OR WAIT null SECS
Surface-enhanced Raman scattering (SERS) is used to improve Raman scattering, often allowing the detection of single molecules. It generates molecularly specific fingerprints of analytes and, with carefully controlled experimental conditions, can be highly quantitative. Roy Goodacre, a professor of biological chemistry at the University of Liverpool in the United Kingdom, first used this technique to achieve whole-organism fingerprinting of bacteria and then explored SERS in a variety of other applications, including within biotechnology, disease diagnostics, quantitative detection, imaging, food security, and more. Goodacre is the 2021 winner of the Charles Mann Award for Applied Raman Spectroscopy. This interview is part of an ongoing series of interviews with the winners of awards that are presented at the annual SciX conference, which will be held this year from September 26 through October 1, in Providence, Rhode Island.
Has your work with SERS solved some of the challenges commonly associated with Raman spectroscopy for quantitative analysis, such as native sample fluorescence and standardization of Raman intensity?
That’s a great question, and many people believe SERS to lack reproducibility. In fact, SERS can be made highly quantitative, which addresses the second half of your question in terms of standardization. Our group have developed SERS for quantitative analysis and embedded design of experiments in order to search the possible experimental landscape in terms of type of metal ions to reduce, how to perform the reduction (which dictates the “capping” agent on the nanoparticle surface), as well as conditions used to initiate aggregation (in terms of salts as well as controlling the pH of the environment). When you do this, you can come up with nice bespoke protocols that lead to high levels of standardization.
In terms of fluorescence, we very rarely see this. This is due I believe (I’m no physicist) to energy transfer occurring during the excitation. In fluorescence the molecules in the system absorb energy and are in the electronic excited state. After non-radiative decay fluorescence is seen as the emission of light at a longer wavelength to the incident radiation. However, in SERS which uses metallic nanoparticles, a type of FRET mechanism occurs where the light is not emitted but this energy is transferred to the metal. I might have that a bit wrong, but anyway SERS effectively quenches the normal fluorescence process. This is why a lot of people still use Rhodamine 6G which is a highly fluorescent dye to illustrate that their SERS conditions work. There are possibly more SERS spectra of this molecule in the literature than your readers have had hot diners!
In your paper “Recent developments in Quantitative SERS: Moving towards absolute quantification” (1) you and your colleagues critiqued the development of quantitative SERS from simple univariate assessment of single vibrational modes to multivariate analysis of the whole spectrum for improved quantification. What knowledge did you and your team gain from this study?
This was actually a review that I wrote along with my colleagues Professors Karen Faulds and Duncan Graham (from the University of Strathclyde in Scotland), rather than a primary paper. I’m glad you ask about this, as reviews are really good ways to synthesize knowledge and thus to refine concepts and to formulate new ideas. In this review, we specifically looked at how fellow scientists had developed SERS to be quantitative and thus used to provide reproducible and robust values for absolute levels of key analytes (or what we may term determinands). There are a number of things that can affect quantification, and these include (among others) variation in laser fluence during the experiment as well as the number of aggregated nanoparticles with associated molecules within the SERS/Raman collection voxel. At the end of this process, the CCD within a Raman spectrometer will give readouts of energy.
SERS is no different from using, say, liquid chromatography-mass spectrometry (LC–MS) to quantify drugs in athletes’ blood and urine during the Olympics. LC–MS only provides absolute levels of these drugs and their metabolites when the detector (the MS in this case) is calibrated appropriately. SERS is no different, one needs to calibrate the system, and once you do this then absolute quantification is readily achieved.
What we learned and reported in this review was that several groups had come to the conclusion that two general experimental strategies can be used to calibrate the whole system so that absolute quantification (for example, µM or ng/mL of a target in, as an example, blood or urine) could be accomplished. Rather interestingly, both these approaches have been borrowed from other fields including mass spectrometry!
The first method involves spiking the sample with a stable isotope of your molecule. This can be 13C or 15N, but what is particularly useful is the use of deuterium (2H or D) as the reduced mass changes (for example, substitution of C–H with C–D) cause very large shifts in vibrational frequencies. This method has been termed isotope dilution surface-enhanced Raman scattering (IDSERS) and the use of isotopologues of the target molecule overcomes any competitive co-adsorption or co-association with the metal surface, since the natural isotope and stable isotope have very similar physical properties. This also overcomes any laser power fluctuations as both analyte and internal standard are recording simultaneously. The isotope standards (isotopologues) can then be used as subtractions or as ratios if key peaks are clearly seen within the spectra for both natural and stable isotope. Alternatively, one can use multivariate analyses such as linear regression techniques like partial least squares (PLS) which uses the full spectrum, or multivariate curve resolution (MCR) which unmixes spectra to generate a series of ‘pure’ spectra as well as corresponding concentration values.
The second method is to use the standard addition method (SAM) where the standard here is the analyte you want to quantify. In SAM the standard is the natural isotope and so will have the same SERS spectrum as the molecule within the sample. The standard is then ‘spiked’ into the sample multiple times at different known levels. The areas of diagnostic peaks are then calculated and plots of these peak areas against the concentration of the spiked standard will yield a straight line. The key thing here is that the zero-level spike will still have a peak from the target molecule and so the “0” spike concentration will intercept the y-axis. The slope can then be used to calculate the level of the molecule in the sample with high precision and accuracy.
These methods both give absolute quantification of targets and are excellent methods “borrowed” from other areas of analytical chemistry.
What are some of the newest uses of SERS for clinical diagnostics?
Well, let’s look at what effect the last 18 months or so have had on the SERS field. Many people have shifted their work towards detecting whether someone has Covid-19 or not. Some of these studies have been based on profiling a body fluid. For example, saliva has been used in a recent study by Carlomagno and colleagues (https://doi.org/10.1038/s41598-021-84565-3), where multivariate analysis was then used for identification of SERS patterns that are associated with infection. Alternatively, the virus can be measured directly using PCR-based approaches or via antibody detection and these methods are coupled with SERS where the read out is SERS of a specific target dye. These methods, including lateral flow devices, feature in a recent review by Saviñon-Flores et al. (https://doi.org/10.3390/bios11030066).
I guess what this pandemic has achieved is to focus people’s minds on diagnostics. This proves that this area, and especially point of care diagnostics, is important and has allowed mass screening of populations. And of course, all these methods of diagnostics are embedded within excellent analytical science.
Do you see SERS being implemented in practical clinical diagnostics? And is quantitative analysis important?
In addition to Covid-19 detection discussed above, other developments of SERS for clinical application are highlighted in our review in Trends in Analytical Chemistry (TrAC) (1). I think one of the most interesting of these includes SERS methods for measuring drugs and their metabolites (xenometabolites) directly in human biofluids (for example, blood and urine). Because no chromatography is used, these methods are very rapid, with results generated in a few minutes. Here, quantification is very important and the measurement of drugs and their xenometabolites also allows one to calculate pharmacodynamics. In addition, SERS can also be used to know if someone has taken an illicit drug several days ago, because drug metabolites are generally cleared from the body at a slower rate than the drug itself.
The speed of analysis here is a huge advantage because many manufacturers are producing handheld and portable Raman instruments that can be taken to the patient, rather than shipping a sample to a central testing facility. I see the future of SERS being within diagnostic point-of-care applications.
In another paper, you described how SERS was used for simultaneous detection of multiple novel psychoactive substances (NPS) as an alternative to various hyphenated methods such as gas chromatography (GC) or liquid chromatography with mass spectrometry (LC–MS) (2). What advantages does the use of SERS have over these hyphenated methods for quantitative analysis of NPS?
The main advantages are similar to those articulated above. SERS can be made portable and is rapid, which means that the methods can be used for on-site for testing. This could be for testing someone for illicit use of these substances.
An alternative application would be to develop and deploy SERS within drug testing centers in large events like music festivals to check the authenticity of NPS. Unfortunately, some of these psychoactive drugs are impure or may contain NPS fentanyl analogs or indeed fentanyl itself. These can be lethal and there have, unfortunately, been several cases in the US, so authentication of drugs prior to use is really important.
What challenges do you think that SERS will have in sample presentation for quantitative detection and discrimination of a range of NPS?
The main challenges will be sample processing in order to extract NPS from body fluids and then ensuring that the NPS associates with the metal surface of the nanoparticles because the analyte needs to be close to or bound to the surface. Neither of these are straightforward, but we can again borrow approaches that we use in our mass spectrometry-based metabolomics. If we want to know the actual concentration in a biofluid then one can adjust for any variance in the extraction efficiency by spiking in isotopes of the NPS target molecule(s) directly into the sample being measured before any sample preparation has taken place. This means that any variation in sample preparation is accounted for as the NPS and its isotope equivalent have the same chemistries and so will be extracted with highly similar efficiencies. Any extraction and measurement errors can then be accounted for by using the ratios of the isotopes in the naturally occurring NPS and comparing this with any stable isotopes in the NPS spike. This approach is also discussed in our review published in TrAC (1).
For the discrimination of a wide range of NPS what we reported in (2) is that the SERS spectra contain vibrational features that reflect their core chemical structures. When we performed unsupervised cluster analysis using PCA (principal components analysis) we found that the NPS samples grouped into three distinct clusters that were identified as methcathinones, aminoindanes and diphenidines. This is potentially very exciting and could be exploited in the future. Clandestine chemists generate new NPS by modifying a designer drug by the addition of chemical groups to a core chemical structure, and that core reflects the psychoactive activity of the drug. Being able to identify the core structure of an NPS is as important as having a method that gives specific spectral fingerprints that allow for the identification of individual NPS.
Is there a specific direction you see this technique developing into over the next few years? In other words, are there unmet challenges to this technique that if resolved would bring broader applications?
SERS, at the moment, is mainly used within research laboratories. However, the realization that this method can be made to be robust, reproducible, highly quantitative, and that it is inherently portable suggests to me that it is only a matter of time before this technique can be translated from the laboratory to the clinical environment.
I’m not the only one to think this. Inspired by employing SERS for therapeutic drug monitoring, Professor Alois Bonifacio (from the University of Trieste in Italy) led a large-scale multi-instrument interlaboratory study, as part of the EU COST Action BM1401 Raman4Clinics to address whether equivalent data can be collected in different laboratories. This study was published last year in Analytical Chemistry(https://doi.org/10.1021/acs.analchem.9b05658) and shows that SERS can indeed be standardized, and that once standard operating procedures (SOPs) are adopted, then SERS is robust enough to allow for the quantification of target molecules. Our group was proud to be part of this study which involved 15 different laboratories throughout Europe. Our study was one of the first multi-center SERS studies.
Of course, central to SERS is the instrument itself. In another study also published last year in Analytical Chemistry (https://doi.org/10.1021/acs.analchem.0c02696) Professor Thomas Bocklitz (from the University of Jena in Germany) lead another Raman4Clinics study that was designed to address how Raman spectrometers can be configured to provide robust results and thus help improve inter-laboratory studies using Raman spectroscopy. This was also a large-scale round robin study involving 15 institutes within seven European countries and compared 35 Raman spectroscopic devices with different configurations. We were also involved in this study, and it led to a series of recommendations that we believe are worthy of future investigations to improve multi-laboratory studies.
I believe that both these studies are very important and show that SERS has great potential for adoption within clinical environments.
What are your next steps in your own work with Raman spectroscopy and SERS?
Lots, but I will just mention one very exciting area that we are developing.
Within the University of Liverpool’s Centre for Metabolomics Research (CMR) we are establishing a variety of Raman-based methods for chemical image analysis of tissues, cells, and bacteria. We have capabilities in spontaneous Raman spectroscopy from Renishaw, stimulated Raman spectroscopy (SRS), and coherent anti-Stokes Raman spectroscopy (CARS) from Leica. We also have a combined Optical Photothermal IR and Raman system from Photothermal Spectroscopy Corp.
Our group is very excited by these techniques, and we hope to combine vibrational spectroscopic imaging with mass spectrometry-based metabolomics. For those interested in the CMR please visit: https://www.liverpool.ac.uk/liverpool-shared-research-facilities/facilities/multi-omics/cmr/
Within these pages there are a few case studies, and to whet your reader’s appetites we have recently published papers on “metabolism in action” where we have been developing stable isotope probing using Raman and infrared imaging for understanding metabolic flux at the single bacterial cell level. This is really pushing the resolution limits of these platforms, and we’re very excited by the results.
Before we close out this interview, I would like to take this opportunity to thank FACSS for the Charles Mann Award. I was really shocked, absolutely thrilled and honored to find out about being given this award. There have been some exceptional scientists that I have looked up to (and indeed still do!) in the Raman field who have been bestowed this award, and I am amazed to be in their company. It’s a wonderful reflection of the work our fantastic research team has achieved over the last 20 years, and they are as deserving as I. Thanks to them all!
I'm also really hoping that I will be able to travel to Providencein September to meet up with friends and colleagues and of course to make new connections. I've certainly missed the international conference scene. I'm reminded of going to ICORS 2018 in Jeju and meeting Professor Luis Liz-Marzán (from CIC biomaGUNE in Spain) during a session or two on SERS. This chance meeting led to a substantial review by nearly 70 authors on SERS. A review that, while only being published two years ago, has already been cited over 500 times. If you’re really interested in SERS, then this tour-de-force in ACS Nano on the “Present and Future of Surface-Enhanced Raman Scattering” (https://doi.org/10.1021/acsnano.9b04224) is a must read.
(1) R. Goodacre, D. Graham, K. Faulds, TrAC, 102, 259–368, (2018).
(2) H. Muhamadali, A. Watt, Y. Xu, M. Chisanga, A. Subaihi, C. Jones, D.I. Ellis, O.B. Sutcliffe, and R. Goodacre, Front. Chem., 7:412, (2019).
Roy Goodacre received his degree in microbiology from the University of Bristol (Bristol, England) and was awarded his PhD in 1992 in mass spectrometry applied to microbiological problems (also from Bristol). He is Professor of Biological Chemistry at the University of Liverpool and leads a team of around 20 researchers in the Department of Biochemistry and Systems Biology within the Institute of Systems, Molecular and Integrative Biology (ISMIB). Additionally, he is the Chair of the Institute's Research and Impact Committee, Deputy Head of Department, and a co-director of the Centre for Metabolomics Research. His research interests include analytical biotechnology and disease diagnostics, detection, imaging, food security, as well as systems and synthetic biology. He has more than 30 years of expertise in mass spectrometry (MS)-based metabolomics, Raman spectroscopy, and advanced multivariate data analysis.
Google Scholar: https://scholar.google.com/citations?user=Dszi0U8AAAAJ&hl=en