JP2017069492A - Evaluation method for silicon single crystal substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method capable of quantitatively evaluating striation of a silicon single crystal substrate.SOLUTION: The present invention relates to an evaluation method for a silicon single crystal substrate. The evaluation method for the silicon single crystal substrate includes the steps of: creating map data by measuring at fixed intervals distribution of a gradient of resistivity within a measurement surface of the silicon single crystal substrate; converting the map data into text data and extracting data in one direction of the silicon single crystal substrate from the text data; calculating a signal that is a moving average for each first width A in the one direction of the silicon single crystal substrate, and a base line that is a moving average for each second width B that is triple or more as wide as the first width A, from the extracted data; calculating a differential between the signal and the base line; and evaluating the calculated differential as a striation strength.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a method for evaluating a silicon single crystal substrate.

  With the recent high integration of devices, there is an increasing demand for silicon wafers having a uniform resistivity in the plane. In the LPS apparatus (Lateral Photovoltaic Scanning, manufactured by LPCon, Germany), the resistance gradient in the measurement surface of the wafer to be evaluated can be displayed as a map to visualize resistance fringes (hereinafter also referred to as striations) ( Non-patent document 1). Conventionally, a map showing the resistivity gradient created by such an apparatus is observed with the human eye to determine the presence or absence of striations and the quality of a wafer to be evaluated.

H. J. et al. Schulze, et al. , J. et al. Electrochem. Soc. 143 (1996) 4105

  However, the conventional method for determining the presence or absence of striations from the appearance of the map as described above and the quality thereof cannot quantitatively evaluate the striations of the silicon single crystal substrate to be evaluated. Therefore, the result of evaluation may vary depending on the operator.

  The present invention has been made in view of the above problems, and an object thereof is to provide a method capable of quantitatively evaluating the striation of a silicon single crystal substrate.

In order to achieve the above object, the present invention provides a method for evaluating a silicon single crystal substrate,
Measuring the distribution of resistivity gradient in the measurement surface of the silicon single crystal substrate at regular intervals to create map data; converting the map data into text data; and converting the text data into the silicon single crystal A step of extracting data in one direction of the substrate, a signal that is a moving average for each first width A in the one direction of the silicon single crystal substrate from the extracted data, and 3 of the first width A Respectively, a step of obtaining a baseline that is a moving average for each second width B equal to or greater than twice, a step of calculating a difference between the signal and the baseline, and a step of evaluating the calculated difference as a striation strength. A method for evaluating a silicon single crystal substrate is provided.

  By doing so, the signal reflects the resistance fringe information of the silicon single crystal substrate to be evaluated, and the baseline reflects the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated. In the striation strength that is the difference, a true resistance stripe state in which the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated are removed appears. Therefore, since the digitized striation strength can be obtained from the map data indicating the resistivity gradient, the striation of the silicon single crystal substrate can be quantitatively evaluated.

At this time, the silicon single crystal substrate to be evaluated is a silicon wafer cut from a silicon single crystal ingot or a silicon single crystal slab for inspection,
The one direction can be a radial direction of the silicon single crystal ingot.

  Thus, the present invention can quantitatively measure the striation of a silicon wafer or a silicon single crystal slab for inspection.

  Moreover, it is preferable that the interval for measuring the distribution of the resistivity gradient in the measurement plane is in the range of 1 μm to 500 μm.

  In this way, while maintaining the positional accuracy of the motor of the measuring device, the measurement of the resistivity gradient within the measurement surface of the silicon single crystal substrate to be evaluated is performed while suppressing the measurement from taking a long time. It can be done more precisely.

Further, the first width A is set to be 10 times or more of an interval for measuring the distribution of resistivity gradient in the measurement plane,
It is preferable that the second width B is 30 times or more the interval for measuring the distribution of the resistivity gradient in the measurement surface.

  In this way, the true resistance fringe state in which the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated are removed more reliably appears in the striation strength.

  Further, it is preferable that the first width A is 0.5 mm or more and the second width B is 5 mm or more. Furthermore, it is preferable that the first width A is 1 mm or more and the second width B is 10 mm or more.

  By using the first width A and the second width B as described above, a more accurate resistance fringe state of the evaluation sample appears in the striation strength.

The first width A is 10 mm or less,
The second width B is preferably 30 mm or less.

  By setting the first width A and the second width B as described above, the peripheral exclusion region becomes large while preventing the fine resistance fringes from becoming unrecognizable, and the resistance fringes on the outer peripheral portion become larger. It is possible to prevent the evaluation from becoming impossible.

  In the method for evaluating a silicon single crystal substrate of the present invention, it is preferable to evaluate the silicon single crystal having a striation strength of a desired value or less as a non-defective product.

  As described above, in the evaluation method of the present invention, the quality of the silicon single crystal substrate can be determined by quantitative evaluation, and only non-defective products with or without resistance fringes can be provided.

  In addition, it is preferable to obtain an arithmetic average Ra of the absolute value of the striation strength and evaluate using the obtained arithmetic average Ra.

  In this way, by using the arithmetic average Ra, it is possible to easily evaluate the striation strength in the entire measurement plane of the silicon single crystal substrate to be evaluated.

  With the silicon single crystal substrate evaluation method of the present invention, since the digitized striation strength can be obtained from the map data indicating the gradient of resistivity, the striation of the silicon single crystal substrate is quantitatively evaluated. can do.

It is process drawing which shows an example of the evaluation method of the silicon single crystal substrate of this invention. It is map data which shows distribution of the resistivity gradient in the measurement surface of the silicon single crystal slab A for inspection. It is map data which shows distribution of the gradient of resistivity in the measurement surface of the silicon single crystal slab B for inspection. It is map data which shows distribution of the resistivity gradient in the measurement surface of the silicon single crystal slab C for inspection. It is map data which shows distribution of the gradient of resistivity in the measurement surface of the silicon single crystal slab D for inspection. It is the graph which showed the relationship between the LPS data in the inspection silicon single crystal slab A, the baseline, the signal, and the striation intensity, and the distance from the center of the inspection silicon single crystal slab. It is the graph which showed the relationship between the LPS data in the inspection silicon single crystal slab B, the baseline, the signal and the striation intensity, and the distance from the center of the inspection silicon single crystal slab. It is the graph which showed the relationship between the LPS data in the inspection silicon single crystal slab C, the baseline, the signal, and the striation intensity, and the distance from the center of the inspection silicon single crystal slab. It is the graph which showed the relationship between the LPS data in the inspection silicon single crystal slab D, the baseline, the signal, and the striation intensity, and the distance from the center of the inspection silicon single crystal slab.

  Hereinafter, although an embodiment is described about the present invention, the present invention is not limited to this.

  As described above, the conventional method of judging the presence or absence of striations and the quality of the silicon single crystal substrate to be evaluated from the visual impression of the map, quantitatively evaluates the striation of the silicon single crystal substrate to be evaluated. I couldn't. Therefore, the present inventors have intensively studied to solve such problems.

  As a result, the map data showing the resistivity gradient reflects the signal that reflects the resistance fringe information of the silicon single crystal substrate to be evaluated and the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated. It was found that the difference was evaluated as the striation strength. In this way, it has been found that the striation of the silicon single crystal substrate can be quantitatively evaluated because the quantified striation intensity can be obtained from the map data indicating the resistivity gradient. And the best form for implementing these was scrutinized and the present invention was completed.

  Hereinafter, a method for evaluating a silicon single crystal substrate of the present invention will be described with reference to FIG.

  First, a silicon single crystal substrate to be evaluated is prepared (SP1 in FIG. 1). Although it does not specifically limit as a silicon single crystal substrate made into evaluation object, For example, it can be set as the silicon wafer or silicon single crystal slab for a test | inspection cut out from the silicon single crystal ingot. The silicon single crystal ingot can be, for example, a cylindrical one manufactured by a known method such as a floating zone (FZ) method or a Czochralski (CZ) method.

  The shape of the silicon single crystal substrate is not particularly limited. For example, in the case of a silicon wafer, the silicon wafer may be divided into ¼. Further, the inspection silicon single crystal slab may be a quadrangular shape obtained by vertically slicing a silicon single crystal ingot (that is, a silicon single crystal ingot cut in the growth axis direction). Moreover, the silicon single crystal slab for inspection can have a thickness of 5 mm or less, for example.

  Next, the distribution of the resistivity gradient in the measurement surface of the silicon single crystal substrate is measured at regular intervals to create map data (SP2 in FIG. 1).

  At this time, the entire in-plane of the silicon single crystal substrate may be measured, but map data may be created by measuring a band-like region such as 30 mm width or 20 mm width, for example.

  There is no particular limitation on an apparatus that can create map data by measuring the resistivity distribution in the measurement plane of the silicon single crystal substrate at regular intervals. For example, an LPS apparatus as described above may be used. it can. Hereinafter, a case where an LPS device is used as a measuring device will be described as an example.

  Here, the principle of measurement by the LPS apparatus is described below (Non-patent Document 1). Carriers are generated by injecting laser light (wavelength λ = 685 nm) from the surface of the sample (silicon single crystal substrate). If there is a gradient in the resistivity of the sample, carriers move by the internal potential. The potential difference generated at that time is detected by electrodes installed at both ends of the sample. Since this potential difference depends on the gradient of resistivity, the resistivity gradient can be visualized by displaying this as a map.

  For this reason, the result of the measurement by the LPS apparatus is greatly influenced by the resistivity of the evaluation sample and the surface polishing state (high brightness surface grinding surface, PW surface (mirror polished wafer surface), etc.). Therefore, the obtained map data includes not only information on resistance fringes but also the influence of resistivity and surface polishing state.

  Next, the map data is converted into text data (SP3 in FIG. 1).

  The text data can be, for example, numerical data when the map data is displayed in gray scale.

  At this time, an interval for measuring the distribution of the resistivity gradient in the measurement plane (hereinafter also simply referred to as a measurement interval) can be set by specifying an arbitrary interval in each of the X direction and the Y direction. A range of 500 μm or less is preferable. More preferably, it is in the range of 10 μm or more and 200 μm or less. In this way, while maintaining the positional accuracy of the motor of the measuring device, the measurement of the resistivity gradient within the measurement surface of the silicon single crystal substrate to be evaluated is performed while suppressing the measurement from taking a long time. It can be done more precisely.

  Next, in order to obtain resistance fringe information in the sample plane, data in one direction of the silicon single crystal substrate is extracted from the text data (SP4 in FIG. 1).

  Since the measurement by the LPS apparatus performs line scanning in the X direction and Y direction on the measurement surface, this data may be extracted by only one line of data, or data of a plurality of lines in the same direction to reduce variation. It may be averaged. For example, assuming that the measurement interval is 50 μm, the data to be extracted can be obtained by averaging 50 lines of data. In this case, since data with a width of 2.5 mm is averaged, data with reduced variations can be extracted.

  When the silicon single crystal substrate to be evaluated is a silicon wafer cut from a silicon single crystal ingot or a silicon single crystal slab for inspection, one direction for extracting data is, for example, the radial direction of the silicon single crystal ingot can do.

  In addition, when a silicon single crystal ingot having a vertically divided silicon single crystal ingot is used as the inspection silicon single crystal slab, one direction in which data is extracted can be the radial direction or the growth axis direction of the silicon single crystal ingot. . In this way, the interface shape can be confirmed.

  Next, from the extracted data, a signal that is a moving average for each first width A in one direction of the silicon single crystal substrate, and a moving average for each second width B that is three or more times the first width A. Each baseline is obtained (SP5 in FIG. 1).

  In the present invention, in this way, the second width B is set to a width that is at least three times as large as the first width A. As a result, the base line sufficiently reflects the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated.

  On the other hand, if the second width B is less than three times the first width A, the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated is not sufficiently reflected in the baseline. Therefore, if the second width B is less than three times the first width A, even if the striation strength is calculated from the difference between the signal and the baseline as will be described later, the striation strength includes the silicon to be evaluated. The true resistance fringe state from which the resistivity and surface polishing state of the single crystal substrate have been removed does not sufficiently appear, so that accurate evaluation cannot be performed.

  Next, the difference between the obtained signal and the baseline is calculated (SP6 in FIG. 1), and the calculated difference is evaluated as the striation strength (SP7 in FIG. 1).

  By doing so, the signal reflects the resistance fringe information of the silicon single crystal substrate to be evaluated, and the baseline reflects the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated. In the striation strength that is the difference, a true resistance stripe state in which the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated are removed appears. In this way, since the digitized striation intensity can be obtained from the map data indicating the resistivity gradient, the striation of the silicon single crystal substrate can be quantitatively evaluated.

  At this time, the first width A is set to 10 times or more of the interval for measuring the resistivity distribution in the measurement plane, and the second width B is set to 30 times the interval for measuring the resistivity distribution in the measurement plane. The above is preferable. In this way, the state of true resistance fringes in which the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated are removed more reliably appears in the striation strength obtained as described above.

  The first width A is preferably 0.5 mm or more, the second width B is preferably 5 mm or more, the first width A is 1 mm or more, and the second width B is 10 mm or more. More preferred. By setting the first width A and the second width B in this way, the resistance fringe state of the evaluation sample appears more accurately in the striation strength.

  Further, it is preferable that the first width A is 10 mm or less and the second width B is 30 mm or less. As described above, by setting the first width A to 10 mm or less, it is possible to prevent the fine resistance fringes from being recognized. Moreover, since it can suppress that a periphery exclusion area | region becomes large because the 2nd width B shall be 30 mm or less, it can prevent that evaluation of the resistance fringe of an outer peripheral part becomes impossible. it can.

  As described above, the map data created in the step of creating map data (SP2) includes not only information on resistance fringes but also the influence of resistivity and surface polishing state. Therefore, it is difficult to perform comparative evaluation among a plurality of samples only by looking at the map. Moreover, quantitative evaluation cannot be performed by judging the appearance of the map in the first place, and it is not possible to respond to recent high demands.

  Further, the text data obtained by converting the map data includes not only the information about the resistance fringes but also the influence of the resistivity and the surface polishing state. Therefore, in order to obtain a more accurate striation strength from the measured map data, it is necessary to remove noise components such as the influence of the resistivity and the surface polishing state from the measured map data.

  In the present invention, the difference between the signal reflecting the resistance fringe information of the silicon single crystal substrate to be evaluated and the baseline reflecting the influence of the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated is striations. Strength. As a result, in the striation strength, a true resistance stripe state in which the resistivity and surface polishing state of the silicon single crystal substrate to be evaluated are removed appears. Therefore, since the digitized striation strength can be obtained from the map data indicating the resistivity gradient, the striation of the silicon single crystal substrate can be quantitatively evaluated.

  In the method for evaluating a silicon single crystal substrate of the present invention, it is preferable to evaluate a silicon single crystal substrate having a striation strength of a desired value or less as a good product. In this way, it is possible to determine the quality of the silicon single crystal substrate by quantitative evaluation, and it is possible to provide only a high quality silicon single crystal substrate with no or no resistance fringes. In addition, the value of the striation strength that is a criterion for pass / fail is not particularly limited, and the wafer manufacturer may set it independently, or the relationship with the device (design rule, device to be manufactured, yield at the time of device manufacture, Etc.).

  At this time, it is preferable that the arithmetic average Ra of the absolute value of the striation strength is obtained and evaluated using the obtained arithmetic average Ra. Thus, by using the arithmetic mean Ra, it is possible to easily evaluate the striation strength in the entire measurement plane. In addition, if a criterion for determining whether or not resistance fringes are good is determined in advance by Ra, wafer selection / pass / fail judgment can be easily performed. In the meaning of evaluation in the entire plane, evaluation can also be performed using the root mean square roughness Rq.

  As described above, since the striation strength can be quantified according to the present invention, Rz (maximum mean square root roughness) is used in addition to Ra (absolute arithmetic mean) and Rq (root mean square roughness). Various evaluation methods such as height), Rp (maximum peak height), and Rv (maximum valley depth) can be used.

  EXAMPLES Hereinafter, although an Example and comparative example of this invention are shown and this invention is demonstrated more concretely, this invention is not limited by these.

(Example)
The production conditions were changed, and four 200 mm diameter silicon single crystal ingots were pulled up (crystals A, B, C, D). At this time, the resistivity and oxygen concentration (average of the whole crystal) of the crystals A to D were the same. A sample to be measured was cut out from each of the crystals A to D at a position where the length from the upper end of the straight body portion was 50 cm. Thereafter, the surface was subjected to high-intensity surface grinding to obtain silicon single crystal slabs A to D for inspection.

  The distribution of the resistivity gradient in the measurement planes of the inspection silicon single crystal slabs A to D was measured with an LPS apparatus, and map data as shown in FIGS. At this time, the measurement intervals in the X direction (horizontal direction in FIGS. 2 to 5) and the Y direction (vertical direction in FIGS. 2 to 5) were each 50 μm.

  Next, map data as shown in FIGS. 2 to 5 is converted into text data, and the data in the radial direction of the silicon single crystal slabs A to D for inspection (hereinafter referred to as LPS data) is obtained from the obtained text data. Extracted. The extracted radial data was an average of 50 lines. Since the measurement interval is 50 μm, that is, data with a radius of 100 mm and a width of 2.5 mm were extracted.

  Then, from the LPS data, a signal that is a moving average for each 1 mm width (first width A) and a baseline that is a moving average for each 10 mm width (second width B) were obtained. Since the measurement interval is 50 μm, in this case, the first width A is 20 times the measurement interval, and the second width B is 200 times the measurement interval.

  Furthermore, the difference between the signal and the baseline was calculated to determine the striation intensity. FIGS. 6 to 9 show the relationship between the LPS data, signals, baseline, striation strength, and distance from the center of the inspection silicon single crystal slab obtained in this manner. It was.

  Furthermore, the arithmetic average Ra of the absolute value of the striation strength calculated in each of the silicon single crystal slabs A to D for inspection was obtained and shown in Table 1 below.

  As shown in Table 1, by obtaining the arithmetic mean Ra of the absolute value of the striation strength of the silicon single crystal slabs A to D for inspection, the strength of each striation could be expressed quantitatively.

(Comparative example)
The map data (FIGS. 2 to 5) showing the resistivity gradient distribution of the inspection silicon single crystal slabs A to D created in the examples were used as comparative examples.

  As a result, as far as the map is viewed, the resistance silicon stripe slabs A (FIG. 2) and B (FIG. 3) are more resistive than the inspection silicon single crystal slabs C (FIG. 4) and D (FIG. 5). However, it is difficult to judge the superiority of A and B and the superiority or inferiority of C and D.

  On the other hand, as shown in Table 1 above, in the examples, the striation strengths of the silicon single crystal slabs A to D for inspection are quantitatively shown. Therefore, the superiority or inferiority of A and B and C and D is It can be clearly seen that A is the best and the quality deteriorates in the order of A, B, C, D in terms of resistance fringes. As can be seen from this result, if the criterion for determining whether or not the resistance fringe is good is determined in advance by Ra, wafer selection / pass / fail judgment can be easily performed.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has any configuration that has substantially the same configuration as the technical idea described in the claims of the present invention and that exhibits the same effects. Are included in the technical scope.

Claims (9)

A method for evaluating a silicon single crystal substrate,
Measuring the distribution of resistivity gradient in the measurement surface of the silicon single crystal substrate at regular intervals to create map data; converting the map data into text data; and converting the text data into the silicon single crystal A step of extracting data in one direction of the substrate, a signal that is a moving average for each first width A in the one direction of the silicon single crystal substrate from the extracted data, and 3 of the first width A Respectively, a step of obtaining a baseline that is a moving average for each second width B equal to or greater than twice, a step of calculating a difference between the signal and the baseline, and a step of evaluating the calculated difference as a striation strength. A method for evaluating a silicon single crystal substrate, comprising: The silicon single crystal substrate to be evaluated is a silicon wafer cut from a silicon single crystal ingot or a silicon single crystal slab for inspection,
The method for evaluating a silicon single crystal substrate according to claim 1, wherein the one direction is a radial direction of the silicon single crystal ingot.   3. The method for evaluating a silicon single crystal substrate according to claim 1, wherein an interval for measuring a distribution of resistivity gradient in the measurement plane is in a range of 1 μm to 500 μm. The first width A is 10 times or more of an interval for measuring the distribution of resistivity gradient in the measurement plane,
4. The silicon according to claim 1, wherein the second width B is 30 times or more of an interval for measuring a distribution of resistivity gradient in the measurement surface. 5. Evaluation method of single crystal substrate. The first width A is 0.5 mm or more,
The method for evaluating a silicon single crystal substrate according to any one of claims 1 to 4, wherein the second width B is set to 5 mm or more. The first width A is 1 mm or more,
The method for evaluating a silicon single crystal substrate according to any one of claims 1 to 5, wherein the second width B is 10 mm or more. The first width A is 10 mm or less,
The method for evaluating a silicon single crystal substrate according to any one of claims 1 to 6, wherein the second width B is set to 30 mm or less.   The method for evaluating a silicon single crystal substrate according to any one of claims 1 to 7, wherein the silicon single crystal substrate having a striation strength of a desired value or less is evaluated as a non-defective product.   9. The silicon single crystal substrate according to claim 1, wherein an arithmetic average Ra of an absolute value of the striation strength is obtained and evaluated using the obtained arithmetic average Ra. Evaluation method.