Method for improving instrument response

• US 6,061,641 A
• Filed: 10/23/1997
• Issued: 05/09/2000
• Est. Priority Date: 10/25/1996
• Status: Expired due to Term

• Current Assignee: National Technology & Engineering Solutions Of Sandia LLC (Honeywell International Inc.)
• Original Assignee: Sandia Corporation (Martin Marietta Corporation)
• Inventors: Hencken, Kenneth R., Hahn, David W., Johnsen, Howard A., Flower, William L.
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): Bui, Bryan
• Application Number:
  US08/956,520
• Time in Patent Office: 929 Days
• Field of Search: 702/85, 702/69, 702/79, 702/191, 356/316, 356/318, 356/346
• US Class Current: 702/85
• CPC Class Codes: G01J 3/443 Emission spectrometry

Abstract

This invention pertains generally to a method for improving the accuracy of particle analysis under conditions of discrete particle loading and particularly to a method for improving signal-to-noise ratio and instrument response in laser spark spectroscopic analysis of particulate emissions. Under conditions of low particle density loading (particles/m³) resulting from low overall metal concentrations and/or large particle size uniform sampling can not be guaranteed. The present invention discloses a technique for separating laser sparks that arise from sample particles from those that do not; that is, a process for systematically "gating" the instrument response arising from "sampled" particles from those responses which do not, is dislosed as a solution to his problem. The disclosed approach is based on random sampling combined with a conditional analysis of each pulse. A threshold value is determined for the ratio of the intensity of a spectral line for a given element to a baseline region. If the threshold value is exceeded, the pulse is classified as a "hit" and that data is collected and an average spectrum is generated from an arithmetic average of "hits". The true metal concentration is determined from the averaged spectrum.

Claims

1. A method for improving the response discrimination of a signal detection instrument, the method comprising the steps of:

a.) detecting and dispersing a random signal, said dispersed signal comprising an energy emission spectrum, each said spectrum further comprising a background signal intensity and a one or more discrete spectral lines, said lines consisting of sharp peaks in said background signal intensity, each said peak having an absolute signal intensity greater than said background signal intensity;

b.) electronically digitizing and storing each said spectrum;

c.) repeating steps a.) and b.) until a predetermined number of spectra have been stored;

d.) computing a time-averaged signal spectrum comprising an arithmetic average of said plurality of stored spectra;

e.) providing a known spectrum for comparison, said known spectrum having a known spectral line;

f.) selecting an analysis region common to both said plurality of stored spectra and to said known spectra, said region chosen to contain said known spectral line;

g.) computing a signal threshold value in said analysis region of said time-averaged spectrum; and

h.) performing a conditional analysis in said analysis region in each of said stored spectra, said analysis for preferentially selecting those said spectra having a signal intensity value which exceeds said signal threshold value.

2. The method of claim 1 wherein the step of computing a signal threshold value further comprises the steps of:

a.) computing a signal response ratio in said analysis region, said ratio consisting of a first average signal intensity, measured across a first region of said time-averaged spectrum, divided by a second average signal intensity, measured across a second region of said time-averaged spectrum, said first region centered over said known spectral line, said second region selected away from said known spectral line so as not to contain said first region; and

b.) multiplying said signal response ratio by a factor of at least about 1.25, thereby providing said signal threshold value.

3. The method of claim 1 wherein the step of performing a conditional analysis further comprises the steps of:

a.) computing a signal response ratio for each of said stored spectra in said analysis region;

b.) comparing said signal response ratio with said threshold value;

c.) summing all said stored spectra exhibiting a signal response ratio greater than said threshold value and discarding all other said stored spectra;

d.) repeating the steps a.) through c.) until all said stored spectra have been analyzed; and

e.) computing an arithmetic average spectrum of said summed spectra.

4. The method of claim 1 wherein said predetermined number is about between 100 and 500.

5. The method of claim 1 wherein said random signal is measured as either a frequency or a wavelength.

6. The method of claim 1 wherein the step of detecting comprises a spectrometer, said spectrometer having means for signal dispersion, said spectrometer further having a detector means comprising a linear diode array, said array comprising a plurality of detecting elements.

7. The method of claim 6 wherein said means for signal dispersion comprises a grating, said grating capable of dispersing said random signal sufficiently to resolve at least one said spectral line.

8. A method for improving the response discrimination of a signal detection instrument, the method comprising the steps of:

a.) detecting and dispersing a random signal, said dispersed signal comprising an energy emission spectrum, said spectrum further comprising a background signal intensity and a plurality of discrete spectral lines, said lines consisting of sharp peaks in background signal intensity, each said peak having an absolute signal intensity greater than said background signal intensity;

b.) electronically digitizing and said spectrum;

c.) zeroing a first and a second electronic register;

d.) electronically adding said spectra to said contents of said first electronic register;

e.) repeating said steps a.), b.) and d.) until reaching a first predetermined number of spectra have been summed;

f.) computing and electronically storing a time-averaged signal spectrum, said time-average spectrum comprising an arithmetic average of said predetermined number of summed spectra;

g.) providing a known spectrum for comparison, said known spectrum having a known spectral line;

h.) selecting an analysis region common to both said time-averaged spectrum and said known spectrum, said analysis region chosen to contain said known spectral line;

i.) computing a signal threshold value in said analysis region of said time-averaged spectrum;

j.) repeating said steps a.) and b.);

k.) performing a conditional analysis in said analysis region of said spectra, wherein the step of performing a conditional analysis further comprises the steps of;

i.) computing said signal response ratio for said stored spectra;

ii.) comparing said signal response ratio with said threshold value;

iii.) selecting those said spectra having a signal intensity value which exceeds said signal threshold value;

l.) electronically adding said selected spectra to said contents of said second electronic register;

m.) repeating steps j.) through l.) until a second predetermined number of spectra have been analyzed; and

n.) computing an arithmetic average spectrum of said summed selected spectra.

9. The method of claim 8 wherein the step of computing a signal threshold value further comprises the steps of:

a.) computing a signal response ratio in said analysis region, said ratio consisting of a first average signal intensity, measured across a first region of said time-averaged spectrum, divided by a second average signal intensity, measured across a second region of said time-averaged spectrum, said first region centered over said known spectral line, said second region selected away from said known spectral line so as not to contain said first region; and

b.) multiplying said signal response ratio by a factor of at least about 1.25, thereby providing said signal threshold value.

10. The method of claim 8 wherein said first predetermined number is about between 100 and 500.

11. The method of claim 8 wherein said random signal is measured as either a frequency or a wavelength.

12. The method of claim 8 wherein the step of detecting comprises a spectrometer, said spectrometer having means for signal dispersion, said spectrometer further having a detector means comprising a linear diode array, said array comprising a plurality of detecting elements.

13. The method of claim 12 wherein said means for signal dispersion comprises a grating, said grating chosen to dispersing said random signal sufficiently to resolve at least one of said spectral lines.