8/7/2019 6108x1353

1/8

Pure&App/. Chem., Vol. 61,No. 8,pp . 1353-1360,1989.Printed in Great Britain.@ 1989 IUPAC

Thermal analysis of polymersThomas R.MANL EYDeparm zent of Mechanical Eng ineering and Manufacturing Systems, Newcastle Polytechn ic,Ellison Building, Newca stle upon Tyne,NE1 8ST, U.K.ABSTRACT:Before 1963there was little thermalanalysisofpolymersin theU.K. Since thenDTA,DS C and TG have become widely used. The combination of TA techxiques with othertechniques including Mass Spectroscopy andFourier Transform Infra Red spectroscopy (FTIR)provide esp ecially powerful analytical tools.Th e sensitivity of TA techniques has increased. Com puters have had a marked im pact on theoperation of TA instrum ents and also on the manipulation of results.Thermo mecha nical techniques are being increasingly used as plastics an d rubbers find app lica-tions in engineering.The thermal degradation of polymers remains a major concern and an alternative method ofstudying the decomposition of polym ers is described.

I N T R O D U C T I O NHeat was used in the earliest days of science; thermal analysis may be said to have been in use in biblicaltimes (St. Paul: Fire shall try every mans work COR. 1,11 1,13 )and even ea rlier, Iron Age Britonshave left traces of knowledge of thermal properties. A ttempts have been made to date thermal analysisfrom the 18th century through the work of Joseph Black and Josiah Wedgewood but true DTA had towait for Roberts-Austen and TG for Honda at the beginning of the 20th century.Historicalevidenceforplasticmaterials is evenolder. Themother of Mosesusedbitumenandbullrushesto make a craft for her baby. This principle of a com posite of a resin and fibrous reinforcem ent is usedtoday in the Royal Navy minesweepers that patrol the Persian Gulf. The beginning of the plasticsindustry can be dated to 1851 ; at the Crystal Palace Exhibition in London on that date P arkesine acellulosenitrateplastic was shown in severa l moulded articles. The natural polymer 1achasbeen usedin India for thousands of years as a varnish (lacquer) and was moulded at the end of the 19th century.Rubber was used in Central America before the time of Columbus and in 1820 Thomas Hancockdiscovered that m astication (shearing) of coagulated rubber made it plastic, whilst heating the plasticrubber with sulphur regenerated elastic material.As with Therm al analysis however it is only in the 20th century that we can talk about the Polymer Age.The application of TA to polymers had to aw ait the second half of the 20th century. After the SecondWorld War there was some work in the US and the first paper in the UK was given at a Society ofChemical Industry Conference in 1962 [l ].In the quarter century since that paper was published the TA of polymers has changed out of allrecognition. In 1962 DTA was conducted on apparatus made in individual laboratories; nowadays thevast majority of workers use comm ercial equipment. TGA and DTA were the on ly techniques used.Home-madeequipmentdoes however still have its uses where explosiveor corrosivematerials are beinginvestigated as the cheap lab made equipm ent is often adequate for these reactions as a good deal ofenergy is available. Inthe intervening yearsDSC has come to the fore, both in the o riginal form of powercompensated DSC or differential enthalpy analysis, and as heat flux DSC. In the polymer field wherelittle work is done above400C, DTA has been substantially replaced by DSC. The use of high pressureDSC has enabled significant advances to be m ade in studies on crystallisation and cu ring [2].

1353

8/7/2019 6108x1353

2/8

1354 T. R . MANLEY

Improvements in instrumentation over the period in question are striking; amplifiers consisting of bulkyboxes and photo galvanometers are now more likely to be found in a museum than in a TA laboratory.The use of TA to study kinetics has been a fruitful field which was opened up with the improvementsin instrumentation. Initially DTA could only be considered as a qualitative technique but this has nowcompletely changed and TA methods are becoming widely accepted as the basis f or British, Europeanand American Standards. DTA alone was long regarded as an indicative technique an d instruments tocombine DTA and TG soon appeared. Other combinations have been investigated and the difficultproblems of "interfacing" have been overcom e.Interfacing of DTA with a mass spectrometer was one of the first combined m ethods . A great deal ofworkhas been done on this by Stanton Redcroft and some unique information has been obtained on fibresand polymers both as resins and finished products.Evolved gas analysis has provided a great deal of information especially on the decomposition ofpolymers; evolved gas analysis as low pressures, the so-called thermal vaporisation analysis has beenextensively used by N. Grassie and I.C.McNeill at Glasgow.When evolved gas analysis is interfaced with Fourier Transform infra red spectroscopy (FT'IR) a newdimension is added to TA which could be one of the most fruitful com binations ever employed. EGAand TG and EGADTA have been employed to provide further means of studying decompositionreactions.Thermo mechanical measurements have come from being of negligible use to one of the most widelyused techniques in the elucidation of the structures and properties of polymers. Therm o mechanicalmethods may be divided into thermodilatometry, static methods (TM A) and dynamic methods (DTMA :dynamic mechanical thermal analysis) and torsional techniques (TBA: torsional braid analysis) and thetorsional pendulum.The torsional pendu lum is probably the oldest form of DMA and themodified form of torsional braidanalyser TBA has provided new insights into the methods of measuring the physical properties ofpolymers and the fundamental behaviour during the process of curing.The most significant change in TA techniques however has resulted from advances in computing.Computers are now used to operate TA equipment, to improve the sensitivity of instruments bymanipulation of results and to interpret the resultant traces.Other changes that have occurred include the use of energy sources other than heat e.g. in photopolym-erisation;[3] the study of explosions and highly exo thermic reactions in special equipment.In the future lies the use of FHR directly measuring the changes taking place in a solid specimen andincidentally providing a true measure of temperature.

80

60s

20

F I G 1.F I R S T F T l R S P E C T R A O FP O L YM ER 5

0 1 " " " " " " " IJ5 00 44 0 380 320 2 6 0 2 0 0 14 0 8 0W A V E N U M B E R , cm-'

8/7/2019 6108x1353

3/8

Thermal analysis of polymers 1355

C O M B I N E D T E C H N I Q U E SConsiderable synergy is obtained by the com bination of DTA, DSC or TG with other techniques.This is exem plified by considering the com binations of TG and FTIR; TG and Mass Spectroscopy;and DTA and smoke estimation measurements [4].TGIFTIRFTIR utilizes an interferometer constructed on principles initiated by Michelson a century ago. Hehimself could only make visual observations and the value of FTIR only became apparent withimprovements in metrology and computing.The major advantage of an interferometer is the greatly improved signal to noise ratio [5 ] . Thisimprovement can be used to give higher resolution, or to obtain spectra where little energy is available,as in astronomy or in studying the far infra red region or to obtain spectra rapidly.For example it was only by the use of FTIR that the spectra of polym ers in the far infra red could beobtained. The first such spectra were obtained in 1965 and are shown in Fig 1. [6].The standard for TG is calcium oxalate and Compton has shown the infra red peaks obtainedcorresponding to the weight losses in the TG trace. These are readily iden tified as being the expectedemissions of water, carbon monoxide and carbon dioxide. In addition howeverFTIR shows that somecarbon dioxide is produced before the carbon monoxide is detected. This suggests that the decompo-sition of calcium oxalate is not so clear cut as had been thought and also it illustrates the sensitivity ofFTIR.In Fig 2 D W D T is the first derivative of weight loss against temperature (orDTG). This correlates withthe evolved gas profile (EGP) but the peak sizes are different because the infra red absorption of theevolved gas is the major factor controlling the EGP [7].The spectral regions (window s) at 1780 - 1900,2050 - 2200 and 2300 - 2380 cm-' show the evolutionduring the experiment of water, carbon monoxide and carbon dioxide respectively. The em ission ofsome CO, at the time of the CO evolution is clearly demonstrated.

raw0cr0CLI

a

a

u

r4CL00cE:0zI

a

u

000- 42 900 SCAN T I M E

r

A0 3 2300 2380LO 1iq4 00O 0 S C A N TIME 4 3 0

v) 0 6. D W / D T9

B 00

ul

c3

o o o o o SCAN T I M E 4 2 9 2 o b o o 2 5 0 0 00 0 0 SCANF I G . 2 F T I R A N D T G O F C A L C I U M O X A L A T E

P O L Y M E R D E G R A D A T I O NIn the study of the degradation of polymers the com bination of TA and FTIR is particularly powerful.Polytetra fluoro ethylene (PTFE) decomposes in nitrogen in a single reaction to give a material witha spectrum looking like that of tetrafluorethylene. The sub traction of a know n spec trum of TFEvapourfrom the experimentalFTIR gave no residue therefore confirming thatTFEwas the sole product of thethermal degradation of PTFE.If PTFE is heated in air thermo oxidative degradation takes place and inaddition to TFE absorptionbands corresponding to CO,,CO, acid fluorides and other oxidationproductsare observed.

8/7/2019 6108x1353

4/8

1356 T. R . MANLEY

FillersThe decomposition of polymers is often complicated by the presence of fillers. DTA can be used toidentify fillers [81.FTIR however can give a great deal more information as was shown [7] in TGA/FTIR studies on woodwhich is an important filler for thermoset plastics. The TG trace showed a small loss at 70OC and a largerloss around 43 0 "C. Th e FTIR shows howev erthat the simple TG curves covers a complex of reactionsas the spectra change over the temperature range. Bands due to carbon monoxide, carbon dioxide, water,carbonyl groups and aliphatic and aromatic hydrocarbons are found.

TA A N D M A S S SPECTROSCOPYIt has long been realized that the analysis of the gases evolved from TG results would greatly facilitatean understanding of the reaction involved.Fig 3 shows a com bined DTA -TG-MS study of a cured phenolic resin. Phenolic (PF) resins are madeby the reaction of phenol and formaldehyde to g ive compounds that on heating produce a hard, "cured",material. At one time it was thought that PFresins were extensively crosslinked th rougha series of cycliccompounds but it is now known that much of the rigidity of the cured mqulding comes fromentanglement of the molecular chains [9]. On the other hand the crosslinking process g ives to PF (andother condensation thermosetting resins) the characteristicmechanical strength that made it pre-eminentamongst insulating engineering m aterials for many years. PF resins also have excellent fire resistanceand they also find application in space vehicles since the thermal degradation of P F resins produces aninsulating char which acts as an excellent heat shield.The TG trace in Fig 3 show little below 400 "C. The lack of sensitivity of TG relative to other TAtechniques was pointed ou t many years ago [101 and can b e seen here by comparison with the DTA trace.Comparison of the TG trace with the trace produced for the evolution of Phenol by the m ass spectrometershows that the TG curve is mainly influenced by the loss of phenol which is the largest molecule (m/z 94) involved.On the other hand the loss of water (m/z 18) and amm onia (m/z 17)which do not appear on the TG traceare much m ore significant because these compounds, and the formaldelyde (mlz 29), are formed by thebreaking of crosslinks which have a disproportionate effect on the mechanical properties of PFm ouldedcompounds. Without cross links PF resins are merely very viscous solids, after cross linking theybecome very strong engineering materials (a tensile strength of 10,000 psi [68.95 MN/m2] is obtainablewith appropriate fillers).The MS traces show that if the PF resin is held at 200 "C then a catastrophic deterioration in mechanicalproperties can be expected. The DTA trace also signals a change around 200 "C,DTA is thus a usefulindicator of temperatures at which engineering properties may change bu t MS show s clearly why thesechanges occur.

W E lG H T /% MOULDED PF

2 0 0 4b0 (00T E M P (''0

8/7/2019 6108x1353

5/8

8/7/2019 6108x1353

6/8

1358 T. R. MANLEY

As temperatures rise more energy is available and further intra-molecular movements become possible.The temperatures at which these transitions occur are marked by a decrease in modulus and a peak intan 6.In addition to Tg ,DMA shows p, y and 6 transitions in PS. The p transition is related to movem ent ofthe large phenyl rings in the side chain. In the ideal structure for PS it is difficult to iden tify convincingmovements for y and 6. No other technique (apart from DETA) is sufficiently sensitive to detect suchtransitions. In addition to fundamental studies on transitions DMTA is widely used to quantify theviscoelastic behaviour of polymers which must be taken into account in the production and use ofengineering materials.Because polymers are viscoelastic the properties are dependent on time and tem perature. The elasticmodulus of a polymer under constant load will decrease with time because the m olecular structurerearranges in order to m inimize local stresses. In the same way a polymer deformed at a high frequencywill exhibit a relatively higher modulus than one deformed at a low frequency.It is possible to prognosticate the properties of viscoelastic materials to meet the requirements ofengineering designers because of the so called super position principle. This states that measurementsmade at a higher temperature (ie and over a sho rter time) may be transposed to a low er temperature anda longer time. it is therefore possible to obtain results applicable to a service life of say 10years in anexperiment lasting a few months. As w ith all extrapolations however caution is needed when they areused.DMTA is a conven ient method for the study of stress relaxation and thus the expected service life.Oxidation is a major factor in the aging of elastomers and the log of m odulus against temperature isshown in Fig 4A and B for nitrile rubber.Above 30 "Cthe rubber became too soft to test. The requirements of the analytical system of manyDMTA devices and the mechanical behaviour of the specim ens restrict the range of modu lus values thatcan be tested.From Fig 4 it can be seen from the similarity of the curves for samples agedup to 86 hours that the stressrelaxation of the rubbers is dominated by physical factors. After aging for 75 hours the rate of fall inmodulus is beginning to decrease;after86 hours of aging the increase inTgmeans that a significant valuefor the modulus remains at the end of the experiment. It is clear that the antioxidant is hindering thechemical reaction and so the differences amongst

9 9 1the samples remains small.

9 91

7 . 1L-6 0 -50 -4 0 -30 -2 0 -10 0 10 20 30 -60 -50 -40 -30 -2 0 -10 0 10 20 30 -

T E M P E R A T U R E OC T E M P E R A T U R E ' CF I G . 4 D M T A O F NITRILE RUBBER

8/7/2019 6108x1353

7/8

Thermal analysis ofpolymers 1359

After 86 hours the effects of chemical relaxation become more obvious, both on the stress relaxationcurve and in the change in modulus. At 1 20hours a slight stiffening has becom e apparent w hile at 235hours the change in physical properties is obvious. Above this exposure time, the hardening proceedsrapidly, until at 648hours the material is very heavily degraded, although it has not yet failed the stressrelaxation test.This would indicate that whilst 25% retention of initial stress may be a convenient criterion for failurein laboratory testing a high value may be preferred when prognosticating the service life of an elastomerWI.

T H E R M A L D E G R A D A T I O N OF P V CPolyvinyl Chloride (PVC) [poly-l-chloroethylene] is one of the most important plastics now available,being second only to polythene. Interestingly it is one of the least thermally stable commercial polymerssince it begins to degrade by means of an unzipping reaction above 70 "C. Whether such a material ifnewly discovered today would ever emerge from the research laboratory is doubtful to say the least.Thermal analysis techniques were shown [131 to have no significant advantages over standard methodsfor the degradation of PVC so a more fundamental change was made.The onset of degradation is of major interest. To study this onset requires high sensitivity and closecontrol over reaction conditions. It was decided to replace the standard DTA and TG furnaces with afluidized bed [141. This has the following advantages:i powdered samples avoid the effects of the solvents or special forms (films, discs, or rods) that

vary the rates of diffusion.the particles do not sinter above the Tgand thus variations in rates of diffusion are obviated [141.precise control of the temperature in the reaction zone is easily achieved and concentration andtemperature gradients are eliminated.the reactor is made of glass to avoid reactions withme tals and to allow the observation of colourchanges in the sample.the hydrogen chloride released is quickly absorbed in ultra-pure water.

iiiiiivvInitially the effects of diffusion in the atmosphere surrounding the particles were examined. Once thesehad been established experimental conditions were chosen to eliminate the effects of this externaldiffusion.Intraparticle diffusion was then studied with reference to the size and porosity of the specimens. Finallythe effect of varying the temperature between 80 and 200 "C on the course of the thermal degradationwas investigated. This last effect was also studied in a thermobalance.

J II

F I G . 5 F L U I D I Z E D BED R E A C T O R

8/7/2019 6108x1353

8/8

1360 T. R . MANLEY

To make absolutely sure that sintering did not occur above the glass transition temperature (Tg)of thePVC an inert pow der was added to separate the PVC particles.Pure nitrogen (99.9992 %), with less than 1ppm , of oxygen was used as the fluidizing gas.The graphs of conductivity versus time show a sm all but abrupt departure from the asymptotic valueof the degassed ultra-pure water when the temperature is raised over 80 C, and a m ilder but progressiveslope ending in a almost straight line when the first hydrogen chloridereleased is detected. This startingzone depends on the final temperature reached, but always ends before reaching a conductivity of 0.2S. Taking this last value as the first clear indicator of degradation the sensitivity of the method for anormal run, involving 0.600 g of PVC and0 .400 1of absorbing water, would be 13ppm. This sensitivitylies among the highest reported and, if necessary, could be easily increased using b igger samples orreducing the absorbing water. Hence good correspondence between the amount of hydrogen chloridedetected at this early stage and that given off by the sample can be assumed. The conductimetricdetection occurs before any appreciab le change in colour can be no ticed in the sample.The Fluidized technique enables one to study the PVC thermal degradation kinetics at low temperatures(80C), and w ith a 30fold increase in sensitivity. Accurate detection of the hydrogen chloride evolvedis possible long before the sam ple changes its colour. Much of the disagreem ent amongst workers whohave used pow dered sam ples in crucibles or in fixed beds can be attributed to neg lecting diffusion,especially of intra particular diffusion. The Arrhenius plot yields an apparent activa tion energy of 29KcaVmol.The use of a fluidized bed in conjunction with MS or FTIR should greatly increase the reliability ofkinetic results in TA.

C O N C L U S I O N SThermal analysis of Polym ers has advanced rapidly in the last 25 years. In conjunction with othertechniques it will be of increasing value to polymer scientists and will be an essential part of anylaboratory for the study of plastics and indeed other materials.

REFERENCESManley. T.R. DTA and its application to Polymer Science in Techniques of Polymer ScienceMonograph, 17, SOC.Chem. Ind., London (1963).Ma nley, T.R., Thermal Stability of hexamethylol melam ine, Polymer J . Japan 4, 111 (1972).Manley, T.R. and Scurr, G. DT A of U.V. cured resins, J.O.C.C.A. 66,43-4 8 (1983)B.R. Glennie and T.R. Manley, Measurement of smoke from smouldering polymers, SOC.Plastics Engineers 34th ANTEC Atlantic C i o , p. 396 (1976).Fellgett, P., J . Phys. Radn. 19 187 (1958)Manley T.R. and Williams D.A, I.U.P.A.C. M acrom olecular Conference Pragu e, A529 (1965).D.A.C. Compton, On line F.T.I.R. analysis of Gaseous Effluent from a ThermogravimetricAnalyser, in press.Manley, T.R., Characterization of filled amino resins by thermal analysis, Trans. Plastics Inst.p . 525 (1967).N.J. Megson, Phenolic Resin Chemistry, Butterworth London (1958).T.R. Manley in H.V applications of Epoxy Resins, Ed. M. Billings, U.M.I.S.T. pres s (1967).R.E. W etton, Dynam ic Mechanical Measurements and their application to Polymer Systems,Analyt. Proceedings, R.S.C. London O ctober 1981 p. 417.Manley, T.R. and C lark L.J., Stress Relaxa tion Studies, Proc. Int. Rubber Con. Goteborg 1,p.103. (1986).Guy. A., M.Sc. Thesis N ewcastle Polytechnic (1980).J.A. Gra cia Ma rtinez, F. Gutierrez Fernandez, T.R. Manley , Initial thermal stab ility of PVC ,Brit. Polymer J., 18, 201 (1986).