Method and apparatus for spectroscopic comparison of compositions

• US 5,046,846 A
• Filed: 03/22/1989
• Issued: 09/10/1991
• Est. Priority Date: 03/22/1989
• Status: Expired due to Fees

• Current Assignee: The Lubrizol Corporation (Berkshire Hathaway Incorporated)
• Original Assignee: The Lubrizol Corporation (Berkshire Hathaway Incorporated)
• Inventors: Ray, James C., Asari, Logan R.
• Primary Examiner(s): McGraw, Vincent P.
• Application Number:
  US07/327,069
• Time in Patent Office: 902 Days
• Field of Search: 356/326, 356/328, 250/339, 364/498
• US Class Current: 356/326
• CPC Class Codes: G01N 2021/3595 using FTIR
• CPC Class Codes: G01N 21/35 using infrared light G01N21/39 takes precedence
• CPC Class Codes: G01N 21/3577 for analysing liquids, e.g. polluted water

Abstract

A method of analysis of and an apparatus for analysing a sample comprising the steps of: (A) examining at least one characteristic of a sample as a function of a variable parameter to obtain data, (B) normalizing the data with respect to a model, (C) making a comparison of the normalized data with reference data, and (D) determining from the comparison the quality of the sample. As a function of such determination of quality, control functions may be carried out with respect to a sample that was analyzed, such as a blend sample in a chemical processing plant.

Claims

1. A method of analysis, comprising the steps of:

(A) examining at least one characteristic of a sample as a function of a variable parameter to obtain data,

(B) normalizing the sample data with respect to a model by;

(1) establishing a plurality of data pairs which are characteristic values for the sample and a reference, both established at a single value of the variable parameter;

(2) performing a regression analysis with the data pairs to obtain normalization parameters by;

(a) performing a first regression, using all data pairs, to obtain a set of regression parameters;

(b) determining acceptability of the regression parameters by comparing variability of the data pairs around the model to predetermined variability limits;

(c) if the regression parameters are not acceptable, excluding data pairs which do not fall within a predetermined statistical limit around the model, and performing a further regression using a diminished set of data pairs to obtain further regression parameters; and

(d) repeating steps (b) and (c) in an iterative manner until acceptable regression parameters are obtained, or until a predetermined number of iterations are completed, whichever first occurs; and

(3) using the normalization parameters to normalize characteristic sample values to reference values,

(C) making a comparison of the normalized sample data with reference data, and

(D) determining from the comparison quality of the sample.

2. The method as claimed in claim 1, wherein the characteristic in step (A) is radiation transmission, the variable parameter is wavelength and the data are absorbance values.

3. The method of claim 1, where the model in step (B) approximates a straight line.

4. The method of claim 1, where in step (C) the comparison is subtraction.

5. The method of claim 4, where in step (C) the reference data are the average absorbance values in the database.

6. The method of claim 4, where in step (C) the reference data are the absorbance values for a reference sample.

7. The method of claim 1, where in step (C) the comparison is a residual value for the normalized data and the reference data.

8. The method of claim 7, where in step (D) the determining is accomplished by comparison of the residual value with the residual database value.

9. The method of claim 7, where the reference data are absorbance values for a reference sample.

10. The method of claim 1, where in step (C) the comparison is accomplished in regions.

11. The method of claim 10, where, the regions are from about to 1 to about 200 wavenumbers long.

12. The method of claim 11, wherein the regions are from about 5 to about 100 wavenumbers long.

13. The method of claim 2, wherein the wavelength is in a range from about 4000 to about 400 wavenumbers.

14. The method as claimed in claim 13, wherein the range is from about 1900 to about 600 wavenumbers.

15. The method of claim 1, wherein the regression analysis is linear regression analysis.

16. The method of claim 1, where in step (D), the determining is accomplished by comparing the comparison in step (C) with acceptance limits.

17. The method of claim 16, where the acceptance limits are from 1 to about 6 times standard deviation of the reference data.

18. The method of claim 17, wherein the acceptance limits are from about 2 to about 4 standard deviations of the reference data.

19. A method of analysis, comprising the steps of:

(A) examining the radiation transmission of the sample as a function of wavelength to obtain absorbance data,

(B) normalizing the sample absorbance data with respect to a model, by;

(1) establishing a plurality of data pairs which are characteristic absorbance values for the sample and a reference, both established at a single value of wavelength;

(2) performing a regression analysis with the data pairs to obtain normalization parameters by;

(a) performing a first regression, using all data pairs, to obtain a set of regression parameters;

(b) determining acceptability of the regression parameters by comparing variability of the data pairs around the model to predetermined variability limits;

(c) if the regression parameters are not acceptable, excluding data pairs which do not fall within a predetermined statistical limit around the model, and performing a further regression using a diminished set of data pairs to obtain further regression parameters; and

(d) repeating steps (b) and (c) in an iterative manner until acceptable regression parameters are obtained, or until a predetermined number of iterations are completed, whichever first occurs; and

(3) using the normalization parameters to normalize characteristic sample values to reference values;

(C) subtracting the noramlized sample data from reference data, and

(D) determining quality of the sample.

20. The method, as claimed in claim 19, wherein the wavelength is examined between about 4000 to about 400 wavenumbers.

21. The method as claimed in claim 20, wherein the wavelength is examined between about 1900 to about 600 wavenumbers.

22. The method as claimed in claim 19, wherein the reference data are average absorbance values in the database.

23. The method of claim 19, where in step (C) the reference data are the absorbance values for a reference sample.

24. The method of claim 19, where in step (C) the subtracting is accomplished in regions.

25. The method as claimed in claim 19, wherein the regions are from about 1 to about 200 wavenumbers long.

26. The method of claim 19, wherein the region are from about 5 to about 100 wavenumbers long.

27. The method of claim 19, where in step (D) the determining is accomplished by comparing or subtracting in step (C) where acceptance limits are from 1 to about 6 times standard deviations in the reference data.

28. The method of claim 19, wherein the acceptance limits are from about 2 to about 4 standard deviations of the reference data.

29. A method of analyzing a sample of a chemical mixture, comprising the steps of:

(A) examining infrared light absorption of the sample as a function of wavelength to obtain data,

(B) normalizing the sample data with respect to a model by;

(1) establishing a plurality of data pairs which are characteristic values for the sample and a reference, both established at a single value of wavelength;

(2) performing a regression analysis with the data pairs to obtain normalization parameters by;

(a) performing a first linear regression, using all data pairs, to obtain a set of regression parameters;

(b) determining acceptability of the regression parameters by comparing variability of the data pairs around the model to predetermined variability limits;

(c) if the regression parameters are not acceptable, exclusing data pairs which do not fall within a predetermined statistical limit around the model, and performing a further linear regression using a diminished set of data pairs to obtain further regression parameters; and

(d) repeating steps (b) and (c) in an iterative manner until acceptable regression parameters are obtained, or until a predetermined number of iterations are completed, whichever first occurs; and

(3) using the normalization parameters to normalize characteristic sample values to reference values,

(C) making a comparison of the normalized sample data with reference data, and

(D) determining from the comparison quality of the sample.