Multivariate optical computing

Multivariate Optical Computing is an approach to the development of spectroscopic instruments, particularly for industrial applications such as process analytical support. "Conventional" spectroscopic methods often employ multivariate methods to extract the concentration (or other analytical information) from data collected at many different wavelengths. Multivariate optical computing uses an optical computer to analyze the data as it is collected. The goal of this approach is to produce instruments which are simple and rugged, yet retain the benefits of multivariate techniques for the accuracy and precision of the result.

An instrument which implements this approach may be described as a multivariate optical computer. Since it describes an approach, rather than any specific wavelength range, multivariate optical computers may be built using a variety of different instruments (including FTIR[1] and Raman[2]).

The "software" in multivariate optical computing is a Multivariate optical element (MOE) which is specific to the particular application. The MOE is designed for the specific purpose of measuring the magnitude of a multi-wavelength pattern in the spectrum of a sample, without actually measuring a spectrum.

Multivariate Optical Computing allows instruments to be made with the mathematics of pattern recognition designed directly into an optical computer, which extracts information from light without recording a spectrum. This makes it possible to achieve the speed, dependability, and ruggedness necessary for realtime, in-line process control instruments.

Multivariate Optical Computing encodes an analogue optical regression vector of a transmission function for an optical element. Light which emanates from a sample contains the spectral information of that sample, wether the spectrum is discovered or not. As light passes from a sample through the element, the normalized intensity, which is detected by a broad band detector, is proportional to the dot product of the regression vector with that spectrum, i.e. is proportional to the concentration of the analyte for which the regression vector was designed. The quality of the analysis is then equal to the quality of the regression vector which is encoded. If the resolution of the regression vector is encoded to the resolution of the laboratory instrument from which that regression vector was designed and the resolution of the detector is equivalent, then the measurement made by Multivariate Optical Computing will be equivalent to that laboratory instrument by conventional means. The technique is making headway in a niche market for harsh environment detection. Specifically the technique has been adopted for use in the oil industry for detection of hydrocarbon composition in oil wells and pipeline monitoring. In such situations, laboratory quality measurements are necessary, but in harsh environments.[3]

References

  1. ↑ Myrick, Michael L.; Haibach, Frederick G. (2004-04-01), "Precision in Multivariate Optical Computing", Applied Optics 43 (10): 2130–2140, doi:10.1364/AO.43.002130, PMID 15074423, retrieved 2006-12-18
  2. ↑ Nelson, MP; Aust, JF; Dobrowolski, JA; Verly, PG; Myrick, Michael L. (1998), "Multivariate optical computation for predictive spectroscopy", Analytical Chemistry 70 (1): 73–82, doi:10.1021/ac970791w
  3. ↑ Jones, Christopher M.; et al. (2014-08-30), "Multivariate Optical Computing enables Accurate Harsh Environment Sensing for the Oil and Gas Industry", Laser Focus World 50 (08): 27–31, doi:10.1364/AO.43.002130, PMID 15074423 |accessdate=2014-08-30

Bibliography of Multivariate Optical Computing

This article is issued from Wikipedia - version of the Wednesday, March 09, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.