JP5141495B2 - Silicon single crystal manufacturing method and silicon wafer - Google Patents

Description

  The present invention relates to a method for producing a silicon single crystal, and more particularly to a method for producing a silicon single crystal that is grown by the Czochralski method (CZ method) and is suitably used for a substrate of a semiconductor device. The present invention also relates to a silicon wafer, and more particularly to a silicon wafer that is cut out from a silicon single crystal ingot grown by the Czochralski method and is suitable as a substrate for a semiconductor device.

  When a silicon single crystal is grown by the Czochralski method, the type and distribution of defects contained in the crystal depend on the ratio of the crystal pulling speed V and the temperature gradient G in the growth direction in the silicon single crystal.

  FIG. 14 is a diagram showing a general relationship between V / G and the type and distribution of defects.

  As shown in FIG. 14, when V / G is large, vacancies become excessive, and microvoids (generally called COP) that are aggregates of vacancies are generated. On the other hand, when V / G is small, interstitial silicon atoms become excessive, and dislocation clusters, which are aggregates of interstitial silicon, are generated. Therefore, in order to produce a crystal containing neither COP nor dislocation clusters, it is necessary to control V / G so as to fall within an appropriate range in the crystal radial direction and length direction. First, since V is constant at any position in the radial direction of the crystal, the structure of the high temperature portion (hot zone) in the CZ furnace must be designed so that the temperature gradient G falls within a predetermined range. Next, with respect to the length direction of the crystal, G depends on the pulling length of the crystal. Therefore, in order to keep V / G within a predetermined range, V must be changed in the length direction of the crystal. At present, even a silicon single crystal having a diameter of 300 mm has been mass-produced by controlling V / G and containing neither COP nor dislocation clusters.

  As described above, COPs pulled by controlling V / G and silicon wafers that do not contain dislocation clusters are mass-produced and used for manufacturing electronic devices. However, these wafers are never uniform over the entire surface, and include a plurality of regions that behave differently when heat-treated. As shown in FIG. 14, there are three regions in order from the largest V / G, the OSF region, the Pv region, and the Pi region, between the region where COP occurs and the region where dislocation clusters occur. The OSF region contains plate-like oxygen precipitates (OSF nuclei) in an as-grown state (a state in which no heat treatment is performed after crystal growth), and at a high temperature (generally 1000 ° C. to 1200 ° C.). This is a region where an OSF (Oxidation Induced Stacking Fault) occurs when thermal oxidation occurs. The Pv region is an area where oxygen precipitate nuclei are contained in an as-grown state, and oxygen precipitates are likely to be generated when two-stage heat treatment is performed at low and high temperatures (for example, 800 ° C. and 1000 ° C.). . The Pi region is a region that hardly contains oxygen precipitation nuclei in an as-grown state and hardly generates oxygen precipitates even when heat treatment is performed.

  Since the difference between V / G at which COP begins to occur and V / G at which dislocation clusters begin to occur is very small, strict control of V is required to produce a crystal that does not contain COPs or dislocation clusters. However, even if the crystal is pulled at the target V, COP and dislocation clusters may occur due to various factors. This is due to the following reason.

  The CZ furnace is composed of members such as a carbon heater, a heat insulating material, and a carbon crucible. These members are continuously used over several tens to hundreds of times. In addition, these members are deteriorated and thinned over time by reaction with silicon melt vapor or droplets, reaction with silicon melt and gas generated from carbon, reaction of quartz crucible, etc. The thermal properties of the inner hot zone also change over time. When such a time-dependent change in the hot zone occurs, the temperature gradient G changes, so that V / G deviates from the design value even if the crystal is pulled up with the target V. For this reason, COP and dislocation clusters are generated even if the crystal is pulled up with the target V.

  Therefore, in order to realize the target V / G, it is necessary to change the profile of the pulling-up speed V in accordance with the change with time of the hot zone.

  Conventionally, a pulling speed profile is set so as to include the OSF region, and a sample cut out from the pulled crystal is subjected to Cu (copper) decoration or heat treatment for OSF evaluation to evaluate the width of the OSF region. Based on this, the speed profile of the subsequent pulling was adjusted (see Patent Documents 1 and 2). That is, if the OSF region is wide, the CZ furnace changes in a direction in which V / G increases (G decreases). Therefore, in subsequent pulling, V is set lower, and conversely if the OSF region is narrow, CZ Since the furnace changes in a direction in which V / G becomes smaller (G becomes larger), V is set higher in subsequent pulling.

  Since these methods adjust the speed profile of subsequent pulling using the size and position of the OSF region as an index, the wafers shipped as products necessarily include the OSF region. So far, the OSF region does not seem to affect the electronic device. However, since the OSF region is a region containing OSF nuclei, that is, plate-like oxygen precipitates even in the as-grown state, it is highly likely to cause deterioration in the characteristics of future electronic devices. Therefore, in the future, it is considered necessary to develop a method for stably pulling up a crystal not including the OSF region without using the width of the OSF region as an index for adjusting the pulling rate.

  As a method that does not use the OSF region as an index for adjusting the pulling speed, the vacancy concentration in the crystal is estimated from the magnitude of the decrease (softening) of the elastic constant of silicon accompanying the cryogenic temperature, and the speed profile of subsequent pulling is calculated. A method of adjusting has been proposed (see Patent Document 3). However, in order to implement this method, a wafer cut out from a silicon single crystal is etched to remove processing strain, ZnO or AlN to be a thin film vibrator is deposited, and an external magnetic field is applied as necessary. In this state, while cooling in a temperature range of 25K (−248 ° C.) or lower, an ultrasonic pulse is propagated, and a change in the sound speed of the propagated ultrasonic pulse is detected. From this change in sound speed, elasticity accompanying a decrease in cooling temperature is detected. It is necessary to take a procedure such as calculating the amount of decrease of the constant and evaluating the concentration of vacancies existing in the silicon wafer from the calculated amount of decrease of the elastic constant. For this reason, expensive evaluation equipment and complicated procedures are required, and it cannot be applied to routine inspection in the manufacturing process of silicon single crystals.

  Evaluation methods based on various principles have been proposed as methods for detecting crystal defects in a silicon single crystal. The wet selective etching method, which has been used for a long time, is an uneven surface on which crystal defects are etched by immersing a sample in a mixed solution of a substance that has an oxidizing action on silicon and a substance that dissolves the oxide. It is a method that manifests as (in many cases etch pits). Nitric acid, chromic acid or the like is used as a substance having an oxidizing action, and hydrofluoric acid is used as a substance having an action of dissolving an oxide. Depending on the type of chemical substance used and its mixing ratio, the normal silicon / defect selection ratio differs, and the sensitivity and the types of detectable defects differ. The wet selective etching method has a lower sensitivity than other methods, but is simple and is still used for evaluating crystal defects. As typical etching liquids, light liquid, seco liquid, dash liquid, etc., all named after the inventor, can be mentioned.

  The infrared tomography method that has been generally used since the 1990s is a method that utilizes the difference in refractive index between silicon and defects. Since infrared rays pass through silicon, defects inside the wafer can be evaluated. This method is characterized by higher sensitivity to oxygen precipitates and COP than the wet selective etching method.

Patent Document 4 describes a defect detection method using reactive ion etching (RIE). This method is a method in which RIE is performed on a sample under a condition with a high Si / SiO ₂ selection ratio after revealing oxygen precipitates such as BMD by heat treatment. As a result, oxygen precipitates (SiO ₂ ) are not etched but are manifested as protrusions. It has been reported that if conditions with a high Si / SiO ₂ selection ratio are selected, defect evaluation with higher sensitivity than infrared tomography is possible.
JP 2005-194186 A International Publication No. 99/40243 Pamphlet JP 2007-261935 A JP 2000-58509 A

  However, Patent Document 4 only reports that it is possible to evaluate oxygen precipitates such as BMD that are manifested by heat treatment, and does not describe evaluation of an as-grown silicon wafer. In the state of the art at the time when Patent Document 4 was filed, only the OSF nucleus was regarded as a defect included in the silicon wafer in the as-grown state, and the OSF nucleus is easily manifested by thermal oxidation. Therefore, it is considered that there was no significance in detecting this in the as-grown state. Further, at the time of the application of Patent Document 4, it was unclear whether the Pv region contained an oxygen precipitation nucleus in the as-grown state.

  The present invention has been made in view of the above circumstances, and without using the position and area of the OSF region as an index, the region included in the pulled crystal is specified by a simple method, and the subsequent pulling is performed. By adjusting the velocity profile, an object is to produce a silicon single crystal that does not contain COP regions or dislocation cluster regions, and preferably does not contain OSF regions. Moreover, it aims at providing the silicon wafer cut out from the silicon single crystal manufactured in this way.

  The present inventor has intensively studied what kinds of defects can be revealed when reactive ion etching is performed on a silicon wafer in an as-grown state. As a result, it has been clarified that when reactive ion etching is performed on an as-grown silicon wafer, OSF nuclei contained in the OSF region and oxygen precipitate nuclei contained in the Pv region become apparent. This means that the boundary between the Pv region and the Pi region can be determined. Therefore, it is possible to change the profile of the pulling-up speed V according to the time-dependent change of the hot zone by using the position and width of the Pv area as an index instead of using the position and area of the OSF area as an index. It is considered possible.

  The method for producing a silicon single crystal according to the present invention is based on such technical knowledge, and a growing step of growing a silicon single crystal ingot that does not contain COP and dislocation clusters by the Czochralski method, By cutting out a silicon wafer from a silicon single crystal ingot and performing reactive ion etching on the silicon wafer in the as-grown state, the grown-in defects including silicon oxide are revealed as protrusions on the etched surface. And an etching process, wherein the growth conditions in the subsequent growth process are adjusted based on the generation region of the protrusions revealed in the etching process.

  According to the present invention, a crystal according to the current V / G can be obtained without performing a time-consuming process such as a heat treatment for revealing defects or a measurement of a decrease amount due to cryogenic temperature reduction of silicon elastic constant. You can quickly know the condition. For this reason, if the subsequent growth conditions are adjusted (feedback) based on the obtained information, the profile of the pulling speed V can be quickly changed. In addition to feedback adjustment of the subsequent growth conditions, it is possible to determine whether the silicon single crystal ingot is acceptable or not.

  In the growth step, it is preferable to grow a silicon single crystal ingot that does not contain OSF nuclei. In the present invention, since V / G detection is performed without using the OSF region as an index, it is suitable for manufacturing a silicon single crystal ingot that does not contain OSF nuclei.

  In the etching step, it is preferable to etch the silicon wafer with an aqueous solution containing hydrofluoric acid and nitric acid and to perform reactive ion etching on the etched surface. According to this, since the processing for the silicon wafer is very simple, it becomes possible to perform feedback of the pulling speed V more quickly.

  In the etching step, the silicon wafer may be cleaved and reactive ion etching may be performed on the cleaved surface. This is because the cleaved surface of the silicon wafer has the same characteristics as an etched surface obtained by etching with an aqueous solution containing hydrofluoric acid and nitric acid.

  Further, in the etching step, the silicon wafer may be mirror-finished and reactive ion etching may be performed on the mirror-finished surface. If the silicon wafer is mirror-finished, defects due to disturbance are almost completely removed, so that it is possible to more accurately evaluate the region where the protrusion is generated.

  The silicon wafer according to the present invention is a silicon wafer which is cut from a silicon single crystal ingot grown by the Czochralski method and does not contain any of COP, OSF nucleus and dislocation cluster, and is reactive in an as-grown state. When a grown-in defect containing silicon oxide is exposed as a protrusion on the etched surface by performing ion etching, the area where the protrusion is exposed becomes a disk shape and / or a ring shape. And Such a silicon wafer can be obtained by the method for producing a silicon single crystal according to the present invention.

  As described above, according to the present invention, the current V / G can be obtained without performing a time-consuming process such as a heat treatment for revealing a defect or a measurement of a decrease amount associated with a cryogenic temperature reduction of silicon. It is possible to quickly know the crystal state according to the above. As a result, it is possible to feed back to the most recent batch, so that it is possible to mass-produce crystals containing neither COP nor dislocation clusters with high accuracy. Moreover, since the OSF region is not used as an index, it is possible to obtain a crystal not including the OSF region as well as the COP and the dislocation cluster.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a schematic diagram showing a configuration of a pulling apparatus applicable to a method for producing a silicon single crystal according to a preferred embodiment of the present invention.

  A silicon single crystal pulling apparatus 10 shown in FIG. 1 includes a chamber 11, a support rotary shaft 12 that passes through the center of the bottom of the chamber 11 and is provided in the vertical direction, and a graphite susceptor fixed to the upper end of the support rotary shaft 12. 13, a quartz crucible 14 accommodated in the graphite susceptor 13, a heater 15 provided around the graphite susceptor 13, a support shaft drive mechanism 16 for moving the support rotation shaft 12 up and down, and a seed crystal Heating of the silicon single crystal ingot 20 by the radiation heat from the seed chuck 17 to be held, the pulling wire 18 for suspending the seed chuck 17, the wire winding mechanism 19 for winding the wire 18, and the heater 15 and the quartz crucible 14. And a heat shielding part for suppressing temperature fluctuation of the silicon melt 21 22, and a control unit 23 that controls each unit.

  A gas inlet 24 for introducing Ar gas into the chamber 11 is provided in the upper part of the chamber 11. Ar gas is introduced into the chamber 11 from the gas introduction port 24 through the gas pipe 25, and the introduction amount is controlled by the conductance valve 26.

  A gas discharge port 27 for exhausting Ar gas in the chamber 11 is provided at the bottom of the chamber 11. Ar gas in the sealed chamber 11 is discharged from the gas outlet 27 through the exhaust pipe 28 to the outside. A conductance valve 29 and a vacuum pump 30 are provided in the middle of the exhaust gas pipe 28, and the pressure inside the chamber 11 is reduced by controlling the flow rate with the conductance valve 29 while sucking the Ar gas in the chamber 11 with the vacuum pump 30. The state is maintained.

  Further, a magnetic field supply device 31 is provided outside the chamber 11. The magnetic field supplied from the magnetic field supply device 31 may be a horizontal magnetic field or a cusp magnetic field.

  2A is a diagram showing the relationship between the pulling speed V of the silicon single crystal ingot 20 and the type and distribution of defects, and FIGS. 2B to 2D are respectively B1 shown in FIG. It is sectional drawing along the -B1 line, the C1-C1 line, and the D1-D1 line. The pulling condition in FIG. 2 is a pulling condition in which the OSF region 42 appears in a disk shape, and is a case where V / G near the center is larger (G is smaller) than the outer periphery of the crystal.

  When the pulling speed V is set to a speed corresponding to the B1-B1 line under the pulling conditions in FIG. 2, the OSF region 42 and the Pv region 43 are formed in the cut silicon wafer 40 as shown in FIG. Appears as a disk. More specifically, the OSF region 42 appears at the center of the silicon wafer 40, the Pv region 43 appears outside thereof, and the outside thereof becomes the Pi region 44. The diameter of the OSF region 42 is the diameter D0, and the diameter of the Pv region 43 is the diameter D1. As described above, when the pulling speed V is set to a speed corresponding to the B1-B1 line, the silicon single crystal ingot 20 that does not include the COP 41 and the dislocation cluster 45 is obtained, but the disk-shaped OSF region 42 is formed in the central axis portion. It is formed. Therefore, in order to obtain a crystal not including the OSF region 42 as well as the COP 41 and the dislocation cluster 45, it is necessary to lower the pulling rate V by the control device 23. In addition, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is rejected.

  Further, when the pulling speed V is set to a speed corresponding to the C1-C1 line under the pulling condition of FIG. 2, as shown in FIG. 2C, a Pv region having a diameter D22 at the center of the cut silicon wafer 40. 43 appears, and the outside thereof is the Pi region 44. Here, the diameter D2 of the Pv region 43 is smaller than the diameter D1 of the Pv region 43 shown in FIG. 2B (D2 <D1). Thus, when the pulling speed V is set to a speed corresponding to the C1-C1 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In addition, since a sufficient margin is secured for the pulling speed V at which the dislocation clusters 45 are generated, it is not necessary to change the pulling speed V by the control device 23. In addition, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  Further, when the pulling speed V is set to a speed corresponding to the D1-D1 line under the pulling condition of FIG. 2, a Pv region 43 having a diameter D3 appears at the center of the silicon wafer 40 as shown in FIG. All the outsides are Pi regions 44. Here, the diameter D3 of the Pv region 43 is further smaller than the diameter D2 of the Pv region 43 shown in FIG. 2C (D3 <D2). Thus, when the pulling speed V is set to a speed corresponding to the D1-D1 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In this case, it is necessary to increase the pulling speed V by the control device 23 because there is only a small margin left with respect to the pulling speed V at which the above occurs. However, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  As described above, in the pulling condition shown in FIG. 2, it is possible to determine the control direction of the pulling speed V based on the diameter of the disk-like Pv region 43. It is also possible to make a pass / fail judgment for the silicon single crystal ingot 20.

  FIG. 3A is a diagram showing the relationship between the pulling speed V of the silicon single crystal ingot 20 and the type and distribution of defects under a condition in which the radial distribution of the temperature gradient G is different from that in FIG. (B)-(d) is sectional drawing along the B2-B2 line | wire, C2-C2 line | wire, and D2-D2 line | wire which are shown to Fig.3 (a), respectively. The pulling condition in FIG. 3 is a pulling condition in which the OSF region 42 appears in a disk shape and a ring shape, and is a case where V / G is large (G is small) in the vicinity of the center of the crystal and the outer periphery.

  When the pulling speed V is set to a speed corresponding to the line B2-B2 under the pulling conditions of FIG. 3, the OSF region 42 and the Pv region 43 are formed in the cut silicon wafer 40 as shown in FIG. Appears in disk and ring form. More specifically, an OSF region 42 appears at the center of the silicon wafer 40, and a Pv region 43, Pi region 44, Pv region 43, OSF region 42, Pv region 43, and Pi region 44 are concentrically formed outside the OSF region 42. Appear in this order. Here, the diameter of the disk-like Pv region 43 is D4. Further, the width of the ring composed of the OSF region 42 and the Pv region 43 is W1. Thus, when the pulling speed V is set to a speed corresponding to the B2-B2 line, the silicon single crystal ingot 20 that does not include the COP 41 and the dislocation cluster 45 is obtained, but the disk-shaped OSF region 42 and the ring-shaped OSF are obtained. Region 42 is formed. Therefore, in order to obtain a crystal not including the OSF region 42 as well as the COP 41 and the dislocation cluster 45, it is necessary to lower the pulling rate V by the control device 23. In addition, when a crystal not including the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is rejected.

  Further, when the pulling speed V is set to a speed corresponding to the C2-C2 line under the pulling condition of FIG. 3, the Pv region 43 is formed in a disk shape on the cut silicon wafer 40 as shown in FIG. And appear in a ring shape. Here, the diameter of the disk-shaped Pv region 43 is D5, which is smaller than the diameter D4 of the disk-shaped Pv region 43 shown in FIG. 3B (D5 <D4). The width of the ring-shaped Pv region 43 is W2, which is narrower than the ring width W1 shown in FIG. 3B (W2 <W1). Thus, when the pulling speed V is set to a speed corresponding to the C2-C2 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In addition, since a sufficient margin is secured for the pulling speed V at which the dislocation clusters 45 are generated, it is not necessary to change the pulling speed V by the control device 23. In addition, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  Further, when the pulling speed V is set to a speed corresponding to the line D2-D2 under the pulling conditions in FIG. 3, the Pv region 43 is formed in a disk shape on the cut silicon wafer 40 as shown in FIG. And the ring-shaped Pv region 43 disappears. Here, the diameter D6 of the disk-like Pv region 43 is smaller than the diameter D5 of the Pv region 43 shown in FIG. 3C (D6 <D5). As described above, when the pulling speed V is set to a speed corresponding to the D2-D2 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In this case, it is necessary to increase the pulling speed V by the control device 23 because there is only a small margin left with respect to the pulling speed V at which the above occurs. However, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  As described above, in the pulling conditions shown in FIG. 3, the control direction of the pulling speed V can be determined based on the diameter of the disk-shaped Pv region 43 and the width (or presence / absence) of the ring-shaped Pv region. It is. It is also possible to make a pass / fail judgment for the silicon single crystal ingot 20.

  FIG. 4A is a diagram showing the relationship between the pulling speed V of the silicon single crystal ingot 20 and the type and distribution of defects under a condition in which the radial distribution of the temperature gradient G is different from that in FIGS. 4B to 4D are cross-sectional views taken along lines B3-B3, C3-C3, and D3-D3 shown in FIG. 4A, respectively. The pulling condition in FIG. 4 is a pulling condition in which the OSF region 42 appears in a ring shape, and is a case where V / G is large (G is small) at the outer periphery of the crystal.

  When the pulling speed V is set to a speed corresponding to the line B3-B3 under the pulling conditions in FIG. 4, the OSF region 42 and the Pv region 43 are formed in the cut silicon wafer 40 as shown in FIG. Appears in a ring shape. More specifically, a ring-shaped OSF region 42 appears between two ring-shaped Pv regions 43, and the other regions become Pi regions 44. The width of the ring composed of the OSF region 42 and the Pv region 43 is W3. Thus, when the pulling speed V is set to a speed corresponding to the B3-B3 line, the silicon single crystal ingot 20 that does not include the COP 41 and the dislocation cluster 45 is obtained, but the ring-shaped OSF region 42 is formed. Therefore, in order to obtain a crystal not including the OSF region 42 as well as the COP 41 and the dislocation cluster 45, it is necessary to lower the pulling rate V by the control device 23. In addition, when a crystal not including the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is rejected.

  Further, when the pulling speed V is set to a speed corresponding to the C3-C3 line under the pulling condition of FIG. 4, the Pv region 43 is formed in a ring shape on the cut silicon wafer 40 as shown in FIG. Appear in Here, the width of the ring-shaped Pv region 43 is W4, which is narrower than the width W3 of the ring shown in FIG. 4B (W4 <W3). As described above, when the pulling speed V is set to a speed corresponding to the C3-C3 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In addition, since a sufficient margin is secured for the pulling speed V at which the dislocation clusters 45 are generated, it is not necessary to change the pulling speed V by the control device 23. In addition, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  Furthermore, when the pulling speed V is set to a speed corresponding to the line D3-D3 under the pulling condition of FIG. 4, as shown in FIG. 4 (d), the Pv region 43 is formed in a ring shape on the cut silicon wafer 40. However, the width W5 of the ring-shaped Pv region 43 is further narrower than the width W4 of the Pv region 43 shown in FIG. 4C (W5 <W4). Thus, when the pulling speed V is set to a speed corresponding to the D3-D3 line, the silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained. In this case, it is necessary to increase the pulling speed V by the control device 23 because there is only a small margin left with respect to the pulling speed V at which the above occurs. However, when a crystal that does not include the OSF region 42 is a non-defective product, the silicon single crystal ingot 20 is passed.

  As described above, in the pulling condition shown in FIG. 4, it is possible to determine the control direction of the pulling speed V based on the width (or presence / absence) of the Pv region 43 appearing in a ring shape. It is also possible to make a pass / fail judgment for the silicon single crystal ingot 20.

  As described above, the lifting condition in which the OSF region 42 appears in a disk shape (FIG. 2), the lifting condition in which the OSF region 42 appears in a disk shape and a ring shape (FIG. 3), and the lifting condition in which the OSF region 42 appears in a ring shape (FIG. Under any of the pulling conditions in FIG. 4), the current pulling speed can be determined by observing the position and width of the Pv region 43 (specifically, the diameter if it is a disk and the width if it is a ring). It is possible to determine whether V is faster or slower than the optimum pulling speed V. Here, the optimum pulling speed V refers to a pulling speed at which a silicon single crystal ingot 20 that does not include any of the COP 41, the OSF region 42, and the dislocation cluster 45 can be obtained and a sufficient margin can be secured. .

  Next, a method for observing the position and width of the Pv region will be described.

  The position and width of the Pv region can be observed by revealing a grown-in defect containing silicon oxide as a protrusion on the etched surface by the RIE method. Specifically, a silicon single crystal ingot that does not contain COPs and dislocation clusters is grown by the Czochralski method (growth process), and a silicon wafer is cut out from the silicon single crystal ingot (cut out process), and an as-grown silicon wafer By performing reactive ion etching on the substrate, grown-in defects including silicon oxide are exposed as protrusions on the etched surface (etching step). Thereby, the position and the width of the Pv region can be observed. As described above, the position and the width of the observed Pv region serve as an index for determining whether the current pulling speed V is faster or slower than the optimum pulling speed V. By feeding back to the growth conditions in the process, it becomes possible to stably mass-produce silicon single crystal ingots having a desired quality. The silicon single crystal ingot having a desired quality is, for example, a silicon single crystal ingot that does not include an OSF nucleus.

In order to make silicon oxide appear as protrusions by RIE, it is necessary to perform RIE under a condition that Si is more easily etched than SiO ₂ , that is, a condition with a higher Si / SiO ₂ selection ratio. As a result, oxygen precipitates (SiO ₂ ) are hardly etched and become apparent as protrusions.

  Here, the relationship between the region where the OSF occurs and the region where the protrusion is detected by the RIE method will be described.

  FIG. 5 is a diagram for explaining the relationship between a region where OSF is generated and a region where protrusions are detected by the RIE method. FIG. 5A is a diagram showing a silicon wafer in which OSF is revealed by heat treatment. ) Is a view showing a silicon wafer in which protrusions are made obvious by the RIE method.

As already described, the OSF region is a region containing plate-like oxygen precipitates in an as-grown state, and this plate-like oxygen precipitate is thermally oxidized at a high temperature of about 1000 ° C. to 1200 ° C. If not, it will not become apparent. The OSF region 42 shown in FIG. 5A shows the position and width of the OSF region that has been revealed by such heat treatment. In this example, the OSF region 42 having a diameter D0 appears at the center of the silicon wafer 40. ing. On the other hand, when RIE is performed under such a condition that the Si / SiO ₂ selection ratio is high with respect to an as-grown silicon wafer without performing such heat treatment, as shown in FIG. A protrusion generation region 46 having a diameter D1 appears at the center of the area. Here, the diameter D1 of the protrusion generation region 46 is wider than the diameter D0 of the OSF region 42 (D1> D0).

  The protrusion generation region 46 is a combined region of the OSF region 42 and the Pv region 43 shown in FIG. That is, when RIE is performed on a silicon wafer in an as-grown state, protrusions are generated in the OSF region 42 and the Pv region 43, and almost no protrusions are generated in the Pi region 44. This means that the boundary between the Pv region 43 and the Pi region 44 can be determined by performing RIE. Therefore, if the growth conditions in the subsequent growth process are adjusted based on the position and size of the protrusion generation region 46, a silicon single crystal ingot that does not contain OSF nuclei can be stably mass-produced.

  The RIE (etching process) may be performed on the etched surface by etching the cut silicon wafer with an aqueous solution containing hydrofluoric acid and nitric acid (case 1). Alternatively, it may be performed (case 2), or the cut silicon wafer may be mirror-finished and performed on the mirror-finished surface (case 3). This means that the projections by RIE are almost equal to any surface of the silicon wafer.

  As described above, according to the present embodiment, it is possible to evaluate the Pv region by a simple method and determine the pass / fail judgment of the crystal and the growth conditions of the crystal grown after the crystal. A single crystal silicon wafer that does not contain COPs or dislocation clusters can be stably manufactured without using the OSF region expected to affect the device as an index.

  The preferred embodiments of the present invention have been described above, but the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention. Needless to say, it is included in the range.

  Hereinafter, although the Example of this invention is described, this invention is not limited to this Example at all.

  In Example 1, the relationship between the region where the OSF occurs and the region where the protrusion is detected by the RIE method was investigated.

First, the polycrystalline silicon lump was put into a quartz crucible, and the polycrystalline silicon lump was heated in an argon atmosphere to obtain a silicon melt. Thereafter, a single crystal having a diameter of 305 mm and a crystal orientation of (100) was pulled at a constant rate of an average pulling rate of 0.50 mm / min to grow a silicon single crystal ingot having a diameter of 305 mm that does not contain COP and dislocation clusters. When the interstitial oxygen concentration was measured by the FT-IR method (ASTM F121-79), it was 9 × 10 ¹⁷ atoms / cm ^{3 to} 11 × 10 ¹⁷ atoms / cm ³ .

Next, as shown in FIG. 6, two wafers 40a and 40b were cut out from adjacent positions of the grown silicon single crystal ingot 20 and subjected to mirror finishing. Since the two wafers 40a and 40b are cut out from the adjacent positions, the defect distribution and the defect density can be regarded as equivalent. One of the two wafers (wafer 40a) is subjected to reactive ion etching (reactive ion etching) under an as-grown condition with a high Si / SiO ₂ selection ratio (ie, SiO ₂ is difficult to etch). gave. The RIE atmosphere was an HBr / Cl ₂ / He + O ₂ mixed gas, and etching was performed at about 5 μm under the conditions set so that the Si / SiO ₂ selection ratio was 100 or more. After RIE, washing with a hydrofluoric acid aqueous solution was performed to remove reaction products adhering at the time of RIE, and the surface etched by RIE was evaluated by visual observation under a condensing lamp and optical microscope observation, and protrusions were generated by RIE. The area width was measured.

  The remaining one (wafer 40b) was heat-treated at 1000 ° C. for 3 hours in an oxygen atmosphere and for 2 hours at 1150 ° C. in an oxygen atmosphere containing water vapor. After removing the oxide film with a hydrofluoric acid aqueous solution, selective etching was performed with a light solution, and the etched surface was evaluated by visual observation under a condensing lamp and observation with an optical microscope, and the area of the OSF was measured. .

  As a result of comparing the evaluation results of the two wafers 40a and 40b, the region where the protrusion is generated by RIE is 7 cm from the center, and the region where the OSF is generated is 3.5 cm from the center, which is more than the region where the OSF is generated. RIE revealed that the region where the protrusions were generated was wider. As a result, in the RIE method, not only OSF nuclei (plate-like oxygen precipitates) but also minute oxygen precipitation nuclei contained in the Pv region adjacent to the OSF region can be detected as projections, and these projections are generated. If the pass / fail judgment of the crystal and the growth conditions of the crystal grown after the crystal are determined using the area size as an index, it means that a crystal containing neither COP, OSF nor dislocation clusters can be produced stably. .

  In Example 2, the difference due to the pretreatment of the sample subjected to RIE was investigated.

  First, three wafers 40c to 40e were further cut out from the position where the two wafers 40a and 40b were cut out in Example 1 as shown in FIG. 7 (see FIG. 7). Since the three wafers 40c to 40e are cut out from the adjacent positions, the defect distribution and the defect density can be regarded as equivalent. Then, these three sheets were etched with a mixed solution containing hydrofluoric acid and nitric acid in order to remove processing strain and smooth the surface.

  The first wafer (wafer 40c) was subjected to SC-1 cleaning (cleaning with ammonia water + hydrogen peroxide water + water) to remove adhered foreign matters. The cleaning for removing the adhering foreign matter is not limited to the SC-1 cleaning, and any cleaning that has an effect of removing the adhering foreign matter may be used. The second sheet (wafer 40d) was first cleaved in half. A 5 mm wide strip is produced by cutting with a dicing saw at a position approximately 5 mm from the cleavage plane of the cleaved wafer in parallel to the cleavage plane, and further, the length of the strip is cut with a dicing saw at the center in the longitudinal direction. A strip 50d was prepared. Therefore, the strip 50d includes the center to the outer periphery of the crystal. The strip 50d thus produced was subjected to SC-1 cleaning to remove adhering foreign matter. The cleaning for removing the adhering foreign matter is not limited to the SC-1 cleaning, and any cleaning that has an effect of removing the adhering foreign matter may be used. The surface opposite to the cleavage surface (the surface cut with a dicing saw) was attached to a support substrate 51 having a diameter of 300 mm with a resist (see FIG. 8). The material used for adhering the strips 50d may be other than a resist, and any material that can withstand treatments such as RIE, hydrofluoric acid cleaning, and SC-1 cleaning. In the present invention, the strip 50d is washed and then attached to the support substrate 51. However, the strip 50d may be attached to the support substrate 51 and then washed together with the support substrate 51. The third sheet (wafer 40e) was mirror-finished.

The three samples subjected to the above treatment were subjected to reactive ion etching under conditions where the Si / SiO ₂ selectivity was high (ie, SiO ₂ was difficult to be etched). The RIE atmosphere was an HBr / Cl ₂ / He + O ₂ mixed gas, and etching was performed at about 5 μm under the conditions set so that the Si / SiO ₂ selection ratio was 100 or more. After RIE, washing with a hydrofluoric acid aqueous solution was performed to remove reaction products adhering at the time of RIE, and the surface etched by RIE was evaluated by visual observation under a condensing lamp and optical microscope observation, and protrusions were generated by RIE. The area width was measured.

  As a result, as shown in FIG. 9, in the first sheet (wafer 40c) and the third sheet (wafer 40e), the region 46 where the protrusion was generated was a disk shape having a radius of 7 cm. However, in the first sample that was not mirror-polished, protrusions that were presumed to be caused by disturbance from the distribution were also observed. Since crystal defects are distributed concentrically, it can be easily estimated that protrusions that are not distributed concentrically are due to disturbance. On the etched cleaved surface of the second strip-shaped sample 50d, protrusions were generated within a range of 7 cm from the center of the crystal, and no protrusions were generated on the outer peripheral side of the crystal. As described above, the width of the region where the protrusion was generated by RIE was consistent among the three samples. Therefore, even if the surface is not a mirror surface, etching caused by a solution containing hydrofluoric acid and nitric acid and the surface subjected to SC-1 cleaning, or the surface subjected to the SC-1 cleaning after cleaving, may cause defects due to defects by RIE. It has become clear that the size of the resulting area can be evaluated.

  As a result of the above examination, the inventor cuts out a wafer for defect evaluation from the pulled crystal at a predetermined interval, and performs RIE on the sample subjected to etching for removal of processing distortion and flattening or cleaved sample. The Pv region can be visualized as a protrusion that can be observed with a visual observation under a condensing lamp or an optical microscope, and the pass / fail judgment of the crystal and growth after the crystal are made using the size of the area where the protrusion is generated as an index. It has been found that if the growth conditions of the crystal to be produced are determined, a crystal containing neither COP, OSF nor dislocation clusters can be produced stably.

  In Example 3, a silicon single crystal was grown using three hot zones having different radial distributions of the temperature gradient G and evaluated.

First, three silicon single crystals were manufactured using three hot zones (HZ-1, HZ-2, HZ-3) having different radial distributions of the temperature gradient G. Specifically, the polycrystalline silicon lump was put into a quartz crucible, and the polycrystalline silicon lump was heated in an argon atmosphere to obtain a silicon melt. Thereafter, a single crystal having a diameter of 305 mm and a crystal orientation of (100) was pulled at a constant speed of 0.50 mm / min. After the pulled crystals are cylindrically ground to a diameter of about 301 mm, five wafers each having a surface substantially perpendicular to the growth axis are cut out at intervals of about 50 mm, and all the wafers are immersed in a mixed solution of hydrofluoric acid and nitric acid for etching. Thus, the processing damage layer generated at the time of cutting was removed. Since these five wafers were cut out from the adjacent positions, the defect distribution can be regarded as equivalent. When the interstitial oxygen concentration was measured by the FT-IR method (ASTM F121-79), it was 9 × 10 ¹⁷ atoms / cm ^{3 to} 11 × 10 ¹⁷ atoms / cm ³ .

  First, COP and dislocation clusters were evaluated for all the crystal parts. One of the five wafers cut out from each part was cleaved in half, and COP was evaluated on one side, and dislocation clusters were evaluated on the other side. COP was evaluated by infrared tomography, and dislocation clusters were evaluated by secco etching. Of the remaining four wafers cut out from each part of each crystal where COP and dislocation clusters were not detected, wafers that were pulled up with HZ-1 and where COP and dislocation clusters were not detected The wafer of “Sample 1”, which was lifted by HZ-2, and where the COP and the dislocation cluster were not detected was designated as “Sample 2”, and the wafer lifted by HZ-3 and the COP and the dislocation cluster were The wafer of the part that was not detected was designated as “Sample 3”. Since there are a plurality of sites where COPs and dislocation clusters are not detected, a plurality of wafers of Samples 1 to 3 exist as one set of four.

  [Evaluation of Sample 1]

  Four sets of samples 1A to 1D were prepared from the sample 1 wafer. As described above, the wafer of sample 1 is a wafer pulled up with HZ-1 and a portion where a COP and a dislocation cluster are not detected.

  The sample 1A of the first set is made by adding etching with a mixed solution of hydrofluoric acid and nitric acid to the sample 1 wafer to further smooth the wafer surface, and SC-1 cleaning (with ammonia water + hydrogen peroxide water + water). This is a sample from which the adhering foreign matter has been removed by washing. In Sample 1, a mixed solution of hydrofluoric acid and nitric acid was used to smooth the wafer surface, but other substances (for example, acetic acid, etc.) may be added. Further, in Sample 1, SC-1 cleaning was performed to remove the attached foreign matter, but it is not necessary to be limited to the SC-1 cleaning, and any cleaning that has an effect of removing the attached foreign matter may be used.

  In the production of the second set of sample 1B, the wafer of sample 1 was first cleaved in half. A 5 mm wide strip is produced by cutting with a dicing saw at a position approximately 5 mm from the cleavage plane of the cleaved wafer in parallel to the cleavage plane, and further, the length of the strip is cut with a dicing saw at the center in the longitudinal direction. A strip of was made. Therefore, this strip includes the center to the outer periphery of the crystal. The strips thus produced were subjected to SC-1 cleaning to remove adhered foreign matters. In sample 1B, SC-1 cleaning was performed in order to remove adhering foreign matter, but it is not necessary to limit to SC-1 cleaning, and any cleaning that has an effect of removing adhering foreign matter may be used. The surface opposite to the cleaved surface (the surface cut by a dicing saw) was attached to a support substrate having a diameter of 300 mm (a silicon wafer was used in Sample 1B) with a resist. In Sample 1B, a resist is used to attach the strip to the support substrate. However, any material other than the resist may be used as long as it can withstand processing such as RIE, hydrofluoric acid cleaning, and SC-1 cleaning. In Sample 1B, the strip is washed and then attached to the support substrate. However, the sample 1B may be washed together with the support substrate after being attached to the support substrate.

  The third set of samples 1C is a sample obtained by performing mirror surface processing on the sample 1 wafer.

  The fourth set of sample 1D is a sample obtained by performing heat treatment on the sample 1 wafer at 1000 ° C. for 3 hours in an oxygen atmosphere and for 2 hours at 1150 ° C. in an oxygen atmosphere containing water vapor.

Next, reactive ion etching (RIE) was performed on the first to third sets of samples 1A to 1C under conditions where the Si / SiO ₂ selectivity was high (ie, SiO ₂ was difficult to be etched). The RIE atmosphere was an HBr / Cl ₂ / He + O ₂ mixed gas, and etching was performed at about 5 μm under the conditions set so that the Si / SiO ₂ selection ratio was 100 or more. After RIE, washing with a hydrofluoric acid aqueous solution was performed to remove reaction products adhering at the time of RIE, and the surface etched by RIE was evaluated by visual observation under a condensing lamp and optical microscope observation, and protrusions were generated by RIE. The area width was measured.

  On the other hand, for the fourth set of samples 1D, the oxide film was removed with a hydrofluoric acid aqueous solution and then selective etching was performed with a light solution, and the etched surface was evaluated by visual observation under a condenser lamp and optical microscope observation. Then, the area of the OSF was measured.

  First, the OSF occurrence status evaluated in sample 1D will be described. In the crystal pulled by HZ-1, there were a sample in which OSF was not generated and a sample in which OSF was generated in a disk shape. This result suggests that in HZ-1, V / G near the center is larger (G is smaller) than the outer peripheral part of the crystal (see FIG. 2).

  Next, the state of occurrence of protrusions after RIE of samples 1A to 1C will be described. In sample 1A and sample 1C, the protrusions occurred in a disk shape. However, in sample 1A that was not mirror-polished, protrusions that were estimated to be caused by disturbance from the distribution were also observed. Since crystal defects are distributed concentrically, it can be easily estimated that protrusions that are not distributed concentrically are due to disturbance. On the etched cleaved surface of Sample 1B, a protrusion was generated on the center side of the crystal, and no protrusion was generated on the outer peripheral side of the crystal.

  FIG. 10 is a plot of the radius of the region where the protrusions occurred in the samples 1A and 1C and the length of the region where the protrusions occurred in the sample 1B (ie, the radius) against the radius of the region where the OSF occurred. is there. The radius of the region where the OSF is generated is a value obtained from the sample 1D.

  As shown in FIG. 10, it was found that the radii in the ranges where the protrusions were observed in the samples 1A to 1C were almost the same, and about 3.5 cm larger than the radius in the range where the OSF was observed in the sample 1D. From this result, the following two things can be said. First, even on a surface etched smoothly with a mixed solution containing hydrofluoric acid and nitric acid or a cleaved surface, the same evaluation as that of the mirror-polished surface (evaluation of the range in which protrusions are generated by RIE) is possible. The radius of the area where the protrusions are generated by RIE is surely wider than the radius of the generated area, and there is a good correlation between the two.

  Therefore, by using the radius of the range in which the protrusion is generated by RIE as an index, it is possible to determine the pass / fail judgment of the crystal and the pulling condition of the crystal grown after the crystal based on the radius. In the case of sample 1, if the pulling speed is set so that the radius of the region where the protrusion is generated by RIE is 3.5 cm or less, the crystal containing neither COP, dislocation clusters nor OSF nuclei can be pulled. For example, when the hot zone changes with time and G becomes smaller (when V / G becomes larger), the radius of the region where the protrusion is generated by RIE becomes larger. Changes in V / G due to changes over time can be detected. If the next crystal is pulled at a lower pulling speed, the V / G can be restored, and the crystal that does not contain OSF nuclei can be subsequently pulled.

  [Evaluation of Sample 2]

  Four sets of samples 2A to 2D were prepared from the sample 2 wafer. As described above, the wafer of sample 2 is a wafer pulled up with HZ-2 and a portion where COP and dislocation clusters are not detected.

  Samples 2A to 2D were produced by the same method as Samples 1A to 1D.

  First, the OSF occurrence status evaluated in sample 2D will be described. In the crystal pulled by HZ-2, there were a sample in which OSF was not generated and a sample in which OSF was generated in a disk + ring shape near the center and the outer periphery of the crystal. This result suggests that in HZ-2, V / G is large (G is small) in the vicinity of the center of the crystal and in the outer periphery (see FIG. 3).

  Next, the generation | occurrence | production situation of the protrusion after RIE of sample 2A-2C is described. In Sample 2A and Sample 2C, the protrusions were generated in a disc + ring shape near the center and near the outer periphery. However, in the sample 2A that was not mirror-polished, protrusions estimated to be caused by disturbance from the distribution were also observed. Since crystal defects are distributed concentrically, it can be easily estimated that protrusions that are not distributed concentrically are due to disturbance. On the etched cleaved surface of sample 2B, protrusions were generated on the center side and the outer peripheral side of the crystal.

  As described above, in Sample 2, protrusions due to OSF and RIE occurred in a disk + ring shape. First, the relationship between the protrusions by OSF and RIE is arranged in the area where the defect occurs in the disk shape. The radius of the region in which the protrusions are generated in the samples 2A and 2C and the length of the region in which the protrusion is generated on the crystal center side in the sample 2B (ie, the radius) are the radius of the region in which the OSF is generated in the disk shape. FIG. 11 is plotted against.

  As shown in FIG. 11, it was found that the radii in the ranges where the protrusions were observed in Samples 2A to 2C were almost the same, and about 4.5 cm larger than the radius where the OSF was observed. From this result, the following two things can be said. First, even on a surface etched smoothly with a mixed solution containing hydrofluoric acid and nitric acid or a cleaved surface, the same evaluation as that of the mirror-polished surface (evaluation of the range in which protrusions are generated by RIE) is possible. The radius of the area where the protrusions are generated by RIE is surely wider than the radius of the generated area, and there is a good correlation between the two. If the pulling speed is set so that the radius of the disk-like region where RIE protrusions are generated is 4.5 cm or less, it is possible to pull a crystal that does not contain COPs, dislocation clusters, and OSF nuclei in the vicinity of the crystal center.

  Next, the relationship between the protrusions by OSF and RIE is arranged for the region where defects are generated in a ring shape. The width of the region where protrusions were generated in a ring shape in Samples 2A and 2C and the width of the region where protrusions were generated on the outer periphery of the crystal in the strip-shaped sample of Sample 2B were plotted against the width of the region where the OSF was generated in a ring shape. This is shown in FIG.

  As shown in FIG. 12, the widths of the ranges where the protrusions were observed in Samples 2A to 2C were substantially the same, and it was found that the width was about 2.5 cm larger than the width of the range where OSF was observed. From this result, the following two things can be said. First, even on a surface etched smoothly with a mixed solution containing hydrofluoric acid and nitric acid or a cleaved surface, the same evaluation as that of the mirror-polished surface (evaluation of the range in which protrusions are generated by RIE) is possible. The width of the area where the protrusions are generated by RIE is necessarily wider than the width of the generated area, and there is a good correlation between the two. If the pulling speed is set so that the width of the ring-like region where protrusions are generated by RIE is 2.5 cm or less, it is possible to pull a crystal that does not contain COPs, dislocation clusters, and OSF nuclei near the crystal periphery.

  Therefore, by using the size of the range in which protrusions are generated in a disk + ring shape by RIE as an index, it is possible to determine the pass / fail judgment of the crystal and the pulling conditions for the crystal grown after the crystal based on the size. In the case of sample 2, the pulling speed is adjusted so that the radius of the region where the disk-like projection is generated by RIE is 3.5 cm or less and the width of the region where the ring-like projection is generated by RIE is 2.5 cm or less. If set, crystals that do not contain COPs, dislocation clusters, and OSF nuclei can be pulled over the entire wafer surface. When G changes with time in the hot zone (when V / G increases), the region in which the protrusion is generated by RIE becomes wider. / G change can be detected. If the next crystal is pulled at a lower pulling speed, the V / G can be restored, and the crystal that does not contain OSF nuclei can be subsequently pulled.

  [Evaluation of Sample 3]

  Four sets of samples 3A to 3D were prepared from the sample 3 wafer. As described above, the sample 3 wafer is a wafer pulled up with HZ-3, and a portion of the wafer where COP and dislocation clusters are not detected.

  Samples 3A to 3D were produced by the same method as Samples 1A to 1D.

  First, the OSF occurrence status evaluated in sample 3D will be described. In the crystal pulled by HZ-3, there were a sample in which OSF was not generated and a sample in which OSF was generated in a ring shape near the outer periphery of the crystal. This result suggests that V / G is large (G is small) in the outer peripheral portion of the crystal in HZ-3 (see FIG. 4).

  Next, the generation | occurrence | production condition of the protrusion after RIE of sample 3A-3C is described. In Sample 3A and Sample 3C, the protrusions occurred in a ring shape near the outer periphery. However, in the sample 3A that was not mirror-polished, protrusions estimated to be caused by disturbance from the distribution were also observed. Since crystal defects are distributed concentrically, it can be easily estimated that protrusions that are not distributed concentrically are due to disturbance. On the etched cleaved surface of the strip-shaped sample of Sample 3B, protrusions were generated on the outer peripheral side of the crystal.

  As described above, in Sample 3, protrusions due to OSF and RIE occurred in a ring shape. The width of the region where the protrusions were generated in the ring shape in Samples 3A and 3C and the width of the region where the protrusions were generated on the outer periphery of the crystal in the strip sample of Sample 3B were plotted against the width of the region where the OSF was generated in the ring shape. This is shown in FIG.

  As shown in FIG. 13, it was found that the widths of the areas where the protrusions were observed in Samples 3A to 3C were almost the same, and about 1.5 cm larger than the width of the area where the OSF was observed. From this result, the following two things can be said. First, even on a surface etched smoothly with a mixed solution containing hydrofluoric acid and nitric acid or a cleaved surface, the same evaluation as that of the mirror-polished surface (evaluation of the range in which protrusions are generated by RIE) is possible. The width of the area where the protrusions are generated by RIE is necessarily wider than the width of the generated area, and there is a good correlation between the two.

  Therefore, if the pulling speed is set so that the width of the ring-like region where the protrusions by RIE are generated is 1.5 cm or less, it is possible to pull a crystal that does not contain COPs, dislocation clusters, and OSF nuclei. That is, using the size of the range in which protrusions are generated in a ring shape by RIE as an index, it is possible to determine pass / fail judgment of the crystal and the pulling conditions for the crystal grown after the crystal based on the size. In the case of the present embodiment, if the pulling speed is set so that the width of the region where the ring-shaped protrusion is generated by RIE is 1.5 cm or less, the COP, the dislocation cluster, and the OSF nucleus are not included over the entire wafer surface. The crystal can be pulled up. When G changes as the hot zone changes over time (when V / G increases), the width of the ring-shaped region where the protrusion is generated by RIE becomes wider. Changes in V / G due to changes over time can be detected. If the next crystal is pulled at a lower pulling speed, the V / G can be restored, and the crystal that does not contain OSF nuclei can be subsequently pulled.

It is a schematic diagram which shows the structure of the pulling apparatus applicable to the manufacturing method of the silicon single crystal by preferable embodiment of this invention. (A) is a figure which shows an example of the relationship between the pulling-up speed V of the silicon single crystal ingot 20, and the kind and distribution of a defect, (b)-(d) is the B1-B1 line | wire shown to (a), respectively. It is sectional drawing along the C1-C1 line and D1-D1 line. (A) is a figure which shows the other example of the relationship between the pulling-up speed V of the silicon single crystal ingot 20, and the kind and distribution of a defect, (b)-(d) is B2-B2 shown to (a), respectively. It is sectional drawing along a line, a C2-C2 line, and a D2-D2 line. (A) is a figure which shows the further another example of the relationship between the pulling-up speed V of the silicon single crystal ingot 20, and the kind and distribution of a defect, (b)-(d) is B3- shown to (a), respectively. It is sectional drawing along B3 line, C3-C3 line, and D3-D3 line. It is a figure for demonstrating the relationship between the area | region where OSF generate | occur | produces, and the area | region where a processus | protrusion is detected by RIE method, (a) is a figure which shows the silicon wafer which revealed OSF by heat processing, (b) is RIE method It is a figure which shows the silicon wafer which made protrusion explicit. It is a schematic diagram for demonstrating the cutting position of the wafers 40a and 40b used in Example 1. FIG. It is a schematic diagram for demonstrating the cutout position of the wafers 40c-40e used in Example 2. FIG. It is a schematic perspective view which shows the state which affixed the strip 50d used in Example 2 to the support substrate 51. FIG. 6 is a schematic diagram showing the evaluation results of Example 2. FIG. 6 is a graph showing an evaluation result of Sample 1 in Example 3. 10 is a graph showing an evaluation result of Sample 2 (disk area) in Example 3. It is a graph which shows the evaluation result of the sample 2 (ring area | region) in Example 3. FIG. 10 is a graph showing the evaluation result of Sample 3 in Example 3. It is a figure which shows the general relationship between V / G and the kind and distribution of a defect.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Silicon single crystal pulling apparatus 11 Chamber 12 Support rotating shaft 13 Graphite susceptor 14 Quartz crucible 15 Heater 16 Support shaft drive mechanism 17 Seed chuck 18 Wire 19 Wire winding mechanism 20 Silicon single crystal ingot 21 Silicon melt 22 Heat shielding member 23 Control Device 24 Gas inlet 25 Gas pipe 26 Conductance valve 27 Gas outlet 28 Exhaust pipe 29 Conductance valve 30 Vacuum pump 31 Magnetic field supply device 40 Silicon wafer 41 COP region 42 OSF region 43 Pv region 44 Pi region 45 Dislocation cluster 46 Protrusion generation region 50d strip 51 support substrate

Claims (8)

A growth step of growing a silicon single crystal ingot containing no COP and dislocation clusters by the Czochralski method;
Cutting out a silicon wafer from the silicon single crystal ingot,
an etching step of revealing grown-in defects including silicon oxide as protrusions on the etched surface by performing reactive ion etching on the silicon wafer in the as-grown state,
A method for producing a silicon single crystal, comprising adjusting a growth condition in a subsequent growth step based on a region where the protrusion is manifested in the etching step.   2. The method for producing a silicon single crystal according to claim 1, wherein whether or not the silicon single crystal ingot is accepted is determined based on a region where the protrusion is manifested in the etching step.   3. The method for producing a silicon single crystal according to claim 1, wherein in the growing step, a silicon single crystal ingot that does not include an OSF nucleus is grown.   The method for producing a silicon single crystal according to any one of claims 1 to 3, wherein a region where the protrusions manifested in the etching step are formed into a disk shape and / or a ring shape.   5. The etching process according to claim 1, wherein, in the etching step, the silicon wafer is etched with an aqueous solution containing hydrofluoric acid and nitric acid, and the reactive ion etching is performed on the etched surface. 6. A method for producing a silicon single crystal.   5. The method for producing a silicon single crystal according to claim 1, wherein in the etching step, the silicon wafer is cleaved, and the reactive ion etching is performed on a cleaved surface. 6.   5. The silicon single crystal production according to claim 1, wherein, in the etching step, the silicon wafer is mirror-finished, and the reactive ion etching is performed on the mirror-finished surface. 6. Method. Based on the width of the region where the protrusion has come to appear in the etching step, adjust the growth conditions in the subsequent growth step,
The method for producing a silicon single crystal according to claim 1, wherein the generation region of the protrusion is a defect-free Pv region that is an oxygen precipitation region.

Images