t.

15.5 ”rim...
rung’ Egg. 9.

 

“mm
:J.Mn...§i
1 LL]. firs...-

.. fJflK
his]: 3...... .
. 1.9.2. i.
83 33....

i
:11092.‘
, .2}. It

 

3 \ $510.11.. Lang...“
. 1:. J 2 (3:15? inap'uwl
1...; 3.5.5. , . , . ”to: p T
13% E... .. x... . , . . .
[21... . 2A. v 13.}: ‘
v ,2]. . ,

I .. . 5.1.15.

 

7“” LIBRARY
Michigan State
University

This is to certify that the
thesis entitled

UV MICROSPECTROPHOTOMETRY OF FIBERS FROM
APPARENTLY WHITE TEXTILES ENCOUNTERED IN
FORENSIC CASEWORK

presented by

ERIN CUNNANE FARR

has been accepted towards fulfillment
of the requirements for the

Cn‘minal Justice
Master of Science degree in with a specialization in
Forensic Science

 

 

 

Major P essor's Signature

2a Quiz? N303

MSU Is an Affirmative Action/Equal Opportunity Institution

 

 

 

PLACE lN RETURN Box to remove this checkout from your record.
1'0 AVOID FINES return on or before date due.
MAY BE RECAU£D with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/CIRC/DateDue.indd—p.15

 

UV MICROSPECTROPHOTOMETRY OF FIBERS FROM APPARENTLY WHITE
TEXTILES ENCOUNTERED IN FORENSIC CASEWORK

By

Erin Cunnane Farr

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Criminal Justice

2005

ABSTRACT

UV MICROSPECTROPHOTOMETRY OF FIBERS FROM APPARENTLY WHITE
TEXTILES EOUNTERED IN FORENSIC CASEWORK

By
Erin Cunnane Farr

The purpose of this study is to determine whether similar looking undyed fibers
can be differentiated by microspectrophotometry. Several classes of undyed fibers will
be analyzed independently, and their UV transmittance properties will be examined.

Undyed fibers are very commonly encountered by forensic fiber examiners.
Identification of the undyed fiber is fairly routine, but the question becomes how to
differentiate between fibers within a class, such as polyester or cotton, for example. One
possibility is that optical brighteners may play an important role in distinguishing fibers,
similar to the way textile dyes are used to differentiate dyed fibers. Textile dyes can be
differentiated through various analytical techniques, and a UV-visible
microspectrophotometer is but one example. A mi crospectrophotometer allows the
forensic scientist to measure the transmission, reflectance, or fluorescence characteristics
of fibers.

Optical brighteners are often used on fibers during the manufacturing process or
acquired through commercial detergents. Perhaps these optical brighteners will cause
similar looking fibers, originating from different sources, to behave differently when

exposed to electromagnetic radiation.

To my father, Jim Cunnane,
who has been constant support throughout my education,
and who taught me never to forget what the solution is:
the value of the variable that makes the equation true.
Here is the solution to my master’s degree.
Thank you for all of your help along the way.

iii

ACKNOWLEDGMENTS

Throughout this project I have received assistance and direction in various forms
from many individuals. A special thanks to Larry Peterson and Tammy Jergovich, as
well as other members of the Georgia Bureau of Identification Trace Evidence Section. I
would also like to thank my advisor at Michigan Sate University, Dr. Jay Siegel, who
accepted me into his program and helped me fulfill my dream of becoming a forensic
scientist. A generous thank you to my parents who helped me get where I am today, with
their continuous support and encouragement—it meant the world to me. To my husband,
Matt, who believed in me in the midst of my frequent procrastination, and to all of my
fiiends who have encouraged me throughout the progression of this project, I am very

gratefiil!

iv

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................ .vi
LIST OF FIGURES ............................................................................................... vii
INTRODUCTION ................................................................................................... 1
Identification of Fibers ................................................................................. 1
Comparison of Fibers ................................................................................... 2
The Importance of Color .............................................................................. 3
Microspectrophotometry .............................................................................. 4
Undyed Fibers ............................................................................................. 5
REVIEW OF LITERATURE .................................................................................. 7
MATERIALS AND METHODS ............................................................................. 9
Sample Preparation ...................................................................................... 9
Microscopes and Instrumentation ............................................................... 10
Instrument Calibration ............................................................................... 12
Data Collection .......................................................................................... 14
RESULTS AND DISCUSSION ............................................................................ 15
Polyester .................................................................................................... 15
Cotton ........................................................................................................ 23
Rayon ........................................................................................................ 28
Nylon ......................................................................................................... 3 1
Acrylic ....................................................................................................... 35
CONCLUSIONS AND RECOMMENDATIONS ................................................. 39
APPENDD( A: OPERATION INSTRUCTIONS FOR THE S.E.E. 2000 ................ 44
APPENDD( B: RAW DATA ..................................................................................... 46
APPENDIX C: MEAN S AND STANDARD DEVIATIONS ................................. 102
REFERENCES ......................................................................................................... 157
BIBLIOGRAPHY .................................................................................................... 158

LIST OF TABLES

TABLE 1: UNDYED FIBER SAMPLES ......................................................................... 11
TABLE 2: QUANTITATIVE ANALYSIS OF POLYESTER FIBERS (REGION 2) ..... 18
TABLE 3: QUANTITATIVE ANALYSIS OF POLYESTER FIBERS (REGION 3) ..... 22
TABLE 4: QUANTITATIVE ANALYSIS OF COTTON FIBERS (REGION 2) ........... 27
TABLE 5: QUANTITATIVE ANALYSIS OF RAYON FIBERS (REGION 2) ............. 31
TALBE 6: QUANTITATIVE ANALYSIS OF NYLON FIBERS (REGION 2) ............. 34

TABLE 7: QUANTITATIVE ANALYSIS OF ACRYLIC FIBERS (REGION 2) ......... 37

LIST OF FIGURES

FIGURE 1: HOLMIUM OXIDE FILTER #D-l 11 CALIBRATION .............................. 13

vii

INTRODUCTION

The importance of forensic fiber analysis relies heavily upon an examiner’s
ability to establish associations between fibers. Fibers are submitted to the laboratory in
the form Of both questioned and known. Questioned fibers are collected by an
investigator, and they are typically recovered from, but certainly not limited to, either a
victim of a crime, a crime scene, or even an object such as a weapon. Known fibers are
typically collected fi'om a person who is believed to be involved in a crime, or from that
person’s textile environment. The forensic scientist will examine the questioned fibers
and the known fibers to see if they can establish any associations between these samples.
The general approach to the examination will consist of both identification and

comparison of the questioned and known fiber samfles.

Identification of Fibers

Identification of fibers in a forensic examination is a twofold process. First, the
examiner must ascertain the type of fiber with which they are dealing. Second, they must
identify as many significant characteristics as possible, such as the color, morphology,
and diameter of the fiber, along with surface detail, amount of delustrant, and any other
distinguishing characteristics. The examiner will generally start with a microscopic
examination, looking for physical features that are characteristic of a particular fiber type.
In addition, it is often helpfiil to compare the questioned fibers to standards from a
reference collection. Generally, natural fibers will show more identifying features
through bright field microscopy than will manufactured fibers. This is because natural

fibers are highly variable whereas manufactured fibers are much more uniform. Thus,

polarized-light microscopy is used to reveal optical properties (i.e. refi'active index, sign
of elongation, birefringence) that are useful for the identification of manufactured

fibers. 1’2 These optical properties are sufficient for the identification of several synthetic
fiber types, whereas other types may need further analysis. If need be, the examiner has
many additional techniques (such as infrared spectroscopy, cross-section analysis, hot-
stage microscopy, pyrolysis gas chromatography, and solubility tests) available for the

purpose of aiding in the identification of the fiber.

Comm—son of Fibers

Comparison of fibers is somewhat more involved than identification, because the
examiner must examine questioned fibers and known fibers in a case until significant
difl‘erences are found, or until all available tests are exhausted. As the number of
associations between the two fiber samples increases, the more likely the possibility that
they originated fi'om the same source. Gaudette says that “if, after conducting several
types of comparative examinations and looking at a large number of comparison
characteristics, no significant differences between the questioned fiber and the known
sample are found, they are said to be consistent with having had a common origin.”3
Several techniques are available to the fiber examiner for comparison purposes. The
fiber type (or identification of the fiber) is the first comparison feature between
questioned and known samples. If the fibers are determined to be of the same type, other
microscopical characteristics should be compared with a comparison microscope. As
mentioned above, microscopic differences can be found in various forms such as color,
morphology, and diameter of the fiber, along with surface detail and amount of

delustrant.1 These differences are important because they can lead to firrther

differentiation within one specific class of fibers, allowing an examiner to say, for
example, that one polyester fiber is different from another polyester fiber. Additional
tests such as FTIR microscopy, UV-visible microspectrophotometry, dye analysis by
thin-layer chromatography, and tests that reveal optical and fluorescent properties of the
fibers are available to the examiner. It is important to remember that not all of these
techniques need to be used in every case. The examiner must consider the
instrumentation available combined with the discriminating power of the techniques

used, to try to obtain as much comparative information as possible.

The Immrtance of Color

The color of the fiber is believed to be a very important and distinguishing
characteristic of a fiber.4 There is a vast array of colors that are produced by textile dye
manufacturers. After a fiber is dyed with a unique dye formula, it has a special chemical
feature that becomes a foundation for distinguishing it fi'om many other fibers within its
same class.

Because of its sensitivity, the human eye is very good at evaluating color
difi‘erences during the initial stages of a fiber examination. It has the ability to detect
very slight differences in shades of the same color from sample to sample. Thus, the eye
becomes a useful tool for both visually screening samples and for the side by side
examinations of fibers under a comparison microscope. However, during later stages Of a
fiber examination the human eye is not as reliable due to subjectivity of the examiner;
each influenced by personal opinion and experience."”’6 Furthermore, many textile dyes
have colors that appear visually indistinguishable, but are made up of entirely different

chemical formulas. This causes some hues of textile dyes to appear to be identical under

one set of lighting conditions, but very different under other conditions. For example, an
examiner may find that fibers from two different garments appear to be similar in color
when viewed through a comparison microscope; however, the UV transmittance data
from the same two fibers will Show different characteristics. This phenomenon is known
as metamerism.6 Spectroscopic analysis is a quick and easy way to prevent metamerism
from going undetected.

Textile dyes are complex mixtures—their chemical compositions become
important when the colors are indistinguishable with the naked eye. A
microspectrophotometer is an instrument that has the capability of detecting these
differences in chemical composition with a degree of Objectivity; thereby eliminating the

subjectivity of the human eye and identifying any possibility of metamerism.

Microspectrophotometg
The method of microspectrophotometry (MSP) has been used for more than

seventy years, and has been used in forensic applications for more than forty.7 MSP has
been used for a variety of microscopic samples (such as biological samples, crystals and
minerals, inks, microchips, paints, fibers and explosives) within a wide range of
disciplines (such as biochemical studies, geology, document analysis, the semiconductor
industry, and forensic trace evidence).8

Microspectrophotometry is one method of UV spectroscopy. A
microspectrophotometer is essentially a spectrophotometer adapted with microscopic
capabilities. When a sample is too small to be analyzed through ordinary spectroscopic
analysis, it can usually be analyzed with a microspectrophotometer. As electromagnetic

radiation interacts with a microscopic sample, some of the light is absorbed by the

sample, and the rest is emitted to the spectrophotometer. The instrument measures the
change in light intensities, relative to wavelength.6’7’8 The light intensity will be
measured at each wavelength over the ultra-violet (UV), visible (vis), and/or near infrared
(N IR) regions. The plotted data will reflect the transmission, reflectance, or fluorescent
light properties of these samples. As electromagnetic radiation interacts with the dyed
fiber, excitation of electrons occurs in the UV region, making UV-vis MSP particularly
useful in resolving the phenomena Of metamerism in dyed fibers.7
Microspectrophotometry is convenient for forensic scientists because it is non-
destructive, quick, and requires very little sample preparation. For example, it is possible
to collect data from fiber samples without even removing the samples from the
microscope Slide. The slide is placed on the microscope stage, and that in turn becomes

the sample compartment Of the spectrophotometer.

Undyed Fibers

Many evidentiary fibers are undyed. This limits the usefulness of
microspectrophotometry. Although the scientist can identify the type Of fiber present,
there are limitations to finding significant differences between fibers Of the same class if
no identifying textile dyes are present. Previous studies have shown that fibers treated
with chemicals such as Optical brighteners and bleaching agents will show different
spectral characteristics in the UV range, even if not discriminated in the visible
range.4’9"° The purpose of this research is to simulate real casework by analyzing undyed
fibers Obtained from clothing samples associated with actual cases. The samples will be
analyzed in the UV range with a microspectrophotometer, and the focus will be to see if

the differences in UV spectra within groupings Of fibers (cotton, polyester, nylon, rayon,

and acetate) are statistically different. The Objective of this work is threefold. First, the
UV microspectral response (transmittance in the approximate range of 250nm to 475nm,
where 11m = nanometers) ofundyed fibers will be examined, and the data will be
collected. Second, the data will be categorized by visual analysis as well as statistical
analysis based on the Observed response of the fibers. And third, it will be determined
whether the transmittance data in the range examined reveals enough information for
evaluating similarities/dissimilarities of fibers to be usefirl for forensic purposes. This is
a preliminary, exploratory study, and by no means all-inclusive. It is simply a process of

discovery to see what features are in the data.

REVIEW OF LITERATURE

In the field of forensic fiber analysis, there has been little investigation into the
analysis of ultraviolet spectral characteristics of undyed fibers using a
microspectrophotometer. A few studies, to date, have researched the ultraviolet spectral
characteristics of fibers at different stages of production or at different stages of chemical
bleaching. This study, on the contrary, has many uncontrolled variables such as the
number of times the garment was washed, the detergents used, and the treatment of the
fibers during manufacturing. Thus, this research is an exploratory study of the effects of
chemical treatments (such as detergents or Optical brighteners) on undyed fibers, along
with an attempt to simulate real casework samples.

In one previous study, Martin used an S.E.E.2100 microspectrometer to analyze
the ultraviolet characteristics of wool fibers at several different stages of the chemical
bleaching and/or dying process.9 The purpose was to determine what kind Of spectral
characteristics are imparted by the fiber structure, as well as the dyes. Six wool samples,
each with different chemical treatments, were tested to see what effects the chemical
treatments would have on the absorption spectra. The first sample was untreated wool,
the second, third and fourth samples were treated with different bleach solutions for
various times and/or pHs. The fifth sample was dyed, and the sixth sample was dyed
then bleached. None of the samples showed any meaningful data in the visible range
spectrum, but all of the samples showed notable differences in their UV spectra, allowing
for separation of the samples based on treatment. Thus, Martin concludes that “not only
can the SEE 2100 microspectrometer detect changes in dyes in fibers, it can also detect

changes in the fibers themselves when they are treated and their structure changes.”9

In another previous study, Desrosiers and Martin used an S.E.E.2100
microspectrometer to analyze the ultraviolet characteristics of nylon fibers at different
stages of production. 1° Again, the purpose was to determine what kind of spectral
characteristics are imparted by the fiber structure, as well as the dyes. In this study, four
samples of undyed fibers were treated with different chemicals and analyzed with the
microspectrometer to determine the effects that the chemical treatments have on the
transmission spectra. Some of the samples were difficult to distinguish when looking at
the visible through short-wave NIR regions because they showed little difference in the
peak positions. However, the UV spectra of these same samples showed very different
spectral characteristics, allowing for clear distinction.

Thus, it is clear that analysis of some fibers in the ultraviolet region has some
advantages. The goal of this study is to determine whether these previous findings are
confirmed with undyed fibers when simulating real casework samples; and furthermore,
if it would be advantageous to incorporate UV microspectrophotometry into real

casework samples involving undyed fibers.

MATERIALS AND METHODS

The undyed fibers for this research were taken fiom thirty-two samples of
apparemly white clothing collected fi'om a morgue. All clothing samples were collected
with as little prescreening as possible, in hopes of simulating real casework. Thus, the
samples had several uncontrolled variables such as environmental exposure (i.e. sunlight,
heat, washing, dry cleaning), manufacturing stress, and the composition of the raw
material. In theory, the number Of times a sample has been washed may impart
differences in the UV spectra of samples due to brighteners or bleaching agents on the
fibers coming from detergents. The only prescreening performed on these samples was a
visual exam to ensure that the samples had no obvious stains on them, and to ensure that
the fibers had no significant colorant dyes or surface dyes present. The fibers used for
this study were collected from samples as “white” as white cotton T-shirt fabric, which is
known to be undyed. In addition, these fibers were selected fiom whole fabrics which
would be noticeably discolored even if only a light beige colorant were present. For this
reason, there is a level of confidence that the fibers used in this testing were undyed as
opposed to white dyed fibers. For the purposes of this paper, the word “undyed” will be
used to describe the fibers as defined above.

Sample Preparation

Using tweezers, fibers were teased out of different areas of a sample to ensure that
a representative sample was collected. The fibers from each sample were then mounted
on individual microscope slides. Traditional means of mounting fibers require glass

microscope slides and cover slips and synthetic mounting media such as XAM and

Permount. Unfortunately, these materials will adsorb too much in the UV range,
interfering with data collection. Thus, collection of transmission spectra in the UV range
requires samples to be mounted on quartz microscope slides in glycerin mounting media
under quartz cover slips. Quartz and glycerin adsorb minimally in the UV range,
allowing for more accurate data collection. Samples for this testing were mounted on
quartz slides under quartz cover slips; both purchased from McCrone Accessories and
Components (W estmont, IL). The glycerin mounting media was purchased fi'om Fisher
Scientific (certified grade).

Representative samples were also collected fiom each sample for the purpose Of
measuring the fiber thickness. These samples were mounted in 99.9% mineral Oil on
glass microscope slides under glass cover slips. Microscope slides and cover slips were

purchased from Fisher Scientific, and the mineral oil was purchased from Safeway.

Microscopes and Instrumentation

Fibers from each sample were identified and measured using an Olympus BH-2
microscope. The BH—2 microscope is a comparison compound light microscope with
polarizing light capabilities. Each quartz slide was mounted on the microscope stage, and
the fiber types within each sample were identified. It is important to note that the
identification scheme was limited to a generic classification of fiber types only, thus no
subtypes were addressed. In the following table, the composition of each sample is

recorded along with the sample origin (See Table 1).

10

Table l: Undyed Fiber Samples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

# ORIGIN FIBER COMPOSITION
1 Warm-up pants Polyester, acrylic

2 Sweatshirt Cotton, polyester

3 Sweatshirt Acrylic

4 Sweatshirt Cotton, polyester

5 Sweatshirt Cotton, polyester

6 Panties Nylon

7 Sweatpants Cotton, polyester, acrylic
8 Button-up shirt Cotton, polyester

9 Sweatshirt Cotton, polyester

10 Towel Cotton, polyester

11 T-shirt Cotton, polyester

12 T-shirt Cotton

13 Pull-over blouse Polyester, some cotton
14 Jacket Cotton, polyester

1 5 Longjohns Cotton

16 Socks Cottorgnylon

17 Athletic socks Cotton, nylon

18 Towel Cotton, rayon (high & moderate delustrance)
19 Bed Sheet Cotton

20 Blanket Polyester, nylon

21 Bed sheet Cotton, polyester

22 Pull-over shirt Cotton, polyester

23 Slacks Polyester

24 Blouse Cotton, polyester

25 Blouse Rayon

26 Nightgown Nylon

27 Socks Acrylic, polyester

28 Button-up shirt Cotton, polyester

29 Blouse Polyester

30 Skirt Rayon (heavy & semi-delustrance)
3 1 Nightshirt Polyester

32 Stretch pants acrylic

 

 

The availability of samples set boundaries to this research. Cotton and polyester
fibers were by far the most abundant within the samples, so 20 of each of those fiber
types were analyzed. Nylon, acetate, and rayon fibers were more limited, so only 5 of
each of those types were analyzed. Although these fiber types were very limited, they

were still analyzed in an attempt to get some preliminary data.

11

The S.E.E. 2000 microspectrophotometer was used for this research. This
microspectrophotometer allows transmission spectra in the 250-850nm range. The power
source is a 100-watt xenon lamp attached to a stabilized power source. The spectrometer
is a dual CCD array detector with a total of 2048 detection wells. The first system has
1024 wells with a spectral range of 400-850nm. The second system has 1024 wells with a
spectral range of 250-400nm. The estimated sensitivity is at approximately 86 photons
per count, with a Signal-to-Noise ratio of 250: 1. The 50x objective was used in this
research, allowing a sampling area of 5 microns by 5 microns.11

All fibers, with the exception of cotton, appeared round during identification, so
fiber thicknesses were measured using a 10X objective with a reticle. Measurements for
the cotton fibers were taken from the widest part Of the untwisted areas. Ten fibers were

measured for each fiber type within a sample, and the measurements were averaged.

Instrument Calibration

Calibration of the S.E.E. 2000 Microspectrophotometer requires a NIST traceable
holmium oxide filter set, which has well documented absorption peaks.6 The filter is
placed on the sample stage, and absorbance data is collected versus wavelength. Accuracy
is reflected if the instrument calls several of the appropriate peaks at low, medium and high
wavelength values. Specifically, the absorption peaks Observed with the holmium oxide
filter should be called over a wide range Of the following wavelengths for UV data
collection: 255.5nm, 286.5nm, 332.5nm, 360nm, 385nm, 417.5nm, 445.5nm, 452.5nm, and
459.5nm. The Holmium Oxide Filter #D-111 was used to calibrate the S.E.E. 2000
microspectrophotometer used for this data collection. The calibration data is included, and

the peaks are labeled (See Figure 1).

l2

83082.82

 

 

 

 

 

On? 600 can on«
_ _ L fllrlrlO
. s.
5
3
1a
6.
9
5
IO?
[on
IOO
— EDGE cause—00 3.33.55“. 00.70 52.5.0:

 

 

 

eouemwsuerl

l3

Data Collection

A UV transmittance plot (transmittance vs. wavelength in nanometers) was taken
for each run, consisting Of 25 scans averaged per fiber, with a sampling frequency of 5.
Before collecting the sample data, it is necessary to run a dark scan and a reference scan
for each sample. A dark scan is collected by blocking the light path to the spectrometer.
It is a measure of the instrument noise and interference. A reference scan is a measure Of
the lamp, slide, and mounting media without the sample Of interest. Both the dark scan
and the reference scan for each sample are stored in the instrument’s memory so that they
can be automatically subtracted fiom the sample run.

Afier collecting the necessary reference and dark scans, data were collected for
the sample fibers. The data consisted Of transmittance spectra (i.e. relative transmittance
vs. wavelength) for ten randomly selected fibers within the same fiber class for every
sample. For each sample, an overlay of ten runs from ten randomly selected individual
fibers Of the same class was collected. The overlays are also considered the raw data for
each fiber type within a sample. For example, sample 1 originated from warm-up pants
with a composition of polyester and acrylic. First, the data were collected for ten
randomly selected polyester fibers from sample 1, and the spectra for all ten runs (the raw
data) were plotted as an overlay on one graph. This process was then repeated for the
data collection Of the acrylic fibers in sample 1. Additionally, means and standard
deviations of the overlay spectra were collected with the S.E.E. software. Step-by-step
instructions for the collection of Transmittance data and operation of the S.E.E. 2000 are
detailed in Appendix A. Appendix B contains the raw data, and Appendix C contains the

means and standard deviations for all fiber types within each sample.

14

RESULTS AND DISCUSSION

The data were analyzed both qualitatively and quantitatively to identify
potentially useful characteristics for separation. Qualitative analysis was based on visual
analysis Of the mean spectra, and quantitative analysis was based on statistical analysis of
the mean spectra. First and foremost, the data (transmittance Spectra overlays and sample
mean spectra) were separated and organized by fiber type. The purpose of this separation
was to facilitate a direct comparison between fibers of the same class that originated from
different samples. The results and discussion section is separated to address each of the
fiber classes individually. Each section will discuss regions of interest in the spectra
established by visual examination. In addition, the statistical analysis is addressed by
excel spreadsheets showing mathematical analysis Of the regions of interest. Finally, the
average fiber thicknesses (in microns”) for each fiber type will be included in the excel
spreadsheets to facilitate an assessment Of whether or not the fiber thickness could be
responsible for imparting differences in the spectrographic data. Unless otherwise noted,

the discussion section will refer to the mean data spectra, located in Appendix C.

Polyester

Visual examination of the UV spectral features of the polyester fibers shows three
main regions of interest. The first region is at wavelength < ~310nm. The second region
of interest is at a wavelength of approximately 310nm, and the final region is at

wavelength > ~310nm.
Within the first region (A < ~310nm), all of the polyester fibers tested showed a

relatively flat spectrum with about 20°/o-30% fluctuations. The fluctuations in this region

15

could be rather insignificant, and attributed to light scattering or noise on the baseline.
However, the features in this region seem to be very similar from one fiber to another
within a sample. Further, when comparing this region to polyester fibers from different
sources, there are obvious differences in the fine features. This suggests that this region
may be useful for comparison or exclusionary purposes. Unfortunately, this hypothesis
would need to be the focus of a further study with an updated S.E.E. that allows
exploration of UV transmittance below 250nm. Thus, no statistical analysis of this
region was attempted in this study.

The second region of Significance for the polyester fibers is around the
wavelength of approximately 310nm. At this wavelength, there is a rapid and very
dramatic change in transmittance where the transmittance suddenly increases. This
region of interest seems to be unique to polyester fibers, as none of the other fiber classes
tested showed this consistent and obvious jump in transmittance. This region was
analyzed quantitatively with a Microsoft Excel spreadsheet, and the goal was to
determine if the jump in transmittance at wavelength 310 was statistically significant,
allowing for separation of any of the polyester fibers. The results are tabulated in Table
2. The values for all spreadsheets were obtained from the mean spectra so that the data
could be reduced to characteristics of the mean population.

The statistical analysis for region two (A ~310nm) was calculated as follows. The
mean transmittance and mean standard deviations (SD) for each polyester sample were
recorded at wavelengths 300nm (T1) and 350nm (T 2). Wavelengths of 300nm and 350nm
were chosen because they were fairly stable areas of the spectra that bracketed the region

of interest. The transmittance values at 300nm and 350nm were used to calculate the

[6

change in transmittance (T2 — T1) at 310nm. To determine the standard deviations (SD)

over the change in transmittance (AT), the following equation was used:

 

\l (SD of T1)2 + (SD of T2 )2 . For the purposes of interpreting the statistics, the average

AT and the average SD over the AT were calculated. Any sample that had a AT outside
of the 95% confidence limit (two average standard deviations) was considered
statistically significant. Six of the twenty polyester samples tested were shown to be
statistically different from the others, and they are highlighted yellow in Table 2. Sample
20 has an asterisk next to it because the data is questionable. The AT for this sample is at
the upper most limits of the 95% confidence limit. Since the value is sill within two
average standard deviations, and to retain a conservative approach to the statistical
analysis in this study, the data for sample 20 will not be considered statistically
significant.

Although the statistical analysis of region two lends some credence to significant
quantitative differences between undyed polyester fibers, it is important to consider how
variations in fiber thickness could be affecting the data. Samples within the 95%
confidence limit have polyester fibers ranging in average thickness fiom 10.0 microns to
26.5 microns. Samples that were calculated to be statistically different had polyester
fibers ranging in average thickness from 10.5 microns to 18.5 microns. The data
collected shows that there is considerable overlap in fiber thickness between samples
within the 95% confidence limit and samples outside of the 95% confidence limit. This
data indicates that variations in fiber thickness between samples are not a cause of the

statistically significant findings in this study.

17

Table 2: Quantitative Analysis of Polyester Fibers (Region 2)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Average 7
Ti 72 IT: ' Tr) FIW

Sample Transmiflance Standard Transmittance Standard Standard Thickness

‘ it 3:5an (T1) Deviation gmnm (T2) Deviation T, - T1 Deviation (microns)
1 7.5 0.4 40.3 1.8 32.8 1.8 16.3
2 12.8 1.8 48.9 3.7 36.1 4.1 12.0
5 13.6 1.2 45.3 2.6 31.7 2.9 15.8
7 22.6 3.8 59.2 6.3 36.6 7.4 10.0
8 14.3 1.3 45.6 1.1 31.3 1.7 12.3
9 13.8 2.0 48.9 2.8 35.1 3.4 11.5
11 10.6 1.3 41.6 1.7 31.0 2.1 11.5
20 ' 5.5 0.1 43.2 1.4 37.7 1.4 10.8
21 18.1 1.7 48.1 0.9 30.0 1.9 11.0
22 10.0 1.0 39.0 2.2 29.0 2.4 10.5
23 6.3 0.4 35.6 2.9 29.3 2.9 26.5
24 9.5 0.4 38.7 1.1 29.2 1.2 11.3
27 7.1 0.6 38.5 1.4 31.4 1.5 16.5
31 9.5 0.2 42.2 5.0 32.7 5.0 17.5
4 8.4 0.4 46.9 1.8 38.5 1.8 15.5
10 10.0 3.7 52.1 3.7 42.1 5.2 12.8
13 8.3 0.2 48.6 1.1 40.3 1.1 15.0
14 11.9 1.7 37.5 1.5 25.6 2.3 11.5
28 11.9 1.0 38.6 1.7 26.7 2.0 10.5
29 7.3 0.2 25.7 1.5 18.4 1.5 18.5

Averages of above columns—> 32.4 2.8
LEGEND
" Borderline statistically significant data
Statistically significant data
Average fiber thickness (microns)
IMO! in this table are presented in color

 

 

For the purposes of this study, the most discriminating data seems to be in region

three. Region three focuses on the area of the spectrum with wavelength greater than

310nm. This region consistently shows a slight gradual increase in transmittance as the

wavelength increases. Visual inspection of the samples shows that as the wavelength

increases, some samples show one or more absorbances (dips in the spectrum, or regions

of decreased transmittance). This observation became the basis for initial binning of the

polyester data. The data were separated into groups according to the number of

18

absorbances (0, 1, and more than 1) in region three. Group P0 has zero absorbances, P1
has one absorbance, and P2 has more than 1.

Region three was also analyzed quantitatively with a Microsoft Excel spreadsheet,
and the results are tabulated in Table 3. The goal was to see if quantitative analysis of
region three justified the initial method of binning based on visual characteristics of the
data. If necessary, the groups were reorganized after quantitative analysis. All anomalies
are marked with an asterisk on the spreadsheet, and some of those samples were re-
grouped in accordance with the criteria for quantitative separation. All samples that were
grouped differently afier quantitative analysis will be addressed.

The statistical analysis for region three (A > 310nm) was calculated as follows.
The mean transmittance and standard deviations (SD) for each polyester sample were
recorded at wavelengths 350nm (T2), 390nm (T3), and 425nm (T4). Wavelengths of
350nm, 390nm, and 425nm were chosen because they were fairly stable areas of the
spectra that bracketed the major areas of absorbance that were of interest for the
statistical analysis. The transmittance values at 350nm and 390nm were used to calculate
the change in transmittance (T3 — T2) at one of the major absorbances, and the
transmittance values at 390nm and 425nm were used to calculate the change in
transmittance (T4 — T3) at a second major absorbance. To determine the standard

deviations over the change in transmittance (AT) at the first and second major

 

absorbances, the following equations were used: J (SD of T2 )2 + (SD of T3 )2 and

 

([(SD of T3 )2 + (SD of T4 )2 respectively.

The above calculations were performed for all of the polyester samples, and the

results are listed in Table 3. Generally, the statistical analysis justified the initial binning

l9

based on visual characteristics. Comparison of the statistical results to the initial
grouping based on visual inspection reveals the following:

1. Group P0 can be identified by looking at the values of T4 - T3 and T3-T2. All
samples that had statistical values below 6 at T4 — T3 and a positive value at T3-T2 also
showed no visual absorbances in the spectra Thus, Group P0 could be separated fi'om the
rest of the polyester fibers by either visual examination or statistical analysis. These samples
are highlighted yellow in Table 3. Sample 22 was the only sample that was not initially
placed in Group P0 based on visual inspection because two shallow absorbances were
Observed visually. The raw data for Sample 22 reveals that the absorbances are not
consistent between all of the individual fibers. Perhaps this sample has a mixture of
polyester fibers that were treated differently before manufacturing the item, or perhaps the
agent that is causing the absorbances was not evenly distributed throughout the fibers.
Whatever the reason, the absorbances that are seen visually are not detected statistically
because the absorbances are so shallow. Therefore, this sample was grouped based upon the
statistical evaluation in attempt to be as conservative as possible in the separation of samples.

2. Group P] can be identified by looking at the value of T3 — T2. All samples that
had negative values at T3 —- T2 also visually showed one large absorbance with a
minimum value at wavelength ~375nm. Thus, Group P1 could be separated from the rest
of the polyester fibers by either visual examination or statistical analysis. These samples
are highlighted blue in Table 3. Sample 20 is the only sample that was not placed in
Group P1 based on visual inspection, but statistically should have been re-grouped into
Group Pl . Reevaluation of Sample 20 showed that all individual polyester fibers within

the sample had consistent features. Sample 20 is a true anomaly because visually, the

20

data shows three regions of absorbance, but statistically only one region of absorbance is
captured. Because the visual features are so drastically different fiom what the statistical
analysis captures, and because the features in the raw data are so consistent, it was
determined that Sample 20 could be separated from the rest of the polyester samples
based on visual analysis instead of statistical analysis. Sample 20 was placed into group
P2 with other samples that showed more than one region of absorbance in the data.

3. Group P2 can be identified by values that do not fit into group P0 or P1. All
samples that had a statistical value above 6 at T4 — T3 in addition to a positive value at T3 —
T2 (with the exception of Sample 20, which had a negative value, as discussed above)
showed more than one absorbance in the spectra. This proves that Group P2 could be
separated from the rest of the polyester fibers by either visual examination or statistical
analysis. These samples are highlighted pink in Table 3. Sample 10 is the only sample that
was not initially placed in Group P2 based on visual inspection. Although sample 10
seems to have only one absorbance in region three by visual inspection, the statistics are
placing sample 10 into group P2 because the transmittance value is much higher at one end
of the absorbance at 425nm than at the other end of the absorbance at 350nm. Thus the
statistics are still able to Show a value above 6 at T4 — T3 and a positive value at T3 — T2
despite having only one obvious absorbance. Upon closer examination, it does appear that

sample 10 could possibly have two very slight absorbances that are very close together.

21

        
 

._o_oo :. tacos! e..- 035 0.5 5 no 2:.
AmcoLQEV $05.25 an: 00.803.
8380QO one ems oLoEINn. 98.0
359.8% 2.01.5. 9.6.0
303903.... Benton. 9.6.0
$255320 .w> gaging Emu E moocecoEo ..
QZImrwlmA _

 

 

 

equ><

 

 

a 5.95 23.". .3822. .o 2522 55.230 a .3...

22

Although the statistical analysis of region three lends some credence to significant
differences between the UV spectra of undyed polyester fibers, it is important to consider
how variations in fiber thickness could be affecting the data. Samples in group P0 have
polyester fibers ranging in average thickness from 10.5 microns to 26.5 microns. Samples
in group P1 had polyester fibers ranging in average thickness fiom 10.0 microns to 15.8
microns. Samples in group P2 had polyester fibers ranging in average thickness fiom 10.8
microns to 18.5 microns. The data collected shows that there is considerable overlap in
fiber thickness between groups P0, P1, and P2. This data indicates that variations in fiber
thickness have no impact on the criteria used for the grouping of polyester fibers in this

study.

9m

Visual examination of the UV spectral features of the cotton fibers shows two
main regions of interest. The first region is at wavelength < ~275nm, and the second
region of interest is at a wavelength > ~275nm.

Within the first region (A < ~275nm), all of the cotton fibers tested showed an
area of fluctuations, similar to the polyester fibers. Again, the fluctuations in this region
could be rather insignificant, and attributed to light scattering or noise on the baseline.
However, the features in this region seem to be very similar from one fiber to another
within a sample. Further, when comparing this region to cotton fibers from different
sources, there are obvious differences in the fine features. This suggests that this region
may be useful for comparison or exclusionary purposes. Unfortunately, this hypothesis

would need to be the focus of a further study with an updated S.E.E. that allows

23

exploration of UV transmittance below 250nm. Thus, no statistical analysis of this
region was attempted in this study.

For the purposes of this study, the most discriminating data for the cotton samples
seems to be in region two. This region focuses on the area of the spectrum with
wavelength greater than 275nm. Visual inspection of the samples shows that as the
wavelength increases, some samples show an area of absorbance (a dip in the spectrum, a
region of decreased transmittance). This observation became the basis for initial binning
of the cotton data. The data was separated into groups according to whether or not the
spectra showed an absorbance in region two. Group C0 has zero absorbances, Cl has
one shallow absorbance, and C2 has one deep absorbance. (Note: The mechanism for
grouping these samples will change after statistical analysis.)

Region two of the cotton fibers was analyzed quantitatively with a Microsoft
Excel Spreadsheet, and the results are tabulated in Table 4. The goal was to see if
quantitative analysis of region two justified the initial method of binning based on visual
characteristics of the data. If necessary, the groups were reorganized after quantitative
analysis. All anomalies are marked with an asterisk on the spreadsheet, and those
samples were re-grouped in accordance with the criteria for quantitative separation. All
samples that were grouped differently after quantitative analysis will be addressed.

The statistical analysis for region two (A > 275nm) was calculated as follows.

The mean transmittance and standard deviations (SD) for each cotton sample were
recorded at wavelengths 375nm (T1) and 450nm (T2). Wavelengths of 375nm and 450nm
were chosen because they were fairly stable areas of the spectra that bracketed the major

area of absorbance that was of interest for the statistical analysis. The transmittance

24

values at 375nm and 450nm were used to calculate the change in transmittance (T2 — T1)

at the absorbance. To determine the standard deviations over the change in transmittance

 

(AT) at the absorbance, the following equation was used: J (SD of T1)2 + (SD of T2 )2 .

The above calculations were performed for all of the cotton samples, and the
results are listed in Table 4. Initially, the cotton samples were separated into three groups
based on visual analysis. As stated earlier, the groups consisted of group C0 which had
zero absorbances, group C1 which had one shallow absorbance, and group C2 which had
one deep absorbance. On the other hand, statistical analysis of the cotton samples
revealed that they could only be separated into two groups with total confidence (Group
C0 and Group C1). Statistically, the groups were identified by the value of T2 - T1.
None of the samples had a T2 — T1 value of 5.3< x <8.3. Thus all samples with a value of
5.3 or lower were placed into group C0, and all samples with a value of 8.3 or higher
were placed into group C1. Although many of these samples had fairly high standard
deviations, this is due to the high variability of fiber thickness in natural fibers such as
cotton. Accordingly, not much weight was placed on the values of the standard
deviations for these samples; the value of T2 - T1 was sufficient for purposes of
separation of the cotton samples.

Comparison of the statistical results to the initial grouping based on visual
inspection reveals the following:

1. All samples initially placed into group CO by visual inspection of the data
remained in group C0. These samples showed no visual evidence of a region of
absorbance, and statistically, the T2 — T1 value was relatively low (below 3.5). These

samples are highlighted yellow in Table 4.

25

2. All samples initially grouped together according to a region of one shallow
absorbance were regrouped into group CO afier statistical analysis. These samples are
highlighted yellow in Table 4 and identified with an asterisk. The reason these samples
were regrouped with samples that showed no visual absorbance is because T2 — T1 values
for all of these samples were fairly continuous (anywhere from 1.3 — 5.3) with no clear
value that could be used for separation purposes. Yet it is interesting to note that the
three samples with no visual absorbance do have the three lowest values for T2 — T1. In
any case, the absorbances that are seen visually in these samples are not detected
statistically because the absorbances are so shallow. For this reason, the statistical values
were used for separation purposes instead of the visual characteristics. Samples that were
initially separated according to either no visual absorbance or one shallow absorbance
were grouped together after statistical analysis in attempt to be as conservative as
possible in the separation of samples. Again, sample size for this project was limited, and
perhaps a more distinct cut-off value for separation purposes would become apparent
with further testing.

3. All samples initially grouped together according to a region of one deep
absorbance, remained in the same group after statistical analysis. These samples all had
T2 — T1 values of 8.3 or higher. Although this group was initially named Group C2 based
on visual analysis, the group name changed to Group C1 after statistical analysis of the
rest of the samples placed them all into Group C0. Samples with one deep absorbance,

belonging to group C1, are highlighted blue in Table 4.

26

Table 4: Quantitative Analysis of Cotton Fibers (Region 2)

 

Average
T1 T2 (T2 ' T1) Fiber
Sample Transmittance Standard Transmittance Standard Standard Thickness

 

 

 

# @ 375nm (T1) Deviation @ 450nm (T 2) Deviation T2 - T,

    

  

       

Deviation (microns)

 

 

LEGEND
Group CO—zero visible absorbances
Group CO—one visible shallow absorbance
I Group C1—one deep absorbance
Average Fiber Thickness (microns)
Imes In this table are presented In color

 

 

 

 

 

 

 

Although the statistical analysis of region two lends some credence to significant
differences between the UV spectra of rmdyed cotton fibers, it is important to consider
how variations in fiber thickness could be affecting the data. Samples in group C0 have
polyester fibers ranging in average thickness from 15.8 microns to 21.3 microns.

Samples in group C1 had polyester fibers ranging in average thickness from 17.8 microns

to 21.5 microns. The data collected shows that there is considerable overlap in fiber

27

thickness between groups C0 and Cl . This data indicates that variations in thickness
have no impact on the criteria used for the grouping of cotton fibers in this study.
R_pypr_1

As mentioned earlier, the availability of samples set boundaries to this research.
Samples made up of polyester and cotton were fairly abundant, whereas the samples
made up of rayon, nylon, and acetate were more limited. Accordingly, only 5 samples
for each of the rayon, nylon, and acetate fibers were available for analysis. Furthermore,
there were actually only three rayon samples available, but two of the samples had both
high and low delusterance fibers within the sample. These were treated as different
samples even though they originated from the same source. Although these fiber types
were very limited, they were still run in attempt to get some preliminary data.

Visual examination of the UV spectral features of the rayon fibers shows two
main regions of interest. The first region is at wavelength < ~275nm, and the second
region of interest is at a wavelength > ~275nm.

Within the first region (A < ~275nm), all of the rayon fibers tested showed an area
of fluctuations, similar to the polyester and cotton fibers. Again, the fluctuations in this
region could be rather insignificant, and attributed to light scattering or noise on the
baseline. However, the features in this region seem to be very similar from one fiber to
another within a sample. Further, when comparing this region to rayon fibers fiom
different sources, there are obvious differences in the fine features. This suggests that
this region may be useful for comparison or exclusionary purposes. Unfortunately, this

hypothesis would need to be the focus of a further study with an updated S.E.E. that

28

allows exploration of UV transmittance below 250nm. Thus, no statistical analysis of
this region was attempted in this study.

For the purposes of this study, the most discriminating data for the rayon samples
seems to be in region two. This region focuses on the area of the spectrum with
wavelength greater than 275nm. Visual inspection of the samples shows that as the
wavelength increases, some samples show an area with a very shallow absorbance (a dip
in the spectrum, a region of decreased transmittance). This observation became the basis
for initial binning of the rayon data. The data was separated into groups according to
whether or not the spectra showed an absorbance in region two. Group R0 has zero
absorbances and consisted of sample number 30 (high and low delust.). Group R1 has
one very shallow absorbance and consisted of sample numbers 25 and 18 (high and low
delust.). (Note: The mechanism for grouping these samples will change after statistical
analysis.)

Region two of the rayon samples was analyzed quantitatively with a Microsofi
Excel spreadsheet, and the results are tabulated in Table 5. The goal was to see if
quantitative analysis of region two justified the initial method of binning based on visual
characteristics of the data. The statistical analysis for region two (A > 275nm) was
calculated as follows. The mean transmittance and standard deviations (SD) for each
rayon sample were recorded at wavelengths 375nm (T1) and 425nm (T2). Wavelengths
of 375nm and 425nm were chosen because they were fairly stable areas of the spectra
that bracketed the major area of absorbance that was of interest for the statistical analysis.
The transmittance values at 375nm and 425nm were used to calculate the change in

transmittance (T2 — T.) at the absorbance. To determine the standard deviations over the

29

change in transmittance (AT) at the absorbance, the following equation was used:

 

(ED of T.)2 +(SD of T2)2 .

The above calculations were performed for all of the rayon samples, and the results
are listed in Table 5. Initially, the rayon samples were separated into two groups based on
visual analysis. As stated earlier, the groups consisted of group R0, which had zero
absorbances, and group R1, which had one very shallow absorbance. However, statistical
analysis of the rayon samples revealed that they could not clearly be separated into two
groups, because there was no clear demarcation in the T2 — T; values. The values ranged
fi'om 0.1_<_ x 5 4.2. However, it is interesting to note that the two samples, which were
placed into group R0 based on visual examination (no absorbance), do have the two
lowest T2 — T1 values. This indicates that statistical analysis may have the ability to
separate these samples, but clearly, there needs to be more data to look at before this could
be supported. The data for fiber thickness is included in the table but will not be discussed
further since the samples could not be separated into groups for the purposes of this study.

The rayon samples that had both high and low delusterance fibers showed the
higher delusterance samples having a lower transmittance (or higher absorbance) as
would be expected. The interesting point to make about these samples is that visual
examination of region one (<275nm) between high and low delusterance fibers within a
sample is significantly different. This indicates that the delusterance is imparting some of
the UV characteristics. However, comparison of high delusterance fibers or low
delusterance fibers from different samples is also significantly different. This indicates
that some of the UV characteristics may be imparted by environmental factors. It is

important to remember that this seems to be the trend with the preliminary data, but

30

clearly, more data would need to be collected before any definitive conclusions could be

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

made.
Table 5: Quantitative Analysis of Rayon Fibers (Region 2)
Average
T1 T: (T 2 ' Tr) Fiber
Sample Transmittance Standard Transmittance Standard Standard Thickness
it Q375nm (T1) Deviation Q 425nm (T2) Deviation T2 - T1 Deviation (microns)
18 (High 46.6 2.6 50.8 2.8 4.2 3.8 12.5
18 (Low) 61.9 3.9 63.9 3.9 2.0 5.5 13.3
25 64.7 3.3 67.4 3.3 2.7 4.7 18.8
wigh 22.6 2.5 22.7 2.5 0.1 3.5 14.5
30 (Low)r 65.6 3.9 67.3 4.0 1.7 5.6 13.8
LEGEND |
Average Fiber Thickness (microns)
urges in this table are presented in color

 

 

1111193

Again, the availability of samples set boundaries to this research, so only 5
samples of nylon fibers were available for analysis. Although the nylon samples were
very limited, they were still run in attempt to get some preliminary data.

Visual examination of the UV spectral features of the nylon fibers shows three
main regions of interest. The first region of interest is at wavelengths < ~300nm, the
second region of interest is around wavelength ~300nm, and the third region of interest is
at wavelengths > ~300nm.

Within the first region (A < ~300nm), all of the nylon fibers tested showed an area
of fluctuations, similar to the polyester, cotton, and rayon fibers. Again, the fluctuations
in this region could be rather insignificant, and attributed to light scattering or noise on
the baseline. However, the features in this region seem to be very similar from one fiber
to another within a sample. Further, when comparing this region to nylon fibers from

different sources, there are obvious differences in the fine features. This suggests that

31

this region may be useful for comparison or exclusionary purposes. Unfortunately, this
hypothesis would need to be the focus of further study with an updated S.E.E. that allows
exploration of UV transmittance below 250nm. Thus, no statistical analysis of this
region was attempted in this study.

For the purposes of this study, the most discriminating data for the nylon samples
seems to be in region two. This region focuses on the area of the spectrum around
wavelength ~300nm. Visual inspection of the samples shows that some samples show a
very gradual increase in transmittance around this region, while others Show a very
drastic increase in transmittance around this region. This observation became the basis
for initial binning of the nylon data. The data was separated into groups according to
whether the spectra showed a gradual or a draStic increase in transmittance in region two.
Group N0 shows a gradual increase in transmittance and consisted of sample numbers 16,
17 and 26, and group N1 shows a drastic increase in transmittance, consisting of sample
numbers 6 and 20.

Region two of the nylon fibers was analyzed quantitatively with a Microsofi
Excel spreadsheet, and the results are tabulated in Table 6. The goal was to see if
quantitative analysis of region two justified the initial method of binning based on visual
characteristics of the data. The statistical analysis for region two (A > ~300nm) was
calculated as follows. The mean transmittance and standard deviations (SD) for each
nylon sample were recorded at wavelengths 275nm (T1) and 325nm (T2). Wavelengths
of 275nm and 325nm were chosen because they were fairly stable areas of the spectra
that bracketed the area of increased transmittance that was of interest for the statistical

analysis. The transmittance values at 275nm and 325nm were used to calculate the

32

change in transmittance (T2 — T1) for region two. To determine the standard deviations

over the change in transmittance (AT) in region two, the following equation was used:

 

J(SD of T1)2 +(SD of T2)2 .

The above calculations were performed for all of the nylon samples, and the
results are listed in Table 6. Statistically, the groups were identified by the value of T2 —
T1, and the statistics seemed to support the original binning based on visual
characteristics. None of the samples had a T2 — T3 value of 9.3< x <22.1. Thus all
samples with a value less than or equal to 9.3 were placed into group N0, and all samples
with a value greater than or equal to 22.1 were placed into group N1. Although many of
these samples had fairly high standard deviations, this is due to the high variability of
fiber thickness in these nylon fibers. Accordingly, not much weight was placed on the
values of the standard deviations for these samples; the value of T2 — T1 was sufficient for
purposes of separation of the nylon samples.

All samples initially placed into group NO by visual inspection of the data
remained in group NO. These samples showed a gradual increase in transmittance around
the 300nm wavelength, and statistically, the T2 — T1 value was relatively low (less than or
equal to 9.3). These samples are highlighted yellow in Table 6. Furthermore, all samples
initially placed into group N1 by visual inspection of the data remained in group N1.
These samples showed a drastic increase in transmittance around the 300nm wavelength
region, and statistically, the T2 — T1 value was relatively high (greater than or equal to

22.1). These samples are highlighted blue in Table 6.

33

Table 6: Quantitative Analysis of Nylon Flbel'e (Region 2)

 

 

 

 

Average
Tr T: (T2 '71) Fiber

Sample Transmittance Standard Transmiflance Standard Standard Thickness

# @ 275nm (T1) Deviation 325nm (T2) Deviation T, - T, Deviation (microns)
16 " 51.9 2.2 60.8 ' 1.9 8.9 2.9 20.0
17 60.2 3.6 66.9 3.8 6.7 5.2 18.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LEGEND
Group N0—gradual increase in Transmittance
. Group N1-dramatic increase in Transmittance
Average Fiber Thickness (microns)
Images in this table are presented in color

 

 

 

 

Although the statistical analysis of region two lends some credence to significant
differences between the UV spectra of undyed nylon fibers, it is important to consider
how variations in fiber thickness could be affecting the data. Samples in group N0 have
polyester fibers ranging in average thickness from 18.5 microns to 20.0 microns.
Samples in group N1 had polyester fibers ranging in average thickness from 19.0 microns
to 20.0 microns. The data collected shows that there is considerable overlap in fiber
thickness between groups NO and N1. Although the limitations in availability of nylon
samples set boundaries to this research, this preliminary data indicates that variations in
fiber thickness have no impact on the criteria used for the grouping of nylon fibers in this
study.

Region three of the nylon fibers focuses on the area of the spectrum with
wavelength > ~300nm. Visual inspection of the samples reveals that all samples have a
very gradual, constant increase in transmittance through this region, but only some of the
samples have an absorbance in this region. Initially, this was believed to be good criteria

for separation purposes. However, the absorbances in nylon samples 16 and 17 were so

34

nominal, that statistical analysis would not even be beneficial. Perhaps this region may
be more useful for statistical analysis if a larger quantity of nylon samples is available for

future studies.

Asfllis

As mentioned earlier, the availability of samples set boundaries to this research,
so only 5 samples of acrylic fibers were available for analysis. Although the acrylic
samples were very limited, they were still run in attempt to get some preliminary data

All five of the acrylic fibers had different UV spectral features that were
noticeable through Visual examination alone. Because the spectra were so different from
each other, there was no common link between any of them to be able to easily separate
them into groups. However, all of the acrylic fibers did show the same area of
fluctuations that the polyester, cotton, rayon and nylon fibers did below wavelength
300nm. Again, the fluctuations in this region could be rather insignificant, and attributed
to light scattering or noise on the baseline. However, the features in this region seem to
be very similar from one fiber to another within a sample. Further, when comparing this
region to acrylic fibers from different sources, there are Obvious differences in the fine
features. This suggests that this region may be useful for comparison or exclusionary
purposes. Unfortunately, this hypothesis would need to be the focus of further study
with an updated S.E.E. that allows exploration of UV transmittance below 250nm. Thus,
no statistical analysis of this region was attempted in this study.

At wavelengths above 300nm, the acrylic fibers were analyzed quantitatively with
a Microsoft Excel spreadsheet, and the results are tabulated in Table 7. The goal was to

see if quantitative analysis of this region would justify binning since visual examination

35

did not. The statistical analysis was calculated as follows. The mean transmittance and
standard deviations (SD) for each acrylic sample were recorded at wavelengths 375nm
(TI) and 450nm (T 2). Wavelengths of 375nm and 450nm were chosen because they
captured a region that had a very large absorbance in one of the samples. The
transmittance values at 375nm and 450nm were used to calculate the change in
transmittance (T2 — T1) over this large absorbance. To determine the standard deviations

over the change in transmittance (AT) in this region, the following equation was used:

 

J(SD of T92 + (SD of T2)2 .

The above calculations were performed for all of the acrylic samples, and the
results are listed in Table 7. Statistically, the groups were identified by the value of T2 —
T1, and the statistics seemed to place the samples into three groups. Acrylic samples 1
and 32 had a T2 — I; value close to zero, which means that they had a fairly constant
transmittance between 300nm and 450nm. These samples are highlighted blue in the
table. Acrylic samples 7 and 27 had a slightly larger T2 — T1 value, close to 6, which
indicates that there was an increase in transmittance between 300nm and 450nm. These
samples are highlighted pink in the table. Acrylic sample 3 had a very large T2 — T1
value at 33.8. The large change in transmittance is due to the large absorbance seen in
the spectrum. This sample (highlighted yellow in the table) is clearly very different fi'om
the others both visually and statistically. Although many of these samples had fairly high
standard deviations, this is due to the high variability of fiber thickness in these acrylic
fibers. Accordingly, not much weight was placed on the values of the standard deviations
for these samples; the value of T2 — T. was sufficient for purposes of separation of the

acrylic samples.

36

Table 7: Quantitative Analysis of Acrylic Fibers (Region 2)

 

 

 

A2—AT - 6
A3—large AT
Fiber Thickness (microns)

 

Although the statistical analysis lends some credence to significant differences
between the UV spectra of undyed acrylic fibers, it is important to consider how
variations in fiber thickness could be affecting the data. Both samples in group A1 had
an average thickness of 17.8 microns. The two samples in group A2 had average
thicknesses of 20.0 microns and 23.5 microns. The sample in group A3 had an average
thickness of 22.0 microns. The data collected shows that there is overlap in fiber
thickness between groups A2 and A3, while group A1 had the smallest diameters.
However, the limitations in availability of acrylic samples set boundaries to this research,
and it would be beneficial to gather more data in future studies before determining
whether variations in fiber thickness had an impact on the criteria used for the grouping
of acrylic fibers in this study.

Although all of the data mentioned herein seems fairly promising for separating
undyed fibers, it must be emphasized that this technique is only in the beginning stages.

Many more samples will need to be tested before any definitive conclusions can be made.

37

Additionally, it is important to remember that the data thus far seems to be more relevant

for exclusionary purposes since the data can only be put into groups and not

individualized.

38

CONCLUSIONS AND RECOMMENDATIONS

The data shows many features in the UV transmittance spectra. Some examples
include a rapid variation in transmittance with wavelength < ~300nm, a steep change in
transmittance at wavelength ~300nm, and / or some characteristic variations with
wavelength > ~300nm. These features became the basis for visual inspection of the data,
and the spectra were analyzed as follows: First, the spectra were organized by fiber type
(i.e. polyester, cotton, rayon, etc.). Second, distinguishing characteristics for each
spectrum were noted, and the samples were grouped based on these characteristics.
Third, quantitative (statistical) analysis was performed on these distinguishing
characteristics based upon a mean response of the data from each fiber type, and samples
were re-grouped accordingly, if needed.

The data from this study indicates that there are potential uses and applications for
this work. One potential use for this data is the ability to analyze the characteristics of
the sample mean with the intention of initial binning. For example, this study indicated
that undyed polyester fibers could be separated into three different groups. With
additional research, it is possible that the average characteristics of these fiber types could
lead to automation by computer in the future. Obviously, in the early stages, this
application would be used as a guide only. Additionally, several caveats should be kept
in mind. For example, averaging within a sample may have a large standard deviation
due to fiber to fiber variation within a sample. Further, there is always the possibility for
a significant amount of information to be lost in averaging the transmittance response.
Normalization of the data should be investigated as a way to minimize the effects of

physical characteristics (i.e. fiber thickness) on the characteristics of a sample mean

39

(inter-sample variation). This would also improve the potential automation of binning.
Finally, a more exhaustive study could be performed on the variation of fiber thicknesses
between samples (intra-sample variation). Beer’s Law could be applied to the statistics to
determine whether variation in fiber thickness had any significant affect on the grouping
system used in this study that was not captured through measuring the fibers.

A recommendation for firture studies is to use an instrument that has the
capability of collecting data extending into the wavelength region around 200nm in order
to take full advantage of the data in the region below 300nm. Exploring the data in this
region, which may be rich in features, could potentially hold answers to further
separation capabilities.

Another intriguing research topic would be to explore the fluorescent properties
of undyed fibers, using a microspectrophotometer. Optical brighteners and other
bleaching agents, which may be responsible for the differences in the UV characteristics
observed in this study, may also cause differences in fluorescent properties. It would be
interesting to see whether fluorescence data would be beneficial to use in conjunction
with UV data. Perhaps this joint approach would allow for further separation within the
groups already mentioned herein.

The final recommendation would be to perform a study that has more controlled
variables (i.e. number of times a garment is washed, detergents used, exposure to UV
light) to see if specific UV characteristics are associated with those particular factors.

A study of these controlled variables, in addition to other studies mentioned above, may
give more insight to whether there is any validity to using UV characteristics of optical

brighteners and bleaching agents as a means of definitively excluding (or more

40

importantly, relating) fibers. Again, it will be emphasized that these are merely
speculations for what may happen in the future, and that the data within this study alone
will not support these speculations without further research. This is a preliminary,
exploratory study to see what kinds of features are in the data. Hopefully the data
gathered herein, as well as the aforementioned applications and recommendations, will be

used as a stepping stone for future studies.

41

APPENDICES

42

APPENDIX A

OPERATION INSTRUCTIONS FOR THE S.E.E. 2000

43

9.

OPERATION INSTRUCTIONS FOR THE S.E.E. 2000

. Turn on “Transmission” xenon lamp power supply. Afier approximately 2 minute

warm-up period, observe the digital display on the power supply, and check that the
reading is approximately “75” and stable.

Turn on the power to the Video monitor and computer. Open the “S.E.E.” operating
software. Allow approximately 20 minutes for system warm-up.

. Remove all filters from the light path. Move the selection knob on the front of the

reflectance head to darkfield (DF) position.
Place the sample Slide on the microscope stage.

Rotate the desired objective into place, and focus on the specimen. The 15X
reflecting objective must be used for UV spectrum collection.

Adjust the transmission substage condenser to approximately 0.2 for the 4X and 20X
objectives and approximately 0.5 for the 50X objective. Install the 15X reflecting
condenser if UV spectrum collection is desired.

If the “Reflectance” and/or “Fluorescence” lamps are on, ensure Shutters for these
lamps are off.

Close the field iris diaphragm of the microscope, and adjust the substage condenser so
that the iris diaphragm is focused.

Center the condenser if necessary.

10. Open the field iris diaphragm until the image is just outside of the field of View.

11. Move the sample of interest under the black box as viewed on the monitor, and focus

on the area to be analyzed. Move to a clear area of slide near the specimen of
interest.

12. Block light from the Specimen, and collect “Dark Scan.”

13. Allow light through the path, and collect “Reference Scan.”

14. Optimize the sampling frequency—counts should be approximately 2,500-3,000 in

the visible region and approximately 1,000 in the UV region. If the sampling

frequency is changed, collect a new “Dark Scan” and “Reference Scan.”

15. Move the sample area to be analyzed under the black square on the monitor.

16. Press “Scan Sample” to begin data collection.

APPENDIX B

RAW DATA

45

 

 

 

eggzflugug.saaim.>n

 

 

 

 

 

._.r2

 

>253 8% decades .N 0155.3

 

lob

 

 

.1.

47

 

 

arose hogan Zoe: .4 0323.3

 

 

 

 

 

 

 

>263 c8338 egos .m 0325.24

a:

 

ran

ran

 

 

eeusprrnsuru;

49

Dug-Z

 

 

\\ . £55 53.58 3.8... c. 29592.

 

 

 

comma;

E80805:

 

 

acres 8838 86...: .o 2955.3

 

 

 

mammalian

51

Inga-.2

 

 

52

eousnnusum,

 

>283 .8828 no»?! .a 99:35:

 

 

 

 

UV-Sample 10, undyed polyester overlay

 

 

L§

 

 

60-
40-

manpower;

'53

 

Nanometers

 

UV-Sa'nple 11. tndyed pelyester overlay

 

.a .
‘ ;\\\ ‘

gi//

 

 

 

l
450
Nehemeters

 

 

 

\l ‘— ..
. rev
5.. a,
r .
\

 

antes season 835 .3. £595

 

re»

 

 

COW],

55

252.352

 

 

 

3:26 .3828 6265 .9 0355.3

 

 

 

 

esuanrursuer;

 

{3238-2

 

23:96
.2328
ensue: .8 ca .
_ Eew.
>3

1:

a8

 

 

so
uamursuu;

57

 

 

 

 

>556 c2328. Zoe: .3 0.9.593

 

 

eeumursuerr

23050302

r5

 

 

 

>mto>o .2328. eo>dcz .NN 29:qu

.1\.\.t.\3\\.3.€~r)~cl\.\rlr’r J .3“ a.
. 3.1.3.1..-. .. , a.
.\ .. . f _

 

 

.I.

59

 

 

 

>353 3.323 nouns: .mu 29:ng

 

 

.l.

2085232 . ,

 

 

 

>355 Bung—on 33:: .3. 29592.

1:

 

 

 

.I.

61

 

 

 

5.96 .3328 326:: Km 29:85:

 

 

.l.

62

282552

 

 

>355 .2328 822.: .mm oEEom.>:

 

 

 

 

 

.l.

88958-2

 

3:26 .2328 Etc... .8 o_nEam.>:

 

 

 

 

 

.l.

 

 

>mzo>o 3.328 33:: .3 0383.):

 

 

 

.l.

 

 

 

 

 

 

 

 

 

 

 

v . ttx
3.38 8:8 33.5 .v egg d

 

 

continuum-u

67

 

 

5.26 82832561953:

 

 

 

 

2825...:

 

 

 

. .. t I! .5 so. ..
\ § . l _ ”—
Ell? k..." / / ..a‘.l.. . a ., ._ .
\\\\IH\\\I ..\< - I.) .:.....z 3.... \ .. .‘(2 x a , M

3:26 588 no»?! K 0.959):

 

 

 

mummy;

282852

“.3.

 

3:26 8:8 365 .u 295m.>:

 

 

 

 

 

 

 

 

WW1

B32552

 

. £233
365 a
3:26 8:8

 

 

 

 

 

 

 

mummy;

71

Hugh-08.2

 

 

>256 :88 .335 .9 0.8.5.2..

 

 

 

 

 

”0W1

 

 

 

 

€55§§8555§>=

 

 

 

 

commas!“

2825...:

93
L

LE

 

 

>585 5:8 3.6:: .w. o..9=um.>:

 

 

 

‘74

 

 

 

 

 

 

 

 

288.232

one , 8? am»
_ _

L§
§

 

 

 

 

 

 

 

 

 

 

 

>§§§8€§i£§3 .

 

 

mama;

'76

 

 

 

 

 

 

282:
23:

.Ffi'

 

»
«:26 5:8 no?
.2 :23.
>3

 

 

 

 

Wu
J.

 

 

 

 

.656 :88 365 .2. gas-$5

 

 

 

”0W1

2833...:

 

586 5.836% .3 283.3

 

 

fico

 

 

mm;

28.5232

L5

 

 

 

 

3:30 8:8 365 .2 £592.

 

 

81

”—ng

r§

 

 

 

 

3:96 :88 no? . «N 03.59):

 

 

mammals;

882.352

 

‘ '
).I\\ (I \.
O a.
‘3) \‘n\\‘.\ I s‘ \‘L\.’
‘ ‘ I. { '
\Ao.’ ‘\an (xiii ’II I \D ’l\.\‘
\u o ‘ . I ‘0’
‘ I .I‘. ‘x
‘
It; I!
I‘ I). I
I

[18,“)
. ‘1‘...
$H.....nnL.ll}...l I... III. )3!)
'5“ “Ir- . ‘5.
5 .
I‘Icl\|/C .\

xmto>
o :88 396:: Na 29%»
.>:

 

  

 

 

 

1

282.352

 

.l.

 

 

 

3:96 5:8 “.262: .3 0.9592.

 

 

 

8308052

 

.3./x 2...
I{§./\I\)l\fl\ I li\|ll|¢&.\..xl.\.\)(ll‘\tu \I. < k (.l\
5“ I,

\/\z\:l!.\,\:\,.\. .

“ I ‘\B\

It‘lllo}. I'll-I" J.‘ \lllltzpi‘ll‘ ’.Il.lllu§l (a

I
'l‘.’\<."|$.a0\-ul\u i.}\ I‘lln‘r‘. ‘.

\Hx‘Qonw.‘ in; .

 

 

.\:I \t.\....\...:\t.

L.<.,\{c<...\. {$3.

.\.\J.\(I\ (/\../\\...2.\...\\!:I\\.<>.

 

.11.]. .3...../<\~./\(

3:96 couoo 355 .3 0355.):

 

 

 

 

 

.I.

28.52.:

r§

 

 

 

>€26 33.8 5.5 5.: no.2: .3 32.3.5

 

 

 

 

comma;

2825...:

 

 

 

 

 

 

5.26 $28 3: E»! g5 .2 29.8.3

 

comma-:1

 

 

230809.62

 

5.26 :93. 3.6:: .3 0385.24

 

 

 

 

 

.l.

282.352

 

{1‘

ll")\|l‘ll )(Itle u “(fi\§”§ni.’\)figl \‘Ida” . \.I!lr... : \- u .. ‘\ I
[I tut-\l‘l" (- f\. I... \\ .\‘-“lc $.15“; t\\( $59!“) .
41...... .. . c- . .. ,...<.. .. .xxtnkfiweflti. 2 .33.. . . . ) .

. ,1: s
. . . l ‘9 ol I
I. {\I\ It, ..
lax .

. I) III
lln' \ll". I’llol‘nnl . ...\-. \ ‘n/\\\- 250....

. . .
a 4d - Q ,. > .
us. t. u , .
\. n.1ol.IXo.\\C..-‘. :0..\).I\Iu..\\ A. . o. s]... x. . . .L
. III! I. \ . .3 . ~. ’
t . .. H. ‘ -' 4‘ .
x .1.. .. 2.1.2.. .3... r k 4 x ,
. \ .3 ~ ' \ s .q ‘

g.“ l‘l‘.\}anl\l|lll\agllll\ ..\I)..I }|\<I.\|.’\I\.I’\3\J/\Ilc/\I.\$I)(\\lfi.. l....\s .. fits}... .k. ..\.\ ./...... .u. }
2<§.EI‘2\1.Jg/I\c1 a taint)... Hg 4
.I\\l\.ll.'.\.llu 1‘! \l \ R I \ (12k

\’ l l\l.\b.l O I. . . ‘5 .f ‘
\li“ oak! . .) ..Ih.al-\..( .. as; . . my
.I [(3). g. (Is u. (idox 1! a... \. u. r...\).. .. . \ \..» . . .. . . _ , ._ .,
1.)?‘O‘JI‘N fl“ .‘ i \ ..\‘ l/\\\II‘).H \./\ql.¢n.\l H. . \..’ . o . . P a . a ‘3‘ 1|. vk...‘ &\ o . J _
I‘blo. . . .. . r .
. . -. .. s \ \x . ,
.9, ...\ \ . _ . _~

32823.8 :95 3:26 :22 8.6:: .8 0353.2.

 

12.

 

 

 

.I.

282.652

 

 

\(3
II. \.‘II\.\.\!{ l\ Ix 2.:3 .I\. K1£Zar .....\\.\
\. .\ 3\/\.\. (k. K. \. .
\... .2... .
- 5. ......... ._..... a {4;
{Igl(‘ u!‘l‘n\l..\ :Kl“..\l3.\ $.13"! 33 s IN...) 3‘. anvil.)\(\ /\.1..\.\}13..u..\\ ...... ...... -

x s... .3. ...

Ill“!

 

. 5.3..
. quu -\ I‘.‘ .
I . III. talc x... .- .\.o‘l.. [.0...\ .0’...o.\looao\oongl3. .‘3‘ 0 03 ..3\ x
5.1. .33).... /\\;\.. Is... .\I.....
\3
\t”? \(Il‘ult .0! ...\l
I\) J \3..\. ‘12.; (ck ./£\\\
\\. n.\/\..~ .. t. 29.3.... :5-.. ...
I, \Il.‘ ‘a‘l‘. ‘- ‘l \ II\
\I‘ct .3]\ .\\ \.I.l\. I.
r'“{ \)n \ I ‘33.
‘31. \v 1“.
.8

30:92.28 :6: 5.26 :28 3.6:: .8 29592..

 

 

 

continuum;

 

.656 SE. 8.65 .o £2.92.

 

 

 

 

000'!”le

91

282.552

r§

 

 

 

>396 c2»: 35:: .2. anamI>D

 

j»

 

 

3803—06-2

 

 

3:96 :03: no»?! .2 0.9593

1K

 

 

monumental

3805052

 

 

3:26 :03: 33:: .3 oEE¢W>D

 

on? 8‘ can 08
IF L _
.\‘_
ion
.\
Kit .I.‘ x. x.
All ‘ IA;
. Iii. \g . \
|I|\\l\l|l\|\l \ .. .c..\|.. l 1}. . .
t al‘.“ l.‘ ’ ~x
low
I}. \\
.l..\..

 

 

288.3
52

 

 

 

>m
=m>
o
co
=8 8%:
a u
on
0
.38.
>2

 

fie.

 

 

mm!
mm
1.

232.852

L§

 

 

 

 

 

5.26 2.5. 365 .P 228$:

 

 

T2.

 

 

continuum;

 

gal

an

 

 

>233 23.8 “.265 .m 195?;

 

 

 

communal,

28050—32

 

 

3:25 2.5g 3&5 g .3233

 

 

 

 

mm; ’

8805052

 

 

 

 

3.05 0.3.80 33:: Km 29:qu

 

 

.l.

 

 

UV-Sample 32, undyed acrylic overlay

- . .
.
"- w
0.1" "'-_..-.,-.-......—P" “I" -o -I‘

.A“..‘

roam—W

 

”WU-L

100

T
350

 

Nanomeum

APPENDIX C

MEANS AND STANDARD DEVIATIONS

101

c8285:

 

 

 

an? cat can as .
_ — h —
....\.\
Ol“\l.\.\\.\ll\ll\n..\
tfilllllizll‘lkiks.l\n\;\ix\1.....k
.Ilul‘l‘lll“ [I'll‘l‘l‘l’l‘la‘l‘llll-l. It: ‘l‘i‘h 4 .\I.\\\.|y....\\_.l..
Hill-1"- . I. .I nllllll‘l‘llc
:- .. iii!
lllll“

 

38.338 2853 new 525 8.828 F i=3 >3

 

 

ooumnusuu;

102

E8082.- .

r3.

 

 

@2838 28:2» 2. 525 5828 u 35% >2

 

 

W1

103

832.652

 

 

 

augusoc 28:8» 25 82$ .2828 e 238 >2

 

13

IS.

 

 

mm;

104

 

 

2.: 9.3 2.” §
(\S :2.
I! {\ank
-l\‘ ’1‘ I
In
I.»
\\
x.\H\.\ .3.

:3

:8

38.83% 23:3. Ba 825 3.828 a £23 >2

5

 

 

 

 

W311.

105

 

3:28.32 p.883 Bu 525 8.82.8 5 one.» >2

 

 

 

flab.

W1

106

8802052

one cow can con
r _ . _ A.

 

 

$82252 23:8» 25 525 8.82.8 2 29:3 >2.

 

 

 

 

 

 

 

 

 

 

 

3:23.52 2833 2o 525 8.828 a £58 >2

 

8805052

 

 

 

 

 

 

$8332 235: 2a 525 53.28 2 .353 >2

fie.

flan

 

 

eon-mama;

109

 

 

68.338 2.222» 25 582 3328 : 295w >2

1:.

 

AK

 

scum],

110

 

 

 

38.538 285.» E. 525 33228 2 2......» >2

 

 

 

000W},

111

288.352

amt

om

 

 

 

38.538 2.85.» 25 825 .2828 3 £28. >2

 

 

mummy;

112

882.652

Pg

 

 

280352 28cm.“ Ba 525 £828 on £88 >2

r2.

1:0

 

 

W;

113

2305232

he:

 

 

 

$822.52 23:2» 93 525 852.8 E 228.6 >3

r2.

 

 

W1

114

280E232

out now can
_ . _ r

 

 

 

 

38.338 32% 25 :85 5828 «u 295» >2

 

 

 

 

sou-quantum;

115

 

£0.32

 

r§

95.55% was...» 8. 525 8328 an 32.8 >2

 

 

 

now;

116

280E052

 

98.533 23%.» Ba 51.523328 3 £58 >2

 

 

 

 

minimum;

117

 

38.538 2853 2a 82.. 8% B one.» >2

 

[2.

1n

 

 

mama-J;

118

2808052

F§
§

 

 

 

 

$8225.. 235.» E. 525 .2823 a 22.8 >2

 

Wm

 

 

119

. 9.82.352

cm»

 

 

 

 

38.53% 2.283.. 8.. 825 .3988 on 2853 >2

1:.

Ian

[on

 

 

.I.

120

882552

 

 

€2.22... 28%.... 8. 525 .2828 3 £88 >2

 

 

 

sou-mammal

121

282.352

8Q 8' con . 86 SN

 

 

6.3.3329 9.09.3» .83 525 .388 .N 29:8 >3

 

 

 

communal;

122

8308052

 

 

$8838 235.» 8a 58... 528 .c 22.8 >2

.loo

1.2.

 

 

manual;

123

2808052

_ _ _ h

 

$8338 2.8:...» Ba :32. 828 a 2..an >2

 

 

 

 

 

W1

124

2.805022

 

.
K {in} :K 3d)... .

. ER
‘0'!“

3.5.538 2.2.5... Ba 28.5 .328 ... 28:...» >2

 

 

 

125

minimum;

3.3-50:32 .

 

 

 

38.33% 238.3 2.3 =32... 8:8 .2 29.32 >2 .

- 000mm;

 

 

126

E33852

eat 8' own

 

\i\5--

 

‘.
’I‘l ‘
’0':
u”
3”
“S
I:
I’l."
‘ I -I
{It I
I,
I, I
. lu-i"
{a}
I.
.0 ll

\
\‘a‘l‘l’
_\ n} I
(I:
‘
n'
\
I1
\.

\§

1.111..

{(1-19

3.3.3.5.. 23833 93 .38.. 328 .a 3.853 >2

\\
A E...

12.

IQO

 

 

now;

127

282.552

 

 

 

3.3.338 23833 .23 .325 .328 .2 .853 >2

?\,\SS\1//>>\c T2.

 

W341

128

230E052

 

 

38.338 333.33 2.3 .325 8:8 .2 2.333 >2

Gav 03. on” can 3N
.. . L . . b V
.60
x\\\l\.\\r./ ./. \. 2../$.20
\ ./.i \\\J\I)\r\/\></\x)\.
\ \\.\.‘\..\ll/ 1 1.5.3:}: 2.}
\ \ .....l l\\ , «(:3 >22 2.
I||\a\\ \\\ ..\l )1!) l/.)......l/). \\.t\x\a < fiO
111-1‘. .. I .\ . .
\\ \\ .\ .1./:1. . I It}.\.\\.
r3

 

 

mammal;

129

 

 

@2333 23:9» 2. 525 8:8 .§ 28.8 >:

 

r§

130

”MIN-L

 

 

a. an . 2.; §

[00

. a>>>

\‘ix’l. slit}!!! \}>\>\<z

\ \(x\\.\(\.()
|\I\.II.II\III\ll|I.III|l|..I\n\.\ \\\I\'\l|\ll’l’lit (1)....Ill. .Ilti))\||\.\ 21...)...k \3¢/>\:>\/\! ION
$I¢\
|\I\||\||I\|IIII\‘I|\I..\I\ I\\:\.\.l|lllll.l..|l:r/.l\)iai s! {iiyi‘l’\\\(s

I‘ll}!!! V ..\l.\.\ . r8
18

 

38.33% E85: 2.. 53.5 5:8 .2 295m >3

 

 

magnum;

131

9.82.352

 

 

 

 

 

803:6 285.» as .325 8:8 .1 29:8 >2

3.. 0.... § §1 4"

13
.8

xi
\s\, #2.

\tt\r\<
“\H skis:
.II. {III ..\.\..\\..\
rt.\\\\l\lil\l.| i 1:} 3.1.1:...

H.” \ifl: i-s\_.\.\. é

 

 

288:0:-

 

 

acofisou 238.» 2a 825 8:8 .2 £58 >3

 

 

 

WI",

133

38088-2

 

 

@2638 2.25» as 525 8:8 .2 .328 >3

1.5

 

 

W;

134

2806232

 

 

{)0}?!

.{Jl‘ {- I‘)

z >.\?x .
\\1 2,3,, -.,., a. ixxkicf
- p1 b«i(l\\\\ 235; :x i, .2. .-.-..?s)7\s\ $8:

$8.58.. 2858 Be essay 5:8 .2 £58 >3

 

 

135

mammal

 

£38.35:

 

38.6.5.3 288.» 2. $25 83.8 .3 29:8 >3

 

W1

 

 

 

136

8805052

 

 

68.538 285» 2a 5.25 8:8 .2 «358 >3

.\.\\,).\k s:
.-..52\-.\3I\\..\ éigl?
a \ Iéh
tsk

 

sou-mammal

$0

 

 

137

L'§

 

 

35.33% 285.» new 525 5:8 .8 29:3 >3

..2..

flan

loo

 

0..

 

000mm;

138

 

230506.:

r§

 

£

111(1)-

 

\.\

\

\Ia\.\.\\.|.\

$3.!
‘0. . .. . z
.-...._.\I{:5.x}.l xtxnszéi (-531... ,...-.z.. i. 7/\/\\

{‘0‘

clxl\.nv. . .
\Il\l\l} (‘Klo I] [III-co k

0’
. O
3|. u o ._
‘
a ..!‘\-a l. - fl

\‘(I‘llSitCC-Jgill. a
l---..i. 2x 1. x z... .

. ‘u‘II-I " o S -

$5.338 28:8» can 525 =o§..u~ {Eon >3

 

12.

 

 

W1

139

 

 

 

1:.

38.638 28%.» 93 52$ 8:8 .3 295m >3

 

 

 

W1

140

9.82.352

an a... m.

r§

 

 

35.538 235» 2. 8.5 8:8 .8 29:8 >3

 

 

 

18.

W1

141

880509.02

 

 

38.338 Esau .23 58.5 .323 5325 .8 one.» >3

 

 

 

ecu-mm;

142

23058.2

 

3....

Ill‘l‘ll‘\l\l\"ll‘l\l\"l‘i(llllvd\‘li'1l a \
. 30‘3". (a‘jl \l‘t‘ I
s3 12"». ..
I l .50 0.. . In, {.3

35.53% 285..» 2. 58.5 .93 22.5%.. .8 £23... >3

. > \
E3..-) xxx..- 22

 

 

”W1

143

 

 

@8333 23:8» Rm 52$ :38 .3 29:5 >3

 

 

 

new”; -

144

9.305.052

 

 

. l | ’1‘!!!‘ ’1’" OI! Its ‘0 \ll '1‘ \{k’raLf\ o
I.\I|II\I‘II 1" I . \( ll‘lllu. . l‘l‘ ($1. Ila-I‘ll, in»
1.88)....\ .\ s 1 188 \ 81K.).J$\, 2 d
{\l’ \u" (lai'.l).)l\/\\ln}’lll f8!\tlualll.lt\o{ (ll-(I I a." 28..
_IIIIII\II‘\ f (In 4‘ I‘ll). ‘51! . {$.- [31‘

68.3.53 3.32% 35 825 33.8 32.1.2.2 .8 £53 >3

 

12.

 

 

”WI-l1

145

9.3050932

 

E3883

 

 

w)\<<fi\

A2325“. Bags» 25 525 $28 32%». .8 £55 >2

 

146

comm;

 

 

232.852 .

 

 

683.6“. 2.83. Ba 825 :2»: .o 28.3 >3

1...: 2R a. a... a.
fig
3 332.
\ a 3%.
332.5;
\x.. .
.\ as
. 8-. a... x..\ \
_ . -8338 .... x... .\.. 1..
(Ex, -
, \8\\.\|\8\
\t 1:.

 

 

autumn;

147

£80535

a: 8*

 

 

 

'I I I ‘ \t l
’
'\*.‘
\‘ll’ll, I’TJ II "OI)III I\‘ \(I! .‘I’( I.)
l E
.

888 888 l ltti‘l ‘3

98.63% 288.» 2a 525 2...»: .2 one.» >:

 

1o»

 

 

mummy;

148

882.652

 

 

8.8888 38.638 2353 2a 525 :2»: .2 anam >3

149

W1

 

 

88.8052

 

 

$co=2>ou 23:3» can 525 :2»: on 33.3 >3

:9.

lav

10m

 

 

”UWUUL

150

280.5232

 

 

\.\l {‘(cl

38.538 235.» v.5 :85. :22 .8 29:8 >3

 

 

 

 

communal

lSl

9.322052

3‘
I.

 

 

 

38:23.0 28:8. us. 52:2 25.8 .. oaeam >2

 

 

1:.

 

 

oouuumsuu;

152

9.32.852

om? . as can can

4

 

6:22.32. 2352» can .325 0.32 .9 29:5 >3

 

 

 

 

aounmmsuul

153

EouoEocaz

one as can
87 _ _

con
F

 

\la‘

8.\n\n\.

 

1888 I
. ‘1
\ul‘ifil‘l’lI’88‘lll 8.81 3l\3|\all

K}...

8 \8) l.)\|t\(1l\.. \..\..\
‘l\lc‘lll\l1\l\l\ \‘
818 138% \‘l

\\

3:28.52 23.5% new 525 grow K anuw >3

[an

:2.

T.

 

 

aoummsuu;

154

2808052

 

 

 

 

38.333 28%.» Ba 525 2.2.8 .R 20.58 >2

 

:2.

 

 

10m

mummy;

155

9.308052

9.3. as aura can om"

 

\3(\ss\

88188:“HMWEK A712 5
8\i\8\|\\li\\.88\8 ..

III‘IIlll/lll (’88 ) .\..\ .I\I\8Inl\

8.888

l‘\

95.338 235% new coo—5 2.2.8 .3 22.5.0; >3

12

rec

 

 

 

aoucngwsuul

156

REFERENCES

l. U.S. Department of Justice, Federal Bureau of Investigation, “Microscopy of
Textile Fibers,” Forensic Science Communications 1, no. 1 (1999):
http://www.fbi.gov/hq/1ab/fsc/backissu/apfil1999/houckch2.htm.

2. Samuel J. Palenik, “Microscopical Examination of Fibers,” in Forensic
Examination of Fibers, ed. James Robertson and Michael Grieve, 2‘“l ed. (Philadelphia:
Taylor & Francis, 1999), 155-56.

3. Barry D. Gaudette, “The Forensic Aspects of Textile Fiber Examination,” in
Forensic Science Handbook, Vol. 2, ed. Richard Saferstein (Englewood Cliffs: Prentice
Hall, 1988), 241 .

4. James Dunlop, “Colour Analysis by Microspectrophotometry,” in Forensic
Examination of Fibres, ed. James Robertson (England: CRC Press, 1992), 127-28.

5. Franz-Peter Adolf, “Microscope Photometry and its Application in Forensic
Science, Part 1,” Presenter’s notes from presentation in Helsinki, Finland, (1986), 8.

6. US. Department of Justice, Federal Bureau of Investigation, “Visible
Spectroscopy of Textile Fibers,” Forensic Science Communications 1, no. 1 (1999):
http://www.fbi.gov/hq/lab/fsc/backissu/april1999/houckch3.htm.

7. Franz-Peter Adolf and James Dunlop “Microspectrophotometry/Colour
Measurement,” in Forensic Examination of Fibers, ed. James Robertson and Michael
Grieve, 2“cl ed. (Philadelphia: Taylor & Francis, 1999), 251-52.

8. CRAIC Technologies, UV-visible-NIR-microscope spectroscopy,
http://www.microspectra.com.

9. Dr. Paul Martin, “Ultraviolet Microspectral Analysis of Bleached Wool,”
S.E. E. Incorporated Applications Report, (1999), 1-8.

10. Andrea Desrosiers and Dr. Paul Martin, “Ultraviolet Microspectral Analysis
of Nylon Fibers,” S. E. E. Incorporated Applications Report, (1999), 1-4.

11. Mitchell Kupferman, letter from applications scientist at S.E.E. Incorporated,
9 September, 2002.

12. Charles R. Woodfall, Microscope Micron Measuring,
http://www.microscopy-uk.net/mag/artdec99/cwnano.html.

157

BIBLIOGRAPHY

Abramowitz, Mortimer. Fluorescence Microscopy: The Essentials. Melville: Olympus
America, 1993.

Adolf, Franz-Peter. “Microscope Photometry and its Application in Forensic Science,
Part 1,” Presenter’s notes fi'om presentation in Helsinki, Finland, (1986), 8.

Adolf, Franz-Peter, and James Dunlop. “Microspectrophotometry/Colour Measurement.”
In Forensic Examination of Fibers, edited by James Robertson and Michael
Grieve. 2'“I ed., 251-90. Philadelphia: Taylor & Francis, 1999.

CRAIC Technologies. UV-visible-NIR-microscope spectroscopy.
http://wwwmicrospecu‘acom.

Desrosiers, Andrea, and Dr. Paul Martin. “Ultraviolet Microspectral Analysis of Nylon
Fibers.” S. E. E. Incorporated Applications Report. (1999): 1-4.

Dunlop, James. “Colour Analysis by Microspectrophotometry.” In Forensic
Exainination of Fibres, edited by James Robertson, 127-40. England: CRC Press,
1992. _

Gaudette, Barry D. “The Forensic Aspects of Textile Fiber Examination,” In Forensic
Science Hancflmok, Vol. 2, edited by Richard Saferstein, 209-62. Englewood
Cliffs: Prentice Hall, 1988.

Martin, Paul. “Ultraviolet Microspectral Analysis of Bleached Wool.” S. E. E.
Incorporated Applications Report. (1999): 1-8.

Palenik, Samuel J. “Microscopical Examination of Fibers.” In Forensic Examination of
Fibers, edited by James Robertson and Michael Grieve. 2'“I ed., 153-78.
Philadelphia: Taylor & Francis, 1999.

US. Department of Justice, Federal Bureau of Investigation. “Introduction to Forensic

Fiber Examination.” In Forensic Science Communications 1, no. 1 (1999).
http://www.tbi.gov/hq/lab/fsc/backissu/april1999/houckch1 .htm.

US. Department of Justice, Federal Bureau of Investigation. “Microscopy of Textile
Fibers.” In Forensic Science Communications 1, no. 1 (1999).
http://www.fbi.gov/hq/lab/fsc/backissu/april1999/houckch2.htm.

US. Department of Justice, Federal Bureau of Investigation. “Visible Spectroscopy of
Textile Fibers.” In Forensic Science Communications 1, no. 1 (1999).
http://www.fbi.gov/hq/lab/fsc/backissu/april l 999/houckch3.htm.

Woodfall, Charles R. Microscope Micron Measuring. http://wwwmicroscopy-
uk.net/mag/artdec99/cwnano.html.

158