Andreas Ruediger, a condensed matter physicist, sat with two new colleagues from a nearby university in the mid-2010s. Ruediger, a Ph.D., is a professor and researcher with the Institut national de la recherche scientifique at the Université du Québec in Montreal, Canada.
The group discussed funding opportunities.
The scientists invited Ruediger to attend an event at Université du Québec à Trois Rivières (UQTR) to show him the work they were doing. A few weeks later, Ruediger was at an affair organized by the forensic section of the chemistry, biochemistry and physics department. It has a connection to École nationale de police du Québec, the police academy of Quebec.
Standing there over a glass of bubbles, the scientists continued to talk about collaborations. Frank Crispino, a former French Gendarmerie officer turned professor, mentioned a problem that plagued law enforcement.
“It's about the recovery of serial numbers in polymers,” Crispino said. “There are more and more firearms made of polymers.”
Law enforcement had reliable techniques for retrieving serial numbers ground off metal guns. But investigators still had a hard time recovering erased numbers from plastic guns.
And plastic guns were becoming popular among the criminal set. Some of these handguns had much larger ammunition capacity, were more durable, lightweight, and compact. And because it’s made of polymers, it was cheaper to manufacture than metal alloy guns.
Law enforcement members, recreational users and criminals adopted the handguns in masses in the 1980s, replacing clumsy six-bullet revolvers with these highly lethal, concealable semi-automatic pistols able to fire many rounds of ammunition without reloading.
Police departments use a standard protocol to reconstruct abraded serial numbers from metal guns, with either electrochemical action or magnetization.
“If you have a distorted metal, you have heaps of dislocations, and these dislocations help you to electrochemically etch it,” Ruediger said.
But that doesn’t work for polymers. Electrochemical etching just eats away the plastic.
That was a moment when Ruediger put things together.
Serial numbers are imprinted, or embossed, in most cases by mechanical force. To conceal the serial number, a bad guy might scratch or grind it off until it can't be read, making the gun untraceable.
That introduces an opportunity, Ruediger thought, after the event.
“You typically don't go that deep (with the grinding), because if it's a firearm, this thing might go off in your hand,” he said. “So you just go as deep as you have to. But when you emboss, you're actually introducing residual information that goes deeper than the profile, because you're compressing the material underneath. And that is information that remains.”
The question was how to access that residual information in polymers.
“My idea was, if there is some stress in there, then the materials should be under strain. And that is something that we should be able to see with Raman spectroscopy,” he said.
Ruediger designed an experiment. He had a joint master’s student with a colleague from UQTR take a cable tie and fold it in the middle to create a point with a strong deformation. Using a Raman spectrometer, he examined one pixel in the under-formed area. A multitude of peaks appeared in a Raman spectrum.
Then the question for Ruediger became, which peak responds to the deformation. Since it’s a polymer, some of these peaks probably wouldn’t change because it doesn’t become heavily distorted. If something is distorted, it will probably produce a response, he theorized.
And so first thing was to chase down the number of peaks that actually does change. That took a lot of time fitting and testing. Eventually, they found something quite significant.
They saw that the peaks changed position by up to one wave number, the units used to measure energies in Raman spectra. A Raman spectrum typically produces a range of 1000 wave numbers. The peak position changed by the tiny amount of one wave number or less.
The next thing was to figure out if they could translate that into an imaging procedure.
“What we did is was pretty low tech,” he said. “We had a colleague manually emboss the letter H, two vertical, one horizontal bars into different kinds of polymers ―polycarbonates, nylon Nylotron, which is fiber reinforced nylon, and polyethylene.”
After making multiple screwdriver etches, they ground out the right side of the letter. The profile was roughly 80 micron deep, and they obliterated it with a grinder into different depths. The deepest, at 200 microns was more than twice as deep as the initial profile. Then they scanned the complete area off the letter H.
The standard pixels appear in red. They found a deep trench in the center, with the pixels changing to blue after the analysis of peak shifts in Raman spectroscopy. The picture above is the superimposed image of the pixels, showing the Raman information extracted and analyzed.
“On the right side of the region which is obliterated, it's showing the shift of the peak as a function of position,” Ruediger said. “So there is actually a signature. Your material has been compressed and it's still there.”
The width of the peaks translates directly into the lifetime of that vibration. By destroying the order of the system or surface, which in this case was the original etching, the Raman signal lifetime goes down and as a result, the peak broadens.
“When you buy a new coffee mug, and the lady at the desk is doing her job, she will basically knock at the coffee mug,” Ruediger said. “And if she gets a nice ‘pling’ from it, then she knows the cup is intact, because the phonon lifetime is very, very long. If the cup is broken, it gives you just a pluck, and that means the phonon lifetime is really short. This is the acoustic analogy to the optical phonons that translates directly into the width of the Raman peak.”
The team used this concept to measure declamations, materials and the degree of order in the ablated polymer material.
“If the degree of order is high, then you get these beautiful long living phonons, and if the degree of order is low, it will basically fade out quickly.”
A law enforcement agent might ask how significant the difference is between something you see and something you don't see. This is actually a contrast, and the researchers ran a histogram of the Raman peaks distribution to illustrate the differences.
“It tells us how significant we can tell the two populations apart,” Ruediger said. “If these two peaks are very far apart, then we have a very strong image contrast, and we can be sure that this is not just a coincidence that we see them as two different colors.”
It’s typically what a lawyer or a judge would want to know, according to the physicist. The histogram allows researchers to express a degree of belief by which they can tell the areas apart. For instance, in this test case, it was about 200 times higher to observe this result if it was due to an erasing than if it’s coincidental.
Ruediger co-authored a 2017 paper on this Raman application. It was publishedin Analytical Chemistry, and titled Reconstruction of Obliterated Characters in Polycarbonate through Spectral Imaging.
According to the paper, the team showed the residual strain and local variations in the structural arrangement of the polycarbonate were nondestructively imaged through peak shifts and variations of the full width at half-maximum of specific Raman lines, respectively. A subsequent quantitative statistical analysis recovered the etching, ground out at depths substantially larger than typical obliteration depths.
The group published a second paper in the same periodical in 2019, called, Contrast Enhancement for theRecovery of Obliterated Serial Numbers in Different Polymers by Correlated Raman Imaging of Strain, Phonon Lifetime, and Strain-Induced Anisotropy. In it, researchers investigated samples of polyethylene by monitoring the vibrational modes which are most susceptible to peak shifts and changes in the full width at half-maximum (fwhm) and peak intensity ratios. They were able to correlate the signals to make them statistically more significant.
Three Raman signals, peak shift, peak broadening and intensity ratio are all independent. The 2019 research combined these measures to increase the accuracy of the results.
Ruediger’s team used a custom Raman Microscopy system from HORIBA’s Optical Spectroscopy Division to conduct the research. The turnkey system is based on HORIBA’s IHR320 imaging spectrometer and Synapse CCD scientific camera. The instrument, using a 320 mm focal length, is suitable for Raman, fluorescence, photoluminescence and absorbance/transmission/reflectance/ emission analysis.
“The system was customized according to our specs, but it was a complete turnkey solution delivered, which I think is kind of cool,” Ruediger said. “Physicists always have their funny own ideas on how they would like to customize things.”
Although the serial number recovery technique has yet to be adopted in a real-world setting as of 2020, Ruediger said law enforcement agencies are interested in it. A Royal Mounted Police officer on his team works with fingerprints. But in law enforcement, the adoption of a new technique is a slow process.
While the research has been replicated, Canadian law prohibits Ruediger from working directly with firearms without permits for himself and all his students.
“We had someone from the forensic lab of the Montreal jurisdiction come with a little suitcase chained to his wrist containing a piece of gun. So we've been testing this on real firearms,” he said.
And yes, it works. That's the good news. But real firearms typically aren’t made in white or transparent colors. They usually come in pitch-black finishes. When you focus a laser on something which is pitch-black, the object heats and burns. Ruediger had to tune down the power of the laser to avoid destroying the sample. That slows the process, but doesn’t affect its accuracy.
The contrasts are still there. Residual mechanical strain and local structural changes can be detected with Raman spectroscopy, a non-destructive technique. The researcher had successfully recovered stamped letters (120 micrometers deep) that were obliterated by milling and said the estimated maximum depth of recovery is approximately 750 to 800 micrometers.
The results of the studies were presented at the 19th Interpol International Forensic Science Managers Symposium, covering the evolution of the state of the art between 2016 and 2019
Ruediger sees the homeland security and law enforcement industries in the U.S. as prime candidates for this technique.
“It is always fun to see an idea transforming into team work and into something meaningful. And we have a few more ideas up our sleeves.”
Do you have any questions or requests? Use this form to contact our specialists.