JP4510948B2 - Manufacturing method of silicon single crystal wafer - Google Patents

Description

【００９０】
【表２】

【００９１】
【表３】

【００９２】
これらの結果から明らかなように、本発明の実施例では、酸素濃度、不純物濃度、ＣＯＰの何れの分布状態も特有の形態を示し、何れのウェーハも信頼性試験の合格率は６０％以上となり、結晶欠陥の少ない、半導体デバイス製造に適したウェーハであることが判った。特に、酸素濃度、不純物濃度、ＣＯＰの何れもが明確な分布形態を示す場合には、信頼性試験の合格率は７０％以上となり、結晶欠陥の極めて少ない、半導体デバイス製造に適したウェーハであることが判った。
【００９３】
一方、比較例では、人為的な変動を加えていないため、酸素濃度、不純物濃度、ＣＯＰの何れもがウェーハ全面に均一に分布（円盤状）しており、信頼性試験の合格率も１０％以下と極めて悪い結果となり、結晶欠陥の極めて多い、半導体デバイス製造に適さないウェーハであることが判った。
【００９４】
【発明の効果】
本発明のシリコン単結晶ウェーハの製造方法は、従来のシリコン単結晶製造装置に改造を加えることなく適用できるため、製造コストを上昇させることなく、高品質のシリコン単結晶ウェーハを歩留良く製造できる。
【図面の簡単な説明】
【図１】 実施例１〜７及び比較例１で用いたＣｚ法シリコン単結晶製造装置の模式図。
【図２】 実施例８〜１０及び比較例２で用いたカスプ磁場印加Ｃｚシリコン単結晶製造装置の模式図。
【図３】 実施例１１〜１２及び比較例３で用いた水平磁場印加Ｃｚシリコン単結晶製造装置の模式図。
【図４】 実施例１１における結晶回転の人為的変動の模式図。
【図５】 ８ｉｎウェーハにおける不純物濃度の変動分布の一例。
【符号の説明】
１ … Ｃｚ無転位シリコン単結晶引上炉
２ … ワイヤ巻上機
３ … 断熱材
４ … 加熱ヒーター
５ … 回転治具
６ … ルツボ
６ａ … 石英ルツボ
６ｂ … 黒鉛ルツボ
７ … ワイヤ
８ … 種結晶
９ … チャック
１０ … ガス導入口
１１ … ガス排出口
２０ … カスプ磁場印加用同軸対向電磁石
３０ … 水平磁場印加用電磁石[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a silicon single crystal wafer by the Czochralski method (hereinafter referred to as Cz method) with few crystal defects. Ha It relates to a manufacturing method. In particular, a silicon single crystal wafer with less COP (Crystal Originated Particle) Ha It relates to a manufacturing method.
[0002]
[Prior art]
In the silicon single crystal pulled by the conventional Cz method, crystal defects are formed in the crystal during the single crystal pulling. Crystal defects formed during single crystal pulling are referred to as grown-in defects, and the grown-in defects are given various names depending on the detection method. For example, what is detected as a pit of a special ripple pattern (flow pattern) appearing on the wafer surface by immersing a silicon wafer in a Seco solution that is a selective etching solution is called a flow pattern defect. An infrared laser incident on a silicon wafer and detected by scattering is called an infrared scattering defect. Furthermore, about 0.05 to 0.20 μm (about 0.15 μm) is formed on the wafer surface after cleaning in a mixed solution of ammonia water and hydrogen peroxide water, which is generally performed for particle removal in the silicon wafer manufacturing process. What is detected by the particle inspection device as a pit having a size of) is called COP. Since COP was first discovered by a particle inspection device, it was named “particle” due to the crystal, but subsequent investigations revealed that the defect was not a particle but a pit. .
[0003]
By the way, when a silicon wafer having a COP is used for a semiconductor device and a gate oxide film having a MOS structure, which is a basic element structure of the semiconductor device, is formed on a position where the COP is present, the reliability is significantly lowered. Further, when the wiring is formed, the pattern is cut off. In particular, since the element structure of the mainstay device (64M-DRAM) is as very fine as 0.3 μm, a pit of about 0.15 μm causes a serious defect and significantly reduces the manufacturing yield of the device. Therefore, it is desirable to reduce the COP of the silicon wafer as much as possible.
[0004]
Recent studies have revealed that the origins of grown-in defects such as flow pattern defects, infrared scattering defects, and COPs are the same, and their substance is an octahedral void. This void is when atomic vacancies (a kind of point defect) taken in immediately after the single crystal is solidified becomes supersaturated as the single crystal is cooled, and eventually passes through a temperature range of about 1100 ° C to 1050 ° C. It has become clear that it is formed by agglomeration.
[0005]
Based on the above knowledge, as a method for reducing Grown-in defects, many proposals have been made for improving the cooling condition of the single crystal after solidification. For example, Japanese Patent Application Laid-Open No. 3-275598 discloses a method of increasing the cooling rate to 900 ° C. after the crystal is solidified to be higher than 1.2 ° C./min in order to reduce COP. Japanese Patent No. 42893 discloses a method of reducing the cooling rate from 1200 ° C. to 800 ° C. below 0.4 ° C./min in order to reduce COP. In order to improve the reliability, a method of setting the temperature gradient in the range from directly above the solid-liquid interface of the silicon melt to 3 cm to 5 ° C./cm or more is disclosed. Japanese Patent Laid-Open No. 6-279188 discloses COP or flow pattern In order to reduce defects or infrared scattering defects, a method of holding a silicon single crystal at a temperature between 1420 ° C. and 1200 ° C. for 1 hour or more is shown. In order to reduce infrared scattering defects or improve the reliability of the gate oxide film, the temperature gradient in the crystal axis direction in the temperature range from the silicon melt to 1300 ° C. is G (° C./mm), and the single crystal Lifting speed is f _p F (mm / min) _p / G is 0.25mm ² A method of setting the cooling rate from 1150 ° C. to 1000 ° C. to 2.0 ° C./min or less is shown.
[0006]
However, all of these methods are techniques for reducing the grown-in defects by changing the cooling conditions of the single crystal. In order to provide the same effect uniformly over the entire length of the single crystal, the structure of the pulling furnace and the pulling up It was necessary to change the internal structure of the furnace, leading to increased costs. In these methods, the size of the void defect, which is the origin of the Grown-in defect, can be apparently reduced or the density can be apparently reduced. However, the atoms that are constituent elements (raw materials) of the void defect The vacancy concentration itself could not be lowered. Therefore, there has been a demand for a method for reducing the atomic vacancy concentration itself during single crystal growth, but there has been no such method in the past.
[0007]
In a general Cz method, polycrystalline silicon is dissolved in a quartz crucible, and a silicon single crystal is seeded in the melt, and then pulled up. Since the melt is always in contact with the quartz crucible, it contains a large amount of oxygen atoms that have melted out of the quartz crucible, and the silicon single crystal pulled up from the melt also contains a large amount of supersaturated solid solution oxygen atoms. ing.
[0008]
On the other hand, in addition to a method for producing a silicon single crystal by a general Cz method, there is a method for pulling up a single crystal by a Cz method while applying a magnetic field to a melt (MCZ method, Magnetic field applied Cz method).
[0009]
The MCZ method includes a cusp magnetic field application MCZ method and a horizontal magnetic field application MCZ method, depending on how the magnetic field is applied. The MCZ method has been proposed as a technique for controlling the oxygen concentration by controlling the flow of the melt with a magnetic field.
[0010]
A single crystal pulling apparatus based on the MCZ method with cusp magnetic field application is disclosed in, for example, Japanese Patent Application Laid-Open No. 58-217493, and the same-pole opposed magnets (for example, superconducting magnets) and magnetic field applying apparatuses Furnace body provided between the two electrodes, a quartz crucible provided inside the furnace body for holding the silicon melt, a heater for heating the silicon melt, and a pulling mechanism for pulling up the silicon single crystal rod while rotating And is configured. An equiaxed symmetrical and radial cusp magnetic field is formed in the melt by the same-pole opposed magnet. By applying a cusp magnetic field, the thermal convection of the silicon melt in the quartz crucible is uniformly suppressed and an axially symmetric temperature distribution is obtained, so that a single crystal having uniform crystallinity can be pulled up. Further, since convection of the melt in the radial direction can be suppressed, mixing of oxygen from the quartz crucible can be reduced. However, even though the conventional cusp magnetic field application MCZ method is used to reduce the oxygen concentration, it has not been applied to the reduction of grown-in defects such as COP.
[0011]
Further, a single crystal pulling apparatus based on the horizontal magnetic field application MCZ method is disclosed in, for example, Japanese Patent Laid-Open No. 56-45889, and between a magnetic field applying apparatus that generates a magnetic field in the horizontal direction and both poles of the magnetic field applying apparatus. Consists of a furnace body provided, a quartz crucible provided inside the furnace body for holding the silicon melt, a heater for heating the silicon melt, and a pulling mechanism for pulling up the silicon single crystal rod while rotating Has been. Suppressing the thermal convection of the silicon melt in the quartz crucible by applying a horizontal magnetic field greatly reduces oxygen contamination from the quartz crucible and keeps the silicon solid-liquid interface in a more static state. be able to. However, the conventional horizontal magnetic field application MCZ method is used for reducing the oxygen concentration, but has not been applied to reducing grown-in defects such as COP.
[0012]
In addition, the COP in-plane distribution of a conventional silicon single crystal wafer in which the grown-in defects are reduced by improving the cooling method or the like has a disk-like form existing inside a circle having the same center as the wafer. However, it was not distributed only on concentric circles, nor was it distributed in a spiral, radial, or combination thereof. When a wafer in which COP is distributed on such a line is used for a semiconductor device, the number of devices affected by COP is less than that of a conventional wafer in which COP is distributed in a disk shape, and the device yield is improved. . Therefore, there has been a demand for a wafer in which COP is distributed on the line as described above, but it has not existed in the past.
[0013]
[Problems to be solved by the invention]
In view of the above-described problems, the present invention provides a silicon single crystal wafer having few crystal defects, particularly few grown-in defects such as COP. Ha An object is to provide a manufacturing method.
[0014]
[Means for Solving the Problems]
The object of the present invention is achieved by the following means.
[0028]
( 1 ) In the manufacturing method of silicon single crystal wafer by Czochralski method, silicon single crystal ingot During the training of Over its entire length, A method for producing a silicon single crystal wafer, wherein a magnetic field strength of a cusp magnetic field or a zero magnetic field position is applied continuously or periodically.
[0029]
( 2 The magnetic field application is an application of a cusp magnetic field of 100 gauss or more on the wall of the quartz crucible containing the melt ( 1 ) A method for producing a silicon single crystal wafer as described above.
[0031]
DETAILED DESCRIPTION OF THE INVENTION
When a silicon single crystal is grown by the Cz method, point defects (atomic vacancies and interstitial atoms) are taken into the single crystal from the solid-liquid interface. As the single crystal is cooled, the vacancies and interstitial atoms decrease by pair annihilation, but the concentration taken into the single crystal is several orders of magnitude higher in the single crystal. The remaining point defects become atomic vacancies. These atomic vacancies become supersaturated as the single crystal is cooled, and when the single crystal is 1100 ° C. or lower, the agglomerates are more stable than the single crystals, and form aggregates. A void defect, which is the origin of a grown-in defect such as COP, is an aggregation of supersaturated atomic vacancies at 1100 ° C. or lower.
[0032]
The inventors of the present invention considered that it is necessary to reduce the concentration of atomic vacancies, which are constituent elements of the void defect, in order to drastically reduce the void defect in the silicon single crystal wafer, and conducted an extensive study. As a result, in order to reduce the concentration of atomic vacancies in the wafer, it has been found that it is important to have a specific form of defect distribution. That is, a wafer obtained from a silicon single crystal grown by the Cz method, and the in-plane distribution of COP, oxygen concentration and / or impurity concentration seen in the wafer is one of concentric, spiral and radial types. Alternatively, when a semiconductor device is manufactured using a silicon single crystal wafer of two or more types, the region where a grown-in defect such as COP exists is linear (band-like), so the influence of defects such as COP The number of devices that receive them is reduced, and the device yield is improved.
[0033]
In the in-plane distribution form of COP, oxygen concentration and / or impurity concentration, when the interval is 1 mm or more, a region where no grown-in defect exists can be secured widely in the wafer, and the influence of the defect can be reduced. Since the number of devices to be received can be further reduced, the device yield is further improved.
[0034]
Further, the area density of such COPs is 1.6 / cm2 on average for the entire wafer. ² In the case of the following, the yield of the semiconductor device manufactured using the wafer is further improved.
[0035]
Further, when the area of the defect-free region where no COP is present is 20% or more of the entire area of the wafer, the device yield is further improved. Here, the COP existence area is an area within a radius of 4 mm centering on the COP position. Therefore, the area where no COP exists refers to an area obtained by subtracting the COP existing area from the entire wafer surface.
[0036]
The size and number of COPs can be measured by a foreign substance inspection apparatus using laser scattering, or can be directly measured by an atomic force microscope (AFM, Atomic Force Microscopy).
[0037]
The manufacturing method is not particularly limited as long as it is a silicon single crystal wafer having such quality, but in order to manufacture efficiently, it is effective to form a point defect absorbing layer in the single crystal during single crystal growth. Thus, it has been found that a defect-free layer is widely formed around the point-defect absorbing layer. Here, the point defect absorbing layer indicates a region having a high concentration in the density distribution of the point defect concentration. That is, a region where the concentration is high because gettering of surrounding point defects, a region adjacent to a region where point defects are reduced by reaction, a region where the concentration of point defects formed when the single crystal is solidified, etc. Is also included in the point defect absorbing layer.
[0038]
Furthermore, the present inventors have found that fluctuations in oxygen concentration / impurity concentration due to fluctuations in the growth rate of micro single crystals at the solid-liquid interface during the growth of silicon single crystals and fluctuations in the melt temperature at the solid-liquid interface are less than 5%. When growing a silicon single crystal with an assembly structure (hot zone structure) inside the pulling furnace as described above, it was found that an absorption layer of point defects can be formed with extremely good control.
[0039]
In order to more actively form a point defect absorption layer, the growth rate of the micro single crystal at the solid-liquid interface during the growth of the silicon single crystal is changed, or the melt temperature at the solid-liquid interface is changed. It has been found that an absorption layer of point defects can be formed perpendicular to the crystal axis at a certain interval in the crystal axis direction by changing the oxygen concentration / impurity concentration by the change. Here, the impurities refer to metal elements such as carbon, aluminum, copper, and tin, including dopants such as boron and phosphorus.
[0040]
By applying this variation continuously or periodically, point defect absorption layers can be formed at regular intervals in the longitudinal direction of the single crystal, which is effective in reducing Grown-in defects over the entire length of the single crystal. I understood. Here, the term “periodic” refers to stepwise or pulse-like (ie, digital) change when changing from one value to another. Further, “continuous” means gradual (analog) fluctuation. Furthermore, it has been found that the greater the variation, the greater the effect. However, when the variation exceeds 5%, large crystal deformation is likely to occur, making it difficult to produce a single crystal.
[0041]
If the interval giving continuous or periodic fluctuation is less than 0.5 mm in the pulling direction of the single crystal, the pulled single crystal is sliced and a silicon single crystal wafer having a thickness of 0.5 mm or more is obtained. When manufacturing, since it becomes difficult to separate the point-defect absorbing layer and the defect-free layer, the interval giving the fluctuation is preferably 0.5 mm or more.
[0042]
Also, during the growth of the single crystal, by changing the rotation speed of the single crystal or the crucible, the growth rate of the micro single crystal at the solid-liquid interface can be changed relatively easily, or the melt temperature at the solid-liquid interface can be changed. This is effective because the oxygen concentration / impurity concentration can be changed by the change of the above. The absorption layer formed by this method has a concentric, spiral or radial form or a combination of these forms.
[0043]
In particular, when the diameter of the silicon single crystal is D (in), the rotational speed of the single crystal is 2000 / D. ² ~ 600 / D ² (Rpm) to 300 / D ² Reduced to ~ 0 (rpm) in transition time 2 minutes to 1 second, and subsequently the rotation speed of the single crystal is 300 / D ² After growing the single crystal at ˜0 (rpm) for 1 second to 1 minute, the rotational speed of the single crystal is set to 2000 / D. ² ~ 600 / D ² It is preferable to change (rpm) to increase the transition time from 2 minutes to 1 second because a point defect absorption layer can be formed in the single crystal with good controllability.
[0044]
In addition, the pulling speed of the silicon single crystal can be changed continuously or periodically by simply changing the growth rate of the micro single crystal at the solid-liquid interface or by changing the melt temperature at the solid-liquid interface. Since the oxygen concentration / impurity concentration can be varied, a point defect absorption layer can be formed effectively together with other variations, which is preferable.
[0045]
Furthermore, the present inventors consider that it is important to control the generation of crystal defects when no fluctuation is applied in order to more accurately control the formation of the point defect absorption layer in the single crystal. It came to. Therefore, when the above-described operation was performed while a magnetic field was applied to the melt during the growth of the silicon single crystal, the state of the melt when the fluctuation was not applied became more stable and extremely controllable. It was found that a point defect absorbing layer can be formed well.
[0046]
When this magnetic field application is a cusp magnetic field of 100 gauss or more on the wall of the crucible containing the melt, the thermal convection of the melt is uniformly suppressed and an axially symmetric temperature distribution is obtained. There is a remarkable effect in stabilizing the melt. In addition, since the effect which applied the cusp magnetic field is not seen if the magnetic field intensity to apply is less than 100 gauss, it is desirable to apply the magnetic field intensity of 100 gauss or more.
[0047]
In the case of a cusp magnetic field, since the same-pole opposed magnet is used, a position where the magnetic field intensity becomes zero exists on the central axis of the cusp magnetic field (zero magnetic field position). The height of the zero magnetic field position can be easily changed by mechanically moving the entire same-pole counter magnet up and down or electrically changing the generated magnetic field strength of each counter magnet. By changing the magnetic field strength or the zero magnetic field position continuously or periodically with the cusp magnetic field applied, the melt state at the interface can be controlled stably. It is possible to accurately change the growth rate of a single crystal or to change the oxygen concentration / impurity concentration by changing the melt temperature at the solid-liquid interface. Therefore, the controllability of the formation of the point defect absorption layer is excellent. In addition, by changing the pulling speed continuously or periodically with the cusp magnetic field applied, it is possible to control the melt state at the interface in a stable manner. It is possible to accurately change the growth rate of the single crystal or change the oxygen concentration / impurity concentration by changing the melt temperature at the solid-liquid interface. Therefore, the controllability of the formation of the point defect absorption layer is excellent.
[0048]
In addition, when the magnetic field application is a horizontal magnetic field of 500 gauss or more on the wall of the quartz crucible containing the melt, the solid-liquid interface is made more static by suppressing the thermal convection of the melt in the crucible. Therefore, there is a remarkable effect in stabilizing the melt. In addition, since the effect of applying a horizontal magnetic field is not seen if the applied magnetic field strength is less than 500 gauss, it is desirable to apply a magnetic field strength of 500 gauss or more.
[0049]
By changing the magnetic field strength continuously or periodically with a horizontal magnetic field applied, the state of the melt at the interface can be controlled with stability, so the formation of the point defect absorption layer can be controlled. Excellent in properties.
[0050]
The dislocation-free silicon single crystal production apparatus used in the present invention is not particularly limited as long as it can be used for the production of dislocation-free silicon single crystals by a normal Cz method, and a production apparatus as shown in FIG. 1 is used. It was.
[0051]
This Cz method silicon single crystal manufacturing apparatus accommodates a crucible 6 composed of a quartz crucible 6a for accommodating a silicon melt M and a graphite crucible 6b for protecting the same, and a grown dislocation-free silicon single crystal ingot S. A crystal pulling furnace 1. The side surface of the crucible 6 is arranged so that the heat insulating material 3 surrounds the heater 4 and heat from the heater 4 to prevent the heat from the crystal pulling furnace from escaping. The crucible 6 is not shown in the drawing. The crucible 6 is moved up and down in order to compensate for the liquid level drop accompanying the decrease of the silicon melt in the crucible 6 while being connected to the apparatus by the rotating jig 5 and rotated at a predetermined speed by the driving device. Yes. A suspended pulling wire 7 is installed in the pulling furnace 1, and a chuck 9 for holding a seed crystal 8 is provided at the lower end of the wire. An upper end side of the pulling wire 7 is provided with a pulling device that is wound around the wire hoisting machine 2 to pull up the dislocation-free silicon single crystal ingot. Then, Ar gas is introduced into the pulling furnace 1 from a gas inlet 10 formed in the pulling furnace 1, flows through the pulling furnace 1, and is discharged from the gas outlet 11. The reason why the Ar gas is circulated in this way is to prevent SiO generated in the pulling furnace 1 from melting into the silicon melt from being mixed into the silicon melt.
[0052]
Moreover, the cusp magnetic field application silicon single crystal manufacturing apparatus used for this invention used the general manufacturing apparatus which arrange | positions the same-pole opposing magnet 20 up and down on a crystal axis, as shown in FIG.
[0053]
Moreover, the horizontal magnetic field application silicon single crystal manufacturing apparatus used for this invention used the general manufacturing apparatus which arrange | positions the electromagnet pair 30 which can generate | occur | produce a magnetic field horizontally as shown in FIG. .
[0054]
The yield of the semiconductor device manufactured using the silicon single crystal wafer of the present invention can be known from the pass rate of the reliability test of the gate oxide film in the MOS structure which is the basic structure of the device. Since the reliability of the gate oxide film is extremely sensitive to COP, the occurrence rate of the semiconductor device failure caused by COP can be known from the pass rate of the reliability test of the gate oxide film. The specific method of the reliability test of the gate oxide film of the silicon single crystal wafer of the present invention is as follows. A MOS diode is formed on a mirror-finished silicon wafer sample, and the electrical characteristics of a gate oxide film (insulating oxide film) that is a 25.0 nm silicon dioxide film formed in a dry oxygen atmosphere at 1000 ° C. are examined. Went by. The MOS diode has 1 × 10 phosphorous on the gate oxide film. ^{twenty one} atoms / cm ^Three It has a two-layer structure of doped polycrystalline silicon and aluminum, and has an area of 20 mm. ² The gold electrode for the ohmic electrode is formed on the back surface. A MOS diode having the above structure was fabricated on the entire surface of the sample wafer. A DC voltage having a polarity in which majority carriers are injected from the substrate silicon was applied to each MOS diode, and the withstand voltage characteristics of the gate oxide film of the sample wafer were evaluated by a voltage wrapping method. The current density flowing through the gate oxide is 1 mA / cm ² The average electric field (dielectric breakdown field) applied to the gate oxide film was measured. At this time, when the dielectric breakdown electric field is less than 8.0 MV / cm, various studies have revealed that the withstand voltage characteristics of the gate oxide film are lowered due to crystal defects such as COP. Yes. That is, a MOS diode having a dielectric breakdown electric field of 8.0 MV / cm or more is a diode that has passed the gate oxide film test.
[0055]
In the case of a general wafer processed from a silicon single crystal manufactured by a conventional method, the ratio of the number of MOS diodes having a breakdown electric field of 8.0 MV / cm or more to the total number (that is, passing the gate oxide film test) Rate) is at most less than 30%. Therefore, when the pass rate of the gate oxide film test is 60% or more, the reliability of the gate oxide film is excellent. That is, it is excellent as a semiconductor device wafer.
[0056]
【Example】
Hereinafter, the present invention will be described with reference to examples, but it goes without saying that the present invention is not limited by the description of these examples.
[0057]
Note that the use of the “stable hot zone” described here means that the oxygen concentration due to fluctuations in the growth rate of the micro single crystal at the solid-liquid interface during the growth of the silicon single crystal and the melt temperature fluctuation at the solid-liquid interface It means a case where a silicon single crystal is grown using an assembly structure (hot zone structure) inside the pulling furnace such that the variation in impurity concentration is less than 5%. In addition, “fluctuation interval” means an interval in which variation is continuously or periodically accompanied with the growth of a silicon single crystal, as illustrated in FIG. A unit may be used. As shown in FIG. 5, the fluctuation distribution in the wafer resulting from this fluctuation is caused by the continuous or periodic fluctuation given artificially together with the macro base fluctuation and the micro base fluctuation caused by the hot zone. The distribution is superimposed with micro-artificial fluctuations.
[0058]
The “COP area density” is an average density of COPs of 0.10 μm or more in terms of diameter in the entire area of the wafer. The “COP non-existing area ratio” means the ratio of the area of the entire wafer surface to the area obtained by subtracting the COP existing area from the area of the entire wafer surface. The COP existence area is an area within a radius of 4 mm with the COP position as the center.
[0059]
Example 1
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0060]
Absorption layer of point defects: Yes
Stable hot zone: none
Single crystal rotation speed: Decrease from 18 rpm to 1 rpm in 2 seconds, hold at 1 rpm for 5 seconds, then periodically raise from 1 rpm to 18 rpm in 2 seconds
Crucible rotational speed: 8 rpm constant
Average pulling speed: 1.0 mm / min
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Variation interval: 0.3 mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0061]
(Example 2)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0062]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 20 rpm
Crucible rotational speed: 8 rpm constant
Pulling speed: Decreasing from 1.0 mm / min to 0.2 mm / min in 3 seconds, holding at 0.2 mm / min for 10 seconds, then increasing from 0.2 mm / min to 1.0 mm / min in 2 seconds Performed periodically
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Fluctuation interval: 0.4mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0063]
(Example 3)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0064]
Absorption layer of point defects: Yes
Stable hot zone: none
Single crystal rotation speed: 19 rpm constant
Crucible rotation speed: Decrease from 10 rpm to 1 rpm in 3 seconds, hold at 1 rpm for 10 seconds, then continuously increase from 1 rpm to 10 rpm in 3 seconds
Average pulling speed: 0.6mm / min
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Variation interval: 0.5mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0065]
Example 4
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0066]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: Decrease from 18 rpm to 1 rpm in 130 seconds, hold at 1 rpm for 20 seconds, and then periodically increase from 1 rpm to 18 rpm in 130 seconds
Crucible rotational speed: 8 rpm constant
Average pulling speed: 0.2mm / min
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Fluctuation interval: 1.5mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0067]
(Example 5)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0068]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 18 rpm
Crucible rotation speed: Decreasing from 8 rpm to 1 rpm in 2 seconds, holding at 1 rpm for 20 seconds, then continuously increasing from 1 rpm to 10 rpm in 2 seconds
Average pulling speed: 1.0 mm / min
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Change interval: 2.0 mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0069]
(Example 6)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0070]
Absorption layer of point defects: Yes
Stable hot zone: none
Single crystal rotation speed: Decrease from 18 rpm to 1 rpm in 3 seconds, hold at 1 rpm for 20 seconds, and then periodically increase from 1 rpm to 18 rpm in 3 seconds
Crucible rotation speed: 9 rpm constant
Average pulling speed: 1.2mm / min
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Fluctuation interval: 3.5mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0071]
(Example 7)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0072]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 18 rpm
Crucible rotation speed: constant 12 rpm
Pulling speed: An operation of decreasing from 1.0 mm / min to 0.0 mm / min in 5 seconds, holding at 0.0 mm / min for 30 seconds, and then increasing from 0.0 mm / min to 1.0 mm / min in 5 seconds Performed periodically
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Variation interval: 4.0 mm
Tables 1 to 3 show the results of cutting the wafer from the ingot grown under these conditions, measuring the oxygen concentration, the impurity (boron) concentration, the distribution state of the COP, and the reliability test of the gate oxide film.
[0073]
(Example 8)
Using the apparatus shown in FIG. 2, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0074]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: Decrease from 18 rpm to 1 rpm in 2 seconds, hold at 1 rpm for 20 seconds, and then periodically increase from 1 rpm to 18 rpm in 2 seconds
Crucible rotational speed: 8 rpm constant
Average pulling speed: 1.2mm / min
Magnetic field application: cusp magnetic field
Magnetic field strength at quartz crucible wall: 1300 gauss constant
Zero magnetic field position: 50mm below the melt surface position
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Variation interval: 5.2 mm
Tables 1 to 3 show the results of cutting the wafer from the ingot grown under these conditions, measuring the oxygen concentration, the impurity (boron) concentration, the distribution state of the COP, and the reliability test of the gate oxide film.
[0075]
Example 9
Using the apparatus shown in FIG. 2, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0076]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 18 rpm
Crucible rotational speed: 6 rpm constant
Average pulling speed: 0.9mm / min
Magnetic field application: cusp magnetic field
Magnetic field strength at the quartz crucible wall: Decrease from 400 gauss to 300 gauss in 115 seconds, hold at 300 gauss for 40 seconds, then periodically increase from 300 gauss to 400 gauss in 115 seconds
Zero magnetic field position: 40mm below the melt surface position
Variation in growth rate at solid-liquid interface: yes
Variation in melt temperature at the solid-liquid interface: yes
Change interval: 6.0 mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0077]
(Example 10)
Using the apparatus shown in FIG. 2, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0078]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 18 rpm
Crucible rotation speed: 7 rpm constant
Average pulling speed: 1.0 mm / min
Magnetic field application: cusp magnetic field
Magnetic field strength at the quartz crucible wall: 1000 gauss
Zero magnetic field position: Lowering from the melt surface from 40 mm down to 50 mm down in 115 seconds, holding at that position for 40 seconds, and then periodically raising from 50 mm down to 40 mm down in 115 seconds
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Variation interval: 7.0 mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0079]
(Example 11)
Using the apparatus of FIG. 3, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0080]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: Decrease from 18 rpm to 1 rpm in 2 seconds, hold at 1 rpm for 20 seconds, and then periodically increase from 1 rpm to 18 rpm in 2 seconds
Crucible rotation speed: constant 4 rpm
Average pulling speed: 1.0 mm / min
Magnetic field application: horizontal magnetic field
Magnetic field strength at the quartz crucible wall: constant 3000 Gauss
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Fluctuation interval: 8.0mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0081]
(Example 12)
Using the apparatus of FIG. 3, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0082]
Absorption layer of point defects: Yes
Stable hot zone: Yes
Single crystal rotation speed: constant 18 rpm
Crucible rotation speed: constant 1.5 rpm
Average pulling speed: 1.0 mm / min
Magnetic field application: cusp magnetic field
Magnetic field strength at the quartz crucible wall: Decrease from 2000 gauss to 1500 gauss in 115 seconds, hold at 1500 gauss for 60 seconds, and then increase from 1500 gauss to 2000 gauss in 115 seconds
Artificial variation in growth rate at solid-liquid interface: yes
Artificial fluctuation of melt temperature at solid-liquid interface: yes
Fluctuation interval: 8.0mm
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0083]
(Comparative Example 1)
Using the apparatus shown in FIG. 1, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0084]
Absorption layer of point defects: None
Stable hot zone: Yes
Single crystal rotation speed: 30 rpm constant
Crucible rotation speed: constant 10 rpm
Average pulling speed: 0.8mm / min
Artificial variation in growth rate at solid-liquid interface: None
Artificial variation of melt temperature at solid-liquid interface: None
Variation interval: None
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0085]
(Comparative Example 2)
Using the apparatus shown in FIG. 2, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0086]
Absorption layer of point defects: None
Stable hot zone: none
Single crystal rotation speed: constant 25 rpm
Crucible rotational speed: 6 rpm constant
Average pulling speed: 0.8mm / min
Magnetic field application: cusp magnetic field
Magnetic field strength at the quartz crucible wall: 1000 gauss constant
Zero magnetic field position: 40mm below the melt surface position
Artificial variation in growth rate at solid-liquid interface: None
Artificial variation of melt temperature at solid-liquid interface: None
Variation interval: None
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0087]
(Comparative Example 3)
Using the apparatus of FIG. 3, a P-type (boron-doped) silicon single crystal ingot having a diameter of 8 inches was grown under the following conditions.
[0088]
Absorption layer of point defects: None
Stable hot zone: none
Single crystal rotation speed: constant 18 rpm
Crucible rotation speed: 1 rpm constant
Average pulling speed: 0.9mm / min
Magnetic field application: horizontal magnetic field
Magnetic field strength at the quartz crucible wall: constant 3000 Gauss
Artificial variation in growth rate at solid-liquid interface: None
Artificial variation of melt temperature at solid-liquid interface: None
Variation interval: None
Tables 1 to 3 show the results of cutting a wafer from an ingot grown under these conditions, measuring the oxygen concentration, impurity (boron) concentration, and COP distribution and performing a reliability test on the gate oxide film.
[0089]
[Table 1]

[0090]
[Table 2]

[0091]
[Table 3]

[0092]
As is clear from these results, in the examples of the present invention, the distribution states of oxygen concentration, impurity concentration, and COP each show a specific form, and the pass rate of the reliability test is 60% or more for any wafer. The wafer was found to be suitable for semiconductor device manufacturing with few crystal defects. In particular, when all of oxygen concentration, impurity concentration, and COP show a clear distribution form, the pass rate of the reliability test is 70% or more, and the wafer is suitable for semiconductor device manufacturing with very few crystal defects. I found out.
[0093]
On the other hand, in the comparative example, since no artificial variation is applied, all of the oxygen concentration, impurity concentration, and COP are uniformly distributed (disk shape) over the entire wafer surface, and the pass rate of the reliability test is 10%. The following results were extremely bad, and it was found that the wafer was extremely unsuitable for manufacturing semiconductor devices with a large number of crystal defects.
[0094]
【The invention's effect】
Book Since the method for producing a silicon single crystal wafer according to the invention can be applied to a conventional silicon single crystal production apparatus without modification, a high-quality silicon single crystal wafer can be produced without increasing the production cost.
[Brief description of the drawings]
FIG. 1 is a schematic diagram of a Cz method silicon single crystal manufacturing apparatus used in Examples 1 to 7 and Comparative Example 1. FIG.
2 is a schematic diagram of a cusp magnetic field applied Cz silicon single crystal manufacturing apparatus used in Examples 8 to 10 and Comparative Example 2. FIG.
3 is a schematic diagram of a horizontal magnetic field applied Cz silicon single crystal manufacturing apparatus used in Examples 11 to 12 and Comparative Example 3. FIG.
4 is a schematic diagram of an artificial variation of crystal rotation in Example 11. FIG.
FIG. 5 shows an example of an impurity concentration fluctuation distribution in an 8-in wafer.
[Explanation of symbols]
1 ... Cz-free dislocation silicon single crystal pulling furnace
2… Wire hoisting machine
3… Insulation
4 ... Heater
5… Rotating jig
6… crucible
6a Quartz crucible
6b Graphite crucible
7 ... Wire
8 ... Seed crystal
9… Chuck
10 ... Gas inlet
11 ... Gas outlet
20 ... Coaxial counter electromagnet for cusp magnetic field application
30 ... Electromagnet for applying horizontal magnetic field

Claims (2)

In the method for producing a silicon single crystal wafer by the Czochralski method,
A method for producing a silicon single crystal wafer, wherein a magnetic field strength of a cusp magnetic field or a zero magnetic field position is continuously or periodically varied during the growth of a silicon single crystal ingot . The magnetic field applied, the method for manufacturing a silicon single crystal wafer according to claim 1 wherein the application of 100 gauss or more cusp field in the wall of the quartz crucible for accommodating melt.

Images