Q&A Spectroscopy gives clear view of data

Spectroscopy is a broad field comprising a variety of methods and modalities, and covering applications from industry process control and security to biomedical. Each modality has its particular advantages and disadvantages, including speed, sensitivity, and ease of implementation. Nevertheless, they share common factors and trends.

Photonics Media spoke with Iwan W. Schie, leader of the Multimodal Instrumentation work group in the Leibniz Institute of Photonic Technology’s (IPHT’s) Department of Spectroscopy and Imaging in Germany. He has a doctorate in biomedical engineering from the University of California, Davis. During his time in the program, his research focused on instrument development and medical applications for multiphoton microscopy and spontaneous Raman spectroscopy. In his current position with Leibniz-IPHT, he is researching high-throughput Raman spectroscopy systems for single-cell classification, multimodal fiber-probe development, and instrumentation for medical in vivo applications.

We asked him to share his take on recent advancements in spectroscopy, exciting areas of research, and what the future holds for the field.

Q: What trends are you seeing in the spectroscopy field?

A: From the research side, the most common trend for spectroscopic modalities is the combining of methods, most frequently with optical modalities, to increase the wealth of information about a sample. Some spectroscopy modalities, such as Raman spectroscopy, tend to be slow but have a spectacular informational depth. For example, in the medical field it is quite advantageous to combine Raman spectroscopy with methods such as optical coherence tomography (OCT), fluorescence lifetime imaging microscopy (FLIM), or even fluorescence spectroscopy. Also advantageous is the use of faster methods as a red-flag technology for the identification of regions of interest, and to use Raman spectroscopy to provide a precise molecular profile, which can be used for diagnostics.

Obvious general trends, including the application of artificial intelligence (AI) approaches for spectroscopic data evaluation, are also more frequently found in the field. However, here the spectroscopy community has already been at the forefront of machine-learning approaches (frequently termed chemometrics) which were applied for decades to better comprehend spectroscopic data, and which in most situations have a multivariate structure. As such, broader availability of larger data sizes will enable a more comprehensive use of these approaches.

On the instrumentation side specifically, the trend is the development of compact and high-performance spectrometers and detectors, which results in an upswing for spectroscopic applications in the consumer market. Today’s consumers are more data-driven, and have sharpened awareness of their habits, using a variety of wearable tracking devices to monitor sleep patterns, calorie intake, or pulse. Here, spectroscopic modalities offer new ways for monitoring — for example, chemically evaluating the freshness of dietary products or determining authenticity of produce. Who would not like to have a mobile phone that provides chemical information about their dietary products or lets them chemically identify things in their surroundings?

Q: What exciting projects are you currently working on?

A: In particular, molecular spectroscopy offers the possibility to identify and characterize chemical signatures of samples rapidly, nondestructively, and without any preparation steps. At least this is the premise for using molecular-sensitive spectroscopy methods such as vibrational spectroscopy, especially in the field of biomedical applications. However, when looking at current devices and commonly performed applications, current spectroscopy devices heavily require human interaction, whether to perform the experiments or to evaluate the data.

Both come at a significant time cost, making it challenging for the casual user who is simply interested in applying the method to get clear-cut information. To overcome this, one of our own specific research topics involves developing devices — or an entire application procedure for that matter — that require little to no user interaction, harvesting all of the advantages of spectroscopic approaches but providing instantaneous and easy-to- understand answers for the user. While this sounds easy, it requires a comprehensive understanding of all of the different aspects of spectroscopy — instrumentation, data processing, and sample handling — to achieve this goal.

On par with this topic is the development of high-throughput Raman devices for a large-scale sampling of cells. Here again, the proper device development and application to real-world scenarios significantly increases the attraction of the method. Another topic related to device translation to clinical applications is that it requires not only extensive engineering know-how but also has to consider the regulatory framework — MDR 2017/745 — which imposes extensive regulatory rules for moving research devices to real in vivo medical applications.

From the point of view of a research institute, this is a particularly challenging and comprehensive undertaking. The combination of spectroscopic modalities, as described above, is also of significant interest.

Q: What implications for the industry are emerging from spectroscopy research and technology?

A: Spectroscopy applications are gaining significant interest in industrial process controlling because they provide immediate information about chemical signatures, concentration profiles, modifications of chemical structures, and changes in chemical ratios. Each set of accessible information is an important factor for process controlling.

Processes are usually analyzed by random sampling of the product, which has to be sent for extensive analysis to external laboratories. This is not only time-consuming but also a significant cost factor. The reader would be surprised to know how many production processes are controlled by the experience of the user or controlling parameters, which have been established empirically and not controlled through closely monitored, in-line analytical tools. Spectroscopy-based methods, however, can nondisruptively profile the process and provide instant information.

Quite frequently, large batches have to be discarded, resulting in significant economic damage to the companies and possible damage to the environment. Here, the spectroscopy-based methods could have a significant impact on different producing industries.

Q: What does the future hold for spectroscopy (in R&D, diagnostics, etc.)?

A: Spectroscopy is currently on the brink of a transformation, primarily driven by the development of sophisticated devices and improved components at reduced cost. This, combined with miniaturization and a general desire for an increase of monitoring and access to information in consumer electronics, will provide significant drive and further increase accessibility and novel applications.

In terms of medical applications in diagnostics, there has been an increasing number of feasible applications shown. However, it is challenging to have studies with hundreds of patients and in different clinical settings to properly validate modalities and to further establish the acceptability in the medical field.

This problem of acceptability is not unique to spectroscopy; it is a challenge for many other optical methods. For industry applications, it is much easier to translate spectroscopy because the regulatory framework and experimental evidence are much easier to attain than in the medical field. Here, actually, the developer of particular industrial equipment can push new in-line and on-line analytical modalities and implement them in their equipment, which will be more acceptable by the customers.

The last decade has seen spectacular developments in spectroscopic applications, and in the next decade there will be even more possibilities. Overall, the next decade will see new excitement in spectroscopy.