16 Chapter 2 Thermocouples emperatu in the svstem is measured by thermocouples Thes oI tw different me sed into ajunction or bead.Thejunction produces a fixed,standard EMF across the junction,varying with tem perature.Strictly a thermocouple system should consist of two such junctions.One is for measuring the sample temperature and this is joined toa second coup held at a constant reference temperature.usually at 0C in melting ice.In modern equipment a referenc e EMF is provided The metals used are carefully chosen to give a very reproducible, accurate EMF at the junction and preferably as high as possible a value of EMF.They are highly refined pure elements or alloys of more than one element.The commonest thermojunction is probably platinum versus platinum allo with 13%rhodi m.This syst n has the adva high melting antage oint and inertne s and purge or product gases Other metals have been used,but these do not stand heating to high temperatures too often and are better restricted to lower temperatures. e.g.only up to 700C. One slight disadvantage of platinum,both for crucibles and ther- mocouples, app ars if th m rature is tak en above 1000C,for stance in glass-making studie where up to 1600Cmay be used.At thes temperatures there is a tendency lor two pieces of platinum to weld together.This might lead to a crucible welding to a thermocouple or the crucible to the hang-down wires or cradle.If this proves to be a problem, then suitable separators must be used.This is often in the form of a thin ceramic plate betweer the two pieces of platin um.A sec proble appears second metal is in contac with the platinum The second metal will have the tendency to dissolve,or alloy,into the platinum causing holes to appear in the platinum in the worst cases.The remedy is again to separate the metals by an intermediate material.A third cause of difficulty is that platinum is a well-known catalyst and may dramatically increase the rate of any reaction taking place in the Temperature Control As well as the thermocouple system for measurement,a second,entirely separate,thermocouple system is provided to sense the furnace tempera- ture and is connected to the furnace control circu ts.The m ther mocouple not use d for the two purposes because the criteria for them are very different.The measuring couple has to be positioned as near to the sample as possible.Sometimes this is just below the sample (see
16 Chapter 2 Thermocouples The temperature in the system is measured by thermocouples. These consist of two different metals fused into a junction or bead. The junction produces a fixed, standard EMF across the junction, varying with temperature. Strictly a thermocouple system should consist of two such junctions. One is for measuring the sample temperature and this is joined to a second couple held at a constant reference temperature, usually at 0°C in melting ice. In modern equipment a reference EMF is provided electronically. The metals used are carefully chosen to give a very reproducible, accurate EMF at the junction and preferably as high as possible a value of EMF. They are highly refined pure elements or alloys of more than one element. The commonest thermojunction is probably platinum versus platinum alloyed with 13% rhodium. This system has the advantages of high melting point and inertness to samples and purge or product gases, Other metals have been used, but these do not stand heating to high temperatures too often and are better restricted to lower temperatures, e.g. only up to 700°C. One slight disadvantage of platinum, both for crucibles and thermocouples, appears if the temperature is taken above lOOO"C, for instance in glass-making studies, where up to 1600°C may be used. At these temperatures there is a tendency for two pieces of platinum to weld together. This might lead to a crucible welding to a thermocouple or the crucible to the hang-down wires or cradle. If this proves to be a problem, then suitable separators must be used. This is often in the form of a thin ceramic plate between the two pieces of platinum. A second problem appears if a second metal is in contact with the platinum. The second metal will have the tendency to dissolve, or alloy, into the platinum, causing holes to appear in the platinum in the worst cases. The remedy is again to separate the metals by an intermediate material. A third cause of difficulty is that platinum is a well-known catalyst and may dramatically increase the rate of any reaction taking place in the crucible. Temperature Control As well as the thermocouple system for measurement, a second, entirely separate, thermocouple system is provided to sense the furnace temperature and is connected to the furnace control circuits. The same thermocouple is not used for the two purposes because the criteria for them are very different. The measuring couple has to be positioned as near to the sample as possible. Sometimes this is just below the sample (see
Thermogravimetry and Derivative Thermogravimetry 17 Figure 3),as near as possible,but not quite touching.In other cases the wires for the two halves of the couple are run down the balance support wires and the junction bead actually touches the crucible.The electrical leads are then arranged in such a way so as not to affect the movement or position of the balance the other hand the furnace thermocouple has to be able to responc rapidly to furnace temperature.If there is a lag in time between furnace power being turned up and temperature rise being detected,then the system will tend to go into temperature swings instead of a steady linear rise.For this reason the furnace measuring couple is positioned as near as of heat,which is the e wire winding.The couple ha 0 electrically insulated from the winding,but is at least embedded in the furnace cement coating on the furnace. Data Collection 0p sseu o 1yo un [ou system when an empty cruci control to dampen swinging of the balance beam.If vibration sets a beam swinging,these swings may continue for a long time before dying away due to the viscosity of the gas in the balance case and will be recorded superimposed on the results.An electronic damping device removes this effect.Sinc e most thermal reactions in the solid state are slow,a high degree of damping may be used.That is,a slow respon of the beam to mass loss is not a problem.However,on some occasions faster reactions, perhaps explosive types,may be studied and in these cases the damping has to be turned down or the balance mechanism will miss fast mass changes.A consequence of this is that the trace will become "noisyand a off bety the conflicting ouput from the controlunitisi Originally data was recorded on chart recorders,but now modern apparatus uses a computer to record the two channels.The analogue signals corresponding to mass and temperature are digitised and the mass and temperature readings are then presented as a graph on the screen and also ste red on a floppy or hard disc.Ther ings may be recalled at a later time to compare with newer res The computer may be pro grammed to automatically convert EMF into temperature using a poly- nomial equation. A typical commercial system is also shown diagrammatically in Figure 4(A),with the furnace shown in more detail in Figure 4(B). In this c ase the ample thermocouple is placed above the sample crucible.There are two purge gas systems.One flow passes over the
Therrnoyravirnetry and Derivative Therrnoyrauirnetry 17 Figure 3), as near as possible, but not quite touching. In other cases the wires for the two halves of the couple are run down the balance support wires and the junction bead actually touches the crucible. The electrical leads are then arranged in such a way so as not to affect the movement or position of the balance. On the other hand the furnace thermocouple has to be able to respond rapidly to furnace temperature. If there is a lag in time between furnace power being turned up and temperature rise being detected, then the system will tend to go into temperature swings instead of a steady linear rise. For this reason the furnace measuring couple is positioned as near as possible to the source of heat, which is the resistance wire winding. The couple has to be electrically insulated from the winding, but is at least embedded in the furnace cement coating on the furnace. Data Collection The control unit often has switches to control mass ranges and to zero the system when an empty crucible is used. There may also be a damping control to dampen swinging of the balance beam. If vibration sets a beam swinging, these swings may continue for a long time before dying away due to the viscosity of the gas in the balance case and will be recorded superimposed on the results. An electronic damping device removes this effect. Since most thermal reactions in the solid state are slow, a high degree of damping may be used. That is, a slow response of the beam to mass loss is not a problem. However, on some occasions faster reactions, perhaps explosive types, may be studied and in these cases the damping has to be turned down or the balance mechanism will miss fast mass changes. A consequence of this is that the trace will become “noisy” and a trade-off between the conflicting requirements has to be made. The output from the control unit is a mV signal representing mass. Originally data was recorded on chart recorders, but now modern apparatus uses a computer to record the two channels. The analogue signals corresponding to mass and temperature are digitised and the mass and temperature readings are then presented as a graph on the screen and also stored on a floppy or hard disc. The readings may be recalled at a later time to compare with newer results. The computer may be programmed to automatically convert EMF into temperature using a polynomial equation. A typical commercial system is also shown diagrammatically in Figure 4(A), with the furnace shown in more detail in Figure 4(B). In this case the sample thermocouple is placed above the sample crucible. There are two purge gas systems. One flow passes over the
Chapter2 LE balance mechanism and then down through the sample area.This keeps hot,maybe corrosive,product gases away from the balance mechanism. A second purge passes across the decomposing sample and the off-gases taken from the left may be led away for analysis.This treatment is mentioned below Isothermal Experiments In many TG experiments,the temperature of the furnace is raised at a constant rate.This type of experiment is referred toas non-isothermal ng or r ising temperatur A alt experimental nique i ,and isof used in kinetic studies.Instead of raising the temperature at a constant rate,the temperature is held constant and the mass loss (or mass gain)observed at this fixed temperature.The results are then presented as mass loss against time,t.In practice the sample has to be placed on the thermobalance and the furnace at first left away from eseWhen thefuace temerature is constantat therd the rna is then run up to the required tempe erature and value,the furnace has to be moved quickly around the sample.There are a number of difficulties with this technique.The sample,crucible,ther- mocouple and cradle have to move rapidly from room temperature to the experimental temperature.They all have a finite thermal capacity,so cannot heat in an ane ously. Th re is a the rmal lag while the sampl temperature rises.The first part of this rise does not matter,because the reaction being studied will not occur rapidly at lower temperatures. However,as the reaction temperature is approached,some reaction will
18 Chapter 2 A TGA Schematic Photodiodes -, Infrared LED ~ Meter movement - Balance arm ~ Tarepan -~ Sample platform Thermocouple ~ Samplepan 1 Furnace assembly - Purge gas outlet J Quartz Liner Off-Gases t B TGA Furnace 4 Balance Purge Sample pan holder Figure 4 A typical commercial thermobalance (With acknowledgement to T A Instruments Ltd. Leatherhead, UK) balance mechanism and then down through the sample area. This keeps hot, maybe corrosive, product gases away from the balance mechanism. A second purge passes across the decomposing sample and the off-gases taken from the left may be led away for analysis. This treatment is mentioned below. Isothermal Experiments In many TG experiments, the temperature of the furnace is raised at a constant rate. This type of experiment is referred to as non-isothermal, scanning or rising temperature. An alternative experimental technique is available, and is often used in kinetic studies. Instead of raising the temperature at a constant rate, the temperature is held constant and the mass loss (or mass gain) observed at this fixed temperature. The results are then presented as mass loss against time, t. In practice the sample has to be placed on the thermobalance and the furnace at first left away from the sample. The furnace is then run up to the required temperature and left to stabilise. When the furnace temperature is constant at the required value, the furnace has to be moved quickly around the sample. There are a number of difficulties with this technique. The sample, crucible, thermocouple and cradle have to move rapidly from room temperature to the experimental temperature. They all have a finite thermal capacity, so cannot heat instantaneously. There is a thermal lag while the sample temperature rises. The first part of this rise does not matter, because the reaction being studied will not occur rapidly at lower temperatures. However, as the reaction temperature is approached, some reaction will
Thermogravimetry and Derivative Thermogravimetry 19 start at temperatures below the chosen value.At higher set temperatures an appreciable amount of reaction may occur at the "wrong"tempera- ture.This leads to doubt about where the zero for time for the main reaction should be set.It is for this reason that small crucibles and thin sed.Also,a lower mass of sat mple should b used in this type of study,and 1 mg samples are common.If the furnace is small,then moving it suddenly round a cold sample system causes a dip in the furnace temperature.and the control system takes a finite time to ethe temperature.Another factor is that the flow of purge gas w be upset as the furnace is moved to close the balance case. furn pen the gas flo ws freely into the atmosphere.When the e syster is closed the gas has to follow a more tortuous path to escape through tubing.This may cause a drop in flow and affect the zero of the balance. Calibration for Mass and Temperature Before a thermobalance is used it should be checked for calibration.The mass reading is relatively easy to check in th same way as for any analytical balance.The balance is first zeroed and then a standard weight, usually in the mg range,is added.If the mass reading is wrong,there are zero adiustments in the control system.The manufacturer's engineer carries out this type of calibration on a routine basis.The measuring therm ocouple curate,but may not qui ebe at the pos of th sample,so there is a slight lag between them.One method of checking temperature readings is to carry out decompositions of known samples.A number of standard materials have been suggested for this purpose.3 The difficulty is that the reaction will not be at a single temperature but spread out over a range of temperatures.A better method is to make use of the Curie point transiti on in n Certa nd all oys are fe netic at room temperature.When these materials are heated,at a tem perature characteristic for each.the material becomes diamagnetic.Ihi change is still not instantaneous but occupies a much shorter range than for a decomposition reaction.If a magnet is placed just above or below the crucible,the sample experiences a magnetic flux in the same direction as gravity At low e this causes a pull on e san and. ss is recorded.A At the Curie temperature there is a sudd apparent mass.Five metals have been quoted as ICTAC Certified Refer ence Materials for this purpose(see Table 1)and all should be used to give a full calibration for all ranges of temperature.These materials are often supplied by,or are available from,a manufacturer. On some instruments the sar mple may be observed visually.It is then possible to place pieces of solid in the crucible and to observe them
Thermoyravimetry and Derivutive Thermogravimetry 19 start at temperatures below the chosen value. At higher set temperatures an appreciable amount of reaction may occur at the “wrong” temperature. This leads to doubt about where the zero for time for the main reaction should be set. It is for this reason that small crucibles and thin wires to hold the crucible are used. Also, a lower mass of sample should be used in this type of study, and 1 mg samples are common. If the furnace is small, then moving it suddenly round a cold sample system causes a dip in the furnace temperature, and the control system takes a finite time to restore the temperature. Another factor is that the flow of purge gas will be upset as the furnace is moved to close the balance case. While the furnace is open the gas flows freely into the atmosphere. When the system is closed the gas has to follow a more tortuous path to escape through tubing. This may cause a drop in flow and affect the zero of the balance. Calibration for Mass and Temperature Before a thermobalance is used it should be checked for calibration. The mass reading is relatively easy to check in the same way as for any analytical balance. The balance is first zeroed and then a standard weight, usually in the mg range, is added. If the mass reading is wrong, there are zero adjustments in the control system. The manufacturer’s engineer carries out this type of calibration on a routine basis. The measuring thermocouple may be accurate, but may not quite be at the position of the sample, so there is a slight lag between them. One method of checking temperature readings is to carry out decompositions of known samples. A number of standard materials have been suggested for this purpo~e.~ The difficulty is that the reaction will not be at a single temperature but spread out over a range of temperatures. A better method is to make use of the Curie point transition in metals. Certain metals and alloys are ferromagnetic at room temperature. When these materials are heated, at a temperature characteristic for each, the material becomes diamagnetic. This change is still not instantaneous but occupies a much shorter range than for a decomposition reaction. If a magnet is placed just above or below the crucible, the sample experiences a magnetic flux in the same direction as gravity. At low temperatures this causes a pull on the sample and a higher mass is recorded. At the Curie temperature there is a sudden loss in apparent mass. Five metals have been quoted as ICTAC Certified Reference Materials for this purpose (see Table 1) and all should be used to give a full calibration for all ranges of temperature. These materials are often supplied by, or are available from, a manufacturer. On some instruments the sample may be observed visually, It is then possible to place pieces of solid in the crucible and to observe them
20 Chapter 2 Table 1 Curie temperatures for ICTAC Certified Reference Materials for Thermogravimetry,GM761 ramsioc Mean Metal temperature/c Standard deviation/C Permanorm3 242-263 253.3 5.3 Nickel 343-36 351. 4 orm 5 43546 451 Trafoperm 728-767 749.5 10.9 melting directly.It is necessary to stop just below the melting point and to increment the temperature in small steps.All of these measurements result in a calib graph,or table,for me asured the rmocouple tem perature versus e ac al temperature from the calibrating sample A blank run should be carried out on a balance before it is used for samples.A run with an empty crucible will check the amount of"noise" (random fluctuation)and base line stability (drift).Whatever gas sur- rounds the sample and crucible it will decrease in density by a large amount as the nperature r .The sample/crucible/cradle system ha nnite volume and displaces this volume of gas,causing a buoyancy effect, which changes with temperature.If the volumes are small and the mass range being used is not too low,the change in buoyancy may not have a measurable effect and may be ignored.To check this,a blank run should be made with the crucible an inert sample,that will show no mass loss du ring the heati It should have the mass,as asample to be used later.Silica sand is a suitable substance but it should be preheated to high temperature first to make sure there is no water to be lost. Effect of Experimental Variables Results Induced by Experimental Conditions The reaction represented by the results in Figure 1 may be seen to occupy a wide span of temperature.This is because a reaction in the solid state is relatively slow compared to gas or solution reactions due to the fact that molecular ement and thesolid state.In some ally contr ol reactions in case there may be a diffusion of one or more of the reacting species through the solid lattice,if temperature is high enough,but this is bound to be slow.The rate of a reaction may also be controlled by diffusion of a product gas or by movement of a reacting
20 Chapter 2 Table 1 Curie temperatures for ICTAC Certijed Reference Materials for Thermogravimetry, G M 761 Temperature of Mean Metal transition1 O C temperature1 O C Standard deviation1 O C Permanorm 3 242-263 Nickel 343-360 Mumetal 363-392 Permanorm 5 435-463 Trafoperm 728-767 253.3 5.3 351.4 4.8 377.4 6.3 451.1 6.7 749.5 10.9 melting directly. It is necessary to stop just below the melting point and to increment the temperature in small steps. All of these measurements result in a calibration graph, or table, for measured thermocouple temperature versus the actual temperature from the calibrating sample, A blank run should be carried out on a balance before it is used for samples. A run with an empty crucible will check the amount of “noise” (random fluctuation) and base line stability (drift). Whatever gas surrounds the sample and crucible it will decrease in density by a large amount as the temperature rises. The sample/crucible/cradle system has a finite volume and displaces this volume of gas, causing a buoyancy effect, which changes with temperature. If the volumes are small and the mass range being used is not too low, the change in buoyancy may not have a measurable effect and may be ignored. To check this, a blank run should be made with the crucible containing an inert sample, i.e. one that will show no mass loss during the heating. It should have the same volume, not mass, as a sample to be used later. Silica sand is a suitable substance but it should be preheated to high temperature first to make sure there is no water to be lost. Effect of Experimental Variables Results Induced by Experimental Conditions The reaction represented by the results in Figure 1 may be seen to occupy a wide span of temperature. This is because a reaction in the solid state is relatively slow compared to gas or solution reactions due to the fact that molecular movement and collision does not normally control reactions in the solid state. In some cases there may be a diffusion of one or more of the reacting species through the solid lattice, if temperature is high enough, but this is bound to be slow. The rate of a reaction may also be controlled by diffusion of a product gas or by movement of a reacting