Methods for characterizing, classifying, and identifying unknowns in samples

• US 6,408,250 B1
• Filed: 08/10/1999
• Issued: 06/18/2002
• Est. Priority Date: 08/10/1999
• Status: Expired due to Term

• Current Assignee: Battelle Memorial Institute, Eigenvector Research Incorporated
• Original Assignee: Battelle Memorial Institute
• Inventors: Grate, Jay W., Wise, Barry M.
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): MILLER, CRAIG S
• Application Number:
  US09/372,641
• Time in Patent Office: 1,043 Days
• Field of Search: 702/22-24, 702/27, 702/30, 73/232, 73/232.1, 73/233.6, 73/233.5
• US Class Current: 702/30
• CPC Class Codes: B82Y 30/00 Nanotechnology for materials or surface science, e.g. nanocomposites
• CPC Class Codes: G01N 2291/0256 Adsorption, desorption, surface mass change, e.g. on biosensors
• CPC Class Codes: G01N 2291/0422 Shear waves, transverse waves, horizontally polarised waves
• CPC Class Codes: G01N 2291/0423 Surface waves, e.g. Rayleigh waves, Love waves
• CPC Class Codes: G01N 2291/0426 Bulk waves, e.g. quartz crystal microbalance, torsional waves
• CPC Class Codes: G01N 2291/0427 Flexural waves, plate waves, e.g. Lamb waves, tuning fork, cantilever
• CPC Class Codes: G01N 2291/106 one or more transducer arrays
• CPC Class Codes: G01N 29/022 Fluid sensors based on microsensors, e.g. quartz crystal-microbalance [QCM], surface acoustic wave [SAW] devices, tuning forks, cantilevers, flexural plate wave [FPW] devices microdevices per se B81B
• CPC Class Codes: G01N 33/0031 comprising two or more sensors, e.g. a sensor array

Abstract

Disclosed is a method for taking the data generated from an array of responses from a multichannel instrument, and determining the characteristics of a chemical in the sample without the necessity of calibrating or training the instrument with known samples containing the same chemical. The characteristics determined by the method are then used to classify and identify the chemical in the sample. The method can also be used to quantify the concentration of the chemical in the sample.

Claims

1. A method for characterizing an unknown sample, wherein said sample is modeled with a plurality of descriptors, comprising the steps of:

a. obtaining a plurality of responses from a multichannel instrument, said plurality of responses equal to or greater than said plurality of descriptors, wherein said plurality of responses is related to each of said plurality of descriptors; and

b. determining said plurality of descriptors from said plurality of responses.

2. The method of claim 1 wherein said plurality of descriptors are selected from the group comprising molecular interaction characteristics of said unknown sample, molecular properties of said unknown sample, molecular structural features of said sample, and combinations thereof.

3. The method of claim 1 wherein said plurality of descriptors are related to a plurality of solubility properties of said samples.

4. The method of claim 1 wherein said plurality of descriptors are vapor solvation parameters.

5. The method of claim 1 wherein said plurality of descriptors are parameters in a linear free energy relationship.

6. The method of claim 1 wherein said plurality of descriptors are parameters in a linear solvation energy relationship.

7. The method of claim 1 wherein said plurality of descriptors are descriptors in a quantitative structure activity relationship.

8. The method of claim 1 wherein said plurality of descriptors are parameters in a principle components equation.

9. The method of claim 1 wherein the response of each channel of said multichannel instrument can be modeled by an equation including a term that is related to said plurality of descriptors.

10. The method of claim 1 wherein the plurality of responses of said multichannel instrument are related to the thermodynamic partitioning of said unknown sample between phases.

11. The method of claim 1 wherein the plurality of responses of said multichannel instrument are related to the partitioning of said unknown sample between the ambient environment and a plurality of sorbent phases.

12. The method of claim 1 wherein said multichannel instrument utilizes a plurality of gas chromatographic columns.

13. The method of claim 1 wherein said multichannel instrument utilizes a plurality of sensors having sorbent phases.

14. The method of claim 13 wherein the sorbent phase is selected from the group comprising a solid surface, a self assembled monolayer, a molecular multilayer, an amorphous solid phase, a liquid, a membrane and a thin film.

15. The method of claim 13 wherein the sorbent phase is a stationary sorbent phase.

16. The method of claim 13 wherein the sorbent phase is a polymer.

17. The method of claim 1 wherein said multichannel instrument utilizes a plurality of acoustic wave sensors selected from the group comprising thickness shear mode devices, surface acoustic wave devices, Leaky surface acoustic wave devices, surface transverse wave devices, Love wave devices, shear-horizontal acoustic plate mode devices, flexural plate wave devices, thin film resonators, and thin rod flexural devices.

18. The method of claim 1 wherein said multichannel instrument utilizes a plurality of acoustic wave sensors coated with polymers and stationary phases.

19. The method of claim 1 wherein said multichannel instrument utilizes a plurality of optical sensors.

20. The method of claim 1 wherein said multichannel instrument utilizes a plurality of chemiresistor sensors.

21. The method of claim 1 wherein said multichannel instrument utilizes a plurality of chemiresitor sensors having a sorbent layer phase and a solid electronic conductor.

22. The method of claim 1 wherein said multichannel instrument utilizes a plurality of sensors selected from the group comprising electrochemical and field effect transistor sensors.

23. The method of claim 1 wherein said multichannel instrument utilizes a plurality of sensors selected from the group comprising microbeam, microbar and microcantilever sensors.

24. A method for characterizing an unknown sample, wherein said sample is modeled with a plurality of descriptors, comprising the steps of:

a. obtaining a plurality of responses from a multichannel instrument, said plurality of responses equal to or greater than said plurality of descriptors, wherein the response from each channel of said multichannel instrument includes a term related to said plurality of descriptors, wherein said term related to said plurality of descriptors contains coefficients for each descriptor; and

b. determining said plurality of descriptors from said plurality of responses.

25. The method of claim 24 wherein said coefficients are coefficients in a linear free energy relationship.

26. The method of claim 24 wherein said coefficients are coefficients in a linear solvation energy relationship.

27. The method of claim 24 wherein said coefficients are coefficients in a quantitative structure activity relationship.

28. The method of claim 24 wherein said coefficients are coefficients in a principle components equation.

29. The method of claim 24 wherein said coefficients are coefficients in a linear free energy relationship related to sorbent phase properties.

30. The method of claim 24 wherein said coefficients are coefficients in a linear free energy relationship based on thermodynamic partition coefficients.

31. The method of claim 24 wherein said coefficients are determined from instrument responses to known compounds.

32. A method for characterizing an unknown sample, wherein said sample is modeled with a plurality of descriptors, comprising the steps of:

a. obtaining a plurality of responses from a multichannel instrument, said plurality of responses equal to or greater than said plurality of descriptors, wherein the response from each channel of said multichannel instrument includes a term related to said plurality of descriptors, wherein said term related to said plurality of descriptors contains coefficients for each descriptor;

b. Defining a matrix P containing said coefficients;

c. determining said plurality of descriptors from said plurality of responses and said matrix P.

33. The method of claim 32 wherein

a. the response is from said multichannel instrument is included in matrix R where R is equal to C 10^(VP+1c)M^{− 1} N

b. said descriptors are determined from matrix V, where V is related to a term of the form {log (C⁻

1 R M N<sup>−

1</sup>)−

1c}P^T(PP^T)<sup>−

1</sup>;

c. C is a diagonal matrix of the concentrations of the vapors

d. M and N are diagonal matrices of particular properties of specific channels of the detector,

e. the superscript of −

1 denotes the inverse of the matrix,

f. N is a diagonal matrix of the Δ

f_s values of the sensors,

g. c is a vector of constants,

h. P^T is the transpose of matrix P,

j. P^T(PP^T)<sup>−

1</sup> is the pseudo-inverse of P.

34. The method of claim 32 wherein

a. the response is matrix R where R is equal to C 10^(VP+1c) D⁻

1 F;

b. said descriptors are determined from matrix V, where V is equal to {log(C<sup>−

1</sup> R D F<sup>−

1</sup>)−

1c}P^T(PP^T)<sup>−

1</sup>;

c. C is a diagonal matrix of the concentrations of the vapors

d. D is a diagonal matrix of the polymer densities,

e. the superscript of −

1 denotes the inverse of the matrix,

f. F is a diagonal matrix of the Δ

f_s values of the sensors,

g. c is a vector of constants,

h. P^T is the transpose of matrix P, and

i. P^T(PP^T)<sup>−

1</sup> is the pseudo-inverse of P.

35. The method of claim 32 wherein matrix P contains LSER coefficients determined from measurements of thermodynamic partitioning.

36. The method of claim 32 wherein matrix V contains solvation parameters for vapors.

37. The method of claim 32 wherein matrix R contains responses of acoustic wave vapor sensors with sorbent interactor layers and wherein matrix P contains LSER coefficients determined from measurements of responses of acoustic wave vapor sensors to known vapors.

38. The method of claim 32 further comprising the step of utilizing one or more of said descriptors to classify said unknown sample as belonging to a class of chemicals with certain properties.

39. The method of claim 32 further comprising the step of utilizing one or more of said descriptors to classify said unknown sample as belonging to a class of chemicals with certain structural features.

40. The method of claim 32 further comprising the step of comparing said descriptors to a table of descriptors of known chemicals to determine the identity of said unknown sample.

41. A method for characterizing an unknown sample at an unknown concentration, wherein said sample is modeled with a plurality of descriptors, comprising the steps of:

a. obtaining a plurality of responses from a multichannel instrument, said plurality of responses equal to or greater than said plurality of descriptors, wherein the response from each channel of said multichannel instrument includes a term related to said plurality of descriptors, wherein said term related to said plurality of descriptors contains coefficients for each descriptor;

b. defining a matrix P_a containing said coefficients and augmented by a vector of ones,

c. determining said plurality of descriptors and concentration where from said plurality of responses

d. wherein the response is matrix R where R is equal to 10^(V_aP_a1c)D<sup>−

1</sup> F;

e. said descriptors and concentration are determined from matrix Va, where Va is equal to {log(R D F<sup>−

1</sup>)−

1c}P_a^T(P_aP_a^T)<sup>−

1</sup>

f. P_a is defined as the matrix P augmented by a vector of ones as given in $P_a=\left[\begin{array}{c}P\\ 1\end{array}\right],$

where P is a matrix containing said coefficients

g. C is a diagonal matrix of the concentrations of the vapors

h. D is a diagonal matrix of the polymer densities,

i. the superscript of −

1 denotes the inverse of the matrix

j. F is a diagonal matrix of the Δ

f_s values of the sensors

k. P_a^T is the transpose of P_a

l. P_a^T(P_aP_a^T)<sup>−

1</sup> is the pseudoinverse of P_a.

42. The method of claim 41 wherein matrix P_a contains LSER coefficients determined from measurements of thermodynamic partitioning.

43. The method of claim 41 wherein matrix V contains solvation parameters for vapors.

44. The method of claim 41 wherein matrix R contains responses of acoustic wave vapor sensors with sorbentinteractor layers.

45. The method of claim 41 wherein matrix P_a contains LSER coefficients determined from measurements of responses of acoustic wave vapor sensors to known vapors.

46. The method of claim 41 further comprising the step of utilizing one or more of said descriptors to classify said unknown sample as belonging to a class of chemicals with certain properties.

47. The method of claim 41 further comprising the step of utilizing one or more of said descriptors to classify said unknown sample as belonging to a class of chemicals with certain structural features.

48. The method of claim 41 further comprising the step of comparing said descriptors to a table of descriptors of known chemicals to determine the identity of said unknown sample.

49. A method for characterizing an unknown sample, wherein said sample is modeled with a plurality of descriptors, comprising the steps of:

a. obtaining a plurality of responses from a multichannel instrument, said plurality of responses equal to or greater than said plurality of descriptors, wherein said plurality of responses is related to each of said plurality of descriptors; and

b. determining one or more of said plurality of descriptors from said plurality of responses using the method of inverse least squares to perform a regression, where an individual descriptor, y, is modeled as a weighted sum of responses according to y=Xb, where X is the measured response and b is a vector of weights, generally determined by regression b=X+y.

50. The method of claim 49 wherein the method of regression is selected from the methods including multiple linear regression, partial least squares, and principle components regression.

51. The method of claim 49 wherein b, the vector of weight for determination of each descriptor, is determined by a regression using responses to known compounds.

52. The method of claim 51 wherein b is used to determine descriptors from the instrument response to unknowns that were not among said known compounds.