482 Analyst, May, 1978, Vol. 103, p p . 482-491 Pyrolysis = Mass Spectrometry of Textile Fibres J. C. Hughes, B. B. Wheals and M. J. Mrhitehouse Metropolitan Police Forensic Science Laboratory, 109 Lirmbeth Road, London, SE 1 7LP A procedure for pyrolysis - mass spectrometry is described and the spectra (mass pyrograms) of various textile fibres are presented. The method is compared with infrared spectroscopy for the forensic characterisation of synthetic fibres. Samples of less than 5 pg can be analysed. Keywords : Textile fibre characterisation ; j~yrolysis - mass spectrometry ; in frayed spectroscopy The forensic analysis of synthetic fibres usually involves the characterisation and comparison of two or more samples and in this laboratory techniques such as microscopy, infrared spectroscopy and thin-layer chromatography combined with ultraviolet spectroscopy of extracted dyes are the methods in current use.In practice the amount of material available may be as little as a single fibre and in such circumstances the infrared method in particular is unable to provide the same amount of information as might be obtainable from a larger sample. Pyrolysis - mass spectrometry (Py - M:S)lp2 is an analytical technique capable of providing useful data on very small polymeric samples and this paper reports on its use for fibre characterisation and compares the method with infrared spectroscopy. Experimental Fibre Pyrolysis Single fibres 0.5-10mm in length were heated in a helium stream using a Curie-point (Pye) or filament pyrolyser (Chemical Data !Systems Pyroprobe, Model 190).On the Pyroprobe samples were generally pyrolysed isothermally for 20 s at a selected temperature (normally 600 or 800 "C), but some of the fibres were sequentially pyrolysed at increasing temperatures in the range 400-900 "C. For the Curie-point method a 15-s pyrolysis time was used. Fibres were mounted for analysis by crimping into a flattened wire for the Curie-point method or by insertion into a quartz tube mounted in the filament coil of the Pyroprobe. Either pyrolyser was connected to an empty glass column, 45 cm x 6.3 mm 0.d. x 2 mm id., heated at 200 "C in the oven of a Varian 2700 gas chromatograph. The pyrolyser was mounted outside the oven and was flushed with helium at 15mlmin-l in order to sweep the pyrolysate into the empty column.The e:mpty column provided a simple means of broadening the pyrolysate band before it entered the mass spectrometer.3 Mass Spectrometry A VG Micromass 12F mass spectrometer was used under standard electron-impact conditions: electron energy 70 eV, emission current 100 PA, accelerating voltage 4 kV and source temperature 240 "C. The pyrolysate emerging from the empty glass column passed into the mass spectrometer via a length of glass-lined stainless-steel tubing and a glass jet separator. A mass range of 25-250 a.m.u. was scanned at I s per decade with a magnet re-set time of 1 s. After pyrolysis 35 scans were collected, with data acquisition and storage achieved with a VG 2040 data system. The 25 most intense spectra of each series were integrated using standard software and a specially written Fortran IV program to produce a composite mass spectrum (i.e., a mass pyrogram).The ions of m / e 28, 32, 40 and 44 were not included in the integration procedure as they are associated with air in the system and only those ions with intensities greater than 0.1% of that of the base peak were included in the processed data. An FIT program was written in Fortran IV to allow the data-processing equipment to compare two sets of data. The FIT equation was defined asHUGHES, WHEALS AND WHITEHOUSE 483 m = 200 1 FIT = 1000 I- I m = 25 m = 200 I m = 25 -I where a and b are the intensities of the same ion of mass m in the spectra of samples A and B. The individual ion intensities were expressed as a fraction of the total ion current ( i e ., they were normalised) in order to compensate for variations in sample size and efficiency of pyrolysis. The FIT factor ranges between 1 000 (for perfectly matched mass pyrograms) and 0 (for completely dissimilar pyrograms). A comparison of the pyrolysers was made by the replicate analysis of ten samples of the same nylon 6 fibre and FIT factors were deter- mined for the resulting 45 pairs of pyrograms for each series. Infrared Spectroscopy Single fibres 1 cm long (mass 3-5 pg) were sealed into glass capillary tubes 2.5 cm x 1 mm i.d., together with 2-3 p1 of solvent, and were heated in an oven at 80-100 "C to effect solution. The solvent selected was dependent upon the type of fibre, e.g., dimethylformamide (acrylics) , m-cresol (nylons and polyesters), acetone (cellulose acetate) and chloroform (cellulose triacetates).The contents of the tubes were poured on to a silicone-coated Petri dish to form a film, which was washed with ethanol, dried and pressed in a 1-mm lead microdisc between layers of potassium bromide. The microdisc was then analysed by using a suitable infrared spectrophotometer with beam condensing facilities. Results and Discussion Fibres examined under the experimental conditions described gave results that were sufficiently reproducible to be used for characterisation. This finding is in agreement with previous studies using the same instrumentation.2J With the exception of Nornex [Fig. l(g)] and Kevlar [Fig. 1 ( h ) ] , which did not pyrolyse efficiently at 600 "C and were pyrolysed a t 800 "C, all of the samples were pyrolysed at 600 "C.This temperature was found experi- mentally to provide the highest yields of characteristic pyrolysis products with both pyro- lysers. The mass pyrograms shown in Figs. 1-9 are reproduced directly from the computer output. They were all obtained by using the Pyroprobe 190, which was operated at 600 "C unless another temperature is stated. In discussing the data an attempt has been made to assign probable identities to significant ions, but it must be noted that these have not been established experimentally but were deduced from the chemical composition of the fibre and experience gained in the pyrolysis - gas chromatography of such materials. Comparison of Pyrolysers The average FIT factors calculated for the 45 comparisons between the ten samples of nylon 6 were as follows: Curie-point pyrolyser, 975.9 (coefficient of variation 1.6%); and filament pyrolyser (Pyroprobe), 994.9 (coefficient of variation 0.3%).The reproducibility of the Pyroprobe was found to be slightly better than that of the Curie-point pyrolyser and fibre handling was easier. Although both pyrolysers produced characteristic mass pyrograms when operated at optimum temperatures, the form of the pyrogram was influenced by the pyrolyser and for comparative analysis it is essential to use the same equipment. Sensitivity with very small samples (ke., less than 1 mm in length). experienced in the analysis of 2-mm lengths of single fibres (mass 0.5-1 pg). The sensitivity of the method was limited principally by the handling difficulties associated No instrumental difficulties were484 HUGHES, WHEALS AND WHITEHOUSE: PYROLYSIS - Analyst, VOZ.103 Comparison of Polymers Nylons features that permit classification are shown in Table I. Typical mass pyrograms for different types of nylons are shown in Fig. 1 and the major J 1'7 I 'II; Y L I I Fig. 1. Mass pyrograms of nylon fibres pyrolysed a t 600 "C. (a), Nylon 4, Tajmir, Alrac Co.; (b), nylon 6, Firestone Synthetic Fiber Co.; (c), nylon 6.10, Grayni, Slack Brothers; ( d ) , nylon 11, Rilsan, Rhodiaceta (Lyon); (e), nylon 6.6, Columbian Rope Co.; (f), nylon, Qiana, Du Pont UK Ltd.; (g), nylon, Nomex, Du Pont Co. (Burlington Industries), pyrolysed a t 800 "C; and ( h ) , nylon, Kevlar, sample provided by the Shirley Institute, pyrolysed at 800 "C.The mass pyrograms of nylon 4 [Fig. l ( a ) ] and nylon 6 [Fig. l ( b ) ] are consistent with these materials pyrolysing to produce large amounts of their respective monomers, vix., butyrolactam and caprolactam.May, 1978 MASS SPECTROMETRY OF TEXTILE FIBRES 485 TABLE I FEATURES PERMITTING THE CLASSIFICATION OF NYLONS BY PYROLYSIS - MASS SPECTROMETRY Nylon type Structural unit 4 . . -[NH(CH,),CO-] ,, 6 . . -[NH(CHJ,CO-] ,, 6.10 . . -[NH(CH,),NHCO(CH,)*CO-] ,, 11 . . -[NH(CH2),,CO-] ,, 6.6 . . -[NH(CH,),NHCO(CH,),CO-] mle I -l Base peak Other prominent peaks 85 30, 41, 42, 84 30 30 or 41 30 or 41 41, 55, 84, 85, 113 Discriminated by minor peaks Discriminated by minor peaks 30 41, 55, 84 Nomex - [ H ~ N H c o ~ c o ] n Kevlar - * [ ~ C O N H ~ C O N H - ] n Discriminated by pyrolysis at 600 "C 103 Nylon 6.10 [Fig.l ( c ) J and nylon 11 [Fig. l ( d ) ] gave results that varied in having either m/e 30 or 41 as the base peak but with m/e 55 as the third most abundant ion. Despite the variability of the base peak, other spectral features in the region of m/e 60-150 are sufficiently characteristic to allow unambiguous classification. Nylon 6.6 [Fig. l ( e ) ] produced pyrograms similar to that of nylon 6, but the weakness or absence of m/e 113 in the spectra from the former material was a characteristic feature. The prominent ion m/e 84 in the nylon 6.6 pyrogram was consistent with the presence of c y clopen t anone. Qiana gave very characteristic pyrograms but it has not been possible to assign identities to the major ions [Fig.l(f)]. Nomex and Kevlar are chemically very similar and it was not found possible to discrimi- nate between these two fibres by Py - MS [Fig. I(g) and (h)]. A base peak of m/e 103 attributable to C,H,CN+ or C,H,NC+ was the dominant feature of both pyrograms. The close similarity of the pyrolysates of these materials is not immediately obvious from the data illustrated in Fig. 1, which reflects the quantitative variability of pyrolysates from these materials. The reason for this variability is not understood but it is far greater than that displayed by other polyamides and may be related to the thermal stability of these fibres or to pyrolysate interactions before entering the mass spectrometer. During experi- ments with the sample of Nomex fibre in our collection it was noticed that heating of this fibre at 400-600 "C using the Pyroprobe gave rise to a product with a mass spectrum matching that of NN-dimethylacetamide.It was assumed that this compound was introduced during the fibre processing and although it is not necessarily distinctive to Nomex such information could be of value in a forensic context. Cellulose acetate and triacetate Fibres of this type differ in the extent to which the fully acetylated product has been hydrolysed and the pyrograms of both [Fig. 2(a) and ( b ) ] show major peaks at m/e 43, 45 and 60, probably indicative of the acetic acid formed on pyrolysis. Variations in the intensities of major and minor peaks were not sufficiently reproducible to allow the discrimi- nation of acetates from triacetates, although the pyrograms were unlike those of any other fibre class investigated.486 HUGHES, WHEALS AND W:HITEHOUSE : PYROLYSIS - Analyst, VoZ.103 Fig. 2. Mass pyrograms of cellulose acetate and triacetate fibres pyrolysed at 600 "C. (a), Cellulose acetate, Silene, SNIA Viscosa; and ( b ) , cellulose triacetate, Tricel, British Celanese Ltd. Polyesters Four types of polyester were studied. The sample of A-Tell was readily distinguishable from other polyesters in giving a pyrogram [Fig. 3(a)] with a base peak of m/e 94 (C,H,O+). The other three materials gave similar mass pyrograms [Fig. 3(b)-(d)] with a base peak of m/e 105 (C,H,CO+) and major peaks at :m/e 77 (c6H5+), 122 (C,H&OOH+) and 149 (OCC,H,COOH+). The relative intensity of the last ion showed great variability between replicate runs of all three fibres, but was always significantly lower in Kodel. This, together with a higher ratio of m/e 43 to 51, appears to allow Kodel to be discriminated from Terylene or Dacron.The method was unable to distinguish between the latter fibres, which differ chemically only in the end grouping of the po1:ymer chains. $1) * c PI 1 Y c 2 n Y 51 > c J Y L I I"! u 5 1 (a 1 'I'l U f E Fig. 3. Mass pyrograms of polyester fibres pyrolysed at 600 "C. ( a ) , Polyester [poly(ethylene oxybenzoate)], A-Tell, Nippon Rayon Co. ; (b) , polyester (cyclohexanedimethanol - dimethyl terephthalate copolymer), Kodel IV, Cane Mills; (c), polyester [poly(ethylene terephthalate)], Dacron, Colurnbian Rope Co.; and ( d ) , polyester [poly(ethylene terephthalate)], Trevira (g-l), Farbwerke Hoechst AG. Acrylic jibres This class of fibres is defined as having a t least 85% of acrylonitrile in the final polymer and as would be expected the incorporation of a comonomer makes only a minor difference to the resulting mass pyrogram. A lOOyo acrylic fibre, Crylor, produced a characteristic pyrogram [Fig. 4(a)] with a base peak of m/e 66 (NCCH,CH=CH+) and major ions at m/eMay, 1978 MASS SPECTROMETRY OF TEXTILE FIBRES 487 105 and 119 (C,H,N+). Zefran (an acrylonitrile - vinylpyrrolidone copolymer) gave an identical pyrogram. Acrilan (an acrylonitrile - vinyl acetate copolymer) was found to give a slightly different result [Fig. 4(b)] with an enhanced peak at m / e 43 (CH,CO+).The other common acrylic fibre, Courtelle (a terpolymer containing acrylonitrile and methyl acrylate), also gave a modified pyrogram with significantly enhanced peaks a t m/e 41 and 54 [Fig. w 58 > n b- U J W LL li: 91 181 U t E Fig. 4. Mass pyrograms of acrylic fibres pyrolysed a t 600 "C. (a), Polyacrylonitrile, Crylor, Rhodiaceta (Lyon) ; ( b ) , polyacrylonitrile - poly(viny1 acetate) copolymer, Acrilan, Israel Chemical Fibres Ltd. ; and (c), terpolymer containing acrylonitrile and methyl acrylate, Courtelle , Courtaulds UK Ltd. iModi$ed acrylics (modacrylics) Modacrylics, i.e., materials containing between 35 and 85% of acrylonitrile, surprisingly gave rise to pyrograms [Fig. 5(a)-(e)] that showed little resemblance to those of acrylic fibres.In each instance a base peak of m / e 41 and intense peaks a t m / e 39 and 27 were the principal features of the spectra. The reproducibility of replicate runs with modacrylics was poorer than for any other class of fibre. This irreproducibility complicates the task of fibre discrimination. The variation in the relative intensity of the minor peaks may provide a basis for discrimination. Polyolefins grams [Fig. 6(a) and ( b ) ] . e g . , m / e 43 (C,H,+), 55 (C,H,+), 57 (C,H,+), 69 (C,H,+) and 83 (C,H,,+). grams can be distinguished from those of the other fibre classes analysed. Both polyethylene and polypropylene fibres break down to give distinctive mass pyro- The major ions are attributable to aliphatic hydrocarbons, The mass pyro- Poly (vinylidene chloride) and poly (vinyl chloride) The mass pyrogram of Saran [poly(vinylidene chloride)] [Fig. 7(a)] is very similar to that obtained in other s t ~ d i e s .~ The major ions are m/e 36, 38 (both associated with HCl+), with smaller contributions from m / e 61, 63 (CH,=C+-Cl), 96, 98 (CH,=C+-Cl,), 146 and 148 In contrast, poly(viny1 chloride) fibres gave a pyrogram notable for the absence of ions due to HCl+ [Fig. 7 ( b ) ] . This difference can possibly be explained by the adsorption of acid on active sites before entry into the mass spectrometer and could be reconciled with the detection of HC1+ in the instance of Saran due to a much higher yield of acid. Whatever the explanation, the mass pyrogram of poly(viny1 chloride) contains many features attributable (C,H,CL+) -488 HUGHES, WHEALS AND WH:ITEHOUSE: PYROLYSIS - Analyst, Vol.103 illl I t l / E 1 1 111 w 51 n > I- -I w Y I 1 6 1 n l E n l E i"l I Fig. 5. Mass pyrograms of modacrylic fibres pyrolysed at 600 "C. (a), Teklan, Courtaulds UK Ltd.; ( b ) , Verel, Type HB, Eastman Chemical International; (c), Verel, Type F, Eastman Chemical International; ( d ) , Dynel, Type 180, Union Carbide Corporation; and (e), Kanekalon (high bulk), Kanegafushi Chemical Industrial Co. to aromatic hydrocarbons, presumably formed by free-radical recombination after pyrolysis. Thus the major ions are m/e 78 (C,H,+), 91 (C,H,+), 104 (C8H8+), 115 (C,H,+) and 128 Two modified PVC fibres, Vinyon (a copolymer with vinyl acetate) and Bristrand (a copolymer with styrene), gave significantly different pyrograms to that of PVC [Fig.7(c) and (41. With Vinyon, ions at m/e 43, 45 and 60 are indicative of the presence of acetic acid in the pyrolysate. The Bristrand fibre produced a pyrogram displaying features that (c10H8+)m )o E HlE Fig. 6. Mass pyrograms of polyolefins. (a), Polyethylene, Suddeutsche CFAG, pyrolysed at 800 "C; and ( b ) , polypropylene, Magyar Viscosa, pyrolysed a t 600 "C.May, 1978 MASS SPECTROMETRY OF TEXTILE FIBRES 489 1 r n v( Y * W I sa n f I / I VI W + U W 58 > U c -I W 8 MlE 188, ‘ I H I E M l E Fig. 7. Mass pyrograms of poly(viny1idene chloride) and poly(viny1 chloride) fil res pyrolysed a t 600 “C. (a), Poly(viny1idene chloride), Saran, Thiokol l‘ibres Canada Ltd. ; ( b ) , poly(viny1 cl Joride), Rhovyl, Rhovyl SA, France ; (c), poly(viny1 chloride) - poly(viny ’ acetate) copolymer, Vinyon, F.M.( ,.Corporation (American Viscose Division) ; and ( d ) , poly(viny1 chloride) - polystyrene copolymer, Bristrand Polymers Incorporated. closely resembled those of polystyrene (Fik-. €9, but it could be differer tiated from the latter by the intensity of the m/e 91 peak. Polystyrene The mass pyrogram of a polystyrene fibr :, Kilmarn (Fig. 8), is distinctive and closely resembles the published mass spectra of styr :ne. The principal ior s are m/e 104 (C,H,+), 103 (C,H,+) and 78 (C6H6+). The intensity of the m/e 91 ion is high6 r in the mass pyrogram than in the normal electron-impact mass spetstrum of styrene ant this difference can be attributed to the formation of toluene during pjyrolysis.Fig. 8. Mass pyrogram of polyst Irene fibre pyrolysed a t 600 “C, Kilmarn, Polymer: Incorpc ‘r- ated. Natural $fibres A selection of natural fibres together with viscose rayoi gave characteristic mass pyro- grams [Fig. 9(a)-(e)], but the complexity of these materials rt-akes it difficult to interpret the data although their value as “fingerprints” is obvious. I! iterestingly, the pyrogram of viscose [Fig. 9(e)] is different to that of natural cellulose [Fig. O(d)] with respect to variation in the ratio of m/e 29 to 43.490 HUGHES, WHEALS AND WHITEHOUSE : PYROLYSIS - Analyst, VoZ. 103 "'1 I c 2 w 5 1 a w W L i i r i z i iii z Y 5 1 a H c w L I ii s i P I Fig. 9. Mass pyrograms of natural fibres and viscose rayon pyrolysed a t 600 "C.(a), Japanese raw silk (Bombyx mori); ( b ) , natural sheep wool; (c), Caucasian head hair, dark brown; (d), cotton (Gossy+m sp.), Brazil; and (e), viscose (bright), Courtaulds (Canada) Ltd. Comparison of Pyrolysis - Mass Spectrometry with Infrared Spectroscopy All of the samples analysed in this study had been previously examined by infrared spectroscopy. When a sufficient amount of material is available the latter technique is superior to Py - MS for discriminating fibres, particularly with those samples in which copolymerisation significantly modifies the infrared spectrum of the parent polymer, e.g., the modacrylics. Nevertheless, in a blind trial conducted with 12 fibres (each 1 cm in length, mainly nylons) Py - MS was used to identify every fibre correctly, whereas infrared spectroscopy led to only two correct assignments.Certainly for the nylons, and probably also for most other fibres, Py - MS has the following advantages over infrared spectroscopy: (1) much smaller samples can be examined; (2) fibres containing appreciable amounts of filler can be assigned with greater confidence than with infrared spectroscopy; (3) the analysis time is shorter, typically 20-30 min, depending on the computer handling facilities and peripherals available. However, it should be pointed out that batch handling of fibres for infrared analysis does significantly reduce this time advantage. Limitations of Pyrolysis - Mass Spectrometry The major factor limiting our approach to Py - MS at present is the poor reproducibility for certain types of polymer.It is not clear how this irreproducibility arises but if it could be overcome then it is likely that far greater discrimination could be achieved by attentionMay, 1978 MASS SPECTROMETRY OF TEXTILE FIBRES 491 to the fine structure of mass pyrograms. Obviously we are constrained by the need to work with equipment normally utilised for gas chromatography - mass spe~trometry,~ rather than a custom-built instrument1 and reaction in the empty glass column used to spread out the pyrolysate band may be contributing to the variability of some of the results. A further disadvantage of the Py - MS procedure described, in comparison with infrared spectroscopy, becomes apparent when the pyrogram is used for interpretation rather than as a fingerprint.The ions contributing to the mass pyrogram result from two vigorous decomposition processes, ie., a thermal and an electron-impact fragmentation, and not surprisingly, it is often difficult to relate the final ions to the starting polymer. The use of chemical ionisation to reduce fragmentation within the mass spectrometer could yield results more amenable to qualitative interpretation,6 and pyrolysis - gas chromatography followed by mass spectrometry is an obvious method of providing information on the significance of particular ions in a mass pyrogram. Conclusions This study indicates that Py - MS provides a rapid and sensitive method for the charac- terisation of synthetic fibres, which could have a wider range of application than infrared spectroscopy for forensic fibre examination. The equipment required for the technique is expensive but those laboratories currently using a mass spectrometer and data system could adapt their instruments at low cost. Provided that the reproducibility of Py - MS can be improved it should become a powerful method for the microanalysis of a wide range of natural and synthetic polymers. The infrared spectroscopic method described in this paper has been in use for some years in this laboratory. The development of this procedure was the work of R. Cook and co-workers at the Metropolitan Police Forensic Science Laboratory. References 1. 2. 3. 4. 5. 6 . Meuzelaar, H. L. C., and Kistemaker, P. G., Alzalyt. Chem., 1973, 45, 587. Hughes, J. C., Wheals, B. B., and Whitehouse, M. J., Forensic Sci., 1977, 10, 217. Hughes, J. C., Wheals, B. B., and Whitehouse, M. J., Analyst, 1977, 102, 143. Senoo, H., Tsuge, S., and Takeuchi, T., J , Chrornat. Sci., 1971, 9, 315. Zeman, A., in Wiedmann, H. J.. Editor, “Thermal Analysis,” Volume 3, Proceedings of the 3rd Saferstein, R., and Manura, J. J., J . Forensic Sci., 1977, 22, 748. International Conference on Thermal Analysis, Rirkauser Verlag, Basle, 1972, pp. 219-227. Received September 22nd, 1977 Accepted November 24th, 1977