JP4082394B2 - Evaluation method of silicon wafer - Google Patents

Description

  The present invention relates to a silicon wafer in which oxygen precipitation can be stably obtained without depending on a device process and a crystal position, a manufacturing method thereof, and a method for evaluating a defect region of a silicon wafer whose pulling conditions are unknown.

  In recent years, with the miniaturization of elements accompanying higher integration of semiconductor circuits such as DRAM, quality requirements for silicon single crystals produced by the Czochralski method (hereinafter sometimes abbreviated as CZ method) as the substrate Is growing. In particular, there are defects called “grown-in” defects such as FPD, LSTD, COP, etc., which deteriorate device characteristics, and their reduction is regarded as important.

  In describing these defects, first, a vacancy-type point defect called vacancy (hereinafter sometimes abbreviated as V) incorporated in a silicon single crystal, and interstitial-silicon (interstitial-Si, hereinafter). What is generally known is a factor that determines the concentration of each interstitial silicon point defect called “I” (sometimes abbreviated as “I”).

  In a silicon single crystal, the V-region is a vacancy, that is, a region having many recesses and holes generated due to a shortage of silicon atoms, and the I-region is generated by the presence of extra silicon atoms. A neutral region (Neutral region, hereinafter referred to as an N-region) that has a shortage of atoms and no (or few) atoms between the V-region and the I-region. May be abbreviated). The grow-in defects (FPD, LSTD, COP, etc.) are generated only when V or I is supersaturated. If there is some atomic bias, It has been found that it does not exist.

  It is known that the concentration of both point defects is determined from the relationship between the crystal pulling rate (growth rate) in the CZ method and the temperature gradient G in the vicinity of the solid-liquid interface in the crystal. In addition, in the N-region between the V-region and the I-region, it has been confirmed that there is a ring-shaped defect called OSF (Oxidation Induced Stacking Fault).

  When these defects due to crystal growth are classified, for example, when the growth rate is relatively high, such as about 0.6 mm / min or more, FPD, LSTD, and COP that are attributed to voids in which vacancy-type point defects are gathered. Groin-in defects such as these are present at high density throughout the crystal diameter direction, and a region where these defects exist is called a V-rich region. Further, when the growth rate is 0.6 mm / min or less, the OSF ring described above is generated from the periphery of the crystal as the growth rate is reduced, and L / D that is considered to be caused by a dislocation loop outside the ring. Defects (Large Dislocation: abbreviations for interstitial dislocation loops, LSEPD, LFPD, etc.) are present at a low density, and a region where these defects are present is called an I-rich region. Furthermore, when the growth rate is lowered to about 0.4 mm / min or less, the OSF ring aggregates and disappears at the center of the wafer, and the entire surface becomes an I-rich region.

Further, an N-region in which there is no vacancy-induced FPD, LSTD, COP, dislocation loop-induced LSEPD, LFPD, or even OSF outside the OSF ring between the V-rich region and the I-rich region recently. The presence of has been discovered. This region is outside the OSF ring, and when oxygen precipitation heat treatment is performed and the contrast of the precipitation is confirmed by X-ray observation or the like, there is almost no oxygen precipitation and is so rich that LSEPD and LFPD are formed. It is not the I-rich region side.
Further, it is confirmed that there are no defects caused by vacancies or dislocation loops inside the OSF ring, and there is an N-region in which no OSF exists.

In the normal method, these N-regions exist obliquely with respect to the growth axis direction when the growth rate is lowered, and therefore exist only in a part of the wafer plane.
In this N-region, in the Boronkov theory (see Non-Patent Document 1), the parameter F / G, which is the ratio of the pulling rate (F) and the temperature gradient (G) in the crystal solid-liquid interface axis direction, is the total concentration of point defects. It is said to decide. Considering this, since the pulling speed should be constant in the plane, since G has a distribution in the plane, for example, at a certain pulling speed, the center is the V-rich region and the N-region is sandwiched around it. Only crystals that would be in the I-rich region were obtained.

  Therefore, recently, when the distribution of G in the plane has been improved, the N-region that has existed only at an angle, for example, when the pulling-up speed F is gradually lowered, the N-region is lateralized at a certain pulling speed. Crystals spread over the entire surface can be manufactured. Further, in order to expand the crystal of the entire N-region in the length direction, it can be achieved to some extent if the pulling rate is maintained while the N-region is expanded laterally. Also, considering that G changes as the crystal grows, if it is corrected and the pulling speed is adjusted so that F / G is constant, the entire surface N- The area crystal can be enlarged.

  When this N-region is further classified, it can be seen that there are an NV region (region with many vacancies) adjacent to the outside of the OSF ring and an NI region (region with a lot of interstitial silicon) adjacent to the I-rich region. Yes.

  Furthermore, in the CZ method silicon substrate, in addition to the importance of reducing such grow-in defects, the control of oxygen precipitation has become increasingly important from the viewpoint of the internal gettering effect on heavy metal impurities. However, since oxygen precipitation strongly depends on the heat treatment conditions, it is extremely difficult to obtain appropriate oxygen precipitation in a device process that is different for each user. Furthermore, the wafer is subjected not only to the device process but also to a heat treatment (crystal thermal history) that is cooled from the melting point to room temperature in the crystal pulling process. Therefore, the oxygen precipitation nuclei (grow-in precipitation nuclei) formed by the crystal thermal history are already present in the as-grown crystal. The presence of the grown-in precipitation nuclei makes it more difficult to control oxygen precipitation.

  The oxygen precipitation process in the device process can be classified into two types. One is a process of growing grown-in precipitation nuclei remaining in the first stage heat treatment of the device process. The other is a process in which nuclei are generated during the device process and the nuclei grow. In the latter case, since it strongly depends on the oxygen concentration, oxygen precipitation can be controlled by controlling the oxygen concentration. On the other hand, in the former case, the thermal stability of the grown-in precipitation nuclei (how much density can remain at the initial temperature of the process) is an important point.

  For example, if the size of the grown-in precipitation nuclei is small but the size is small, it becomes thermally unstable and disappears in the first stage heat treatment of the device process, so oxygen precipitation cannot be secured. The problem here is that the thermal stability of grown-in precipitation nuclei strongly depends on the crystal thermal history, so even in the wafer with the same initial oxygen concentration, the oxygen in the device process depends on the crystal pulling conditions and the position in the crystal axis direction. The precipitation behavior varies greatly. Therefore, in order to control the oxygen precipitation in the device process, it is important to control the thermal stability of the grown-in precipitation nuclei by controlling not only the oxygen concentration but also the crystal thermal history.

V. V. Voronkov; Journal of Crystal Growth, 59 (1982) 625-643.

The technology for reducing the aforementioned grow-in defects is currently under development, but the low-defect crystal produced by such a method controls the crystal thermal history in order to reduce the grow-in defects. This is considered to change the thermal stability of the grown-in precipitation nuclei. However, it is completely unknown how much it has changed.
Therefore, it is expected that the oxygen precipitation behavior in the device process of such a low defect crystal varies greatly, resulting in a decrease in device yield.

  In addition, in the case of a wafer with an unknown defect region, a method for judging from which defect region the wafer was made has not been established, so it is predicted what oxygen precipitation behavior will be exhibited in the device process It was difficult to do.

  The present invention has been made in view of such problems, and an object of the present invention is to provide a silicon wafer in which oxygen precipitation can be stably obtained without depending on a crystal position or a device process, and a manufacturing method thereof. Another object of the present invention is to provide a method for evaluating a defect region of a silicon wafer whose pulling condition is unknown and whose defect region is unknown.

  The present invention has been made to achieve the above object, and the entire surface of the silicon wafer is one of the NV region, the NV region including the OSF ring region, and the OSF ring region, and the interstitial oxygen concentration is 14 ppma ( This is a silicon wafer characterized in that it is below the Japan Electronics Industry Promotion Association (JEIDA) standard).

  Thus, if the entire surface of the silicon wafer is the NV region, or the entire surface is the OSF ring region or a region in which these are mixed, there is an appropriate amount of thermally stable large grown-in precipitation nuclei. Therefore, BMD (oxygen precipitate called Bulk Micro Defect) can be obtained stably. Further, if the interstitial oxygen concentration is 14 ppma or less, the density of small grown-in precipitation nuclei is low, so that a silicon wafer with reduced variation due to crystal positions of oxygen precipitates is obtained.

  The present invention also provides a silicon wafer obtained by slicing a silicon single crystal rod grown by doping with nitrogen by the Czochralski method, wherein the entire surface of the silicon wafer includes an NV region and an OSF ring region. A silicon wafer characterized by being either an NV region or an OSF ring region.

  As described above, if the silicon wafer is doped with nitrogen and the entire surface is an NV region or an OSF ring region or a region in which these are mixed, thermally stable large grown-in precipitation nuclei can be obtained at a high density. The silicon wafer can obtain a sufficient gettering effect in the device process.

Furthermore, the present invention is the silicon wafer, wherein the doped nitrogen concentration is 1 × 10 ^{10 to} 5 × 10 ¹⁵ atoms / cm ³ .
That is, in order to obtain an extremely high density BMD due to the effect of nitrogen doping, a nitrogen concentration of at least 1 × 10 ¹⁰ pieces / cm ³ or more is necessary, and at 5 × 10 ¹⁵ pieces / cm ³ or more, the CZ method is used. This is because, when the single crystal rod is pulled up, it may hinder single crystallization.

  Also, in the case of a nitrogen-doped wafer, if the interstitial oxygen concentration is 14 ppma or less, the density of small grown-in precipitation nuclei is low, so that variations in oxygen precipitates due to crystal positions can be reduced.

Next, according to the present invention, when a silicon single crystal is grown by the Czochralski method, the pulling rate is F [mm / min], and the temperature gradient in the crystal in the pulling axis direction between the melting point of silicon and 1400 ° C. When the average value is expressed in G [° C./mm], the distance D [mm] from the crystal center to the crystal periphery is taken as the horizontal axis, and the value of F / G [mm ² / ° C./min] is taken as the vertical axis. When the crystal is pulled in the NV region or the OSF ring region of the defect distribution diagram showing the above, the silicon wafer is manufactured by pulling the crystal so that the interstitial oxygen concentration is 14 ppma or less.

Thus, based on the defect distribution diagram of FIG. 8 obtained by analyzing the results of the experiment / investigation, the boundary between the V-rich region and the NV region and the region surrounded by the boundary between the NV region and the NI region. If the crystal is pulled by controlling the pulling rate F of the crystal and the average value G of the temperature gradient in the crystal in the pulling axis direction between the melting point of silicon and 1400 ° C., the grown single crystal rod is sliced. The entire surface of the silicon wafer obtained in this way can be made into either the NV region, the NV region including the OSF ring region, or the OSF ring region. At the same time, the interstitial oxygen concentration is controlled to 14 ppma or less to pull the crystal. Can be raised.
Accordingly, in such a region, there are moderately large grown-in precipitation nuclei that are thermally stable, so that even if the device process is different, there is little variation in oxygen precipitation, and BMD can be obtained stably. In addition, since the interstitial oxygen concentration is 14 ppma or less, the density of small grown-in precipitation nuclei can be lowered, and variations in oxygen precipitates due to crystal positions can be reduced.

In the present invention, when growing a silicon single crystal by the Czochralski method, the pulling rate is F [mm / min], and the average temperature gradient in the crystal in the pulling axis direction between the melting point of silicon and 1400 ° C. When the value is expressed in G [° C./mm], the distance D [mm] from the crystal center to the crystal periphery is taken as the horizontal axis, and the value of F / G [mm ² / ° C./min] is taken as the vertical axis. In the case of pulling up a crystal in the NV region or the OSF ring region of the defect distribution diagram shown, the silicon wafer manufacturing method is characterized by pulling up the crystal while doping nitrogen.

When the crystal is pulled under such conditions, the silicon wafer obtained by slicing the grown single crystal rod is doped with nitrogen, and the entire surface is covered with the NV region, the NV region including the OSF ring region, and the OSF. It can be any of the ring areas.
As described above, if the silicon wafer is doped with nitrogen and the entire surface is an NV region or an OSF ring region or a region in which these are mixed, thermally stable large grown-in precipitation nuclei can be obtained at a high density. A wafer capable of obtaining a sufficient gettering effect in the device process can be manufactured.

In this case, the nitrogen concentration to be doped can be set to 1 × 10 ^{10 to} 5 × 10 ¹⁵ atoms / cm ³ .
Further, in this case, when the crystal is grown by the CZ method, the crystal can be pulled up so that the interstitial oxygen concentration is 14 ppma or less.

Thus, in order to obtain an extremely high density BMD by nitrogen doping, a nitrogen concentration of 1 × 10 ¹⁰ pieces / cm ³ or more is necessary, and if it exceeds 5 × 10 ¹⁵ pieces / cm ³ , a single crystal is obtained by the CZ method. When pulling up the rod, it may hinder single crystallization, so it is preferable to set it to 5 × 10 ¹⁵ pieces / cm ³ or less.
In addition, even when nitrogen is doped, if the interstitial oxygen concentration is 14 ppma or less, the density of small grown-in precipitation nuclei is low, so that variations in oxygen precipitates due to crystal positions can be reduced.

The present invention also relates to a method for evaluating a defect region of a silicon wafer produced by the CZ method, wherein the defect region of a silicon wafer to be evaluated is compared by comparing the density of at least two oxygen precipitates measured by the following steps. It is a method to evaluate.
(1) A wafer to be evaluated is divided into two or more wafer pieces (A, B,...).
(2) The wafer piece A of the divided wafer is put into a heat treatment furnace held at a temperature T1 [° C.] selected from a temperature range of 600 to 900 ° C.
(3) The temperature is increased from T1 [° C.] to a temperature T2 [° C.] of 1000 ° C. or more at a temperature increase rate t [° C./min] until the oxygen precipitate in the wafer piece A grows to a detectable size. Hold (however, t ≦ 3 ° C./min).
(4) The wafer piece A is taken out from the heat treatment furnace, and the oxygen precipitate density inside the wafer is measured.
(5) Another wafer piece B of the divided wafer is put into a heat treatment furnace held at a temperature T3 [° C.] selected from a temperature range of 800 to 1100 ° C. (provided that T1 <T3 <T2). .
(6) The temperature is increased from T3 [° C.] to the T2 [° C.] at the rate of temperature increase t [° C./min] and maintained until oxygen precipitates in the wafer grow to a detectable size.
(7) The wafer piece B is taken out from the heat treatment furnace, and the oxygen precipitate density inside the wafer is measured.

  Conventionally, in the case of a wafer whose defect area is unknown, a method for judging from which defect area the wafer was made has not been established, so it is predicted what oxygen precipitation behavior will be exhibited in the device process However, according to the above-described defect region evaluation method, it is possible to evaluate the defect region of the silicon wafer whose pulling condition is unknown and the defect region is unknown. It became possible to predict.

Hereinafter, the present invention will be described in more detail.
The present inventors investigated the thermal stability of grown-in precipitation nuclei by conducting the following experiments.
First, several types of wafers having different defect areas were prepared, and these wafers were subjected to the following heat treatment.
After inserting the wafer into a furnace set at T ° C. (T = 700, 800, 900, 1000), the temperature was raised from T ° C. to 1050 ° C. at a rate of 1.5 ° C./min, and then at 1050 ° C. for 4 hours. Retained. In this heat treatment, the growth speed of a slow rate is increased to a size at which a stable grown-in precipitation nucleus at T ° C. or higher does not disappear at 1050 ° C., and further maintained at 1050 ° C. for 4 hours. Grow until.

  The important point is that the conditions are such that growth-in precipitation nuclei are sufficiently grown by optimizing the heating rate and that no new precipitation nuclei are generated in the heat treatment step. Therefore, the oxygen precipitate (BMD: Bulk Micro Defect) density after this heat treatment shows a stable grown-in precipitation nucleus density at T ° C. or higher. The BMD density after the heat treatment was measured by infrared scattering tomography (LST). The measurement position is 10 mm from the edge to the center at 10 mm intervals, and the depth is an area of about 50 μm to 180 μm from the surface.

  As a result of the experiments described above, the thermal stability of the grown-in precipitation nuclei can be influenced by the defect region (inside the ring, on the ring, outside the ring), the oxygen concentration, and the position in the crystal axis direction using the OSF ring as an index. all right. The results are shown below.

(1) Relationship between Grown-In Precipitation Nuclei and Defect Region FIG. 1 shows the relationship between the heat treatment start temperature T ° C. and the BMD density. The symbol indicates the low oxygen product (12-14 ppma), and the white symbol indicates the high oxygen product (15-17 ppma). The difference in symbol shape indicates the difference in wafer type (crystal pulling condition), but these differences are not discussed here.
For example, when the BMD density is 1 × 10 ⁹ / cm ³ at 700 ° C., the density of grown-in precipitation nuclei that can remain at 700 ° C. is 1 × 10 ⁹ / cm ³ . . Theoretically, the higher the temperature, the larger the size (critical size) of precipitation nuclei that can remain at that temperature. Large precipitation nuclei that can remain at high temperatures can remain at low temperatures. Therefore, the BMD density at 700 ° C. is the density of all precipitation nuclei that can remain at a temperature of 700 ° C. or higher.

(1-1) OSF ring inner area (V-rich area)
FIG. 1A shows the result inside the OSF ring. The higher the heat treatment start temperature, the lower the BMD density. That is, the larger the size of the precipitation nuclei, the lower the density. In particular, at 900 ° C. or higher, it is 10 ⁶ / cm ³ or less and is extremely low. This shows that the density of thermally stable relatively large grown-in precipitation nuclei is extremely low in the inner region of the OSF ring. Further, since the temperature dependence is strong, it is estimated that the BMD density is greatly different when the device process conditions (first-stage heat treatment temperature) are different.

(1-2) OSF ring and NV region The results on the OSF ring and the NV region are shown in FIGS. 1B and 1C, respectively. Both showed the same tendency. It can be seen that the BMD density at 900 ° C. or higher is clearly higher than that inside the OSF ring. That is, the thermally stable precipitation nucleus density is high. Since the temperature dependence is weak, it can be seen that the BMD density does not change greatly even if the device process conditions are different. There is a significant difference between whether or not OSF is generated by high-temperature oxidation between the OSF ring and the NV region. This difference is considered to be due to the difference in the density of precipitation nuclei stable at a temperature higher than 1000 ° C.

(1-3) NI and I-rich region The results in the NI and I-rich region are shown in FIG. Although the number of data is small, the tendency is almost the same as the case inside the OSF ring.

  From the above results, it was found that, for example, by measuring the BMD density in the heat treatment at the start temperatures of 800 ° C. and 1000 ° C., it can be determined which defect region the wafer is in.

(2) Dependence of oxygen concentration (2-1) Considering the influence of crystal position The oxygen concentration dependency of BMD density is shown in FIG. Differences in symbols (circles, triangles, and squares) indicate differences in crystal positions, which were classified from 0 to 40 cm, 40 to 80 cm, and 80 cm from the crystal shoulder.
FIG. 2A shows the difference between the BMD density at 700 ° C. and the BMD density at 800 ° C. This difference indicates the density of only grown-in precipitation nuclei that can remain at 700 ° C. but cannot remain at 800 ° C., that is, very small precipitation nuclei. From this result, it can be seen that the density of small precipitation nuclei is strongly dependent on the oxygen concentration, and the density decreases as the oxygen concentration decreases. Further, there is crystal position dependency, and it can be seen that the density decreases at a position of 80 cm or more from the crystal shoulder.

As shown in FIG. 2B, the oxygen concentration dependence is observed even in the density of the precipitated nuclei stable at 800 to 900 ° C., but the influence of the crystal position does not appear clearly.
On the other hand, as shown in FIGS. 2C and 2D, the density of only large precipitation nuclei that is stable at 900 to 1000 ° C. and 1000 ° C. or higher has almost no oxygen concentration dependency or crystal position dependency. I understand.

(2-2) Considering Effect of Defect Region FIG. 3 shows the result of considering the effect of the defect region in the oxygen concentration dependence of the BMD density. Data near the boundary between defect areas was omitted. The influence of the defect region clearly appears as the temperature region increases, that is, as the precipitation nucleus size increases. As shown in FIGS. 3C and 3D, the density of large precipitation nuclei that are stable at 900 to 1000 ° C. and 1000 ° C. or higher is clearly higher on the OSF ring and in the NV region. However, there is almost no oxygen concentration dependency.

When the results of (2-1) and (2-2) are combined, the following can be understood.
The density of relatively small grown-in precipitation nuclei depends strongly on the oxygen concentration and the crystal position, but is hardly affected by the defect region. On the other hand, the density of large grown-in precipitation nuclei that are stable at high temperatures hardly depends on the oxygen concentration or crystal position, but strongly depends on the defect region.

(3) In-plane uniformity of oxygen precipitation As described above, it was found that the thermal stability of the grown-in precipitation nuclei strongly depends on the defect region using the OSF ring as an index. Therefore, it can be easily imagined that in-plane uniformity of oxygen precipitation is deteriorated in a wafer including a plurality of defect regions. The results are shown in FIGS.

(3-1) In the case of a high oxygen product FIG. 4 shows the results of a high oxygen product (15 to 17 ppma). The difference in the symbols indicates the difference in the heat treatment start temperature. In a wafer including a plurality of defect regions (FIGS. 4C to 4F), the in-plane uniformity of the BMD density is deteriorated when the temperature is high. This is because, as described in (2), the density of large grown-in precipitation nuclei that is stable at high temperatures is strongly influenced by the defect region. However, the in-plane uniformity improves as the temperature decreases. This is because the density of small grown-in precipitation nuclei is hardly affected by the defect region and strongly depends on the oxygen concentration. Considering the influence of the device process from this result, the in-plane uniformity of the BMD density is not deteriorated in a low temperature process (700 to 800 ° C.) with a low initial stage temperature, but in a high temperature process (to 900 ° C.) with a high initial stage temperature. It is presumed that the uniformity will deteriorate. This is considered to be a problem in the conventional low defect crystal.

(3-2) Case of Low Oxygen Product FIG. 5 shows the result of a low oxygen product (12 to 14 ppma). Compared with the high oxygen product, the low oxygen product has a poor in-plane distribution of BMD density even when the heat treatment start temperature is low. This is because, when the oxygen concentration is low, the density of small precipitation nuclei decreases, and large precipitation nuclei that are strongly influenced by the defect region become dominant at any temperature. From this result, it is suggested that the in-plane uniformity of the BMD density is deteriorated in any device process in the low oxygen product.

  The present inventor has repeatedly studied earnestly based on the knowledge obtained by the above (1) to (3), and a method for stably obtaining a BMD density in all device processes from a high temperature process to a low temperature process is described below. In view of the above, the present invention has been conceived. A conceptual diagram related to the following consideration is shown in FIG.

<Discussion 1> Method for reducing variation in BMD density due to crystal position A major problem in oxygen precipitation control is that there is large variation due to crystal position. In this experiment, it was found that the influence of the crystal position is remarkable for the density of small precipitation nuclei that are stable at 700 to 800 ° C. This precipitation nucleus is considered to be formed in a temperature zone of 700 ° C. or lower of the crystal thermal history. In other words, in order to reduce the variation due to the crystal position, the heat history of 700 ° C. or less may be made the same at the top portion (shoulder side (K side)) and the bottom portion (tail side (P side)) of the crystal. This is extremely difficult. Therefore, it is considered that the variation in crystal position is reduced if the density is lowered. From the result of FIG. 2, in order to reduce the density of small precipitation nuclei, the oxygen concentration needs to be 14 ppma or less. When it exceeds 14 ppma, it becomes impossible to reduce the variation due to the crystal position, which is the object of the present invention. The idea of eliminating the crystal position dependence of oxygen precipitation is also applied to the case of nitrogen doping, and the oxygen concentration may be made 14 ppma or less.

<Discussion 2> Method for forming thermally stable grown-in precipitation nuclei To obtain BMD stably even if the device process is different, large thermally stable precipitation nuclei are required. The density of large precipitation nuclei depends strongly on the defect region, and increases on the OSF ring and in the NV region. However, since there is a possibility that OSF may occur in the case of a high-temperature process on the OSF ring, the optimum region is considered to be the NV region.

In combination with consideration 1, a wafer in which BMD can be stably obtained without depending on the crystal position and device process has an oxygen concentration of 14 ppma or less and the entire surface is an NV region (or the entire surface is an NV region including an OSF ring region), or It can be said that this is a wafer whose entire surface becomes an OSF ring region. However, in the conventional NV region, the density of precipitation nuclei is on the order of 10 ⁷ / cm ^3, which is not necessarily sufficient.

<Consideration 3> Method of increasing the density of precipitation nuclei in the NV region As shown in FIG. 3, the density of thermally stable large precipitation nuclei hardly depends on the oxygen concentration. Cannot be expected.
Here, consider the mechanism by which stable precipitation nuclei are formed in the NV region. FIG. 7 shows a conceptual diagram thereof. By controlling the crystal pulling condition: F / G (F: pulling rate, G: temperature gradient near the growth interface), the supersaturation degree of vacancies is reduced in the NV region, and the formation of voids is suppressed. As a result, at a temperature lower than the void formation temperature zone, the NV region becomes more vacant than the region where the void is formed. The phenomenon that oxygen precipitation nucleation at a relatively high temperature is promoted by excessive vacancies has been confirmed by various experiments. That is, in the NV region, it is considered that the formation of precipitation nuclei at a high temperature (which seems to be in the range of about 1000 to 750 ° C.) is promoted by excess vacancies. When nuclei are formed at a high temperature, they can grow sufficiently in the subsequent cooling process, so that they become thermally stable and large sized precipitation nuclei.

  Based on the mechanism described above, the density of thermally stable precipitation nuclei is increased by increasing the vacancy concentration. However, since the void formation is promoted when the vacancy concentration is increased, the vacancy concentration contributing to the formation of precipitation nuclei is lowered as a result. Therefore, there is a need for some method that can suppress the aggregation even when the pore concentration is high.

  Therefore, the idea was to dope nitrogen. As shown in FIG. 6, if nitrogen doping is performed, the aggregation of vacancies is suppressed and the remaining excess vacancy promotes the formation of precipitation nuclei at a high temperature, resulting in the formation of thermally stable precipitation nuclei. The density will increase. However, the OSF ring inner region cannot be used because it deteriorates device characteristics due to frequent occurrence of minute COPs (minute void defects). Therefore, the NV region of the nitrogen-doped crystal is preferable, and there is an effect that a stable oxygen precipitate can be obtained without depending on the device process.

  According to the present invention, stable oxygen precipitation can be obtained without depending on the crystal position and the device process, so that a wafer having a stable gettering ability with little variation in oxygen precipitate density can be obtained. Furthermore, by using the evaluation method of the present invention, it is possible to relatively easily determine a defect region of a silicon wafer whose pulling condition is unknown and whose defect region is unknown.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
First, a configuration example of a single crystal pulling apparatus by the CZ method used in the present invention will be described with reference to FIG. As shown in FIG. 9, the single crystal pulling apparatus 30 includes a pulling chamber 31, a crucible 32 provided in the pulling chamber 31, a heater 34 disposed around the crucible 32, and a crucible that rotates the crucible 32. Holding shaft 33 and its rotation mechanism (not shown), seed chuck 6 holding silicon seed crystal 5, wire 7 pulling up seed chuck 6, and winding mechanism (not shown) for rotating or winding wire 7 Z). The crucible 32 is provided with a quartz crucible on the inner side containing the silicon melt (hot water) 2 and on the outer side with a graphite crucible. A heat insulating material 35 is disposed around the outside of the heater 34.

Moreover, in order to set the manufacturing conditions related to the manufacturing method of the present invention, an annular solid-liquid interface heat insulating material 8 is provided on the outer periphery of the solid-liquid interface 4 of the crystal 1, and an upper surrounding heat insulating material 9 is disposed thereon. Yes. The solid-liquid interface heat insulating material 8 is provided with a gap 10 of 3 to 5 cm between its lower end and the molten metal surface 3 of the silicon melt 2. The upper surrounding heat insulating material 9 may not be used depending on conditions. Further, a cylindrical cooling device (not shown) that cools the single crystal by blowing cooling gas or blocking radiant heat may be provided.
In addition, recently, a magnet (not shown) is installed outside the pulling chamber 31 in the horizontal direction, and a magnetic field in the horizontal direction or the vertical direction is applied to the silicon melt 2 to suppress the convection of the melt. The so-called MCZ method is often used to achieve stable growth.

  Next, a nitrogen-doped single crystal growth method will be described as an example of a single crystal growth method using the single crystal pulling apparatus 30 described above. First, in a crucible 32, a high-purity polycrystalline raw material of silicon is heated to a melting point (about 1420 ° C.) or higher and melted. At this time, in order to dope nitrogen, for example, a silicon wafer with a nitride film is introduced. Next, the tip of the seed crystal 5 is brought into contact with or immersed in the approximate center of the surface of the melt 2 by unwinding the wire 7. Thereafter, the crucible holding shaft 33 is rotated in an appropriate direction, and the winding seed crystal 5 is pulled up while rotating the wire 7, thereby starting single crystal growth. Thereafter, the single crystal rod 1 doped with substantially cylindrical nitrogen can be obtained by appropriately adjusting the pulling speed and temperature.

  In this case, in the present invention, in order to control the temperature gradient in the crystal, as shown in FIG. 9, the gap 10 between the lower end of the solid-liquid interface heat insulating material 8 and the molten metal surface 3 of the silicon melt 2 is used. In the outer peripheral space of the liquid portion in the single crystal rod 1 on the molten metal surface of the pulling chamber 31, the temperature of the crystal near the molten metal surface is, for example, in a temperature range from 1420 ° C to 1400 ° C. The liquid interface heat insulating material 8 is provided, and the upper surrounding heat insulating material 9 is disposed thereon. Furthermore, if necessary, a device for cooling the crystal is provided above the heat insulating material, and the crystal can be cooled by blowing a cooling gas from above, and a radiant heat reflector is attached to the bottom of the cylinder for control It may be.

Hereinafter, specific embodiments of the present invention will be described by way of examples, but the present invention is not limited to these examples.
Example 1
By using the CZ method, raw material polycrystalline silicon is charged into a quartz crucible having a diameter of 24 inches, and the F / G is controlled so that a single crystal rod having a region where the entire surface becomes an NV region is formed. The single crystal rod of the type, orientation <100>, and interstitial oxygen concentration of 12 to 14 ppma (converted to JEIDA (Japan Electronic Industry Association)) was pulled up. At this time, the oxygen concentration is controlled by controlling the rotation of the crucible during pulling, and two kinds of single crystal rods are added depending on whether or not a silicon wafer having a predetermined amount of silicon nitride film in advance is put in the polycrystalline raw material. Thus, mirror-polished wafers (nitrogen doped wafer and nitrogen non-doped wafer) whose entire surface is composed of the NV region were produced from these single crystal rods.

The nitrogen-doped wafer is sliced from a position where the nitrogen concentration calculated from the amount of nitrogen introduced into the raw material polycrystal and the segregation coefficient of nitrogen is 1 × 10 ¹⁴ pieces / cm ³ and processed into a wafer. .

  For these wafers, instead of the first stage heat treatment in the device process, the wafer was inserted into a furnace set at 1000 ° C. and then heated from 1000 ° C. to 1050 ° C. at a rate of 1.5 ° C./min. Then, heat treatment was performed for 4 hours. And the BMD density after heat processing was measured by the infrared scattering tomography method (LST), respectively. The measurement position is 10 mm from the edge to the center at 10 mm intervals, and the depth is an area of about 50 to 180 μm from the surface.

As a result, BMD density, nitrogen-doped wafer is 3 × 10 ⁹ ~8 × 10 ⁹ pieces / cm ^3, a nitrogen undoped wafer was 2 × 10 ⁷ ~5 × 10 ⁷ cells / cm ^3. Therefore, it was found that even if a relatively high temperature heat treatment was performed in the first stage of the device process, all the wafers had a considerable amount of BMD density with excellent in-plane uniformity. That is, this shows that large grown-in oxygen precipitation nuclei having a stable size at a high temperature are uniformly formed in both the wafers. Further, in the case of a nitrogen-doped wafer, a considerably high density BMD was obtained, and it was found that the gettering effect was extremely high.

(Example 2)
A wafer having an unknown defect area is divided into two, and a piece is inserted into a furnace set at 800 ° C., and then heated from 800 ° C. to 1050 ° C. at a rate of 1.5 ° C./min. Holding for a time, the nuclei were grown to a detectable size.
Similarly, after inserting the remaining piece into a furnace set at 1000 ° C., the temperature was raised from 1000 ° C. to 1050 ° C. at a rate of 1.5 ° C./min, and kept at 1050 ° C. for 4 hours to detect precipitation nuclei. Grown to possible size.

  And the BMD density after heat processing was measured by the infrared scattering tomography method (LST), respectively. The measurement position is 10 mm from the edge to the center at 10 mm intervals, and the depth is an area of about 50 to 180 μm from the surface.

The BMD density was 1 × 10 ⁹ pieces / cm ^{3 when} heated from 800 ° C., and 3 × 10 ⁶ pieces / cm ^{3 when} heated from 1000 ° C. That is, since the difference in BMD density between 800 ° C. and 1000 ° C. is two orders of magnitude or more, it can be determined that the defect region of the wafer is a V-rich region inside the OSF ring.

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

  For example, in the above-described embodiment, the case where a silicon single crystal having a diameter of 8 inches is grown has been described as an example. However, the present invention is not limited to this, and the present invention is not limited to this. The present invention can also be applied to the above silicon single crystal.

It is a figure showing the relationship between heat processing start temperature and BMD density. (A) OSF ring inner region, (b) On the OSF ring, (c) NV region, (d) NI region and I-rich region. It is a figure which shows the oxygen concentration dependence of BMD density. (A) Density difference between BMD density at 700 ° C. and BMD density at 800 ° C., that is, density distribution by crystal position of only minimal precipitation nuclei, (b) BMD density difference between 800 ° C. and 900 ° C. and by crystal position Density distribution, (c) BMD density difference between 900 ° C. and 1000 ° C. and density distribution by crystal position, (d) BMD density distribution by crystal position at 1000 ° C. or higher. It is a figure which shows the result which considered the influence of the defect area | region in the oxygen concentration dependence of BMD density. (A) BMD density difference between 700 ° C. and 800 ° C. and BMD density distribution by defect region, (b) BMD density difference between 800 ° C. and 900 ° C. and BMD density distribution by defect region, (c) 900 ° C. and 1000 ° C. BMD density difference and BMD density distribution by defect area, (d) BMD density distribution by defect area at 1000 ° C. or higher. It is a figure which shows in-plane distribution of the BMD density of a high oxygen product [(a)-(f)]. It is a figure which shows in-plane distribution of the BMD density of a low oxygen product [(a)-(h)]. It is explanatory drawing which shows the method of reducing the variation by the crystal position of oxygen precipitation nucleus density. It is explanatory drawing which shows the method of increasing the oxygen precipitation nucleus density in NV area | region. FIG. 5 is a distribution diagram of various defects when the radial position of the crystal in the silicon single crystal is the horizontal axis and the F / G value is the vertical axis. It is a schematic explanatory drawing of the single crystal pulling apparatus by CZ method used by this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Growing single crystal rod, 2 ... Silicon melt, 3 ... Hot water surface, 4 ... Solid-liquid interface,
5 ... Seed crystal, 6 ... Seed chuck, 7 ... Wire, 8 ... Solid-liquid interface heat insulating material,
9: Upper Go insulator, 10 ... Gap between hot water surface and lower end of solid-liquid interface insulator,
30 ... Single crystal pulling device, 31 ... Pulling chamber, 32 ... Crucible,
33 ... crucible holding shaft, 34 ... heater, 35 ... heat insulating material.

Claims (1)

A method for evaluating a defect region of a silicon wafer manufactured by the Czochralski method, wherein a defect region of a silicon wafer to be evaluated is evaluated by comparing the density of at least two oxygen precipitates measured by the following steps Method.
(1) A wafer to be evaluated is divided into two or more wafer pieces (A, B,...).
(2) The wafer piece A of the divided wafer is put into a heat treatment furnace held at a temperature T1 [° C.] selected from a temperature range of 600 to 900 ° C.
(3) The temperature is increased from T1 [° C.] to a temperature T2 [° C.] of 1000 ° C. or more at a temperature increase rate t [° C./min] until the oxygen precipitate in the wafer piece A grows to a detectable size. Hold (however, t ≦ 3 ° C./min).
(4) The wafer piece A is taken out from the heat treatment furnace, and the oxygen precipitate density inside the wafer is measured.
(5) Another wafer piece B of the divided wafer is put into a heat treatment furnace held at a temperature T3 [° C.] selected from a temperature range of 800 to 1100 ° C. (provided that T1 <T3 <T2). .
(6) The temperature is increased from T3 [° C.] to the T2 [° C.] at the rate of temperature increase t [° C./min] and maintained until oxygen precipitates in the wafer grow to a detectable size.
(7) The wafer piece B is taken out from the heat treatment furnace, and the oxygen precipitate density inside the wafer is measured.

Images