Manufacturers in many industries depend upon processes for combining, separating or identifying chemical and biological compounds. Makers of pharmaceuticals, foods, beverages, chemicals, plastics and agricultural products mix or blend multiple substances during their production processes. Producers of oils, gases, and environmental substances often separate chemicals to ensure the purity of their process outcome. Even the homeland security industry must identify chemical or biological compounds.
Food and beverage manufacturers have become increasingly dependent upon control processes. The industry has stepped up efforts to assure product and quality control, and as such, food and beverage producers have been continually searching for an ideal method for testing and analyzing their products.
Looking Back on Limitations
For many years, scientists have widely used optical spectrometers to fulfill their process control needs. Use of these instruments has become a choice method by which scientists have chosen to analyze product samples. This process occurs in the laboratory setting, removed from industrial process lines.
Several factors are putting in question the viability of sampling methods, as means to ensure process control and final product assurance:
- High-speed and high-volume industrial production require real-time, in-line process control for minimizing the costly waste generated by the long feedback time associated with sampling and laboratory analysis;
- Public opinion and federal regulation demand strict control of final product characteristics;
- Safety awareness and regulation restrict exposure of personnel to many chemicals
- Competitive pressure requires highly automated production lines with minimum personnel attendance;
- Consistent quality necessitates elimination of human errors and guesswork.
In order to address those issues, analytical scientists and instrument manufacturers have naturally focused on taking optical spectrometers out of the labs, and finding ways to bring them as close as possible to process lines, in at-line, on-line, and hopefully in-line modes of application. In doing so, they have encountered a number of challenges and difficulties, which have even posed questions on the viability of spectroscopy as a process control technique:
- Optical spectrometers are typically slow and hardly suited for taking the high-frequency measurements required by high-speed production lines;
- Optical spectrometers are delicate laboratory-grade instruments, hard to directly couple to process lines, and of problematic use in industrial environments;
- Optical spectrometers need complex and frequent calibration, and the result of their measurements needs interpretation by highly-qualified resources;
- Presentation of compounds being processed to optical spectrometers is complex and costly, requiring from remote transmission of light through signal deteriorating fiber optic probes, to material handling techniques for automated sampling;
- The total cost of installation and operation of an optical spectrometer to an industrial process line can be a manifold of the original cost of the instrument.
Evidently, trying to apply laboratory instruments to in-line applications is not the best course of action. But until recently, it has been the only course of action due to lack of a suitable optical spectroscopy technology expressly conceived for real-time, in-line process control.
New Technology Falls in Line
Over the course of 10 years, Dr. Michael Myrick of the University of South Carolina has developed a breakthrough technology. This technology leverages the power of optical spectroscopy by making it extremely fast, insensitive to the rigors of industrial environments, and fully automated in its operation.
Called Multivariate Optical Computing (MOC), this technology has been refined and translated into products for industrial applications by Ometric Corp. (Columbia, S.C.). Pattern recognition techniques can be applied to optical spectra of complex mixtures for the purpose of detecting the presence of specific compounds, measuring their concentrations, or estimating properties that depend on sample chemistry. Based on multivariate (multi-wavelength) calibration, this approach to chemical measurement is powerful, but it is used infrequently because of the expense and complexity of the spectroscopic tools required.