The Transfer and Persistence of Fibres on Bare Skin by Hilary J Burch September 2008 Centre for Forensic Science The Forensic Science Service University of Strathclyde Huntingdon Laboratory Royal College Hinchingbrooke Park 204 George Street Huntingdon Glasgow, G1 1WX Cambridgeshire, PE29 6NU
The Transfer and Persistence of Fibres on Bare Skin by Hilary J Burch A thesis submitted to the Centre for Forensic Science, University of Strathclyde, in part fulfilment of the regulations for the degree of Master of Science in Forensic Science. September 2008
Acknowledgements First and foremost I would like to thank my supervisor Ray Palmer for making this project possible and for sharing his fibre knowledge with me. Thank you also to Jon Butcher for his help with the MSP and to Gavin Foad for taking the photographs in this thesis. Thank you also to my parents for their financial support throughout this MSc. Finally thank you to all those people at the Forensic Science Service in Huntingdon who kindly volunteered to be covered in fibres and then taped. i
Contents 1 Abstract 1 2 Introduction 2 2.1 Introduction to fibres................................ 2 2.1.1 Classification................................ 2 2.1.2 Textiles.................................... 2 2.2 Fibres in forensic science.............................. 4 2.2.1 Population studies............................. 5 2.2.2 Target fibre studies............................. 6 2.2.3 Fibre transfer studies............................ 7 2.2.4 Fibre persistence studies.......................... 7 2.3 Aims of this project................................. 8 2.4 Previous relevant work............................... 8 2.4.1 Transfer of fibres to garments....................... 8 2.4.2 Persistence of fibres on garments..................... 9 2.4.3 Redistribution................................ 10 2.4.4 Differential shedding............................ 10 2.4.5 Fibre persistence on skin.......................... 11 2.4.6 Fibre persistence in head hair....................... 12 3 Experimental 14 3.1 Transfer and persistence experimental design.................. 14 3.1.1 Target fibres................................. 14 3.1.2 Taping.................................... 17 ii
3.1.3 Length.................................... 19 3.1.4 Differential shedding............................ 19 3.1.5 24 and 48 hour persistence experiments................. 19 3.2 Population study experimental design...................... 20 3.2.1 Perceived colour.............................. 21 3.2.2 Generic class................................ 22 3.2.3 Length.................................... 22 3.2.4 Delustrant.................................. 22 3.3 Analytical techniques................................ 23 3.3.1 Comparison microscopy.......................... 23 3.3.2 Microspectrophotometry......................... 23 3.3.3 Polarized Light Microscopy........................ 25 4 Results and discussion 28 4.1 Transfer and persistence.............................. 28 4.1.1 Transfer experiments............................ 28 4.1.2 Persistence experiments.......................... 29 4.1.3 24 and 48 hour persistence experiments................. 32 4.1.4 Differential shedding............................ 33 4.1.5 Length.................................... 33 4.2 Background population study........................... 37 4.2.1 Perceived colour.............................. 37 4.2.2 Length.................................... 39 4.2.3 Fibre generic class............................. 41 4.2.4 Delustrant.................................. 46 iii
5 Conclusions 47 5.1 Transfer and persistence.............................. 47 5.2 Population study.................................. 48 6 Suggestions for further work 49 6.1 Transfer and persistence.............................. 49 6.2 Population study.................................. 49 7 Appendix 50 7.1 Significance tests.................................. 50 7.2 Raw Data....................................... 50 iv
1 Abstract Cotton, polyester and wool fibres were transferred to the bare arms of male and female volunteers, and their persistence determined at intervals up to 48 hours, during which normal office/laboratory work was undertaken. Decay curves for the persistence of each fibre type on bare skin showed an initial rapid loss followed by an approximately exponential decay. After 5 hours approximately 15% of cotton and polyester fibres and 5% of wool fibres remained on bare skin. The length distribution of the fibres was also monitored during this time and showed a shift towards shorter fibre lengths after 5 hours. For a garment composed of a cotton/polyester mix, its differential shedding properties were determined upon initial transfer, after 2 hours, and again after 5 hours. At each stage a small bias towards the retention of cotton was observed. No target fibres were found to remain after 24 or 48 hours where volunteers had bathed during that interval. The skin of volunteers was blanked by taping before the transfer of target fibres, and these tapings were used to determine the background population of coloured fibres on bare skin. Background fibres were classified according to perceived colour, length, generic class (cotton, polyester, regenerated cellulosic, acrylic and wool) and for synthetic fibres, the presence or absence of delustrant. The majority of fibres were natural and black/grey in colour, with the most common fibre type observed on bare skin being black/grey cotton. Most man-made fibre types were comparatively rare, except for polyester. 1
2 Introduction 2.1 Introduction to fibres 2.1.1 Classification Textile fibres are usually broadly classified as either occurring naturally or being manmade. Within these two classes are various sub-classes as shown in fig. 1 many of which can be further sub-divided. Only a small number of the generic fibre types listed in fig. 1 were encountered during this project. Natural fibres such as cotton and wool are derived respectively from plants and animals, but are usually subject to further processing before their use in textiles, such as the chemical cleaning and scouring of wool. 1 Man-made fibres are manufactured from fibre-forming polymers. These polymers may be entirely synthetic or they may be regenerated from natural polymers. For example, polyester is a synthetic polymer, but viscose is formed from regenerated cellulose. 2.1.2 Textiles Textiles are made by twisting fibres into yarns which are then woven or knitted together into a fabric. Fibres shed from textiles when they are pulled out of the weave or knit by contact with another surface, broken into fragments by friction, 2 or fall out because they are short. Short fibres are known as staple fibres, of which cotton is an example. Staple fibres shed easily from garments because they are held only weakly by the weave or knit. In comparison, the length of man-made fibres can be controlled, and as a result these fibres shed less easily from textiles. The strength of fibres (as determined by their degree of crystallinity) also affects their propensity to shed from textiles. Staple fibres are generally more brittle and susceptible to breakage than man-made fibres. 2
3 Figure 1: A basic classification scheme for textile fibres adapted from Robertson. 1
2.2 Fibres in forensic science Many crimes involve some form of physical contact between offender and victim, or between offender and scene. 2 Any fibres transferred during these contacts can provide valuable evidence, 3 because they establish associations between people, locations and objects, 4,5 and can also give an indication of the time frame for the contact. The main sources of fibres in forensic investigations are clothes, carpets, bedding and upholstery. Casework will often involve the examination of an item for fibres which may match control fibres, taken from a known source. The analysis consists of a number of sequential steps. First, the item will be examined using a search microscope. Any possible matches for the control fibres (referred to as recovered fibres) are then removed from the item and mounted onto microscope slides. Next, a comparison microscope (section 3.3.1) is used to compare the physical characteristics (such as the colour and fluorescence) of the recovered fibres with the control fibres. If the fibres cannot be distinguished at this stage, microspectrophotometry (section 3.3.2) and thin layer chromatography can be used to analyse the fibre dye, and infra-red spectroscopy can be used to analyse the polymer type of man-made fibres. If the recovered fibres are found to match the control fibres, the evidential value of this match must be considered within the context of the case. The evaluation of fibre evidence is complex because many interdependent factors must be considered. 6 These include the colour and type of fibre, the number of fibres, and the situation from which they were recovered. 7 Other factors such as changing textile markets, climate, 8 fashion and tradition will affect background fibre populations, and should also be taken into account. 9 Over the past decade there has been a move towards employing Bayes theorem in the evaluation of fibre evidence. 10 12 This involves the generation of a likelihood ratio, 13 4
which is defined as: p(e H) p(e H) (1) In equation (1) H and H are two competing hypotheses such that the numerator is the probability of recovering the matching fibre types given that the suspect did commit the offence and the denominator is the probability of recovering the matching fibre types given that the suspect did not commit the offence. If numbers can be assigned to both parts of the likelihood ratio, an overall numerical value can be obtained which gives an estimate of the strength of the fibre evidence. The key task in this approach is the calculation of the probabilities. A number of fibre studies have been published which can assist in the evaluation of fibre evidence, whether a Bayes treatment is employed or not. These broadly fall into three categories: population studies, target fibre studies, and transfer and persistence studies. 2.2.1 Population studies The number of suspect fibres recovered in a forensic context is often only a small proportion of the total number of extraneous fibres found on the surface of an item under examination. As such it is often argued that the presence of these suspect fibres is a pure coincidence. 5 A key factor therefore in the interpretation of fibres evidence is that of estimating fibre frequency. 8 Population studies provide fibre frequency data by investigating and reporting the components of a fibre population on a chosen surface. 4 As a general rule, the less frequent the fibre, the higher the potential value of the evidence. 6 Although useful, population studies must be interpreted with care, as they can only indicate the relative rarity of a generic match. The very basic classification methods used (generic type and perceived colour) can misrepresent the evidential value of certain fibre types (both common fibre types e.g. blue cotton and more unusual fibre types e.g. acetate 5
or silk 8 ) because only the initial stages of the detailed examination outlined in section 2.2 are considered. When applied to a case scenario, the origin of any population data used should be carefully considered. Ideally the population chosen should be the fibre population in the vicinity of the crime at the time of the offence 14 although in practice this is rarely possible. Population studies have been reported for a variety of surfaces including human head hair, 15 cinema seats, 16 washing machines, 4 car 6 and bus 9 seats and outdoor surfaces (e.g. lamp posts, park benches and roof tiles). 17 The results of these studies will be considered in more detail in section 4.2. Fibre populations from human skin have yet to be reported in the literature. 2.2.2 Target fibre studies Target fibre studies assess the degree of random occurrence of certain, specific fibres (known as target fibres) in the general extraneous fibre population. 7 For example, Palmer and Chinherende assessed the random occurrence of red acrylic and green cotton fibres on car and cinema seats. 18 Target fibre studies are much more specific than population studies because the full range of experimental tests (see section 2.2) is used to produce the frequency data. In a target fibre study, the colour, dye and sub-type (e.g. nylon 6,6 rather than polyamide ) of the target fibres and any possible matches are examined. Target fibre studies reported in the literature have generally concluded that to find any more than a small number of positive matches at this level of analysis, by pure coincidence is very unlikely. 5 However, again the data most pertinent to a case scenario, are those which are obtained when the type of surface, the season and the geographic location of a study are very similar to those in the case under investigation. 6 6
2.2.3 Fibre transfer studies In this type of study, fibre transfer between objects or people is simulated and the number of fibres transferred is subsequently counted. For example, in a series of experiments by Pounds and Smalldon 2,19,20 red and brown, wool and acrylic fibres were transferred to recipient garments and the effect of variables such as pressure, surface area and duration of contact were investigated. These, and subsequent studies will be considered in more detail in section 2.4. 2.2.4 Fibre persistence studies In persistence studies target fibres are transferred to the object (known as seeding) and allowed to remain for successive intervals of time, after which the remaining fibres are counted and a decay curve is produced. For example the persistence of wool and acrylic fibres (originating from ski masks) in head hair was investigated by Ashcroft et al. 21 Persistence studies are used to estimate the number of fibres which are likely to remain on a surface following a certain time period after an offence. 19 They can be limited in their usefulness however, as the time elapsing before recovery of the clothes of a suspect may vary from a few hours to several months after the crime. 22 All fibre persistence studies are further limited because of the difficulty in producing exactly the same absolute populations in repeat experiments. This difficulty arises because of the numerous variables involved in fibre transfer, and the difficulty in controlling these variables. However, valid generalisations can still be made using fibre transfer and persistence experiments. 3 Fibre persistence studies will be dealt with in more detail in section 2.4. 7
2.3 Aims of this project There are two main aims to this project: The first is to investigate the transfer and persistence of fibres on bare skin with respect to the factors of fibre type, gender, differential shedding and length. The second is to investigate the background population of fibres on bare skin, and classify any background fibres by perceived colour, length, and generic type. 2.4 Previous relevant work 2.4.1 Transfer of fibres to garments Garments have been the most thoroughly studied recipient surface for transfer and persistence experiments. This is because fibre transference from the outer clothing of an assailant to a victim and vice versa is of particular interest in the investigation of many crimes. 19 The number of fibres transferred depends upon the nature of both the donor and recipient surface e.g. its texture. 23 In the experiments of Pounds and Smalldon 19 fibres were transferred to recipient garments via a polystyrene block. The results showed that the number of fibres transferred increased with the area of contact and with pressure, but decreased with the number of contact passes, suggesting some back transference was occurring. Kidd and Robertson 23 transferred acrylic, wool, cotton, polyester and viscose fibres to recipient garments using an abrasion tester, and found that there was a threshold pressure above which no further increase in the number of fibres was observed. Grieve et al. 24 used more realistic transfer methods (including a simulated struggle) in their investigation of the transfer of red acrylic fibres. They found that the number of fibres transferred during the struggle greatly exceeded those counted after more casual contacts or as a re- 8
sult of secondary transfer. The mechanism of fibre transfer is predominantly mechanical, with some contribution from electrostatic forces. 2 2.4.2 Persistence of fibres on garments Previous studies into fibre persistence on garments during normal wear, show an initial rapid loss of fibres followed by a subsequent slower loss, 25 with 0-10% of fibres remaining after 8 hours of wear 20 (see fig. 2). Thus for many garments examined in casework, only a few fibres at most can be expected to remain from any initial contacts. 20 Figure 2: A typical persistence decay curve for fibres on garments, reproduced from Pounds et al. 20 Fibres which have been transferred to a garment via a high pressure contact persist longer than those transferred by a low pressure contact. 22,25 The shape of the decay curve in fig. 2 suggests that two or more parallel first order processes may occur during fibre loss, reflecting the possible different states of fibres e.g. strongly or weakly bound. 9
It is intuitive that fibres which are weakly bound will be rapidly lost from the surface of garments, such that as wearing time increases, a larger proportion are likley to be strongly bound thus more difficult to remove. 2,20 The rate at which fibres were lost from the surface of garments did not depend on the type or length of the donor fibre, 20 but did depend on the recipient garment texture. 25 Krauß 22 monitored the persistence of yellow cotton, yellow polyester, and pink wool fibres on garments which were subsequently left in open-air conditions. When compared to the losses recorded during normal wear, a high percentage of fibres was found to persist on these garments. This reflects the importance of post-transfer activity in estimating fibre persistence. 2.4.3 Redistribution The term fibre loss when used in persistence experiments may be misleading, as some fibres will be redistributed away from the contact area, rather than lost altogether from the recipient garment. 26 Redistribution occurs both to other garments worn by the person involved and to other areas of the same garment. Therefore, considerable caution should be exercised in placing too much significance on the topographical distribution of a small number of fibres. 25 2.4.4 Differential shedding Many garments are composed of two or more different fibre types. It is important therefore to know whether the respective fibre types will transfer in proportion to their distribution in the garment. The actual numerical proportion of fibres in the fabric will depend upon their respective densities, diameters and lengths (usually made equivalent in blended garments in order to facilitate spinning). 27 It is important to note that the gar- 10
ment label does not always accurately reflect the fibre distributions counted upon transfer. 5,23,25,27,28 Often, deviation from the label is biased in favour of the cellulosic or wool fibres. 27 For example, in an 80:20 wool/nylon garment, the wool was found to consistently shed proportionately more that the nylon. 5 Bias in favour of cellulosic fibres, although not always observed, 23 may be due to greater fragmentation of these fibres during contact as compared to the synthetic fibres, and due to their relative strengths. 29 However, other factors such as recipient surface texture 27 and the construction of the weave or knit of the shedding garment are also likely to be significant. 2.4.5 Fibre persistence on skin As there are no studies on the transfer and persistence of fibres on human skin in the literature, no direct comparisons can be made. As a first approximation however, the skin of both living and dead humans can be considered to be a smooth surface. It has been reported that the number of fibres transferred to smooth surfaces (such as cotton laboratory coats) is lower than the number transferred to rough surfaces. 20 This is because smooth surfaces cannot readily accept loose fibres and therefore any loose fibre equilibrium is strongly in favour of the transferring material. Direct fragmentation also seems unlikely on a smooth surface. 2 In addition, for smooth garments decay is rapid, 25 with only a few fibres remaining after 2 hours. 20 The behaviour of human skin can also be approximated using pig skin. 30 Krauß and Hildebrand 31 seeded pig skin with wool, cotton and polyester fibres and left it in open-air conditions for a number of days. The number of persisting fibres was found to depend on the weather conditions, such that when the combination of wind and precipitation was recorded, fibre loss increased dramatically. The experiments never showed a total loss of fibres, suggesting that the probability of finding fibres originating from the offender s 11
clothing on the skin of a homicide victim is very high, even when the corpse has been exposed to the elements for several days (up to two weeks). 31 Davidson and Riley 30 seeded samples of wet and bloodstained pig skin with acrylic and wool fibres, and recovered the fibres with adhesive tape. In comparison to dry surfaces, there was a lower fibre recovery rate from the wet and bloodstained skin. This may have been due to the water and blood reducing the adhesive properties of the tape, and suggests that taping of bloodstained skin should be carried out more than once to maximise the recovery of fibres. 30 2.4.6 Fibre persistence in head hair The transfer and persistence of fibres in human head hair is of particular interest because violent crimes are often carried out by a perpetrator wearing some form of mask, such as a balaclava. 21 Although traditionally any recovered masks have been searched for hairs and saliva, another approach is to search the hair of the suspect for fibres originating from the mask. 32 Ashcroft et al. 21 found that the longer a (wool or acrylic) mask was in contact with the hair, and the longer the hair of the person, the higher the number of transferred fibres. Fibre loss followed an approximately exponential decay to leave a residual 8-10% of the fibres after 24 hours. If the hair was not washed, fibres could persist for up to 6 days. Salter and Cook 32 also transferred wool and acrylic fibres to head hair, and found that on average, fibre persistence in hair was greater than for garments. In a secondary transfer study, the head hair of volunteers was seeded with wool and acrylic fibres, and the persistence of the target fibres on their pillow cases was determined. Although the persistence did not follow a classic exponential decay, the results showed that where a suspect is apprehended outside of the fibre persistence window ( 6 days), there is value in seizing and taping the suspect s pillow case. 28 This window of persistence can be very 12
useful in estimating a time frame for the wear of an item such as a balaclava, particularly if the suspect concedes ownership of the item but denies contact with it around the time in question. 13
3 Experimental 3.1 Transfer and persistence experimental design Subjects were asked to attend wearing clothing that exposed their upper and lower arms. These areas of skin were taped to remove any background fibres (known as blanking) and the tapes retained. The skin was then seeded with target fibres by contact between the skin and a target garment. The contact involved wrapping the sleeves of the garment around the arms of the subject (see fig. 3) and moving the sleeves along the length of each arm and over the hands. The contact was standardised as far as possible for all subjects, and was intended to represent a prolonged and forceful contact. The subject s skin was taped immediately (fig. 4) and the number of target fibres counted to establish an initial (t=0) value. Subjects were then seeded again and asked to return after a certain time interval (from 0.5 to 5 hours). Subjects were asked not to cover their skin during this time interval, as it has been shown previously that the presence of an overgarment results in a more rapid loss of fibres. 25 After the designated time period had elapsed, the subjects were taped again and the number of remaining target fibres counted. A note was made of the colour and fibre content of the upper garments worn by each subject on each occasion, and a tapelift of the garment taken. Experiments were approved by the University of Strathclyde Ethics Committee. 3.1.1 Target fibres Two target garments were used for the transfer and persistence experiments. A bright blue knitted hooded top (80% cotton, 20% polyester) shown in fig. 5, and a bright pink knitted jumper (100% wool) shown in fig. 6. These were chosen from a selection of garments used for training exercises at the FSS. Although both garments were knitted, the 14
Figure 3: A photograph of the contact between a subject and one of the target garments. Figure 4: A photograph of the taping of the arms of a subject. 15
inside of the hooded top had a brushed texture [fig. 7(b)] and this surface was used for transfer experiments due to its preferential shedding properties and high transfer potential. Brightly coloured garments were selected for ease of identifying and counting target fibres. Figure 5: The blue 20% polyester and 80% cotton hooded top. Scale bar = 10 cm. Bulk samples of fibres from each garment were mounted [fig. 8(a) and fig. 8(b)] and examined by high power microscopy to confirm the fibre types listed on the garment labels. In fig. 8(b) the polyester fibres are bright blue and delustered with a round cross section, and the cotton fibres are pale blue and convoluted. The bulk sample fibres were also examined by microspectrophotometry (J & M MSP 400 upgraded for ultra-violet use down to 240 nm) used in the range 320-710 nm (spectra are shown in fig. 11). Several MSP measurements were made to encompass the entire range of shades caused by variation in dye uptake (10 spectra for cotton and wool fibres 16
Figure 6: The pink 100% wool jumper. Scale bar = 10 cm. and 5 for polyester fibres). 3.1.2 Taping Fibre recovery was achieved by pressing high adhesive tape (J-LAR, 72 N/25 mm, 2 cm wide 33 ) onto the skin of the subjects (fig. 4). The pieces of adhesive tape known as tapings were then attached to clear A5 acetate sheets which had been appropriately labelled. 28 The taping procedure was not reported to be uncomfortable by any of the subjects, in agreement with previous studies, 15,32 although some arm hairs were removed, particularly from male subjects. Cells from the stratum corneum the outermost layer of skin, were also removed by the taping procedure. 34 In this study zonal taping (as opposed to 1:1 taping) was used, with the following zones: right upper arm, right lower arm, right hand, left upper arm, left lower arm and 17
(a) The pink wool jumper. (b) The cotton/polyester blue hooded top. Figure 7: Micrographs showing the knitted and brushed surfaces of the two target garments. Scale bars = 2 mm. (a) The pink wool fibres. (b) The blue cotton and polyester fibres. Figure 8: Micrographs of the bulk fibre samples at x10 magnification. In (b) the bright blue fibres are polyester and pale blue fibres are cotton. Scale bars = 200 µm 18
left hand. Larger zones could not be used, as the adhesive properties of the tape quickly became exhausted, 35 and the resulting efficiency of the retrieval reduced. The technique of 1:1 taping (where strips of adhesive tape are used only once) was not used because it is time-consuming and produces a very large number of tapings to search. 36 1:1 taping can provide detailed information on the distribution of fibres on a surface (known as fibre mapping), but only where the recipient surface remains undisturbed after transfer. In this experiment redistribution of fibres (see section 2.4.3) by subjects touching their skin, was thought to be highly likely. Tapings were examined using a search microscope (Nikon SMZ645) and target fibres were recorded by drawing circles on the acetate sheets using a permanent marker. 3.1.3 Length The lengths of target fibres at t=0 and t=5 hours were estimated by comparison with a mm scale under the search microscope. 17 3.1.4 Differential shedding The respective numbers of cotton and polyester fibres for the blue hooded top were recorded at t=0, t=2 hours and t=5 hours. 3.1.5 24 and 48 hour persistence experiments In an extension of the persistence study, subjects were seeded with target fibres and asked to return in 24 or 48 hours. They were instructed to carry out their normal activities, which should include a bath or shower (subjects were provided with a white cotton towel for use after bathing to speed up searching and minimise spurious blue cotton matches). Anti-contamination measures were taken including hand washing, wearing gloves and 19
conducting the tapings in a room separate to the seeding room. Fibres from the tapings which were superficially similar to the two targets were removed from tapes by cutting through the back of the tape using a scalpel, and were mounted in Entellan, 12 fibres to a slide under individual 10 mm glass cover slips. Each fibre was numbered, 17 and then compared to the bulk target sample using low power microscopy, comparison white light and fluorescence microscopy (section 3.3.1) and MSP (section 3.3.2). Failure in any one of the three stages eliminated the fibre from any further comparison. 3.2 Population study experimental design The blanking process at the start of the transfer and persistence experiments removed any background fibres present on subject s arms. These tapings were retained, transferred to labelled acetates sheets, and used to determine the population of coloured fibres on bare skin. The fibres were classified according to perceived colour, generic class, length and delustrant status (see Table 1). Type Colour Cotton, wool, acrylic, polyamide, polyester, regenerated cellulosic (including acetates and viscose/rayon), other Black/grey, blue, red, orange/brown, purple/pink, green, other Length (mm) 0.5, 0.6-1.0, 1.1-3.0, 3.1-5.0, >5.0 Delustrant status Present, absent (for man-made fibres only) Table 1: Summary of the categories used for the classification of fibres on bare skin. 20
Figure 9: An example of a set of background tapings showing fibres marked by perceived colour. Scale bar = 2 cm. 3.2.1 Perceived colour The perceived colour of each fibre under the search microscope was recorded by circling the acetate sheet with different colour permanent markers (see fig. 9). The classification chosen for perceived colour was in agreement with previous population studies. 15 All fibres were assigned a colour subjectively. An objective method for classifying colour would be MSP (see section 3.3.2), but this analysis is too detailed and time-consuming for general information. 6 Only coloured fibres were considered in this study as colour is one of the most important parameters in the comparison of textile fibres, and these fibres tend to be given preference in forensic examinations. 17 21
3.2.2 Generic class Given the large number of fibres recovered in this study, generic class identification was carried out by randomly sampling two 1 cm 2 windows on each taping using a specially made stencil. 6,15 Fibres both wholly within the window and partially within the window were mounted as before. 9 The perceived colour of the sub-sample of fibres was recorded and compared to the overall colour population to ensure a representative sample. Fibres were classified by polarised light microscopy (section 3.3.3) using a Leitz Ortholux II POL-BK polarising microscope equipped with a 20 order tilting compensator. Mean fibre diameters were determined using a calibrated eyepiece graticule. 3.2.3 Length The lengths of background fibres were estimated by comparison with a mm scale under the search microscope. 3.2.4 Delustrant Particles of delustrant (sub-micrometre crystals of TiO 2 ) are often added to man-made fibres to reduce their brightness or lustre. 1 The delustrant status of fibres encountered in this population study was limited to just two categories (absent or present). It was not possible to be any more specific, as there is currently no reliable way to quantify or grade the delustrant content of individual fibres. 6,37 22
3.3 Analytical techniques 3.3.1 Comparison microscopy In comparison microscopy two samples are compared side-by-side on different optical stages, which are joined by an optical bridge with a single ocular head. 1 Comparison microscopy can be used to compare the appearence of fibres under white light [fig. 10(a)] ultra-violet light [fig. 10(b)], and blue light [fig. 10(c)]. In a typical experiment a single recovered fibre is mounted and compared to the bulk target fibre sample at x10, x20 and x40 magnifications. If the two fibres cannot be discriminated when viewed under white light, the other lighting conditions are employed. (a) White light. (b) Ultra-violet light (340-380 nm). (c) Blue light (450-490 nm). Figure 10: Comparison micrographs of a recovered pink wool fibre (left) and one of the target pink wool fibres (right) under different lighting conditions. 3.3.2 Microspectrophotometry The technique of microspectrophotometry is a combination of microscopy and absorption spectroscopy. It allows the absorption spectrum of a fibre fragment viewed through a microscope to be obtained. It is applicable to both natural and synthetic fibres, as the absorption represents the dye rather than the fibre. 23
(a) MSP spectra of the pink wool target fibres and a recovered pink wool fibre (dotted pink line). In this case the spectrum of the recovered fibre falls outside the range of control fibre spectra and has a different lineshape, so the fibre is judged not to be a match. (b) MSP spectra of the blue cotton target fibres and a recovered blue cotton fibre (dotted pink line). In this case the spectrum of the recovered fibre falls within the range of control fibre spectra and has a similar lineshape, so the fibre is judged to be a match. Figure 11: Microspectrophotometry spectra for both pink wool (a) and blue cotton (b) target fibres. 24
MSP is generally used in the latter stage of fibre examination, after the morphological, colour and fluorescence characteristics of two fibres have been shown to match under appropriate conditions. A wavelength range encompassing both ultraviolet and visible light is employed, with the former being used to discriminate between very dark blue and black coloured fibres. The ultra-violet range below 320 nm was not used in this work because this requires the re-mounting of fibres onto quartz slides with quartz cover slips. In a typical MSP experiment a single recovered fibre is mounted and focussed under the x40 objective of an optical microscope, which is attached to a spectrometer via an optical path. A small area of the background near to the fibre is selected using a box known as a diaphragm and a background or reference spectrum is taken. The diaphragm is then moved on to the fibre and a further spectrum taken. The absorbance due to the dye is then calculated. The spectrum of the recovered fibre can be compared to the spectra of the control fibres (section 2.2) by plotting absorbance vs wavelength for both on the same scale. Features of the spectra such as maxima, minima, peak height/width ratio, shoulders and plateaux are then compared (fig. 11). If the spectrum of the recovered fibre falls outside the control range [as in fig. 11(a)] it is excluded. If the spectrum of the recovered fibre falls within the control range it is said to be a match [as in fig. 11(b)]. 3.3.3 Polarized Light Microscopy Polarized Light Microscopy (PLM) is used to distinguish between synthetic fibres that have the same visual appearance but have different polymer compositions. It can also be used in the examination of natural fibres (although this is more difficult because of the variation in thickness of natural fibres). PLM was used only for synthetic fibres in this project and the fibres were examined without dye stripping. PLM operates on the following basis: plane polarised light passing through a synthetic 25
fibre will interact with the fibre polymer, with the interaction being different depending on the chemistry of the polymer. If the fibre is at an angle to the plane of polarisation, patterns of colours may be observed, which are known as birefringence colours [fig. 12(a)]. The colours are produced because the fibre has two different refractive indicies, one along its length (n ) and one at right angles to its length (n ). When plane polarised light passes through the fibre it is slowed down (or retarded) differently along these two axes. When these slowed components of the light are recombined, the difference in phase between them is known as the retardation (R) and this can be measured using PLM. The value of the birefringence (Γ), which is used to identify the polymer type, is calculated using the following equation: Γ = R 1000T where T is the thickness of the fibre in µm. In a typical PLM experiment an individual fibre is mounted and focussed under transmitted light using the x40 objective. The fibre is then viewed under crossed polars (placed between two polarising filters) and any birefringence colours observed. A tilting compensator is then placed between the fibre and the second polarising filter. This is a device containing an optical plate which can be tilted out of the microscope optical path. 38 The compensator is tilted until the birefringence colours are extinguished and the fibre appears black [fig. 12(b)]. The angle is noted and compared to a reference table to obtain a value for the retardation. The ratio of the retardation and the average fibre diameter (in nm) is the birefringence. For acetate and acrylic fibres which appear grey under crossed polars [fig. 12(c)] a different type of compensator is used. When a 1λ compensator is placed between the fibre and the second polarising filter the background shifts from grey to purple and acetate fibres aligned perpendicular to the compensator will appear red, whereas those parallel to the compensator will appear blue [fig. 12(d)]. Acrylic fibres produce the opposite effect. (2) 26
(a) (b) (c) (d) Figure 12: (a) A ramie fibre under crossed polars showing birefringence colours, (b) extinction of birefringence colours using a tilting compensator, (c) acetate fibres under crossed polars and (d) acetate fibres under crossed polars with a 1λ compensator, reproduced from Olympus. 38 27
4 Results and discussion 4.1 Transfer and persistence 4.1.1 Transfer experiments The average numbers of target fibres initially transferred to participants for both garments are shown in Table 2. Twenty three initial transfer experiments were performed for the blue hooded top, and the average number of blue fibres initially transferred was 245 ± 142 (range = 106-730). Twenty one initial transfer experiments were performed for the pink wool jumper, and the average number of pink fibres initially transferred was 133 ± 50 (range = 48-214). The difference between the initial transfer values for blue and pink garments was significant at the 5% level (see appendix). As the same type of contact was used in both sets of experiments, these results can be attributed to the fact that the wool garment was less sheddy than the polyester/cotton garment. The lower sheddability of wool compared to cotton was also noted by Roux et al. 3 The brittle nature and convoluted structure of cotton is thought to be responsible for its propensity to be transferred during contact. 1 The higher numbers of transferred blue fibres may also be due to the fact that the blue hooded top was new and unwashed, 29 whereas the pink wool jumper was a second-hand donation. The structure of the knit of each fabric will also have influenced the propensity of the fibres to shed from each garment, with fibres being more tightly held in the knitted pink jumper than the brushed blue hooded top [fig. 7(a) and fig. 7(b)]. The smaller number of transferred wool fibres, compared to cotton fibres is in agreement with the transfer of fibres to human head hair. 28 There was also variation within transfer experiments using the same garment by gender, with fewer blue fibres being transferred to women. The difference in the number of pink fibres transferred to men and women was not significant at the 5% level. 28
It is likely that the number of fibres transferred was mainly dependent on two variables; the arm surface area (as discussed in section 2.4) and the density of arm hair (the number of initially transferred fibres was consistently higher for hirsute subjects). Garment Blue hooded top Pink jumper Average 245 ± 142 133 ± 50 Men 312 ± 163 135 ± 57 Women 172 ± 60 131 ± 46 Table 2: Numbers of target fibres initially transferred to the bare skin of male and female subjects. 4.1.2 Persistence experiments Persistence experiments were carried out for intervals of 0.5, 1, 2, 3, 4 and 5 hours for each garment. Forty eight persistence experiments were performed for the blue hooded top and 37 persistence experiments were performed for the pink wool jumper. The decay curves for both garments are shown in fig. 13. Both decay curves show one standard deviation limits for each time interval. The standard deviations are high for both garments, which is a result of several factors. Firstly, the high standard deviations reflect the variation in persistence between individuals found in this type of study. Variations in persistence are due to individual differences in skin surface texture (with more hirsute subjects being more retentive) and also variation in subject activity following the seeding. 32 Secondly, the high standard deviations reflect the inherent variability of initial primary transfer. 28 Although the simulated contact was standardised in all transfer experiments, it is impossible to produce exact duplicates 29
Figure 13: Decay curves for fibres on bare skin for (a) the pink wool jumper and (b) the blue hooded top. The y-axis shows the number of fibres remaining as percentage of the initial value. Both decay curves show one standard deviation limits for each time interval. At t=0 the percentage of fibres remaining is 100% with zero standard deviation. 30
in this type of study because of natural variations in the force applied during contact with the garment. It is also possible that the propensity of the garments to shed decreased with repeated use. The standard deviations are higher at the earlier time intervals because they reflect the larger experimental error when the rate of fibre loss is greatest. 20 When the rate of loss of fibre is high, small timing errors are magnified. The overall shape of the decay curves in fig. 13 is approximately exponential after an initial rapid loss. The difference in persistence between male and female subjects was not significant at the 5% level for either garment. After 5 hours approximately 15% of the blue cotton and polyester fibres remained. The rate of loss of the the pink wool fibres was higher however, with only 5% of fibres remaining after 5 hours. The difference in rate of loss between blue and pink garments was significant at the 5% level. In most of the previous persistence studies no difference between the persistence of wool and other fibre types was reported, 20,21,25,32 but the greater persistence of cotton fibres over wool fibres is in agreement with results obtained for human head hair. 28 It has been suggested that woollen fibres might persist in human hair for longer than other fibre types, because of hair-to-hair interactions between both rough scaled surfaces. 21 This does not appear to be the case for human skin. During some persistence experiments, subjects were unexpectedly called into the laboratory and were required to wear a lab coat (white 35% cotton and 65% polyester). When subjects were subsequently taped, only 20% of the expected percentage of fibres for that time interval remained. This is in agreement with the findings for the persistence of fibres on garments, where the presence of an overgarment resulted in a more rapid loss of fibres. 25 Overall, these persistence results are within the range reported for non-smooth garments (see section 2.4.2). 20,35 Therefore the treatment of skin as a smooth garment (sec- 31
tion 2.4.5) can be considered to be too simplistic. Human skin also appears to differ from human hair, which retains fibres for longer than garments. 32 4.1.3 24 and 48 hour persistence experiments For the blue hooded top, 24 hour persistence experiments were undertaken for 10 subjects and 48 hour persistence experiments for 7 subjects. The number of apparent blue matches after 24 and 48 hours was initially high ( 10 per subject with the majority being blue cotton), but these apparent matches were eliminated by fluorescence microscopy. 18 No matching fibres were recovered after 24 or 48 hours. For the pink jumper, 24 hour persistence experiments were undertaken for 8 subjects and no matching fibres were found. In contrast to the cotton/polyester garment, only one apparent pink match was found across the entire study and this was also eliminated by fluorescence microscopy. This difference in the number of apparent matches reflects the difference in frequency of the two fibre types on skin (see section 4.2.3). Vigorous activity has been shown to modify the persistence of fibres, 32 therefore the activity undertaken during the 24 or 48 hours was recorded for each subject. Although subject activity was found to vary from sports to sedentary activities, zero matching fibres were recovered after 24 or 48 hours, therefore any correlation with activity could not be inferred from the results. The lack of matching fibres after 24 or 48 hours is mainly due to the fact that subjects showered or bathed during this time, as washing has previously been shown to remove the vast majority of fibres. 32 21 32
4.1.4 Differential shedding The label inside the blue hooded top stated that the garment was composed of 80% cotton and 20% polyester. The generic class of approximately half of the initially transferred blue fibres (2801), and all of the fibres recovered at t=2 (268) and t=5 hrs (176) was determined. The cotton:polyester ratios were found to be 69:31, 72:28 and 75:25 respectively. The results showed a modest and increasing bias towards cotton being retained at each stage. The greater loss of polyester fibres may be due to the smooth nature of polyester, 25 with the convoluted structure of cotton producing a more tenacious contact between fibre and skin and possibly between fibre and arm hair. 4.1.5 Length The lengths of the 4038 blue target fibres and the 2470 pink target fibres initially transferred to subjects were estimated and are shown as histograms in fig. 14 and fig. 15. For both garments the size grouping with the highest frequency initially was 1.1-2.0 mm, but the distributions were different with only a small proportion of the pink target fibres having length > 3.0 mm. The histograms also show the lengths of the 343 blue target fibres and the 46 pink target fibres recovered after 5 hours. For the blue garment after 5 hours, the size grouping with the highest frequency was still 1.1-2.0 mm, but the distribution had changed such that shorter fibres were more highly represented than longer fibres. The same overall trend was recorded for the pink wool garment, but with a more dramatic shift towards shorter lengths after 5 hours, and the 0.5 mm size group having the highest frequency after 5 hours. No pink wool fibres with length > 6 mm were recorded after 5 hours. The frequency of fibres in the final size grouping of > 10 mm is misleading in both histograms as this group size is much larger than the others. Some disagreement exists in the literature over the effect of fibre length on persistence, 33
Figure 14: The length distribution of the blue cotton and polyester target fibres on bare skin initially and after 5 hours. The histogram shows that the size group with the highest frequency initially is 1.1-2.0 mm. After 5 hours this is still the size grouping with the highest frequency but the distribution around it has changed to favour shorter fibres. 34
Figure 15: The length distribution of the pink wool target fibres on bare skin initially and after 5 hours. The histogram shows that the size group with the highest frequency initially is 1.1-2.0 mm. After 5 hours the majority of fibres fall into the 1.0 mm group and the frequency distribution has shifted dramatically towards shorter lengths. 35
with one study claiming that size distribution remains constant, 20 and another reporting a confused picture with no clear trends. 32 The results above are in agreement with the studies by Krauß 22 and Robertson 25 which concluded that longer fibres are lost more quickly. This is thought to be because long fibres are more likley to suffer disturbance than short fibres. 2 36
4.2 Background population study Any background fibres present on the bare skin of subjects were counted and classified according to Table 1. The average numbers of coloured fibres recovered from the bare skin of subjects was 241 ± 143 fibres for women, and 458 ± 246 fibres for men. This difference was significant at the 5% level. As noted for the transfer of target fibres (section 4.1.1) higher numbers of background fibres were found on both hirsute subjects and those with a larger arm surface area. 4.2.1 Perceived colour 12,399 fibres taken from 21 subjects were classified according to perceived colour under reflected light (fig. 16). Black/grey fibres made up almost three-quarters of the fibres classified (72.3%) followed by blue (12.4%), red (6.3%) brown (4.4%) and green (3.8%). The remaining colour groups made up less than 1% of the total, and consisted of pink/purple (0.8%) and other (including yellow, tiger tail and multi-coloured fibres) making up only 0.03%. The very high percentage of black fibres recorded probably reflects the subjective nature of the classification method, as in reality many of these fibres would have appeared very dark green or blue under transmitted light. Direct comparison with some previous population studies (Table 3) shows that the dominance of black fibres, followed by blue and red was also observed for fibres in head hair, 15 on car seats 6 and in washing machines. 4 The four remaining minor colours do not always agree between published studies, probably because they will be heavily influenced by climate, season, 15 and fashion. Direct comparison is not possible with all previously published population studies because the full colour results are not always reported. 17 For population studies using cinema seats 16 and white t-shirts, 39 only the two most popular colours, (black and blue) were given, although this ranking agrees with the 37
Figure 16: Percentage distribution of fibres on bare skin according to colour (n = 12399). Location Head hair 15 Car seats 6 Washing machines 4 Bare skin Bare skin corrected Black/grey 48 40 42 73 50 Blue 29 33 28 12 29 Red 13 12 19 6 6 Orange/brown 3 4 7 4 8 Yellow 0 2 1 0 0 Pink/purple 5 4 1 1 2 Green 2 4 2 4 5 Table 3: Comparison of the percentages colours of fibres obtained from published population studies. indicates that the values have been corrected for the contribution of colourless fibres. 38
work described here. In addition, direct comparisons cannot be made where population studies include colourless fibres. 9 Although not outside of the range previously reported, the number of green fibres reported in this study may have been inflated by one set of background tapings which contained 145 green fibres out of 282 (51%). 4.2.2 Length Of the 12,399 fibres classified according to perceived colour, 491 fibres were randomly selected (see section 3.2.2) and classified according to length (fig. 17). The size grouping with the highest frequency in this sub-sample was 1.1-3.0 mm (with a frequency of 43%), followed by the size grouping 0.5-1.0 mm (27%). A quarter of the fibres had a length 0.5 mm, 4% of the fibres fell into the 3.1-5.0 mm size grouping, and only 1% of fibres had length > 5.0 mm. It is important to note that the size groupings are unequal in fig. 17. Although the absolute percentages vary between studies, comparison with previous population studies (Table 4) on human head hair, 15 car seats 6 and washing machines 4 shows agreement with the length rankings reported here (although fibres with length < 0.5 mm were not included in the washing machine study). The length results do not agree with the population of fibres on outdoor surfaces which showed higher percentages of shorter fibres. 17 The authors of that study suggested that the fibres had become short as a result of damage over some time, and were not necessarily representative of recent transfer. 39
Figure 17: Percentage distribution of fibres on bare skin according to length (n = 491) showing that the size grouping with the highest frequency was 1.1-3.0 mm. Location Head hair 15 Outdoor surfaces 17 Car seats 6 Washing machines 4 Bare skin 0.5 24 35 18-25 0.6-1 33 30 25 29 27 1.1-3 37 28 46 53 43 3.1-5 4 4 8 11 4 > 5.0 2 3 4 8 1 Table 4: Comparison of the length distributions obtained from published population studies showing that a group size of 1.1-3.0 mm has also been reported to have the highest frequency for other surfaces. 40
4.2.3 Fibre generic class The sub-sample of 491 fibres was classified according to generic fibre type. Analysis of the data showed that 79% of the fibres were natural and 21% were man-made. This is in agreement with previous population studies (Table 5) which have all reported a higher percentage of natural fibres. Within each of the natural and man-made classes the fibres were further sub-divided. The largest group of natural fibres was cotton (72%) followed by wool (8%). Polyester accounted for the highest proportion of man-made fibres (15%) followed by regenerated cellulosic and acrylic fibres (2%). Most of the other group (2%) consisted of man-made fibres which were too darkly dyed or pigmented to be identified at this level of analysis. Further analysis such as infra-red spectroscopy would be used in a casework scenario to identify the polymer type. Comparison with the previously published population studies (Table 6) shows that cotton fibres have consistently been found to be the most abundant fibre type. This is perhaps unsurprising given that a clothing database compiled in Germany showed that 74% of summer clothing in the database was composed of cotton fibres. 40 Polyester was found to be the second most common fibre type in this study, which is also in agreement with the clothing database which found 15% of summer clothing to be composed of polyester fibres. 40 The proportion of polyester fibres also agrees with the population of fibres in head hair, 15 but not with any other published poulation study. Discrepancies in the percentages and the order among the less-common fibre types can generally be accounted for by climatic differences in the region of study and the difference in sample sizes between the studies. The sub-sample of 491 fibres was also classified according to both colour and generic class (fig. 19). Cotton dominates the population with the two most prevalent combinations of black/grey cotton and blue cotton accounting for over half (56%). 41
Figure 18: Percentage distribution of fibres on bare skin according to generic class (n= 491) showing that the majority of fibres are cotton. The fibre population studies of head hair, 15 cinema seats, 16 washing machines, 4 car seats 6 and white t-shirts 39 all support the finding that black/grey cotton is the most popular colour/class combination followed by blue cotton. This study does not agree with a population study of Polish bus seats which found both blue cotton and green cotton to be more popular than black/grey cotton. 9 The third most common grouping in this study was black/grey polyester, a position not supported by previous studies which have reported either red cotton 4,15 or black/grey wool 6,16 in this position. The next most common fibre class/colour combinations were blue wool (6%) and red cotton (5%) followed by green cotton, blue polyester and brown cotton (4%), brown polyester (3%) and black wool (2%). The remaining 16 groups had populations of 1% or smaller and in total accounted for less than 10% of the population. The grouping blue cotton includes both denim and other blue cotton fibres. 6 42