Introduction

P. J. Haines

Oakland Analytical Services, Farnham, UK

MATERIALS, HEAT AND CHANGES

Whenever a sample of material is to be studied, one of the easiest tests to perform is to heat it. The observation of the behaviour of the sample and the quantitative measurement of the changes on heating can yield a great deal of useful information on the nature of the material.

In the simplest case, the temperature of the sample may increase, without any change of form or chemical reaction taking place. In short, it gets hotter. For many other materials, the behaviour is more complex. When ice is heated, it melts at 0°C and then boils at 100°C. When sugar is heated, it melts, and then forms brown caramel. Heating coal produces inflammable gases, tars and coke. The list is endless, since every material behaves in a characteristic way when heated.

Thermal methods of analysis have developed out of the scientific study of the changes in the properties of a sample which occur on heating. Calorimetric methods measure heat changes.

Some sample properties may be obvious to the analyst, such as colour, shape and dimensions or may be measured easily, such as mass, density and mechanical strength. There are also properties which depend on the bonding, molecular structure and nature of the material. These include the thermodynamic properties such as heat capacity, enthalpy and entropy and also the structural and molecular properties which determine the X-ray diffraction and spectrometric behaviour.

Transformations which change the materials in a system will alter one or more of these properties. The change may be physical such as melting, crystalline transition or vaporisation or it may be chemical involving a reaction which alters the chemical structure of the material. Even biological processes such as metabolism, interaction or decomposition may be included.

Sometimes a change brought about by heating may be reversed by cooling a sample afterwards. A pure organic substance melts sharply, for example benzoic acid melts at 122°C and it recrystallises sharply when cooled below this temperature. Ammonium chloride dissociates into ammonia and hydrogen chloride gases when heated, but these recombine on cooling. At high temperature, calcium carbonate splits up to yield calcium oxide and carbon dioxide gas, and these too will recombine on cooling if the carbon dioxide is not removed. The system reaches an equilibrium state at a particular temperature.

[FORMULA NOT REPRODUCIBLE IN ASCII]

To raise the temperature of any system heat energy must be supplied and when sufficient energy is available the system will change into a more stable state. The mechanical properties of a material change as it is heated. Often it expands and becomes more pliable well below the melting point. These are fundamental, important changes on a molecular level, and their study enables the analyst to draw valuable conclusions about the sample, its previous history, its preparation, chemical nature and the likely behaviour during its proposed use.

The temperature at which a particular event occurs, or the temperature range over which a reaction happens, are often characteristic of the nature and history of a sample, and sometimes of the methods used to study it. Sharp transitions, such as the melting of pure materials, may be used to calibrate equipment and as the "fixed points" of thermometry and of the International Practical Temperature Scale (IPTS).

For example, how does the simple, pure inorganic compound potassium nitrate, KNO3, behave when heated? At room temperature, say 20°C, this is a white, crystalline solid. To raise its temperature to 30°C at constant presssure, we must supply an amount of heat depending on the specific heat capacity, Cp approximately 1 J K-1 g-1 at this temperature, the mass m of the sample and the change in temperature. So, for 1 g heated 10°C, we must supply 10 J. To complicate matters, the heat capacity changes with temperature as well. When the temperature reaches 128 °C, the crystals change their structure, and this needs more energy, about 53 J g-1. Then the new crystals are heated, when Cp ≈ 1.2 J K-1 g-1, until the melting point of 338 °C, when more heat must be supplied to melt the sample. Raising the temperature above the melting point eventually causes the sample to decompose to form potassium nitrite, KNO2, so that the mass of the sample is decreased by around 16% and oxygen gas is given off.

This example illustrates the importance of thermal techniques and measurements. Calorimetry measures the amounts of heat, while appropriate thermal methods give the temperatures of phase changes, the temperatures of decomposition and the products of the reaction. Other methods will show the expansion, mass and colour changes on heating.

The analysis of thermal events may be approached in two ways, which overlap considerably. Either the experiment may be designed to measure thermal properties (heat capacity, enthalpy, entropy and free energy) with high precision and accuracy at particular temperatures and conditions, or we may study properties, including thermal properties, over a wider range of temperatures using a controlled heating procedure.

Which experiment is chosen depends on the sample to be analysed. There would be little point in obtaining highly accurate heat capacities on a polymeric or cement sample of complex composition, but its behaviour on heating would be informative. Theoretical work on organic structure and kinetics might require precise knowledge of equilibrium thermal properties which could not easily be obtained using variable temperature methods. Therefore, the techniques are complementary.

Since the worldwide adoption of the SI system of units it is perhaps useful to stress the symbols and units to be used for the physical quantities involved in these methods. The major quantities are given in Appendix 1A and the others may be found in the references.

DEFINITIONS OF THERMAL AND CALORIMETRIC METHODS

Formal definitions are not essential, but those accepted by the scientific community may be found in the literature.

Calorimetry is the measurement of the heat changes which occur during a process. The calorimetric experiment is conducted under particular, controlled conditions, for example, either at constant volume in a bomb calorimeter or at constant temperature in an isothermal calorimeter.

Calorimetry encompasses a very large variety of techniques, including titration, flow, reaction and sorption, and is used to study reactions of all sorts of materials from pyrotechnics to Pharmaceuticals.

Calorimetric methods may be classified either by the principle of measurement (e.g. heat compensating or heat accumulating), or by the method of operation (static, flow or scanning) or by the construction principle (single or twin cell). These will be discussed further in Chapter 5. Thermal analysis is a group of techniques in which one (or more) property of a sample is studied while the sample is subjected to a controlled temperature programme. The programme may take many forms:

(a) The sample may be subjected to a constant heating (or cooling) rate (dT/dt = β), for example 10 K min-1.

(b) The sample may be held isothermally (β = 0).

(c) A "modulated temperature programme" may be used where a sinusoidal or other alteration is superimposed onto the underlying heating rate.

(d) To simulate special industrial or other processes, a stepwise or complex programme may be used. For example, the sample might be equilibrated at 25°C for 10 min, heated at 10 K min-1 up to 200°C, held there for 30 min and then cooled at 5 K min-1 to 50°C.

(e) The heating may be controlled by the response of the sample itself.

THE FAMILY OF THERMAL METHODS

Every thermal method studies and measures a property as a function of temperature. The properties studied may include almost every physical or chemical property of the sample, or its products. The more frequently used thermal analysis techniques are shown in Table 1 together with the names most usually employed for them.

INSTRUMENTATION FOR THERMAL ANALYSIS AND CALORIMETRY

The modern instrumentation used for any experiment in thermal analysis or calorimetry is usually made up of four major parts:

• The sample and a container or holder;

• sensors to detect and measure a particular property of the sample and to measure temperature;

• an enclosure within which the experimental parameters (e.g. temperature, pressure, gas atmosphere) may be controlled;

• a computer to control the experimental parameters, such as the temperature programme, to collect the data from the sensors and to process the data to produce meaningful results and records.

This is shown schematically in Figure 1, and specific applications and instrumentation will be considered in the following chapters.

THE REASONS FOR USING THERMAL AND CALORIMETRIC METHODS

Novice analysts may enquire why yet another technique is needed when gas chromatography, molecular and atomic spectrometry and electrochemical analysis plus many other powerful analytical tools are available. The answer might best be given by considering two practical examples.

First, how can you analyse a mixture of processed minerals such as a cement? Although X-ray diffraction might tell you the different minerals present and atomic absorption spectrometry could measure the elements quantitatively, this does not help to analyse how the cement would behave in practice. For this we need to compare the behaviour under conditions of mechanical and thermal stress and the thermoanalytical techniques of TG, DTA and TMA are important tools for doing this.

Second, the preparation of new chemicals for new pharmaceutical products, synthetic materials and foods could add to the hazards which workers and customers face. Thermal instability and explosive behaviour can be extremely destructive and costly events. Reaction calorimetry and similar techniques can help to predict the likely behaviour of chemicals when reactions, transport and storage are concerned. Physiological behaviour may vary with the nature and form of a drug, and the nature and interconversion of these forms is often studied by thermal and calorimetric methods.

Many analytical techniques require samples in a particular form. For example, gas–liquid chromatography and mass spectrometry need volatile samples and UV–VIS spectrometry usually uses solutions. Therefore, in analysing we destroy the structure of the matrix containing the sample. This has two disadvantages: (i) the behaviour of the sample in its original matrix may be different and (ii) it is time-consuming to alter the form. It is possible to use thermal methods to study the sample "as received". This avoids laborious preparation, does not change the thermal and molecular history of the sample and gives information to the analyst about the real sample and how it would behave in the situation or process where it is actually used.

THE NEED FOR PROPER PRACTICE

Some analytical techniques are sample specific. The "group frequency" bands in an infrared spectrum are largely independent of the method used to obtain the spectrum, whether it is run as a solid KBr disc, a Nujol mull or a solution and whether it is obtained by a dispersive or a Fourier transform instrument. Similarly the titration of an acid with a base should give the same result whether the end-point is detected by an indicator or electrochemically.

This is not always so in the case of thermal methods. The results obtained depend upon the conditions used to prepare the sample, the instrumental parameters selected for the run and the chemical reactions involved. That is not to say that results are not reproducible provided similar conditions are selected. For example, it is possible to compare samples of a polymer to see if their behaviour is "good" or "bad" according to their potential use, but the experimental parameters used for running each sample must be the same.

The useful acronym "SCRAM" (sample–crucible-rate of heating–atmosphere–mass) will enable the analyst to obtain good, reproducible results for most thermal methods provided that the following details are recorded for each run:

The sample: A proper chemical description must be given together with the source and pre-treatments. The history of the sample, impurities and dilution with inert material can all affect results.

The crucible: The material and shape of the crucible or sample holder is important. Deep crucibles may restrict gas flow more than flat, wide ones, and platinum crucibles catalyse some reactions more than alumina ones. The type of holder or clamping used for thermomechanical methods is equally important. The make and type of instrument used should also be recorded.

The rate of heating: This has most important effects. A very slow heating rate will allow the reactions to come closer to equilibrium and there will be less thermal lag in the apparatus. Conversely, high heating rates will give a faster experiment, deviate more from equilibrium and cause greater thermal lag. The parameters of special heating programmes, such as modulated temperature or sample control, must be noted.

The atmosphere: Both the transfer of heat, the supply and removal of gaseous reactants and the nature of the reactions which occur, or are prevented, depend on the chemical nature of the atmosphere and its flow. Oxidations will occur well in oxygen, less so in air and not at all in argon. Product removal by a fairly rapid gas flow may prevent reverse reactions occurring.

The mass of the sample: A large mass of sample will require more energy, and heat transfer will be determined by sample mass and dimensions. These include the volume, packing, and particle size of the sample. Fine powders react rapidly, lumps more slowly. Large samples may allow the detection of small effects. Comparison of runs should preferably be made using similar sample masses, sizes and shapes.

Specific techniques require the recording of other parameters, for example the load on the sample in thermomechanical analysis. Calorimetric methods, too, require attention to the exact details of each experiment. In the following chapters the principles and practice of thermal analysis and of calorimetry will be described and illustrated with some of the many examples of its use in industry, academic research and testing.

Thermogravimetry and Derivative Thermogravimetry

G. R. Heal

Formerly of Department of Chemistry and Applied Chemistry, University of Salford, UK

INTRODUCTION AND DEFINITIONS

Thermogravimetry (TG) is an experimental technique used in a complete evaluation and interpretation of results when it is known as Thermog-ravimetric Analysis (TGA). The technique has been defined by ICTAC (the International Confederation for Thermal Analysis and Calorimetry) as a technique in which the mass change of a substance is measured as a function of temperature whilst the substance is subjected to a controlled temperature programme. The temperature programme must be taken to include holding the sample at a constant temperature other than ambient, when the mass change is measured against time. Mass loss is only seen if a process occurs where a volatile component is lost. There are, of course, reactions that may take place with no mass loss. These may be detected by the allied techniques of Differential Thermal Analysis (DTA) and Differential Scanning Calorimetry (DSC) which are described in Chapter 3. Results are presented as a plot of mass, m, against temperature, T or time, t. The mass loss then appears as a step. This is shown in Figure 1(A).

The temperature range shown in this plot has been restricted to 400 to 600 °C to show the detail of the step. In a normal experiment the temperature might be run from room temperature to 1000°C or higher. It should be noted that the shape is sigmoid in nature, that is, although most mass loss occurs around one temperature, where the line is steepest, some reaction starts well before the main reaction temperature. Similarly there is still some residual mass loss well after the main reaction.

An alternative presentation of results is to take the derivative of the original experimental curve to give dm/dt, or rate of mass loss against time, and to plot that against temperature, T or time, t. Alternatively the derivative may be against temperature T giving dm/dT. The production of such curves is called Derivative Thermogravimetry (DTG). Such a curve is shown in Figure 1(B); the spread of the reaction over a wide temperature range appears here as a relatively broad peak. The DTG curve is of assistance if there are overlapping reactions. Double peaks or a shoulder on a main peak appear in these cases. Slow reactions, with other fast reactions superimposed, then appear as gradient changes in the DTG curve. The area under the DTG peak is proportional to the mass loss, so relative mass losses may be compared. Measurements of just relative peak heights may suffice for some purposes. The position of the peak may not be indicative of any characteristic point in the mechanism of the reaction, only where mass loss is fastest. However, it may be used, if all that is required is to use the peak as a "finger print" of the presence of a substance in a mixture, e.g. a particular mineral in a rock or soil sample.

INSTRUMENTATION

Balance

In the essential form of the apparatus, the substance is placed in a small inert crucible, which is attached to a microbalance and has a furnace positioned around the sample. The furnace may be positioned in several places relative to the balance. This is shown in Figure 2.