Method and apparatus for identifying a substance using a spectral library database

Abstract

A spectroscopic detector for identifying the presence of a first substance in the presence of another substance includes a laser for illuminating the substances at a plurality of wavelengths to induce the emission of radiation characteristic of the substance; a spectrometer for measuring the emitted radiation to obtain a plurality of spectral measurement data; and a processor for processing the data. An algorithm combines the data into a composite spectrum and a parameter characteristic of the first substance is identified while information in the composite spectrum contributed by emission of radiation from the other substance is removed to identify the presence of the first substance and obtain a characteristic spectral signature of the first substance. The signature is compared to signatures in a spectral library database, wherein at least some of the library signatures have spectral characteristics differentiated from each other by identifiable spectral characteristics caused by environmental factors.

Description

CROSS REFERENCE TO RELATED APPLICATIONS [0001] The present application claims the benefit of the priority filing date of provisional patent applications No. 60 / 535,179, filed Jan. 7, 2004, and No. 60 / 601,180, filed Aug. 13, 2004, both incorporated herein by reference. [0002] The present application is related to U.S. Ser. No. ______ , entitled “METHOD AND APPARATUS FOR IDENTIFYING A SUBSTANCE”, filed concurrently herewith.FIELD OF THE INVENTION [0003] This invention relates to a method and apparatus for identifying a substance in the presence of other substances. More particularly, the invention relates to a method and apparatus for identifying chemical, biological or other constituents of interest in the presence of other substances by comparing spectral signatures in a spectral Library database. BACKGROUND OF THE INVENTION [0004] Spectroscopic identification of bio-organisms utilizing resonance or near-resonance-Raman spectroscopy, described in U.S. Pat. No. 4,847,198, incorporate...

AI Technical Summary

Benefits of technology

[0012] In these embodiments, the invention overcomes the prior art limitations noted above by acquiring spectra, preferably resonant and near-resonant Raman Spectra, that are more complete and contain more information. It also overcomes prior art limitations noted above by utilizing a powerful code to analyze the information contained in these spectra. For example, in an application in which one hundred different illumination wavelengths are used, the acquired spectra contain as much as 100 times the information of the traditional single illumination wavelength Raman spectrum. This provides an increase of specificity and a greater resistance to interference from background clutter. The embodiment requires no pre-enrichment or only minimal pre-enrichment.

[0014] The invention provides a new tool with which to study the protein or DNA / RNA markers of microorganisms and cells, as well as a tool to rapidly identify and count organisms that cause diseases. The invention also provides a tool to identify compounds, including chemicals in the presence of other chemicals. Specific applications include air monitoring, water monitoring, monitoring during food production, monitoring during production of pharmaceuticals, rapid detection of targeted disease organisms such as tuberculosis, pre and post sterilization monitoring, monitoring of allergens, identification of folded proteins (prions), and monitoring of blood constituents.

Problems solved by technology

This prior art technique has very limited ability to identify species.

The ability to distinguish gram+from gram−bacteria has been demonstrated, but more specific identification has not been possible.

Excitation at a single wavelength may excite not just the substance of interest but also some or all of the other substances present, spectrally masking its signature and making it difficult to interpret the emitted spectra and to identify the substance of interest.

Other approaches have applied spectral data processing algorithms to data resulting from a single illumination wavelength of pure samples but have demonstrated only a limited capability to distinguish between spectral “fingerprints”, that is, the ability to identify a signature of a particular species, organism, or substance.

None have been successful identifying an organism in the presence of other substances and / or organisms.

Some require obtaining sets of training data and do not therefore lend themselves effectively to real time processing needs.

Others exhibit limited ability to identify organisms due to the inherent limitations of their spectral data processing methodology.

Images