Silicon single crystal manufacturing process

Abstract

A process for producing a single crystal of silicon includes placing a feed rod of silicon having a diameter of 230 mm to 270 mm in a float zone apparatus, placing a first hollow cylinder having a bottom edge and an inner diameter 30 mm to 50 mm greater than the diameter of the feed rod, placing a second hollow cylinder having a top edge and an inner diameter 20 mm to 60 mm greater than a target diameter of the single crystal, and pulling a cylindrical portion of the single crystal having a target diameter of 290 mm to 310 mm inclusive, wherein the feed rod on the melting side forms an outer melting edge and the single crystal ingot on the growing side forms a crystallized edge, the pulling speed is 1.3 mm / min to 1.5 mm / min, preferably 1.35 mm / min to 1.45 mm / min, the vertical distance of the bottom edge of the first hollow cylinder from the outer melting edge is less than 2 mm, and the top edge of the second cylinder protrudes 1 mm to 10 mm beyond the crystallized edge.

Description

[Technical Field]

[0001] The subject of the present invention is a process for producing monocrystalline ingots of silicon in a float zone plant, and the ingot pieces produced therefrom. [Background technology]

[0002] Float zone pulling of single crystals is known from the prior art (J. Bohm, et al., "Handbook of Crystal Growth," ed.: DTJ Hurle, vol. 2, Part A, pp. 213-257, 1994) and is used on an industrial scale for the production of single-crystal materials. In this process, an induction coil containing a high-frequency current is used to melt the starting material in a zone. This material then solidifies in the form of a single crystal as the material is pulled vertically, and the resulting single crystal is usually rotating. Depending on the specific embodiment of the process, the single crystal can be pulled upward or downward. The electromagnetic field of the induction coil generates a flow in the melting zone with a double vortex structure. This flow is continuously directed inward in the center of the zone, while near the two ends of the melting zone, the flow is always directed radially outward. The resulting flow in the melting zone is generated not only by electromagnetic forces, but also by buoyancy and Marangoni forces, as well as by the rotation of the ingot or crystal. The geometry of the solidification phase boundary is established according to the temperature distribution prevailing within the ingot, which distribution is influenced by the flow conditions.

[0003] Flow control during float-zone pulling, and the associated improvement of crystal quality and operational stability, have been the subject of research that proposes optimizing the operating parameters of the induction coil geometry, induction coil current, ingot or crystal rotation, and pulling speed (A. Muhlbauer et al., Journal of Crystal Growth, vol. 151, pp. 66-79, 1995; S. Otani et al., Journal of Crystal Growth, vol. 66, pp. 419-425, 1984; S. Y. Zhang et al., Journal of Crystal Growth, vol. 243, pp. 410-418, 2002). Therefore, attempts have been made to homogenize the dopant distribution by varying the crystal rotation, by displacing the induction coil relative to the crystal axis, or by optimizing the induction coil geometry.

[0004] On an industrial scale, float zone pulling is used in particular for the production of single crystals of silicon, where the single crystals are obtained from polycrystalline feed rods, the use of single crystal silicon feed rods being another possible option.

[0005] For this process, a feed rod undergoes initial melting at one end using a high-frequency coil (inductor), and a single-crystal seed crystal is attached to the resulting molten droplet. The gradually melted material from the feed rod serves as a continuous source of single crystals that subsequently grow on the seed crystal. A length, called the neck, is first crystallized to bypass dislocations from the crystal lattice. The diameter of the growing single crystal is then expanded to the target diameter in a length called the initial cone (seed cone). A length of single crystal with the target diameter is then produced. At the end of the process, a length called the end cone is also produced. The process may optionally be terminated without an end cone; however, in that case, the end portion of the length with the target diameter would exhibit dislocations and therefore be unusable for the intended further processing.

[0006] Here, the feed rod is mounted at one end on a rotatable shaft (the pulling shaft) so that it does not slip even if the direction of rotation changes suddenly. Furthermore, it is required that the center of the other end of the feed rod is located on the rotation axis of the pulling shaft at all times during crystal pulling. If the center of the other end moves away from the rotation axis of the pulling shaft, it can cause significant melting by the pulling coil, which can have a negative impact on the entire pulling process.

[0007] Currently, the only commercially available monocrystalline silicon ingots are those produced by the float zone method and have a nominal diameter of up to 200 mm.

[0008] Achieving a target diameter greater than 200 mm combined with maximum pull speeds has long been an unmet need in the industry, as it promises greater returns, particularly in the production of components on semiconductor wafers produced from the process.

[0009] Prior art DE 101 37 856 A1 discloses a crucible-less float zone pulling process for producing single silicon crystals having a diameter of at least 200 mm over a length of at least 200 mm, which region of this length is dislocation-free, and in which a fusion neck forms during float zone pulling between the feed rod and the single crystal.

[0010] Nevertheless, it becomes apparent that the pull rate is too low for economically reasonable pulling of the crystal.

[0011] EP 2142686 discloses a process for producing single crystals by guiding a polycrystalline rod through a heated region to create a molten zone, applying a magnetic field to the molten zone, and inducing single crystal growth during solidification of the molten material on the single crystal seed. The growing single crystal is rotated alternately clockwise and counterclockwise. This process is useful for producing single silicon crystals with uniform electrical properties. An apparatus for carrying out the process is also disclosed. While the patent claims crystals larger than 200 mm, no specific process is provided for 300 mm diameter crystals.

[0012] U.S. Patent No. 2016,053,401 discloses an auxiliary heating device for a zone melting furnace and a heat conservation method for a single crystal rod. The auxiliary heating device includes an auxiliary heater disposed below a high-frequency heating coil in a zone melting furnace. The heater is formed by winding a hollow metal circular pipe. The auxiliary heating device's starting end is located at the top and its ending end is located at the bottom, with the upper and lower ends extending from both ends. A hollow cylindrical heating load is disposed inside the auxiliary heating device, and an insulating section is disposed between the heating load and the auxiliary heating device. This invention can solve the problem of uneven distribution of heat fields and cracks in the single crystal rod caused by excessive thermal stress during the growth process of zone melting silicon single crystals exceeding 6.5 inches.

[0013] DE 3 805 118 A1 discloses an induction heating coil suitable for use in a crucible-less pulling process. A method is likewise shown that allows optional adaptation of the coil. Summary of the Invention [Problem to be solved by the invention]

[0014] The object of the invention is to provide a process that allows for the production of single crystal ingots by Czochralski pulling, with minimal axial variation of the dopants.

[0015] Likewise, the objective is to provide corresponding crystal pieces. [Means for solving the problem]

[0016] This object is achieved by the processes and products set forth in the claims. [Brief explanation of the drawings]

[0017] [Figure 1] 1 shows an axial section of a float zone plant during crystal pulling required for the process of the present invention. [Figure 2] The contour of the growth strip (201) determined as a function of the radius of the crystal and the length D in the direction of crystal growth is shown. DETAILED DESCRIPTION OF THE INVENTION

[0018] DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS IN ACCORDANCE WITH THE INVENTION The subject of the present invention is a process for producing single crystals of silicon by the float zone method.

[0019] The inventors have recognized that to achieve the goal of maximum diameter combined with high growth rate, it is necessary to install a feed rod in the pull plant with a diameter of at least 230 mm and up to 270 mm.

[0020] Furthermore, it is clearly necessary to have a first hollow cylinder installed in advance, which surrounds the feed rod while the crystal is being pulled. The inner diameter of the first hollow cylinder must be at least 30 mm and no more than 50 mm greater than the diameter of the feed rod. Essentially, the longitudinal axes of the first hollow cylinder and the feed rod are positioned above and below each other. While smaller radial deviations of less than 3 mm may be unavoidable, it is advantageous to minimize these deviations during installation.

[0021] Furthermore, before pulling the crystal, a second hollow cylinder must be installed to surround the single crystal ingot to be pulled later. The inner diameter of this second hollow cylinder must be 20 mm to 60 mm larger than the target diameter of the single crystal ingot. For example, if a crystal with a target diameter of 300 mm is being pulled, the inner diameter should preferably be 320 mm to 360 mm.

[0022] The cylindrical portion of the single crystal preferably has a diameter of between 290 mm and 310 mm and a length not less than 15 cm. The maximum length of the cylindrical portion of the ingot depends substantially on the dimensions of the crystal pulling plant.

[0023] For operational reasons, the diameter of the single crystal is subject to slight variations which can be minimized but cannot be completely eliminated, and the concept of target diameter is therefore understood to refer to the average diameter of the single crystal.

[0024] As shown in Figure 1, the feed rod forms an outer melting edge on the melting side, and the single crystal ingot forms a crystallization edge on the growth side. p The system includes a feed rod (102) having an inner diameter dc1, a first hollow cylinder (105) having an inner diameter dc1, a second hollow cylinder (109) having an inner diameter dc2, a single crystal ingot (101), a melt (106), and a coil (104).

[0025] The parameter h2 here defines the vertical distance of the top edge of the second hollow cylinder (109) from the crystallization edge (103). The parameter h1 here defines the vertical distance of the bottom edge of the first hollow cylinder (105) from the outer melt edge (110) of the feed rod. The vertical distance between the outer melt edge (110) and the crystallization edge (103) is h ak During crystal pulling, the feed rod (102) melts at the melt surface (107). The point in the diagram where the feed rod, melt, and gas space coincide is called the internal triple point (108).

[0026] The inventors have recognized that during crystal pulling, the vertical distance of the bottom edge of the first hollow cylinder from the outer melt edge of the feed rod is preferably less than 2 mm, where the bottom edge of the first hollow cylinder is positioned above the outer melt edge of the feed rod, and thus the first hollow cylinder is offset upward relative to the melt edge.

[0027] The length of the first hollow cylinder is more preferably at least 10 cm and less than 50 cm. The material from which the first hollow cylinder is made preferably consists of silver, and very preferably a material with high emissivity is required as a coating on the inner surface of the first hollow cylinder.

[0028] The emissivity of an object indicates how much radiation it emits compared to a black body, which is an ideal radiant heat emitter.

[0029] Here, gold, silver, silver alloys, carbon or copper etc. are good candidates for carrying out the coating, the use of gold, silver or silver alloys being preferred, as in these cases there is no risk of contamination of the melt or the single crystal.

[0030] More preferably, the first hollow cylinder may be made up of two hollow cylinders, in which case the lower hollow cylinder may preferably be equipped with active heating, preferably in the form of a device similar to that described in US 2016 053 401.

[0031] The inventors particularly focused on ensuring that the upper edge of the second cylinder protrudes above the crystallization edge, and it is particularly preferred that the vertical distance between the crystallization edge and the upper edge of the second hollow cylinder is between 1 mm and 10 mm.

[0032] The material from which the second hollow cylinder is made should be selected in relation to the first hollow cylinder so that the emissivity inside the hollow cylinder is as high as possible. Preferably, the second hollow cylinder is further comprised of two hollow cylinders made from different materials.

[0033] The lower part of the second hollow cylinder, which is the part further from the coil, is preferably made of silver and very preferably includes a surface treatment such as an inner coating of silver or gold or their alloys to maximize emissivity.

[0034] The upper part of the second hollow cylinder, i.e., the hollow cylinder facing the coil, is preferably made of a material that has high emissivity on the inside and is at the same time robust to high temperatures (i.e., above 1000°C). Recommended candidates for this part include ceramic materials, or platinum or platinum-coated ceramics.

[0035] The second hollow cylinder also preferably has passages and holes that allow the image processing system to freely view the crystallized edge of the single crystal. These passages and holes should be as small as possible and as large as necessary, as they may adversely affect the pulling operation.

[0036] The pull rate is preferably between 1.3 mm / min and 1.5 mm / min, and more preferably between 1.35 mm / min and 1.45 mm / min. The pull rate is understood to be the rate at which the single crystal ingot grows axially. For a given pull rate, the rate at which the feed rod must be fed can be easily calculated by the corresponding mass balance.

[0037] The length of the second hollow cylinder is preferably more than 10 cm and not more than 40 cm. The wall thickness of the two hollow cylinders is preferably 10 mm or less and 3 mm or more.

[0038] As is customary in the prior art of the float zone process, the gas space contains nitrogen, which enters the crystal being pulled.

[0039] Crystals pulled by the process just described may be further processed like conventional crystals from the Czochralski pulling process.

[0040] Further processing preferably includes circular grinding of the single crystal, removing lengths of the ingot to form ingot pieces, sawing the ingot pieces into wafers, and grinding and polishing the single crystal into wafers.

[0041] Semiconductor wafers produced from ingots produced according to the process just described are highly suitable for use in the manufacture of power components, with very low defect counts. The reliability of this process lies in the fact that interstitial oxygen is virtually absent from the crystal lattice to form oxygen precipitates. The nominal diameter of 300 mm and the high pulling speed make this process very economical and thus unattainable until now.

[0042] After applying the process of the present invention, a single crystal ingot having a nominal diameter of 290 mm to 330 mm is obtained, which is then cut into ingot pieces having a length preferably of 15 cm to 50 cm.

[0043] For example, if the ingot piece thus obtained, 300 mm in diameter and 20 cm in length, is cut along its length (i.e., in the axial direction), a so-called plank having a width of 300 mm and a length of 20 cm can be obtained.

[0044] Measurements can be made on the plank that characterize both the crystal and the pulling method used to produce the crystal.

[0045] Dopants, typically added in gas form to the melt during the float-zone process, preferably contain boron or phosphorus, and are randomly incorporated into the crystal, resulting in locally non-uniform resistivity distribution in the silicon known as "striation."

[0046] Although great efforts are made to avoid striation so that it is not adversely affected during the operation of the component, striation is nevertheless always measurable as soon as the pulled ingot is doped in the float-zone process.

[0047] Since the dopant is incorporated into the crystal from the melt along the melt / crystal interface, the original morphology of the interface between the crystal and the melt can be confirmed in the morphology of the growth strip by analyzing the measured resistance distribution. By way of example, two references dedicated to this measurement and evaluation process can be referred to:

[0048] "Investigation of defects and striations in as-grown Si crystals by SEM using Schottky diodes" Appl.Phys.Lett.27, 313 (1975); https: / / doi.org / 10.1063 / 1.88482, AJRde Kock, SD Ferris, LCKimerling, and HJLeamy and Ludge, A., Riemann, H "Doping inhomogeneities in silicon crystals detected by the lateral photovoltage scanning (LPS) Method" Inst. Phys. Conf. Ser. 160, 145-148 (1997).

[0049] The latter paper (Ludge et al.) describes a method of "lateral photovoltage scanning" (LPS) that is also suitable for reconstructing the interface between the crystal and the melt, i.e. the growth strip, when the resistance induced by the doping is high and therefore the dopant concentration is low.

[0050] When the method of "lateral photovoltage scanning" (LPS) is applied to the above mentioned Planck, it is possible to see the contours of the growth strip, which accurately reproduce the deviation of the interface between the melt and the crystal.

[0051] The profile of the growth strip of a crystal (201) pulled by the process of the present invention is shown in Figure 2. The value d0 in Figure 2 indicates the maximum deviation of the growth strip. A characteristic parameter of the process of the present invention is the angle of incidence β, determined at a radial position of 85 mm between the horizontal and the tangent applied to the growth strip.

[0052] It is possible to derive two characteristic variables that describe the properties of the crystal flakes. (1) the maximum deviation d0 of the growth strip, and (2) The angle of incidence β between the horizontal and the tangent applied to the growth strip, determined at a radial position of 85 mm.

[0053] A silicon ingot piece containing dopant and having a diameter and an axial length of 15 cm to 50 cm is preferred, the diameter being 290 mm to 330 mm, the ingot piece including the radial extent of a growth strip resulting from the dopant, the maximum deviation of the growth strip being 55 mm to 45 mm.

[0054] Furthermore, it is particularly preferred if the angle of incidence β is greater than or equal to 14° and less than or equal to 16°, the angle of incidence β being located at a radial position of 80 mm between the horizontal and the tangent applied to the growth strip.

[0055] Ingot pieces are 5x1015 at / cm 3 (ASTM Standard F121-83) and an interstitial oxygen content of 1×10 15 at / cm 3 Over 7.5 x 10 15 at / cm 3 It is particularly preferred that the interstitial nitrogen content is: [Explanation of symbols]

[0056] 101 Diameter D c Single crystal ingot having 102 Diameter D p Silicon feed rod having 103 Crystallization edge of single crystal 104 Coil 105 a first hollow cylinder having an inner diameter dc1 106 Melt 107 Melting surface of supply rod 108 Internal triple point 109 A second hollow cylinder having an inner diameter dc2 110 outer molten edge of feed rod h1 Vertical distance of the bottom edge of the first hollow cylinder from the outer molten edge h2 vertical distance of the second cylinder from the crystallization edge h ak The vertical distance between the molten edge and the crystallized edge 201 The contour of the growth strip determined as a function of the radius of the crystal and the length D in the direction of crystal growth. 202 Cylindrical surface of crystal β The angle of incidence determined at a radial position of 85 mm between the horizontal and the tangent applied to the growth strip d0 Maximum deviation of growth strip

Claims

1. 1. A process for producing single crystals of silicon, said process comprising: placing a silicon feed rod having a diameter of at least 230 mm and not more than 270 mm in a float zone device; providing a first hollow cylinder having a bottom edge and an inner diameter that is at least 30 mm and no more than 50 mm greater than the diameter of the feed rod; providing a second hollow cylinder having an upper edge and an inner diameter that is at least 20 mm and no more than 60 mm greater than the target diameter of the single crystal; pulling a cylindrical portion of the single crystal having the target diameter of 290 mm or more and 310 mm or less; the feed rod on the melting side forms an outer melting edge, and the single crystal ingot on the growth side forms a crystallization edge; The pulling speed is 1.3 mm / min or more and 1.5 mm / min or less, the vertical distance of the bottom edge of the first hollow cylinder from the outer fused edge is less than 2 mm; The upper edge of the second hollow cylinder protrudes from the crystallized edge by 1 mm to 10 mm, removing a length of said single crystal to form an ingot piece having a length of at least 15 cm and not more than 50 cm.

2. 2. The process of claim 1, wherein the vertical distance between the molten edge and the crystallized edge is greater than or equal to 35 mm and less than or equal to 40 mm.

3. The process of claim 1 , wherein the feed rod is manufactured by the CZ process.

4. The process comprises: grinding the single crystal into a circular shape; sawing the ingot pieces into wafers; 10. The process of claim 1 further comprising grinding and polishing the wafer.

5. having a diameter The axial length is between 15cm and 50cm. A piece of silicon ingot containing a dopant, The ingot piece has a diameter of 290 mm to 330 mm, the single crystal includes a radial extent of a growth strip that is dominated by the dopant, and the maximum deviation of the growth strip is 55 mm to 45 mm.

6. 6. The ingot piece of claim 5, wherein there is an angle of incidence β between the horizontal and a tangent to the growth strip that is greater than or equal to 14° and less than or equal to 16° at a radial position of 80 mm.

7. The single crystal is 5×10 １５ at / cm ３ 6. The ingot piece of claim 5, having an interstitial oxygen content of less than or equal to ASTM standard F121-83.

8. The single crystal is 1×10 １５ at / cm ３ 7.5 x 10 １５ at / cm ３ 6. The ingot piece of claim 5 having an interstitial nitrogen content of:

Images