In Situ Analysis of Ink Lines Made by Blue and Black Ballpoint Pens by Reflectance and Luminescence Spectroscopy Using the VSC6000HS

1.1. Ballpoint Pens: Ink Composition

The transfer from a single dye dissolved in an oilbased solvent to a mixture of dyes dissolved in a glycol-based solvent was the most important makeover in ballpoint ink formulation. These changes offered document examiners the chance of greater success in differentiating inks. The historical developments in ballpoint ink composition are explained in detail in the literature23, 24, 30.

Currently, ballpoint ink is typically comprised of a mixture of synthetic dyes dissolved in a vehicle that includes solvents, resins and oils23, 31, 32. Additional materials are added into the vehicle such as ball lubricants, elasticity modifiers and corrosion inhibitors to optimize the properties of the ink32. Table 1.1 shows the main components of ballpoint inks.

Ballpoint ink, overall, has distinguishable characteristics; it is typically paste, more viscous, thicker and contains more organic dyes than for example roller ball and gel inks31.

Ink composition differs from one manufacturer to another and between inks made by the same manufacturer. This forms the basis of ink analysis. The chemical composition of ink plays a role in terms of the response obtained when the ink is exposed to illumination. However, it is not well understood which of the ink’s component has the greater significance.

1.2. Composition of paper

Even though papers are generally composed of similar materials (e.g. matted fibers obtained from cotton, linen, sisal and pulp and other chemicals added to optimize properties such as fluorescence, gloss and smoothness)33, the composition and the concentration of these components differ.

Type of papers may interfere with the luminescence properties of inks24, 33, 34. Therefore, interpretation of infrared luminescence of entries on different substrates should be carried out with extreme caution. This interference is much diminished when comparing two ink entries made on the same substrate3. However, it is not eliminated since surface inhomogeneities exist3, 24.

1.3. Infrared Reflectance/Absorption

As illustrated in Figure 1.1, the object is illuminated by a floodlight that contains a high infrared component (400 – 1000 nm). The object transmits, absorbs, reflects and/or absorbs then re-emits the light (explained more in section 1.4.). Different inks theoretically respond differently; some reflect more light (are brighter) or at a different wavelength (are a different color). This response can be seen by the human eye if the reflected light falls within the visible spectrum.

Figure 1.1.

However, the reflected light can extend into the infrared (IR) region since the illumination source has an infrared component. To visualize this reflected infrared light, a filter is used to block the transmission of visible light and only allow IR light to pass through to be recorded by an IR sensitive camera. This technique can be used to differentiate between inks that are indistinguishable to the naked eye6.

1.4. Infrared luminescence

Infrared luminescence has been an effective, rapid, direct, simple and non-destructive technique for differentiating inks that appear similar to the unaided eye35 since at least 19638. This phenomenon occurs when molecules of an object are irradiated by a source of light.3. They absorb energy from that source and re-emit it as light at longer wavelengths 3, 35.

As shown in Figure 1.2, to observe and record the property of infrared luminescence the object has to be radiated by an excitation beam. The emitted radiation produced from the interaction between the object and the excitation beam goes through a filter that only allows light in the infrared or far-red range to pass through to be recorded by an infrared-sensitive camera7. The differentiation is based on the different responses that inks show when they absorb and re-emit light. Some inks remit light strongly or at different wavelengths35 while some ink dyes do not have the property of infrared luminescence7. However, differentiating inks by visual comparison to determine the degree of luminescence poses several difficulties including the subjectivity of this analysis and impossibility of knowing the exact wavelength visually. For this reason, ink’s luminescence responses in this study will be based on their luminescence spectra obtained using the VSC6000 spectrometer rather than merely visually.

Figure 1.2.

In addition, infrared luminescence cannot be considered in isolation of other methods to differentiate among inks8. A single ink contains components that are luminescent and components that are not. If an ink is not luminescing it may mean that its non-luminescent components are masking the luminescent components. Introducing a substance such as water, body lotion or milk could cause the non-luminescent components to diffuse allowing the luminescent components to become visible8.

1.5. VIS-IR Reflectance and VIS-IR Luminescence Spectroscopy

The subjectivity of the visual assessment of infrared reflectance and luminescence behavior has driven the need for methods that allow a more objective way of drawing conclusions. Spectroscopy, defined as “a technique for resolving electromagnetic radiation into its component wavelengths and measuring the radiation as it interacts with matter”24, allowed document examiners to base their findings on spectral analysis rather than on visual observations.

The VSC6000HS has an automated spectrometer that uses a video camera and bandpass filter to analyze the light originating from the object. Spectral analysis is then performed over the wavelength band 400 – 1000 nanometer38. The results are displayed as a spectral graph and chromaticity values.

The spectrometer facility of the VSC has been used in several studies6, 15, 24, 34, 39. Spectra obtained from a VSC2000 demonstrated repeatability and successful discriminations between different inks were made34. In another study that included roller-ball and ballpoint inks, the VSC2000 discriminated 98.24% of the black non-liquid inks and 90.90% of the blue non-liquid inks24.

Similar to most spectroscopic analysis, several factors affect the results of reflectance and luminescence spectroscopy. The age of the IR light source, the writing substrate and the variation between the “standard white tile” may influence the spectra34. In addition, contamination and weathering may cause non-luminescent inks to exhibit infrared luminescence and consequently interfere with the spectra3. These factors should be noted prior to performing any ink differentiation method, particularly surface analyses.

1.6. Aims and Objectives

The main objective of this project is to determine the discrimination power of the VSC6000HS in differentiating among ink lines made by blue and black ball point pens.

A study performed on the previous version of the VSC (VSC5000) have concluded that reflectance spectroscopy has no value in differentiating ink lines made by different pens.15

Foster + Freemen Limited implemented an integral grating-type spectrometer with the new version VSC6000HS36. Discrimination of inks, based on reflectance and luminescence spectroscopy, utilizing the new version has not been researched yet.

The project aims to:

• Analyze blue and black ballpoint ink lines by non-destructive objective methods utilizing the latest version of the Video Spectral Comparator VSC6000HS

• Determine the discriminating power of VSC6000HS reflectance and luminescence spectroscopy

• Determine if the VSC6000HS spectrometer can differentiate between ink lines that appear similar by conventional methods (e.g. the visual assessment of reflectance and luminescence behavior)

• Determine the effect differences in paper has on the results of the spectra obtained by the VSC6000HS

2.1. Sample Collection

30 blue and 30 black ballpoint pens were collected from stationery stores around Wellington, New Zealand for the purpose of this project. Table 2.1 and 2.2 show the pen code given to each pen and description including brand name.

Paper substrate study: A variety of papers, as shown in Table 2.3, were provided by New Zealand Police Document Examination Section. The purpose of this study was to determine the effect the paper substrate had on the reflectance and luminescence spectra obtained from VSC6000HS.

2.2. Sample Preparation

An ink line from each pen used in the study was drawn on a piece of white A4 80 GSM Staples photocopy paper. In order to reduce the effect of paper on the results obtained, all ink lines of the same color were drawn on the same sheet of paper that was divided into 30 sections by using a pencil and ruler. The blue ink lines were on one sheet of paper and the black ink lines were on another.

Paper substrate study—On each sheet of paper used in this study, an ink line was drawn from a blue ballpoint pen (bb3—Papermate 4 Ball, Business RT BP 1mm) and a black ball point pen (bk18—Impact retractable ASST 0.7mm).

2.3. Instrumentation

The Video Spectral Comparator VSC6000HS, Foster + Freeman Ltd., was used to study the luminescence behaviors and to obtain reflectance and luminescence spectra of all samples. The experiment was conducted in a dark room where the VSC was the only light source available.

3.1. Reproducibility Study of Reflectance and Luminescence Spectroscopy

A single ink line is not totally homogenous because the ink’s components are not evenly distributed across the ink line. This is mainly due to the skipping characteristic associated with ballpoint implements. Thus, it was essential to record the spectra from different areas of the ink line in order to acquire the intra-sample variations. 3 replicate spectra of 5 different areas within the same ink line were taken. This allowed the variation within the same ink line to be obtained.

Figure 3.1 shows 15 reflectance spectra of (bb9) and Figure 3.2 shows 15 luminescence spectra of (bb11) taken from 5 different areas – each area was scanned three times. These spectra were taken on two different days – three areas in one day and two areas on another.

Figure 3.1

Figure 3.2

Analysis of these spectra demonstrated spectral reproducibility with minor variations. All spectra of the same ink line showed consistent pattern shape, number of peaks and wavelengths of peak maxima. Slight differences in intensities represent the variations within the ink line. Thus, signal intensity was not used as a central differentiation factor when comparing different inks.

3.2. Ink Differentiation Analysis

3.2.1. Ink lines of blue ballpoint pens

The Visual Assessment of the Luminescence Behavior

Based on their infrared luminescence behaviors exhibited across the full range of the excitation wavelengths, blue ink lines were divided into 4 groups as shown in Table 3.1. Within each group the individual pens were undistinguishable from each other by their infrared luminescence behavior.

Table 3.1

Groups of blue ballpoint pens based on the visual assessment of luminescence behavior.

Group	Pen Code (bb)	Luminescence Behavior as the excitation wavelengths increase
1	4, 6, 9, 12, 13, 17, 18, 19, 20, 28, 29	No Luminescence across the full range of the excitation wavelengths
2	1, 3, 5, 8, 15, 21, 22, 24, 27, 30	Strong Luminescence across the full range of the excitation wavelengths
3	23, 25, 26	Light luminescence across the full range of the excitation wavelengths
4	2, 7, 10, 11, 14, 16	Luminescence gradually increased

The 30 ink lines of blue ballpoint pens exhibited different responses as the excitation wavelengths increased. Group 1 ink lines had no luminescence across the full range of the excitation wavelengths. This was an initial indication that it would not be beneficial obtaining luminescence spectra for this group.

On the other hand, the inks in Groups 2, 3 and 4 all exhibited infrared luminescence. A uniform strong luminescence was recorded for Group 2. Group 3 ink lines exhibited light luminescence that was uniform across the full range of the excitation wavelengths.

Group 4 ink lines had a proportional relationship with the excitation wavelengths. They gradually increased their luminescence with the increase of the excitation wavelengths.

The comparison of inks’ luminescence behaviour is widely practiced and considered as a valid method of differentiating among inks7, 8. However, it should be approached with caution since a degree of subjectivity exists in this examination, and the luminescence response could be influenced by other factors3. Therefore, the above classification was not conclusive. The purpose of this visual classification was a preparation step for the subsequent spectral comparison of the ink lines. The results of this initial examination explained the overall infrared luminescence behaviour of each ink line and established the appropriate excitation wavelengths and camera filters for recording luminescence spectral profiles.

Spectral Comparison

The 30 blue ballpoint ink lines produce 435 possible pairs of compared samples. Based on their reflectance spectra alone, all except 52 pairs could be differentiated as shown in Figure 3.3. The luminescence spectra differentiated all except 100 pairs. Thus, with this sample of 30 pens, the discriminating power for the blue ballpoint pens by reflectance and luminescence spectroscopy was approximately 88% and 77% respectively.

Figure 3.3

By combining the results of the two methods it was possible to differentiate another 18 pairs of inks, increasing the discriminating power to 92%. Table 3.2 demonstrates the final groupings of the blue ballpoint pens after applying both methods.

Table 3.2

Groups of blue ballpoint pens based on VSC6000HS reflectance and luminescence spectroscopy.

Group	Pen Code (bb)
1	1, 3, 8, 15, 22
2	4, 9, 17
3	6, 7, 10
4	2, 14, 16
5	11
6	23
7	25
8	26
9	5, 21
10	24, 27, 30
11	28, 29
12	12, 13, 18, 19, 20

Figure 3.4 is a clear example of pairs that were undistinguishable by reflectance spectroscopy. It shows the spectra of bb1, 3, 8, and 15. These ink lines that were made by different pens demonstrated a similar spectral shape, number of peaks and wavelengths of peak maxima. Slight differences in signal intensities were expected within the same ink line as illustrated in the reproducibility study. Thus, they should not be weighed as significant differences that differentiate ink lines. These spectra were undistinguishable from each other.

Figure 3.4

On the other hand, bb2, 5 and 9 demonstrated different responses in reflected light, as can be seen from their spectra in Figure 3.5. Bb9’s spectrum had 3 peaks whereas bb2 and 5 had only two. In addition, they are significantly different in reflected light within the infrared region of the spectrum. Therefore, these ink lines were distinguishable from each other. This is an example showing pairs that were distinguishable by reflectance spectroscopy.

Figure 3.5

These ink lines also had different infrared luminescence behaviors and so could be distinguished by infrared luminescence. Reflectance spectroscopy, however, provided a more objective, and still relatively simple way of clearly demonstrating the differences between these ink lines.

Reflectance spectroscopy also succeeded in discriminating some ink lines that had no luminescence and accordingly could not be discriminated by infrared luminescence. For example, bb6 and 9 were not differentiated by infrared luminescence, but were discriminated by reflectance spectra as shown in Figure 3.6. Therefore, the VSC6000HS reflectance spectroscopy provided an additional powerful tool in discriminating inks when the conventional method – infrared luminescence – was useless.

Figure 3.6

Figure 3.7 shows luminescence spectra of two ink lines made by 2 different pens (bb24 and 27). They exhibited similar spectral pattern with one peak at the same wavelength. This is an example demonstrating a lack of differentiation of ink lines using luminescence spectroscopy. This resemblance does not confirm that the two inks have come from the same source, just that they cannot be differentiated at this level of analysis.

Figure 3.7

Figure 3.8 shows luminescence spectra of bb14, 21 and 26. These ink lines all visually showed a strong luminescence across the full range of the excitation wavelengths. Accordingly they were undistinguishable by infrared luminescence. Luminescence spectroscopy, however, discriminated these ink lines. Differences in the number of peaks, spectral shape and wavelengths of peak maxima were significant differences that did not only differentiate these ink lines, but also provided an objective way of measuring luminescence and easily illustrating these differences.

Figure 3.8

3.2.2. Ink lines of black ballpoint pens

The Visual Assessment of Luminescence Behavior

Similar to the blue ink lines’ results, black ink lines showed variable infrared luminescence behaviors as the excitation wavelengths increased. Table 3.3 shows the different groupings of the black ink lines based their luminescence behavior. Within each of Groups 1, 2, 3, and 4 the ink lines produced consistent responses with the increase of the spot light filter. In Group 5, however, the ink lines exhibited irregular luminescence behaviors. For example, some of these ink lines produced light luminescence with short excitation wavelengths. Then, they suddenly luminesced strongly when excited with mid-range wavelengths. However, they started to lose luminescence with longer excitation wavelengths. Therefore, their relationship with the excitation wavelengths was not fully understood.

Table 3.3

Groups of black ballpoint pens based on luminescence Behavior.

Group	Pen Code (kb)	Luminescence Behavior as the excitation wavelengths increase
1	2, 3, 6, 7, 10, 17, 21, 29, 30	No Luminescence across the full range of the excitation wavelengths
2	11, 14, 18, 20	Strong Luminescence across the full range of the excitation wavelengths
3	26	Light luminescence across the full range of the excitation wavelengths
4	13, 19, 22, 23, 24, 28	Luminescence gradually increases
5	1, 4, 5, 8, 9, 12, 15, 16, 25, 27	Irregular luminescence behavior as the wavelength increases

Excitation wavelengths within the mid-range of the VSC6000HS settings produced the most variant infrared luminescence of ink lines. There fore, these groupings have established that these excitation wavelengths were suitable to be used for obtaining luminescence spectra.

Spectral Comparison

The 30 black ballpoint ink lines produce 435 possible pairs of compared samples. Based on their reflectance spectra alone, all except 54 pairs could be differentiated as shown in Figure 3.9. The luminescence spectra differentiated all except 72 pairs. Thus, with this sample of 30 pens, the discriminating power for the black ballpoint pens by reflectance and luminescence spectroscopy was approximately 88% and 83% respectively.

Figure 3.9

By combining the results of the two methods it was possible to differentiate another 29 pairs of inks, increasing the discriminating power to approximately 94%. Table 3.4 demonstrates the final groupings of black ballpoint pens after applying both methods.

Table 3.4

Groups of black ballpoint pens based on VSC6000HS reflectance and luminescence spectrometry.

Group	Pen Code (kb)
1	3, 21, 29, 30
2	2
3	13, 25
4	1, 5, 9, 12
5	8, 15
6	19
7	22
8	6, 7, 10, 17
9	11
10	4, 14
11	20
12	18, 27
13	23, 24, 26
14	28
15	16

Figure 3.10 shows the reflectance spectra of kb21, 29 and 30. These spectra have similar pattern shapes indicating that they demonstrated similar responses in reflecting light. This is a clear example of pairs that are undistinguishable by VIS-IR reflectance spectroscopy. The differences in signal intensities are found even within the same ink line. Thus, they are not counted as differentiating differences.

Figure 3.10

On the other hand, the ink lines of kb2, 3 and 6 exhibited different spectral pattern shape, as can be seen from their spectra in Figure 3.11. Therefore, they were easily discriminated. This is an example of distinguishable pairs.

Figure 3.11

It is noteworthy to mention that these ink lines did not produce infrared luminescence and so were undistinguishable from each other using that technique.

Figure 3.12 shows the spectra of three different ink lines. This is an example showing a lack of differentiation of ink lines by luminescence spectroscopy on the VSC6000HS. These ink lines have a similar spectral pattern, wavelength of peak maxima and number of peaks. This similarity means that these ink lines produced similar luminescence at that excitation wavelength, though they are known to be from different pens. For this reason, among others, this methodology cannot be used for identification.

Figure 3.12

The luminescence spectra of kb11 and 18 are shown in Figure 3.13. These ink lines produced strong luminescence and were undistinguishable from each other using infrared luminescence. However, luminescence spectroscopy succeeded in discriminating them. Furthermore, it offered a more objective and reliable way of demonstrating the differences. This is a clear example distinguishable pairs.

Figure 3.13

3.5. Paper substrate study

The paper substrate study was conducted to determine the effect the paper substrate had on the reflectance and luminescence spectra obtained from the VSC6000HS.

3.5.1. Paper substrate effect on VIS-IR reflectance spectra

Blue ink line

No significant differences were found between the reflectance spectra of the blue ink lines (using bb3) that were drawn on each of the different white paper substrates used in this study. Agreements in overall spectral pattern shape, wavelengths of peak maxima and number of peaks made these spectra undistinguishable from each other. The slight differences of signal intensities were expected to be found even within the same ink line on a single substrate.

Accordingly, the white paper substrates did not affect the reflectance spectroscopy’s results. The ink lines from the same pen (bb3) on the different white paper substrates used in this study demonstrated similar reflectance spectra.

Figure 3.14 is an example showing the similarity of the blue ink spectra on two different white papers. Based on these results, comparing blue ink entries on different white paper substrates using reflectance spectroscopy is satisfactory as a method of differentiation. The method, similar to other optical methods, should not be used as an identification technique since a number of variables prevent achieving that.

Figure 3.14

The spectra of the blue ink lines on the colored papers had variable results. The pink and blue papers had similar results to those obtained from the white paper substrates. However, the yellow and green papers had different responses within the visible region of the spectrum.

Figure 3.15 shows reflectance spectra of the blue ink on a white paper substrate and a blue paper substrate. The ink lines on these two different colored substrates behaved similarly in terms of reflecting light. Both spectra have similar overall spectral pattern shape, number of peaks and wavelengths of peak maxima. Therefore, they would not be differentiated from each other.

Figure 3.15

The yellow and green paper substrates on the other hand, showed different spectral appearances in comparison to the spectra taken from the other substrates (e.g. white, blue and pink), as demonstrated in Figure 3.16. Even though they all have a similar response in reflecting light in the infrared region, the green and yellow substrates showed spectral differences within the blue region of the visible spectrum. The blue ink lines drawn from the same pen on these two colored substrates seemed to absorb more light rather than reflect in that region, resulting in diminishing the first peak. This could lead to an incorrect differentiation.

Figure 3.16

Black ink line

Spectra of the black ink lines (using kb18) drawn on all of the white paper substrates demonstrated similar reflectance spectra. Figure 3.17 shows the reflectance spectra of the black ink on 3 different white paper substrates. These spectra from different papers are undistinguishable from each other due to the similarity in comparison criteria e.g. pattern shape and wavelength of peak maxima. Accordingly, the white paper substrates did not affect the reflectance spectroscopy’s results.

Figure 3.17

The black ink lines drawn on all of the colored papers, apart from the yellow paper, showed spectra similar to the white papers’ results. They were undistinguishable from each other. The yellow paper substrate, however, had a different spectral appearance in the visible region as shown in Figure 3.18. The ink line absorbed more light, resulting in diminishing the small peaks as can be seen from the white substrate’s spectra. Even though the ink line came from the same pen, the reflectance spectra had differences due to the paper substrate used.

Figure 3.18

3.5.2. Paper substrate effect on VIS-IR luminescence spectra

Blue ink line

The blue ink lines on all the white substrates except “EXP Green – the 50% recycled paper” exhibited similar spectra. As can be seen from Figure 3.19, these spectra are undistinguishable from each other.

Figure 3.19

The luminescence spectra of the ink line on “EXP Green” paper had peak maxima at different wavelengths compared to the rest of the white paper substrates as shown in Figure 3.20. This paper was the only paper that is 50% recycled. When examined using visual infrared luminescence, strongly luminescing inclusions were noted in the paper, possibly from the mixture of recycled materials.

Figure 3.20

Even though the ink was the same, the luminescence spectra obtained were different. An incorrect differentiation could be made based on these differences. The luminescence spectra of the blue ink lines drawn on the colored papers exhibited differences in wavelength of peak maxima and signal intensity. Figure 3.21 shows an example of the luminescence spectra of the blue ink on a yellow paper substrate in comparison to a white paper substrate. These spectra have significant differences that make them distinguishable from each other. Therefore, the color of the paper influenced the luminescence spectra. Comparing ink entries on different colored substrates using luminescence spectroscopy was found not to be useful.

Figure 3.21

Black ink line

No significant differences were found between the luminescence spectra of the black ink on all of the white paper substrates as shown in Figure 3.22.

Figure 3.22

The luminescence spectra of the black ink lines drawn on all the colored papers also exhibited differences. As shown in Figure 3.23, the luminescence spectra of the black ink on a yellow paper substrate has significant differences compared to the spectra from a white paper substrate. These spectra are distinguishable from each other. Thus, it is not possible to compare an ink entry on white paper with an entry on a colored paper using this method.

Figure 3.23

Table 3.7 summarizes the results of the paper substrate study whose aim was to determine if the paper type would have an influence on the reflectance and luminescence spectra obtained by the VSC6000HS spectrometer facility.

Table 3.7

A summary of the paper substrate study’s results. Did the paper type affect the spectra obtained in comparison to a standard white “Staples” A4 paper?

Paper type	Reflectance spectra		Luminescence spectra
Blue ink line	Blue ink line	Blue ink line	Blue ink line
EXP (50% recycled)	No	No	Yes	No
Officemax	No	No	No	No
APRIL Fine Paper	No	No	No	No
REFLEX	No	No	No	No
Staples	No	No	No	No
Sapphire offset	No	No	No	No
Mondi	No	No	No	No
Mataura (blue)	No	No	Yes	Yes
Mataura (pink)	No	No	Yes	Yes
Mataura (green)	Yes	No	Yes	Yes
Mataura (yellow)	Yes	Yes	Yes	Yes

“Yes” indicates that the resultant spectra looked different than the spectra obtained from a standard white “Staples” A4 paper and that accordingly the paper substrate has affected the resultant spectra. “No” means the spectra were similar and accordingly the paper did not affect the spectra obtained.