raman spectroscopy of ink on paper - PROBLEMS OF FORENSIC ...

RAMAN SPECTROSCOPY OF INK ON PAPER
Thomas ANDERMANN
Bundeskriminalamt, Wiesbaden, Germany
ABSTRACT: Blue and black ballpoint pen as well as blue, black, green and red fluid
<strong>ink</strong>s on paper are examined by Raman <strong>spectroscopy</strong>. The quality <strong>of</strong> the Raman spectra
is strongly dependent on the excitation wavelength. Therefore, two different
spectrometers with different excitation wavelengths have been tested. In addition,
Surface Enhanced Resonance Raman Spectroscopy (SERRS) was used. Application
<strong>of</strong> SERRS helps to avoid fluorescence and enhance certain Raman signals. Because <strong>of</strong>
the minute amount <strong>of</strong> reagent applied to the <strong>ink</strong>-line, this procedure may still be regarded
as non-destructive. The resulting Raman spectra are compared and classified.
Classification results are compared with a classification <strong>of</strong> the <strong>ink</strong>s by thin layer
chromatography. An evaluation <strong>of</strong> the results will show to what extent Raman <strong>spectroscopy</strong>
may be useful in questioned document examination.
KEY WORDS: Raman <strong>spectroscopy</strong>; Ink; Document examination.
Problems <strong>of</strong> Forensic Sciences, vol. XLVI, 2001, 335–344
Received 27 November 2000; accepted 15 September 2001
INTRODUCTION
Recently we see an increasing number <strong>of</strong> applications in Raman <strong>spectroscopy</strong>.
Thermoelectrically cooled CCD array detectors and techniques like
surface enhanced resonance Raman <strong>spectroscopy</strong> (SERRS) enable an increase
in sensitivity sufficient for applications in forensics [1]. In Raman
<strong>spectroscopy</strong> the analyte is excited with monochromatic laser-light. The
scattered light (rayleigh scattering) contains the excitation wavelength,
which can be removed by holographic filters, and signals at longer (Stokes
shift) and shorter wavelength (anti stokes shift). Usually the Stokes signals
are more intensive and therefore monitored as the Raman spectrum.
Position and intensity <strong>of</strong> Raman signals depend on the maximum absorption
wavelength <strong>of</strong> the analyte. Therefore appropriate excitation wavelengths
are needed. Fluorescence <strong>of</strong> the analyte or the substrate may cover
the Raman signals, which can be avoided by blocking the fluorescence.
In document examination we are mainly interested in <strong>ink</strong>s and dyes
therein. We are confronted with a great variety <strong>of</strong> dyes <strong>of</strong> different hues,
which means they have different absorption maxima. The dyes belong to different
chemical classes such as inorganic and organic pigments, acid and ba-

336 T. Andermann
sic dyes, substantive and reactive dyes, triphenylmethane dyes, azo dyes
etc.
EXPERIMENTAL
For this study we used two Raman spectrometers: A FORAM 685 from
Foster & Freeman Ltd. and a LabRam Infinity from Dilor GmbH. The
FORAM 685 <strong>of</strong>fers one excitation wavelength at 685 nm. The spectral range
is between 400 and 2000 cm –1 with a resolution <strong>of</strong> 8 cm –1 . The laser power on
the samples was about 1.5 mW. The LabRam Infinity <strong>of</strong>fers several laser
sources, from which we used 514, 633 and 785 nm. The spectral range we
used is between 60 and 3000 cm –1 with a resolution <strong>of</strong> 4 cm –1 . The laser power
on the samples can be tuned from 0 to 6 mW.
We examined <strong>ink</strong> lines drawn on white <strong>of</strong>fice paper <strong>of</strong> 26 blue and
26 black ball point pen <strong>ink</strong>s and 63 blue, 60 black, 36 green and 48 red fluid
<strong>ink</strong>s. The <strong>ink</strong>s were examined directly without any sample preparation and
after subsequently treating the samples with a solution <strong>of</strong> poly-L-lysine and
a silver colloid to obtain SERRS spectra (SERRS 1). In some cases additional
application <strong>of</strong> an ascorbic acid solution (SERRS 2) leads to better results.
RESULTS AND DISCUSSION
The discrimination power obtained by the FORAM 685 is shown in
Table I. In most cases the application <strong>of</strong> the SERRS reagents leads to a distinct
increase in the discrimination power. The decrease in case <strong>of</strong> the black
ball point pen <strong>ink</strong>s is due to the fact, that in the native spectra the shape <strong>of</strong>
the fluorescence can be used as a criterion for differentiation. The SERRS
spectra show no fluorescence.
TABLE I. DIFFERENTIATION OF INKS BY FORAM 685
Ink Raman SERRS1 SERRS2
Blue ball point 47.1% 80.3% –
Black ball point 57.8% 33.5% –
Blue fluid <strong>ink</strong> 47.4% 82.7% 84.2%
Black fluid <strong>ink</strong> 74.3% 92.4% 83.2%
Green fluid <strong>ink</strong> 34.6% 67.1% 72.7%
Red fluid <strong>ink</strong> 28.2% 87.9% 92.7%

Raman <strong>spectroscopy</strong> <strong>of</strong> <strong>ink</strong> on paper 337
Among optical examination TLC is commonly used for differentiation <strong>of</strong>
<strong>ink</strong>s. Figure 1 shows the TLC results <strong>of</strong> 19 blue ball point <strong>ink</strong>s some <strong>of</strong> which
we are able to distinguish and some which show identical chromatograms.
We take an exemplary look at the first four <strong>ink</strong>s. We cannot distinguish between
<strong>ink</strong>s 1 and 2 or <strong>ink</strong>s 3 and 4, but we can easily see the differences between
the two pairs.
Fig. 1. TLC results <strong>of</strong> 19 blue ball point <strong>ink</strong>s.
The Raman spectra obtained with the FORAM 685 and SERRS 1 lead to
the same conclusion: No differentiation between <strong>ink</strong>s 1 and 2 or <strong>ink</strong>s 3 and
4 respectively, but we see distinct differences between the two pairs (Figures
2–4).
Blue Ball Point 1
Blue Ball Point 2
2000 1916 1832 1748 1664 1580 1496 1412 1328 1244 1160 1076 992 908 824 740 656 572
Raman Shift Wavenumbers
Fig. 2. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 1 (blue) and 2 (magenta), FORAM 685,
SERRS 1.

338 T. Andermann
2000 1919 1838 1757 1676 1595 1514 1433 1352 1271 1190 1109 1028 947 866 785 704 623 542
Raman Shift Wavenumbers
Blue Ball Point 3
Blue Ball Point 4
Fig. 3. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 3 (blue) and 4 (magenta), FORAM 685,
SERRS 1.
2000 1919 1838 1757 1676 1595 1514 1433 1352 1271 1190 1109 1028 947 866 785 704 623 542
Raman Shift Wavenumbers
Blue Ball Point 3
Blue Ball Point 1
Fig. 4. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 3 (blue) and 1 (magenta), FORAM 685,
SERRS 1.
Figure 4 shows differences in the region <strong>of</strong> lower wavelengths between the
Raman spectra <strong>of</strong> blue points 1 and 2 obtained with the LabRam Infinity
(ex- 633 nm, SERRS 1). But it has not been ball checked yet, if these differences
are due to different spectroscopic properties <strong>of</strong> the <strong>ink</strong>s or caused by an
artefact. Figure 6 shows no difference between ball point 3 and 4. Whereas
Figure 7 shows a distinct difference between ball points 1 and 3.

Raman <strong>spectroscopy</strong> <strong>of</strong> <strong>ink</strong> on paper 339
1998
1929
1859
Blue Ball Point 2
Blue Ball Point 1
1788
1717
1645
1572
1499
1425
1350
1274
1198
1120
Figures 8 and 9 show superimposed spectra obtained with both different
spectrometers <strong>of</strong> ball point 1 and 3 respectively. The quality <strong>of</strong> the spectra
looks comparable. The LabRam Infinity <strong>of</strong>fers greater spectral range and
higher resolution, which may contain additional information.
An advantage <strong>of</strong> Raman <strong>spectroscopy</strong> compared with TLC is its non-destructive
application to documents. Furthermore, for TLC we need to extract
the <strong>ink</strong> from the paper, which is not always possible. Figures 10 and 11 show
SERRS 1 spectra <strong>of</strong> blue gel pen <strong>ink</strong>s, which are easily to distinguish.
1042
963
884
803
Raman Shift Wavenumbers
Fig. 5. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 1 (blue) and 2 (magenta), LabRam Infinity,
ex- 633 nm, SERRS 1.
1998
1923
1848
1771
1694
1616
1537
1458
1377
1295
Blue Ball Point 3
Blue Ball Point 4
1213
1130
1045
960
874
787
721
Raman Shift Wavenumbers
Fig. 6. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 3 (magenta), and 4 (blue), LabRam Infinity,
ex- 633 nm, SERRS 1.
699
639
556
609
472
519
387
428
301
335
214
242
126
147

340 T. Andermann
1998
1923
1848
1771
1694
1616
1537
1458
Blue Ball Point 3
Blue Ball Point 1
1377
1295
1213
1130
1045
960
874
787
Raman Shift Wavenumbers
Fig. 7. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 1 (blue) and 3 (magenta), LabRam Infinity,
ex- 633 nm, SERRS 1.
699
609
Blue Ball Point 1
FORAM 685 (SERRS1)
LabRam Infinity (SERRS1)
Fig. 8. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong> 1, FORAM 685 (magenta) and LabRam
Infinity (blue), SERRS 1.
519
428
335
242
147

Raman <strong>spectroscopy</strong> <strong>of</strong> <strong>ink</strong> on paper 341
Blue Ball Point 3
FORAM 685 (SERRS1)
LabRam Infinity (SERRS1)
Fig. 9. Raman spectra <strong>of</strong> blue ball point <strong>ink</strong>s 3, FORAM 685 (magenta) and LabRam
Infinity (blue), SERRS 1.
2000
1946
1892
Pentel Hybrid Roller
Standard B'Gel
Pilot G-1 100
1838
1784
1730
1676
1622
1568
1514
1460
1406
1352
1298
1244
1190
1136
1082
1028
Raman Shift Wavenumbers
974
920
866
812
758
704
Fig. 10. Raman spectra <strong>of</strong> blue gel pen <strong>ink</strong>s, Pentel Hybrid Roller (magenta), Standard
B’Gel (red), Pilot G-1 100 (blue), FORAM 685, SERRS 1.
650
596
542

342 T. Andermann
1998
1923
1848
1771
1694
1616
1537
1458
1377
1295
1213
1130
The advantage <strong>of</strong> testing several sample preparation techniques is
shown by comparing figures 12 and 13. Only after application <strong>of</strong> SERRS 2
the black Pilot G-1 100 gel pen shows Raman signals. The same results are
obtained with the LabRam Infinity (Figure 14).
1045
Pentel Hybrid Roller
Pilot G-1 100
960
874
787
Raman Shift Wavenumbers
Fig. 11. Raman spectra <strong>of</strong> blue gel pen <strong>ink</strong>s, Pentel Hybrid Roller (magenta), Pilot
G-1 100 (blue), LabRam Infinity, SERRS 1.
2000
1946
Zebra J-Roller Medium
Pilot G-1 100
Standard B'Gel
1892
1838
1784
1730
1676
1622
1568
1514
1460
1406
1352
1298
1244
1190
1136
1082
1028
974
Raman Shift Wavenumbers
699
920
866
609
519
812
758
428
335
704
650
242
596
542
Fig. 12. Raman spectra <strong>of</strong> black gel pen <strong>ink</strong>s, Zebra J-Roller Medium (blue), Standard
B’Gel (red), Pilot G-1 100 (magenta), FORAM 685, SERRS 1.
147

Raman <strong>spectroscopy</strong> <strong>of</strong> <strong>ink</strong> on paper 343
2000
1958
Zebra J-Roller Medium
Pilot G-1 100
Standard B'Gel
1916
1874
1832
1790
1748
1706
1664
1622
1580
1538
1496
1454
1412
1370
1328
1286
1244
1202
1160
1118
1076
1034
992
950
908
866
824
782
740
698
656
614
572
530
Fig. 13. Raman spectra <strong>of</strong> black gel pen <strong>ink</strong>s, Zebra J-Roller Medium (blue), Standard
B’Gel (red), Pilot G-1 100 (magenta), FORAM 685, SERRS 2.
1800
1760
1720
1680
1639
1599
1558
1517
1475
1434
1392
1350
1308
1265
Pilot G-1 100 633nm SERRS2
Pilot G-1 100 633nm SERRS1
1222
1179
1136
1092
1049
1004
960
916
Fig. 14. Raman spectra <strong>of</strong> black gel pen <strong>ink</strong> Pilot G-1 100, SERRS 1 (blue), SERRS 2
(magenta), LabRam Infinity, ex- 633 nm.
871
826
780
735
689
642
596
549
502
455
407
359
311
263
214
165
116
66

344 T. Andermann
CONCLUSIONS
Raman <strong>spectroscopy</strong> <strong>of</strong> <strong>ink</strong>s on paper is a virtually non-destructive technique
and gives additional information to “standard” methods. It should not
be used as a single technique for <strong>ink</strong> examination alone, which may lead to
false conclusions. Figure 15 shows two very similar spectra <strong>of</strong> a blue and
a black ball point <strong>ink</strong>.
Blue and Black Ball
Point Pen Ink.
2000
1958
1916
1874
1832
1790
1748
1706
1664
1622
1580
1538
1496
1454
1412
1370
1328
1286
1244
1202
1160
1118
1076
1034
992
950
908
866
824
782
740
698
656
614
572
530
Fig. 15. Raman spectra <strong>of</strong> blue (blue) and black (black) ball point pen <strong>ink</strong>s, FORAM
685, SERRS 1.
It has been shown that different sample preparation techniques and excitation
wavelengths lead to different results, which may all contain valuable
information. Therefore we currently cannot recommend a certain sample
preparation technique or excitation wavelength but to collect and evaluate
all the information taken from different techniques. Further research work
is needed on Raman <strong>spectroscopy</strong> <strong>of</strong> pure <strong>ink</strong> dyes to see in which way their
signals appear in <strong>ink</strong> spectra. Libraries <strong>of</strong> dyes and <strong>ink</strong>s should be helpful in
identification <strong>of</strong> certain <strong>ink</strong> formulations.
Acknowledgements:
The author wishes to express his gratitude to Foster & Freeman and Dilor GmbH for
giving us their spectrometers for this work. I would like to thank Emma Wagner and
Norbert Jaufmann for examining all the <strong>ink</strong>s with the FORAM 685 and for the evaluation
<strong>of</strong> resulted spectra. Carolin Schumann, Dr. Ursula Hendriks and Dr. Jochen
Geyer-Lippmann from the Polizeitechnische Untersuchungsstelle <strong>of</strong> the Landeskriminalamt
Berlin did the same work with the LabRam Infinity.
References:
1. W h i t e P. C., SERRS <strong>spectroscopy</strong> – a new technique for forensic science?, Science
& Justice 2000, vol. 40, pp. 113–119.