Richard Crocombe and colleagues have caused something of a stir in the spectroscopy community with their surprising solution to Raman’s well-known problem of evaluating fluorescence interference – LEGO blocks.
They found, in the first of a trilogy of papers, that Legos make unexpectedly effective “standard” samples for evaluating fluorescence avoidance and mitigation in Raman instruments. They then compared the same set of LEGO blocks across ten handheld Raman instruments – the most susceptible to fluorescence interference – finding that mitigation strategies had a dramatic impact on spectral quality. And most recently, the researchers once again reached into their toybox to evaluate a promising low-light detection technique: photon avalanche photodiode (SPAD).
Although their results highlighted the limitations with the current generation of SPAD arrays, Crocombe (Crocombe Spectroscopic Consulting, USA), Pauline Leary (Noble, Inc., USA), and Brooke Kammrath (University of New Haven, and Henry C. Lee Institute of Forensic Science, USA) believe that the technology has the potential to make a significant impact in forensic, environmental, and planetary exploration fields.
Here, the three co-authors tell us more about their work with LEGO, the future of SPAD-based Raman, and discuss whether we’ll ever fully eliminate the fluorescence interference problem.
What is the problem of fluorescence for portable spectrometers?
Fluorescence is the dominant challenge for Raman spectrometers. Raman scattering is inherently weak, with only 1 in a million (106) photons undergoing Raman scattering – while fluorescence is a much stronger effect, with quantum yields that can approach unity. That means even minor impurities or components in a sample can produce overwhelming fluorescence, masking the Raman signal completely.
This is especially problematic in portable instruments, which are typically used by non-scientists in field settings for rapid material identification. These users rely on built-in spectral libraries and automated search algorithms to deliver actionable results. But if fluorescence obscures the Raman signal, the instrument may fail to identify the material or return misleading results. That’s why strategies to avoid or mitigate fluorescence – whether through excitation wavelength selection, time-resolved detection, or advanced data processing – are critical to ensure reliable performance in real-world conditions.
In your first two papers, you showed that LEGO blocks make unexpectedly effective “standard” samples for evaluating fluorescence avoidance and mitigation in Raman instruments. When did it become clear that LEGO would work so well for this purpose?
In the hazardous materials, safety, security and law enforcement fields, vendors used samples such as sesame seed oil and whiskey as “difficult” samples to demonstrate fluorescence mitigation in portable Raman instruments – but those are complex, variable mixtures that degrade over time and can’t be standardized across labs. In the analytical space, colored polymer pellets were already known to be challenging. It was a small, creative step from polymer pellets to Legos as an idea for a set of “standard” samples.
The advantages of LEGO blocks were immediately clear: they’re low-cost, rugged, non-toxic, stable, and easy to transport. Most importantly, they’re manufactured using a consistent process, which means they offer reproducibility – a critical requirement for any “standard” sample. Their range of colors and tones (white, yellow, red, blue, gray, black) also mimics the types of materials encountered in the field, from straightforward to extremely difficult. That combination made them ideal for evaluating fluorescence avoidance and mitigation schemes, especially in handheld Raman instruments used by non-scientists in high-stakes environments.
Your second paper compared the same set of LEGO blocks across ten handheld Raman instruments. What did that comparative study reveal?
We are fortunate in having access to many portable spectrometers, ranging from Raman to GC-MS. For portable Raman instruments, there has been a trend towards longer wavelength excitation, from 785 nm, to 830 and 850 nm, and onto 1064 nm. Our comparative study revealed just how differently handheld Raman instruments perform when challenged with fluorescence. Using the same set of LEGO blocks across ten instruments, we saw that excitation wavelength and data processing strategies had a dramatic impact on spectral quality. We found that longer wavelength excitation yielded generally better results with these samples. Some instruments mitigation schemes, like sequentially shifted excitation, can be effective – but they can also generate spectra that look “odd” to a trained laboratory spectroscopist. Also, the details of these schemes are not always disclosed, and this could be a problem if a result is challenged in a court of law.
For field users – often non-scientists working in safety, security, or law enforcement – this matters. These instruments are used in high-stakes environments, often with limited operator control and time constraints. Our study helps clarify how different technologies respond to a variety of challenging samples, and why fluorescence mitigation isn’t just a technical detail – it’s central to reliable identification.
This new study focuses on pulsed-laser excitation and time-resolved detection with a SPAD array. What motivated you to make SPAD-based time gating the focus of the third paper?
Single photon avalanche photodiode (SPAD) arrays are excellent low light detectors, and therefore could be suitable for Raman spectroscopy; moreover, because they produce a time-sequence of data they can be used with a pulsed laser, and temporally discriminate against fluorescence. In addition, there are large commercial drivers behind the development of SPAD arrays – automotive LiDAR and night vision – and so we felt that this was a technology of the future. A commercial instrument using these technologies, with 532 nm excitation, is available (TimeGate, based in Finland), and we were lucky enough to have a professional colleague who had one on a short-term loan. So we jumped at the opportunity to run the Legos on that instrument.
Based on your results, what do you see as the broader potential of SPAD technology in Raman spectroscopy?
There is already commercial interest in SPAD-based Raman instruments, with a second company (Renishaw, UK) recently introducing one. They even used the spectra of LEGO blocks in their brochure! That’s a strong signal that this technology is gaining traction. SPAD-based time-resolved detection gives both laboratory spectroscopists and field operators of portables an additional tool for interrogating difficult samples, especially those with strong fluorescence.
That said, in our third paper we found that for the LEGO blocks, the pulsed laser-SPAD method didn’t outperform 1064 nm excitation. The main limitation was the low data acquisition duty cycle – the SPAd-based spectrometer was only collecting Raman signals for a tiny fraction of the total time. So while the technology is promising, current systems are still maturing.
We believe SPAD-based Raman will first make an impact in laboratory settings, where acquisition time and power can be optimized. As fast-pulse lasers and SPAD arrays continue to evolve – driven by high-volume markets like LiDAR and night vision – we expect portable systems to follow. For forensic applications, SPAD could be transformative: enabling fluorescence-tolerant Raman scans of colored or complex materials in the field. Industrial and environmental sectors would benefit from cleaner spectra and better reproducibility. Planetary exploration is currently dominated by deep-UV excitation, but SPAD-based systems could eventually offer a complementary approach – especially if fluorescence discrimination becomes critical.
What do you see as the main takeaways of your most recent paper for end-users who routinely confront fluorescence in real-world samples?
For non-scientist operators in the field, 1064 nm excitation – for instance in Rigaku instruments – is the current method of choice. For scientists operating in the field – for instance in cultural heritage – the Bruker BRAVO instrument, with 852/785 nm excitation, provides spectra that compare best to those obtained in the laboratory. And spatially-offset Raman spectroscopy (SORS), implemented on the Agilent Resolve instrument, has its niche as well in obtaining spectra through packaging.
Your results also highlight that no single excitation approach works for every sample, even within a standardized set like LEGO. How should this shape how we think about instrument design or selection for portable Raman?
We may never be able to completely remove the fluorescence problem in Raman spectroscopy, especially for real-world field samples. Our results reinforce that no single approach works for every sample – even within a standardized set like LEGO. That’s a critical insight for instrument design and selection. For example, 1064 nm excitation gave clean spectra for most blocks, but the blue block still showed multimodal fluorescence. Meanwhile, 785 nm excitation revealed resonance Raman bands from the blue pigment, but struggled with fluorescence in other blocks. This variability means that users should not expect one wavelength to solve every problem, and that future instrument design needs to be more adaptive.
Looking ahead, we see real promise in SPAD-based systems. Within the next five years, a portable instrument based on 785 nm excitation, with a 1 MHz repetition rate, and ~50 ps pulse width, combined with a SPAD array and manufactured in high volumes, is feasible and could offer the best of both worlds: long enough wavelength to avoid severe fluorescence, but still operating in the silicon detector region. That kind of innovation could reshape field protocols and expand Raman’s reach into more challenging environments.
Looking ahead, you mention plans to explore deep-UV excitation using the same LEGO standards. What do you hope to learn from that work, and why is deep-UV an important next step?
Deep-UV excitation, pioneered by Prof. Sandy Asher at the University of Pittsburgh, USA, separates the fluorescence and Raman signals in wavelength space and can also provide resonance enhancement. But the components required for that work are immature, may have limited lifetimes, and can photolyze samples due to UV radiation. There’s published evidence from studies on street drugs that resonance enhancement alters the spectra of the active ingredients, compared with those obtained using conventional excitation. This implies that a complete set of new spectral libraries would be required, so deep-UV excitation also may not be a complete solution. But for field operation, this approach can be immune to ambient light because solar-blind detectors are used, which is especially attractive for stand-off applications, along with the short measurement times required. It will be interesting to see how spectra of the Legos obtained using this technique compare with those we have already measured.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
