Ceglia v. Zuckerberg et al

Filing 416

DECLARATION signed by Larry Stewart re 348 Order on Motion to Stay, Scheduling Conference, Oral Argument,,,,,,,,,,,, 414 Declaration filed by Paul D. Ceglia Expert Report. (Attachments: # 1 Supplement 46-55, # 2 Supplement 56-66, # 3 Supplement 67-90, # 4 Exhibit p. 1-82, # 5 Exhibit p. 83-167, # 6 Exhibit p. 168-229, # 7 Exhibit p. 230-290)(Boland, Dean)

Exhibit 14

Declaration of Larry F. Stewart

Exhibit 15

Declaration of Larry F. Stewart

Lesnovich has implied that there was some change in the pages of the WFH document
sometime between his Q1 and Q2 images and his Q3 and Q4 images. Following is an
experiment comparing the position of the interlineation writings to the printed document
on the best (clearest) sample of each of his two groups.
Q1 from Lesnovich. Tif sent by Ceglia to Argentieri 6/27/10.

Q3 from Lesnovich. Aginsky scan from 1/13/11 examination.

Exhibit 16

Declaration of Larry F. Stewart

Comparison of poorly resolved scanned image of the WFH Pg1 to high resolution scanned
image taken after forensic testing began and the document was in secure storage:

Here we have two
images that were
simply scanned at
different settings
creating similar
anomalies to what
Lesnovich is pointing
out. This doesn’t
imply wrongdoing,
but instead computer
scanner setting
differences.

Lesnovich’s
Q1 - Image
from Ceglia
email to
Argentieri
6/27/10 (Note:
Poor
resolution)

High
resolution
scanned
image taken
7/15/11
Look at the
difference in
thickness within the
vertical staff of the
printed “P” in the
lower right corner.
The difference in
thickness can cause
the letters to merge or
appear to close in on
each other.

Exhibit 17

Declaration of Larry F. Stewart

Exhibit 18

Declaration of Larry F. Stewart

TEST REPORT
December 13, 2011
Page 1 of 2
IPS FI 02956-11
Report to:

Larry Stewart
Stewart Forensic Consultants
793 A East Foothill Blvd.
San Luis Obispo, CA 93405

Sample identification: 2 Vials
Date received: November 1, 2011
Test requested: Fiber Identification
Purchase Order: Credit Card
Report of Fiber Analysis
Enclosed are the results of the analysis performed on the sample we received with your
Test Services Request Form.
If you have any questions concerning this work, please do not hesitate to contact us.
Authorized By: ______________
Gregory J. Fox
Lab Manager

Signed _______________________________
Walter J. Rantanen
Technical Leader, Fiber Science
[phone] Ext. 127

WJR/jml

3211 E. Capitol Drive ~ Appleton, WI 54911
[phone] ~ Fax: [phone] ~ www.ipstesting.com

Report to Stewart Forensic Consultants
IPS FI 02956-11
Fiber Identification

December 13, 2011
Page 2 of 2

The paper samples did not have any detectable mechanical (high lignin) pulp fibers
which would be effected by photodegredation from UV light. There is a strong UV
fluorescence in both samples, which indicates optical brightening agents. In the small
punch outs, significant fluorescence differences were not detected. It could not be
determined if these samples were effected by contact with UV light, but long exposure
to UV light has been known to lower the whiteness of paper. A noticeable particulate
material was observed on one side of the punch outs. This particulate may also affect
the UV fluorescence of paper. The main inorganic substance in these particulates was
found to be iron. The EDS spectra are enclosed. The nature of this material implies
contact on one surface of the papers.
Spot tests show the same consistent reactions for starch and pH levels between the two
samples. The fiber content of the two vials is consistent with coming from the same mill
and production run.
Table 1. Fiber Identification of Vial 7
Hardwood Bleached Kraft – Principally Redgum and Oak with some Blackgum, YellowPoplar, Cherry, Southern Magnolia
Softwood Bleached Kraft – Hard Pine (Except Red & Pine)

Table 2. Fiber Identification of Vial 9
Hardwood Bleached Kraft – Principally Oak and Redgum with some Yellow-Poplar,
Blackgum, Cherry
Softwood Bleached Kraft – Hard Pine (Except Red & Pine)

Method: TAPPI Test Method T 401 om-03 “Fiber Analysis of Paper and Paperboard.”
Analyzed by WJR
Quality review by JML, KTM
Date(s) of testing November 8, 2011
Notes: These results relate only to the item(s) tested. This test report shall not be
reproduced, except in full, without written consent of IPS. See the TAPPI test
method(s) cited for estimates of measurement uncertainty.

Exhibit 19

Declaration of Larry F. Stewart

Lat7 y F. Stewart, 1 M . F . S .

Ballpoint Ink Age Determination by Volatile
Component Comparison--A Preliminary Study

REFERENCE: Stewart, L. F., "BaHpolntInk Age Determination by Volatile Component Comparlson--A PreliminaryStudy," Journal of Forensic Sciences, JFSCA, Vol. 30, No. 2, April 1985,
pp. 405-411.
ABSTRACT: Ballpoint pen inks consist primarily of a mixture of dyes. resins, and vehicle components. The vehicles are used to solubilize or suspend dyes. resins, and other compoueuts as well as
to provide smooth ball movement and flow of ink onto writing surfaces. These vehicles are relatively
volatile and make up approximately 50% of the ink by weight. Extraction and fornmlation identification of the questioned ink is performed. Once identified, the volatile components of the ink are
measured quantitatively by gas chromatography. PreliminalS stvdies show that the relative proportions of these volatile ingredieuts decrease as the ink ages. How long an ink has been on paper is
determined by comparison of the relative concentrations of the volatile components of the questioned ink with those of known inks (age) of the same formulation. The relationship between age of
ink. storage conditions, aud paper will also be discussed.
KEYWORDS: questioned documents, inks, gas chromatography, pens, age determination
The a m o u n t of time an ink has been on a d o c u m e n t has been a question t h a t has plagued
many forensic science examiners since writing inks were first introduced. The conventional approach to dating an ink entry has been the identification of certain ink components which may
indicate gross formulation changes. Ballpoint pen inks, first developed in the 1930s [1], used
oils for the vehicles. It was not until the 1950s that glycol-based inks were widely used [2]. Additions to formulas such as the introduction of copper phthalocyanine dye (1954) in ballpoint pen
inks, fluorescent dyes ( 1955 to 1957) in fountain pen inks, and the introduction of entirely new
markets, for example, felt and fiber tip pen inks (1961) have aided the forensic science investigator in determining the "age" of an entry. 2 O t h e r methods for dating an ink entry included
determining the presence or absence of a dye. for example, the blue dye in blue-black writing
inks of the early 1900s 131, and differentiating between the amounts of a c o m p o n e n t extracted
from two entries through the use of chemical reagents (for example, oxalic acid) in the 1920s [4].
During the mid-1960s, in an attempt to improve upon the conventional method by increasing the knowledge of known changes in formulations, W e r n e r Hoffman, Zurich Cantonal
Police, Zurich, Switzerland, began collecting samples of European ballpoint pen inks. He
began comparing questioned inks with his collection for purposes of showing similarities or differences between formulas [5]. In the mid-1960s, Richard Brunelle, Bureau of Alcohol, ToNote: the experimental work for this paper was conducted at the Bureau of Alcohol. Tobacco and Firearms Forensic Science Branch. National Laboratory Center. Rockville, MD.
Presented at the 34th Annual Meeting of the American Academy of Forensic Sciences, Orlando, FL.
8-11 Feb. 1982 and the Spring 1982 Joint Meeting of the Mid-Atlantic Association of Forensic Scientists/
Northeast Association of Forensic Scientists. Harrisburg, PA, April 1982. Received for publication 21
May 1984; revised manuscript received 28 June 1984; accepted for publication 29 June 1984.
IDocument analyst, United States Secret Service, Washington, DC.
2A. A. Cantu, private communication. 1980.
405

Copyright © 1985 by ASTM International

406

JOURNAL OF FORENSIC SCIENCES

bacco & Firearms, National Laboratory Center, Washington, DC, began collecting a library of
standard inks from U.S. manufacturers. This library has been maintained and expanded to its
present-day status of being the largest single collection of inks in the world consisting of over
4000 domestic and foreign inks. 3
The Bureau of Alcohol, Tobacco and Firearms (ATF) also initiated a national ink tagging
program (1971 to 1974) in an effort to determine more closely the age of an entry produced by
an ink whose formulation is not often changed by the manufacturer [6]. Even with these technical advances it is often necessary to determine more closely (less than a few years span) the actual age of an entry. This work will address only ballpoint pen inks because of their amenability
to drying determinations.

Composition of Ballpoint Inks
Ballpoint ink is a high viscosity (nonfluid) writing medium. It consists primarily of three
components [7]:
(1) vehicles,
(2) dyes or pigments or both, and
(3) resins or polymers.

Vehicles
Vehicles are added to an ink for purposes of solubilizing(or carrying) the dyes/pigments and
for ease of flow over the cartridge ball.
Vehicles in ballpoint inks have had only one dramatic formulation change since their inception in the 1930s. Before 1950, inks contained oil as the primary vehicle; after the early 1950s.
glycol-based inks were developed and quickly became the favorite among the population.
These inks usually contain one or more of the following vehicle solvents: 4
1,3 propylene glycol
Diethyl glycol phenyl ether
Benzyl alcohol
2 ethyl hexoic acid
Ethylene glycol
2,3-butylene glycol
Monophenylether
1.2-propylene glycol
Ethylene and diethylene glycol monomethyl ether

Hexylene glycol
Octylene glycol
1.3 butylene glycol
Di and triethylene glycol
Dipropylene glycol
Glycerine
Phenoxyethanol
Phenoxyethylene glycol

The volatile components of the ink make up approximately 50% of its composition.

Dyes and Pigments
Dyes and pigments are the color giving components of an ink. Some of the more common
ones used in ink formulations include:4
Methyl violet
Victoria blue
Crystal violet
Copper pbthalocyanine
Nigrosine
Solvent fast blue
Luxol fast orange
Dyes and pigments make up approximately 25% of the ink's composition.

3A. A. Cantu, private communication, 1982.
4Private communicationswith ink manufacturers. 1982.

STEWART , BALLPOINT INK AGE DETERMINATION

407

Reshts and Polymers
Resins a n d polymers are added to ballpoint inks for purposes of " e x t e n d i n g " the ink (used as
a filler) and for thickening the ink. Some resinous components found in writing inks include: 4
Vinsol|
Nevillac Hard +
Pyrrolidone (PVP)
K r u m b h a a r K-1717 ~
Phthalopal SEB ~
Synthetic Resin SK
The resinous additives usually make up approximately 25% of the total ink volume.
The vehicle components are of primary interest in this work. The ink cartridge is considered
a "closed" system; essentially no drying takes place in the cartridge. The ink on the paper surface is an " o p e n " system; the ink drying process begins as soon as the ink is placed on the paper.
The vehicles evaporate with time leaving the dyes,, pigments and resins ,'polymers adhering
to the writing surface.
This work is based on the fact that volatile components evaporate with time. Ballpoint pen
inks contain volatile components that begin evaporating when placed on a document. This indicates that the age of a ballpoint pen ink entry stored u n d e r some " c o n s t a n t " conditions could
be determined if the a m o u n t of volatile components per weight,volume of ink was measured
(see Fig. 1). If the temperature and humidity do not remain constant, then only the "relative"
age of an entry as compared to another entry (stored on the same paper) may be determined.
Materials and Equipment

The materials and e q u i p m e n t used were:
9 Temperature programmable gas chromatograph equipped with a flame ionization detector
9 Stainless steel column 1.8 m (6 ft) packed with 3 % Tenax GC on 60-80 mesh Supelcoport
9 Ten-microlitre syringe
9 Micro vials (0.5 dr tapered)
9 High purity methanol
9 Micro pipets, 10/xL
9 Ice bath
9 High purity vehicle standards
9 •
gauge hypodermic needle
9 Plunger
9 Timer

"New" Entry

'i"

~as~rlng

Apparatus

"Old" Entry

FIG. l - - Theo O' o/' work.

Consists of X + Y %
Volatile Conponent s

408

JOURNAL OF FORENSIC SCIENCES

Method
The first step in determining the age or "relative" age of a ballpoint pen ink entry is the identification of the ink formulation. Identification is necessary so the examiner can determine the
quality control from the manufacturer and the "uniqueness" of a formula. The method used
involves thin-layer chromatographic comparisons of questioned to known ink samples [5]. The
known ink samples used in this work are stored and maintained in the standard ink library at
the Bureau of Alcohol, Tobacco & Firearms, National Laboratory Center, Roekville, MD.
Once the questioned ink formulation has been identified, the volatile components and percentages present in kr~own "fresh" ink of the same formula are obtained by gas chromatographic analysis.
Fresh ink samples of the same formulation were placed on a single sheet of paper on various
dates. This sheet was stored under "standard" conditions (that is, room temperature and
humidity) in a file drawer.
Samples of ink were removed from the paper by a micro-pellet technique. This technique
utilizes a blunted 20-gauge hypodermic needle fitted with a shortened "syringe type" plunger.
The micro-plugs of ink and paper ( : 15 plugs) are placed in tapered microvials. Approximately 10 to 15 p,L of methanol is slowly added (being careful not to disturb the pellets) by syringe through the capped/stoppered lid of the vial. The vial is placed in an ice bath to minimize
"travel" of the methanol up the sides of the tapered vial, The vials remain in the ice bath
undisturbed for 5 rain. At the end of the extraction process a 5- to 10-pL aliquot is removed
for injection into the gas chromatograph.
A gas chromatograph equipped with a flame ionization detector was chosen as the analysis
instrument because of the need for reproducible detection and quantitation of micro-amounts
of volatile components.
Fresh ink samples containing different combinations of volatile components were chromatographed using various extraction methods, gas chromatograph columns, and temperature
programs. A suitable method for analysis was obtained. The gas chromatographic conditions
chosen are as follows:
Temperature programmable gas chromatograph (Perkin-EImer Sigma 3B)
Flame ionization detector
3% Tenax GC on 60-80 mesh Supelcoport (stainless steel, 1.8 m [6 ft])
N2 gas flow at 25 cm3/min
Initial hold, 0-min
12~
ramp
50 to 280~
Final hold, 10 min
Chart speed, 12.7 ram/rain (0.5 in./min)
Injections of 5 to 10 #L using methanol as the extracting solvent
Attenuation: • K till methanol peak, then • 100
Once a suitable chromatogram was obtained the vehicle peaks were identified by using
known standards and formulation information obtained from the ink's manufacturer.
An "'aging" curve for each ink was obtained (Fig, 2). This was done by finding two sufficiently
resolved vehicle peaks, quantitatively determining the peak areas, and then ratioing one peak
to the other. The ratio of Peak A/Peak B is plotted versus actual age (days). This gives the aging
curve for that particular ink formulation (Fig. 3).
The questioned entry is analyzed in the same way. The peak areas are taken, a ratio is obtained, and, using the previously calculated "aging curve." the age of the Q entry is determined
(Fig. 4).
This calculated "age" of the Q entry is absolute only if the storage conditions of both the Q
and K entries are identical. The storage conditions of the inks used to obtain the aging curve
should be equal or better (that is, slower aging process) than those of Q.

STEWART

9 BALLPOINT

INK AGE DETERMINATION

409

GO COMPARISION OF TWO INKS
OF THE SAME FORMULATION
WRITTEN AT DIFFERENT TIMES

&

fresh

writing
t w o month o I d
writing
/
_~

FIG. 2--Aging curve.for etwh ink.

8.0

7"01

RATIO = (PeakA) / (PeakB)

6.0
5.0

4.O

3.0
2.O
I.

0.0

0

10

20

30

40

50

60

70

80

90

lob

110

120

130

1~I0

150

160

FIG. 3--Agblg curve calculated from the ratio of Peak A/Peak B versus actual age.

DY
AS

410

JOURNAL OF FORENSIC SCIENCES

8.0

7.0

6.0

5.0

4.0

3.0

2,0

5.0

0.0

0

I0

20

30 ~"~ 40

50

60

70

8n

90

Ioo

11F)

]2F)

"3,s

130

I~0

]50

260
DY
AS

FIG. 4--Aging curve for d w Q ('toO'.

Conclusions
Ideally at least two inks of the same formula should be compared. They should be on the
same paper and stored u n d e r the same conditions.
If two inks of the same formulation on the same document have different ratios of the volatile
components, then one ink can be determined to be fresher than tile other, that is,
(A : first eluting comparison peak)
(B :

second eluting comparison peak) new

>

(A)

(B) old

If two inks of the same formulation are found on different paper, then tile paper type is probably not a factor but storage conditions are. The "willingness" of tile paper to allow these components to be extracted in the same ratio slmuld not be affected by a paper's porosity, thickness, type, or age. However, this must be further tested.
Certain ballpoint pen ink formulations were shown to have reproducible aging curves up to
one-and-one-half years after placement on paper. Differences in peak ratios for known inks
stored under standard conditions were detected over as small a time frame as a few days. Some
ink formulations tested have evaporation rates or vehicle components not ameuable to this
technique.
Ratioing the chromatograph peaks eliminates tile necessity of removing equal masses of
"questioned and known age" ink when performing an age comparison.
Further work that should be performed includes testing the paper iqdependence theory and
developing a laboratory technique for "controlled" artificial aging of ink standards to obtain
" i m m e d i a t e " aging curves for known standard inks.

STEWART * BALLPOINT INK AGE DETERMINATION

411

Acknowledgments
This work was greatly assisted by the knowledge and cooperation of the following: Dr.
Antonio A. Cantu, the ink industry, Dr. Phillip M. Daugherty, Richard L, Brunelle, and
Connie Lee.

References
[1] "The Battle of the Bali-Point Pen," Reader's Digest, Dec. 1946, pp. 59-63.
[2] Daugherty, P. M.. "Composition of Ball Pen Inks," presented at the First Georgetown University
Conference on Surface Analysis. Washington, DC. Oct. 1969.
[3] Waters, C. E., "'Blue Dye as Evidence of the Age of Writing," hzdustrialandEngmeering Chemist13,,
Vol. 25, No. 9, Sept. 1933. pp. 1034-1035.
[4] Mitchell, C. A. and Hepworth, T. C.. "'Inks: Their Composition and Manufacture," 3rd ed., Griffin
& Co., London, 1924, p. 181.
[5] Crown, D. A., Brunelle, R. L.. and Cantu, A. A., "The Parameters of Ballpen Ink Examinations."
JournalqfForensic Sciences. Vot. 21, No. 4, Oct. 1976. pp. 917-922.
[6] Brunelle, R. L., Cantu, A. A., and Lyter, A. H., "Current Status of Ink Analysis," presented at the
1975 Annual lnterpol Meeting. St. Cloud. France.
[7] Witte, A. H., "The Examination and Identification of Inks." in Methods of Forensic Science, Vol. 2,
[nterscience Publishers, London, 1963, p. 35.
Address requests for reprints or additional information to
Larry F. Stewart
United States Secret SeB'ice
Forensic Services Division
1800 G St., N.W., Room 929
Washington. DC 20223

Exhibit 20

Declaration of Larry F. Stewart

JOURNAL OF
CHROMATOGRAPHY A

ELSEVIER

Journal of Chromatography A, 678 (1994) 119-125

Determination of the age of ballpoint pen ink by gas and
densitometric thin-layer chromatography*
Valery N. Aginsky
Forensic Science Centre, Ministry of the Interior, 22 Raspletina Street, Moscow 123060, Russian Federation

(First received October 8th, 1993; revised manuscript received March 31st, 1994)

Abstract

Two procedures for dating ballpoint inks are considered that use gas chromatography (a combination of the
technique for determining the extent of extraction of ink volatile components and of the accelerated ageing
technique) and densitometric thin-layer chromatography (separation of ink components and evaluation of the
resulting chromatograms using a specially developed mass-independent
technique that is also a very effective tool
for the comparative TLC examination of similarly coloured inks, paints, fibres and other materials of forensic
interest). The procedures have been used in many real case situations and the results of the examinations were
accepted as conclusive evidence by courts of law.

1. Introduction

Gas chromatography
(GC) and densitometric
thin-layer chromatography
(TLC) have been
demonstrated to be useful tools for the solution
of many problems frequently encountered in ink
analysis, including ink dating problems [l-4].
Recently, five new procedures for dating ballpoint inks have been described [5,6]. Two of
them, based on using chromatographic methods,
are as follows.
(1) A GC method is used to determine the
extent of extraction of ink volatile components,
which decreases as ink ages on paper. The
procedure considered in this paper combines the
capabilities of this method and of the accelerated
ageing technique. The procedure allows discrimi*Presented in part at the
Germany,

August 1993.

13th IAFS

Meeting,

Dtiseldorf,

nation between “fresh” (age less than several
months) and “old” ballpoint ink entries and it
does not need dated reference entries written
with ink having the same formula as that of the
questioned ink.
(2) A TLC method is used for determining
age changes in resins and other colourless nonvolatile ballpoint ink components; these changes
are detected by observing the resulting thin-layer
chromatograms
under UV illumination
and
evaluated by using scanning densitometry.
The
modified TLC procedure described in this paper
includes a new, mass-independent
approach to
evaluating thin-layer chromatograms that allows
one to obtain the values of an “ink ageing
parameter” [7] directly proportional to the ratios
of the masses of the separated ink components
(dyes, resins, etc.). For this reason, the proposed
procedure gives more reliable results for ink age
determination
than those obtainable with the

0021-9673/94/$07.00
0 1994 Elsevier Science B.V. All rights reserved
SSDZ 0021-9673(94)00322-Z

V.N. Aginsky.

120

I J. Chmmatogr.

widely
used peak signal-to-peak
signal ratio
technique.
The described
approach
is also a
powerful
tool for the comparative
examination
of similarly
coloured
inks, paints,
fibres and
other materials
of forensic
interest
as its discriminating
power is much greater
than that
usually produced
by the peak ratioing technique
[S-10].

2. Experimental
2.1.

Materials

Up to 15year-old
entries written with Soyuz
ballpoint
inks of different colours having similar
compositions
of colourless
components
were
analysed by GC. Entries of known ages (1 day, 1
month,
1, 2, 3 and 6 years old) written with a
Parker blue ballpoint ink were analysed by TLC.
Camag N-11 polypropylene
micro vials with
cone-shaped
interiors
and a lo-~1 Hamilton
syringe were used.
2.2.

Gas chromatography

A Hewlett-Packard
Model 5890 gas chromatograph equipped
with a flame ionization
detector
and an HP split-splitless
injection
system was
used. A SCOT column containing
SP-1000 (polyethylene
glycol
20M terminated
with nitroterephthalic
acid) (Supelco)
(25 m x 0.5 mm
I.D.) was used with nitrogen
(4 p.s.i.) as the
carrier
gas at a flow-rate
of 40 mlimin.
The
column oven temperature
was programmed
from
50°C (held for 0.5 min) at lO”C/min
to 220°C
(held for 6 min). The injection
volume was 2 ~1
(splitless) at 250°C. A flame ionization
detector
was used at 250°C.
Each sample was obtained by cutting out a ca.
l-cm sliver of ink of approximately
equal thickness from the paper using a safety razor and
placed in a micro vial. A lo-~1 volume of carbon
tetrachloride
as a “slowly extracting
weak solvent”, containing
10 pgiml of benzyl alcohol as
an internal stranded (if benzyl alcohol is detected
in ink samples in significant
amounts,
another
appropriate
substance can be used as an internal

A 678 (1994) 119-125

standard)
was added and the vial was capped.
After 30 min a ca. 2-pl aliquot of each sample
was removed and analysed by GC.
The samples were removed
from the extraction solutions,
dried and placed into other micro
vials. A second extraction
was carried out for 1
min, stirring with a needle, with 10 ~1 of chloroform (“fast-extracting
strong solvent”)
also containing benzyl alcohol in the same concentration.
About 2 ~1 of each extract were removed
and
analysed by GC.
The masses of a vehicle component
determined in each of the two extracts analysed (M,
and M2 for the first and second extractions,
respectively)
were calculated
by means of the
internal
standard
method.
The percentage
extraction [l,S], that is, the percentage
of the mass
of the ink vehicle component,
%M, extracted in
the “weak” solvent (relative to its total amount
contained
in the sample analysed),
was calculated as follows:
%M = [M, /(M, + M2)]. 100
The values of %M obtained
analysed
were plotted
against
known ink entries (see Fig. 1).
2.3.

for all samples
the age of the

Thin-layer chromatography

To obtain an “ageing curve”, samples as two
l-cm slivers of ink of approximately
equal thickness were taken from five entries of known ages
(X,-X,).
Three more samples were taken from
a 2-year-old
entry
that was analysed
as a
questioned
(Q) entry. Each sample was placed in
a micro vial and extracted for 2 min with 15 ~1 of
chloroform,
stirring with a needle. A calibration
standard solution was prepared by the treatment
of eight l-cm slivers taken from a l-month-old
entry with 60 ~1 of chloroform.
Volumes of 10 ~1
of the obtained extracts and 5,8,11 and 15 ~1 of
the calibration
standard
solution
(calibration
standards,
S,-S,)
were applied to a 20 x 10 cm
precoated
Merck HPTLC silica gel 60 F,,, plate
as X-mm bands by means of a Camag Linomat-3
applicator.
One-dimensional
ascending
development was performed
with ethanol-acetone-hex-

V.N. Aginsky.

I J. Chromatogr.

ane (1:5:20, v/v/v). The development distance
was 50 mm.
The resulting chromatograms contained zones
of two ink components,
A and B. For five
samples, X,-X,, the relative proportion of these
components was obviously linked with the age of
the ink (see Fig. 2). The chromatograms were
scanned densitometrically
by reflectance in the
absorbance mode for fluorescence quenching at
254 nm using a Camag TLC/HPTLC
scanner
(with a mercury lamp, monochromator
bandwidth 30 nm, slit dimensions 0.3 x 5 mm and
scanning speed 1 mm/s) connected to an SP4100
integrator (Spectra-Physics).
The densitometric
data obtained were evaluated with external standards in the following way.
For the calibration standards S,-S,, it was
assumed that the contents of components A and
B per zone (their real values are unknown, as
follows from the procedure used for preparing
the calibration standard solution) were equal to
the corresponding values of the volumes of the
calibration standard solution applied to the plate
(see Table 1).
For each calibration standard component, A
and B, a logarithmic (this function gave the best
correlation coefficient in all non-linear calibrations that were tested in the given case) approximated calibration graph was constructed and
then the contents C, and C, of components A
and B per zone were determined for the chromatograms of the samples taken from the known
and Q aged entries.
The ratio C, lC, was calculated for each

A 678 (1994) 119-125

121

sample. The values obtained
were plotted
against the actual age of the known ink entries
and the age of the Q ink entry was determined
(see Table 3 and Fig. 3).

3. Results and discussion
3.1. Gas chromatography
Figure 1 shows ageing curves obtained for
Soyuz ballpoint inks of different colours having
similar compositions of colourless components.
The curves show that significant ageing taking
place over a period from about 6 months to more
than 2 years for different inks. After this period
until the age of 15 years the extent of the
extraction of the volatile component, phenoxyethanol, from the ink entries remained at about
20? 10%.
An explanation of this result characterizing the
mechanism of evaporation of volatile compophenoxynents
(such as phenoxyethanol,
ethoxyethanol
and other high-boiling vehicles
frequently used as the ingredients of ballpoint
inks) from ageing inks has been given previously
[5]. It was considered that the evaporation process includes a limiting stage of diffusion of a
vehicle from the interior layers of the ink body
PERCrn

Kx!lmcl!IoN
mp1)

Table 1
Contents of components A and B in the chromatographic
zones of the calibration standards
Standard”

Volume
applied
W)

Component content
(mg per spot)
A

S,
S*
S,
S,

5
8
11
15

B

5
8
11
15

5
8
11
15

a Extract from the l-month ink entry.

Aat

CF INK

(Years)

Fig. 1. Ageing curves obtained for violet, blue, green and
black Soyuz ballpoint inks: 1 and 2 relate to the maximum
and minimum values of %M obtained for the inks analysed.

122

V.N. Aginsky.

I J. Chromatogr.

to the surface of the film. For the same vehicle
and thickness
(depth) of the ink film, the efficiency
of the diffusion
process
is mainly
a
function of the nature of ingredients
of inks such
as resins and polymers.
Moreover,
if the resin is
capable
of polymerizing,
i.e., of cross-linking,
the diffusion
process slows as an ink ages on
paper,
and at a certain stage of ageing it can
virtually stop. For this reason the remaining
ink
volatile components
can be detected
in the ink
line even after a long period of time; this is
shown in Fig. 1 for up to 15year-old
entries
written with Soyuz ballpoint
inks.
In such situations,
a “weak”
solvent
(with
regard to hardened
ink resins), being unable to
penetrate
inside an old ink line, extracts the ink
volatile components
only from its exterior layers.
However,
the newer the ink, the more exterior
layers of the ink become available
to the weak
solvent,
and hence a greater
amount
of the
volatile components
is extracted.
Fig. 1 is a good illustration
of the above
observation
that the extraction
efficiency
of a
“weak”
solvent decreased
from about 90% for
fresh writings to about 20% for old writings.
It should be noted that the proposed
method
includes
also an important
stage that is carried
out if the values of %M determined
for the Q
ink entry are larger than cu. 60%. In this event,
another
sample (l-cm sliver) is taken from the
ink entry, heated moderately,
e.g., at 80°C for 5
min,
and analysed
as described
under
Experimental.
The percentage
extraction
value,
%M,, is calculated
for the heated sample and
compared
with the value of %M that was determined
for the unheated
sample. If the difference between
%M and %M, is ca. 10% or
larger, it can be concluded
that the ink entry
analysed
is a fresh one. If the difference
is less
than lo%, it means that a more suitable “weak
solvent”
should be chosen for a given ink.
For example. as a result of studying the ageing
process
of many
ballpoint
inks of different
formulae
by using the proposed
method,
it has
been established
that if, for a given ink, the
analytical
results
are %M > 70% and %M then the age of the ink
%O”~=XO”C.min > lo%,
5
analysed is less than cu. 6 months (depending
on

A 678 (1994) 119-125

the ink formula, this value may decrease to cu. 2
months).
As an example, Fig. 1 shows the results of the
age determination
obtained
for the Q entry (in
fact, it was a 3-month old entry written with a
Soyuz blue ballpoint
ink) using the proposed
method.
The method demonstrated
high efficiency
in
many actual case situations
when it was necessary to determine
whether the age of the Q entry
was less than several months or not less than 1
year. Such cases are fairly typical when the
investigator
suspects
that the given entry or
signature
was made after the time the investigation began. Some similar examples have been
presented
by Cantu [ 111.

3.2.

Thin-layer chromatography

Fig. 2 demonstrates
the view under UV illumination of the fragment
of the thin-layer
chromatogram
(without
the chromatographic
zones
of the paper’s ingredients)
and corresponding
densitograms
obtained
for samples taken from
entries written with Parker blue ballpoint
ink. A
and B represent
separated
colourless
components of interest in the ink analysed.
It is clearly seen in Fig. 2 that there is an

I

W light
(254 nm)
_______________________-___________

(B

----

_

a_---

I Front

1

Fig. 2. Fragment of the thin-layer chromatogram
(UV detection at 254 nm) and corresponding
densitograms
obtained for
Parker blue ballpoint ink entries of different ages. S,-S, are
calibration
standards:
Q relates to an entry of questionable
age; X,-X,
relate to known ink entries: X, = 1 day, XL = 1
month, X, = 1 year. X, = 3 years, X, = 6 years old.

V.N. Aginsky.

I J. Chromatogr. A 678 (1994) 119-125

obvious link between the relative proportion of
substances A and B and the age of the ink
writings examined: the substance A/substance B
ratio gradually increases as the ink ages (it is a
minimum for a fresh, l-day-old entry, the X,
track, and maximum for a 6-year-old entry, the
X, track).
As a rule, the relative proportions of the
components separated by TLC are evaluated by
obtaining related densitometric data and further
by calculating the ratios of the components’ peak
signals [l-3,8,9].
However, this approach has
been shown to produce
erroneous
results
because, in densitometric TLC, when chromatograms are scanned by reflectance in ‘the absorbance mode, the relationship between signal
output (peak height or peak area) and the
content of a separated zone is hardly ever a
directly proportionality
[ 10,12,13].
In this connection, a more reliable approach is
offered here. It can be considered as a version of
the external standard method for evaluating thinlayer chromatograms for cases typical in forensic
analysis when information on the quantitative
and even qualitative composition of samples to
be analysed is not available and, therefore,
calibration graphs cannot be obtained for the
analytes. (Another way to avoid erroneous results produced by the signal-to-signal ratio technique includes the application of the approach
based on the mass-independent
version of the
peak ratioing technique [10,13].)
The proposed method allows one to obtain the
actual mathematical functions of signal versus
content for any two components, A and B, of
the materials analysed within a certain calibration range of the contents of these components,
This calibration range is formed by
Cmin-Cmaxe
applying at least four or five calibration standards on a TLC plate as follows.
If samples are sprayed on as narrow bands,
different volumes of only one standard solution
can be applied to form a calibration range,
An important characteristic of the
‘min
coax.
method is that the real values of Cmin and C,,,
can be unknown to the examiner: only the values
of c max/Cmin and CiI&,
(where i relates to a
calibration standard characterized by the content

123

of a component per zone that is less than C,,,
and larger than Cmin) must be known, as was
described under Experimental.
If samples are applied as spots, calibration
standards should be prepared in different concentrations and spotted as a fixed constant volume: multiple spotting of a single standard
solution to generate a calibration graph is not
acceptable for accurate quantification as there is
no simple correlation between signal response
for a constant amount of substance and spot size
in scanning densitometry [12]. In this case, the
contents of components A and B per zone are
assumed to be equal (or directly proportional, if
only dilution factors, not real concentrations, are
known for the calibration standard solutions) to
the corresponding concentrations of the calibration standard solutions. Hence the value of the
ratio of the contents of any component in the
chromatographic zones corresponding to any two
prepared calibration solutions will be equal to
the value showing how many times one of these
solutions is more (or less) concentrated than the
other.
Further, for each component A and B, an
appropriate approximation function is found and
used as a calibration function for calculating the
contents, C, and C,, of components A and B
per spot of the samples taken from the entries of
the known and questionable
ages. Although
these content values are not real, this is not
sufficient for the considered method: the main
point is that the ratios of these values, C,IC,,
are independent of mass, in contrast to the peak
ratioing technique that is based on using the
mass-dependent
values of signal, /signal, (see
Table 2).
Tables 2 and 3 show peak-height values calculated by an integrator for the separated components of the Parker ink analysed, PH, (for
component A) and PH, (for component B), and
the values characterizing the relative proportions
of the components A and B calculated by using
the peak rationing technique (fourth column)
and the proposed “content” ratioing method
(last column).
Fig. 3 shows ageing curves obtained for the
Parker ink by plotting the values of the ratios

V.N. Aginsky.

124
Table 2
Data obtained
Standard

for calibration

s,

A 678 (1994) 119-12.5

standards

reading”
PH,

31032
51 716
73 261
88 782

90
99
107
112

506
126
745
056

Mean
R.S.D.

Ratio of
peaks,
PH, IPH,

Ratio of
contentsa.

0.34
0.52
0.68
0.79

1.02
0.98
0.96
1.04

0.58
0.34

Integrator
PH,

S,
S,
S,

I J. Chromaiogr.

1.00
0.04

C,‘CB

a Peak heights, PH, and PH,, were plotted against the contents C, and C, (see Table I). As a result, the following regression
equations
and correlation
coefficients
(r”) of the logarithmic
calibration
graphs were obtained:
PH, = -56 496 + 53 529 log C,,
(r” = 0.9929) and PH, = 57 816 + 20 258 log C, (r’ = 0.9894). Using these equations,
the values of CA and C, were recalculated
for each standard,
S,-S,,
and used for calculating
the content ratio values listed in the last column.

listed in the last two columns of Table 3 against
the actual age of the known ink entries.
The results of determining
the age of the Q
entry (in fact, the age of this entry was 2 years)
are also shown in Fig. 3 and presented
in Table
3.
It is clearly follows from Fig. 3 and the data in
Tables 2 and 3 that, in comparison
with the peak

Table 3
Data obtained
Ink
entry

ratioing technique,
the proposed
mass-independent content ratioing method gives a significant
increase in the accuracy and precision of ink age
determination.
It should also be noted that this method can be
successfully
applied to a comparative
TLC cxamination
of similarly
coloured
inks, paints,
fibres and other materials of forensic interest, as

for ink entries
Integrator

reading

PI-I,

Xl
XZ
%
X,
Xi
Q

Age determined

PH,

Ratio of
peaks,
PH,/PH,,

39 659
64 644
70 73s
84 472
88351

110 334
103 435
94817
94 388
93 093

0.36
0.63
0.84
0.8’)
0.95

0.45
1.01
2.04
2.27
2.60

Xl 887
87 963
SO 168

93 529
96 975
Y3 394

0.88
0.91
0.85

2.26
2.14
2.20

1.3
2.6
4.1

1.4
2.4
3.0

for the Q entry

(years)

Ratio of
contents,
c*ic,

V.N. Aginsky.

I J. Chromatogr.

NASS-DEPmwT

NASS-INDKPmm
NAT10 OF “CONTmrs”

MT10 OF PEAK SIGNALS(b)

(0)

2.5

1

.o

I
QL\
2.0
I

.-

:

lQ
/

:

0.8

1.5
:

1.0

Ii

0

0.7

I

x

:

AGE OF INK

Fig. 3. Ageing curves
the content
ratioing
technique.

obtained
method

(Years)

A 678 (1994)

119-125

125

The method using TLC allows the detection of
age changes in resins and other non-volatile ink
components. It includes a new procedure for
evaluating thin-layer chromatograms
of separated ink components. Being mass-independent,
this procedure gives much more correct results
for dating inks than those obtained with the aid
of the widely used signal-to-signal ratio technique.
Both methods, together and separately, have
been used in many actual case situations and the
results of the examinations have been accepted
as conclusive evidence by courts.
Further work is necessary to evaluate the
limits of the applicability of the methods to
numerous inks that are on the market.

for the Parker ink using (a)
and (b) the peak ratioing

References
Cantu
and R.S. Prough,
J. Forensic
Sci., 32
(1987) 1151.
PI R.L. Brunelle and A.A. Cantu, J. Forensic Sci., 32
(1987) 1522.
[31 R.L. Brunelle, J. Forensic Sci., 37 (1992) 113.
[41 L.F. Stewart, J. Forensic Sci., 30 (1985) 405.
PI V.N. Aginsky, J. Forensic Sci., 38 (1993) 1134.
PI V.N. Aginsky, in Proceedings of the 13th IAFS Meeting,
Diisseldorf, Germany, August 1993, in press.
171 A.A. Cantu, J. Forensic Sci., 33 (1988) 744.
181 R.L. Brunelle and H. Lee, J. Forensic Sci., 34 (1989)
1166.
[91 G.M. Golding and S. Kokot, J. Forensic Sci., 35 (1990)
1310.
1101 V.N. Aginsky, J. Forensic Sci., 38 (1993) 1111.
IllI A.A. Cantu, Anal. Chem., 63 (1991) 847A.
1121 C.F. Poole and S.K. Poole, J. Chromatogr., 492 (1989)
539.
u31 V.N. Aginsky, J. Planar Chromatogr., 4 (1991) 167.

[II A.A.

its discriminating power is much greater than
that usually produced by the widely used massdependent signal-to-signal ratio technique.

4. Conclusions
Two complementary methods for dating ballpoint inks have been considered. The method
using GC allows discrimination between fresh
(age not greater than a few months) and old
ballpoint inks, including inks with formulae
unknown to the examiner. It is effective for
analysing ballpoint inks that contain phenoxyethanol, phenoxyethoxyethanol
or similar highboiling vehicles.

Exhibit 21

Declaration of Larry F. Stewart

Exhibit 22

Declaration of Larry F. Stewart

Exhibit 23

Declaration of Larry F. Stewart

J Forensic Sci, Jan. 2004, Vol. 49, No. 1
Paper ID JFS2003217–491
Available online at: www.astm.org

TECHNICAL NOTE
Gerald M. LaPorte,1 M.S.F.S.; Jeffrey D. Wilson,3 B.S.; Antonio A. Cantu,2 Ph.D.;
S. Amanda Mancke,3 B.S.; and Susan L. Fortunato,1 B.A.

The Identification of 2-Phenoxyethanol in Ballpoint
Inks Using Gas Chromatography/Mass
Spectrometry—Relevance to Ink Dating∗

ABSTRACT: Developing and implementing a generally accepted procedure for the dating of ink found on documents using dynamic approaches
has been a very formidable undertaking by forensic document examiners. 2-Phenoxyethanol (PE), a common volatile organic compound found in
ballpoint inks, has been recognized for over a decade as a solvent that evaporates as ink ages. More recently, investigations have focused on the
solvent loss ratio of PE prior to and after heating. To determine how often PE occurs in ink formulations, the authors analyzed 633 ballpoint inks
utilizing a gas chromatograph/mass spectrometer. 2-Phenoxyethanol was identified in 85% (237/279) and 83% (293/354) of black and blue inks,
respectively.
KEYWORDS: forensic science, questioned documents, phenoxyethanol, ballpoint ink, ink dating, ink aging, GC/MS, volatile analysis, SPME

An examination to determine the age of ink on a document can
be quite challenging. Cantu (1,2) outlines two approaches to determine the age of ink on a questioned document. The first of these
is the static approach, which generally applies to methods based
on the comparison of various ink components to a standard reference collection to determine the first date of production. In fact,
the United States Secret Service (USSS) and the Internal Revenue
Service (IRS) jointly maintain the largest known collection of writing inks from around the world. These inks date back to the 1920s
and include over 8000 inks obtained from various manufacturers
throughout the world. Annually, contact is made with the pen and
ink manufacturers requesting that they send any new formulations
of inks, along with appropriate information, so that the submitted
standards can be chemically tested and added to the reference collection. Writing pens are also obtained from the open market and
compared to the library of standards to verify and identify additional
inks. This is a formidable task that obviously requires significant
resources and maintenance. Indeed, this is not always a practical
solution for every forensic facility to achieve.
Ideally, ink tags would be the most reliable method for the dating
of inks. Tags can be added to formulations in the form of fluorescent
1 United States Secret Service, Forensic Services Division, Questioned Document Branch, 950 H Street NW, Washington, DC.
2 United States Secret Service, Forensic Services Division, Research Section,
950 H Street NW, Washington, DC.
3 United States Secret Service, Forensic Services Division, Questioned Document, chemist contractor, 950 H Street NW, Washington, DC.
∗ This work was presented, in part, at the Mid-Atlantic Association of
Forensic Scientists Annual Meeting, May 8, 2003, Annapolis, MD. All references pertaining to manufacturers and their products do not imply endorsement
by the United States Secret Service or the authors.
Received 21 June 2003; and in revised form 4 Sept. 2003; accepted for publication 4 Sept. 2003; published 17 Dec. 2003.

Copyright

C

compounds or rare earth elements and were evident in some formulations from about 1970 until 1994. Factors have precluded some
ink manufacturers from participating in such a program, including,
but not limited to, insufficient resources, low priority, and/or disagreement about the type of tag utilized. This is not to say that a
widespread tagging agenda is not achievable. On the contrary, efforts do continue to convince ink companies to add tags to their
formulations. As recently as November 2002, a dominant ink manufacturer has begun adding tags to their ink in collaboration with
the U.S. Secret Service.
With stringent demands from the forensic community to develop
and validate scientifically reliable laboratory techniques, implementing other methods for ink dating is an arduous endeavor as
well. Such methods include those involving the dynamic approach,
which incorporates procedures that measure the physical and/or
chemical properties of ink that change with time. The changes that
occur over a given period of time can generally be referred to as
aging characteristics. Ballpoint inks mainly consist of colorants
(dyes and/or pigments) and vehicles (solvents and resins). There is
also a wide array of other ingredients, which may include antioxidants, preservatives, and trace elements, but these are typically a
small component of the overall ink composition. Nevertheless, the
importance of their presence should not be diminished since the
combination of all ingredients may play a pivotal role in the aging characteristics of an ink formulation. However, the subject of
this paper will focus on the vehicles found in ballpoint inks. More
specifically, the authors have chosen to investigate a single volatile
compound that has been reported by the industry to be in many
formulations of inks.
Volatile analysis of ballpoint inks, using GC/MS, for determining the age of inks on paper has been studied and reviewed in the
literature for more than a decade (3–8). These authors have laid the

2004 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.

Copyright by ASTM Int'l (all rights reserved); Tue Sep 15 09:34:34 EDT 2009
Downloaded/printed by
Tina Tam (Good and Cormier) pursuant to License Agreement. No further reproductions authorized.

1

2 JOURNAL OF FORENSIC SCIENCES

FIG. 1—The chemical structure and formula for 2-phenoxyethanol.

groundwork for what may be a very promising dynamic approach
to the future of ballpoint ink age determination. These works have
honed in on the analysis of 2-phenoxyethanol (PE), a common
volatile organic compound found in some ballpoint writing inks.
2-Phenoxyethanol, also referred to as ethylene glycol monophenyl
ether, 1-hydroxy-2-phenoxyethane, beta-hydroxyethylphenyl ether,
Dowanol EP, and Phenyl Cellosolve, is a glycol ether and is used as
the principal solvent in many ballpoint ink formulations. It is a colorless, slow evaporating, viscous liquid with a faint aromatic odor
and is used in most ballpoint ink formulations because it is stable in
the presence of acids and alkalis. It is also nonhygrosopic (does not
absorb water, making it amenable to hot, humid climates), nonhazardous, economical, and especially good at solubilizing resins and
nigrosine (a common solvent soluble black dye used in the writing
ink formulations). It is recognized as Chemical Abstracts Service
(CAS) number 122-9-6 and has a molecular weight of 138.17 with
a boiling point of 245.2◦ C (9). Figure 1 depicts the chemical formula and structure of PE. Beshanishvily et al. (4) were the first to
discuss the identification of PE as it relates to the aging of inks.
Since then, Aginsky (5) reported that, “. . . significant aging [takes]
place over a period of about 3 months. After this period until the age
of 15 years the extent of the extraction of the volatile component
(phenoxyethanol) from the ink entries has been kept at a level about
20%.” Aginsky also describes the ink-drying process and surmises
that volatile components stop emitting from a dried sample of ink
until they are freed by heating or a solvent extraction.
More recently, Gaudreau and Brazeau (10) presented their findings on an extensive research effort that focuses on how PE levels
change over time following an ink entry placed on paper. They
discuss solvent loss and state that the “. . . phenoxyethanol in ink
evaporates at a high rate during the first six to eight months following its application on paper. The rate of evaporation stabilizes
over a period of six to eighteen months. This process is no longer
significant after a period of about two years.” Given the chemical
properties of PE, its loss due to evaporation is most affected by
heat. With these caveats, they developed a dynamic approach to ink
dating that incorporates comparing the PE ratio of an ink prior to
and after heating.
In addition, Brazeau et al. (11) have experimented with solid
phase micro-extraction (SPME), which utilizes a specially coated
silica fiber that is mounted in a syringe-like device. A small glass
vial is placed over the ink entry with the SPME device inserted in the
sealed environment. Volatile solvents that emit from the ink adsorb
onto the fiber for a set time, i.e., until an equilibrium is achieved
within the system. The fiber is then withdrawn and injected into
a gas chromatograph (GC), whereby the volatile components are
desorbed due to the high temperature (e.g., 250◦ C) in the injection
port. The analytes are then separated in the GC and identified using
an appropriate analytical instrument such as a mass spectrometer
(MS). SPME has proven to be an efficient and effective method for
the extraction of volatile components (12,13) and has been utilized

in the authors’ laboratory for the detection of PE in some ballpoint
inks.
Chemical analysis of writing inks by means of thin layer chromatography (TLC) is viewed by the scientific community as a valid
procedure to compare inks (14–18). Since TLC is an effective and
efficient method for separating and identifying various colored components such as dyes, and nearly all ink formulations are proprietary,
forensic examinations that employ TLC analysis are invaluable. For
instance, two or more questioned inks can be compared to determine if they are the same, or questioned inks can be associated to
a known standard to determine the age of an ink, i.e., the static approach. With respect to this latter instance, appropriate information
and documentation acquired from a manufacturer regarding their
ink will allow a forensic document examiner to make significantly
reliable conclusions, assuming there is access to a thorough ink
collection. Although obtaining a supplemental volatile profile may
increase the degree of discrimination, limitations include solvent
loss over time or other external factors such as exposure to high
temperatures, light, and/or humidity.
With the benefit of having a large collection of standards, the
authors determined that it would be advantageous to begin conducting volatile profiles of writing inks to investigate the percentage
of ballpoint inks that actually contain PE since it is an important
compound of interest for the determination of ink age. An extensive
search of the literature was conducted, but no studies investigated
a large population of inks to determine how often PE is present
in ballpoint ink formulations. Thus, the focus of this paper will
be on the examination for the presence of 2-phenoxyethanol in
633 ballpoint inks.
Materials and Methods
Ink Standards
As stated, ink standards are received by the USSS from all over
the world and date back to the 1920s. As new ink formulations
are received, samples of the ink are placed onto WhatmanTM filter
paper No. 2 (also referred to as scribble sheets), allowed to air dry,
placed in a protective sheet and binder, and finally stored in dark
cabinets to ensure minimal degradation due to environmental factors
such as light, temperature, and humidity. Many of the ink standards
are received as a liquid in a bottle and permanently retained, and
others are received in pens, pen refills, or as samples on paper. For
this study, whenever possible, ballpoint ink samples in liquid form
were analyzed directly from the bottle or pen. Other ink samples
were taken off the scribble sheets; however, volatile profiles of
scribble sheet samples were closely examined to determine if they
were suitable to include in the study since some were over 30 years
old. This topic will be discussed under Results and Discussion. A
total of 279 black ballpoint inks from 31 companies and 354 blue
ballpoint inks from 26 companies were chosen for analysis using a
PerkinElmer TurboMassTM gas chromatograph/mass spectrometer.
Extraction
Liquid inks were sampled utilizing a disposable capillary glass
pipette in order to minimize sample handling. The pipette was then
placed into a glass vial containing 1 mL of acetonitrile. The ink and
solvent were agitated/stirred to ensure a homogenous mixture. Dried
ink samples from scribble sheets were sampled using a 5-mm hole
punch. The punches were taken from a highly dense area and allowed to extract in a vial with 1 mL of acetonitrile for approximately
1 min. The solvent was decanted and placed into a separate vial.

Copyright by ASTM Int'l (all rights reserved); Tue Sep 15 09:34:34 EDT 2009
Downloaded/printed by
Tina Tam (Good and Cormier) pursuant to License Agreement. No further reproductions authorized.

LAPORTE ET AL.

r ANALYSIS OF 2-PHENOXYETHANOL 3

FIG. 2a—The total ion chromatogram for 2-phenoxyethanol standard (J.T. Baker TM product No. T-319-07).

Gas Chromatography/Mass Spectrometry (GC/MS)
All of the extracted ink samples were analyzed using an auto
sampler attached to a PerkinElmer TurbomassTM GC/MS. Onemicroliter samples were injected into the GC. The column used
was an HP-5 (30 m × 0.32 mm × 0.25 µL) cross-linked 5% phenylmethylsiloxane. The injector temperature was set to 260◦ C and the
flow rate was 1.2 mL/min split mode at 20 mL/min. The temperature program started at 50◦ C for 1 min and increased at a rate
of 10◦ C/min to 200◦ C with a 2-min hold. The second rate was
25◦ C/min to 300◦ C with a final 2-min hold. The mass spectrometer
detector was set for full scan from 1.8 to 24 min with the 1.8-min
delay set to begin following the solvent elution, i.e., solvent delay.
The detector was programmed to scan compounds ranging from 28
to 500 atomic mass units (amu).
Results and Discussion
A review of the standards library indicated that at the inception
of this project there were 516 black ballpoint inks from 53 companies and 854 blue ballpoint inks from 65 companies. All the liquid
ink samples that were obtained from bottles were extracted and exhibited significant volatile profiles with sufficient peak abundance
for accurate integration. In contrast, there were numerous scribble
sheet samples that did not produce a significant, or very limited,
chromatographic profile. The lowest level of detection for sufficient interpretation was estimated to be 0.1 ppb. It was determined
that insignificant peak area was the result of the age of the ink on
the scribble sheet (e.g., some scribble sheets were 20 to 30 years
old). The results for the samples determined to have poor chromatographic profiles were recorded, but not used to calculate the
statistics presented in this paper. A total of 279 black ballpoint inks

FIG. 2b—The mass spectrum for Peak 1 at retention time (RT) = 8.79 min.

FIG. 2c—The mass spectrum for Peak 2 at RT = 12.45 min.

from 31 companies and 354 blue ballpoint inks from 26 companies were determined to have significant chromatographic profiles
necessary for peak integration and, hence, accurate identification
of chemical composition. 2-Phenoxyethanol was identified in 85%
(237/279) and 83% (293/354) of the black and blue inks, respectively. Furthermore, 20 of the 31 companies that have manufactured
black ballpoints used PE in the vehicle in 100% of their ballpoint
formulations and 11 of the 26 manufacturers incorporated PE in all
of their blue ballpoint inks.
The 2-phenoxyethanol standard (J.T. Baker product No. T31907, Lot No. N35622) contained two major peaks on the TIC (total
ion chromatogram), and the GC/MS results are depicted in Figs. 2a,
2b, and 2c. In addition to the PE at retention time (RT) 8.79 min,

Copyright by ASTM Int'l (all rights reserved); Tue Sep 15 09:34:34 EDT 2009
Downloaded/printed by
Tina Tam (Good and Cormier) pursuant to License Agreement. No further reproductions authorized.

4 JOURNAL OF FORENSIC SCIENCES
2-(2-phenoxyethoxy)ethanol, also referred to as diethylene glycol
phenyl ether (DGPE) or phenyl carbitol, was detected at a RT of
12.45 min. It is identified as CAS number 104-68-7. The SigmaAldrich catalogue (2003-2004) indicates that Dowanol EPH, i.e.,
PE, can contain up to 10% DGPE.
DGPE was detected in 21.5% (60/279) and 12.1% (43/354) of
the black and blue ballpoint inks that contained PE, respectively.
There was no evidence that the level of DGPE was directly related
to the level of PE (e.g., DGPE was occasionally absent in samples
with relatively high levels of PE and was present in samples with
relatively low levels of PE). However, no further study was conducted to examine if the ratio of PE/DGPE changed with aging.
The authors did infer that there may be differences in the composition of 2-phenoxyethanol that may be attributable to a particular
chemical manufacturer. Indeed, this information could be utilized
to differentiate ink manufacturers depending on their supplier.
GC/MS is an obviously powerful analytical tool, not only for the
dating of inks, but for the identification of some components of
inks. Brunelle and Crawford (19) recently wrote, “GC-MS shows
great promise for strengthening ink identifications, because it can be
used to identify both volatile and non-volatile ingredients of inks.”
Accordingly, the authors were cognizant of this at the outset of the
study and maintained data of all the identifiable components in the
633 black and blue ballpoint inks to determine if there are chemical
class characteristics specific to a manufacturer. A thorough review
of the results and all subsequent conclusions pertaining to the use of
GC/MS to profile company ink formulations was considered a secondary objective. The authors determined that this analysis would
be better suited in a future work with an extensive and dedicated
discussion to the GC/MS analysis of a large population of ballpoint
inks.
The analysis of volatile components such as PE to determine the
age of an ink is promising, especially when using methods that are
based on the relative loss of a solvent between heated and unheated
samples. There are different scenarios of a document examination
that an examiner may encounter that will significantly affect the
degree of qualification of a conclusion. For example, an examiner
may be requested to compare two or more inks on the same document to determine if they were produced contemporaneously. If
prior examinations indicate that the inks are matching formulations,
but they are suspected of being made at two different time periods
(e.g., several months apart), then factors such as storage conditions,
the type of paper, exposure to a variety of environments, and differences in formulation should not preclude a forensic examiner from
making conclusions with limited qualifications. It is important to
note that the Merck Index (9) indicates that PE is used as a fixative
in perfumes, which would require handlers to be cognizant so as
not to possibly contaminate a questioned document. As well, the
approaches discussed in this paper are not mass independent when
sampling the ink; therefore, care and accuracy need to be administered when removing ink plugs for analysis. One final caveat that
requires some consideration is the rate of PE migration on paper
once an entry is made. Since PE is a liquid solvent, it is feasible to
ascertain that it may dissipate through the paper into a questioned
entry if ink is present on the reverse side of a page. Ink may also migrate from nearby adjacent entries, but taking blank samples (e.g.,
samples of the paper with no ink) in close proximity may aid the
examiner in understanding the extent of PE migration.
Another scenario that may be encountered is the analysis of ink
entries on a document that are not of the same formulation. Although one may argue that the document is likely to have been
stored under the same conditions, the level of PE may exist in different levels in different formulations from the same manufacturer.

Finally, one may be requested to date entries on multiple pages that
are part of the same document submission (e.g., multi-page wills
or contracts) to determine if they were produced at, or around, the
same time. Differences in paper, storage conditions, and how the
document is arranged (e.g., the presence of ink solvents on subsequent or overlying pages that transfer to adjacent pages) should be
taken into consideration. Indeed, more research and validation into
these unknown effects will be fundamental in developing standard
allowable variations. Standard error can then be incorporated to account for human error and experimental deviation that are necessary
to make qualified conclusions of forensic significance.
Conclusion
The identification of PE in over 80% of black and blue ballpoint
ink formulations has shown that studies investigating PE as it relates
to the aging of writing inks have been and continue to be significant.
As our field undergoes necessitated scrutiny of forensic examinations, GC/MS is an excellent and well-proven analytical tool for
the identification and quantification of chemical compounds. Validation of the instrumentation and the procedures utilized to identify
PE should therefore be minimal. This will allow future researchers
to concentrate their efforts on the development and implementation
of a generally accepted procedure for a dynamic approach to ink
dating.
Acknowledgments
The authors wish to extend their gratitude to Dr. Ben Fabian from
Sensient Imaging Technologies for the information he provided regarding the use of 2-phenoxyethanol in the writing ink industry. As
well, we wish to thank Mr. Marc Gaudreau and Mr. Luc Brazeau of
the Canada Customs Revenue Agency for their efforts to accommodate our many questions regarding the analysis of volatile solvents
in ballpoint writing inks.
References
1. Cantu AA. A sketch of analytical methods for document dating. Part I:
The static approach: determining age independent analytical profiles. Int
J Forensic Doc Examiners 1995;1(1):40 –51.
2. Cantu AA. A sketch of analytical methods for document dating. Part II:
The dynamic approach: determining age dependent analytical profiles.
Int J Forensic Doc Examiners 1996;2(3):192–208.
3. Stewart LF. Ballpoint ink age determination by volatile component
comparison—a preliminary study. J Forensic Sci 1985;30(2):405–11.
4. Beshanishvily GS, Trosman EA, Dallakian PB, Voskerchian GP. Ballpoint ink age—a new approach. Proceedings of the 12th International
Forensic Scientists Symposium; 1990 Oct 15–19; Adelaide, Australia.
5. Aginsky VN. Some new ideas for dating ballpoint inks—a feasibility
study. J Forensic Sci 1993;38(5):134–50.
6. Aginsky VN. Determination of the age of ballpoint ink by gas and densitometric thin-layer chromatography. J Chromatogr A 1994;678:117–25.
7. Aginsky VN. Dating and characterizing writing, stamp pad and jet printer
inks by gas chromatography/mass spectrometery. Int J Forensic Doc
Examiners 1996;2(2):103–15.
8. Aginsky VN. Measuring ink extractability as a function of age—why
the relative aging approach is unreliable and why it is more correct to
measure ink volatile components than dyes. Int J Forensic Doc Examiners
1998;4(3):214–30.
9. Budavari S, editor. The merck index: an encyclopedia of chemicals, drugs,
and biologicals. 12th ed. Whitehouse Station (NJ): Merck Research Laboratories Division of Merck & Co., Inc., 1996:7410.
10. Gaudreau M, Brazeau L. Ink dating using a solvent loss ratio method.
Proceedings of the 60th Annual Conference of the American Society of
Questioned Document Examiners; 2002 Aug 14 –18; San Diego (CA).

Copyright by ASTM Int'l (all rights reserved); Tue Sep 15 09:34:34 EDT 2009
Downloaded/printed by
Tina Tam (Good and Cormier) pursuant to License Agreement. No further reproductions authorized.

LAPORTE ET AL.

[PubMed]
[PubMed]

11. Brazeau L, Chauhan M, Gaudreau M. The use of micro-extraction in
the development of a method to determine the aging characteristics
of inks. Proceedings of the 58th Annual Conference of the American
Society of Questioned Document Examiners; 2000 Aug 24 –29; Ottawa,
Ontario.
12. Vu DTT. SPME/GC-MS characterization of volatiles associated with
methamphetamine: toward the development of a pseudomethamphetamine training manual. J Forensic Sci 2001;46(5):1014 –24.
13. Vu DTT. Characterization and aging study of currency ink and currency canine training aids using headspace SPME/GC-MS. J Forensic
Sci 2003;48(4).
14. Aginsky VN. Forensic examination of “slightly soluble” ink pigments using thin-layer chromatography. J Forensic Sci 1993;38:
1131–3.
15. Brunelle RL, Pro MJ. A systematic approach to ink identification. J AOAC
Int 1972;55:823–26.

r ANALYSIS OF 2-PHENOXYETHANOL 5

16. Kelly JD, Cantu AA. Proposed standard methods for ink identification.
J AOAC Int 1975;58:122–25.
17. The American Society for Testing and Materials (ASTM) E 1422-98:
Standard Guide for Test Methods for Forensic Writing Ink Comparison,
1998;530–7.
18. The American Society for Testing and Materials (ASTM) E 1789-96:
Standard Guide for Writing Ink Identification, 1996:722– 6.
19. Brunelle RL, Crawford KR. Advances in the forensic analysis and dating
of writing ink. Springfield (IL): Charles C. Thomas, 2003;164.
Additional information and reprint requests:
Gerald M. LaPorte, M.S.F.S.
United States Secret Service
Forensic Services Division
950 H Street NW
Washington, DC 20223
E-mail: [e-mail]

Copyright by ASTM Int'l (all rights reserved); Tue Sep 15 09:34:34 EDT 2009
Downloaded/printed by
Tina Tam (Good and Cormier) pursuant to License Agreement. No further reproductions authorized.