor rect not only the temperature but even the measured heat flow rate for possible influences of the asymmetries (e. g., a curved zeroline, which otherwise must be subtracted afterwards) during the measurement. Certain heat transfer problems and a temperature gradient inside the sample and its influence on the peak shape can be taken into account by introducing further thermal resistances and capacities in the network of Fig. 3.7. A thicker sample with poor conductivity can be considered as having been split up into different layers which are linked with one another by heatconducting boundary layers. In the end an additional differential quotient in the differential equation and another time constant re sult for each additional thermal resistance in connection with a heat capacity. The solution of such a refined approximation is as follows:
1
iPr(t)=
Reff
[
dflT
d ^{2} flT
flT+ko+k _{1} +k 2 _{2} +k _{3}
dt
dt
d ^{3} flT dt
 _{3}
+
J
3.1
Heat Flux DSC
41
In reality, the constants k _{i} are terms into which the thermal resistances and ca pacities (and thus the time constants) of the arrangement enter. Calculation of the reaction heat flow rate r/J _{r} presupposes that all k _{i} and the measurement signal AT and its time derivatives are known. It can be shown (Loblich, 1985) that, for practical application, the 2nd order differential equation is a sufficiently good approximation to calculate the true desmeared reaction heat flow rate; only the time constants f1 and f2 must be known for this purpose. For a detailed treatment how to determine the time constants experimentally, see Loblich, 1994. So far we have only considered the case where it is assumed that no heat ex change takes place between sample and reference sample, which is true for the turrettype DSC as well as for some of the cylindertype DSCs. This simplifica tion is certainly not permissible for the widespread disktype measuring sys tems. Figure 3.8 is a more realistic representation of the measuring system of such a calorimeter. When the differential equation is to be set up for this system, it may be convenient to use the analogue electric circuit represented in Fig. 3.9 for the 2nd approximation (Rohne, 1983). In order to formulate the desired dif ferential equation, according to Kirchhoff's laws, the voltage balance and current balance are made up for each loop and each node of the analogue electric cir cuit. For the circuit of Fig. 3.9, five equations for both voltages and currents are then obtained. On the basis of the laws of electricity, these 10 equations are sim plified and combined to form one differential equation which is retranslated into the language of heat transport. The result for a symmetric twin design (i.e., R _{F}_{M}_{S} = RFMR = R, CFMS = CFMR = C) is as follows:
/
T,/
(3.6)
'~==:='~~'%
/JT
 ¢'R
Fig.3.8. Disktype measuring system of a heat flux DSC (to calculate the tIIr(~ndependence). Ms temperature measurement point on sample side, MR temperature measurement point on reference side, tIIr heat flow rate produced/consumed by the sample
42
3 Theoretical Fundamentals of Differential Scanning Calorimeters
U F
s
Fig.3.9. Equivalent electric circuit for the disktype measuring system according to Fig. 3.S. S sample, R reference sample, M temperature measurement point, F furnace, U voltage, i current, R resistance, C capacitance (cf.legend to Fig. 3.7)
This equation is similar to that of the 2nd approximation [Eq. (3.5)]. However, as
there is a thermal resistance RMM between sample system and reference sample system, a thermal effect in the sample will also affect the reference side and thus T _{R} _{•} Only in the steadystate case and at a sufficiently great distance from peaks is d TRid t equal to the heating rate f3, and the second derivative of the reference sample temperature is equal to zero only there. For the steadystate (st) case,
dilT
d 2 ilT
with tP _{r} = 0,  = 0 and 2 = 0 the following is then valid and characterizes
dt
dt
the baseline:
C _{R} _{}
ilTst =
1
Cs
2
+
R
RMM
. f3
When Eq. (3.6) is integrated to get the area of a transition peak, the following is valid with the approximation dTR/dt::::: f3:
f
12
II
tPr (t)dt = Qr =
(
2) r12 f ilT(t)dt f
12
II
II
1

+ 
R RMM
C R _{}
1
C s
2
+
=
R
(
1
2) 12
  + . f (ilT(t)  ilT.t) dt
R
RMM
II
RMM
. f3dt
]
Mult mai mult decât documente.
Descoperiți tot ce are Scribd de oferit, inclusiv cărți și cărți audio de la editori majori.
Anulați oricând.