US 5,668,373 AGranted

Methods and apparatus for analysis of complex mixtures

AssigneeTrustees of Tufts College

InventorsAlbert Robbat, Jr., Yuriy V. Gankin

Life of this patent

Estimating the term…

Loading patent…

Abstract

This invention provides methods and apparatus for determining the presence or absence of a specific constituent in a mixture. The method includes the steps of (a) applying energy to the mixture to cause constituents in the mixture to produce characteristic patterns of spectral lines and detecting quantity values at the spectral lines, wherein a quantity value may represent a contribution of two or more constituents; (b) comparing detected quantity values for a number N of spectral lines from a pattern of spectral lines characteristic of the specific constituent with known quantity values for the N spectral lines; (c) determining whether the detected quantity values for the N spectral lines fall within a predetermined error range, and; (d) determining the presence of the constituent in the mixture when detected quantity values of a number of spectral lines less than N fall within the predetermined error range.

Claims24

Dependent claims

An apparatus for determining the presence or absence of a specific constituent in a mixture, said apparatus comprising

(a) an energy source arranged to apply energy to the mixture to cause constituents in the mixture to absorb, emit, or fluoresce particles of energy;
(b) a spectral separation unit arranged to separate said particles of energy into patterns of spectral lines, wherein each line represents one particle, the size of each line represents the quantity of each type of particle, and each pattern of spectral lines is characteristic for one constituent, and wherein a quantity value may represent a contribution of two or more constituents;
(c) a detector arranged to detect the particles and produce data representing said patterns of spectral lines;
(d) a memory arranged to store the data, a library containing a known set of ratios for spectral lines of a pattern characteristic of said constituent, and predetermined error ranges;
and
(e) a processor arranged to analyze the data, wherein the processor is programmed to:
(f)(i) calculate a set of relationships R.sub.N of detected quantity values for a number N of spectral lines within a pattern characteristic of said specific constituent;
(f)(ii) compare relationships R.sub.N with a known set of relationships for said N spectral lines of said constituent in a library, wherein correspondence of said relationships R.sub.N to said known library relationships within a predetermined error range is taken to indicate that said constituent is present in the mixture;
wherein if said relationships R.sub.N do not correspond to said known library relationships within said error range,
(g)(i) select a first subset N.sub.s of N spectral lines wherein N.sub.s is less than N;
(g)(ii) calculate a set of relationships R.sub.S of detected quantity values for said N.sub.s spectral lines in said first subset;
(g)(iii) compare said first subset relationships to said library relationships and record a first subset variation value based on variations between the first subset relationships and said library relationships;
(g)(iv) repeat said selecting, calculating, and comparing steps (i-iii) for different subsets N.sub.s of N spectral lines, and record a variation value for each subset;
and
(g)(v) determine if a subset has a variation value within a predetermined error range and if so, using said subset to indicate that said specific constituent is present in the mixture.

12 more elements

A method for determining the presence or absence of a specific constituent in a mixture, said method comprising

(a) applying energy to the mixture to cause constituents in the mixture to produce characteristic patterns of spectral lines and detecting quantity values at said spectral lines, wherein a quantity value may represent a contribution of two or more constituents;
(b) calculating a set of relationships R.sub.N of detected quantity values for a number N of spectral lines within a pattern characteristic of said specific constituent;
(c) comparing relationships R.sub.N with a known set of relationships for said N spectral lines of said constituent in a library, wherein correspondence of said relationships R.sub.N to said known library relationships within a predetermined error range is taken to indicate that said constituent is present in the mixture;
wherein if said relationships R.sub.N do not correspond to said known library relationships within said error range,
(d)(i) selecting a first subset N.sub.s of N spectral lines wherein N.sub.s is less than N;
(d)(ii) calculating a set of relationships R.sub.S of detected quantity values for said N.sub.s spectral lines in said first subset;
(d)(iii) comparing said first subset relationships to said library relationships and recording a first subset variation value based on variations between the first subset relationships and said library relationships;
(d)(iv) repeating said selecting, calculating, and comparing steps (i-iii) for different subsets N.sub.s of N spectral lines, and recording a variation value for each subset;
and
(d)(v) determining if a subset has a variation value within a predetermined error range and if so, using said subset to indicate that said specific constituent is present in the mixture.

7 more elements

A computer program for analyzing data representing patterns of spectral lines of a constituent in a mixture, wherein each line represents one particle from the constituent, the size of each line represents the quantity of each type of particle in the mixture, and each pattern of spectral lines is characteristic for one constituent, said program residing on a computer-readable medium and comprising instructions for causing a processor to:

(a) select a number N of spectral lines from a pattern characteristic of said specific constituent;
(b) calculate a set of relationships R.sub.N of detected quantity values for said N spectral lines;
and
(c) compare relationships R.sub.N with a known set of relationships for N spectral lines of a pattern characteristic of said constituent in a library, wherein correspondence of said relationships R.sub.N to said known library relationships within predetermined error ranges is taken to indicate that said constituent is present in the mixture;
wherein if said relationships R.sub.N do not correspond to said known library relationships within said error ranges,
(d)(i) select a first subset N.sub.s of N spectral lines wherein N.sub.s is less than N;
(d)(ii) calculate a set of relationships R.sub.S of detected quantity values for said N.sub.s spectral lines in said first subset;
(d)(iii) compare said first subset relationships to said library relationships and record first subset variation values based on variations between the first subset relationships and said library relationships;
(d)(iv) repeat said selecting, calculating, and comparing steps (d)(i-iii) for different subsets N.sub.s of N spectral lines, and record variation values for each subset;
and
(d)(v) determine if a subset has variation values within predetermined error ranges, and if so, use said subset to indicate that said specific constituent is present in the mixture.

8 more elements

Dependencies are inferred from the claim text.