JP2011038939A - Spectrum classification method and apparatus of energy-dispersive x-ray spectroscope - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a spectrum classification method and an apparatus of an energy-dispersive X-ray spectroscope which does not require the correlation between spectra and is capable of high-speed processing. <P>SOLUTION: The spectrum classification apparatus of an energy-dispersive X-ray spectroscope is constituted of: a spectrum acquisition means for irradiating a sample with a charged particle beam, detecting X rays generated from the sample for every prescribe section, and determining an X-ray spectrum which indicates energy in a horizontal axis and the number of generated X rays in a vertical axis on the basis of detected X rays; a numeral value computation means for dividing the spectrum acquired from the spectrum acquisition means by dividing the horizontal axis by M and the vertical axis by N, determining a numerical value in the vertical axis for every section of the horizontal axis divided by M, and acquiring a sequence of determined numerical values for every section; a coloring means for collecting up sections having the same sequence of numerical values and coloring, by the same color, sections determined as being of the same composition on the basis of the same sequence of numerical values; and a distribution map acquisition means for synthesizing an individual distribution determined as being of each composition to acquire a distribution map. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a spectrum classification method and apparatus for an energy dispersive X-ray analyzer, and more specifically, it is possible to identify a distribution in a substance unit such as a compound rather than a distribution state in an element unit such as an element map. The present invention relates to a spectrum classification method and apparatus for an energy dispersive X-ray analyzer.

  In a scanning electron microscope (SEM) and a transmission electron microscope (TEM), an energy dispersive X-ray analyzer (EDS) is attached to classify spectra based on X-rays. EDS performs an analysis (hereinafter referred to as an element map) to obtain an element distribution state. An element map image is analyzed using a statistical analysis technique, and a distribution is determined by obtaining a correlation between pixels. FIG. 19 shows a correlation diagram. For example, the horizontal axis represents the strength of the A element, and the vertical axis represents the strength of the B element. When the relationship between these A elements and B elements is plotted, a correlation diagram as shown in FIG. 19 is obtained.

  FIG. 20 is a diagram showing the phase distribution, and shows what kind of composition is distributed. The element map described above is obtained by drawing the X-ray intensity in a specific energy range (hereinafter referred to as ROI) based on the coordinate state of electron beam scanning. In recent EDS, in addition to this element map image, spectrum information at each coordinate is also measured simultaneously. Based on this spectrum information, the element map image is redrawn by changing the ROI. Spectral information of a specific region is extracted, or quantitative correction calculation is performed based on spectral information of each coordinate, and a distribution map is drawn (hereinafter referred to as a quantitative map) based on the result.

  19 and 20 show the element map image analyzed using a statistical analysis technique, and the distribution is determined by obtaining the correlation between the pixels. Alternatively, the spectrum generated at each pixel is analyzed using a statistical analysis technique, and the distribution is determined by obtaining the correlation between the spectra.

  As a conventional apparatus of this type, the spectrum database is emptied and measurement is started. The X-ray spectrum obtained by the measurement and the X-ray spectrum in the spectrum database are compared by the spectrum comparison means, and the results match. A technique is known in which an X-ray spectrum obtained by measurement is additionally registered in a database and specified measurement is repeated when no X-ray spectrum exists (see, for example, Patent Document 1).

  In addition, the method includes obtaining X-ray data indicating X-ray radiation characteristics to be monitored of a sample in response to an incident energy beam, and composition data of a plurality of materials, wherein the material of the sample is contained inside Obtaining a data set; calculating predicted X-ray data for each of the substances in the data set using the composition data; and comparing the obtained X-ray data and the predicted X-ray data. There is known a method for identifying a substance that includes a step of determining a possible feature of the substance of the sample based on the comparison (see, for example, Patent Document 2).

  In addition, the spectrum is measured by irradiating an unknown sample with infrared rays, the spectrum is compared with the spectrum of the known material, and the gradient of each spectrum or each first derivative is obtained for a number of wavelength sections of the spectrum. Each first derivative number is classified into one of a normalized multi-step numerical range, and each gradient class classified for multiple wavelength segments of the measured spectrum is pre-evaluated for known materials. A technique is known in which similarities and differences of individual gradient classes are obtained as gradient distances between gradient classes by comparing with respective gradient classes classified in a plurality of wavelength segments that are the same as the spectrum (see, for example, Patent Document 3). ).

JP 2002-365246 A (paragraphs 0016 to 0017) JP 2007-3532 A (paragraphs 0032 to 0045) JP-A-9-297062 (paragraphs 0019 to 0046)

  In the conventional apparatus, there is a problem that a trace amount component is easily overlooked in the analysis using the element map image. In addition, analysis based on a spectrum consumes a large amount of storage area and is difficult to process at high speed. The present invention has been made in view of such problems, and provides a spectrum classification method and apparatus for an energy dispersive X-ray analyzer that can perform high-speed processing without the need to obtain correlation between spectra. The purpose is to do.

In order to solve the above problem, the present invention has the following configuration.
(1) The invention according to claim 1 is an energy dispersive X-ray analyzer attached to a scanning electron microscope or a transmission electron microscope, and applies X-rays generated from the sample by irradiating the sample with a charged particle beam. From the detected X-rays, a spectrum representing energy on the horizontal axis and the number of X-rays on the vertical axis is obtained. The spectrum is divided into M on the horizontal axis and N on the vertical axis. The numerical value of the vertical axis is obtained for each M divided area of the axis, the obtained numerical sequence is obtained for each section, the sections of the same numerical sequence are grouped, and the sections determined to be the same composition by the same numerical sequence are the same color And a distribution map is obtained by synthesizing individual distributions determined as the respective compositions.

  (2) The invention according to claim 2 is an energy dispersive X-ray analyzer attached to a scanning electron microscope or a transmission electron microscope, wherein X-rays generated from the sample by irradiating the sample with a charged particle beam are predetermined. Spectrum acquisition means for obtaining a spectrum representing energy on the horizontal axis and the number of X-rays on the vertical axis from the detected X-rays, and the spectrum obtained from the spectrum acquisition means M on the horizontal axis. Dividing, dividing the vertical axis into N, obtaining the numerical value of the vertical axis for each M-divided area of the horizontal axis, and combining the numerical value sequence calculating means for obtaining the obtained numerical value sequence for each of the partitions, It is composed of coloring means for coloring the sections determined to be the same composition by the same numerical sequence with the same color, and a distribution chart obtaining means for obtaining a distribution chart by combining the individual distributions determined as the respective compositions. It is characterized by that.

(3) The invention described in claim 3 is characterized in that a specific divided area of the M divisions is further divided into a plurality of areas to obtain a numerical string.
(4) The invention described in claim 4 is characterized in that a numerical sequence is obtained by considering a plurality of specific areas in the M division as one area.

(5) The invention according to claim 5 is characterized in that, when the sample surface has an uneven shape, the low energy portion of the spectrum is excluded from the judgment target.
(6) In the invention according to claim 6, the waveform resolution FWHM of the energy dispersive X-ray analyzer is represented by a function of energy E and pulse processing time τ, and division is performed based on this function when dividing the horizontal axis region. It is characterized by varying the energy width of the.

  (1) According to the first aspect of the present invention, the obtained spectrum is divided into M on the horizontal axis and N on the vertical axis, and the numerical value on the vertical axis is obtained for each M-divided area on the horizontal axis. A column is obtained for each section, the sections of the same numerical sequence are grouped together, the sections determined to be the same composition by the same numerical sequence are colored with the same color, and individual distributions determined as each composition are synthesized Since the distribution map is obtained, it is not necessary to obtain the correlation between the spectra, and it is possible to provide a spectrum classification method for an energy dispersive X-ray analyzer that can be processed at high speed.

  (2) According to the invention described in claim 2, the obtained spectrum is obtained by dividing the obtained horizontal axis into M and dividing the vertical axis into N, and obtaining a numerical value on the vertical axis for each M division on the horizontal axis. Is obtained for each section, the sections of the same numerical sequence are grouped, the sections determined to be the same composition by the same numerical sequence are colored with the same color, and the individual distributions determined to be the respective compositions are combined and distributed. Since the figure is obtained, it is not necessary to obtain the correlation between the spectra, and it is possible to provide a spectrum classification device for an energy dispersive X-ray analyzer that can be processed at high speed.

(3) According to the invention described in claim 3, the feature of the shape can be expressed in the result by partially making the divided region of the horizontal axis fine.
(4) According to the invention described in claim 4, by making the plurality of divided regions on the horizontal axis into one region, it is possible to reduce the numerical sequence obtained without reducing the amount of information.

  (5) According to the invention described in claim 5, when it is detected that the sample surface is uneven, an accurate X-ray analysis can be performed by excluding the low energy portion from the judgment target. .

  (6) According to the invention described in claim 6, by calculating the Fano factor and the process time in advance, it is possible to set the optimum energy width of the divided region on the horizontal axis in each apparatus.

It is a figure which shows the structural example of a scanning electron microscope provided with the energy dispersive X-ray analyzer which concerns on this invention. It is a figure which shows the example of the observation image of a certain sample. It is a figure which shows the spectrum data observed from the whole range of an observation image. It is a figure which shows the state which divided | segmented the observation area | region into the division. It is a figure which shows the spectrum for every composition. It is a figure which shows the normalized spectrum data. It is a figure which shows the state which divided | segmented the vertical axis | shaft and the horizontal axis into 10 divisions, respectively. It is a figure which shows evaluation of the area a. It is a figure which shows evaluation of the area b. It is a figure which shows evaluation of the area c. It is a figure which shows the state which completed the evaluation about the area a to j. It is a figure which shows the method of the synthesis | combination of distribution of each composition. It is explanatory drawing of a variable division area. It is a figure which shows the discharge | release state of the X-ray according to the shape of the sample surface. It is a figure which shows the difference in the spectrum data by the difference in a sample surface. It is a figure which shows the example of the difference in resolution by energy (DELTA) Ed. It is a figure which shows the difference of the spectrum by process time. It is a figure which shows the example which set the division area of the horizontal axis to the unequal space | interval based on the function of FWHM. It is a figure which shows a correlation diagram. It is a figure which shows a phase distribution map.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a diagram showing a configuration example of a scanning electron microscope (SEM) provided with an energy dispersive X-ray analyzer (EDS) according to the present invention. In the figure, 1 is a scanning coil, 2 is an electron beam, and the scanning coil 1 scans the electron beam 2 in the X and Y directions. Reference numeral 6 denotes a sample, and an electron beam 2 is finely focused on the surface by the scanning coil 1 and irradiated and deflected.

  3 is an X-ray spectrometer that detects X-rays emitted from the sample 6 in accordance with the deflection operation, and 4 is an electron beam detector that similarly detects reflected electrons or secondary electrons emitted from the sample 6. A control device 5 receives the outputs of the X-ray detector 3 and the electron beam detector 4 and performs predetermined arithmetic processing. For example, a computer is used as the control device 5.

A monitor 7 receives the output of the electron beam detector 4 and displays an electron beam image, or receives the output of the X-ray spectrometer 3 and displays an X-ray spectrum. As the monitor 7, for example, a liquid crystal display (LCD) or a CRT is used. The scanning coil 1 receives a scanning signal from the control device 5. The present invention will be described in detail using the apparatus configured as described above. The main body of the operation is the control device 5.
Example 1
The control device 5 gives a scanning signal to the scanning coil 1. As a result, the electron beam 2 scans over the sample 6, and X-rays are emitted from the sample 6. This X-ray enters the control device 5 and is subjected to image processing to obtain spectral data. The obtained spectrum data is displayed on the monitor 7.

  FIG. 2 is a diagram showing an example of an observation image of a certain sample. This observation image is obtained by performing image processing on the signal detected by the electron beam detector 4 in FIG. In the figure, three types of images (A), (B), and (C) are obtained. This image is an image obtained by observing a sample made of three different compounds or simple substances. FIG. 3 is a diagram showing spectral data observed from the entire range of the observed image. The horizontal axis is energy, and the vertical axis is the X-ray count value. 2 corresponds to (A) in FIG. 2, B corresponds to (B) in FIG. 2, and C corresponds to (C) in FIG.

  Next, as shown in FIG. 4, the observation region is divided into fine sections, and the spectrum data observed in each section is extracted. In FIG. 4, since it is divided into 64 × 48 sections, 3072 pieces of spectral data are extracted. At this time, since the shape of the spectrum data depends on the composition, the shape of the spectrum data extracted from the respective sections included in the compositions (A), (B), and (C) in FIG. 2 is as shown in FIG. Suppose that it was a simple shape.

FIG. 5 is a diagram showing a spectrum for each composition. (A) is spectral data collected with the composition of (A), (b) is spectral data collected with the composition of (B), and (c) is spectral data collected with the composition of (C). Since the spectral data extracted from the compartments having the same composition are substantially similar, the shape of the spectral data is digitized by the following procedure.
A) The height of the extracted spectrum data is normalized with the maximum value. The normalized spectrum data is as shown in FIG. (A) is the spectrum of FIG. 5 (a) with normalized height, (b) is the spectrum of (b) of FIG. 5 with normalized height, and (c) is FIG. 5 with normalized height. (C) of the spectrum.
B) For the spectrum data with normalized height, the horizontal axis and the vertical axis are each divided into 10 areas as shown in FIG. Needless to say, the number of areas to be divided may be other than ten. In FIG. 7, the horizontal axis is divided into 10 areas a to j, and the vertical axis is divided into 10 areas 0 to 9.
C) For the area a, the maximum intensity reached by the spectrum data within the range is examined. FIG. 8 is a diagram showing the evaluation of the area a. As is apparent from the figure, the maximum intensity is “0”.
D) Next, for the area b, the maximum intensity that the spectrum data has reached within the range is examined. FIG. 9 is a diagram showing the evaluation of the area b. As is apparent from the figure, the maximum intensity is “0”.
E) Next, for the area c, the maximum intensity that the spectrum data has reached within the range is examined. FIG. 10 is a diagram showing the evaluation of the area c. As is apparent from the figure, the maximum intensity is “7”.
F) When such an operation is performed from zone a to zone j, the result is as shown in FIG. Then, a numerical value “0079111100” in which the determined area numbers on the vertical axis are arranged is obtained as an ID representing the shape of this spectrum.

  When the procedures A) to F) are performed on the spectrum data for the above-mentioned 3072 sections, the ID is determined for each spectrum data. This ID represents the shape of the spectrum data. Since the shape depends on the composition, the spectrum data generated from the same composition has the same ID.

By having the same ID, sections determined to have the same composition are colored with the same color.
A) The results of coloring the sections determined to be the compositions (A), (B), and (C) with different colors are (a) to (c) in FIG. FIG. 12 is a diagram showing how to synthesize the distribution of each composition.
B) When the individual distributions (a) to (c) are synthesized, a distribution diagram (d) which is a result to be obtained is obtained.

As described above, according to this embodiment, the obtained spectrum is divided into M on the horizontal axis and N on the vertical axis, and the numerical value on the vertical axis is obtained for each M division on the horizontal axis. Collecting the sections of the same numerical sequence, collecting the sections determined to be the same composition by the same numerical sequence with the same color, and combining the individual distributions determined to be the respective compositions to obtain a distribution map Therefore, it is not necessary to obtain the correlation between spectra, and it is possible to provide a spectrum classification method and apparatus for an energy dispersive X-ray analyzer that can be processed at high speed.
(Example 2)
In the case of spectral data in which the peak positions are concentrated in a narrow range, the feature of the shape does not appear in the result because the equally divided area is too wide for the peak interval. In such a case, the divided region on the horizontal axis is partially made finer automatically or by user instruction. Regardless of the size of the divided areas, the spectrum shape is digitized by assigning a single digit value to each area.

  FIG. 13 is an explanatory diagram of the variable division region. In this characteristic, the area C is further divided into 10 areas (c0, c1, c2,..., C9). And the numerical value of spectrum data is calculated | required for every 10 divisions in the area c. As a result, a total of 19 digits is obtained as “00000099635354000000”. In addition to obtaining detailed information in a small divided region for a region with a large amount of information, it is conceivable to simplify without dividing a region with a small amount of information.

  For example, if the entire spectrum is examined and no characteristic peak is found in the areas ej, the area is regarded as one area without being divided. Then, it is only necessary to consider the five areas of areas a, b, c, d, and ej, and the result in the case of FIG. 13 is “00940” and a 5-digit numeric string. As a result, the obtained numerical sequence can be shortened without reducing the amount of information.

It is also conceivable to combine the subdivision of the divided areas and the process for grouping the divided areas. That is, the subdivision of the area c and the grouping process of the areas e to j are combined into numerical values. When quantifying in this way, the areas to be quantified are considered as 14 areas a, b, c0 to c9, d, ej,
A numerical string “00000000043540” can be obtained.
(Example 3)
FIG. 14 is a diagram showing an X-ray emission state corresponding to the shape of the sample surface. As shown in FIG. 2B, when the sample surface is uneven, the distance that X-rays generated inside the sample must pass through the sample structure is longer than that of a flat sample as shown in FIG. Therefore, X-rays are absorbed by the sample itself, and the detected X-ray dose decreases. Since this is more noticeable with low energy X-rays with low transmission capability, the observed spectral data has a shape in which the low energy portion is depressed as shown in FIG.

  In such a case, even if the spectrum data is observed from the same composition, the shape cannot be recognized as a similar shape, and the composition is determined to be different. Therefore, when it is detected that the sample surface is uneven, it may be possible to exclude the low energy portion from the determination target. For example, it is conceivable that low energy (for example, 1 keV or less) is automatically deleted from the divided region by monitoring the shape of the background. In the example of FIG. 15, it can be determined that the composition is the same if the area a is excluded and only the areas b to j are determined.

According to this embodiment, when it is detected that the sample surface is uneven, an accurate X-ray analysis can be performed by excluding the low energy portion from the determination target.
Example 4
Since the waveform resolution (FWHM) of EDS can be expressed by a function (FWHM = F (E, 1 / τ)) of energy (E) and pulse processing time (τ), the region of the second embodiment is divided. In this case, it is also possible to change this function.

The function (FWHM = F (E, 1 / τ)) can be expressed by the following equation. The detector-specific resolution ΔEd is expressed by the following equation.
ΔEd = 2.355√ε · F · E
E: Incident X-ray energy (eV) to the detector
ε: Average energy (eV) required to generate a pair of electron / hole pairs
F: Fano factor (a substance-fixing constant that limits the fluctuation of the amount of charge generated inside the detector)
That is, it can be expressed by a function (√E) proportional to the square root of the detected X-ray energy (E). Furthermore, the resolution obtained by EDS takes into account ΔEp associated with electrical processing. This is the resolution of the charge collection type preamplifier and pulse processor that converts the charge generated from the detector into a voltage signal (the analog signal output from the preamplifier is finally converted into a digital signal), and mainly the process time. It is expressed by a function depending on (pulse processing time) (τ: time for processing one X-ray signal). If the process time is increased, the S / N ratio is improved, and the resolution ΔEp is decreased. Finally, the waveform resolution of EDS is
FWHM = √ (ΔEd + ΔEp ² )

  FIG. 16 is a diagram showing an example of the difference in resolution depending on the energy ΔEd. (A) shows the case of Al-K, and (b) shows the case of Cu-K. FIG. 17 is a diagram showing the difference in spectrum depending on the process time. (A) shows the case where the process time τ is τ1, and (b) shows the case where the process time τ is τ4. Here, τ4 is collected with a process time of about 64 times τ1.

  In this way, the divided region on the horizontal axis when the shape of the spectrum data is digitized is variably set based on the detected X-ray energy (E) and the process time (τ). Since the Fano factor (F) and process time (τ) vary depending on the type of detector and individual differences, it is possible to set an optimal horizontal axis division region for each device by calculating it in advance. It becomes possible.

FIG. 18 is a diagram showing an example in which the divided regions on the horizontal axis are set at unequal intervals based on the FWHM function. (A) shows the case of process time τ4, and (b) shows the case of process time τ1. Since the spectrum of the process time τ4 has higher energy resolution, the spectrum of the process time τ1 shows small spectra such as CKa and OKa that have been mixed into the large spectrum. Therefore, the divided region on the horizontal axis is set more finely in the case of the spectrum of τ4 than in the case of τ1. In this way, by making the divided regions on the horizontal axis unequal, the spectrum can be classified correctly reflecting the characteristics of the shape of the low energy portion. According to the present invention, the following effects can be obtained.
1) It is not necessary to consider the correlation between spectra, and the processing is completed at high speed.
2) Since the intensity is considered as the relative intensity in a single spectrum, the obtained feature quantities can be compared with each other even when the counting rate differs depending on the analysis position.
3) By changing the number of sections based on judgment by the user or automatically, the amount of information included in the feature amount can be adjusted.
4) By the methods of Example 2 to Example 4, it is possible to analyze with an emphasis on the energy region where there is a large amount of information depending on the tendency of the spectrum.
5) By changing the divided area using a function of the waveform resolution, it is possible to more appropriately discriminate even when the measurement mode (variing the pulse processing time) or the divided area is arbitrarily changed.
6) By omitting the low energy region from the divided region automatically / manually, it is possible to prevent erroneous determination as a different phase even when the sample surface is rough.

DESCRIPTION OF SYMBOLS 1 Scan coil 2 Electron gun 3 X-ray spectrometer 4 Electron beam detector 5 Control apparatus 6 Sample 7 Monitor

Claims (6)

In an energy dispersive X-ray analyzer attached to a scanning electron microscope or a transmission electron microscope, a sample is irradiated with a charged particle beam to detect X-rays generated from the sample for each predetermined section. From the line, an X-ray spectrum representing energy on the horizontal axis and the number of X-rays on the vertical axis is obtained.
The spectrum is divided into M on the horizontal axis and N on the vertical axis, and the numerical value on the vertical axis is obtained for each M-divided area on the horizontal axis, and the obtained numerical sequence is obtained for each section.
Collecting the sections of the same numerical sequence, coloring the sections determined to be the same composition by the same numerical sequence with the same color,
Combining individual distributions determined as each composition to obtain a distribution map,
A method for classifying spectra of an energy dispersive X-ray analyzer. In an energy dispersive X-ray analyzer attached to a scanning electron microscope or a transmission electron microscope, a sample is irradiated with a charged particle beam to detect X-rays generated from the sample for each predetermined section. Spectrum acquisition means for obtaining an X-ray spectrum representing energy on the horizontal axis and the number of X-rays on the vertical axis from the line;
A numerical value obtained by dividing the spectrum obtained from the spectrum acquisition means into M on the horizontal axis and N on the vertical axis to obtain a numerical value on the vertical axis for each M-divided area on the horizontal axis, and obtaining the obtained numerical sequence for each section Column calculation means;
Coloring means for grouping sections of the same numerical sequence and coloring sections determined to have the same composition by the same numerical sequence with the same color;
Distribution map acquisition means for obtaining a distribution map by combining individual distributions determined as respective compositions;
An apparatus for classifying a spectrum of an energy dispersive X-ray analyzer characterized by comprising:   3. The spectrum classification apparatus for an energy dispersive X-ray analyzer according to claim 2, wherein a specific divided area of the M divisions is further divided into a plurality of areas to obtain a numerical string.   3. The spectrum classification apparatus for an energy dispersive X-ray analyzer according to claim 2, wherein a numerical sequence is obtained by considering a plurality of specific areas in the M division as one area.   5. The spectrum of the energy dispersive X-ray analyzer according to claim 2, wherein when the sample surface has an uneven shape, the low energy portion of the spectrum is excluded from the object of determination. Classification device.   The waveform resolution FWHM of the energy dispersive X-ray analyzer is expressed as a function of energy E and pulse processing time τ, and the division energy width is varied based on this function when dividing the horizontal axis region. The spectrum classification apparatus of the energy dispersive X-ray analyzer according to any one of claims 2 to 5.