Detecting Doctored Documents
Written by Lewis Mitchell   

How Raman spectroscopy can help identify forged documents

This article appeared in the May-June 2021 issue of Evidence Technology Magazine.
You can view that full issue here.


IN THIS EVER-PROGRESSING AGE of technological advancement, more and more we witness the old remnants of technologies recede into history. It now appears, in the advent of the modern era, that print is endangered. Personal computers with access to the internet now permeate the daily life of people across the world, offering easy access to literature that only decades before would have required printing. Your very reading of this article demonstrates that point.

Understandably, advancements in the way we access and handle information has led many people to even question the future validity of the ultimate written medium, old-fashioned pen and paper. Where print media is stubbornly falling to alternatives, handwriting persists as the most convenient and accessible way of conveying information. We can safely say pen and paper will be sticking around for a few years yet.

This does present some interesting challenges, notably when it comes to the validation of written works. A prime example presents itself within the realm of legal documentation—contracts and agreements. Despite the growth of e-commerce, the norm for signing legal documents remains a handwritten signature on paper. Moreover, authenticated documents are required to be signed before a notary public, all seals must be original, and all certifications must match.

How do we go about validating these documents, and how can we ensure that a signature is original? Unlike electronic documents, there are no methods of tracking contributors or alterations. What if an entrepreneurial rogue changes an initial on a mortgage agreement, or adds an additional zero on the check they received? For these reasons, it is vitally important that we can determine the authenticity of writing. The answer to this challenge comes in one elegant form: spectroscopic Raman analysis.

Crossing Ink Analysis
A technique is required to determine ink composition to establish if a document contains changes made with different pens. There are many different types of inks. Colors may be the same, but chemically they can be different. Raman analysis allows for rapid, non-destructive testing of questioned areas with the specificity to distinguish similar ink types that may visually look identical.

Recently, researchers have been taking this investigation further, exploring how to determine the crossing order of two ink lines originating from different pens. Traditionally this has been a difficult challenge to approach; the inconsistent background of paper and the requirement to interpret chemical images introduce large degrees of uncertainty. How then, can Raman spectroscopy reliably determine the origin of additional ink lines on paper?

This is possible using a new method to measure the ink order through the implementation of several techniques. The premise lies in the analysis of each ink's coverage within the region where they cross. The crossing region is a mixture of both ink components, logically, where the uppermost ink appears in the largest quantity when analyzing using a laser impinging on the top surface. This is revealed using false-color images and concentration-estimation techniques.

The process begins by capturing a white-light image of the crossing region and surrounding area. To capture the composition of the crossing region, the area is scanned using chemical imaging. This collects spectral information from the inks with a continually moving line-focused laser. The line focus has a lower-power density than a traditional spot focus, which allows for a greater laser power without damaging the ink layer.

Once the data is gathered, distinct regions of the image are masked and pure references for each ink are obtained. The masking tool is used to limit the data to be processed as defined by the thresholding of an image (e.g. white light or Raman) or by manual selection of an area of the scan.

Figure 1. White-light image of ink crossing. The region between the two lines defines the crossing. The uncrossed ink sections were masked, and pure references were obtained for component analysis.

Now that we have a reference for each ink, component analysis can be performed to obtain false-color images that display the distribution of each component on top of the white-light image. Previously at this stage, the user attempting to identify the crossing order would threshold the false-color image and establish the crossing order. As darker areas indicate less similarity to the reference, the darker ink image in the crossing region was assumed to be on the bottom layer; however, there was no guarantee two independent users would threshold the image the same way and arrive at the same conclusion. Depending on the image, the ink order can potentially be interpreted either way. So, how can we alleviate this previously encountered pitfall? The answer, of course, lies in another exclusive investigative tool... introducing concentration estimates.

Figure 2. False-color Raman images of the two different pen strokes. These images demonstrate the chemical specificity of Raman spectroscopy, ink species are identified, and relative concentrations of each ink can be determined.

The concentration estimates tool determines the total percentage contribution from each ink based on the component-analysis generated image. The process doesn’t account for any changes to the image thresholds, therefore ensuring highly consistent results from user to user. The larger concentration estimate value corresponds to the ink that is more prevalent in the crossing area being analyzed. The larger the concentration estimation difference between the two references, the more confidence we can have in the deposition order.

Similarly, the concentration estimate of the pure regions indicates the high specificity between the references. In this way, we have a statistically meaningful value to provide confidence in our determination of the ink deposition order.

Figure 3. This table displays the concentration estimate value of each ink at the pure ink and crossing regions. The pure regions confirm the chemical specificity of the technique. The large difference in concentration at the crossing gives confidence in teh proposed deposition order.

And there you have it: You now know how disputes of forgery or altercation are spectrographically analyzed and resolved.

About the Author
Lewis Mitchell is an applications scientist in the Raman spectroscopy department at Renishaw. There, he works on cutting-edge applications including the development of new hardware and software with recent work involving auto-focus tracking, univariate stage movement, and averaged imaging techniques. He graduated Heriot Watt university with a Master of Science in chemical physics and is now taking his first steps into the world of Raman spectroscopy.

< Prev   Next >

Interview with an Expert

One of the more specialized areas of crime-scene investigation has to do with searching for evidence of arson. To get some background in this area, we spoke with an individual who has had more than 46 years in fire service, 24 of which have focused specifically on fire/arson investigation.