Method for pulling single crystal ingot

Abstract

A method of pulling a single crystal ingot from a silicon melt according to the Czochralski technology comprises the steps of: (i) providing a silicon melt and a single crystal silicon seed crystal over the silicon melt wherein the silicon melt and / or seed crystal optionally comprise a dopant; (ii) lowering the seed crystal until contact is created between the seed crystal and the silicon melt; and (iii) sequentially pulling the seed cone, the slender neck, the starting cone and the cylindrical portion of the single crystal ingot by raising the seed; the seed crystal cone is characterized in that the upper diameter of the seed crystal cone is not smaller than 12 mm and not larger than 22 mm, the lower diameter of the seed crystal cone is not smaller than 2 mm and not larger than 10 mm, and the axial length of the seed crystal cone is not smaller than 75 mm and not larger than 300 mm.

Description

Technical Field

[0001] The subject of this invention is a method for pulling single crystal ingots, a method for producing semiconductor material wafers, and a single crystal silicon wafer. Background Technology

[0002] Semiconductor single-crystal wafers (slices) are the foundation of modern electronic devices. They are produced in multiple process steps, including pulling single-crystal ingots from melt, grinding and slicing the crystals into wafers, and surface finishing of the wafers.

[0003] Semiconductor material single-crystal wafers (especially silicon semiconductor wafers) are typically produced by first pulling single-crystal ingots according to the so-called Chuklaski method or Czochralski (CZ) process. The resulting ingots are then diced into wafers using a saw suitable for this purpose (e.g., a wire saw, inside-diameter saw, or band saw), and these wafers are typically processed into semiconductor wafers using a wire saw or inside-diameter saw. Following further mechanical, chemical-mechanical, and / or chemical steps, an epitaxial layer may optionally be deposited on the wafer using chemical vapor deposition (CVD).

[0004] The CZ process typically requires lowering a seed crystal (often referred to as a seed crystal) toward a silicon melt at a temperature above 1410°C until contact is made with the melt. Crystal growth is achieved by slowly pulling the seed crystal upwards again while rotating, simultaneously solidifying the melt at the growth interface. In this process, a starting cone of a single crystal with an increasing diameter is first pulled up, followed by a cylindrical portion with the desired diameter, and finally a tail cone with a decreasing diameter.

[0005] When a seed crystal comes into contact with hot silicon melt, stress is generated in the seed crystal, leading to the development of dislocations within it. These dislocations may propagate (or extend) along the entire length of the crystal ingot.

[0006] In most cases, these dislocations can be eliminated from the single crystal by forming a neck. To form a neck, also known as a Dash neck, a spinning seed crystal is pulled upwards at high speed to reduce the diameter of the essentially cylindrical neck to a typical 2 to 6 mm. Starting from the dislocation-free neck, the diameter of the single crystal ingot increases again, and a dislocation-free single crystal with the desired diameter can be pulled. This method is also known as the Dash technique or Dash necking technique.

[0007] Due to the Das necking technique, edge dislocations can usually be avoided. However, the situation is different for spiral dislocations. These dislocations cannot always be eliminated by the Das necking technique.

[0008] For example, EP 1 498 517 A1 describes how dislocations at the center of a seed crystal with a crystal orientation of ‹110› cannot be easily eliminated by the Das necking technique. In the case of a seed crystal with a ‹110› crystal orientation, i.e., if the pulling direction of the crystal extends perpendicularly to a crystal plane having a Miller index of (110), the immersion of the seed crystal into the melt may be accompanied by the formation of a slip dislocation at the center, which propagates along the entire length of the ingot in the axial direction (i.e., along the pulling direction of the single crystal ingot). In the case of a ‹100› crystal orientation, dislocations appear in some pulled single crystal ingots even when the Das necking technique is employed. EP 1 498 517 A1 describes a method for pulling single crystal ingots in which the formation of dislocations in the seed crystal is prevented. Therefore, dislocation elimination is not required, and thus the method operates without the need for the Das necking technique.

[0009] EP 1 498 516 B1 describes a method for producing silicon single crystals with a crystal orientation of ‹110›, wherein dislocations in the seed crystal (seed crystal) are eliminated by tilting the seed crystal at a predetermined angle relative to the axial pulling direction, whereby the pulling direction is not perfectly perpendicular to the crystal plane having a Miller index of (110). Subsequent formation of a fine neck according to the Das necking technique allows dislocation elimination. However, semiconductor wafers with surfaces extending perfectly parallel to the ‹110› crystal plane cannot be produced from single crystal ingots according to EP 1 498 516 B1, thus limiting the applicability of these semiconductor wafers to electrical components requiring high switching frequencies.

[0010] US 4,002,523 A describes a complex method for pulling a single-crystal ingot with a crystal orientation of ‹110›, wherein a neck with alternating thin and thick regions is pulled, causing dislocations extending parallel to the growth direction of the single crystal (pulling direction) to gradually move away from the crystal axis and toward the periphery. This increases the likelihood that dislocations terminate at the surface and thus disappear from the crystal. However, this complex and cumbersome method requires gradually thinning the neck to 2.5 mm, thus reducing the stability of the neck and making it more difficult to remove the crystal from a large ingot.

[0011] Furthermore, JP 9 255 490 A2 proposes using seed crystals and melts with equal boron concentrations in the context of producing boron-doped silicon single crystals, thereby preventing the development of dislocations caused by different lattice constants.

[0012] JP 2982053 B2 describes a single-crystal pulling method that operates without the need for das necking. In this method, a seed crystal is lowered just above the melt, heated at this point, and then a portion of the seed crystal is immersed in the melt at a rate of 0.05 to 2 mm / min, after which the immersed portion is melted. The vertical length of this immersed portion corresponds at least to the diameter of the seed crystal. The melting operation also melts the portions of the seed crystal containing dislocations caused by thermal stress. Therefore, the portion of the seed crystal remaining after melting no longer contains any defects.

[0013] The methods described above are all relatively complex and inefficient, and therefore have only limited usefulness in industrial production. Furthermore, these methods are insufficient for producing single-crystal ingots without near-center defects and with a crystal orientation of ‹110›. Summary of the Invention

[0014] Technical objective and beneficial effects of the present invention The objective of this invention is to provide an efficient method for pulling single-crystal ingots of semiconductor materials using the Chuklaski technique without near-center defects. Specifically, the objective of this invention is to provide an efficient method suitable for producing dislocation-free single-crystal ingots of semiconductor materials with crystal orientations ‹100›, ‹110›, or ‹111›.

[0015] This invention is based on the following surprising observation: after a seed crystal has been immersed, a single crystal ingot with a crystal orientation ‹110› without near-center defects (especially without near-center spiral dislocations) can be produced by pulling a seed crystal cone with a defined geometry under defined conditions.

[0016] According to a first aspect of the invention, this objective is achieved by a method of the invention for pulling single-crystal ingots from a silicon melt according to the Chuklaski technique. The method comprises the following steps: (i) A silicon melt and a single-crystal silicon seed crystal above the melt are provided, wherein the silicon melt and / or seed crystal may optionally include a dopant; (ii) Lower the seed crystal until contact is established between the seed crystal and the molten silicon; and (iii) By raising the seed crystal, the seed crystal cone (or seed cone), the neck, the starting cone and the cylindrical part of the single crystal ingot are pulled up in sequence; The seed crystal cone has an upper diameter (or upper diameter) of not less than 12 mm and not more than 22 mm, a lower diameter (or lower diameter) of not less than 2 mm and not more than 10 mm, and an axial length of not less than 75 mm and not more than 300 mm.

[0017] Preferred embodiments of the method of the invention according to the first aspect are the subject of dependent claims 2 to 10.

[0018] It has been surprisingly determined that, using the method of the present invention, single crystal ingots with a crystal orientation of ‹110› without near-center defects (especially without near-center screw dislocations) can be produced. Furthermore, the method of the present invention is particularly efficient and therefore economically attractive because the seed crystal can move at a surprisingly high speed before crystal pulling begins (since a holding phase is not required).

[0019] Near-center defects are linear combinations of various dislocations (e.g., multiple spiral dislocations and / or edge dislocations) connected by so-called kinks.

[0020] Near-center defects arise when the seed crystal comes into contact with the melt, and they propagate through the neck. In order to propagate through the neck, such defects must continue through the center of the seed crystal, across the neck, and to the center of the grown single crystal ingot. Therefore, they appear at the very top of the cylindrical portion of the single crystal, and thus in the wafer with the desired nominal diameter obtained from the very top of the cylindrical portion (at or at least near the center). Any possible deviation of the pull direction from the desired crystal orientation causes dislocations to move away from the center as the length of the grown single crystal increases. Therefore, as the distance from the neck increases, dislocations propagating through the single crystal ingot can move further from the center of the grown single crystal ingot. In the context of this invention, a near-center dislocation is a dislocation occurring at a distance of up to 50 mm from the center of the semiconductor wafer (i.e., from the midpoint of the wafer). In the context of this invention, a single crystal ingot is considered dislocation-free if the wafer from the very top of the cylindrical portion (more specifically, from the very top 10 cm of the cylindrical portion of the single crystal ingot) does not contain near-center dislocations.

[0021] According to a second aspect, the present invention relates to (or pertains to) a method for producing semiconductor material wafers. The method includes the following steps: Single crystal ingots are pulled according to the method of the first aspect; Grinding the single crystal ingot; The ground single crystal ingot is cut into wafers; Grinding and / or polishing the wafer; Polishing the wafer; and The wafer is chemically cleaned.

[0022] In a preferred embodiment of the method of the invention according to the second aspect, the method additionally includes the step of depositing a semiconductor material epitaxial layer on a semiconductor material wafer.

[0023] According to a third aspect, the present invention relates to a single-crystal silicon wafer, wherein... The wafer has a nominal diameter of 300 mm, and the crystal orientation (or crystal direction) of the principal face deviates from the orientation ‹110› by no more than 5°. The wafer may optionally include dopants, and There are no dislocations within a radius of 50 mm around the midpoint of the wafer. The following techniques are used to perform dislocation detection on wafers: The mirror-etched or polished wafer is placed on a turntable (or stage) in a dark-field transmission polarizer (or polarizing mirror). Rotate the wafer on the turntable. A linearly polarized infrared beam is used to illuminate a rotating semiconductor wafer on a turntable, while the semiconductor wafer is continuously or discontinuously shifted in the radial direction. Determine the location-related depolarization and transmission, and Points where the beam is depolarized are identified as dislocations. Attached Figure Description

[0024] Figure 1 The measured depolarization distribution of a wafer produced according to the working example is shown, and it is shown that the wafer has no near-center defects.

[0025] Figure 2 The measured depolarization distribution of a wafer produced according to a comparative example is shown, which exhibits near-center defects. Detailed Implementation

[0026] In the method of the invention according to the first aspect, in step (i), a silicon melt and a single-crystal silicon seed crystal are provided above the silicon melt. The terms "seed crystal" and "seed crystal" are used synonymously. The seed crystal is preferably rotationally symmetric and disposed above the silicon melt such that its axis of rotation is vertical (i.e., along the pulling direction). Thus, the axis of rotation is arranged parallel to the pulling direction. More preferably, the seed crystal is cylindrical, and its cylindrical axis is arranged parallel to the pulling direction.

[0027] The seed crystal preferably has a diameter of not less than 5 mm and not more than 25 mm. In a particularly preferred embodiment, the seed crystal has a diameter of not less than 14 mm and not more than 22 mm. The axial length of the seed crystal preferably corresponds to at least three times the diameter, more preferably at least four times the diameter, and most preferably at least five times the diameter. The axial length of the seed crystal extending downward from the support of the seed crystal preferably corresponds to at least three times the diameter of the seed crystal, more preferably at least four times (before immersion, i.e., before contact is formed between the seed crystal and the molten silicon). The support here is an element at the lower end of the pull shaft or cable used to fix the upper portion of the seed crystal to the pull shaft or cable.

[0028] The silicon melt and seed crystal are essentially composed of silicon. In a preferred embodiment, the silicon fraction in the melt and seed crystal is not less than 99.0% (by weight), more preferably not less than 99.8% (by weight), and most preferably not less than 99.9% (by weight). The silicon concentration in the melt and seed crystal can vary independently of each other.

[0029] The silicon melt and / or seed crystal here may optionally contain, for example, p-type and / or n-type dopants. Dopants are, for example, at least one element selected from boron, aluminum, gallium, indium, carbon, germanium, nitrogen, phosphorus, arsenic, antimony, and oxygen. The seed crystal preferably has a density of less than [amount missing] cm³. 3 10 15 The dopant concentration is approximately 10 atoms per cubic centimeter. The silicon melt preferably has a dopant concentration of less than 10 atoms per cubic centimeter. 15 The dopant concentration per atom. The dopant concentration here is the sum of the concentrations of dopants B, Al, Ga, In, C, Ge, Sn, Pb, N, P, As, Sb, Bi, and O.

[0030] The silicon melt here has a temperature of not less than 1410°C, preferably not less than 1420°C. The melting point of the silicon melt is about 1410°C. In one embodiment, the melt is superheated. This means that the temperature of the melt is higher than its melting point.

[0031] In step (ii), the seed crystal is lowered until contact is established between the seed crystal and the molten silicon. In the context of this invention, contact can be at least one of the following: touch between the bottom of the seed crystal and the surface of the melt, thermal contact, and electrical contact. Here, a portion of the seed crystal may also be immersed in the molten silicon. Preferably, the seed crystal is immersed in the molten silicon for no more than 5 mm, more preferably no more than 2 mm. The seed crystal may be fixed to a lifting shaft or cable, by means of which the seed crystal can move vertically downwards along the lifting axis and come into contact with the melt.

[0032] In step (ii), the seed crystal is lowered and brought into contact with the melt at a rate not exceeding 30 mm / min, preferably not less than 5 mm / min and not more than 20 mm / min. Lowering the seed crystal at this rate reduces the rate of thermal stress in the seed crystal as it is lowered, thereby reducing the development and propagation of dislocations.

[0033] The seed crystals are arranged along the pulling direction. The axis of rotation of the seed crystal (which is preferably rotationally symmetric) thus extends in the vertical direction. Therefore, the cylindrical axis of the particularly preferred cylindrical seed crystal extends in the vertical direction. The seed crystal preferably has a ‹100› crystal orientation, a ‹110› crystal orientation, or a ‹111› crystal orientation. In the case of a ‹100› crystal orientation, the crystal plane (or crystal face) extends perpendicular to the pulling direction with a Miller index of (100) (and therefore, in the case of a cylindrical seed crystal, perpendicular to the cylindrical axis of the seed crystal). Correspondingly, in the case of a seed crystal with a ‹110› crystal orientation, the cylindrical axis and the pulling direction extend perpendicular to the crystal plane with a Miller index of (110). The desired crystal orientation in the sense of the present invention also exists when the angle between the plane extending perpendicular to the pulling direction and the plane having the desired crystal orientation is not greater than 5°. This means that the axis perpendicular to the crystal plane having the desired orientation (e.g., ‹100› crystal orientation, ‹110› crystal orientation, or ‹111› crystal orientation) can deviate from the lifting axis (i.e., the lifting direction) by up to 5°. The deviation of the lifting direction from the desired crystal orientation is preferably no more than 4°, more preferably no more than 3°, and most preferably no more than 2°. The deviation of the lifting direction from the ‹100› or ‹110› crystal orientation is preferably no more than 2°. Therefore, in this embodiment, the angle between the lifting direction and the axis perpendicular to the crystal plane having a Miller index of (100) or (110) is no greater than 2°. In another particularly preferred embodiment, the deviation of the lifting direction from the desired crystal orientation is no more than 1.5°.

[0034] After the seed crystal is lowered and comes into contact with the silicon melt, in step (iii) of the method of the invention according to the first aspect, the single crystal ingot can be pulled from the melt by raising the seed crystal. Here, the seed crystal cone, the neck, the starting cone, and the cylindrical portion of the single crystal ingot are pulled in sequence. In the sense of the invention, the seed crystal cone is part of the single crystal growth, which tapers towards the bottom and preferably has the form of a truncated (circular) cone. Here, the upper diameter preferably corresponds to the maximum diameter of the truncated cone portion, and the lower diameter corresponds to the minimum diameter of the truncated cone portion.

[0035] When contact is established between the seed crystal and the molten silicon, and the immersed portion of the seed crystal is melted under appropriate conditions, a uniform boundary layer is formed between the molten silicon and the solid portion of the seed crystal. In step (iii), the pull shaft or cable on which the seed crystal is fixed can then be slowly pulled upwards with rotation, while the molten silicon solidifies at the growth (or growth) interface. By changing the pull speed, rotation speed, and temperature, a specific desired diameter can be obtained for the grown, rotated ingot (or crystal ingot). The diameter of the ingot can be adjusted very precisely by means of suitable rules.

[0036] First, in step (iii), the seed crystal cone is pulled up; the seed crystal cone has an upper diameter of not less than 12 mm and not more than 22 mm, a lower diameter of not less than 2 mm and not more than 10 mm, and an axial length of not less than 75 mm and not more than 300 mm. The seed crystal cone thus tapers downward. The seed crystal cone preferably has a length of not less than 150 mm and not more than 225 mm, an upper diameter of not less than 14 mm and not more than 20 mm, and a lower diameter of not less than 4 mm and not more than 6 mm. The upper diameter of the seed crystal cone here preferably corresponds to the diameter of the seed crystal before contact is formed between the seed crystal and the silicon melt. The angle between the surface line of the seed crystal cone and the axis of rotation of the seed crystal cone is preferably not less than 2° and not more than 15°, more preferably not less than 4° and not more than 13°. Pulling up a seed crystal cone with the above geometry allows the elimination of near-center dislocations, especially screw dislocations, from a single crystal ingot. The seed crystal cone preferably has the shape of a truncated cone. Therefore, the diameter of the seed crystal cone preferably decreases uniformly with increasing length during ingot pulling until the lower diameter of the seed crystal cone is reached and the neck is pulled. During pulling, the crucible preferably rotates at 5 to 20 revolutions per minute, more preferably at 10 to 15 revolutions per minute.

[0037] After pulling the seed crystal cone, a neck is pulled; the diameter of the neck does not exceed the lower diameter of the seed crystal cone. The neck preferably has a diameter of not less than 2 mm and not more than 10 mm, more preferably not less than 4 mm and not more than 6 mm. In a preferred embodiment, the diameter of the neck corresponds to the lower diameter of the seed crystal cone. The neck preferably has a length of not less than 100 mm, more preferably not less than 200 mm and not more than 600 mm.

[0038] After necking, the starting cone is pulled. To this end, the diameter of the grown single crystal ingot is continuously increased until the desired diameter of the single crystal ingot is reached.

[0039] The desired diameter of a single crystal ingot depends on the desired nominal diameter of the wafer to be produced from the single crystal ingot. Due to the diameter reduction during wafer manufacturing, as a result of machining steps such as grinding, the desired diameter of the single crystal ingot is larger than the nominal diameter. Preferably, the desired diameter of the single crystal ingot is not less than 1 mm and not more than 20 mm above (or greater than) the nominal diameter of the semiconductor material wafer to be produced, and more preferably not less than 2 mm and not more than 10 mm.

[0040] The portion of a single crystal ingot having the desired diameter is also referred to as the cylindrical portion of the single crystal ingot. In this portion, the diameter falls within the range of desired diameters at each location along the axial direction (i.e., along the pulling direction). Therefore, the cylindrical portion of a single crystal is the portion where the diameter at each location along the axial direction is greater than the nominal diameter of the semiconductor material wafer to be produced. At each location along the axial direction, the diameter in the cylindrical portion is preferably not less than 1 mm and not more than 20 mm above the nominal diameter of the semiconductor material wafer to be produced, more preferably not less than 2 mm and not more than 10 mm. Therefore, the cylindrical portion of a single crystal ingot is the portion of the single crystal ingot suitable for producing wafers with a desired nominal diameter.

[0041] The diameter of the cylindrical portion can vary along its axial length, provided that the diameter at each point is at least 1 mm above the desired nominal diameter of the semiconductor wafer to be produced. In a particularly preferred embodiment, the diameter of the cylindrical portion varies by less than 20 mm, preferably less than 10 mm, along its axial length, corresponding to the difference between the maximum and minimum diameters along the axial length.

[0042] The nominal diameter is preferably not less than 150 mm and not more than 300 mm, more preferably not less than 200 mm and not more than 300 mm, and most preferably 300 mm. The wafer has a desired nominal diameter when the actual diameter deviates from the value listed in the nominal diameter by no more than 5 mm, preferably no more than 3 mm, and the nominal diameter, more preferably no more than 1 mm.

[0043] After pulling the cylindrical portion of a single crystal ingot with the desired diameter, the tail cone can be pulled and the ingot removed.

[0044] In step (iii), the pulling speed is preferably in the range of 0.5 to 10 mm / min (more preferably 2 to 6 mm / min). The rotational speed of the crystal is preferably in the range of 1 to 20 rpm (more preferably 5 to 15 rpm).

[0045] During step (iii), the temperature of the silicon melt is preferably not lower than 1410°C, more preferably not lower than 1420°C, and most preferably not lower than 1425°C. The temperature at the start of the lifting process is preferably not higher than 1450°C, more preferably not higher than 1440°C, and most preferably not higher than 1435°C. In a particularly preferred embodiment, the temperature at the start of the lifting process is not lower than 1420°C and not higher than 1440°C.

[0046] The temperature of the melt can be measured using a measuring unit with an accuracy of 0.1°C. The measurement here is preferably of the temperature at the surface of the silicon melt. The measuring unit used to measure the temperature at the surface of the silicon melt is preferably a pyrometer. The temperature of the silicon melt at a defined measuring point can be measured with an accuracy of 0.1°C and adjusted with an accuracy of 1.0°C. The temperature of the melt can be adjusted by one or more heating elements, preferably one or more resistance heaters, arranged around the crucible. The heating elements are controlled via a programmable control unit connected to the measuring unit and adjusts the temperature at the surface of the silicon melt at the measuring point as a function of the measured temperature with an accuracy of 1.0°C.

[0047] According to a second aspect of the invention, a semiconductor material wafer can be produced by performing the following steps after the method for pulling a single crystal ingot according to the first aspect: grinding the single crystal ingot; dicing the ground single crystal ingot into wafers; grinding and / or polishing the wafers; polishing the wafers; and chemically cleaning the wafers. Thus, after pulling is completed, the ingot can be diced and ground, diced into wafers, and then ground or polished, polished, and chemically cleaned in this sequence to obtain a semiconductor material wafer. A semiconductor material epitaxial layer can optionally be deposited on the wafer thus obtained. Deposition is preferably performed in the vapor phase by means of CVD. The semiconductor material epitaxial layer is preferably composed of optionally doped silicon, germanium, silicon carbide, or gallium nitride.

[0048] Dislocation checking can then be performed on the polished wafer using the stress-induced birefringence effect. This effect is based on the generation of a stress field caused by dislocations and defects. If a material is exposed to shear stress, its refractive index becomes anisotropic, and the material becomes birefringent. Therefore, incident linearly polarized light is depolarized by the sample during transmission or reflection. Thus, the obtained depolarized distribution can be interpreted as a shear stress distribution.

[0049] Polished wafers can be measured, for example, by means of SIRD (Scanning Infrared Depolarization). SIRD measurements are performed using a dark-field transmission polarimeter. In this measurement, a linearly polarized beam (preferably a laser beam) with a fixed position is irradiated onto the semiconductor wafer being inspected. For measurement purposes, the wafer is placed on a turntable, rotated, and continuously displaced in the radial direction. If the wafer's crystal structure has no dislocations and stress, the polarization of the beam remains unchanged. However, if dislocations are present, the beam undergoes local depolarization caused by stress-induced birefringence.

[0050] Near-center defects / dislocations in semiconductor wafers produced by the method of the present invention can be inspected using instruments such as PSI-type instruments from manufacturer Semilab or instruments from manufacturer PVA Metrology & Plasma Solutions. The lateral resolution of such measurements is typically in the range of 50 μm to 100 μm. Therefore, the determination of position-dependent (or site-dependent) depolarization and transmission in the wafer according to the third aspect of the invention is performed with a lateral resolution of at least 100 μm, preferably 50 μm.

[0051] With the aid of this measurement, dislocation checks can be performed on semiconductor material wafers (more particularly, single-crystal silicon wafers with a diameter of 300 mm and a ‹110› crystal orientation) produced according to the second aspect of the invention to determine whether the wafer is free of near-center defects.

[0052] The single-crystal silicon wafer according to a third aspect of the invention has a nominal diameter of 300 mm, and the crystal orientation of its main facet deviates from orientation ‹110› by no more than 5°. The wafer may optionally include dopant, preferably at a concentration of less than 10 ppm. 15 Each atom. The dopant concentration corresponds to the sum of the concentrations of dopants B, Al, Ga, In, C, Ge, Sn, Pb, N, P, As, Sb, Bi, and O.

[0053] According to the third aspect, the wafer has no dislocations within a radius of 50 mm around its midpoint, and therefore no near-center defects. Dislocation inspection of the wafer is performed using the following technique: The mirror-etched or polished wafer is placed on a turntable in a dark-field transmission polarizer. Rotate the wafer on the turntable. A linearly polarized infrared beam (preferably an infrared laser beam) is used to irradiate a rotating semiconductor wafer on a turntable, while the semiconductor wafer shifts continuously or discontinuously in the radial direction. Determine the location-related depolarization and transmission, and Points where the beam is depolarized are identified as dislocations. The lateral resolution of location-dependent depolarization and transmission is preferably in the range of 50 μm to 100 μm.

[0054] The monocrystalline silicon wafer according to the third aspect of the present invention can be produced by the method of the present invention according to the second aspect. A ‹110› crystal orientation exists in the sense of the third aspect of the present invention when the orientation of the main face of the wafer deviates from the ‹110› crystal orientation by no more than 5°, preferably no more than 2°, and more preferably no more than 1.5°. Therefore, the orientation of the main faces of the front (or front side) and back (or rear side) of the wafer according to the third aspect of the present invention can deviate from the orientation of the crystal plane having the Miller index (110) by a maximum of 5°, preferably no more than 3°.

[0055] Detailed description of inventive working examples and comparative examples In each working example, a cylindrical monocrystalline silicon seed crystal with a ‹110› crystal orientation and a diameter of 19 mm was used to extract an ingot with a ‹110› orientation using the Chuklaski technique. The seed crystal was secured to the pulling shaft of an EKZ 3000 crystal pulling system from manufacturer PVA TePla. A quartz crucible filled with solid high-purity polycrystalline silicon from manufacturer Wacker Chemie was arranged between resistance heaters in the crystal pulling system. After the crystal pulling system had been shut down, purged, and evacuated, the quartz crucible containing silicon was heated to provide a silicon melt. The crystal pulling system used additionally included a camera system for observing phase boundaries and a pyrometer for determining the temperature of the melt. Thus, the camera system could be used to determine the precise position and shape of the seed crystal relative to the melt, as well as the shape of the grown ingot (particularly the seed crystal cone).

[0056] The seed crystal is then moved downwards at a speed of 10 mm / min until a portion of it is immersed in the melt. It is then slowly moved upwards with rotation, and in this way, first, a seed crystal cone with a length of 175 mm to 200 mm, an upper diameter of 19 mm, and a lower diameter of 5 mm is drawn up by correspondingly increasing the drawing and rotation speeds. Subsequently, a thin neck with a diameter of 5 mm and a length of 300 mm is drawn up. The starting cone is then drawn up, and its diameter is increased to 310 mm.

[0057] When a crystal diameter greater than 310 mm is achieved, a cylindrical portion with a roughly constant diameter is pulled using the Chuklaski technique. The pulling speed at this point is approximately 4 mm / min, and the crystal rotation speed is 10 rpm. Finally, a tail cone with a decreasing diameter is pulled. At the end of the ingot pulling process, the completed ingot is removed and diced into wafers with a thickness of 900 μm. Wafers with a diameter of 200 mm are then produced, ground to a thickness of 750 μm, and mirror-etched.

[0058] These wafers were then subjected to stress testing. For this stress testing, each wafer was placed on a turntable of a PSI-type instrument from the manufacturer Semilab. This instrument features a fixed-position linearly polarized infrared light source and a turntable. During measurement, the wafer, rotating on the turntable, continuously shifts radially, allowing the entire surface of the wafer to be scanned by the light source. This revealed no stress on the wafer, and more specifically, no stress within a 50 mm radius around the midpoint of the wafer. Figure 1 The depolarization distribution determined by measurements of wafers produced in the working example is shown, and it is shown (or indicated) that these wafers have no near-center defects. Therefore, in the working example, a single-crystal silicon wafer with a ‹110› crystal orientation and no near-center defects is produced.

[0059] The only difference between the comparative example and the working example is that the seed cone having the geometry according to the method of the invention is not pulled. As in the working example, the ingot produced in the comparative example is also removed and diced into wafers with a thickness of 900 μm. Wafers with a diameter of 200 mm are then produced, ground to a thickness of 750 μm, and mirror-etched. These wafers are then subjected to stress checks in the same manner as the wafers produced in the working example. Figure 2 The depolarization distribution determined for a single-crystal silicon wafer produced in the comparative example is shown. Figure 2 In the wafer, stress is significant at the center. Therefore, stress around the wafer midpoint and consequently near-center defects were detected.

Claims

1. A method for pulling a single crystal ingot from a silicon melt according to the Chuklaski technique, comprising the following steps: (i) A silicon melt and a single-crystal silicon seed crystal above the melt are provided, wherein the silicon melt and / or seed crystal may optionally include a dopant; (ii) Lower the seed crystal until contact is made between the seed crystal and the silicon melt; as well as (iii) By raising the seed crystal, the seed crystal cone, the neck, the starting cone and the cylindrical portion of the single crystal ingot are sequentially lifted; The seed crystal cone has an upper diameter of not less than 12 mm and not more than 22 mm, a lower diameter of not less than 2 mm and not more than 10 mm, and an axial length of not less than 75 mm and not more than 300 mm. Furthermore, the angle between the surface line of the seed crystal cone and the rotation axis of the seed crystal cone is not less than 2° and not greater than 15°.

2. The method for pulling single crystal ingots as described in claim 1, wherein, The seed crystal cone has a length of not less than 150 mm and not more than 225 mm, an upper diameter of not less than 14 mm and not more than 20 mm, and a lower diameter of not less than 4 mm and not more than 6 mm.

3. The method for pulling single crystal ingots as described in claim 1 or 2, wherein, The upper diameter of the seed crystal cone corresponds to the diameter of the seed crystal before it comes into contact with the silicon melt.

4. The method for pulling single crystal ingots as described in any one of claims 1 to 3, wherein, The axis of rotation of the seed crystal cone extends parallel to the pulling direction.

5. The method for pulling single crystal ingots as described in any one of claims 1 to 4, wherein, The seed crystal has a crystal orientation of ‹100›, ‹110›, or ‹111›.

6. The method for pulling single crystal ingots of semiconductor materials as described in any one of claims 1 to 5, wherein, The lifting direction deviates from the orientation by no more than 3°.

7. The method for pulling single crystal ingots of semiconductor materials as described in any one of claims 1 to 6, wherein, The lifting direction deviates from the orientation ‹110› by no more than 2°.

8. The method for pulling a single crystal ingot as described in any one of claims 1 to 7, wherein, The seed crystal has a density of less than 10 per cubic centimeter. 15 The dopant concentration per atom.

9. The method for pulling single crystal ingots as described in claim 8, wherein, The dopant concentration is the sum of the concentrations of dopants B, Al, Ga, In, C, Ge, Sn, Pb, N, P, As, Sb, Bi, and O.

10. A method for producing semiconductor material wafers, comprising the following steps: Single crystal ingots are pulled using the method described in any one of claims 1 to 9. Grinding the single crystal ingot; The ground single crystal ingot is cut into wafers; Grinding and / or polishing the wafer; Polish the wafer; as well as The wafer is chemically cleaned.

11. The method for producing semiconductor material wafers as described in claim 10, wherein, The method additionally includes the following steps: An epitaxial layer of semiconductor material is deposited on the semiconductor material wafer.

12. A single-crystal silicon wafer, wherein, The wafer has a nominal diameter of 300 mm, and the crystal orientation of the main face deviates from the orientation ‹110› by no more than 5°. The wafer may optionally include dopants, and There are no dislocations within a radius of 50 mm around the midpoint of the wafer. The following techniques are used to perform dislocation detection on wafers: The mirror-etched or polished wafer is placed on a turntable in a dark-field transmission polarizer. Rotate the wafer on the turntable. A linearly polarized infrared beam is used to illuminate a rotating semiconductor wafer on a turntable, while the semiconductor wafer is continuously or discontinuously shifted in the radial direction. Determine the location-related depolarization and transmission, and Points where the beam is depolarized are identified as dislocations.

Images