JP3645533B2 - X-ray analysis method and X-ray analysis system using the method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an X-ray analysis method and apparatus for qualitative analysis and quantitative analysis of a sample by detecting characteristic X-rays generated by irradiating the sample with an electron beam, and particularly suitable for analyzing a very small region of the sample. The present invention relates to an X-ray analysis method and an X-ray analysis system.
[0002]
[Prior art]
In general, X-ray analysis is often used for qualitative analysis for knowing the constituent elements of a sample and quantitative analysis for knowing the abundance of elements contained in the sample. This is because when the sample surface is irradiated with an electron beam or X-ray accelerated at a high voltage, the inner electron of the atom of the sample substance element is excited by the energy of the incident electron beam or X-ray and repels out of the shell. Is done. At this time, electrons in the outer shell fall, and energy corresponding to the energy difference between the electrons in the outer shell and the inner shell is emitted as characteristic X-rays. Since this value has an element-specific value, the emitted X-ray also has a specific wavelength or energy and is called a characteristic X-ray. By measuring the wavelength or energy of these characteristic X-rays and their intensities, qualitative and quantitative analysis of the elements of the sample surface material can be performed.
[0003]
Due to the spread of high-luminance and ultra-fine electron beam probes such as field emission electron guns in recent years, evaluation of ultra-fine regions in semiconductor device evaluation and the like has been strongly demanded. Due to such an increase in demand, the opportunity to use a high-intensity and extremely minute electron beam probe as a primary probe for X-ray analysis is rapidly increasing. However, the problem with X-ray analysis using a high-intensity and ultra-fine electron beam probe is that the reliability of the data depends on the sample misalignment due to drift of the sample and the irradiated electron beam and the change in the sample state due to electron beam damage. There was a technical problem of being damaged.
[0004]
That is, since an electron beam probe having a beam diameter of 1 nmφ or less can be used as a primary probe for X-ray analysis by using a field emission electron gun, 1 nmφ is often required as an evaluation region. In order to realize a simple evaluation region, the accuracy of nanometer order is required as an analysis position. The stability of the sample stage and the electron beam probe is not always sufficient for such a requirement, and there is a high possibility that a drift of nanometer order will occur in the sample stage and / or the electron beam probe during spectrum integration. Even if an extremely small region is evaluated under such a situation, the actual evaluation region takes into account the drift amount of the sample stage and irradiation probe.
[0005]
Further, when a sample is irradiated with a high-intensity and extremely minute electron beam, a gas component such as hydrocarbon adsorbed on the sample or the sample holder is attracted to the electron beam probe, and carbon contamination called contamination is generated. Since contamination grows in the irradiation region of the electron beam probe, it causes a local increase in sample thickness. The irradiated electron beam probe is scattered by contamination, and the electron beam irradiation area is enlarged. Therefore, X-rays excited by electron beam irradiation are generated from a wider range than a preset very minute electron beam irradiation region, and correct information on the target region cannot be obtained.
[0006]
Further, depending on the sample to be evaluated, knock-off (Knock Off) occurs in which atoms in the sample are knocked out of the sample by the irradiated high acceleration electron beam, and the ratio of elements constituting the sample may vary with the electron beam irradiation. is there. The probability that the atoms in the sample are knocked out by knock-off includes the scattering cross section for the sample's highly accelerated electrons and the sample thickness for the ejected atoms to be released into the vacuum without remaining in the sample. However, the phenomenon is not limited to a sufficiently thin sample used for high-resolution image observation by a transmission electron microscope, and is a phenomenon that occurs even in a relatively thick region suitable for X-ray analysis. When Knock Off occurs, the sample density decreases with the irradiation of the high-intensity electron beam, and vacancies may eventually be formed. That is, since the sample state changes due to the electron beam irradiation during the integration time required for the analysis, correct information on the electron beam irradiation region cannot be obtained.
[0007]
For example, JP-A-5-144402 and JP-A-6-213830 are known as solutions to technical problems that occur when the above-described X-ray analysis method is applied to a minute portion. In other words, the former compares the transmitted electron dose measured at the start of analysis of a micro area with the transmitted electron dose during analysis when analyzing a micro part such as a microcrystal or a grain boundary. By interrupting the capture of the signal that is generated next, the displacement of the analysis position due to the thermal drift of the sample itself, the drift of the sample holder, or the like is detected and corrected. In the latter case, the electron beam is used only for X-ray analysis during analysis in the conventional system, and is not used for secondary electron image formation, so it is difficult to confirm and correct the deviation of the analysis position caused by sample drift. In view of the above, time-division of the electron beam is performed, and X-ray analysis position irradiation and X and Y two-dimensional irradiation for obtaining a secondary electron image are alternately performed to detect a time-series position shift. This is to be corrected.
[0008]
[Problems to be solved by the invention]
However, the techniques disclosed in JP-A-5-144402 and JP-A-6-213830 described above are exclusively the thermal drift of the sample itself and the drift of the sample holder that occur when performing an X-ray analysis of a very small portion of the sample. There is a major technical problem that no consideration has been given to dealing with the contamination of the electron beam probe and dealing with the knock-off phenomenon in the sample. .
[0009]
Therefore, an object of the present invention is made to solve the above technical problem, and even if X-ray analysis is performed on a very small region portion of the sample, drift or contamination of the sample or the sample stage, knock-off, etc. It is an object of the present invention to provide an X-ray analysis method and an analysis system thereof capable of detecting the influence of the above phenomenon on an X-ray analysis spectrum and evaluating and correcting the reliability of the spectrum.
[0010]
[Means for Solving the Problems]
That is, the X-ray analysis method of the present invention is an X-ray analysis method that converges and irradiates an electron beam to an analysis point on a sample and measures a characteristic X-ray to be excited until a preset time elapses or is set in advance. Until the peak intensity is reached, X-ray analysis spectra obtained by accumulating the accumulated characteristic X-ray intensities are divided and collected as at least two temporally continuous X-ray analysis spectra, and there is a difference between the collected X-ray analysis spectra. When the fluctuation exceeds a predetermined range, it is determined that the analysis point is an unsuitable place for measurement, and feedback is performed to change the position to another analysis point and perform measurement again. In some cases, it is possible to confirm whether or not the electron beam is stably convergently irradiated to the analysis point before performing this position change, or after performing a position change described above after a certain time for confirmation. Therefore, it is possible to provide feedback that enables more reliable and accurate measurement.
[0011]
In this way, by extracting and displaying information that may vary with the passage of time simultaneously with the acquisition of the spectrum, it is possible to determine in real time whether or not to continue integration during analysis. In addition, the amount of change over time can be obtained by using a temporally continuous spectrum in which a change in the sample state is confirmed by irradiation with high brightness and a very small electron beam.
[0012]
In the X-ray analysis method of the present invention, the number of detected characteristic X-rays, the detection efficiency of the X-ray detector, and the quantitative calculation value in two or more temporally continuous X-ray analysis spectra are calculated in real time. It is characterized by detecting fluctuations in the X-ray analysis spectrum during measurement. As a result, the fluctuation that occurred during the measurement is estimated from the detection efficiency in each spectrum, the detected count number, or the fluctuation status of the quantitative calculation value obtained from each spectrum, and the state at the start of measurement or before the fluctuation is predicted. It becomes possible.
[0013]
Furthermore, in the X-ray analysis method of the present invention, when the intensity distribution of the X-ray analysis spectrum fluctuates with time in two or more time-sequential X-ray analysis spectra, the analysis is caused by the drift of the sample or the electron beam probe. When a change in the point is detected and a carbon peak appears on two or more time-sequential X-ray analysis spectra, the occurrence of contamination in the electron probe is confirmed, and the characteristic X-ray intensity is remarkable. If a significant decrease is confirmed, it is estimated that a knock-off phenomenon has occurred, and feedback to the measurement is performed. As a result, analysis and correction of data corresponding to a failure associated with irradiation of a very small portion of the sample can be performed without delay.
[0014]
Furthermore, in the analysis system of the present invention, the X-ray analysis spectrum in which the characteristic X-ray intensity is accumulated and accumulated is continuously continuous for at least two times until a preset time elapses or until a preset peak intensity is reached. The X-ray analysis spectrum is divided and collected, and when a difference between the collected X-ray analysis spectra exceeds a predetermined range and fluctuates, feedback to measurement is performed. And, preferably, in two or more time-continuous X-ray analysis spectra, the number of detected characteristic X-rays, the detection efficiency of the X-ray detector, and the quantitative calculation value are evaluated in real time to measure. It is characterized by detecting fluctuations in the X-ray analysis spectrum. Further, preferably, when the intensity distribution of the X-ray analysis spectrum changes with time in two or more time-sequential X-ray analysis spectra, it is detected that the analysis point has changed due to drift of the sample or the electron beam probe. When a carbon peak appears on two or more temporally continuous X-ray analysis spectra, the occurrence of contamination in the electron probe is confirmed, and a marked decrease in the characteristic X-ray intensity is confirmed. Is characterized by estimating the occurrence of the knock-off phenomenon and performing feedback to the measurement.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the X-ray analysis method and analysis system of the present invention will be described below with reference to the drawings. In this embodiment, the present invention is applied to knock-off in a transmission electron microscope. In the drawings, the same portions are denoted by the same reference numerals.
[0016]
FIG. 1 shows the configuration of a transmission electron microscope system to which an X-ray analysis method according to the present invention is applied. This X-ray analysis system includes a transmission electron microscope (not shown) having a function of irradiating an observation and analysis point with a high-intensity and ultrafine electron beam, and an energy dispersive X-ray analysis provided in the transmission electron microscope. Of the semiconductor detector 1, a pulse height analyzer (hereinafter referred to as PHA) 2 for converting a signal detected by the semiconductor detector 1 into an X-ray intensity, and a signal from the PHA 2, This is an energy dispersive X-ray analysis system including a computer 3 for performing analysis spectrum display and quantitative calculation.
[0017]
In such a system, the sample 4 is irradiated with a high-intensity focused electron beam 5. In this embodiment, an epitaxial SiGe cross-sectional sample formed on a Si single crystal was used as the sample 4. The characteristic X-ray 6 is excited from the region irradiated with the high-intensity electron beam 5 and is taken into the semiconductor detector 1. In the semiconductor detector 1, an electric signal corresponding to the X-ray intensity is formed, transmitted to the PHA 2, and converted into a digital signal. The signal digitized in the PHA 2 is sent to the data processing computer 3. At this time, the computer 3 divides and stores the X-ray analysis spectrum for each preset integration time t, for example. In this embodiment, the integration time of one spectrum is 50 seconds, and the total integration time of 400 seconds is divided and acquired as eight temporally continuous spectra. In addition, the measurement was carried out twice at an analysis point close to the same sample, with an electron beam probe diameter on the sample 4 of 100 nmφ (analysis point 1) and 1 nmφ (analysis point 2). From two sets of temporally continuous spectra obtained by two measurements, the transition of the count numbers of Si and Ge at each measurement time and the Ge concentration (Atom%) calculated from these count numbers are shown in FIG. Analysis points 1 and 2 are shown in FIG. As apparent from FIG. 2, it can be determined that the fluctuations appearing in the Si count number and the Ge count number at the analysis point 1 are within the statistical error range. On the other hand, the Si count number at the analysis point 2 decreases almost uniformly from the start of measurement, confirming that the Si element is knocked out by knock-off. Such fluctuations in the number of counts are not caused by changes in detection efficiency, but are clearly caused by changes in the sample state in the electron beam irradiation region, and even those detections are difficult with conventional X-ray analysis methods. Met. However, in this embodiment, a continuous spectrum is acquired in time, so that information that may vary over time can be extracted and displayed at the same time as the acquisition of the spectrum, so that integration can be continued or interrupted during measurement. Can be determined in real time.
[0018]
In addition, the presence or absence of a change in the sample state associated with irradiation with high brightness and a very small electron beam can also be determined by determining the amount of change over time. For example, assuming that the Si count number and the Ge count number shown in the graph of the analysis point 2 in FIG. 2 are decreasing at a uniform rate, fitting by the least square method was performed. The formula for giving the Si count number and Ge count number obtained by the fitting and the formula for giving the Ge concentration (Atom%) at each time t are as shown in FIG. By obtaining the value when the curve indicating the Ge concentration (Atom%) with the passage of time is t = 0, the state before the sample state is changed by the electron beam irradiation can be estimated. In the quantitative calculation by the conventional analysis method, the Ge concentration was 20.2 Atom%. However, by predicting the amount of change with the passage of time according to the present embodiment, the value before the sample state is changed is 16. A result of 0 Atom% was obtained. However, in this embodiment, fitting by the least square method is performed on the assumption that the variation amount is constant with respect to the Si count number and the Ge count number as time elapses. Since the interaction with the line does not always work constantly, it is also assumed that the amount of fluctuation per unit time repeatedly increases, decreases, or increases / decreases with time. For such variations, it is possible to set a function used for fitting according to the situation. As a method for confirming the possibility that the measured count number or the variation in the spectrum depends on the stability of the apparatus, it is also effective to confirm the time lapse of the saturation time ratio (Deadtime) of the detector.
[0019]
In the above-described embodiment, the present invention is described with an example in which the present invention is applied to knock-off in a transmission electron microscope. It goes without saying that various modifications are possible within the scope of the invention described in the claims, and these are also included in the scope of the present invention.
[0020]
The present invention will be described below with reference to flowcharts. FIG. 3 is a flowchart of the X-ray analysis method of the present invention. As can be seen in the figure, the sample is first prepared and the TEM is started. After determining the position of the sample to be analyzed, a means for suppressing the drift of the sample is implemented. If drift cannot be eliminated by this means, other drift suppression means is employed. If the drift is successfully suppressed, electron beam convergent irradiation is performed. Then, it is determined whether or not there is a drift in the irradiation beam. If it is determined that beam drift has occurred in this process, beam drift suppression means is performed. When it is confirmed that beam drift does not occur, measurement is started.
[0021]
In the process of measurement, impurity contamination of the measurement system, that is, generation of contamination, decrease in detection efficiency and spectrum fluctuation are monitored, and if these occurrences are observed, it is determined whether or not re-experiment is possible. In other words, if the contamination is widespread and the influence of the contamination cannot be eliminated even if the analysis point is changed, the measurement is continued as it is, and if the analysis point that can eliminate the influence of the contamination is found, the analysis point can be found. Change to start measurement. Similarly, in the case of other reductions in detection efficiency or significant fluctuations in the spectrum, it is determined whether or not re-experiment is possible, and the continuation of measurement is determined.
When the measured analysis spectrum has sufficient intensity, the measurement is terminated and the peak intensity measurement of the component elements is started. On the other hand, when sufficient measurement intensity is not obtained, the measurement is continued further retroactively.
[0022]
The peak intensity of the component elements is then measured and the spectrum is divided into appropriate time intervals. If there is a significant difference in intensity change between the divided spectra, it is determined whether or not the intensity change is uniform. If this change is uniform, the rate of change is estimated and the peak intensity ratio at the beginning of the experiment is calculated. Based on this result, quantitative calculation is performed to determine the composition ratio.
On the other hand, if the intensity fluctuation between spectra is not uniform, the cause of the fluctuation is estimated, and if the data is reliable, an appropriate evaluation time range is determined, and the change in spectrum intensity is evaluated within that range. To do. If there is a significant difference in the change in the spectrum intensity and the change is uniform, the peak intensity ratio at the initial stage of the experiment is calculated from the change rate, and the composition ratio is determined by metric calculation.
In addition, if there is no significant difference in intensity fluctuation between spectra or the intensity does not change uniformly, the true peak intensity ratio is estimated in consideration of the estimated fluctuation factors, and the composition is similarly calculated by quantitative calculation. Determine the ratio.
[0023]
As described above, according to the present invention, when performing a quantitative analysis of a minute region of a sample using a high-energy electron beam, the analysis is performed while appropriately monitoring the change in the state of the sample, and the measurement is performed. Since it becomes possible to make an appropriate judgment about the change of the condition or the validity of the result, the reliability of the result is improved.
[0024]
【The invention's effect】
According to the X-ray analysis method and the analysis system of the present invention, it is possible to detect fluctuations in measurement conditions caused by irradiation with high-intensity and ultra-fine electron beams during integration of X-ray spectra. Reliability can be confirmed. In addition, it is possible to estimate the state before the fluctuation occurs by estimating the fluctuation amount with time from two or more temporally continuous spectra.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a configuration of a transmission electron microscope system to which an X-ray analysis method according to the present invention is applied.
FIG. 2 is a graph showing the transition of the Si and Ge count numbers for each measurement time and the Ge concentration (Atom%) calculated from the count numbers at the analysis points 1 and 2;
FIG. 3 is a flowchart for explaining an embodiment of the present invention.
1: Semiconductor detector 2: PHA
3: Computer 4: Sample 5: Irradiated electron beam 6: Characteristic X-ray

Claims (4)

In an X-ray analysis method for measuring an excited characteristic X-ray by irradiating an electron beam to an analysis point on a sample and measuring a characteristic X-ray until a preset time elapses or a preset peak intensity is reached. When X-ray analysis spectrum accumulated and accumulated in intensity is divided and collected as at least two temporally continuous X-ray analysis spectra, and the difference between the collected X-ray analysis spectra exceeds a predetermined range and fluctuates. The X-ray analysis method is characterized in that the analysis point is determined to be a place where measurement is inappropriate and feedback is performed to change the position to another analysis point and perform measurement again. In the two or more time-sequential X-ray analysis spectra, the number of detected characteristic X-rays, the detection efficiency of the X-ray detector, and the quantitative calculation value are evaluated in real time, so that X The X-ray analysis method according to claim 1, wherein a change in the line analysis spectrum is detected. When the intensity distribution of the X-ray analysis spectrum fluctuates with time in the two or more time-sequential X-ray analysis spectra, it is detected that the analysis point has fluctuated due to drift of the sample or electron beam probe, When a carbon peak appears on two or more temporally continuous X-ray analysis spectra, confirm the occurrence of contamination in the electron probe, and if a significant decrease in characteristic X-ray intensity is confirmed, The X-ray analysis method according to claim 1, wherein the occurrence of a knock-off phenomenon is estimated and feedback to measurement is performed. 4. An X-ray analysis system incorporating the X-ray analysis method according to claim 1.

Images