You are on page 1of 4

44. K. W. Street, Jr., W. E. Acree, Jr., J. C. Fetzer, P. H. Shet, and C.

F. Poole, Appl. Spectrosc. 43, 1149 (1989).
45. S. A. Tucker, L. E. Cretella, R. Waris, K. W. Street, Jr., W. E. Acree,
Jr., and J. C. Fetzer, Appl. Spectrosc. 44, 269 (1990).
46. U. Breymann, H. Dreeskam p, E. Koch, and M. Zander, Chem.
Phys. Lett. 59, 68 (1978).
47. S. A. Tucker, J. M. GrifŽ n, W. E. Acree, Jr., P. P. Mulder, J. Lugtenburg, and J. Cornelisse, Analyst (Cambridge, U.K.) 119, 2129 (1994).
48. L. C. Sander and S. A. Wise, Polycyclic Aromatic Hydrocarbon
(NIST Special Publication 922, Washington, 1997).
49. S. L. Murov, I. Carmichael, and G. L. Hug, Handbook of Photochemistry (Marcel Dekker, New York, 1993), 2nd ed.

Forensic Analysis of Single Fibers by
Raman Spectroscopy*
JASON V. MILLER† and
EDWARD G. BARTICK‡
Department of Forensic Sciences, The George
W ashington University, W ashington, District of
Columbia 20052 (J.V.M.); and Forensic Science
Research Unit, FBI Academy,§ Quantico, Virginia,
22135 (E.G.B.)

Index Headings: Forensic analysis; Fibers; Raman spectroscopy;
Ram an m icroscopical analysis; Ram an m icroprobe; Ram an
microspectroscopy.

INT RODUCTIO N
Polymeric identiŽ cation is an important step in the forensic analysis of transferred Ž bers collected as evidence.
In addition to physical and optical comparison of known
and questioned Ž bers, polymeric identiŽ cation of the Ž ber
provides the forensic scientist with information on which
to base conclusions, such as transfer behavior and prevalence in the community.1 After collecting potentially signiŽ cant Ž bers and screening them with a stereo microscope, they are mounted on glass slides for further examination by comparison, polarized light (PLM), and
 uorescence microscopes. Optical properties such as index of refraction, birefringence, and optical sign of elongation are used to determine the generic class of the Ž ber.
As Ž ber technology has progressed, new generic and subgeneric classes of Ž bers have been developed. PLM cannot be used to differentiate between subgeneric classes;
therefore, a method with additional discrimination is nec* The bases of this note was presented at the 26th Annual Conference
of the Federation of Analytical Chemistry and Spectroscopy Societies
(FACSS), Vancouver, British Columbia, Canada, Oct. 24 –29, 1999.
† Current Address: Explosives Unit, Laboratory Division, Federal Bureau of Investigation, 935 Pennsylvania Avenue, N.W., Washington,
DC 20535.
‡ Author to whom correspondence should be sent.
§ This is publication number 01-21 of the Laboratory Division of the
Federal Bureau of Investigation. Names of comm ercial manufacturers
are provided for identiŽ cation only and inclusion does not imply endorsement by the Federal Bureau of Investigation.

essary to m ore fully characterize the chemical composition of Ž bers.
M icroscopical analysis by FT-IR spectroscopy has
been the preferred m ethod for obtaining additional information on the polymeric composition of a Ž ber. 2– 4 It allows unequivocal determination of the generic class and
frequently permits subclassiŽ cation.5,6 FT-IR analysis
cannot be conducted while the Ž ber is mounted due to
the strong absorption of infrared (IR) radiation by glass.
The Ž ber must be rem oved from the slide and rinsed to
clean off the mounting medium before preparing the Ž ber.
This increases sample preparation time and the chance
for sample loss.
Glass produces little response to Raman spectroscopy.
Therefore, Raman analysis of Ž bers can be conducted
while the Ž ber remains mounted on a glass slide under
coverslips. The theoretical beam width of the laser is 0.8
mm; thus, Raman will potentially permit m easurement of
microŽ bers used in some fabrics in the current market.
The samples do not need to be thinned or  attened, as
with IR, because the photons are scattered from the surface and not transm itted through the sample.
The current literature contains limited reports and citations on the use of Raman m icroprobe spectroscopy for
forensic analysis of Ž bers. Research by Lang et al. 7 has
shown that it is possible to scan Ž bers while they are
taped onto a glass slide. By the same sample preparation
method, Keen et al.8 have shown that it is possible to
differentiate between several generic and subgeneric classes of Ž bers. This report will demonstrate that it is possible to scan through a glass coverslip and m ounting medium to obtain polymeric identiŽ cation of nondyed Ž bers.
EXPERIMENTAL
Undyed Ž ber samples were obtained from the Collaborative Testing Ser vices (CTS) (Herndon, VA) 1987 Collection and from an FBI collection acquired from manufacturers. Three sample preparation methods were used:
(1) Ž bers were taped directly to glass m icroscope slides
for analysis, (2) samples were taped to the shiny side of
aluminum foil covering microscope slides, and (3) samples were m ounted in Perm ount, produced by Fisher ScientiŽ c (Pittsburgh, PA), under glass coverslips on microscope slides. Including the subclasses, a total of 70 Ž ber
types were analyzed by these methods.
Raman spectra were collected using a Renishaw (Gloucestershire, United Kingdom) Model 2000 spectrograph
attached to an Olympus (M elville, NY) BHSM metallurgical m icroscope. A 785 nm, 300 m w, continuous wave,
wavelength-stabilized laser diode (SDL, San Jose, CA)
was used as the excitation source. For all scans, a 503
microscope objective was used that produced a theoretical 0.8 mm spot size on the sample. Using 1808 backscattering geometry, the scattered light was collected by
the objective and passed through a holographic notch Ž lter to remove the Rayleigh line. The detector was a thermoelectrically cooled charge-coupled device (CCD) detector of 576 3 384 pixels. Spectra were collected in the
range of 200 to 2000 cm 2 1 except for acrylic and modacrylic Ž bers, which were collected from 200 to 2400
cm 2 1 . Spectral acquisition, presentation, and analysis
were performed with the Renishaw W IRE and
APPLIED SPECTROSCOPY

1729

Nylon 6. Permount is a formulation of pinene.6-di-tert-butyl-P-cresol. The Ž nal spectrum after baseline  attening is shown in Fig. are within one wavenumber of Hendra’s results. Along with the large broad peak near 1400 cm 2 1 due to the  uores1730 Volume 55. several additional peaks from the Permount are evident. 785. Raman spectra of a nylon 6 Ž ber mounted on a glass slide in Permount with a coverslip. After baseline  attening. 2D shows the nylon 6 spectrum obtained on alum inum foil. Attempts to reduce the intensity of these residual bands by changing the thickness compensation factor resulted in negative nylon 6 bands. Spectra of a blank glass slide and Permount between a coverslip and a glass slide were collected daily for subtraction from the Perm ount-mounted Ž ber spectra. therefore. (B) nylon 6 taped on an aluminum foil slide. 2C. the high-quality spectrum in Figure 1C results. Hydrogen bonding with the Ž ber appears to have broadened the P erm o unt band s. The ability to distinguish nylon 6 from nylon 6. and a second subtraction of the glass slide was necessary to remove the contribution from the glass. Perturbation from intermolecular interaction due to apparent hydrogen bonding between the Permount and the Ž ber10 was observed on the Ž nal subtraction spectrum shown in Fig. The spectrum is shown with major peak assignments as described by Hendra et al. no interference from the tape occurred. (B) Permount on aluminum foil. To remove the interference contributed by the peak. and 2. with the exception of the 1078 cm 2 1 peak. Figure 1A shows a spectrum of a nylon 6 Ž ber that is typical of Ž bers taped directly on a glass slide. The use of aluminum foil to back the Ž ber eliminated the  uorescence of glass and the subtraction step. thu s p rodu cing seco nd d erivativeshaped curves after subtracting the spectrum of Permount. Salem. (A) nylon 6 taped on a glass slide. (D ) nylon 6 spectrum obtained on aluminum foil-covered slide.6 is also frequently seen as evidence in forensic laboratories. The spectra were acquired using scan time settings of 120 s for the aluminum foil slides and four coadded scans of 30 s for the Permount-mounted slides. Note that the underlying broad peak around 1400 cm 2 1 is now absent.9 who analyzed nylon pellets using an FT-R aman system. because the software time settings are for the portion of the spectrum for each grating position. cence of glass. Figure 1B shows a spectrum of a nylon 6 Ž ber taped on an aluminum foil-covered slide. obtained by the Gramsy peak-picking routine. Total analysis time in both cases was approximately eight m inutes. The bands do not interfere with the interpretation of the spectra and are m inimal compared to perturbation effects that are often obser ved in spectral subtraction calculations. The instrum ent was calibrated for frequency daily to within one wavenumber of naphthalene. Nevertheless. Artifacts are indicated by arrows where the m ajor Permount bands were subtracted. Often. Figure 2A shows a spectrum of a nylon 6 Ž ber typical of Ž bers mounted in Permount. The portions of the spectrum are digitally stitched together to complete the total spectral range covered. GRAM Sy/32C (Galactic Industries.6 provides an additional subgeneric discriminat- . and 525 cm 2 1 . and the peak frequencies are within one wavenumber when comparing the subtracted Permount slide with the aluminum slide and the spectra obtained by Hendra et al. A background spectrum taken adjacent to the Ž ber was subtracted to eliminate the Permount peaks. Next. Number 12. This is also true after a 13-point Savitsky–Golay sm ooth is perform ed on the Ž nal spectrum from the glass mount. 2B with the major peaks near 1005. the broad peak from the glass  uorescence was still present. No instrum ent response corrections were made. characteristic bands are still obser vable. RESULTS AND DISCUSSIO N Fibers were initially taped on top of glass slides and analyzed directly where there was no tape. a background spectrum of the glass slide alone must be subtracted from the Ž ber spectrum acquired on the glass. toluene. Raman spectra of a nylon 6 Ž ber on different mounts. A spectrum of neat Perm ount is shown in Fig. 1. Fig. which is two wavenumbers different. 2C. Raman spectra of Ž bers mounted in Permount generally have a lower signal-tonoise ratio than those m easured on an aluminum foil slide.F IG . 2001 F IG . (C ) nylon 6 on an aluminum foil slide after baseline  attening. 2. For comparison. Fluorescence from the glass slide produces a large broad underlying peak near 1400 cm 2 1 . studies were conducted on Ž bers mounted under glass coverslips. (C ) nylon 6 mounted in Perm ount after subtraction of Permount and glass slide and baseline  attening plotted full scale (arrows show subtraction artifacts). The gratings are moved incrementally to cover the spectral range speciŽ ed. NH) software. The peak frequencies. (A) nylon 6 mounted in Permount (arrows show major Permount peaks).

Likewise. but the carbonyl and the C–O stretches do not appear in these spectra. Figure 4 shows IR spectra of nylon 6 and nylon 6. No other features are observable to identify them as glass Ž bers. M inor variations in band shapes are the only differences. The comonomers shown in Figs. Additional comonomers and additives show characteristic features in the IR that permit further discrimination between the subclasses of acrylic Ž bers. Figure 5 contains four IR spectra of m ajor subgeneric classes of acrylic Ž bers. and 5D F IG . 5C. Sarans differ only in the amount of plasticizer used in the pro- F IG . ing factor that could prove useful in a forensic investigation. (A ) p o ly a c r y lon itrile h o m op o ly m e r. frequently used in sweaters. (C ) poly(acr ylonitrile : m ethyl m ethacr ylate).F IG . the features distinguishing the subgeneric classes are m ore subtle. (B) nylon 6. respectively.6 and nylon 6 is by the presence of the weak crystalline band near 935 cm 2 1 in the nylon 6.6. (D ) poly(acr ylonitrile : vinyl acetate). are shown in Figs. have only m inor notable differences. Raman acrylic Ž ber spectra. it is necessary to limit the computer search to 1150 –900 cm 2 1 to obtain the correct results for the nylon subclasses. Raman analysis is not useful for identifying glass Ž bers while mounted on m icroscope slides. Generally. The differentiating features in these spectra lie in the C–O stretching region between 1300 –1100 cm 2 1. Therefore. 5. The major characteristic spectral feature for acrylics is the nitrile band near 2240 cm 2 1 shown in Fig. An exception is glass Ž bers on m icroscope mounts. it is useful to compare the information gained from Raman spectroscopy to that from the IR. 6. with their polymeric structures.6 in Figs. can readily transfer during contact. 4A and 4B. Com parison of infrared spectra of four m ajor acr ylic su b-g e ne ric c la sse s. IR analysis is useful for discriminating over 20 known subclasses. and the C–C skeletal stretches are shifted from the peak locations in the nylon 6 spectrum. 5B. 11 The nylon 6. C–C bonds found in the skeletal backbone structure of many polymeric Ž bers readily polarize and will be good Raman scatterers. (A) nylon 6. 3. 5A of polyacrylonitrile homopolymer. (B) nylon 6.12 Acrylics are composed of at least 85 percent by weight of polyacrylonitrile along with other comonomers and ionic end groups. 3A and 3B.9.6 spectrum. Acr ylic Ž bers are another important generic class in forensic analysis because these Ž bers. In addition to acrylics.13 As a result of variations in composition. acrylic Ž bers are not likely to be subclassiŽ ed using Raman analysis. The most convenient way by IR to differentiate between nylon 6. Infrared spectra of two nylon Ž bers with the distinguishing peaks labeled. To determine the differences between the different numbered series of nylons. APPLIED SPECTROSCOPY 1731 .6 spectrum does not have an amide III band (C–N stretch and N–H bend) at 1281 cm 2 1. 4. W hen conducting IR spectral searches of nylons. Application of m ultivariate analysis to Raman spectra may lead to improved discrimination of such spectra. The major bands are characteristic for polyamides and m ake the identiŽ cation of the generic class a simple task. In comparison to the IR spectra. The peak assignments used to differentiate between the nylons are listed on the spectra. shown in Fig. and they are easier to distinguish. Comparison of Raman spectra of two nylon Ž bers mounted in Permount with the distinguishing peaks labeled.6. contain carbonyls that absorb near 1730 cm 2 1. The nitrile stretch is obser vable near 2240 cm 2 1. Since Raman spectroscopy provides inform ation that is complimentary to IR. respectively.6. However. M ost other Ž ber types studied were found identiŽ able by Raman analysis. producing many characteristic peaks. Spectra of nylon 6 and nylon 6. but with the addition of dye features. (B ) poly(acr ylonitrile : m ethyl) acr ylate. it is necessary to carefully compare the small skeletal-region bands between 1150 –1000 cm 2 1 . acquired in the slide mount. saran Ž bers are the only class that cannot be subclassiŽ ed by Raman spectroscopy. Differences are m ore apparent in the Raman spectra. Glass Ž bers produce the same  uorescence response in their spectra as the glass coverslips. strong dipole moments from functional group vibrations will absorb strongly in the IR. The two nylon spectra are very similar. (A) nylon 6.

2001 Raman spectra of single Ž bers were obtained directly from glass slide mounts. Royaud.. M. p. O’Keefe. 2. 6. Maehly and R. p. Tungol. 41. (Springer–Verlag. J. Humecki. 10.. Appl. G. Microchem. New York. and A. M. Lang. Montaser. A. (Taylor and Francis Inc. Part A 47. J. 1999). Soc. W.. spectral differences are too subtle for clear distinction between subclasses. Katon. in Forensic Examination of Fibers. Chap. Philadelphia. Schiering. F. 13. Subgeneric identiŽ cation by Raman analysis was successful for nondyed Ž bers with the exception of acrylics and sarans. 11. C. Comparison of Raman spectra of four m ajor acr ylic subgeneric classes. C. 8. A. E. (B) poly(acrylonitrile : methyl acrylate). 179. ‘‘Fibres and Their Examination in Forensic Science’’. 319 (1986). H. and A. M. 3. C. Grieve. Spectrosc. M. 12. W. 1732 Volume 55. 543 (1990). 2nd ed. Tungol. 1535E (1991). J. M . Grieve. 747 (1990). W. Raman spectra. ‘‘Infrared Microspectroscopy of Fibres’’. A. J. Royaud. ‘‘From the Crime Scene to the Laboratory’’. and V. in Forensic Science Progress. Spectrosc. E. W. p. J. p. 44. Forens. W. 179 (1995). Smalldon. Maddams and I. 245. M ontaser. 14. Kizer. 5. F. Masters Thesis. 9. Glass Ž bers are an exception because they cannot be analyzed by this method. Sci. G. and therefore. Bartick. P. 7. I. 34. G. 89. 345 (1975). and A. Part A 46. Table I shows the generic classes of Ž bers with the ability to identify generic classes and subclasses of nondyed Ž bers based TABLE I. Zichy. 4. M. IR Raman Fiber class Class Subclass Class Subclass Acetate Acr ylic Aramid Azlon Fluorocarbon Glass Modacrylic Natural Novaloid Nylon Polycarbonate Polyester OleŽ n Rayon Saran Spandex Sulfar Vinyl Vinyon good good good good good good good good good good good good good good good good good good good poor/difŽ cult good good good good NA a good good NA good NA good good NA good good NA NA NA good good good good good poor good good good good good good good good good good good good good good poor good good good NA a good good NA good NA good good NA poor good NA NA NA a NA indicates that there are no subclasses in this generic class. 15. no additional sample preparation was required after visual light microscopy. DC (1999). P. Robertson and M. Philadelphia. (Marcel Dekker. duction of the Ž bers. A. J. Science & Justice 35. 1999). Spectrochim. 1327 (1991). Acta.. The ester bands of the plasticizer are not observable in the Raman spectra. ‘‘Forensic Examination of Synthetic Textile Fibers by M icroscopic Infrared Spectrometry’’. I.. Appl. Bartick. The IR and Raman spectra were compared by M iller 14 for the entire 70 Ž bers analyzed. Ed. 1995). 1. Grieve. Hendra.. (C ) poly(acrylonitrile : methyl methacr ylate). 82 (1998). CO NCLUSIO N F IG . 6. in Forensic Examination of Fibers. Fredericks. New York. Eds. Inc. 29. F. 43. 17 (1975). Ability to use IR vs. 8. J. (A) polyacrylonitrile homopolymer. M. 2nd ed. G. Montaser. L. Grieve. Forensic Sci. Willis. Chap. Williams. Washington. J. With these Ž ber classes. L. W hite. in Practical Guide to Infrared M icrospectroscopy. Tungol. P. Keen. Part B 46. IR analysis is recommended for subclassiŽ cation of acrylics and sarans. Roux. W. T. Chap. 5. E. Robertson and M. Raman analysis in Ž ber class and subclass identiŽ cation of nondyed Ž bers.on IR vs. W. A study dealing with Raman analysis of dyed Ž bers will be addressed in a forthcoming publication. Acta. Eds. . P. 7. W. Spectrochim. Kirkbride and M. and P. E. J. Eds. therefore. Acta. Spectrochim. Hirschfeld and K. W. The George Washington University.. and D. Robertson and C. Number 12. K. 1990). V. Miller. M . Pounds and K. Maddams. J. Bartick. J. H. (D ) poly(acrylonitrile : vinyl acetate). (Taylor and Francis Inc. Tungol. A.