JPH0675003B2 - Spectral separation processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of Industrial Application The present invention separates an energy spectrum obtained by X-ray photoelectron spectroscopy (XPS), emission spectroscopy, X-ray diffraction, etc. into a reference spectrum from a reference spectrum. In this case, the present invention relates to a spectrum separation processing method suitable.

(B) Prior art and its problems In general, an energy spectrum obtained by X-ray photoelectron photoanalysis or the like is given as a distribution curve including many peaks. Then, the substance indicated by each peak is identified and quantified from the position, height, and half width of each peak of this energy spectrum.

By the way, in the measured energy spectrum, peaks for each substance rarely appear separately,
In many cases, adjacent peaks overlap each other. As a result, the measured spectrum peak is different from the spectrum pattern of the substance alone, and the peak height, half-value width, etc. cannot be determined as they are. Therefore,
In particular, it becomes difficult to quantify the substance.

Therefore, conventionally, the measured energy spectrum is displayed on a CRT, etc., and the displayed energy spectrum is regarded as a set of distribution curves such as Gaussian distribution or Lorentz distribution, and the distribution curve is expanded or reduced. Then, the reference spectrum is set by matching each peak of the energy spectrum, and then the matched reference spectrum is subtracted from the energy spectrum. Then, the energy spectrum finally measured is separated into the reference spectra by repeating the above-mentioned operation in sequence.

In such a method of separating the energy spectrum, if the fitting of the distribution curve is not appropriate, as a result of the separation, many peaks will remain at positions other than the assumed positions. When there is a residual peak in this manner, conventionally, the measured spectrum of the original energy spectrum is returned to, the distribution curve is selected again from the beginning, and the reference spectrum is separated again for each peak.

However, in such a conventional method, the separation operation must be repeated many times, and the separation operation is complicated and extremely inefficient.

The present invention has been made in view of such circumstances, and it is an object of the present invention to reduce the number of operations for separating an energy spectrum obtained by measurement into a reference spectrum and to efficiently perform the separation operation. And

(C) Means for Solving the Problems In order to achieve the above object, the present invention provides a distribution curve such as a Gaussian distribution or Lorentz distribution for each peak included in an energy spectrum obtained by measuring a sample. After selecting a reference spectrum by adjusting the peak position and expansion rate of the distribution curve so that the selected distribution curve fits the peak of the energy spectrum, subtract the reference spectrum created from the energy spectrum. In the method of separating the energy spectrum into each reference spectrum, when there is a residual peak at the separated proximity position when the reference spectrum is separated from the energy spectrum, the separated reference spectrum is added again to the separated spectrum, While again selecting the distribution curve for this added spectrum, , The selected distribution curve is fitted to the added spectrum to newly create a reference spectrum, and a separation operation is performed to subtract the newly created reference spectrum from the immediately preceding spectrum.

(D) Embodiment FIG. 1 is a block diagram of an analytical data processing device for applying the present invention. In the analysis data processing device 1 of this embodiment, the data measured by X-ray photoelectron spectroscopy etc. is temporarily stored in the memory 4 via the CPU 2, and the measurement data stored in the memory is read out by the CPU 2 and stored in the CRT 6. The energy spectrum is displayed with energy on the horizontal axis and intensity on the vertical axis. Further, the memory 4 stores the data of distribution curves of Gaussian distribution and Lorentz distribution when a reference spectrum is set in advance. An operation unit 8 is operated when the energy spectrum is separated into the reference spectrum.

Next, a method for separating the energy spectrum into the reference spectrum will be described with reference to the flowchart shown in FIG.

For example, when the energy spectrum is measured by X-ray photoelectron spectroscopy analysis or the like (step 1), the following operation is performed on this energy spectrum. First, it is assumed that a section having an energy spectrum to be separated from the reference spectrum has, for example, a shape as shown by the third (a) solid line. This energy spectrum is 2
Since the two peaks P ₁ and P ₂ overlap each other, the peak height of each reference spectrum cannot be determined as it is. Therefore, the operation unit 8 is operated to set an initial condition for separating the reference spectrum (step 2). That is, a distribution curve is set in consideration of the characteristics of the apparatus such as the energy section separated into the reference spectrum, the number of peaks to be separated, the Gaussian distribution, and the Lorentz distribution. Further, when performing quantitative analysis, the peak positions are also known in advance, so the peak positions λ ₁ and λ ₂ are also set.

Next, through step 3, by setting the peak height, the half-value width, etc. of the selected distribution curve at the peak position λ ₁ set as described above, the pan is set so that the distribution curve is the closest to the energy spectrum. Enlarge, shrink, and distribute the distribution curve to one peak P _{1 of the} energy spectrum
To fit (step 4). For this purpose, the distribution curve stored in the memory 4 is read out and displayed on the CRT 6, and the operation section 8 is operated while viewing this distribution curve. When the reference curve S ₁ (shown by the broken line in FIG. 3A) is created by fitting the distribution curve in this way (step 5), the created reference spectrum is subtracted from the energy spectrum. This is CPU2
Done by. Subtracting the reference spectrum S ₁ gives
The energy spectrum is as shown by the solid line in FIG.

Next, if the separation has not been completed for the designated peaks P ₁ and P ₂ (step 7), the process returns to step 3 and the above separation operation is repeated. That is, in this example, the reference spectrum S ₂ is separately created at the peak position λ ₂ , and the created reference spectrum S ₂ is subtracted from the energy spectrum. The energy spectrum after the subtraction is shown in Fig. 3 (c).
The pattern is as shown in.

Reference spectrum from measured energy spectrum
When separation is insufficient when separating S ₁ and S ₂ , and residual peaks P ₁ , P ₂ and P ₃ are present at the separated close positions, re-separation is required in step 3. Judgment is made, and then the separated reference spectrum (for example, S ₁ ) is added again to the separated spectrum (step 8). The added spectrum has a pattern as shown by the solid line in FIG. Therefore, the separation operation from step 4 to step 7 is repeated for this added spectrum. That is, the distribution curve is selected again, the peak height of the selected distribution curve,
The half value width is reset and the distribution curve is fitted to the added spectrum. In this way, a new reference spectrum S ₁ ′ is created, and the newly created reference spectrum S ₁ ′ is subtracted from the immediately preceding spectrum. The energy spectrum after the subtraction has a pattern as shown in FIG.

As described above, when the separation operation is completed for all the designated peaks included in the energy spectrum, the process proceeds to step 9, and if there are no remaining peaks, the separation operation is completed (step 10).

If residual peaks are still present, set which peak to reseparate before separation (see step
11) Return to step 3 and repeat the above separating operation again.

In this way, based on the peak height and half width data for the reference spectrum finally obtained
The content of the sample is calculated by CPU2 and the data is displayed on CRT6.

(E) Effect According to the present invention, when a residual peak is present after separation into the reference spectrum, the conventional spectrum is returned to the pattern of the energy spectrum of the group, and the distribution curve is reselected from the beginning to obtain the reference spectrum. Instead of performing the separation operation of step 1, the separated reference spectrum is added again to the separated spectrum, and the separation operation is performed on this added spectrum. Decrease. Therefore, excellent effects such as efficient separation operation can be achieved.

[Brief description of drawings]

The drawings show an embodiment of the present invention. FIG. 1 is a block diagram of an analytical data processing device, FIG. 2 is a flow chart for explaining the method of the present invention, and FIG. 3 is a measured energy spectrum and a reference thereof. It is a figure which shows the pattern after isolate | separating a spectrum. S ₁ , S ₂ …… Reference spectrum, P ₁ , P ₂ …… Peak, λ ₁ ,
λ ₂ ... peak position.

Claims (1)

[Claims] 1. A distribution curve such as a Gaussian distribution or Lorentz distribution is selected for each peak included in an energy spectrum obtained by measuring a sample, and the selected distribution curve is fitted to the peak of the energy spectrum. It is a method of separating the energy spectrum into each reference spectrum by subtracting the reference spectrum created from the energy spectrum after creating the reference spectrum by adjusting the peak position and expansion rate of the distribution curve. If there are residual peaks in the separated positions when the spectra are separated,
The separated reference spectrum is added again to the separated spectrum, the distribution curve is selected again for this added spectrum, and a new reference spectrum is created by fitting the selected distribution curve to the added spectrum.
A spectrum separation processing method characterized by performing a separation operation of subtracting a newly created reference spectrum from the immediately preceding spectrum.