Mineral identification using mineral definitions having compositional ranges

• US 9,091,635 B2
• Filed: 10/26/2012
• Issued: 07/28/2015
• Est. Priority Date: 10/26/2012
• Status: Active Grant

• Current Assignee: FEI Company (Thermo Fisher Scientific Incorporated)
• Original Assignee: FEI Company (Thermo Fisher Scientific Incorporated)
• Inventors: Owen, Michael James
• Primary Examiner(s): MASKELL, MICHAEL P
• Application Number:
  US13/661,826
• Publication Number:
  US 20140117234A1
• Time in Patent Office: 1,005 Days
• Field of Search: 250/306, 250/307, 250/310, 250/311, 250/281, 250/282, 378/45, 378/46, 378/48, 702/28, 702/32, 850/10
• US Class Current: 1/1
• CPC Class Codes: G01N 2223/616 earth materials
• CPC Class Codes: G01N 23/2252 Measuring emitted X-rays, e.g. electron probe microanalysis [EPMA]

Abstract

Determining the composition of a mineral sample entails comparing measurements of a sample to mineral definitions and determining a similarity metric. The mineral definition includes a subspace of compositional values defined by end members. The similarity metric is related to a projection of the measured data point onto the subspace or onto an extension of the subspace.

Claims

1. A method of spectrum data analysis, the method comprising

acquiring a spectrum of an unknown mineral sample by generating x-rays from the unknown mineral sample and detecting the x-rays with a x-ray detector;

decomposing the spectrum of the unknown mineral sample to yield a compositional value;

determining a similarity metric between the unknown mineral sample and a mineral definition, the mineral definition providing a range for at least one element in th definition and a variation measure and including a variable mineral composition defined by a subspace with end members, the similarity metric being determined using a sample variation determined by a variation in the mineral definition and the number of x-ray counts in the unknown spectrum; and

if the computed value is within a user-defined threshold distance to an extrapolation of the subspace, extending the subspace to include a computed value outside bounds of the subspace.

2. The method of claim 1 in which determining a similarity metric includes determining individual element probability matches of the compositional value of the unknown mineral sample to the mineral definition.

3. The method of claim 1 in which the mineral definition includes multiple disconnected element series.

4. The method of claim 1, in which determining a closest point in the mineral definition to the compositional value includes projecting the compositional value onto the subspace such that a distance between the compositional value and the closest point in the mineral definition is minimized.

5. The method of claim 1 in which the end members are either determined experimentally or theoretically.

7. The method of claim 1, in which the mineral definition is created by measuring known samples of a mineral multiple times to create multiple mineral definitions, the multiple mineral definitions comprising variable mineral compositions.

8. The method of claim 7 in which determining a closest point in the mineral definition to the compositional value is repeated for each of the multiple mineral definitions and the mineral definition with a highest overall probability match is displayed.

9. The method of claim 1 further comprising displaying an overall probability match and identifying the unknown mineral sample.

11. A system for spectrum data analysis, comprising:

an electron beam column for directing an electron beam toward a sample;

an x-ray spectrometer for measuring energies of x-rays emitted from the sample in response to the electron beam;

a processor for executing computer instructions to analyze a mineral sample; and

a computer memory for storing instructions for carrying out the method of claim 1.

18. The method of claim 1 wherein generating x-rays from the unknown mineral sample comprises stimulating an emission of the x-rays from the unknown mineral sample by focusing an electron beam onto the unknown mineral sample using an electron microscope, and

further comprising counting the x-rays detected by the x-ray detector.

6. A method of spectrum analysis, the method comprising:

acquiring a spectrum of an unknown mineral sample by generating x-rays from the unknown mineral sample and detecting the x-rays with a x-ray detector;

decomposing the spectrum of the unknown mineral sample to yield a compositional Value;

determining a similarity metric between the unknown mineral sample and a mineral definition, the mineral definition providing a range for at least one element in the definition and a variation measure, the similarity metric being determined using a sample variation determined by a variation in the mineral definition and the number of x-ray counts in the unknown spectrum as well as the distance from the unknown mineral sample to the closest point in a subspace representing a mineral definition,

in which the closest point of the subspace representing the mineral definition to the compositional value is determined using a least-squares method to find a minimum distance between the compositional value and the closest point in the mineral definition.

10. A method of spectrum data analysis, the method comprising:

acquiring a spectrum of an unknown mineral sample by generating x-rays from the unknown mineral sample and detecting the x-rays with a x-ray detector;

decomposing the spectrum of the unknown mineral sample to yield a compositional value; and

determining a similar metric between the unknown mineral sample and a mineral definition, the mineral definition providing a range for at least one element in the definition and a variation measure, the similarity metric being determined using a sample variation determined by a variation in the mineral definition and the number of x-ray counts in the unknown spectrum

in which the mineral definition includes a variable mineral composition defined by a subspace with end members and

in which the subspace is a plane, such that the mineral definition can comprise two or more variable elemental compositions.

12. A method of spectrum data analysis, the method comprising:

acquiring a spectrum of an unknown mineral sample by generating x-rays from the unknown mineral sample, detecting the x-rays using an x-ray detector, and counting the x-rays detected by the x-ray detector;

fitting elemental spectra from a mineral definition to the spectrum of the unknown mineral sample to calculate elemental concentrations of the unknown mineral sample; and

determining from the calculated elemental concentrations a similarity metric between the unknown mineral sample and the mineral definition, the mineral definition including a range of elemental compositions between known end points and a variation measure, the similarity metric being using a sample variation determined by a variation in the mineral definition and the number of x-ray counts in the unknown spectrum, in which the range of elemental compositions between known end points is extended if the calculated elemental concentrations of the unknown mineral sample lies within a certain user-defined threshold of an extrapolation of the range of elemental compositions between known end points.

13. The method of claim 12, further comprising repeatedly determining from the calculated elemental concentrations similarity metrics, each between the unknown mineral sample and one of additional mineral definitions, and identifying the unknown mineral sample with the mineral definition having the closest similarity metric.

14. The method of claim 12, in which determining a similarity metric for the unknown mineral sample to the mineral definition from the calculated elemental concentrations includes determining a closest point in the mineral definition by projecting the closest point onto the range of compositions between known end points.

15. The method of claim 14, in which the closest point in the mineral definition is used to calculate individual element probability matches of the elemental concentrations of the unknown sample to the elemental spectra of the mineral definition, the individual element probability matches being used to calculate the similarity metric for the unknown sample to the mineral definition.

16. The method of claim 12, in which the known end points are calculated from experimental or theoretical end values.

17. The method of claim 12 in which the mineral definition includes a list of elements and a range of relative quantity of at least one of the elements.