     

 

LIBRm
Michigan State
University

This is to certify that the
thesis entitled

DETERMINATION OF TRACE ELEMENTAL

CONCENTRATIONS IN DOCUMENT PAPERS FOR
FORENSIC COMPARISON USING INDUCTIVELY COUPLED

MS.

PLASMA MASS SPECTROMETRY

presented by

ELIZABETH ANN MCGAW

has been accepted towards fulﬁllment
of the requirements for the

degree in Forensic Science

MUM

Major PTr essor’s Signature

 

01/1233 /0 7
Date

MSU is an afﬁnnative-action, equal-opportunity employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRC/DateDue.indd-p.1

 

DETERMINATION OF TRACE ELEMENTAL CONCENTRATIONS IN
DOCUMENT PAPERS FOR FORENSIC COMPARISON USING INDUCTIVELY
COUPLED PLASMA MASS SPECTROMETRY
By

Elizabeth Ann McGaw

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

2007

ABSTRACT
DETERMINATION OF TRACE ELEMENTAL CONCENTRATIONS IN
DOCUMENT PAPERS FOR FORENSIC COMPARISON USING INDUCTIVELY
COUPLED PLASMA MASS SPECTROMETRY
By
Elizabeth Ann McGaw

Traditional document paper analysis involves examining mostly physical
Characteristics of the paper: weight, size, color, opacity, brightness, and
watermarks. While these are useful Characteristics they are usually uniform for a
speciﬁc brand and type of paper because these are carefully controlled by the
manufacturer. Trace elemental concentrations are not speciﬁcally controlled and
can vary between batches of the same brand and type of paper making this type
of analysis more selective.

In this work microwave digestion was used to dissolve the paper samples
prior to analysis by inductively coupled plasma mass spectrometry (ICP-MS)
This technique was successfully applied to paper samples from two vendors
(New Leaf Paper® and Staples®). Five reams from each vendor were sampled
and it was determined that the elemental concentration (Mg, Al, Mn, Fe, Sr, Y,
Ba, Ce and Nd) was consistent across single sheets as well as throughout a
single ream of paper. Statistically Signiﬁcant differences were observed between
reams and between vendors. The most useful elements for discrimination were

AI and Ba for reams of New Leaf Paper® and Mg, Mn, and Sr for reams of

Staples® paper. The best element to discriminate between the vendors was Ba.

‘ ACKNOWLEDGMENTS

There have been many people that have helped me during my time here
at Michigan State. First, I would like to thank Dr. Jay Siegel, for your guidance
and encouragement. It is because of you that I decided to pursue a PhD in
Chemistry, you believed in me before I believed in myself and for that I will
always be grateful. Second, Dr. Ruth Waddell, thank you for “adopting” me into
your group. I appreciate all of the guidance and support over the past year while
working to finish this project.

I also need to acknowledge all of the help I received from Dave Szymanski
on the lCP-MS analysis. Thank you for putting so much time into testing my
preliminary project ideas and giving up time on your weekends to run my
samples! Thank you for offering ideas and suggestions and being willing to
answer all of my crazy questions from lCP-MS to geology.

And ﬁnally I need to that my friends and family for their support. For Mel,
you’ve been here with me since the beginning and it has been wonderful to share
everything with such a great friend. To everyone else, Amber, Audrey, Luther,
and JB, thanks for listening when I had silly questions and for keeping me sane
while I was trying to ﬁnish two degrees at once. To my parents for all of their
love and support and for teaching me a strong work ethic that has gotten me
through graduate school. To Scott, thank you for all of your love throughout all of
this and being willing to tolerate me coming home late day after day, I couldn’t

have done this without you by my side.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................. vi
LIST OF FIGURES ............................... vii
CHAPTER 1
INTRODUCTION ....... ' .......................................................................................... 1
Background ..................................................................................................... 1
Document Paper Production ............................................................................ 2
Document Analysis .......................................................................................... 5
Traditional Analysis .................................................................................... 5
Elemental Analysis ..................................................................................... 8
Inductively Coupled Plasma Mass Spectrometry (lCP-MS) ........................... 11
Thesis Overview ............................................................................................ 16
References .................................................................................................... 18
CHAPTER 2
EXPERIMENTAL MATERIALS AND METHODS ................................................ 21
Paper Samples .............................................................................................. 21
Microwave Digestion ..................................................................................... 22
Particulate Characterization .......................................................................... 25
Fourier Transform Infrared Spectroscopy (FTIR) ..................................... 25
X-ray Diffraction Spectroscopy (XRD) ...................................................... 26
Dissolution/Hydroﬂuoric Acid Digestion .................................................... 26
Inductively Coupled Plasma Mass Spectrometry (lCP-MS) ........................... 27
Chemicals ...................................................................................................... 29
Statistical Analysis ......................................................................................... 29
References .................................................................................................... 33
CHAPTER 3
PARTICULATE CHARACTERIZATION .............................................................. 34
Introduction .................................................................................................... 34
Particulate Identiﬁcation ................................................................................ 35
Fourier Transform Infrared Spectroscopy (FTIR) ..................................... 35
X-ray Diffraction Spectroscopy ................................................................. 36
Elemental Analysis of Hydroﬂuoric Acid (HF) Digest ................................ 40
Effect of Particulate on Elemental Analysis ................................................... 41
Comparison of Nitric Acid/Hydrogen Peroxide and Hydroﬂuoric Acid
Digestion .................................................................................................. 41
Effect of Filtering HNOg/HzOz Digest ........................................................ 43
Elemental Absorption by Particulates ....................................................... 44
Conclusions ................................................................................................... 46
References .................................................................................................... 47

CHAPTER 4

RESULTS AND STATISTICAL COMPARISONS ............................................... 48
Introduction .................................................................................................... 48
Method Development .................................................................................... 48

Full Mass Scan ......................................................................................... 49
Calibration ................................................................................................ 52
Procedural Blanks .................................................................................... 53
Relative Standard Deviations ................................................................... 54
Variation Between Reams ............................................................................. 56
New Leaf Paper® ...................................................................................... 56
Staples® ................................................................................................... 62
Variation Between Vendors ........................................................................... 69
Conclusions ................................................................................................... 73
References .................................................................................................... 76
CHAPTER 5
CONCLUSIONS AND FUTURE WORK ............................................................. 77

Table 2.1.

Table 3.1

Table 3.2

Table 3.3.

Table 3.4

Table 3.5.

Table 3.6.

Table 4.1.

Table 4.2.

Table 4.3.

Table 4.4.

LIST OF TABLES

Microwave digestion temperature program ................................... 24

Comparison of absorption frequencies in IR spectra of particulates
and reference materials ..... . ............................................................ 36

X-ray diffraction peaks for Ti02 ...................................................... 39

Comparison of AI, Si and Ti concentrations, determined by lCP-MS,
in solutions from HNO3/H202 and HF digestions ........................... 41

Comparison of elemental concentrations in solutions from
HNO3/H202 and HF digestions ...................................................... 42

Comparison of elemental concentrations for ﬁltered and unﬁltered
samples. Errors represent standard deviations ............................ 43

Elemental concentration paper samples digested with HNOa/H202
and allowed to equilibrate in the presence of the particulates for
various amounts of time ................................................................ 46

Instrument quantitation limits for elements chosen for paper
analysis. Sc was not linear so no detection limit is listed .............. 52

Elemental concentration of procedural blanks, compared to lowest
concentration of the element in any paper digest solutions. ND
indicates concentration was below the detection limit ................... 54

Two-way ANOVA results for New Leaf Paper®. Signiﬁcant
differences deﬁned as p-value s 0.05 ............................................ 57

Two-way ANOVA results for Staples® paper. Signiﬁcant
differences deﬁned as p-value s 0.05 ............................................ 62

vi

Figure 1.1

Figure 1.2.

Figure 1.3.

Figure 1.4.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 4.1 .

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

LIST OF FIGURES

Overview of the papermaking process for both virgin and secondary
fibers. The steps where chemicals are added are shown in italics
and the additives are shown in the boxes ........................................ 5

ICP torch and sampling interface. (A) solution aerosol injection, (B)
RF induction coils to sustain plasma, (C) shielding box, (D)
induction region, (E) initial reaction zone of the plasma, (F)
analytical zone, (H) expanding gases sampled from plasma, (I)
sampling cone and (J) skimmer cone ............................................ 12.

Calculated values for degree of ionization (MVM) reported as a
percentage. Elements with * also form a signiﬁcant portion of
M + ................................................................................................. 13

Diagram of a quadrupole. An example of a stable ion’s oscillatory
path is drawn down the center of the four parallel rods. ................ 15

FTIR spectrum of particulates from paper digestion ...................... 35

X-ray diffraction spectrum of undigested material from the

HNO3/l'l202 digestion ..................................................................... 38
Kaolin peak in the x-ray diffraction spectrum of undigested material

from the HN03/H202 digestion ....................................................... 40
Full mass scan of New Leaf Paper® (NL131) digest ...................... 50

Expanded region (m/z 20 — 30) of full mass scan for New Leaf
Paper® (NL131) digest ................................................................... 50

Expanded region (m/z 50 - 150) of full mass scan for New Leaf
Paper® (NL131) digest ................................................................... 51

Expanded region (m/z 150 — 240) of full mass scan for New Leaf
Paper® (NL131) digest ................................................................... 51

Mg concentration in New Leaf Paper®, reported as ug Mg per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.05 (*) or p s 0.01 (**) ........... 59

vii

Figure 4.6.

Figure 4.7.

Figure 4.8.

Figure 4.9.

Figure 4.10.

Figure 4.11.

Figure 4.12.

Figure 4.13.

Figure 4.14.

Figure 4.15.

Figure 4.16.

Al concentration in New Leaf Paper®, reported as pg Al per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.01 (**) ................................. 59

Sr concentration in New Leaf Paper®, reported as pg Sr per 9
paper. Error bars represent standard deviation. Significant
difference between values with p s 0.05 (*) or p s 0.01 (**) ........... 60

Ba concentration in New Leaf Paper®, reported as pg Ba per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.01 (**) ................................. 60

Mg concentration in Staples® brand paper, reported as pg Mg per 9
paper. Error bars represent standard deviation. Significant
difference between values with p s 0.05 (*) or p s 0.01 (**) ........... 63

Al concentration in Staples® brand paper, reported as pg Al per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.01 (**) ................................. 64

Mn concentration in Staples® brand paper, reported as pg Mn per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.05 (*) or p s 0.01 (**) ........... 64

Fe concentration in Staples® brand paper, reported as pg Fe per 9
paper. Error bars represent standard deviation. No signiﬁcant
differences between pairs .............................................................. 65

Sr concentration in Staples® brand paper, reported as pg Sr per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.05 (*) or p .<. 0.01 (**) ........... 65

Y concentration in Staples® brand paper, reported as pg Y per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p S 0.01 (**) ................................. 66

Ba concentration in Staples® brand paper, reported as pg Ba per 9
paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s. 0.05 (*) ................................... 66

Ce concentration in Staples® brand paper, reported as pg Ce per 9

paper. Error bars represent standard deviation. Signiﬁcant
difference between values with p s 0.05 (*) ................................... 67

viii

Figure 4.17.

Figure 4.18.

Figure 4.19.

Concentration of trace metals found in the paper samples from New
Leaf Paper® and Staples® brand. The error bars represent
standard deviation of the values n=15. Signiﬁcant difference
between values with p s 0.05 (*) or p s 0.01 (**) ........................... 7O

Ternary plot of the ratio of Ba, Ce and Nd. The Ba concentration
was signiﬁcantly higher than the Ce or Nd concentrations therefore
it was divided by 10 to make the plot easier to read. Each point
represents an average ream value, 5 reams from each

vendor are plotted .......................................................................... 72

Ternary plot of the ratio of Ba, Sr and Nd. The Ba and Sr
concentration were signiﬁcantly higher than the Nd concentration
therefore there were divided by 10 and 100, respectively, to make
the plot easier to read. Each point represents an average ream
value, 5 reams from each vendor are plotted ................................ 73

CHAPTER 1
INTRODUCTION
1.1. Background

Documents have become an important part of our society. Nearly every
major life event involves a document: birth certiﬁcate, death certiﬁcate, property
‘ deed, marriage certiﬁcate. Documents are used in all sorts of transactions daily
for example, Checks, passports, wills, contracts and loans. While documents
may not be common evidence in violent crimes, crimes committed with false
documents (e.g., forged Checks, embezzlement, tax fraud, altered wills) cost our
country billions of dollars each year.1

Questioned document analysis involves a wide variety of analyses
including handwriting, ink, typescript, printed and copied materials and the paper
itself. Document examiners are also trained to visualize entries that have been
erased or obliterated and even to visualize indented writing. Recently there has
been considerable emphasis in research on comparing the ink2'9 and toner”‘14
used in a document and even studies to determine age of the ink.15 21 While
identifying the ink is certainly important there are other aspects of document
analysis that can be equally valuable.

The research presented herein will focus on the chemical composition of
the document paper itself, speciﬁcally the trace elemental components. Since
paper is created from natural materials (e.g., wood pulp, clays) and recycled
consumer waste, it is highly unlikely that different paper manufacturers will

produce a product with identical Chemical composition.‘ While only Chemical

analysis of the paper will be discussed in this work this is just one portion of a
comprehensive document analysis. There are many different types of papers
available commercially and the samples Chosen for this work were document
papers (multipurpose copy paper) made from 100% recycled materials, since
such paper is commonly used in everyday activities.

1.2. Document Paper Production

Paper is composed of natural ﬁbers, usually cellulose ﬁbers (e.g., wood
pulp, hemp, cotton), that are combined and ﬂattened by heat or pressure.
Although other fibers such as asbestos, silk, wool, glass, and plastic can be used,
cellulose is the least expensive and therefore the most commonly used in paper
manufacture.22 The samples used in this research were made of cellulose ﬁbers
so only this production method will be discussed.

Pulp is the raw material that makes up the paper. This pulp can come
from a number of sources however in the United States wood is typically the
source of virgin raw ﬁber.23 Most paper mills now use a mixture of virgin and
secondary (recycled wastepaper) pulps.23 The wood for pulp is usually material
left over from creation of higher value lumbers and the residual wood from these
processes is then chipped into small pieces (approximately 20 mm x 5 mm),
which are turned to pulp by either mechanical or Chemical methods. Mechanical
methods grind the wood and typically create much shorter ﬁbers that make a
much weaker paper but 90-95% of the input material is retained in the ﬁnal
pulp.23 The most common chemical pulping method, called the Kraft process, is

accomplished by cooking the wood Chips in a solution of sodium hydroxide

(NaOH) and sodium sulﬁde (NaZS). The solution is very basic (pH >13) and will
attack lignin, the highly polymerized substance found in the intercellular spaces of
the wood. Since lignin is undesirable in high quality papers, the Kraft process is
used to solublize the lignin and free up the cellulose fibers. Chemical pulping
creates longer ﬁbers but has lower yields (45-50%) so usually some combination
of mechanical and chemical pulping is used to obtain the highest yield of pulp
while maintaining longer ﬁbers for good paper strength.

Once the pulp is created it is reﬁned and must go through a series of
processing steps. The pulp is washed to remove residual Chemicals (e.g., NaOH,
NaZS) from the Chemical pulping process. It is then screened to remove any
large (> 318 inch) or unwanted particles from the “good” ﬁbers. The pulp is then
bleached to obtain the desired brightness. Chlorine was commonly used as a
bleaching agent until the 19903. However, since it is known to be a major
producer of toxic Chlorinated organic compounds, chlorine has been phased
out.23 Now the most common bleaching sequence includes oxygen extraction,
alkaline extraction, and peroxide steps. After completion of these steps the pulp
can then be shipped to paper manufacturers.

Wastepaper can be repulped for use and goes through many of the same
steps described above. The wastepaper is ﬁrst dispersed and broken up into a
water solution. It is then cleaned and chemically deinked followed by a very
similar screening and bleaching processes as the virgin ﬁbers. The difference is

that the recycled pulp samples will always contain some amount of contaminants;

waxes, glues, inks, and various other dissolved solids that may affect the end
paper product.

The paper manufacturer will generate a wet stock by dispersing the pulp
ﬁbers into a water slurry. At this point the manufacturer mixes the desired
amount of virgin and secondary (recycled) paper pulp. To this wet stock a variety
of paper chemicals are added to give the end product certain qualities.23 Acids or
bases are added to control the pH of the paper to prevent yellowing. Starches
are added to improve the tensile strength of the paper. Fillers such as Clay, talc,
and titanium dioxide are added to improve the opacity. Chemical dyes are also
added at this point for whitening or color.

The pulp is then placed on large screens and the water is allowed to drain
off. Physical watermarks may be introduced onto the paper by adding a pattern
to the screen on which the paper is formed. The resulting paper sheet is then
squeezed between rollers to remove remaining water and ensure smoothness
and uniform thickness. The paper will also pass through large heated rollers to
remove any remaining water before being cut to the appropriate dimensions.

There are many points during this process in which different Chemicals
and trace components can be introduced into the paper. An overview of this
process is shown in Figure 1.1, the steps where Chemicals may be added to the

process are displayed in italics and possible chemicals that made be added are

noted in the boxes.

Virgin Secondary (Recycled)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wood Post-consumer waste paper adﬁﬁiisves
l l
Wood Chips Slurry
l I
NaOH , l t
Nazs PulpIng Deinking 52:36:58
\Washing / enzymes
1
Filtering
Cl 1
H202 3,9530an acids/bases
03 l, starch
Chemical additives Clay
talc
1' , mica
ScreenIng 1102
l
Drying

Figure 1.1. Overview of the papermaking process for both virgin and secondary
ﬁbers. The steps where Chemicals are added are shown in italics and the
additives are shown in the boxes.

1.3. Document Analysis
1. 3. 1. Traditional Analysis

The physical properties of the paper are typically examined ﬁrst.1 These
include the size, weight, opacity, color, brightness and ﬂuorescent properties of
the paper. Opacity is a measure of the amount of light that passes through the
paper. The color of many papers is slightly yellow so some manufacturers may
add blue dyes to make the paper appear whiter, ﬂuorescent dyes are also used
for this purpose. The color, and presences of visible dyes, can be measured
using UV/visible spectroscopy. The presence of ﬂuorescent dyes and the

intensity of the ﬂuorescence can also be measured by spectroscopy. Most of

these physical Characteristics are carefully controlled by the manufacturers during

their quality control procedures and will not be very discriminating for paper
produced by the same manufacturer.

Watermarks, imparted during paper manufacture, also serve as an
important Characteristic of the paper. These watermarks are usually trademarked
property of a manufacturer and many manufacturers will use different watermarks

for different grades of paper.22

Watermarks are generated in several different
ways: imprinted into the wet sheet (conventional method), mechanically
embossed after drying, or Chemically added. The watermarks that are
mechanically embossed cause a compression of the fibers and will have a raised
appearance and feel, different from the conventional or chemical methods. A
conventional and Chemical watermark can be differentiated under ultraviolet light
where the conventional one will appear lighter and the Chemical one darker than
the paper. The watermarks are not only helpful in identifying the source of the
paper but may also narrow dates of prodUction based on manufacturer’s codes or
design changes.

The next step in the analysis would be to examine the ﬁber content.
Different papers have different ﬁber types (e.g., wood, cotton) and an examiner
can even determine if different pulping processes (e.g., mechanical, Chemical)
were used. This is typically done using microscopic analysis and observing the
morphology of the ﬁbers present. The paper sample must be deﬁbered, which
involves tearing it into small pieces, boiling in water and shaking vigorously to

produce a suspension. This suspension is then placed on a carefully cleaned

microscope slide. Once the sample is mounted different stains can be used to

differentiate different wood types, for example a Stellger stain (calcium nitrate
and potassium iodide) can differentiate coniferous and hardwood pulps.22 Other
stains can be used to determine the content of bleached pulp by determining the
amount of lignin still present. While staining can identify the types of ﬁbers
present this is still a Class Characteristic and many papers of the same type may
also contain the same mixture of ﬁbers. The technique is tedious, requires an
experienced scientist, and is destructive.

The last part of the document analysis is the analysis of the Chemical
composition of the paper. Many Chemical additives are used during the
production of the paper to improve its color, strength, opacity, adjust pH, and
even to prevent growth of microorganisms (e.g., fungi).23 Infrared reﬂectance
spectroscopy can be used to determine the coating contents based on
characteristic adsorption bands for the different coating materials (e.g., starch,
calcium carbonate, casein). The ﬁllers and pigments in papers (e.g., titanium
dioxide) are typically crystalline in nature and therefore can be determined by x-
ray diffraction analysis. Typical ﬁllers such as kaolin Clay, calcium carbonate,
talc, mica, and titanium dioxide can be identiﬁed at concentrations of less than
2% by weight.22 X-ray diffraction is also a non-destructive method. There are
many other methods that can be used to determine speciﬁc Chemical
components through extraction methods and additional information about these
tests can be obtained from Browning’s work.”

While all of the tests described above are useful for the paper analysis

there may be many papers that have all the same Characteristics, for example,

two paper samples of the same type from the same manufacturer.
Manufacturers strive to produce a product that is very uniform and have greatly
improved the quality control of the manufacturing processes. Therefore there are
fewer differences among paper samples of the same type. However, even if the
main components are the same there can still be variation at the trace level
especially in the case of recycled materials.
1. 3. 2. Elemental Analysis

Elemental analysis of paper samples can increase the signiﬁcance of an
association and can also be helpful in determining approximate production times
if suitable references are available. Elemental analysis has been performed on
paper samples using a variety of techniques including neutron activation analysis
(NAA),2“'26 scanning electron microscopy coupled with electron dispersive x-ray
spectroscopy (SEM-EDS),27 atomic absorption spectroscopy (AA),28 and
inductively coupled plasma mass spectrometry (ICP-MS).29' 3°

Schlesinger and Settle did some of the initial elemental Characterization of
paper samples using NAA in the early 19703.24 A total of 102 pairs of paper
samples were obtained from nine manufacturers, each pair containing a sample
from the beginning and end of a production run. Relative standard deviations for
elements in these pairs were relatively high, ranging from 12-20%. However,
when only sample pairs containing high Al and Ti concentrations (ﬁller
components added deliberately by the manufacturers) were considered, relative
standard deviations of <10% were obtained. No specific elemental markers were

identiﬁed for any of the manufacturers.

Blanchard and Harrison developed a method to determine a “ﬁngerprint”
based on the Clay content of the specific papers using NAA.25 Initially 12 clay
samples (8 kaolin and 4 non-kaolin) were analyzed and it was determined that
the trace elemental proﬁle of these Clays was different. Five paper samples were
then prepared from pulp with 15% of a different Clay ﬁller added to each, then
analyzed by NAA. Elements with relatively high concentrations were marked as
key identifying elements, the ratio of the key elements to other key elements was
calculated, and the presence of unique elements was noted. These features
were used to compare the results in the paper samples to the results from the
Clay samples. The ﬁve paper samples were correctly associated with the Clays
used to produce them. While this is a useful analysis for papers with a high clay
content more modern copy papers have replaced much of the clay with calcium
carbonate.29

Brunelle and coworkers also used NAA and found that some elements
were more likely to be present (Na, Mn, Ag, and Cu) but elements that were less
common (Ta, As, and Sn) were more valuable for distinguishing samples.26
Brunelle was able to distinguish papers from different manufacturers, 600
samples from 10 manufacturers. The papers were distinguished based on which
elements were present as well as the concentration of the elements. The
disadvantage to NAA analysis is the need for special irradiation facilities that
makes this a less universal method and not commonly available in forensic crime

labs.

Polk and coworkers used scanning electron microscopy with electron
dispersive x-ray spectroscopy (SEM-EDS) to determine elemental composition of
several writing papers.27 SEM is typically used for high power magniﬁcation and
can be used as an imaging technique for ﬁber morphology. However, in
conjunction with EDS this technique is also able to give elemental analysis. The
most abundant elements detected in the paper samples were Al, Si, S, Ca, and
Ti. The relative standard deviation on a single sheet and within a box of paper for
all these elements was low (<5% for sheet, <10% for box). They were able to
distinguish sheets from different but identically labeled boxes. The disadvantage
of SEM-EDS is that it has limited sensitivity for elements with atomic numbers
larger than Na and relative, rather than absolute, concentrations are used for
discrimination.

Copeland and coworkers ashed and directly analyzed paper samples
using atomic absorption spectroscopy (AA).28 They determined ten elemental
concentrations (CU, Pb, Mn, Sb, Cr, Co, Cd, Fe, Mg, and Ag) present in the mid
ppb to ppt range. They then used Clustering algorithms (Q mode and Zahn
minimal spanning tree) to group the samples. Using Co, Cr, Cu, Mn and Sb
concentrations 12 of the 19 papers were differentiated using the Zahn minimal
spanning tree algorithm and 16 of the 19 were differentiated using the Q-mode
algorithm. The 3 samples that could not be differentiated, using Q-mode, had
high relative standard deviations (>10%) for all of the elements used, which likely
affected the differentiation capability. The limitations of this method seem to be

due to variation in the measurements which may be limited by the reproducibility

10

of the dry ashing technique. AA is also typically a single elemental analysis and
can be time consuming to perform multi-element analysis.
1.4. Inductively Coupled Plasma Mass Spectrometry (ICP-MS)

ICP-MS has been used as a method for trace elemental analysis since the
19805.3“33 ICP is an efﬁcient ion source used to generate ions to be separated
and detected by MS. This technique has several advantages, rapid multielement
analysis, low detection limits (ppt), wide linear dynamic range (5-6 orders of
magnitude), and high precision (0.5 — 5%).33

A plasma is a gas Cloud of particles including radicals, free electrons, ions
and neutral species. The plasma is formed in a torch that consists of 3
concentric quartz tubes, see Figure 1.2, and the end of these tubes are placed in
a copper induction coil. Argon gas ﬂows between the two outermost tubes and
an electric spark is used to introduce free electrons. The induction coil has an
oscillating (RF) frequency that will accelerate the electrons in alternating
directions. These accelerated electrons will collide with Ar atoms and can cause
them to lose an electron. This will continue until the rate of Ar ions forming is
equal to the rate of electrons recombining with Ar ions and the plasma reaches a
steady state. The plasma consists mostly of Ar atoms with a small fraction of free
electrons and Ar ions.

The plasma is very efficient at generating ions which results in low
detection limits for most elements. Solution samples can be directly analyzed
using ICP as an atomization and ionizationmethod. An aerosol is generated from

the solution using a nebulizer. This aerosol is introduced into the plasma where

11

the temperature ranges from 6000 - 10,000 K and is typically 6000 — 6500 K in
the analytical zone where ions are sampled.34 The particles will spend
milliseconds in the plasma which is sufﬁcient time to desolvate, atomize, excite,
and ionize the atoms, forming a high density of positive sample ions.31 The
ionization efﬁciency is described by the Saha equation.35 Thompson and Houk36
calculated the degree of ionization (MN) for many of the elements using this

equation and these calculations are shown as percentages in Figure 1.3.

 

 

 

 

 

Mass Analyzer

 

 

 

 

 

 

Figure 1.2. ICP torch and sampling interface. (A) solution aerosol injection, (B)
RF induction coils to sustain plasma, (C) shielding box, (D) induction region, (E)
initial reaction zone of the plasma, (F) analytical zone, (H) expanding gases
sampled from plasma, (I) sampling cone and (J) skimmer cone.

12

 

 

0.1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

98 90

Th Pa
100*

Li Be 8 c N o F Ne
100 75 58 5 0.1 0.1 9x10" 6x1 0“5
Na Mg Al Si P S Cl Ar
100 98 98 85 33 14 0.9 0.04

K Ca Sc TI V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se 8r Kr
1 00 99 100 99 99 98 95 96 93 91 90 75 98 90 52 33 5 0.6
Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te l Xe
100 96 98 99 98 98 96 94 93 93 85 99 96 78 66 29 8.5
Cs Ba La Hf Ta W Re 05 Ir Pt Au Hg Tl Pb Bi Po At Rn
100 91 90 98 95 94 93 78 62 51 38 100 97 92

Fr Ra Ac . 9 , , .

Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
99’ 91 92

 

 

99* 97 190* 93 99* 100*
P Am c Bk cr E r d N
1310‘ NpluL lm' islmlMlo Lr
Figure 1.3. Calculated values for degree of ionization (MVM) reported as a
percentage.36 Elements with * also form a signiﬁcant portion of M”.

 

 

 

—#

 

 

 

 

 

The calculations from this equation indicate that the efﬁciency of forming
singly Charged ions is very high and these are the most abundant ionic species
generated. This is important because it greatly simpliﬁes the mass spectra that
are obtained.

The ICP generates ions at atmospheric pressure however mass
spectrometers require low operating pressures (<10'5 torr) therefore these ions
must be introduced to the MS through a differentially pumped vacuum system.
The interface (Figure 1.2) is designed to extract a small amount of the plasma
gas along with the ions into the vacuum system. A sampling cone is positioned
to extract just above the hottest part of the plasma where the highest density of
ions exist. The sample gas is directed through the sampling cone to a region
where the pressure is approximately 1 torr. The gas expands in the vacuum

region forming a plume. The skimmer cone is positioned just behind the sampler

13

cone and is designed to transmit as much of the sample as possible into a
second vacuum chamber.

After the ions have entered the second vacuum chamber they are focused
with ion lenses and accelerated toward the mass analyzer. In some cases a
collision cell is used to remove some of the interfering species. Interfering
Species (Ari, ArN+, ArO“ etc.) are likely to be formed in Clustering reactions that
occur in the ion sampling interface.32 The collision cell is a ,multipole (4, 6, or 8)
chamber with only AC voltage at RF frequency applied to the poles to act as a
high-pass mass ﬁlter allowing all ions over a predetermined mass pass through.
The chamber is ﬁlled with a reaction gas, typically hydrogen or helium. The
hydrogen and helium promote collision induced dissociation (CID) of the
Clustered ions and therefore reduce the interference of these molecules in the
ﬁnal mass spectrum.37

The typical mass analyzer used in commercial ICP-MS instruments is a
quadrupole (Figure 1.4). The quadrupole consists of 4 parallel metal rods with
opposite pairs electrically linked together. On each of these pairs a DC and AC
voltage (with RF frequency) are applied. For one pair the DC voltage is positive
and for the other it is negative and the AC voltages have the same amplitude but
are 180° out of phase with one another. The ions enter the quadrupole region
and the oscillating electric ﬁelds cause the ions to travel in oscillatory paths along
the z-axis. At particular settings for the DC and AC voltages only ions with a
particular mass to charge (m/z) ratio will have a stable path and will travel

through the ﬁeld and reach the detector. All other ions with different m/z values

14

than the selected value will impact the rods and become neutralized and will not
be detected. By scanning the voltage (DC and AC) settings, ions of different m/z

values will be selected for detection, allowing full mass scan of ions.

 

Figure 1.4. Diagram of a quadrupole. An example of a stable ion's oscillatory
path is drawn down the center of the four parallel rods.

ICP-MS has been used by Spence and co-workers to differentiate
document paper from 17 manufacturers around the world based on trace
elemental concentrations.29' 30 A Single sample (30 mm x 40 mm) was taken from
5 random sheets in each ream, only 1 ream was selected from each
manufacturer, except the Australian manufacturer where 6 reams were obtained.
These samples were microwave digested and analyzed using ICP-MS. Nine
elements (Na, Mg, AI, Mn, Sr, Y, Ba, La and Ce) were selected as potential
discriminating elements because they showed low relative standard deviations
over the sheets from one ream and had minimal interferences from polyatomic
ions. All 17 paper manufacturers could be distinguished based only on elemental
concentrations of Mn and Sr. Papers from the same Australian manufacturer but

four different batches (sampled monthly) were able to be statistically

15

distinguished using Al, Zr, and Mn; however, differences in consecutively
manufactured rolls could not be determined.
1.5. Thesis Overview

Forensically it would be useful if paper of the same type from the same
manufacturer or vendor could be differentiated. Many of the current document
analysis techniques are based on Characteristics that vary between
manufacturers and paper types but do not explore differences between batches
or reams of the same paper. Being able to detect differences between reams of
the same paper would add another level of speciﬁcity to the document analysis
and serve to narrow the number of potential sources.

This thesis describes studies of the trace elemental composition of 100%
recycled document (multipurpose copy) paper as determined by ICP-MS. These
studies were intended to build on previous work and expand the study of
differentiation of papers from the same vendor, Speciﬁcally looking at
differentiation of papers from separate reams.

Samples were studied from two US. vendors, ﬁve reams from each.
Several levels of comparisons were made to assess the potential of
differentiating paper samples based on elemental composition. First the
homogeneity of the element composition across a single sheet and within a ream
of paper was determined to Choose elements suitable for comparison between
reams. Next a statistical comparison was made between reams from the same

vendor to differentiate the reams. Finally a comparison among papers from the

16

two vendors was made, aiming to distinguish different vendors based on

elemental content of the papers.

17

1.6. References

(1)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

Brunelle, R. L. In Forensic Science Handbook, 2nd ed.; Saferstein, R.,

Ed.; Prentice Hall: Upper Saddle River, New Jersey, 2002; Vol. I, pp 697-
744.

Grim, D. M.; Siegel, J.; Allison, J. Journal of Forensic Sciences 2001, 46,
1411-1420.

Dunn, J. D.; Siegel, J. A.; Allison, J. Journal of Forensic Sciences 2003,
48, 652-657.

Dunn, J. D. The detection of ink dyes by laser desorption mass
spectrometry coupled with thin-layer Chromatography and the use of
photochemistry for dye Characterization. MS. Thesis, Michigan State
University, East Lansing, MI, 2004.

Wilson, J. D.; LaPorte, G. M.; Cantu, A. A. Journal of Forensic Sciences
2004,49, 364-370.

Siegel, J.; Allison, J.; Mohr, D.; Dunn, J. D. Talanta 2005, 67, 425—429.

LaPorte, G. M.; Arrendondo, M. D.; McConnell, T. S.; Stephens, J. C.;
Cantu, A. A.; Shaffer, D. K. Journal of Forensic Sciences 2006, 51, 689-
692.

Zieba-Palus, J.; Kunicki, M. Forensic Science International 2006, 158,
1 64-172.

Jones, R. W.; Cody, R. B.; McClelland, J. F. Journal of Forensic Sciences
2006, 51, 915-918.

Armitage, S.; Saywell, 8.; Roux, C.; Lennard, C.; Greenwood, P. Journal
of Forensic Sciences 2001, 46, 1043-1052.

DeKoeijer, J. A.; DeMoel, J. J. M. Z Zagadnien Nauk Sadowych 2001,
2001, 413-427.

Merrill, R. A.; Bartick, E. G.; Taylor, H. J. Analytical and Bioanalytical
Chemistry 2003, 376, 1272-1278.

Egan, W. J.; Morgan, S. L.; Bartick, E. G.; Merrill, R. A.; Taylor, H. J.
Analytical and Bioanalytical Chemistry 2003, 376, 1279-1285.

Egan, W. J.; Galipo, R. C.; Kochanowski, B. K.; Morgan, S. L.; Bartick, E.

G.; Miller, M. L.; Ward, D. C.; Mothershead, R. F. Analytical and
Bioanalytical Chemistry 2003, 376, 1286-1297.

18

(15)
(15)

(17)

(18)

(19)

(20)
(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

Stewart, L. F. Journal of Forensic Sciences 1985, 30, 405-411.

Aginsky, V. N. International Journal of Forensic Document Examiners
1998, 4, 214-230.

Grim, D. M.; Siegel, J. A.; Allison, J. Journal of Forensic Sciences 2002,
47, 1294-1297.

Grim, D. M. Laser desorption/ionization mass spectrometric analysis of
methyl violet: a new approach to relative ink age determination. MS.
Thesis, Michigan State University, East Lansing, MI, 2002.

LaPorte, G. M.; Wilson, J. D.; Cantu, A. A.; Mancke, S. A.; Fortunato, S. L.
Journal of Forensic Sciences 2004, 49, 155-159.

Hofer, R. Journal of Forensic Sciences 2004, 49, 1353-1357.

Xu, Y.; Wang, J.; Yao, L. Forensic Science lntemational 2006, 162, 140-
143.

Browning, B. L. Analysis of Paper, 2nd ed.; Marcel Dekker, Inc.: New
York, 1977.

Smook, G. A. Handbook for Pulp & Paper Technologists, 3rd ed.; Angus
Wilde Publications Inc.: Vancouver, 2002.

Schlesinger, H. L.; Settle, D. M. Journal of Forensic Sciences 1971, 16,
309-330.

Blanchard, D. 8.; Harrison, S. H. Journal of Forensic Sciences 1978, 23,
679-686.

Brunelle, R. L.; Washington, W.; Hoffman, 0.; Pro, M. Journal of the
Association of Ofﬁcial Analytical Chemists 1971, 54, 920-924.

Polk, D. E.; Attard, A. E.; Giessen, B. C. Journal of Forensic Sciences
1977, 22, 524-533.

Simon, P. J.; Giessen, B. C.; Copeland, T. R. Analytical Chemistry 1977,
49, 2285-2288.

Spence, L. D.; Baker, A. T.; Byrne, J. P. Journal of Analytical Atomic
Spectrometry 2000, 15, 813-819.

Spence, L. 0.; Francis, R. B.; Tinggi, U. Joumal of Forensic Sciences
2002, 47, 648-651.

Houk, R. S.; Fassel, V. A.; Flesch, G. D.; Svec, H. J.; Gray, A. L.; Taylor,
C. E. Analytical Chemistry 1980, 52, 2283-2289.

19

(32)
(33)
(34)
(35)

(35)

(37)

Hou k, R. S. Analytical Chemistry 1986, 58, 97A-105A.

Olesik, J. W. Analytical Chemistry 1991, 63, 12A-21A.

Thomas, R. Spectroscopy 2001, 16, 26-30.

Miller, M. In Inductively Coupled Plasmas in Analytical Atomic
Sgggtrometry, Golightly, D. W., Ed.; VCH Publishers, Inc.: New York,

Houk, R. S.; Thompson, J. J. In Mass Spectrometry Reviews 7; Gross, H.
L., Ed.; Wiley: New York, 1988, pp 425-461.

Tanner, S. D.; Baranov, V. l.; Bandura, D. R. Spectrochimica Acta, Part B:
Atomic Spectroscopy 2002, 57, 1361-1452.

20

CHAPTER 2
EXPERIMENTAL MATERIALS AND METHODS
2.1. Paper Samples

Paper samples consisted of two brands of 20 lb. weight white document
paper (81/2 x 11 inches) made from 100% post-consumer waste which were acid-
free and Chlorine-free processed. The brands chosen were New Leaf Paper® -
Encore 100 (San Francisco, CA) and Staples® 100% Recycled Copy Paper (Item
#620016) (Framingham, MA). These brands were chosen because they were
readily available in the East Lansing, MI area.

Five reams (500 sheet units) were selected from each brand. The New
Leaf Paper® had the following markings on the ream wrappers for reams one
through ﬁve respectively: JD6041015, JC6041115, J06041015, JD6041115 and
JD6041015. It is suspected that these codes relate to batches or production time
frames. The company was contacted but the meaning of these codes was not
positively determined. New Leaf Paper® reams one through four were obtained
from different ofﬁces in the Chemistry building at Michigan State University and
ream ﬁve was obtained from Kinko’s Copy Center, East Lansing, MI. The
Staples® brand papers had no individual markings on ream wrappers and all of
these reams were obtained from Staples®, Lansing, MI. From each of the reams,
for both vendors, 3 sheets were selected, the top sheet, a sheet in the middle
and the bottom sheet of each ream. These 30 sheets of paper (3 sheets from 5

reams from each of 2 vendors) were used in this study.

21

From each sheet of paper, 5 samples were obtained, one from each
corner and one from the center of the page. Samples (approximately 23 mm x
18 mm) were cut from the paper using plastic scissors and handled using plastic
tweezers and gloves to prevent any elemental contamination. Each sample was
weighed on a Mettler H80 balance by ﬁrst weighing a quartz vessel, then the
paper sample was added and the total mass recorded. The sample mass was
determined as the difference between these masses and each sample weighed
approximately 0.03 g (0.029 1 0.001).

A numbering system was used to easily identify the origin of each
individual sample. The ﬁrst Ietter(s) in the numbering system referred to the
vendor, NL for New Leaf Paper® and S for Staples®. The ﬁrst number referred to
the ream (1-5) and the second number referred to the location of the sheet in the
ream with 1 indicating the top, 2 the middle, and 3 the bottom sheet in the ream.
The last number in the numbering system referred to the sampling position on
the sheet, 1 was the top left corner, 2 was the top right corner, 3 was the bottom
left corner, 4 was the bottom right corner, and 5 was the center of the page. For
example, the sample NL335 referred to a sample from New Leaf Paper®, ream 3,
bottom sheet cut from the center of the page. On some occasions replicates
were sampled from a Single digest of one paper sample and lower case letters
were used after the numbers to indicate these replicates.

2.2. Microwave Digestion
Microwave digestion is a sample preparation technique used for trace

metal analysis. This technique involves a high temperature, closed vessel acid

22

digestion that is able to shorten digestion times and use less acid than
conventional hot plate digestions. Because the digestion is done in a sealed
vessel, elevated pressures can be used with the elevated temperature to help
digest difﬁcult matrices.

Paper samples were digested using a microwave digestion unit (Ethos EX,
Milestone, lnc., Shelton, CT) equipped with an internal temperature probe.
Quartz liner vessels (Milestone, Inc.) were used to minimize contamination and
allow for smaller volumes of acids to be used. The procedure used for
microwave digestion was suggested by the manufacturer for the type of samples
used in this work. First the Clean quartz vessels were weighed then a paper
sample added and re-weighed to obtain an accurate sample mass. To the quartz
vessel, 1.5 mL of Optima grade nitric acid (69%) and 0.75 mL of hydrogen
peroxide was added. In the outer Teﬂon® vessel 11 mL of DI water and 1 mL of
hydrogen peroxide were added. The quartz vessel was placed inside the Teﬂon®
vessel (Milestone, Inc.) and capped with a quartz top. The Teﬂon® vessel was
then sealed with a Teflon® cap. This assembly was placed in a segmented rotor
and held in place with a screw tightened to the manufactures speciﬁcations using
a torque wrench. The sealed assembly can hold up to 100 bar of pressure and
retains volatile components produced during digestion.

Five vessels were loaded onto the rotor that rotates during digestion to
evenly heat all vessels. The temperature probe was placed in a reference vessel
to monitor temperature. The system automatically adjusted applied microwave

power (wattage) to obtain and maintain the desired temperature. The

23

temperature program used to digest the paper samples is shown in Table 2.1. A
maximum wattage is set as a safety precaution in case the temperature probe is
not working. The wattage set is based on the number of vessels loaded and the

desired maximum temperature.

Table 2.1. Microwave digestion temperature program

 

Step Time (min) Temperature Maximum Wattage
1 15 Ramp to 210 °C 300
2 10 Hold at 210 °C 300

 

 

 

 

 

 

 

 

Following digestion, the Teflon® vessels were allowed to cool to below 100 °C
before being opened in order to prevent loss of solvent. The quartz vessels were
then removed from the Teﬂon® vessels and allowed to cool to room temperature
with the quartz tops in place to prevent evaporative changes in volume. The
liquid in the Teﬂon® vessels was discarded.

The paper contained in the quartz vessel was partially digested; the
resulting solution was clear with a small amount of white particulate matter. The
particulate matter was Chemically characterized and is discussed in more detail in
Chapter 3. The solution initially consisted of 1.5 mL nitric acid (69%) and 0.75
mL hydrogen peroxide, during the digestion the hydrogen peroxide was broken
down to water and the resulting solution was 47% nitric acid.

The digested sample was further diluted to a nitric acid concentration of
2% for ICP-MS analysis by adding 429 pL of the decanted digestion solution to
10 mL of Ultra pure water. The diluted digest was stored in 15 mL polystyrene

conical vials (BD FalconTM, BD Biosciences, Franklin Lakes, NJ) that were

24

previously acid washed. The digests were stored in the vials for up to 4 weeks
prior to analysis.

The acid washing procedure for the conical vials was a 3 step process: 1)
vials were ﬁlled with 15% hydrochloric acid, soaked in a water bath at 45 °C for
24 hours, 2) acid was removed, vials were triple-rinsed and soaked in Ultra pure
water for 24 hours 3) vials were triple-rinsed in ultrapure water and allowed to
dry.

Procedural blank samples were run daily along with the paper digest
samples. The same procedure described above was used to prepare these
samples, except no paper was added to the quartz vessel. Over the course of
the digestions 11 procedural blank samples were created.

2.3. Particulate Characterization

The particulate matter was collected from the bottom of twenty vessels
from the standard nitric acid/hydrogen peroxide digestion. The collected
particulate matter was washed with ultrapure water and allowed to dry. The
collected particulates were used for spectroscopic analysis, aiming to
characterize the particulates.

2.3.1. Fourier Transform Infrared Spectroscopy (FTIR)

A potassium bromide (KBr) pellet was made of the dried particulates using
approximately a 1:100 dilution of the sample particles with KBr. The mixture was
pressed (Carver Laboratory Press, Fisher Scientiﬁc, Pittsburg, PA) and the

resulting pellet analyzed by FTIR.

25

An IR spectrum was collected on a Mattson Galaxy Series 3000 FTIR
(Madison, WI) in the transmittance mode. The sample was scanned from 350
cm‘1 to 4000 cm‘1 with the instrument resolution set at 4 cm'l. WinFirst software
(Mattson Instruments, Madison, WI) was used to record the spectrum.

2.3.2. X—ra y Diffraction Spectroscopy (XRD)

The particles were mounted onto a glass slide using double sided tape. A
spectrum was taken on a Rigaku Rotaﬂex RU 200B XRD instrument (Tokyo,
Japan). The x-ray beam was generated by accelerating electrons (45 W, 100
mA) at a rotating CU-anode. The Cu K1-a line was used and the instrument was
run in the 29/9 mode. Scans were performed at a speed of 1 degree per minute.
2.3.3. Dissolution/Hydroﬂuoric Acid Digestion

In order to Characterize the elemental make up of the particles, a complete
digestion was performed to be analyzed by ICP-MS. Microwave digestion was
again used to obtain a complete dissolution of the paper components but
hydroﬂuoric acid (HF) was added to the digestion matrix. Because HF can
dissolve glass and quartz vessels, only the Teﬂon® vessel was used without the
quartz insert. The Teﬂon® vessel is much larger than the quartz and requires a
minimum solution volume of 10 mL. In order to make sure approximately the
same elemental concentration was obtained using this procedure, compared to
that described in section 2.2, a paper sample, approximately four times larger,
was necessary.

The Teﬂon® vessels were initially weighed then a paper sample was

added and the vessel was reweighed. The sample mass was calculated from the

26

differences in weights and was approximately 0.12 9. To these vessels, 6 mL of
Optima grade nitric acid, 3 mL of hydrogen peroxide, and 1 mL of hydroﬂuoric
acid were added. The vessels were then sealed with the Teﬂon® tops, placed in
the segmented rotor, and tightened to the manufacturer’s speciﬁcations using a
torque wrench. The same temperature program described in Table 2.1 was used
for these samples. Following digestion a Clear solution with no particulate matter
remained in the Teﬂon® vessels.

The hydroﬂuoric acid was removed from the sample solutions before ICP-
MS analysis since HF is a highly corrosive acid that will shorten the lifetime of the
instrument. To remove HF, the samples were evaporated to dryness and
reconstituted to the proper dilution with nitric acid. From the digestion solution
427 pL was removed and placed in a small Teﬂon® container. These were
placed on a hot plate in a hood and allowed to evaporate over approximately one
hour. Following evaporation of all the solution, 10 mL of 2% nitric acid (Optima
grade) was added to reconstitute the sample. This reconstituted sample was
used for the ICP-MS analysis.

2.4. Inductively Coupled Plasma Mass Spectrometry (ICP-MS)

Samples were analyzed on a Micromass Platform quadrupole ICP-MS
(Thermo Electron Corp, Waltham, MA) with hexapole collision cell using a
CECTAC ASX-500 autosampler (Omaha, NE). Tune conditions were optimized
using a 10 ppb solution of Be, Co, In, Ce, Bi, and U prepared in a 2% nitric acid

solution. Nebulizer (Ar gas) ﬂow rate, was 0.71 - 0.73 L/min and hexapole

27

helium and hydrogen gas ﬂow rates ranged from 1.8 - 4.0 mL/min. Samples
were scanned for 2.75 min using a dwell time of 0.1 S for each element.

Paper digests were prepared in a 2% nitric acid solution as described
above. Each sample was mixed 2:1 with an internal standard solution containing
20 ppb of Bi and In. The instrument response was corrected using 115In as an
internal standard for 24Mg, 27Al, 45Sc, 55Mn, 56Fe, 662m 88Sr, 89Y, 90Zr, 120Sn,
138Ba, 140Ce and 146Nd while 209Bi was used for 208Pb, 232Th and 238U. Final
sample concentrations were quantiﬁed using a set of multi-element external
calibration standards prepared in 2% nitric acid solUtion. Three ranges of
calibration standards were used because the elements had a wide range of
concentrations. The ranges were 5 to 1000 ppb for Mg, Al and Fe, 0.1 to 100
ppb for SC, Mn, Zn, Sr, Zr, Sn, Ba, Pb and U, and 0.05 to 10 ppb forY and Nd.

The autosampler probe was ﬂushed with 2% nitric acid, for 90 seconds,
between every sample, “blank” samples (pure 2% nitric acid) were run between
every 10 samples, and the calibration standards were run approximately every
20 samples. To correct for any instrument drift, new calibration curves were
generated after each calibration standard run. Each calibration curve was
generated using two sets of standards, the standards run just before and just
after the set of samples. Both sets of data were plotted on the same curve and
the line generated was representative of both sets of standards. The
concentrations for the paper digest samples were calculated using the calibration
curve generated using calibration standards run immediately before and after the

sample. A single digest sample was run in triplicate and yielded concentrations

28

that had less than 3% relative standard deviation, indicating that the
concentrations determined by ICP-MS were consistent run-to-run.
2.5. Chemicals

All solutions were prepared using ultra high purity water (> 17 MQ—cm)
from a Barnstead E-pure or Millipore Milli-Q system. Nitric acid (69%) was
obtained from Fischer Chemical (Pittsburg, PA) and was Optima grade (low ppt
levels of metal contaminants) to reduce trace element contamination of the
sample. Hydrogen peroxide (30%) was obtained from J.T. Baker (Phillipsburg,
NJ) and was ACS reagent grade and used without additional puriﬁcation.
Hydroﬂuoric acid (48%) was obtained from CCI (Columbus Chemical Industries,
Columbus, WI) and used without additional puriﬁcation. Potassium bromide was
IR spectral grade, stored dry, was Obtained from Spectrum Chemical (Gardena,
CA).

Elemental standards were ICP standard grade (Spex Certiprep,
Metuchen, NJ) and were premixed in 1000 pg/L concentration in a 2% Optima
nitric acid solution. One multielement standard was used for Ce, Nd, and Y but
all other standards were single element solutions. These standards were used
without any additional puriﬁcation.

Ar used for ICP-MS was reagent grade and obtained from Linde Gas
(Independence, OH).

2.6. Statistical Analysis
Prior to any statistical analysis, the elemental concentrations obtained

from the ICP-MS were corrected for dilution factors and normalized to the paper

29

weight measured prior to digestion giving the concentration in pg element/g
papen

The average elemental concentration was calculated for each sheet (n=5)
along with relative standard deviation (RSD). Q-tests were performed on all data
sets that had RSDs higher than 15%, removing the data points that were
determined statistically to be outliers (conﬁdence limit of 95%). After the q-test
the relative RSDs of elemental concentrations for each Sheet was re-calculated.
Any elements with an RSD greater than 15%, for any of the sheets tested were
not used in any further analysis.

At this point, the samples are grouped by vendor and are treated as two
separate groups. Two-way analysis of variance (ANOVA) was applied to the
average elemental concentration for each sheet tested. The two variables in
ANOVA were sheets within a ream (n=3) and the reams (n=5). ANOVA was
used to determine which elements showed signiﬁcant differences within a ream
and between reams. To investigate these differences further, Tukey’s HSD
(honestly signiﬁcant difference) test was performed to determine which reams
differed from each other. This test has also been used by Almirall and co-
workers for his work on forensic glass samples also analyzed by ICP-MS." 2

Tukey’s HSD completes painrvise comparisons of all the data sets. The

equation (2.1) is shown below.

 

t,= ‘ I (2.1)

 

3O

where M - M; is the difference between the ith and jth means, MSE is the mean
square error and nh is harmonic mean (equation 2.2) of sample sizes of groups i
and j. If the sample sizes are even as they are in this case nh is equal to the

number of items per group.
__ 2’11”;
- ni + n

 

nh (2.2)

I

where n; is the number of samples in group i and n] is the number of samples in
group j. The Tukey statistic (ts) is compared to a table of values based on the
number of samples and degrees of freedom to determine the level of statistical
signiﬁcance of the difference between the two values.

To determine signiﬁcant differences between the elemental concentrations
between the two vendors, a t-test analysis was performed. The average
concentration of each element was calculated for all of the paper samples from
the vendor (n=75). This was then compared to the average for the other vendor.
An F-test was performed for each element to compare the variance of the
elemental concentrations for each vendor (conﬁdence limit 95%). If the variances
were statistically different the degrees of freedom (df) for the t-test comparison

were calculated with equation 2.3.3

df = n: n] _ 2 (2.3)

Where Sf is the variance of group i, ni is the number of data points in group i and

Similar for group j.

31

The t-statistic was calculated with equation 2.4.

XI

t= ”xi
2
.9.
”I

.25
‘+

(D

23

(2.4)

Where R, is the mean of group i and similar for group j. The t-statistic was

compared to tables of values at the 95% and 99% conﬁdence limits (two-tail)

using 8 (n-2, n=10) or the value calculated from equation 2.3 for the degrees of

freedom.

All statistical analyses were performed with either OriginPro for windows

7.5 SR 4 (v7.5853, OriginLab Corp., Northampton, MA) or Excel 2000 (Microsoft®

Corp.)

32

2.7. References

(1) Trejos, T.; Montero, S.; Almirall, J. R. Analytical and Bioanalytical
Chemistry 2003, 376, 1255-1264.

(2) Montero, S.; Hobbs, A. L.; French, T.‘ A.; Almirall, J. R. Journal of Forensic
Sciences 2003, 48, 1-7.

(3) Harris, D. C. Quantitative ChemicalAna/ysis, 6th ed.; W. H. Freeman and
Company: New York, 2003.

33

CHAPTER 3
PARTICULATE CHARACTERIZATION
3.1. Introduction

With a standard microwave digestion matrix Of HN03 and H202, the paper
samples were not completely digested. There was a white particulate matter
remaining that settled to the bottom of the quartz vessels. During manufacture
several inorganic additives (e.g., clay, titanium dioxide, mica, talc) are added to
paper to change the physical properties such as the brightness or texture of the
paper. These inorganic additives are likely the portion of the paper sample that
did not break down during the standard digestion procedure. Based on what is
known about the additives in paper, it was suspected that the particulate matter
include kaolin and TiOz.

Kaolin is an aluminosilicate clay material (A128i205(OH)4) that is commonly
used as a filler material for paper to give it color, opacity and good printing
quality. About one half of the kaolin mined today is used in paper making
processes.1 Ti02 is an inorganic material that is commonly used as a white
pigment in a wide variety of materials such as paint, cosmetics, and toothpaste,
which is also added to papers to enhance their whiteness.2

Kaolin and Ti02 are likely to be found in paper samples but are not
expected to be digested just with HN03 and H202. Two types of spectroscopy,
infrared and x-ray diffraction were used to determine the composition of the
undigested particulate matter. These two types of spectroscopy were Chosen

because they can give information about the structure of the material and identify

34

the materials present. Additional digestions and analysis with ICP-MS were
conducted to determine the effect of the particulate matter on the composition
and reproducibility of the digestion solution.
3.2. Particulate Identiﬁcation
3.2.1. Fourier Transform Infrared Spectroscopy (FTIR)

A KBr pellet was prepared from the particulate material and an IR

spectrum was collected (Figure 3.1).
80 a
75 -

70-

Tio2 <2

65-

Kaolin <
TiO2

60-

% Transmittance

 

55-

 

 

 

50 I I r ﬁr I I F —I

3850 3350 2850 2350 1850 1350 850 350

Wavenumbers (cm'l)

Figure 3.1. FTIR spectrum of particulates from paper digestion

The spectrum was compared to reference IR spectra of kaolin (clay). TiOZ,
and silicon dioxide (8102); possible materials that would be expected in paper
samples. 8102 is a ﬁller material used in papers that also will not dissolve in the

HN03/H202 digestion. The absorption frequencies from the particulate sample

35

and the reference materials are listed in Table 3.1. From the table it is Clear that
the IR spectrum of the particulate sample is a combination of both kaolin and
TiO2. The strongest peaks are from the kaolin and the T102 peaks are not as
strong, suggesting the particulate sample contains a larger portion of kaolin.

Table 3.1 Comparison of absorption frequencies in IR spectra of particulates and
reference materials

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Kaolin (clay)3 Titanium Silicon Dioxide5
(cm'l) (cm'l) Dioxide4 (cm'l) (cm")
3695 3690 - -
3619 3620 - -
3414 - 3400 3425
2919 - 2915 -
2850 - 2840 -
1636 - - 1620
- - 1455 -
1384 - 1375 -
1 101 1 100 - 1 100
1031 1030 - -

- 1005 - -
913 910 -
669 695 800 - 400 (broad -
538 535 peak) -
468 460 475

 

There is a strong peak in the sample IR spectrum at 1384 cm'1 that could
possibly be from the TiO2 but it is much strong than the other peaks. This peak is
in the range of a normal vibration of an inorganic nitrite.6 Since the paper
samples were digested in HN03 it is possible that an insoluble inorganic nitrite
species was formed, and absorbs at 1384 cm'1.

3.2.2. X-ray Diffraction Spectroscopy
X-ray diffraction spectroscopy is based on the principle of diffraction and is

commonly used to identify crystalline materials. In order for diffraction to occur

36

the Spacing of scattering centers must be similar to the wavelength (A) of the light
being scattered. The spacing between atoms in crystal structures is
approximately the same as the wavelength of x-rays, in the range of A.

When an incident beam of x-rays bombard an atom they can be elastically
scattered (i.e., maintains same A as incident beam). When several atoms are
close to one another, such as in a crystal lattice, theyx-ray waves scattered from
the different atoms can interfere either constructively or destructively.
Constructive interference occurs when the lattice spacing is an integer multiple of
the wavelength of the incident radiation as described by Bragg’s law (equation
3.1),

nA=2d°sin9 (3.1)
where n is an integer, A is the wavelength of incident light, d is the distance
between scatters (lattice spacing), and 9 is the angle of refraction. Based on this
equation, if the wavelength and the scattering angle are known then the lattice
spacing can be determined.

In an x-ray diffraction experiment a single wavelength is focused onto the
sample and the x-ray scattering is measured at a range of angles. The data is
plotted as x-ray intensity versus angle. The angles that have x-ray scattering
intensity above the background result in a peak. From the angle where a peak
occurs the lattice spacing (d-spacing) can be determined using equation 3.1.
Every crystalline material has a speciﬁc lattice structure in which the atoms are in
a certain conﬁguration (e.g., face centered cubic) and there are several distances

within a single lattice (e.g., between atoms in the same plane, atoms in different

37

planes). The pattern of d-Spacing and the relative intensity of these peaks are
Characteristic to the crystal and can be used to identify the material. If the
sample is a mixture of multiple materials the intensity of the x-ray scattering is
related to both the concentration as well as how strongly the material scatters.
The particulates that were collected from the HNOg/HzOz digestions were
mounted on a glass slide. The sample was scanned between 2 and 65° at a rate

of 1°/min and the resulting XRD spectrum is shown in Figure 3.2.

J 3.5148 A

X-ray Counts (104)
N

1.6996 A 1.6655 A

 
      

2.4288 A 23307 A

 

 

 

0 10 20 30 40 50 60 70
29 (°)
Figure 3.2. X-ray diffraction spectrum of undigested material from the
HNOa/HzOz digestion.

The peaks expected for TiO2 and their relative intensities are shown in Table 3.2,

all of the peaks visible in Figure 3.2 are consistent with TiO2.7

38

Table 3.2. X-ray diffraction peaks for TiO27

 

 

 

 

 

 

 

 

 

d (A) IrI1

3.52 100
2.431 10
2.378 20
2.332 10
1.892 35
1.5999 20
1.5555 20
1.4930 4
1.4808 14

 

 

 

 

 

There is one small peak correlating to a lattice spacing of 7.187 A that can be
seen in the expanded spectrum (Figure 3.3). From the FTIR spectroscopy it was
apparent that some component 'of this mixture is kaolin. The kaolinite group of
Clays has a lattice spacing of 7.15 — 7.20 A.8 The peak in the sample spectrum is
in the middle of this range at 7.187 A and therefore it was determined the kaolin
was the likely source of this peak. Because clay materials are not perfectly
crystalline, relatively weak signals are observed in x-ray diffraction.8 This is
apparent in this case based on the relative intensity of the kaolin peak to the TiO2
peaks, which is in contrast to the stronger kaolin signals observed in the FTIR
spectrum. Neither XRD nor FTIR was used to quantify the TiO2 and kaolin

concentration in the undigested particulate matter.

39

X-ray Counts (102)

1 ‘1 7.187A

 

 

T T 1 I I

0 5 10 o 15 20 25
Figure 3.3. Kaolin peak in the x-ray diffigétion spectrum of undigested material
from the HNO3/H2O2 digestion.

3.2.3. Elemental Analysis of Hydroﬂuoric Acid (HF) Digest

In order to determine the concentration of TiO2 and kaolin in the
particulate material, paper samples were completely digested with the addition of
HF to the digestion matrix, as described in section 2.3.3. Levels of Ti, Al and Si
levels were determined by ICP-MS for both digestion procedures: HNO3/l'l202
(particulates present) and HF (no particulates present). The same sheet of paper
was used for both digestions therefore it was assumed that any difference in the
concentrations of these elements was due to the material in the particulate

matter. The concentration of Al, Si and Ti under both digestion conditions are

presented in Table 3.3.

40

Table 3.3. Comparison of Al, Si and Ti concentrations, determined by ICP-MS, in
solutions from HNO3/H2O2 and HF digestions.

 

 

 

 

Al (.19ng Si 1 Ti 1
HNO3IH2O2
Digest 380 d: 40 800 i 100 50 :l: 50
(n=4l
HF
Digest 1020 :l: 40 0 179 i 3
(n=3)

 

 

 

 

 

There was a difference between the concentrations determined with each
of these digestion methods. The Al and Ti were about three times higher in the
HF digest, as expected, however the Si was no longer detectable in the HF digest
solution. An HF digest cannot be use to quantify Si because during the digestion
the Si and F will form a volatile complex SiF4.9‘11 The boiling point of this
compound is -86 °C hence the Si is lost from the digestion solution prior to
quantification.12

Assuming the additional Al present in the HF digest is from kaolin the
approximate concentration is 3100 pg kaolin per 9 paper. A similar calculation for
TIO2 yields a concentration of approximately 220 pg TiO2 per 9 paper. Based on
these rough calculations kaolin is the major component of the particulate matter,
conﬁrming results obtained from FTIR.
3.3. Effect of Particulate on Elemental Analysis
3.3.1. Comparison of Nitric Acid/Hydrogen Peroxide and Hydroﬂuoric Acid
Digestion

In order to determine if complete dissolution of the particulate matter had

any signiﬁcant effects on the element concentrations being determined, a

41

comparison was made between the two digestion methods. Four paper samples
were digested using HNO3/H202 and three were digested using HF; all samples
were from the same sheet. All digests were then analyzed by ICP-MS, using the
same calibration curve, to quantify elemental concentrations. The results are
shown in Table 3.4. Signiﬁcant differences between the digests, based on a t-
test with a 95% confidence limit, were observed in the concentrations of Mg, Al,
Fe and Pb.

Table 3.4 Comparison of elemental concentrations in solutions from HNO3/H202
' and HF digestions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HNO3IH202 HF Signiﬁcant
Digest (n=4) Digest (n=3) Difference
(pg/9) lug/9) (95% CL)

Mg 1200 j: 100 1460 :I: 60 yes
Al 390 :I: 40 1200 j: 300 yes
Mn 4.7 :I: 0.4 6.0 :t 0.1 no
Fe 100110 129:3 yes
Zn 12 i 2 15 :I: 5 no
Sr 38 i 3 43 :I: 1 no
Y 0.14 :I: 0.02 0.16 i 0.02 no
Ba 1211 13.9:I:0.6 no
Ce 0.34 :I: 0.01 0.7 :I: 0.2 no
Pb 0.11 :I: 0.01 1.21 :I: 0.03 yes
Nd 0.19 :I: 0.01 0.22 :I: 0.03 no

 

The increase in concentration of Al was expected, as it was determined a
large portion of the particulate sample consisted of an aluminosiliate Clay. Other
sheet silicate minerals are known to contain Mg (e.g., serpentine
(Mg3Sl205(OH)4)) and Fe (e.g., mica (K(Mg,Fe)3(AlSi3Oto)(OH)2)) and could be
minor components of the Clay as well. There was variation in Pb concentration

across a single sheet of paper (RSDS 20 - 100%) but the increase in the HF

42

digest is 1000% so it is more likely that the Pb was a contaminant present in the
HF used. No procedural blank was done with the HF solution so this cannot be
determined for sure.
3.3.2. Effect of Filtering HNOg/HzOz Digest

Following digestion in HN03/H2O2, a small portion of the digestion solution
was decanted to be diluted for ICP-MS analysis. It is possible that some small
particulates were also sampled and could affect the elemental concentrations and
the reproducibility of the analysis. To test this hypothesis, two digested samples
from the same sheet of paper were selected. From each of these samples, one
aliquot was ﬁltered and a second was not. A new syringe filter (Millex-Ha, 0.45
pm, Millipore) and an acid wash syringe were used to ﬁlter each of the samples.
Both the ﬁltered and unﬁltered solutions were then analyzed by ICP-MS and the
elemental concentrations normalized to the mass of the sample. The results are
shown in Table 3.5.

Table 3.5. Comparison of elemental concentrations for ﬁltered and unﬁltered
samples. Errors represent standard deviations.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Unﬁltered Filtered Signiﬁcant
(n=2) (n=2) Difference
(11919) (Hg/g) (95% CL)

Mg 1100170 1130130 no
AI 670 1: 70 680 i 60 no
Mn 4.42 i 0.09 4.5 1: 0.2 no
Fe 90 1: 4 92 i 6 no
Zn 13 :I: 6 1000 1: 1000 yes
Sr 30.5 1: 0.2 31.2 1: 0.5 no
Y 0.13 i 0.01 0.13 i 0.01 no
Ba 6.49 :I: 0.04 22 1: 10 no
Ce 0.34 1: 0.02 0.34 :t 0.02 no
Pb 0.08 i 0.01 0.15 i 0.05 no
Nd 0.18 i 0.02 0.18 1: 0.01 no

 

 

 

 

 

43

From the results it was clear that ﬁltering the digest did not signiﬁcantly
affect the concentrations of the elements of interest with the exception of Zn,
which showed a signiﬁcant increase. There may be Zn present in the filters or
syringes that is washed into the solutions when the 2% HNO3 matrix passes
through the filter causing the increase Zn concentration in the filtered samples.
This particular filter material was not rated for trace metal analysis. Because
none of the other elemental concentrations were affected by the ﬁltering
procedure and to avoid the possibility of sample contamination decanting of the
digestion liquid was determined to be sufﬁcient.

3.3.3. Elemental Absorption by Particulates

One other concern of having undigested particulates is that the
particulates could have absorbed some of the elements in the digestion sOIution
and subsequently altered the concentrations determined. The particulate matter
did contain clay material that could absorb cations. In soil analysis a cation
exchange capacity (CEC) is used to gauge how well the soil will hold nutrients
(e.g., Ca”, Mg”, K‘). Since this property is a measure of how well the material
will absorb cations in this application CEC was related to how much effect the
Clay had on the digestion solution.

The previously described studies determined that the type of clay present
in the papers is kaolin. Kaolin has a relatively low CEC of 3.0 - 3.7 meq/100 g
Clay, where meq is mol M"“/n.13 The average CEC of the Clay Mineral Society
source clays is 61 meq/100 9.13 While kaolin does not have very strong

exchange capability for cations it is still possible that it could absorb some of the

44

elements in the solution. The amount of kaolin that was present is relatively
small in regard to the weight of the paper samples; based on estimations of the
concentration of kaolin (3100 pglg paper) it makes up less than 1% of the paper
by weight. Even if the kaolin does absorb some of the elements from the
digestion solution, if the system has reached equilibrium, the concentration of the
elements in the digestion solution will be relatively constant with time.

To determine if there was a time dependent component to the elemental
concentrations, a single sheet of paper was sampled and the digestion solution
remained in contact with the particulates for various amounts of time before
decanting for ICP-MS analysis. Because the digestions are done at an elevated
temperature in a sealed vessel the minimum amount of time was 20 minutes to
allow the solution to cool below the boiling point before opening the vessel. The
results of the time study ICP-MS quantiﬁcations are shown in Table 3.6, based on
elemental concentrations determined by ICP-MS.

The values for each element do not appear to signiﬁcantly Change with
time: there is no pattern of decreasing or increasing concentration with time. It
seems that by the ﬁrst time point (20 min) the particulate matter has reached
equilibrium with the elements in the solution and no changes are seen with time.

The particulate matter does not affect the elemental concentration in the

digestion solution.

45

Table 3.6. Elemental concentration paper samples digested with HN03/H2O2 and
allowed to equilibrate in the presence of the particulates for various amounts Of

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

time.
Equilibration Time
20 min 60 min 120 min 180 min

Mg (pglg) 791 861 800 848

Al (pglg) 431 328 370 373
Mn (pglg) 3.84 4.12 3.94 4.18
Fe (pglg) 70.1 74.3 75.2 77.9
, Zn (pglg) 8.00 4.12 9.31 5.07
Sr (pglg) 63.0 69.6 63.6 67.2
Y (pglg) 0.094 0.094 0.148 0.099
Bajpglg) 1.99 2.18 2.20 2.12
Ce (pglg) 0.244 0.263 0.259 0.258
Pb (pglg) 0.656 0.752 0.721 0.674
Sn (pglg) 8.39 9.25 8.44 9.67
Zr (pglg; 1.46 1.92 1.55 1.72
Ndeglg) 0.131 0.132 0.129 0.139

 

 

 

 

 

 

 

3.4. Conclusions

With a digestion solution that contained only HN03 and H202, complete
digestion of paper samples was not possible. The undigested portion was
Characterized as a mixture of kaolin (Clay) and Ti02 by FTIR and XRD. The
presence of this particulate matter does not appear to have an effect on the
reproducibility of the concentrations determined in the digest, as determined by
comparing the HN03/H2O2 digestion (particulates) with an HF digestion (no
particulates). HF is a very corrosive acid; it is hazardous to work with, requires a
larger paper sample, and an additional evaporation step prior to ICP-MS
analysis. Since an HF digest is not necessary the elemental concentrations used
for comparisons for the rest of this work were determined using the HN03/H2O2

digestion.

46

3.5. References

(1)

(10)

(11)

(12)

(13)

Moore, D. M.; Reynolds Jr., R. C. X-Ray Diffraction and the Identiﬁcation
and Analysis of Clay Minerals, 2nd ed.; Oxford University Press: Oxford,
1997.

Smook, G. A. Handbook for Pulp & Paper Technologists, 3rd ed.; Angus
Wilde Publications Inc.: Vancouver, 2002.

In The Sadtler Infrared Spectra Handbook of Minerals and Clays; Ferraro,
J. R., Ed.; Sadtler Research Laboratories: Philadelphia, PA, 1982, pp 354.

In The Sadtler Infrared Spectra Handbook of Minerals and Clays; Ferraro,
J. R., Ed.; Sadtler Research Laboratories: Philadelphia, PA, 1982, pp 400.

In The Sadtler Infrared Spectra Handook of Minerals and Clays; Ferraro,
J. R., Ed.; Sadtler Research Laboratories: Philadelphia, PA, 1982, pp 399.

Nakamota, K. Infrared and Raman Spectra of Inorganic and Coordination
Compounds Part B: Applications in Coordination, Organometallic, and
Bioinorganic Chemistry, 5th ed.; John Wiley 8. Sons Inc.: New York, 1997.

In Powder Diffraction File; McClune, W. F., Ed.; JCPDS International
Center for Diffraction Data: Swathmore, PA, 1980; Vol. 21 to 22, pp 21-
1272.

Brown, G.; Brindley, G. W. In Crystal Structures of Clay Minerals and their
X-ray Identiﬁcation; Brown, G., Ed.; Mineralogical Society: London, 1980.

Jarvis, I. In Handbook of Inductively Coupled Plasma Mass Spectrometry,
Houk, R. 8., Ed.; Chapman and Hall: New York, 1992, pp 172-224.

Taylor, V. F.; Toms, A.; Longerich, H. P. Analytical and Bioanalytical
Chemistry 2002, 372, 360-365.

Pruseth, K. L.; Yadav, S.; Mehta, P.; Pandey, D.; Tripathi, J. K. Current
Science 2005, 89, 1668-1671.

Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 87th ed.;
Taylor and Francis: Boca Raton, FL, 2007.

Borden, D.; Giese, R. F. Clays and Clay Minerals 2001, 49, 444-445.

47

CHAPTER 4
RESULTS AND STATISITICAL COMPARISONS
4.1. Introduction

Paper samples from 2 vendors (New Leaf Paper® and Staples®), 5 reams
each, were collected to compare the trace elemental content for differences.
From each of the reams, 3 sheets were selected from the top, middle, and
bottom. On each sheet 5 samples were collected for analysis. The paper
samples were digested in a solution of nitric acid and hydrogen peroxide using a
microwave system. Following digestion, they were analyzed using inductively
coupled plasma mass spectrometry to determine trace elemental concentrations.
Several comparisons were made; within sheet, between sheet, between ream
and between vendors.
4.2. Method Development

The ultimate goal of the project was to be able to differentiate papers.
Therefore it was necessary to determine if signiﬁcant differences existed in the
elemental concentrations of reams of the same type of paper. First a suite of
elements to compare had to be determined. This suite of elements was based on
the elements that were present above background levels, were not present in the
background, and were a consistent concentration throughout the ream. For the
comparison to be useful, the elemental concentrations must not be signiﬁcantly

different between Sheets from the same ream.

48

4.2.1. Full Mass Scan

Initially a list of potential elements in the papers was developed from a
literature survey."7 This list included: 7Li, 23Na, 24Mg, 27AI, 45Se, 55Mn, 52Cr, 55Rh,
56Fe, 5900’ 63CU, 662“, 88ST, 89Y, 902T, 107Ag 114Cd, 1218b, 13883, 139La, 14009,
146Nd, 208Pb, 232Th, and 238U. A full mass scan was measured for a digest from a
single sheet of paper (New Leaf Paper®), scanning the range of m/z values,
between 7 and 240, to determine if all the elements listed above were present
and if there were any additional elements to include in the analysis. Figure 4.1
shows the full range of masses, this mass spectrum is dominated at the lower
m/z values by smaller atomic weight atoms such as C, Na, Ca and O as well as
Ar, H and He from the source and hexapole gases and other interferences.
Figures 4.2 through 4.4 show expanded portions of the mass spectrum to identify
the other elements present that may be useful for comparisons.

Based on the full mass scan some elements were eliminated from the list
(52Cr, 55Rb, 59Co, 63Cu, 1O7Ag, 114Cd, 121Sb, and 139La) because the concentrations
were below the instrument detection limit. 7Li and 23Na were eliminated because
the instrument has high levels of contamination at these masses. Additional
elements were observed (12°Sn and 184W) in the paper digest and were added to
the list.

The full mass spectrum as well as previous work on the elemental
analysis of document papers was used to develop a list of elements to quantify in

the papers. The final list included Mg, AI, Sc, Mn, Fe, Zn, Sr, Y, Zr, Sn, Ba, Ce,

49

Nd, Pb, and U. One of the elements present in the full mass scan (W) was

initially overlooked and was not included in the suite of elements to be tested.

 

 

 

Relative Intensity (%)

 

 

 

5*3Fe 88Sr
AI ‘LI I J TI l AT I I T 1—'
20 40 60 80 100 120 140 160 180 200 220 240

mlz

Figure 4.1. Full mass scan of New Leaf Paper® (NL131) digest.

24Mg

Relative Intensity (%)

 

I I I I I

20 22 24 26 28 30

Figure 4.2. Expanded region (m/z 20 — 30) of full mass scan for New Leaf
Paper® (NL131) digest.

5‘5Fe

88Sr

Relative Intensity (%)

1388a

146
Mg m 9027 120?” 14°C Nd
n e /
Vw—{W I I I A / Mm /

50 60 70 80 90 100 110 120 130 140 150
mlz

Figure 4.3. Expanded region (m/z 50 — 150) of full mass scan for New Leaf
Paper® (NL131) digest.

U'l
Ul

 

 

 

 

 

 

208Pb

1 84W

1
i . 2 38
Wiiliwrtmwjiil i M

150 160 170 180 190 200 210 220 230 240
mlz

 

 

Relative Intensity (%)

 

 

 

Figure 4.4. Expanded region (m/z 150 - 240) of full mass scan for New Leaf
Paper® (NL131) digest.

51

4.2.2. Calibration

An internal standard (20 ppb Bi and In) solution was added to all samples
prior to analysis. All signals were ratioed to either Bi or In signal, depending on
mass, to correct for any ﬂuctuation in instrument response. Multi-element
calibration standards were prepared containing all of the elements listed above.
Because these elements existed in the paper samples over a wide range of
concentrations three different ranges were used: 5 - 1000 ppb (Mg, Al, Fe), 0.1 —
100 ppb (SC, Mn, Zn, Sr, Zr, Sn, Ba, Pb, U) and 0.05 — 10 ppb (Y, Ce, Nd).

Calibration curves were generated for each element based on the signals
ratioed to the internal standard. All of the curves were linear and the limits of

quantitation for each element are listed in Table 4.1

Table 4.1. Instrument quantitation limits for elements chosen for paper
analysis. Sc was not linear so no detection limit is listed

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Element Quantitation Limit (ppb)
24M 9 <5
27AI 10
458C _
“Mn <0.1
56Fe <5
66Zn 0.25
88Sr 0.25
89v <0.95
90Zr <0.1
“'20Sn <0.01
138Ba 0.25
140Ce <0.05
”6Nd <0.05
208Pb <0.1
238U <0.1

 

 

One element, Sc, did not have a linear calibration curve therefore no

detection limit is listed. For many of the elements the calibration curve was linear

52

over the entire range of standards tested so a true quantitation limit could not be
calculated and is listed as less than the concentration of the lowest elemental
standard analyzed. Because Sc did not have a linear calibration curve it was
eliminated from the suite of elements.
4.2.3. Procedural Blanks

Due to the various steps involved in the digestion and analysis procedures
as well as the sensitivity of the ICP-MS, some of the elements in the proﬁle could
have been environmental contaminants. To determine possible contaminants,
procedural blanks were run along with the samples. These samples were
handled the same as the paper digests except no paper was added to the quartz
vessels prior to the digestion step. A blank sample was generated every day
samples were digested giving a total of 11 blank samples.

The procedural blanks were analyzed along with the paper samples and
the elements from the suite of elements identiﬁed above were quantiﬁed with a
calibration curve. Shown in Table 4.2 is the average concentration of each
element in the 11 blanks along with the lowest concentration of the element found
in any of the 150 paper samples analyzed (2 vendors, 5 reams each, 3 sheets
per ream, and 5 samples from each sheet). In the last column of Table 4.2 the
ratio of the blank to the lowest paper concentration is reported as a percentage.
This percentage represents the maximum contribution the background level of an

element will have on the concentration determined in a digest sample.

53

 

Table 4.2. Elemental concentration of procedural blanks, compared to lowest
concentration of the element in any paper digest solutions. ND indicates
concentration was below the detection limit

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Element Blank Value (ppb) Lowest Value in Ratio of Blank to
(n=11) Paper (ppb) Lowest Paper (%)
Mg 1.3 1: 0.7 313 0%
Al 1.0 i 0.8 131 1%
Mn 0.02 i 0.01 1.69 1%
Fe 1.0 :I: 0.8 30.1 3%
Zn 3 i 2 1.97 138%
Sr 0.1 1: 0.1 12.6 1%
Y ND 0.04 -
Ba 0.03 i 0.02 0.94 3%
Ce ND 0.11 -
Pb 0.03 1: 0.04 0.05 69%
Sn 5.0 1: 0.2 3.84 129%
Zr 0.09 i 0.04 0.19 46%
Nd 0.01 i 0.00 0.05 20%

 

 

 

 

 

For some of these elements (i.e., Zn, Pb, Sn, and Zr) a signiﬁcant portion
of the signal is coming from a procedural contamination source, as indicated by
the high ratio listed in Table 4.2. Because the background levels could be highly
variable and were not representative of the Characteristics of the papers, these
elements were not used in any future comparisons.

4.2.4. Relative Standard Deviations

The last criterion used to Choose elements for statistical comparison
among paper samples was the variability in elemental concentration across a
single sheet. Because only a portion of the sheet would be sampled in a forensic
analysis scheme it is important that no significant differences occur within the
Single sheet so that the sample analyzed is representative of the whole sheet.

For each sheet of paper, ﬁve samples were analyzed, one from each corner and

54

one from the center. These areas were Chosen to detect Spatial variations in the
elements across the page.

Relative standard deviations (RSD) were calculated for each element
concentration on a sheet; this was done for all sheets that were sampled. Any
element concentration with a RSD larger than 15% was checked for outliers
using a q-test. If the RSD was still larger than 15%, after removing any outliers,
the element was not considered useful. An RSD Of > 15% indicates an elemental
concentration that is variable over a single sheet, which Is not desirable for
forensic analysis.

In general, the RSD for an elemental concentration within a sheet was 5 —
10%, highlighting the acceptable precision in paper, digestion procedure and
ICP-MS analysis. Zn had very high variation across the sheets with RSDS of 20
— 50%. Pb and Zr also had high variation within a sheet with RSD values of up to
130% for Pb and around 25% for Zr. The elements with high RSD values (Zn,
Pb, and Zr) also had a high contribution from the blank samples. This suggests
that the high RSD values are due to varying levels of contamination. Based on
the high RSD values as well as the high blank levels, these elements were not
used for any statistical comparisons.

The instrumental detection limits, calibration curves, and procedural blank
levels were used to narrow the suite of elements for statistical analysis to include:
Mg, Al, Mn, Fe, Sr, Y, Ba, Ce and Nd.

It is unclear where these elements come from exactly however Mg and Mn

are known micronutrients in plants (i.e., tree cells) and the variation in these

55

elements could come from variation in the amount present in the trees used for
the pulp.8 Sr is similar to Ca and can be found in cells in place of Ca. Al is a
component of some of the ﬁller materials (e.g., clay) however this portion of the
paper does not digest in the HN03/H2O2 digest used for these measurements
and so the origin of the Al that was freed during digestion is unknown. Calcium
carbonate (C3003) is another common ﬁller material for paper9 and commercial
supplies of this Chemical can have chemical impurities, Sigma-Aldrich listed
potential impurities to include Mg, Fe, and Ba, all are trace elemental
components found in the paper digests.1o Because both the vendors sampled in
this work used recycled materials it is impossible to be certain of the source of
the trace elements.
4.3. Variation Between Reams

In this section the variation of elemental concentration between the reams
of paper is discussed. For simplicity the reams from each vendor, New Leaf
Paper® and Staples®, are discussed individually.
4.3.1. New Leaf Paper®

The ﬁrst analysis performed was a two-way analysis of variance (ANOVA)
based on the elemental concentrations of the sheets from five different reams.
The two variables tested were the variance among sheets (average of 5 samples
from each sheet) and the variance between the reams (average of 15 samples, 3
sheets with 5 samples each). The null hypotheses were that there was no
difference in the means of the elemental concentrations of the sheets and that

there was no difference in the means of the elemental concentrations of the

56

reams. The alternate hypotheses were the means of the elemental
Concentrations of the sheets and reams were not equal. The test generated a p-
value, which is a statistical measure of the probability that the null hypotheses
were true. A p-value S 0.05 was Chosen to deﬁne a signiﬁcant difference
between elemental concentrations, at this p-value the probability that the means
Of the elemental concentrations were the same was 5% or less. Shown in Table
4.3 are the resulting p-values from the two-way ANOVA analysis on the New Leaf
Paper®.
Table 4.3. Two-way ANOVA results for New Leaf Paper®. Signiﬁcant

as S
reams

    

Element

None of the elements showed signiﬁcant variation in concentration among
all the sheets tested from a single ream; all the p-values were > 0.05. This
indicated that the elemental concentrations for all the elements listed were not
statistically different for the three sheets in a single ream. If a single sheet Of
unknown origin were sampled it should have the same elemental concentration
as any other sheet in the ream it came from and could be excluded from as

originating from a ream that had different elemental concentrations.

57

The average elemental concentration within an entire ream (15 samples, 3
sheets with 5 samples each) was compared for the 5 reams and the resulting p-
values are also displayed in Table 4.3. There are four elements (Mg, Al, Sr, and
Ba) that have a significant variation in concentration among the reams. In order
to further investigate these differences among the reams (i.e., determine
Speciﬁcally which reams differ from one another), a statistical Tukey test was
performed. The Tukey test is a pairwise comparison between each group
member, in this case reams, to determine where the statistical differences occur.
The null hypothesis was there was no difference between the elemental
concentration of the reams and the alternate hypothesis was there is a difference
in the mean elemental concentration of the reams. The p-values represent the
probability that the null hypothesis is correct. A p-value s 0.05 indicates the
mean concentrations are different at a 95% conﬁdence level and similarly for a p-
value S 0.01 the conﬁdence level is 99%. The results of the Tukey test are
displayed graphically in Figures 4.5 through 4.8.

In Figures 4.5 through 4.8 the bars represent the average element
concentration (n=15) for each of the reams. The error bars represent the
standard deviation of these measurements. Mean values that are signiﬁcantly
different from one another are connected with a line and the * (p-value S 0.05)
and ** (p-value s 0.01) indicate the signiﬁcance level of the difference. For
example in Figure 4.5, the Mg concentration in ream 1 iS signiﬁcantly different

than the concentration in ream 2 at a 99% conﬁdence level.

58

 

 

1200 i II N IF
1000 1
800 -
600 -
400 -

200 -

Mg Concentration (pg/g)

 

O
l

1 2 3 4 5
Ream Number

Figure 4.5. Mg concentration in New Leaf Paper®, reported as pg Mg perg paper. Error
bars represent standard deviation. Signiﬁcant difference between values with p S 0.05
(*) or p s 0.01 (**).

700 - H H
500 -
500 -
400 -
300 -
200 -
100 -

0 ..

Al Concentration (pg/g)

 

1 2 4 5

3
Ream Number

Figure 4.6. Al concentration in New Leaf Paper®, reported as pg Al per 9 paper.
Error bars represent standard deviation. Signiﬁcant difference between values
with p s 0.01 (**).

59

S Concentration

r
.0"
o

 

.0
O
I

**

 

** * *

 

 

I ll

1 2 3 4 5

Ream Number

Figure 4.7. Sr concentration in New Leaf Paper®, reported as pg Sr per 9 paper.
Error bars represent standard deviation. Significant difference between values

14.0

12.0

10.0 -
8.0 -
6.0 -
4.0 -
2.0 -

 

Ba Concentration (pg/g)

0.0 ‘

with p s 0.05 (*) or p s 0.01 (**).

**

 

 

 

1 2 3 4 5
Ream Number

Figure 4.8. Ba concentration in New Leaf Paper®, reported as pg Ba per 9
paper. Error bars represent standard deviation. Signiﬁcant difference between

values with p s 0.01 (**).

60

For the New Leaf Paper®, the same pattern was apparent for all of the
elements Chosen. The Tukey test was able to group the reams into a few groups.
Reams 1 and 3 were distinguished from reams 2 and 4 at the 95% conﬁdence
level or higher and vice versa based on all of the elements used for discrimination
(Mg, AI, Sr and Ba). Reams 1 and 3 were not distinguished from each other and
reams 2 and 4 were not distinguished from each other. Ream 5 was not
distinguished from any of the reams. In looking at the elemental concentrations
ream 5 has concentrations that were in between the concentrations seen in the
group of ream 1 and 3 and the group of ream 2 and 4. This made it statistically
indistinguishable from any Of the other reams. Interestingly reams 1, 3, and 5
had the same marking on the ream wrapper (JD6041015) and reams 2 and 4 had
very similar marking (J06041115 and JD6041115, respectively). This suggested
that these ream markings had some association with a batch or manufacturing
time frame because the elemental concentrations were statistically the same for
some of the similarly marked reams. It is also interesting to note that ream 5 had
the same markings as 1 and 3 and, while it was not a statistically signiﬁcant
difference, the mean concentrations were different than 1 and 3 suggesting that
there could be some variation within a single batch.

The other criterion to look at was the level of statistical signiﬁcance
indicated by the * (p s 0.05) or ** (p s 0.01) in the ﬁgures. Both the Al and the Ba
have all of the differences signiﬁcant at the 99% conﬁdence level. The larger
difference in the mean concentration of Al and Ba (Al: 338 1: 35 pglg for ream 1

and 561 1: 48 pg/g for ream 2, Ba: 12.0 i 0.9 pglg for ream 1 and 7.0 i 0.6 pglg

61

for ream 2, vs. Mg: 1070 i 71 pglg for ream 1 and 956 i 72 pglg for ream 2) led
to higher statistical signiﬁcance of the comparison. Because only a small
population of reams was studied it is impossible to make overall conclusions but
the preliminary results of this study suggest that Al and Ba would be the best
choice for discrimination between reams of New Leaf Paper®.
4.3.2. Staples®

Elemental concentrations in sheets and reams of Staples® paper were
statistically compared using the previously described procedures. The two-way
ANOVA results are shown in Table 4.4.

Table 4.4. Two-way ANOVA results for Staples® paper. Signiﬁcant differences
deﬁned as p-value s 0.05.

Element

0.139 no
Al 0.512 no
Mn 0.539 no
Fe 0.193 no 0.0384

Sr 0.384 no
Y 0.407 no
Ba 0.544 no 0.0258
Ce 0.480 no 0.0302
Nd 0.391 no 0.498 no

 

Both the New Leaf Paper® and the Staples® showed no statistical difference
among concentrations of elements throughout a single ream. There was also no
effect of the paper being in contact with the ream wrapper as the very top and
bottom sheets were sampled. This is important because in forensic sampling
only a small portion of the document will be sampled and it is important that the
sampled analyzed is representative of the whole Sheet. Also it is important that

the ream is consistent so a single sheet that has been removed from the ream

62

(i.e., questioned document) will have the same elemental concentration of the
other sheets in that ream used for comparison.

The variation between the mean elemental concentrations of the reams
was calculated (n=15, 3 sheets with 5 samples each) and the resulting p-values
are displayed in Table 4.4. Every element, with the exception Of Nd, had
statistical differences. To investigate further the Tukey test was employed to
Obtain pairwise comparisons to determine which reams could be distinguished.
The results Of the Tukey test are displayed graphically in Figures 4.9 through
4.16 for each of the elements that showed a signiﬁcant difference in

concentration among the 5 reams with ANOVA.

 

** **
1200 4 ** ** 1

1000 -

 

 

800 -

 

 

600 -

400 -

200 -

Mg Concentration (pg/g)

 

O
1

1 2 4 5

3
Ream Number
Figure 4.9. Mg concentration in Staples® brand paper, reported as pg Mg per 9

paper. Error bars represent standard deviation. Significant difference between
values with p S 0.05 (*) or p s 0.01 (**).

63

 

800 T *9: T I **

O)
O
O

400 -

200 '

AI Concentration (pg/g)

 

O
l

1 2 3 4 5
Ream Number
Figure 4.10. Al concentration in Staples® brand paper, reported as pg Al per 9

paper. Error bars represent standard deviation. Signiﬁcant difference between
values with p s 0.01 (**).

 

 

. * ** **
.ll.

 

 

.01 .01
O O
I l

P
o

I" 9’
o o
l

1.0-

Mn Concentration (pg/g)

 

0.0 -
3 4 5
Ream Number

Figure 4.11. Mn concentration in Staples® brand paper, reported as pg Mn per 9
paper. Error bars represent standard deviation. Signiﬁcant difference between
values with p s 0.05 (*) or p s 0.01 (**).

120 ~

6‘» j
E)
3100
C
it“:
,5 50-
C
d.)
O 40~
C
O
O 20-
CD
LI. 0_

 

1 2 4 5

3
Ream Number

Figure 4.12. Fe concentration in Staples® brand paper, reported as pg Fe per g
paper. Error bars represent standard deviation. No significant differences
between pairs.

 

 

 

*

A 751 ** **
9 | ** **
c»
3 601
8
'4: 45-
(0
J:
5 30-
o
C
8 15-
L-
U)

0.1

 

 

1 2 3 4 5
Ream Number

Figure 4.13. Sr concentration in Staples® brand paper, reported as pg Sr per 9
paper. Error bars represent standard deviation. Significant difference between
values with p s 0.05 (*) or p S 0.01 (**).

65

A 0.30 '1

 

0.25 1

.0
N
o

Y Concentration (pg/g
p o o

 

0.00 -

1 2 4 5

a
Ream Number

Figure 4.14. Y concentration in Staples® brand paper, reported as pg Y per 9
paper. Error bars represent standard deviation. Signiﬁcant difference between
values with p s 0.01 (**).

3.0 1
2.5 -

2.0 -

0.5 '

Ba Concentration (pg/g)

 

 

0.0 -
1 2 3 4 5

Ream Number
Figure 4.15. Ba concentration in Staples® brand paper, reported as pg Ba per 9

paper. Error bars represent standard deviation. Significant difference between
values with p s 0.05 (*).

 

 

0.3 -

0.2 ‘

0.1 -

Ce Concentration (pg/g)

 

   

0.0 -
1 2 3 4 5

Ream Number
Figure 4.16. Ce concentration in Staples® brand paper, reported as pg Ce per 9

paper. Error bars represent standard deviation. Signiﬁcant difference between
values with p s 0.05 (*).

Using the results Of the Tukey test, two of the reams of paper (1 and 3)
were able to be discriminated from all other reams based on elemental
concentration. Ream 3 had different elemental concentrations than all other
reams for Mg, AI, Mn, Sr and Y, at a 99% conﬁdence level. Ream 1 was also
differentiated from all other reams; the Mn concentration was significantly
different from reams 2 and 5 at the 95% conﬁdence level, the Mg concentration
was different from ream 4 at the 95% conﬁdence limit and Mg, Al, Mn, Sr and Y
concentrations were different from ream 3 at the 99% conﬁdence limit. Ream 2
was found to have a different Sr concentration from ream 5 at a 95% conﬁdence
limit. Reams 2 and 4 could not be differentiated from each other nor could reams
4 and 5.

Because nearly all of the elemental concentrations Of ream 3 were so

different from the other reams, this suggests that ream 3 came from an entirely

67

different batch. This is not known for sure because there were no markings on
the packages. Ream 1 could also be differentiated from all the others but every
elemental concentration did not differ from all other reams. This is promising
because only a very small population (n=5) of reams was tested, all purchased at
the same time, and a Signiﬁcant difference was observed. However, it should
also be noted that there were pairs of reams in this sample set (e.g., reams 2 and
4, reams 4 and 5) that could not be differentiated.

Based on these Observations it seems that Mg, Mn, and Sr were the most
useful elements for comparison of Staples® paper because these elements were
able to discriminate the largest number of pairs. No statistically different pairs
were observed for Fe, despite a p-value of 0.04 from ANOVA. Even though
ANOVA indicated that there was statistically signiﬁcant variance between the
reams the Tukey test did not identify any pairs that were statistically different from
one another. The concentration of Ba was relatively constant across all of the
reams and the only statistical difference was observed between the highest and
lowest values (reams 3 and 5). Ream 3 had very different elemental
compositions from all of the other reams. AI and Y were very different in ream 3
as compared to the others but were unable to differentiate any other pairs. While
Al and Y were not very useful in this small population (n=5 reams) there was one
ream in this population that was different than the other suggesting that if a larger
number of reams were tested it is possible that more variations in Al and Y would

be observed. Additionally the suite of elements was optimized using New Leaf

68

Paper® digests so it is possible that elements only present in Staples® were not
studied and these elements could offer more discrimination.
4.4. Variation Between Vendors

In these studies, two paper types manufactured in the United States, both
of 100% recycled materials, were investigated. Students t-tests were used to
determine if there were signiﬁcant differences in the means of the elemental
concentrations between the two sample sets. The average concentrations of
each of the samples (n=75, 5 reams, 3 sheets from each and 5 samples per
sheet) and the results of the t-test are'displayed in Figure 4.17. In this ﬁgure, the
bars represent the average concentration of the element throughout all the
samples taken from the vendor and the error bars represent the standard
deviation of these measurements. Similar to the previous ﬁgures, pairs of
samples with signiﬁcant differences are connected by a line and the * (p s 0.05)
or ** (p s 0.01) indicates the signiﬁcance level. The t-test showed there were
differences in the Ba and Nd concentrations at the 99% conﬁdence level and Sr
and Ce at the 95% conﬁdence limit.

Ternary plots were used to graphically Show the differences in the
elemental composition of the papers. These types of plots make it very easy to
visually see similarities and differences between the samples and are useful for
court presentations to juries.11 Only three elements can be used for each ternary
plot, so two plots were made one with Ba, Nd, and Ce and the second with Ba,
Nd, and Sr. The elemental ratios are plotted on three axes as the proportion of

each element so the sum was equal to 1.0 (e.g., Ba + Nd + Ce = 1.0).

69

1200 1

1000 1

800 -

600 -

400 -

Concentration (pg/g)

14.0 -

12.0 -

10.0 -

8.0 -

6.0 ‘

4.0 -

Concentration (pg/g)

2.0 -

 

0.0 -

0.40 -
0.35 -
0.30 ‘
0.25 ‘
0.20 -
0.15 -
0.10 4

Concentration (pg/g)

0.05 -

 

0.00 -

 
   

Ce

I New Leaf Paper®
El Staples®

**

Nd Y

Figure 4.17. Concentration of trace metals found in the paper samples from New
Leaf Paper® and Staples® brand. The error bars represent standard deviation of
the values n=15. Significant difference between values with p s 0.05 (*) or p s

0.01 (**).

Each of the five reams from each vendor is plotted in Figures 4.18 and 4.19.
Two groups are clearly formed, each containing the 5 reams from the different
vendors, as illustrated in Figures 4.18 and 4.19.

The elements that were most useful for discriminating the reams from the
same vendor were Al and Ba for the New Leaf Paper® and Mg, Mn, and Sr for
Staples® paper. The Ba (New Leaf Paper®) and Sr (Staples®) were also
elements that signiﬁcantly different concentrations between the vendors. Note in
Figure 4.18 the New Leaf Paper® samples are more spread out across the Ba
axis than the Staples® because Ba is also a discriminating element for the New
Leaf Paper®. Similarly in Figure 4.19 the Staples® samples are more spread
across the Sr axis because this is a discriminating element for the Staples®
papers. Of the two plots, Figure 4.18 is more useful to display differences
between the vendors because the Sr variation in the Staples® makes that
grouping much more spread out.

Both the Ba and the Sr have differences between the reams within a
vendor and between the vendors. These two elements may originate from trace
components of items (e.g., kaolin, Ti02, pulping chemicals) added during the
paper manufacturing process. There iS some variation within a vendor of Ba and
Sr but there is a larger variation between the vendors suggesting that the
concentration of these elements is affected by the manufacturing process used
by each vendor. Ba is a trace component in the filler material (CaCOa) which
may be added to the paper during the manufacturing. Small variations in Ba

concentration among the vendors may occur based on the lot of CaC03 that is

71

used. If the different manufacturers who supplied each of the vendors used
different amounts of CaCOa this could cause the large difference in the Ba
concentration. If these trace elements were from variation in the wood pulp (i.e.,

naturally occurring in the tree cells) a difference between the vendors would not

be expected.

 

 

 

Figure 4.18. Ternary plot of the ratio of Ba, Ce and Nd. The Ba concentration
was significantly higher than the Ce or Nd concentrations therefore it was divided
by 10 to make the plot easier to read. Each point represents an average ream
value, 5 reams from each vendor are plotted.

The Ba concentration of the New Leaf Paper® is approximately 280%
greater than the concentration in the Staples® paper (9.3 i 2.3 pg/g vs. 2.5 :t 0.3
pglg) but the Nd is only about 10% greater in the New Leaf Paper® (0.16 :I: 0.1
pg/g vs. 0.14 1: 0.1 pglg). Although both of these differences are statistically

signiﬁcant at the 99% confidence limit, Ba would be much easier to use as a

72

discriminating element for these paper samples because of the larger increase.
This is also evident in the ternary plot because the biggest separation in the two
groups is along the Ba axis. The other beneﬁt of using the Ba is the
concentration is approximately 10 times higher than the Nd, making Ba much

higher above the detection limit of the method.

 

 

 

 

 

0 500.75

Bal10
Figure 4.19. Ternary plot of the ratio of Ba, Sr and Nd. The Ba and Sr
concentration were significantly higher than the Nd concentration therefore there
were divided by 10 and 100, respectively, to make the plot easier to read. Each
point represents an average ream value, 5 reams from each vendor are plotted.
4.5. Conclusions
The trace elemental concentrations appeared to be consistent across a

single sheet as well as within a single ream. This indicated that each ream can

be treated as a Single source and it can be assumed that a single sample from

73

the ream will have the same elemental concentrations as any other sample from
the ream.

There were statistically signiﬁcant differences between the different
reams. The different vendors had different discriminating elements. For the New
Leaf Paper® the most discriminating elements appeared to be Al and Ba but for
the Staples® paper Mg, Mn and Sr were more beneﬁcial.

The papers that came from different vendors were easily discriminated
from one another. In the samples that were used here the Ba concentration was
the best element to compare, New Leaf Paper® had a concentration that was
about 280% higher than the Staples® paper. Other elements (Sr, Ce, and Nd)
also had significant differences.

If this technique were to be applied to a case sample the ﬁrst analysis
should be a full mass scan (mlz 7 — 240) for both the questioned document and
the exemplar sample. This would determine if the samples had any immediate
differences (i.e., presence or absence of element). If there were no immediate
differences, based on this study the suggested elements to quantify would be
Mg, Al, Mn, Sr, Ba, Ce, and Nd. The digestion and quantiﬁcation would need to
be done on the questioned document as well as exemplar samples from possible
reams (sources). The elemental concentrations of the question document would
be compared to the exemplar samples using a t-test. If signiﬁcant differences
were found between the questioned document and the exemplar then the paper

could be excluded as coming from that source.

74

This study was only completed on a small number of ream samples
collected over a relatively short time frame. A ream is an arbitrary packaging of
500 sheets of paper and it may not be known to a consumer, or forensic scientist,
if two reams are from the same batch which may produce thousands of reams of
paper. It is possible that the reams from each vendor sampled in this study were
from the same batch; however, differences were still observed between the
reams. More extensive sampling would be able to determine if the
concentrations were consistent for a single batch or if smaller populations existed
within a batch. Also additional research of different paper vendors and
manufacturers may reveal element concentrations that are characteristic of a

particular vendor or manufacturer.

75

4.6. References

(1)

(3)

(4)

(5)

(10)

(11)

Schlesinger, H. L.; Settle, D. M. Journal of Forensic Sciences 1971, 16,
309-330.

Blanchard, D. 8.; Harrison, S. H. Journal of Forensic Sciences 1978, 23,
679-686. ‘

Brunelle, R. L.; Washington, W.; Hoffman, 0; Pro, M. Journal of the
Association of Official Analytical Chemists 1971, 54, 920-924.

Polk, D. E.; Attard, A. E.; Giessen, B. C. Journal of Forensic Sciences
1977, 22, 524-533.

Simon, P. J.; Giessen, B. C.; Copeland, T. R. Analytical Chemistry 1977,
49, 2285-2288. '

Spence, L. D.; Baker, A. T.; Byrne, J. P. Joumal of Analytical Atomic
Spectrometry 2000, 15, 813-819.

Spence, L. D.; Francis, R. B.; Tinggi, U. Journal of Forensic Sciences
2002, 47, 648-651.

Campbell, N. A. Bio/091’. 4th ed.; Benjamin/Cummings Publishing
Company, Inc.: Menlo Park, CA, 1996.

Smook, G. A. Handbook for Pulp & Paper Technologists, 3rd ed.; Angus
Wilde Publications Inc.: Vancouver, 2002.

Sigma-Aldrich. Specification Sheet - Calcium Carbonate, ACS reagent.
http://www.siqmaald riCh.Com/catalog/Search/ProductDetail/SlAL/398101
(accessed 2007)

Watling, R. J.; Lynch, B. F.; Herring, D. Journal of Analytical Atomic
Spectrometry 1997, 12, 195-203.

76

CHAPTER 5
CONCLUSIONS AND FUTURE WORK

In this work the trace elemental concentrations of document papers were
examined to determine the forensic utility of this measurement. The paper that
was examined was a common multipurpose copy paper made from 100% post-
consumer waste. Two vendors (New Leaf Paper® and Staples®) were selected.
From each, 5 reams were sampled and 3 sheets were sampled from each of
these reams. A number of comparisons were made to determine the
homogeneity of element concentration across a sheet of paper as well as
throughout a ream. Statistical comparisons were made of the elemental
concentrations of different reams as well as different vendors.

In order to determine the trace elemental concentrations of the paper
sample a microwave digestion was performed followed by analysis with
inductively couple plasma mass spectrometry (ICP-MS). In a digestion solution
that contained only nitric acid (HN03) and hydrogen peroxide (H202), a complete
digestion of paper the paper samples was not possible and a small amount of
white particulate matter remained in the bottom _of the vessel. The undigested
portion was Characterized by infrared spectroscopy (FTIR) and x-ray diffraction
spectroscopy (XRD) and determined to be a mixture of kaolin (clay) and Ti02,
which are common ﬁller materials in papers.

The paper could be fully digested if hydroﬂuoric acid (HF) was added to
the digestion solution. However, this method is not desirable because HF is a

very corrosive acid, requires larger paper samples, and additional evaporation

77

steps prior to ICP-MS analysis. A comparison of the digestion methods
(HNOa/HzOz digestion in which particulates were present and HF digestion in
which no particulates were present) showed that the reproducibility of the
elemental concentrations was not affected by the presence of this particulate
matter. Since an HF digest was not necessary, the elemental concentrations
used for comparisons in this work were determined using the HNOg/HzOz
digestion.

Once the digestion procedure was developed a suite of elements to
analyze needed to be determined. Based on previous works as well as a full
mass scan of one paper sample an initial list of element was generated (Li,
23Na, 24Mg, 27 Al, 458‘), 55Mn, SZCr, sst’ 56Fe’ 5900’ sac”, 662m aasr’ 89Y’ 902V.
1°7Ag 114cc, 1218b, 1388a, 139La, 14"Ce, 146Nd, 208%, 232m and 238m. A full mass
scan was performed on a single paper sample and some of these elements were
eliminated (520r, 55Rb, 59Co, 63Cu, 107Ag, 114Cd, 121Sb, and 139La) because they
were at concentrations below the instrument detection limit and additional
elements were added to the list (12°Sn and 184W). The procedural blank samples
were analyzed and elements with a signiﬁcant contribution from the background
were eliminated as well as elements that were not homogeneous over a single
sheet (relative standard deviation >15°/o). The ﬁnal element suite was narrowed
to include only 24Mg, 27Al, 55Mn, 56Fe, 88Sr, 89Y, 138Ba, 140Ce, and 146Nd.

Statistical analysis was completed using analysis of variance (ANOVA)
and Tukey tests to determine if any statistical differences existed and if so,

between which pairs of samples. In all cases the trace elemental concentrations

78

were consistent across a single sheet as well as throughout a ream. This is
important because when sampling only a small portion of a single sheet will need
to be analyzed and it will be representative of the entire sheet. Likewise a single
sheet that has been removed from a ream (i.e., questioned document) will have
the same elemental concentrations of other sheets from the same ream.
Therefore it is appropriate to sample a small section of a document for analysis
and comparison.

There were statistically signiﬁcant differences between the reams. For the
New Leaf Paper® (NL), reams NL1 and NL3 were distinguished from reams NL2
and NL4 at the 95% conﬁdence level or higher and vice versa based on the
elemental concentrations of Mg, Al, Sr and Ba. Reams NL1 and NL3 were not
distinguished from each other and reams NL2 and NL4 were not distinguished
from each other. Ream NL5 was not distinguished from any of the reams.
Interestingly, reams NL1, NL3, and NL5 had the same markings on the ream
wrapper and NL2 and NL4 had very similar markings. This suggested that the
reams that were grouped with one another were from the same batch and the
ream wrapper markings were also correlated to a speciﬁc batch.

For the Staples® paper (S), two of the reams (S1 and S3) were able to be
discriminated from all other Staples® reams based on elemental concentration.
Ream 83 had different elemental concentrations than all other reams for Mg, Al,
Mn, Sr and Y, at a 99% confidence level. Ream St was also differentiated from
all other reams; the Mn concentration was significantly different from reams 82

and SS at the 95% conﬁdence level, the Mg concentration was different from

79

ream S4 at the 95% confidence limit and Mg, Al, Mn, Sr and Y concentrations
were different from ream S3 at the 99% conﬁdence limit. Ream 82 was found to
have a different Sr concentration from ream 85 at a 95% conﬁdence limit.
Reams 82 and S4 could not be differentiated from each other nor could reams
S4 and SS. The ream S3 was very different from all of the other reams
suggesting that it may have come from an entirely different batch.

The different vendors had different elements that offered the best
discrimination. For the New Leaf Paper® the best elements were Al and Ba but
for the Staples® paper Mg, Mn and Sr were more useful. The Staples® paper had
more elements that varied between the reams, but one ream (SB) was different
from all others and accounted for a large number of the variations. It is promising
to see that a paper of the same brand can vary so much in elemental
concentration even when it was purchased from the same store and at the same
time as the other reams tested.

A student’s t-test was used to compare the elemental concentrations of
papers that came from the different vendors. The vendors were easily
discriminated from one another. In the samples that were tested in this work, Ba
concentration was the best element to compare; New Leaf Paper® had a
concentration that was about 280% higher than the Staples® paper. However,
Sr, Ce, and Nd also showed signiﬁcant differences in concentration between the
vendors and could also be used for comparison.

Forensically this test is useful because it is known that the elemental

concentrations within a ream of paper are consistent but differences exist

80

between reams therefore two papers can be compared and, if they have different
elemental concentrations, they can be excluded from having come from the
same source (ream). For example if it was suspected that a page in a will had
been added after the original document was created, these two paper samples
could be analyzed and compared. If it was found that the two pages had
different trace elemental composition they could be excluded from having come
from the same source. Based on the small sample used in this study, the
recommended elements for comparison would be Mg, Al, Mn, Sr, Ba, Ce, and
Nd.

Because this study only use a very small population there are many
potential studies that can be done to further assess the utility of this method. First
of all it would be difﬁcult to determine how many reams have the same elemental
concentrations. Studies need to be completed on papers from the same batch.
For this, cooperation of a paper manufacturer would be necessary to obtain
several samples from the same batch (i.e., a single pulp mixture) to determine
how much, if any, variation is observed. It would also be interesting to
investigate the elemental concentrations over time (e.g., months, years) to see
how much they varied. I

If more samples could be obtained for a larger number of vendors or
manufacturers, it would be useful to determine if certain elemental concentrations
are characteristic of a particular vendor or manufacturer. For example, all New
Leaf Paper® may have a Ba concentration of around 9 pglg while all other

vendors have a much lower Ba concentration. This would be useful to narrow

81

the source of the paper to a particular vendor or manufacturer. Along the same
lines, perhaps a study of different types of paper from the same vendor (e.g.,
copy paper, laser printer paper, inkjet paper) may point to some characteristic
elemental concentrations that are characteristic of the paper type. These
variations may be due to different fillers or sizing chemicals used to give the
different paper types different characteristics.

In this work only paper used from 100% post consumer use was
evaluated. Another study could examine the effect of the percentage of recycled
material used in the paper on the variation seen in the elemental concentrations.
In preliminary studies done for this work a few sheets of virgin and 30% recycled
paper and these samples contained all of the elements that were used for the
comparison of the 100% recycled paper, however, no assessment was made as
to the discrimination ability of these elements in the other paper types. It is also
possible that in paper that is made from recycled material or recycled paper may
be easier to distinguish because it has a larger variability in the trace
components.

Many of the trace elements in the paper may come from the pulp source
(e.g., tree cells). Therefore there could be some variation in the trace element
concentration depending on the type of ﬁber source used (e.g., tree, cotton,
hemp). Also the geographical location where the tree or plant was grown may
affect the trace elements present based on the composition of the soil.

Overall ICP-MS is promising for forensic analysis of paper. This chemical

analysis of the trace elements present would be an addition to the traditional

82

document analysis. This test appears to be able to distinguish papers of the
same type from the same vendor that traditional analysis would not be able to

distinguish base only on physical characteristics.

83

 

lllllllgillllljijjlililyi