Multimodal Spectroscopy: Production Workhorse

The European Commission’s smart manufacturing vision “Manufuture for 2020” and the US FDA PAT/QbD-platform (Process Analytical Technology, Quality by Design) are both helping to increase interest in the concept of intelligent manufacturing. It is a transdisciplinary technology where process chemists, process engineers, chemometricians and many other technologists must work together. In short, the holistic process analysis component will be the bedrock that supports the production of smart materials in smart factories! Indeed, process analytics by spectroscopy can improve understanding of how the process operates, and can be used to determine potential targets for process improvement by removing waste and increasing efficiency (1).

I have no doubt that optical spectroscopy – together with chemometrics – will play an important role in transforming industry from reactive to proactive production. Because spectroscopic techniques can detect morphological (from scatter) and chemical features (from absorption) simultaneously, the complete fundamental functionality of a compound is inherent in every spectrum.

However, we must recognize that sensitivity, selectivity and robustness of each individual technology, in combination with the wavelength range used, has limitations because of the structure of the measured species and the optical configuration selected. Furthermore, in any application, a key issue is finding the causal link between the measured spectral features and the final target quality. I believe multivariate data analysis of big data will be a key technology in the future. Proper process analysis means understanding the causal relation between measurement and response, and with spectral imaging, the spatial distribution in the x-, y- and, possibly, z-direction may also be of interest (1).

Many of you will know that ultraviolet- and visible (UV/Vis) spectroscopy is a highly sensitive technique for electronic transitions, while mid infrared (MIR) spectroscopy is specific for vibrational transitions. However, we also know that near infrared (NIR) spectroscopy is less sensitive than MIR due to lower cross sections of higher order vibrational transition probabilities. Clearly, the major advantage with NIR is that even at higher concentrations no sample preparation (for example, dilution) is necessary, but I think it is important to emphasize that both NIR and MIR spectroscopy are highly sensitive to water absorption. And in recent years, Raman spectroscopy has developed into a highly sensitive and versatile technique, proving to be a very good process-monitoring tool, especially in aqueous systems such as those found in biotechnology.

Currently, NIR- and Raman-spectroscopy are the workhorses in PAT applications, and multimodal spectroscopy will be the all-in-one sensor of the future. Samples containing phase boundaries that display simultaneous and superimposed wavelength dependent absorption and scattering effects cannot be characterized by a single measurement. Here, multimodal spectroscopy will come to the fore because of its ability to deal with combinations of measurements in different wavelength regions or in different optical set ups (for example, transmission and reflection).

In the future, special emphasis will be given to measuring not only the chemical entities but also their lateral distribution in an object. Spectral imaging (also known as chemical imaging) is an emerging field for a wide range of applications; for example, finding biomarkers in a tissue or controlling and qualifying pharmaceutical tablets or food. In addition, push broom imaging (line scanning) technology will allow multiplexing, thus reaction tomography or measurements in micro-reactor systems will be possible.

Therefore, I am very confident that process analysis, together with spectroscopy and intelligent data analysis, will play a more important role in modern manufacturing and processing. The German government’s “Industry 4.0” concept describes the future of industrial automation as being arbitrarily modifiable and expandable (flexible), able to connect different components from multiple producers, enabling those components to perform tasks related to a context independently (self-organizational), with an emphasis on ease of use (user-oriented). Spectroscopy – particularly the workhorse techniques of vibrational spectroscopy – will be an important set of tools to realize this concept (2).