JP7172878B2 - Method for measuring resistivity of single crystal silicon - Google Patents

Description

The present invention relates to a method for measuring the resistivity of single crystal silicon, and more particularly to a method for measuring the resistivity of a sample wafer cut from a single crystal silicon ingot grown by the Czochralski method (CZ method).

Silicon wafers are widely used as substrate materials for semiconductor devices. In the production of silicon wafers, a single-crystal silicon ingot grown by the CZ method is ground to adjust its diameter, and then the top and tail portions are cut off. Make a silicon block. At this time, a sample wafer (slug) for quality inspection is also cut from both ends of the silicon block at the same time, and the quality of the silicon block is inspected for resistivity, oxygen concentration, carrier recombination lifetime, presence or absence of crystal defects, etc. A pass/fail decision is made.

If the silicon block passes the quality inspection, the silicon block is processed into products. In processing a silicon block, a wire saw is used to slice the silicon block into a plurality of silicon wafers at once. After that, the wafer product is completed through processes such as surface grinding, etching, surface polishing, and cleaning.

Regarding a technique for evaluating a single crystal silicon ingot, for example, in Patent Document 1, a single crystal silicon ingot is cut into blocks with a band saw or the like, sample wafers are cut from both ends of the silicon block, and resistivity, oxygen concentration, crystal defects, etc. are measured. It is described that the pass/fail judgment of the silicon block is performed by evaluating the .

Further, Patent Document 2 describes a first outer circumference grinding process for grinding the outer circumference of a cylindrical ingot to a diameter larger than that of an ingot block for manufacturing wafers, and a cylindrical ingot after the first grinding process. A block cutting step of cutting into a plurality of ingot blocks, a sample cutting step of cutting out silicon inspection samples from the plurality of ingot blocks, a quality evaluation step of performing quality evaluation using the cut out inspection samples, and an ingot block a second peripheral grinding step of grinding the outer periphery to a diameter dimension for the wafer manufacturing process; a notch forming step of forming a notch in the outer periphery of the ingot block after the second peripheral grinding step; and a wafer manufacturing process for slicing the wafer.

Patent Document 3 describes a method for measuring the resistivity of a high-resistivity silicon wafer having a resistivity of 2000 Ωcm or more. In this method of measuring resistivity, the oxide film on the surface of the silicon wafer to be measured is removed by non-aqueous treatment such as buffing after at least two hours have passed after the donor killer heat treatment of the silicon wafer. After that, an electrode needle is brought into contact with the surface to be measured to measure the resistivity. According to this method, the resistivity of the wafer can be accurately measured by preventing the dopant from being inactivated due to the contact of the surface to be measured with the hydrogen ions.

JP 2014-201458 A Japanese Patent No. 6332422 JP 2015-26755 A

In general, a band saw or a slicing machine using an inner peripheral blade or an outer peripheral blade is used to cut out silicon blocks or sample wafers from a single crystal silicon ingot. These cutting machines use a single blade with diamond abrasive grains electrodeposited on the cutting edge, and the ingot is cut by sending this blade from the upper end to the lower end in the radial direction of the single crystal silicon ingot. be.

Sample wafers cut with a blade such as a band saw have deeper scratches on the surface than product wafers cut with a wire saw, and the thickness distribution is uneven. Therefore, conventionally, various inspections are performed after the surface of the sample for quality inspection is removed by etching by immersing the sample for quality inspection in an etchant.

However, when etching is performed on a cut wafer using a blade such as a band saw, although processing strain on the wafer surface can be removed, it is not possible to correct even the waviness of the surface. In addition, in the etching process, since the etching of the outer peripheral portion of the wafer progresses rapidly, the shape of the outer peripheral portion of the wafer is sagging, and the thickness of the outer peripheral portion becomes thin. For this reason, when the resistivity is measured by the four-probe method or the like, the contact between the electrode needle and the wafer surface becomes unstable, resulting in a problem of lowering the accuracy of repeated resistivity measurement. The problem of sagging at the outer periphery of the wafer also occurs when the wafer is polished using a polishing cloth.

Furthermore, when the wafer surface is an etched surface, there is also the problem that the measurement accuracy of the resistivity is lowered due to the surface roughness being too small. Such problems have a greater influence on silicon wafers with higher resistivity, and it is very difficult to accurately measure resistivity.

As a method of suppressing the influence of sagging on the outer periphery of the wafer due to etching, the pulling diameter of the single crystal silicon ingot is made sufficiently larger than the product wafer, and a larger diameter sample wafer cut from this ingot is used for quality evaluation. There is (see patent document 2). However, in this method, the machining allowance for grinding the outer periphery for processing the silicon block to the diameter of the product wafer becomes large, which causes problems such as waste of raw materials and an increase in crystal pulling time.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a resistivity measuring method capable of increasing the repeatability of measuring the resistivity of an inspection sample wafer cut from a single crystal silicon ingot using a band saw or the like.

The present inventors have extensively researched a wafer flattening process that does not rely on etching. It was found that the accuracy of rate measurement was improved. Further, when a sample wafer cut out by a band saw or the like is divided into smaller sizes and then subjected to surface grinding, cracking of the wafer during grinding can be prevented.

The present invention is based on such technical findings, and the method for measuring the resistivity of single crystal silicon according to the present invention includes the steps of radially cutting a single crystal silicon ingot to cut out sample wafers; The method is characterized by comprising a step of cutting out a sample piece, a step of grinding the surface of the sample piece, and a step of measuring the resistivity of the sample piece after the grinding.

According to the present invention, by surface grinding a sample piece cut out from an ingot as it is, warpage and undulation of the surface can be reduced as much as possible, and an appropriate surface roughness can be ensured. Therefore, it is possible to improve the accuracy of repeated measurement of the resistivity of single crystal silicon. Moreover, since the grinding process is performed after cutting the sample wafer into sample pieces, cracking of the wafer during the grinding process can be prevented.

Preferably, the method for measuring the resistivity of single crystal silicon according to the present invention further comprises the step of subjecting the sample piece to donor killer heat treatment before measuring the resistivity of the sample piece. As a result, it is possible to prevent the deterioration of the resistivity measurement accuracy due to the influence of the thermal donor, and in particular, it is possible to improve the resistivity measurement accuracy of a high-resistivity product having a resistivity of 1000 Ωcm or more.

In the method for measuring the resistivity of single crystal silicon according to the present invention, the donor killer heat treatment is preferably performed after the surface of the sample piece is ground. By performing the donor killer heat treatment after grinding the sample piece, the metal impurities adhering to the surface of the sample piece can be removed by grinding, and the etching treatment and cleaning for removing the metal impurities can be omitted. . As a result, it is possible to prevent impurity contamination of the sample piece due to the donor killer heat treatment.

In the present invention, the step of cutting out the sample wafers includes cutting out a silicon block from the single crystal silicon ingot using a band saw, an inner peripheral blade or an outer peripheral blade, and a step of cutting out the sample wafer from the end of the silicon block. is preferably included. In this case, the step of cutting out the sample pieces preferably includes a step of dividing the sample wafer by dicing, and it is particularly preferable to equally divide the wafer into quarters. After cutting out a sample wafer using a blade such as a band saw, when etching is used to remove processing scratches on the surface of the sample wafer, warpage and undulation of the sample wafer remain without being removed, and sagging at the outer periphery of the wafer also occurs. is generated, the accuracy of resistivity measurement is lowered. However, when the resistivity is measured after the surface of the sample piece is ground as described above, it is possible to prevent the accuracy of the resistivity measurement from decreasing due to processing scratches and unevenness (undulation).

In the present invention, the step of grinding the surface of the sample piece is a step of grinding so that the surface roughness Ra (arithmetic mean roughness) of the sample piece is 0.01 μm or more and 0.5 μm or less. preferable. By setting the surface roughness Ra of the sample piece to 0.01 μm or more and 0.5 μm or less, it is possible to prevent a decrease in resistivity measurement accuracy.

In the present invention, the step of grinding the surface of the sample piece is preferably a step of grinding the sample piece having a resistivity of 1000 Ωcm or more so that the surface roughness Ra is 0.1 μm or less. If the surface roughness Ra of the sample piece is 0.1 μm or less, it is possible to prevent a decrease in resistivity measurement accuracy even in a high-resistivity product having a resistivity of 1000 Ωcm or more.

In the present invention, the step of grinding the surface of the sample piece is preferably a step of grinding so that the flatness TTV of the sample piece is 6.0 μm or less, or the flatness Wa is 0.1 μm or less. . As a result, it is possible to suppress warpage, undulation, sagging of the outer peripheral portion, etc. of the sample piece, improve the accuracy of repeated measurement of resistivity, and prevent cracking of the sample piece during grinding.

In the present invention, the step of measuring the resistivity of the sample piece preferably measures the resistivity of the sample piece by a four-probe method. In the measurement of resistivity by the four-probe method, the contact resistance between the electrode needle and the sample piece has a large effect on the measurement result. descend. The contact resistance changes depending on the surface condition of the sample piece, the wear condition of the tip of the electrode needle, the deviation of the contact angle between the electrode needle and the sample piece from the perpendicular, etc. The higher the resistivity of the sample piece, the greater the change in contact resistance. . However, when the resistivity is measured after grinding the sample piece, the contact state between the electrode needle and the sample surface can be improved to improve the repeatability of the resistivity measurement.

ADVANTAGE OF THE INVENTION According to this invention, the resistivity measuring method which can improve the repeatable measurement precision of the resistivity of the sample wafer for an inspection cut out from the single crystal silicon ingot using a band saw etc. can be provided.

FIG. 1 is a flow chart showing a method for measuring the resistivity of single crystal silicon according to the first embodiment of the present invention. FIG. 2 is a schematic diagram for explaining the method of measuring the resistivity of single crystal silicon according to the first embodiment. FIG. 3 is a flow chart showing a resistivity measuring method for single crystal silicon according to a second embodiment of the present invention. FIG. 4 is a graph showing the relationship between the surface roughness Ra of a silicon wafer and the CV value of resistivity. FIGS. 5A and 5B are graphs showing the relationship between the surface roughness Ra of silicon wafers and resistivity.

Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

FIG. 1 is a flow chart showing a method for measuring the resistivity of single crystal silicon according to the first embodiment of the present invention. FIG. 2 is a schematic diagram for explaining the method for measuring the resistivity of single crystal silicon according to the first embodiment.

As shown in FIGS. 1 and 2, the method for measuring the resistivity of single-crystal silicon according to the present embodiment includes cutting out a sample piece 13 from a single-crystal silicon ingot 10, flattening the sample piece 13, and then making a sample. It is characterized by measuring the resistivity of the piece 13 .

Therefore, first, a single crystal silicon ingot 10 is prepared (step S11). The single crystal silicon ingot 10 is grown by the CZ method, and includes a top portion 10a having a gradually increasing crystal diameter, a straight body portion 10b having a substantially constant crystal diameter, and a tail portion 10c having a gradually decreasing crystal diameter. and In this embodiment, the single-crystal silicon ingot 10 is preferably a low resistance product with a resistivity of 0.1 Ωcm or less, but may be a normal resistance product with a resistivity of 0.1 Ωcm or more and 1000 Ωcm or less. When manufacturing a silicon wafer (product wafer) with a diameter of 300 mm, the diameter of the straight body portion 10b of the ingot is 308 mm or more.

In the processing of the single crystal silicon ingot 10, first, the straight body portion 10b of the ingot is radially cut to cut out the sample wafer 12 (step S12). Specifically, after cutting off the top portion 10a and the tail portion 10c of the ingot and processing it into a cylindrical shape, the outer circumference is ground to adjust the diameter (for example, 304 mm) to be sufficiently larger than that of the product wafer. Next, a plurality of silicon blocks 11 are cut out from the ingot by alternately performing a block cutting step of cutting out silicon blocks 11 from a cylindrical ingot and a wafer cutting step of cutting out sample wafers 12 from both ends of each silicon block 11. A sample wafer 12 is cut. For cutting out the silicon block 11 and the sample wafer 12, a cutting machine using a band saw, an inner peripheral blade, an outer peripheral blade, or the like is used. In the wafer cutting step, two or more sample wafers may be continuously cut.

Next, each sample wafer 12 is diced to produce a quarter size sample piece 13 (step S13). Although the sample wafer 12 is divided into four parts in this embodiment, the number of divisions is not particularly limited, and any number of divisions of two or more may be used. One of the thus-fabricated fan-shaped sample pieces 13 is used as a sample for measuring the resistivity, which is one of the quality indexes of silicon wafers. Other sample pieces are used as samples for evaluating oxygen concentration, carrier recombination lifetime, crystal defects, and the like.

Next, the surface of the sample piece 13 used for resistivity measurement is ground (step S14). Grinding is a type of mechanical processing in which a grindstone rotating at high speed is pressed against a surface to be processed to remove the surface layer and obtain a smooth surface. The grinding process is preferably a single-sided surface grinding process in which one side is ground one by one, but is not particularly limited as long as it is a mechanical process that can flatten the surface of the sample piece. good too.

The surface of the wafer cut by the band saw is a cut surface with a rough cut edge, and there are deep processing scratches and irregularities. Therefore, when the resistivity of a sample piece is measured by, for example, the four-probe method, the measured value varies due to the poor surface shape. However, when the sample piece is flattened by grinding, processing scratches and irregularities (undulations) on the surface can be removed, so that the resistivity measurement accuracy can be improved.

It is desirable to grind the sample piece on both sides of the sample piece. As a result, it is possible to remove metal impurities adhered during the cutting process of the wafer, and to prevent fluctuations in resistivity due to metal contamination of the sample piece, which is a particular problem when the donor killer heat treatment described later is performed. However, if etching is performed in advance to remove surface contaminants, it is also possible to grind only one side of the object to be measured.

When the sample piece is ground, it is desirable that the surface roughness Ra (arithmetic mean roughness) of the surface to be processed of the sample piece is 0.01 μm or more and 0.5 μm or less. This is because if the surface roughness Ra is greater than 0.5 μm, processing scratches and the like cannot be sufficiently removed. This is because the measurement accuracy decreases.

When grinding the sample piece, the flatness Wa (arithmetic mean waviness) of the surface to be processed of the sample piece should be 0.1 μm or less, or the flatness TTV (Total Thickness Variation) should be 6.0 μm or less. is preferred. The flatness TTV is one of the flatness indices also called GBIR (Global Backside Ideal Range), and is the maximum and minimum thickness of the flatness application area (distance from the back surface reference plane) when the wafer is fixed by suction. Defined as the difference in values. In this way, by ensuring a certain level of flatness, it is possible to remove warpage and undulation of the sample piece and improve the accuracy of resistivity measurement. Further, by removing the warp and undulation of the sample piece, cracking of the sample piece, which will be described later, can be prevented.

The grinding process is performed on fan-shaped sample pieces 13 obtained by dicing the circular sample wafer 12 . As described above, the surface of the sample wafer 12 cut by the band saw has processing scratches and unevenness (undulations). Therefore, when the sample wafer 12 having a large area is subjected to grinding processing, the stress caused by the processing scratches and the like is increased. Wafer cracking is likely to occur due to concentration. However, when the sample wafer 12 is processed to a small size in advance and then ground, cracking of the wafer can be prevented.

In the grinding process, after setting a silicon wafer sample piece 13 cut to a quarter size, for example, on a suction stage 21 of a single-sided surface grinder 20, the grindstone 22a of a grinding head 22 is pressed against the upper surface of the sample piece 13. Rotate while grinding. When grinding both sides of the sample piece 13, after grinding one surface, the sample piece 13 is turned over and the opposite surface is ground. In order to remove processing scratches and irregularities on the surface, the grinding amount of the sample piece 13 is preferably 50 μm or more, particularly preferably 70 μm or more, per side.

Next, the ground sample piece 13 is washed (step S15). As a cleaning method, ultrasonic cleaning or hydrofluoric acid cleaning can be used. As a result, metal impurities adhering to the surface of the sample piece 13 can be removed, and fluctuations in resistivity due to impurity contamination can be prevented.

After that, the resistivity of the sample piece 13 is measured (step S16). The resistivity of the sample piece 13 is preferably measured by the four-probe method according to JIS_H_0602-1995, but may be measured by other methods such as the spreading resistance method.

In the four-probe method, four electrode needles arranged in a straight line are brought into pressure contact with the surface of the object to be measured. is measured, and the resistivity is calculated from this potential difference and the distance between the pair of measuring electrode needles. If the surface of the sample piece has large irregularities (undulations), the contact between the electrode needle and the sample surface becomes unstable, resulting in a large error in the resistivity measurement value. However, in this embodiment, since the surface of the sample piece is flattened by grinding, the electrode needle can be reliably brought into contact with the surface of the sample piece. In addition, since the surface has moderate roughness as compared with the case of etching, the variation in contact resistance between the electrode needle and the surface of the sample piece can be reduced, and the accuracy of repeated measurement of resistivity can be improved.

The sample piece 13 that has been flattened by grinding has much less sag at the outer periphery than a conventional sample piece that has been etched, so the resistivity can be accurately measured from the center to the outermost periphery of the wafer. It is possible. In the conventional method of measuring the resistivity of a sample wafer, it was necessary to prepare a wafer with a sufficiently large diameter in anticipation of an increase in peripheral sagging due to etching. However, when the outer peripheral sag is very small as in this embodiment, the resistivity of the outer peripheral portion of the wafer can be accurately measured without preparing a wafer with a sufficiently large diameter. Therefore, it is possible to reduce the machining allowance when grinding the outer circumference of the silicon block, thereby preventing waste of raw materials and an increase in crystal pulling time.

As a result of measuring the resistivity of the sample piece 13, if the resistivity satisfies a predetermined condition, the sample piece is accepted as an acceptable product in terms of resistivity. Other quality items such as oxygen concentration are also inspected, and when all the quality items are passed, the silicon block 11 from which the sample wafer is cut is also certified as a passed product and delivered to the subsequent process. In the post-process, the outer periphery of the silicon block 11 is ground so as to have the diameter of the product wafer, a notch or an orientation flat is formed, and then slicing using a wire saw is performed to obtain a plurality of silicon wafers from the silicon block. cut out at the same time. After that, each silicon wafer is subjected to surface grinding, lapping, etching, double-side polishing, single-side polishing, cleaning, and the like to complete the silicon wafer.

As described above, in the method for measuring the resistivity of single-crystal silicon according to the present embodiment, the surface of a sample wafer cut out from a silicon ingot by a band saw or the like is ground, and then the ground surface is subjected to resistivity measurement. The measurement accuracy of resistivity can be improved. In addition, since the sample wafer cut out from the silicon ingot is divided into smaller sizes and then ground, cracking of the sample during grinding can be prevented.

FIG. 3 is a flow chart showing a resistivity measuring method for single crystal silicon according to a second embodiment of the present invention.

As shown in FIG. 3, the method for measuring the resistivity of single-crystal silicon according to the present embodiment is characterized by being particularly suitable for measuring high resistance products with a resistivity of 1000 Ωcm or more, and measuring the resistivity of a sample piece. The point is that the donor killer heat treatment (step S17) is performed before the heat treatment. Other steps are the same as in the first embodiment.

Single crystal silicon grown by the CZ method contains supersaturated oxygen, and when heat-treated at a low temperature of around 450° C., several oxygen atoms aggregate to form oxygen clusters, which act as thermal donors that emit electrons. Therefore, it becomes a cause of lowering the measurement accuracy of the resistivity. However, by extinguishing the thermal donors by the donor killer heat treatment, it is possible to suppress fluctuations in resistivity due to the influence of the thermal donors, and to improve the accuracy of resistivity measurement.

The donor killer heat treatment is better to be performed when measuring a normal resistance product with a resistivity of 0.1 Ωcm or more and 1000 Ωcm or less, and is essential when measuring a high resistance product with a resistivity of 1000 Ωcm or more. On the other hand, when measuring a low-resistance product with a resistivity of 0.1 Ωcm or less, the resistivity fluctuation due to the influence of oxygen donors can be ignored, so the donor killer heat treatment is unnecessary, and the resistivity is measured according to the first embodiment. It is sufficient to measure by the method.

The donor killer heat treatment is a short-time heat treatment performed in an inert gas atmosphere at 600° C. to 700° C. in order to eliminate oxygen donors, and the heat treatment time is 10 minutes or more and 240 minutes or less. In the case of lamp heating, it is also possible to set the heat treatment time to 1 second. NO donors also occur in nitrogen-doped silicon wafers. The donor killer heat treatment for annihilating the NO donors is heat treatment at a temperature of 1000° C. to 1200° C. for 30 minutes or more and 240 minutes or less.

Before the resistivity of the sample piece 13 is measured, the donor killer heat treatment may be performed after grinding the sample piece as in the present embodiment, or before grinding the sample piece. However, if the donor killer heat treatment is performed before the sample piece is ground, it is necessary to previously remove processing scratches and metal impurities by hard etching, which requires a large removal amount. This is because if the wafer surface has processing scratches or warpage, the wafer may crack during the donor killer heat treatment, and metal impurities adhered during sample cutting may diffuse into the wafer due to the donor killer heat treatment.

On the other hand, when the donor killer heat treatment is performed after grinding as in the present embodiment, there is no need to perform the hard etching as described above, and the surface of the ground sample piece should be normalized by chemical cleaning before the donor killer heat treatment. Enough. That is, when the donor killer heat treatment is performed after grinding, the process can be shortened and the cost can be reduced.

When measuring the resistivity of the sample wafer 12 with a high resistivity of 1000 Ωcm or more, it is desirable to grind the sample piece 13 so that the surface roughness Ra is 0.01 μm or more and 0.1 μm or less. When measuring the resistivity of the high-resistivity sample piece 13 by the four-probe method, the contact resistance between the electrode needle and the sample piece 13 has a large effect on the measurement result, and the surface roughness Ra is greater than 0.1 μm. This is because, in some cases, the accuracy of repeated measurement of resistivity is lowered due to the influence of contact resistance. The contact resistance changes depending on the surface condition of the sample piece 13, the wear condition of the tip of the electrode needle, and the deviation of the contact angle between the electrode needle and the sample piece 13 from the perpendicular. growing. Also, as described above, when the surface roughness Ra is less than 0.01 μm, the surface roughness is too small, and the resistivity measurement accuracy decreases.

As described above, from the viewpoint of preventing contamination of the sample piece during the donor killer heat treatment, it is preferable to grind both surfaces of the sample piece. However, if the etching treatment is performed in advance before the donor killer heat treatment, it is also possible to grind only one side to be measured for the resistivity. Since the metal impurities adhering to the surface of the sample piece are removed by etching, the grinding process for the sole purpose of removing metal impurities can be omitted.

As described above, in the method for measuring the resistivity of single crystal silicon according to the present embodiment, the donor killer heat treatment is performed before the resistivity of the sample piece is measured. A remarkable effect can be obtained in the resistivity measurement of a high resistance product with a resistivity of 1000 Ωcm or more. In addition, in the present embodiment, since the donor killer heat treatment is performed after the sample piece is ground, metal impurity contamination accompanying the donor killer heat treatment can be prevented.

Although the preferred embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention. Needless to say, it is included within the scope of the invention.

The influence of the surface roughness Ra of the ground silicon wafer on the resistivity measurement accuracy was evaluated. First, a single crystal silicon ingot for a 300 mm wafer grown by the CZ method was prepared. For the diameter of the ingot, a sample wafer for inspection was cut from each single crystal silicon ingot, and the sample wafer was further divided into four to produce fan-shaped sample pieces #1 to #8. After that, both surfaces of each sample piece #1 to #8 were ground with a single-sided surface grinder, and the surface roughness was measured with a surf tester (SJ-400).

Next, the resistivity distribution in the radial direction of these sample pieces #1 to #8 was measured by the four-probe method. The resistivity measurement points are the wafer center position (0 mm), the radial middle position of the wafer (R/2 mm, where R is the wafer diameter), the position 10 mm radially inward from the outermost periphery of the wafer (R-10 mm), Five points were set at a position 5 mm radially inward from the outermost periphery of the wafer (R-5 mm) and a position 3 mm radially inward from the outermost periphery of the wafer (R-3 mm).

The measurement of the resistivity distribution of the sample pieces #1 to #8 was repeated 20 times, and the average value and standard deviation of the resistivity at each measurement point were obtained. Furthermore, the coefficient of variation CV (standard deviation/average value×100), which is an index of the accuracy of repeated measurement of resistivity, was determined, and the average value of the CV values at each measurement point was determined.

As a result, the average CV value of sample piece #1 with surface roughness Ra = 0.01 µm was about 1.1%, and the average CV value of sample piece #2 with surface roughness Ra = 0.02 µm was about The average CV value of the sample piece #3 with a surface roughness Ra of 0.1 μm is about 0.3%, and the average CV value of the sample piece #4 with a surface roughness Ra of 0.25 μm is 0.6%. is about 0.45%, the average value of the CV value of sample piece #5 with surface roughness Ra = 0.45 µm is about 0.6%, and the CV value of sample piece #6 with surface roughness Ra = 0.48 µm The average value is about 0.7% The average value of the CV value of sample piece #7 with surface roughness Ra = 0.6 µm is about 1.2%, and the CV value of sample piece #8 with surface roughness Ra = 0.8 µm was about 1.6.

FIG. 4 is a graph showing the relationship between the surface roughness Ra of a silicon wafer and the CV value of resistivity.

As is clear from FIG. 4, it was found that the CV value of the resistivity measurement value can be reduced to 1% or less if the surface roughness Ra of the wafer is within the range of 0.02 to 0.5 μm. On the other hand, it has been found that when the surface roughness Ra of the wafer is 0.6 μm or more, the surface roughness is too large and the precision of repeated measurement of resistivity deteriorates, resulting in a CV value of 1% or more. Furthermore, it was found that when the surface roughness Ra of the wafer was 0.01 μm, the surface roughness was too small and the precision of repeated measurement of resistivity deteriorated, resulting in a CV value of 1% or more.

Next, the effect of the surface roughness Ra on the repeatability of resistivity measurement was evaluated when a high resistance silicon wafer having a resistivity of 1000 Ωcm or more was ground. First, a single crystal silicon ingot for a 150 mm wafer grown by the CZ method so as to have a resistivity of about 7000 Ωcm was prepared, sample wafers were cut, 1/4 sample pieces were cut, sample pieces were ground, and donors were processed. A killer heat treatment was performed in order, and the resistivity distribution of the sample piece was measured. A #6000 grindstone was used in the grinding process. The donor killer heat treatment was heat treatment at 650° C. for 30 minutes.

5A and 5B are graphs showing the relationship between the surface roughness Ra and resistivity of silicon wafers, in which the horizontal axis is the distance (mm) from the center of the wafer, and the vertical axis on the right side is the surface roughness. The height Ra (μm) and the vertical axis on the left side indicate the resistivity (%). In addition, the dot graph in the figure is the measured value of resistivity, and the solid line graph is the measured value of surface roughness Ra.

As shown in FIG. 5(a), when the surface roughness Ra of the wafer is within the range of 0.11 to 0.15 μm, the measured values of resistivity vary greatly within the range of 7000 to 25000 Ωcm. became.

On the other hand, as shown in FIG. 5B, when the surface roughness Ra of the wafer is within the range of 0.05 to 0.08 μm, the variation in the measured values of resistivity is small, and more than half of the measured values was about 7000Ωcm.

From the above results, by setting the surface roughness Ra of the wafer to 0.1 or less, the variation in the measured value of resistivity can be reduced, and even a high resistance product with a resistivity of 1000 Ωcm or more can be repeatedly measured. It has been found that the accuracy can be improved.

10 single-crystal silicon ingot 10 straight body portion 10a top portion 10b straight body portion 10c tail portion 11 silicon block 12 sample wafer 13 sample piece 20 single-sided surface grinder 21 suction stage 22 grinding head 22a grindstone S11 ingot preparation step S12 sample wafer cutting step S13 Sample piece cutting step S14 Grinding step S15 Cleaning step S16 Resistivity measurement step S17 Donor killer heat treatment

Claims (6)

a step of cutting a silicon block from a single crystal silicon ingot using a band saw, an inner peripheral blade or an outer peripheral blade;
a step of radially cutting the silicon block using a band saw, an inner peripheral blade, or an outer peripheral blade to cut out a sample wafer from an end portion of the silicon block ;
cutting the sample wafer into fan-shaped sample pieces;
a step of grinding the surface of the sample piece;
and measuring the resistivity of the sample piece after the grinding process. 2. The resistivity measurement of single crystal silicon according to claim 1, further comprising the step of subjecting the sample piece to donor killer heat treatment after grinding the surface of the sample piece and before measuring the resistivity of the sample piece. Method. The step of grinding the surface of the sample piece includes:
3. The method for measuring resistivity of single crystal silicon according to claim 1 , wherein the step of grinding the sample piece so that the surface roughness Ra of the sample piece is 0.01 [mu]m or more and 0.5 [mu]m or less. The step of grinding the surface of the sample piece includes:
4. The method for measuring resistivity of single crystal silicon according to claim 3 , wherein the step of grinding a sample piece having a resistivity of 1000 Ωcm or more so that the surface roughness Ra is 0.1 μm or less. The step of grinding the surface of the sample piece includes:
3. The method of measuring resistivity of single crystal silicon according to claim 1 , wherein the step of grinding the sample piece so that the flatness TTV of the sample piece is 6.0 [mu]m or less. 3. The method for measuring resistivity of single crystal silicon according to claim 1 , wherein the step of measuring the resistivity of said sample piece measures the resistivity of said sample piece by a four-probe method.

Images