SURFACE TEMPERATURE CALCULATION METHOD AND CONTROL METHOD FOR POLYCRYSTALLINE SILICON ROD, METHOD FOR PRODUCTION OF POLYCRYSTALLINE SILICON ROD, POLYCRYSTALLINE SILICON ROD, AND POLYCRYSTALLINE SILICON INGOT

Abstract

An average diffraction intensity ratio (y=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) for a rotation angle () is obtained from a first diffraction chart and a second diffraction chart, and a surface temperature during deposition is calculated based on this average diffraction intensity ratio. Based on data on the surface temperature of a polycrystalline silicon rod calculated and supplied current and applied voltage during the deposition of the polycrystalline silicon rod, the supplied current and the applied voltage when newly manufacturing a polycrystalline silicon rod is controlled to control a surface temperature during the deposition process. By using such a temperature control method, it is also possible to control the difference T (=T.sub.cT.sub.s) between the center temperature T.sub.c and the surface temperature T.sub.s of a polycrystalline silicon rod during a deposition process to control the value of residual stress in the polycrystalline silicon rod.

Claims

1. A method for calculating a surface temperature of a polycrystalline silicon rod grown by a Siemens process, during a deposition process, comprising: taking a plate-shaped sample having a cross section perpendicular to a radial direction of the polycrystalline silicon rod as a major surface from a position corresponding to a radius R from a center line of a silicon core wire on which the polycrystalline silicon rod is deposited; disposing the plate-shaped sample at a position where Bragg reflection from a Miller index plane (h1, k1, l1) is detected, and in-plane-rotating the plate-shaped sample around a center of the plate-shaped sample as a rotation center at a rotation angle of so that an X-ray irradiation region determined by a slit -scans the major surface of the plate-shaped sample, thereby obtaining a first diffraction chart showing dependence of an intensity of Bragg reflection from the Miller index plane (h1, k1, l1) on the rotation angle () of the plate-shaped sample; disposing the plate-shaped sample at a position where Bragg reflection from a Miller index plane (h2, k2, l2) is detected, and in-plane-rotating the plate-shaped sample around the center of the plate-shaped sample as the rotation center at the rotation angle of so that an X-ray irradiation region determined by the slit -scans the major surface of the plate-shaped sample, thereby obtaining a second diffraction chart showing dependence of an intensity of Bragg reflection from the Miller index plane (h2, k2, l2) on the rotation angle () of the plate-shaped sample; obtaining an average diffraction intensity ratio (y=(h1, k1, l1)/(h2, k2, l2)) for the rotation angle () from the first diffraction chart and the second diffraction chart; and calculating a surface temperature at the position corresponding to the radius R of the polycrystalline silicon rod during deposition of polycrystalline silicon based on the average diffraction intensity ratio.

2. The method for calculating a surface temperature of a polycrystalline silicon rod according to claim 1, wherein the calculation of the surface temperature is made based on a conversion table of an average diffraction intensity ratio (y) to a surface temperature previously obtained.

3. The method for calculating a surface temperature of a polycrystalline silicon rod according to claim 2, wherein the conversion table is based on a conversion equation obtained by expressing a relationship between an estimated temperature x and the average diffraction intensity ratio y as a regression equation when the estimated temperature based on a resistivity of a polycrystalline silicon rod calculated from a diameter of the polycrystalline silicon rod and supplied current and applied voltage to the polycrystalline silicon rod is x.

4. The method for calculating a surface temperature of a polycrystalline silicon rod according to claim 1, wherein the Miller index plane (h1, k1, l1) and the Miller index plane (h2, k2, l2) are (111) and (220).

5. A method for controlling a surface temperature of a polycrystalline silicon rod while manufacturing the polycrystalline silicon rod by a Siemens process, the method comprising: based on data on a surface temperature of the polycrystalline silicon rod calculated by the method according to claim 1 and supplied current and applied voltage during deposition of the polycrystalline silicon rod, controlling supplied current and applied voltage when newly manufacturing the polycrystalline silicon rod, to control a surface temperature during a deposition process.

6. A method for manufacturing a polycrystalline silicon rod, comprising controlling a difference T (=TcTs) between a center temperature Tc and a surface temperature Ts of a polycrystalline silicon rod during a deposition process using the temperature control method according to claim 5, to control a value of residual stress in the polycrystalline silicon rod.

7. The method for manufacturing a polycrystalline silicon rod according to claim 6, wherein the T during the deposition process is consistently controlled at 70 C. or less.

8. A polycrystalline silicon rod which is grown by controlling the T at 160 C. or more in the method for manufacturing a polycrystalline silicon rod according to claim 6 and in which a remaining Roll compressive stress, tensile stress, or both, is noted.

9. A polycrystalline silicon ingot obtained by fracturing the polycrystalline silicon rod according to claim 8.

10. A polycrystalline silicon rod which is grown by controlling the T at less than 160 C. in the method for manufacturing a polycrystalline silicon rod according to claim 6 and in which a remaining compressive stress is noted, but a remaining tensile stress is not noted.

11. The method for calculating a surface temperature of a polycrystalline silicon rod according to claim 2, wherein the Miller index plane (h1, k1, l1) and the Miller index plane (h2, k2, l2) are (111) and (220).

Description

BRIEF DESCRIPTION OF DRAWINGS

[0038] FIG. 1A is a diagram for explaining an example of the taking of plate-shaped samples for X-ray diffraction measurement from a polycrystalline silicon rod deposited and grown by the Siemens process.

[0039] FIG. 1B is a diagram for explaining an example of the taking of plate-shaped samples for X-ray diffraction measurement from a polycrystalline silicon rod deposited and grown by the Siemens process.

[0040] FIG. 2 is a diagram for explaining the outline of an example of a measurement system when obtaining an X-ray diffraction profile from a plate-shaped sample by a scan method.

[0041] FIG. 3 is one example of charts obtained by performing the scan measurement shown in FIG. 2 for Miller index planes (111) and (220).

[0042] FIG. 4 is a flow diagram for explaining the outline of a method for calculating the surface temperature of a polycrystalline silicon rod according to the present invention.

[0043] FIG. 5 shows the ratio between a first diffraction chart from a Miller index plane (1,1,1) and a second diffraction chart from a Miller index plane (2,2,0) (=(1,1,1)/(2,2,0)) obtained using plate-shaped samples taken from various R.

[0044] FIG. 6 is a diagram showing the relationship between an estimated temperature x and (111)/(220) ratio when the diameter of a polycrystalline silicon rod is in the range of 10 to 30 mm.

[0045] FIG. 7 shows diffraction charts from a Miller index plane (111) and a Miller index plane (220) obtained by -scanning a plate-shaped sample taken from a position that is roughly R.sup.0/2 in a polycrystalline silicon rod.

DESCRIPTION OF EMBODIMENTS

[0046] Embodiments of the present invention will be described below with reference to the drawings.

[0047] For the purpose of developing a new method for controlling with high precision the surface temperature of a polycrystalline silicon rod during a deposition process when manufacturing a polycrystalline silicon rod by the Siemens process, the present inventors have evaluated the crystallinity of polycrystalline silicon synthesized at various CVD temperatures by an X-ray diffraction method.

[0048] FIG. 1A and FIG. 1B are diagrams for explaining an example of the taking of plate-shaped samples 20 for X-ray diffraction profile measurement from a polycrystalline silicon rod 10 deposited and grown by the Siemens process. In the figure, numeral 1 denotes a silicon core wire on the surface of which polycrystalline silicon is deposited to form a silicon rod. In this example, in order to confirm the radial dependence of the surface temperature of the polycrystalline silicon rod during deposition, the plate-shaped samples 20 are taken from three sites (CTR: a site close to the silicon core wire 1, EDG: a site close to the side surface of the polycrystalline silicon rod 10, and R.sup.0/2: an intermediate site between CTR and EGD), but the taking is not limited to taking from such sites.

[0049] The diameter of the polycrystalline silicon rod 10 illustrated in FIG. 1A is roughly 120 mm (radius R.sup.060 mm), and a rod 11 having a diameter of roughly 19 mm and a length of roughly 60 mm is hollowed perpendicularly to the longitudinal direction of the silicon core wire 1 from the side surface side of this polycrystalline silicon rod 10.

[0050] Then, as shown in FIG. 1B, plate-shaped samples (20.sub.CTR, 20.sub.EDG, and 20.sub.R/2) having a cross section perpendicular to the radial direction of the polycrystalline silicon rod 10 as a major surface and having a thickness of roughly 2 mm are taken from a site close to the silicon core wire 1 (CTR), a site close to the side surface of the polycrystalline silicon rod 10 (EDG), and an intermediate site between CTR and EGD (R/2) in this rod 11 respectively.

[0051] The site where the rod 11 is taken, the length, and the number should be appropriately determined according to the diameter of the silicon rod 10 and the diameter of the rod 11 to be hollowed, and the plate-shaped samples 20 may also be taken from any sites of the hollowed rod 11, but the sites are preferably at positions where the properties (that is, the surface temperature during deposition) of the entire silicon rod 10 can be rationally estimated.

[0052] For example, when two plate-shaped samples are obtained, the plate-shaped samples are preferably obtained from two places, positions on the center side and outside of a point half the radius from the center, with respect to the radius of the circumference of the silicon rod. Further, for example, when the positions where two samples to be compared are obtained are a position on the center side of a point one third of the radius from the center and a position on the outside of a point two thirds of the radius from the center, higher precision comparison can be performed. In addition, the number of plate-shaped samples to be compared should be two or more, and there is no particular upper limit.

[0053] In addition, the diameter of the plate-shaped sample 20 is roughly 19 mm only as an illustration, and the diameter should be properly determined in a range in which no hindrance occurs during X-ray diffraction measurement.

[0054] When the crystallinity (that is, the surface temperature during deposition) of the plate-shaped sample 20 taken by the above-described procedure from a position corresponding to a radius R from the center line of the silicon core wire 1 on which the polycrystalline silicon rod 10 is deposited is evaluated by an X-ray diffraction method, first, the above plate-shaped sample 20 is disposed at a position where Bragg reflection from a first Miller index plane (h.sub.1, k.sub.1, l.sub.1) is detected, and in-plane-rotated around the center of the plate-shaped sample 20 as the rotation center at a rotation angle of so that an X-ray irradiation region determined by a slit -scans the major surface of the plate-shaped sample 20, thereby obtaining a first diffraction chart showing the dependence of the intensity of Bragg reflection from the Miller index plane (h.sub.1, k.sub.1, l.sub.1) on the rotation angle () of the plate-shaped sample 20.

[0055] FIG. 2 is a diagram for explaining the outline of an example of a measurement system when obtaining an X-ray diffraction profile from the plate-shaped sample 20 by a scan method, and in the example shown in this figure, an X-ray beam 40 (CuK ray: wavelength 1.54 ) emitted from a slit 30 and collimated is allowed to enter a narrow rectangular region determined by the slit in a region between both circumferential ends of the plate-shaped sample 20. Then, the plate-shaped sample 20 is rotated (=0 to 360) in the YZ plane around the center of the disk-shaped sample 20 as the rotation center so that this X-ray irradiation region scans the entire surface of the plate-shaped sample 20, thereby obtaining a first diffraction chart showing the dependence of the intensity of Bragg reflection from the Miller index plane (h.sub.1, k.sub.1, l.sub.1) on the rotation angle () of the plate-shaped sample 20.

[0056] Following this, by the same procedure as the above, the plate-shaped sample 20 is disposed at a position where Bragg reflection from a second Miller index plane (h.sub.2, k.sub.2, l.sub.2) is detected, and in-plane-rotated around the center of the plate-shaped sample 20 as the rotation center at the rotation angle of so that an X-ray irradiation region determined by the slit -scans the major surface of the plate-shaped sample 20, thereby obtaining a second diffraction chart showing the dependence of the intensity of Bragg reflection from the Miller index plane (h.sub.2, k.sub.2, l.sub.2) on the rotation angle () of the plate-shaped sample 20.

[0057] FIG. 3 is one example of charts obtained by performing the above scan measurement for Miller index planes (111) and (220).

[0058] Then, an average diffraction intensity ratio (y=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) for the rotation angle () is obtained from these first diffraction chart and second diffraction chart, and a surface temperature at the position corresponding to the radius R of the polycrystalline silicon rod 10 during the deposition of polycrystalline silicon is calculated based on this average diffraction intensity ratio.

[0059] FIG. 4 is a flow diagram for explaining the outline of a method for calculating the surface temperature of a polycrystalline silicon rod according to the present invention.

[0060] That is, in the method for calculating the surface temperature of a polycrystalline silicon rod according to the present invention, by the above-described procedure, a plate-shaped sample having a cross section perpendicular to the radial direction of a polycrystalline silicon rod as a major surface is taken (S101), the intensity of Bragg reflection from the Miller index plane (h.sub.1, k.sub.1, l.sub.1) of this plate-shaped sample is obtained to obtain a first diffraction chart showing rotation angle () dependence (S102), and then the intensity of Bragg reflection from the Miller index plane (h.sub.2, k.sub.2, l.sub.2) of the plate-shaped sample is obtained to obtain a second diffraction chart showing rotation angle () dependence (S103). Then, an average diffraction intensity ratio (y=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) for the rotation angle () is obtained from the above-described first diffraction chart and second diffraction chart (S104), and a surface temperature at a position corresponding to the radius R of the polycrystalline silicon rod during the deposition of polycrystalline silicon is calculated based on this average diffraction intensity ratio (S105).

[0061] In this manner, the method for calculating the surface temperature of a polycrystalline silicon rod according to the present invention is a method for calculating the surface temperature of a polycrystalline silicon rod grown by the Siemens process, during a deposition process, comprising the steps of taking a plate-shaped sample having a cross section perpendicular to the radial direction of the above polycrystalline silicon rod as a major surface from a position corresponding to a radius R from the center line of a silicon core wire on which the above polycrystalline silicon rod is deposited; disposing the above plate-shaped sample at a position where Bragg reflection from a Miller index plane (h.sub.1, k.sub.1, l.sub.1) is detected, and in-plane-rotating the above plate-shaped sample around the center of the above plate-shaped sample as the rotation center at a rotation angle of so that an X-ray irradiation region determined by a slit -scans the major surface of the plate-shaped sample, thereby obtaining a first diffraction chart showing the dependence of the intensity of Bragg reflection from the above Miller index plane (h.sub.1, k.sub.1, l.sub.1) on the rotation angle () of the above plate-shaped sample; disposing the above plate-shaped sample at a position where Bragg reflection from a Miller index plane (h.sub.2, k.sub.2, l.sub.2) is detected, and in-plane-rotating the above plate-shaped sample around the center of the above plate-shaped sample as the rotation center at the rotation angle of so that an X-ray irradiation region determined by the slit -scans the major surface of the above plate-shaped sample, thereby obtaining a second diffraction chart showing the dependence of the intensity of Bragg reflection from the above Miller index plane (h.sub.2, k.sub.2, l.sub.2) on the rotation angle () of the above plate-shaped sample; obtaining an average diffraction intensity ratio (y=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) for the above rotation angle () from the above first diffraction chart and the above second diffraction chart; and calculating a surface temperature at the position corresponding to the radius R of the above polycrystalline silicon rod during the deposition of polycrystalline silicon based on the above average diffraction intensity ratio.

[0062] The calculation of the surface temperature in step S105 is made, for example, based on a conversion table of an average diffraction intensity ratio (y) to a surface temperature previously obtained.

[0063] Such a conversion table is obtained, for example, based on a conversion equation obtained by expressing the relationship between an estimated temperature x and an average diffraction intensity ratio y as a regression equation when the estimated temperature based on the resistivity of a polycrystalline silicon rod calculated from the diameter of the polycrystalline silicon rod and supplied current and applied voltage to the polycrystalline silicon rod is x.

[0064] The Miller index plane (h.sub.1, k.sub.1, l.sub.1) and the Miller index plane (h.sub.2, k.sub.2, l.sub.2) are preferably (111) and (220).

[0065] FIG. 5 shows one example of the ratio between a first diffraction chart from a Miller index plane (h.sub.1, k.sub.1, l.sub.1) and a second diffraction chart from a Miller index plane (h.sub.2, k.sub.2, l.sub.2) (=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) obtained using plate-shaped samples taken from various R, and here, the Miller index plane (h.sub.1, k.sub.1, l.sub.1)=(1,1,1), and the Miller index plane (h.sub.2, k.sub.2, l.sub.2)=(2,2,0).

[0066] This figure shows the diffraction intensity ratio (y=(111)/(220):left vertical axis) obtained from each sample and the converted surface temperature corresponding to the diffraction intensity ratio (right vertical axis). The samples are obtained as follows. A polycrystalline silicon rod having a diameter of about 160 mm (R.sup.080 mm) is grown, and 10 (a total of 20) plate-shaped samples are taken at intervals of 8 to 12 mm in the radial direction from the center line of the silicon core wire used for the deposition of the polycrystalline silicon rod.

[0067] This polycrystalline silicon rod is grown by making the surface temperature during deposition constant by a current value control method, a conventional method, and it is seen that the (111)/(220) ratio (that is, crystallinity) is different depending on the site. This means that the surface temperature of the polycrystalline silicon rod is different depending on the site. As the surface temperature during deposition becomes lower, the (111) diffraction becomes more dominant. On the other hand, as the surface temperature during deposition becomes higher, the (220) diffraction becomes more dominant.

[0068] That is, by obtaining a first diffraction chart and a second diffraction chart and obtaining an average diffraction intensity ratio (y=(h.sub.1, k.sub.1, l.sub.1)/(h.sub.2, k.sub.2, l.sub.2)) for a rotation angle () by the above-described method, it is possible to calculate a surface temperature at a position corresponding to the radius R of a polycrystalline silicon rod during the deposition of polycrystalline silicon.

[0069] For such calculation of the surface temperature, it is necessary to previously confirm the correspondence relationship between an average diffraction intensity ratio (y) and a surface temperature.

[0070] Therefore, the present inventors have performed the following experiment. In a state in which the diameter of a polycrystalline silicon rod is small, the difference between the center temperature and the surface temperature is extremely small. Therefore, when the estimated temperature based on the resistivity of a polycrystalline silicon rod calculated from the diameter of the polycrystalline silicon rod and supplied current and applied voltage to the polycrystalline silicon rod is x, the estimated temperature x is a value close to the actual surface temperature. That is, when the relationship between the above estimated temperature x and the above (111)/(220) ratio in a state in which the diameter of a polycrystalline silicon rod is small is found, the surface temperature in the state can be calculated based on the relationship from the (111)/(220) ratio in a state in which deposition proceeds.

[0071] Therefore, the relationship between the above estimated temperature x and the above (111)/(220) ratio when the diameter is in the range of 10 to 30 mm is obtained.

[0072] FIG. 6 is a diagram showing the relationship between an estimated temperature x and (111)/(220) ratio when the diameter of a polycrystalline silicon rod is in the range of 10 to 30 mm. The equation shown in the figure is a conversion equation obtained by expressing the relationship between the estimated temperature x and the above average diffraction intensity ratio y as a regression equation.

[0073] The results shown in this figure show that when the relationship between an average diffraction intensity ratio (y) and a surface temperature (referred to as a conversion table for convenience) is previously obtained, the surface temperature of a polycrystalline silicon rod grown by the Siemens process, during a deposition process, can be calculated. Such a conversion table can, for example, be based on a conversion equation obtained by expressing the relationship between an estimated temperature x and the above average diffraction intensity ratio y as a regression equation when the estimated temperature based on the resistivity of a polycrystalline silicon rod calculated from the diameter of the above polycrystalline silicon rod and supplied current and applied voltage to the polycrystalline silicon rod is x.

[0074] In the actual CVD process, when the trichlorosilane as concentration and flow rate and the hydrogen as concentration and flow rate are changed, the surface temperature of the synthesized silicon polycrystal also changes naturally, and the change is directly reflected in a change in crystallinity. Therefore, the change appears in (111)/(220) ratio.

[0075] Therefore, when the temperature when a polycrystalline silicon rod is manufactured by the Siemens process is controlled, it is possible to, based on data on the surface temperature of a polycrystalline silicon rod calculated by the above-described method and supplied current and applied voltage during the deposition of the polycrystalline silicon rod, control supplied current and applied voltage when newly manufacturing a polycrystalline silicon rod, to control a surface temperature during the deposition process.

[0076] By using such a temperature control method, it is also possible to control the difference T (=T.sub.cT.sub.s) between the center temperature T.sub.c and the surface temperature T.sub.s of a polycrystalline silicon rod during a deposition process to control the value of residual stress in the polycrystalline silicon rod.

[0077] For example, it is also possible to consistently control T during a deposition process at 70 C. or less or control T at less than 160 C. (for example, eliminate the difference T between the center temperature and the surface temperature) or, on the contrary, control T at 160 C. or more to grow a polycrystalline silicon rod, and the like.

[0078] As described above, when the use of polycrystalline silicon is a raw material for monocrystalline silicon manufacture by the CZ method, the polycrystalline silicon preferably has moderate ease of breaking so as to be easily crushed into a nugget shape (polycrystalline silicon ingot). Therefore, a polycrystalline silicon ingot obtained by fracturing a polycrystalline silicon rod grown by controlling T at 160 C. or more is suitable for this use.

[0079] On the other hand, when the use of polycrystalline silicon is a raw material for monocrystalline silicon manufacture by the FZ method, polycrystalline silicon that is not easily fractured so that a polycrystalline silicon rod does not fall, collapse, or the like in a state of being set in an FZ furnace, and has low residual stress is preferred. Therefore, a polycrystalline silicon rod grown by controlling T at less than 160 C. is suitable for this use.

[0080] According to an experiment by the present inventors, when T is 160 C. or less, the stress remaining when a CVD process is completed and the polycrystalline silicon rod is cooled to room temperature is only compressive stress, and no tensile stress occurs. In the measurement of residual stress in this experiment, a method of precisely measuring an interplanar spacing value d by an X-ray diffraction method is adopted. For the measurement directions, measurement is performed in three directions, the rr direction that is the growth direction, the direction that is a direction at a right angle to the rr direction, and the zz direction that is the vertical direction.

Examples

[0081] The methods for calculating and controlling the surface temperature of a polycrystalline silicon rod according to the present invention will be described below by Examples.

Example 1 (Surface Temperature During Deposition and Average Diffraction Intensity Ratio)

[0082] Disk-shaped samples (diameter 19 mm, thickness 2 mm) used for the calculation of the surface temperature were sampled according to a method described in Japanese Patent Laid-Open No. 2014-1096 (Patent Literature 3). The diameter of each of polycrystalline silicon rods obtained by deposition on silicon core wires constructed in an inverted U shape was 160 mm, and the height from the lower end to the upper end (the vicinity of the bridge) was about 1,800 mm. In addition, the above silicon core wires were disposed in the central portion and its periphery in a furnace as multi-ring type rod disposition, and polycrystalline silicon was deposited on these silicon core wires.

[0083] A cylindrical core having a diameter of 19 mm whose center was in the right angle direction (growth direction) was hollowed from each of sites in the vicinities of the bridges of three polycrystalline silicon rods obtained in this manner and sites 300 mm from the lower ends, and the above disk-shaped samples were obtained at regular intervals of 8 to 12 mm.

[0084] For X-ray diffraction measurement, the sample surface needs to be flat. Therefore, in order to remove the slicing marks, the surface was polished with an abrasive (Carborundum #300), and after the polishing, etching was performed with a mixed acid with HF:HNO.sub.3=1:5 (HF=50 wt %, HNO.sub.3=70 wt %) for 1 minute for mirror finish.

[0085] For each of these disk-shaped samples, according to the method described in Japanese Patent Laid-Open No. 2014-1096 (Patent Literature 3), scan X-ray diffraction charts from the Miller index planes (111) and (220) were obtained, and the average value of diffraction intensity was calculated for each sample. For the average value of diffraction intensity, when no peaks are present in the diffraction chart, it is possible to make a visual determination and read the average value on the chart, but when many peaks are detected in the diffraction chart, the diffraction intensities of these peaks are also included in the amount of detection for averaging.

[0086] According to these measurement results, as the surface temperature during deposition becomes lower, the diffraction from the Miller plane (111) becomes more dominant, and as the surface temperature during deposition becomes higher, the diffraction from the Miller plane (220) becomes more dominant. The present inventors understand the reason as follows.

[0087] The electronic structure of Si is 1s.sup.22 s.sup.22p.sup.63 s.sup.23p.sup.2, and a total of four valence electrons, that is, outermost electrons, two in the 3s orbital and two in the 3p orbital, are present. Therefore, for example, when two Si molecules are formed by a CVD reaction, a total of eight electrons, four electrons present in the outermost shell of one molecule and four electrons present in the outermost shell of the other molecule, form a closed shell structure for stabilization.

[0088] Also when silicon deposits as crystals at relatively low temperature, the same as this occurs. As is well known, electron orbitals in which s orbitals and p orbitals are mixed form four equivalent orbitals forming the apexes of a regular tetrahedron at an angle of 109.5 with each other. Four apexes of these orbitals correspond to the apexes of the regular tetrahedron, and each plane of it corresponds to {111}. The {111} plane of a face-centered cubic lattice is the most dense plane in which the number of atoms per unit area is the largest and is the most stable crystal face. Therefore, crystal growth on the {111} plane is dominant at the initial stage of crystal growth, and in the CVD reaction of a trichlorosilane system, crystal growth on the {111} plane is confirmed even at a considerably low temperature such as about 600 to 700 C.

[0089] But, when the deposition temperature increases, the deposition rate increases significantly, and the number of silicon atoms involved in crystal formation increases. Therefore, from the viewpoint of structural stability as the entire crystal bulk, crystal growth on other crystal faces (for example, the {100} plane) including the {110} plane is dominant.

[0090] FIG. 7 shows diffraction charts from a Miller index plane (111) and a Miller index plane (220) obtained by -scanning a plate-shaped sample taken from a position that is roughly R.sup.0/2 in a polycrystalline silicon rod with diameter R.sup.0=160 mm. No peaks are observed in the diffraction chart for the Miller index plane (111), and on the other hand a large number of peaks are observed in the diffraction chart for the Miller index plane (220). The presence of these diffraction peaks means that acicular crystals grow locally in the [220] direction during deposition.

[0091] As described above, scan X-ray diffraction charts from Miller index planes (111) and (220) are obtained, and the average value of diffraction intensity is obtained for each sample, and the average value is checked with the relationship between an estimated temperature x and (111)/(220) ratio shown in FIG. 6 to calculate a surface temperature during deposition.

[0092] From the results, the following facts became clear. First, the surface temperature of the site in the vicinity of the bridge is higher than the surface temperature of the site 300 mm from the lower end. Second, the above surface temperature difference is smaller on the center side of the furnace. Third, the temperature difference T in the growth direction is lower in the vicinity of the bridge than at the site 300 mm from the lower end. These findings are facts clarified for the first time by the present invention.

[0093] The diffraction chart ratio (=(1,1,1)/(2,2,0)) shown in FIG. 5 shows results obtained using the plate-shaped samples taken from the site 300 mm from the lower end described above, in this Example.

[0094] In the example shown in this figure, the surface temperature of the central portion (the site close to the silicon core wire) during deposition is relatively low, and the surface temperature during deposition becomes relatively higher as the site becomes closer to the outermost surface side. The difference T reaches 164 C.

[0095] A polycrystalline silicon rod grown under such conditions breaks easily, and is in a state in which compressive stress and tensile stress coexist in all sites of the polycrystalline silicon rod according to residual stress measurement.

[0096] A polycrystalline silicon rod was grown with temperature control for decreasing the above surface temperature difference T performed and other conditions unchanged. Specifically, during deposition in the vicinity of the silicon core wire, current was supplied so that the surface temperature was 1180 C., and in all steps of deposition, the supplied current was controlled so that the surface temperature was in the target temperature range of 1150 to 1180 C.

[0097] For the polycrystalline silicon rod grown under such conditions, the above-described diffraction intensity ratio was obtained and converted to temperature. The result was that the surface temperature difference T was controlled at 48 to 73 C. in all sites. In addition, the residual stress of this polycrystalline silicon rod was measured, and only compressive stress was noted in all sites.

[0098] A polycrystalline silicon rod for obtaining polycrystalline silicon ingots (nuggets) for silicon single crystal growth by the CZ method is desirably easily fractured. For this purpose, a higher value of tensile stress in residual stress in the polycrystalline silicon rod is more advantageous. But, drawbacks of such a polycrystalline silicon rod are that the polycrystalline silicon rod collapses easily in the reaction furnace during the cooling step after the completion of the deposition step, and the like. Therefore, there is an appropriate upper limit value of tensile stress remaining in the polycrystalline silicon rod.

[0099] In order to set residual tensile stress in the polycrystalline silicon rod equal to or less than the above-described appropriate upper limit value, the difference T between the surface temperature of the central portion (the site close to the silicon core wire) during deposition and the surface temperature of the outermost surface portion during deposition during the deposition step needs to be controlled at 200 C. or less.

[0100] On the other hand, a polycrystalline silicon rod for obtaining polycrystalline silicon for silicon single crystal growth by the FZ method or polycrystalline silicon for recharge during silicon single crystal growth by the CZ method is desirably not easily fractured, and a smaller value of the above T is better.

[0101] According to the present invention, it is possible to control with high precision the surface temperature of a polycrystalline silicon rod during a deposition process when manufacturing a polycrystalline silicon rod by the Siemens process, and therefore it is possible to also control the above T with high precision.

[0102] That is, by manufacturing a polycrystalline silicon rod by controlling the difference T (=T.sub.cT.sub.s) between the center temperature T.sub.c and the surface temperature T.sub.s of a polycrystalline silicon rod during a deposition process using the above-described temperature control method, to control the value of residual stress in the polycrystalline silicon rod, it is possible to distinctively make a polycrystalline silicon rod for obtaining polycrystalline silicon ingots (nuggets) for silicon single crystal growth by the CZ method, and a polycrystalline silicon rod for obtaining polycrystalline silicon for silicon single crystal growth by the FZ method or polycrystalline silicon for recharge during silicon single crystal growth by the CZ method.

[0103] For example, in the case of a polycrystalline silicon rod for obtaining polycrystalline silicon ingots (nuggets) for silicon single crystal growth by the CZ method, the above T is controlled at 160 C. or more to grow the polycrystalline silicon rod.

[0104] On the other hand, in the case of a polycrystalline silicon rod for obtaining polycrystalline silicon for silicon single crystal growth by the FZ method or polycrystalline silicon for recharge during silicon single crystal growth by the CZ method, the above T is controlled at less than 160 C. to grow the polycrystalline silicon rod. Preferably, the above T is consistently controlled at 70 C. or less.

Example 2 (Surface Temperature Control During Deposition Process)

[0105] A polycrystalline silicon rod having a diameter of 160 mm was newly grown while, based on data on the surface temperature of a polycrystalline silicon rod calculated based on the results shown in FIG. 5 and supplied current and applied voltage during the deposition of the polycrystalline silicon rod, the supplied current and the applied voltage when newly manufacturing a polycrystalline silicon rod was controlled to control a surface temperature during the deposition process.

[0106] Plate-shaped samples were taken from various sites of this polycrystalline silicon rod, and the surface temperature during deposition was calculated. The results are summarized in Table 1.

TABLE-US-00001 TABLE 1 Vicinity of 300 mm from lower Radial direction bridge end Height direction (mm) ( C.) ( C.) T ( C.) +80 1154 1155 1 +60 1156 1156 0 +50 1142 1148 6 +40 1155 1110 45 +30 1159 1157 2 +20 1180 1162 18 +10 1180 1183 3 0 10 1188 1160 28 20 1190 1180 10 30 1162 1160 2 40 1160 1135 25 50 1150 1137 13 60 1155 1145 10 80 1154 1150 4 Growth direction 48 73 T ( C.)

[0107] The residual stress in this polycrystalline silicon rod was only compressive stress in any of the above-described three directions. In addition, a silicon single crystal was grown by the FZ method using as a raw material a polycrystalline silicon rod grown under the same conditions, and troubles such as collapse and fall did not occur.

Example 3 (Surface Temperature Difference T During Deposition Process and Residual Stress)

[0108] The relationship between a surface temperature difference T ( C.) during a deposition process and residual stress was obtained. The results are summarized in Table 2.

[0109] When T is 160 C. or more, the remaining of compressive stress and tensile stress is noted. On the other hand, when T is less than 160 C., only the remaining of compressive stress is noted, and the remaining of tensile stress is not noted.

[0110] In addition, when T exceeds 170 C., collapse in a CVD reaction furnace may occur. When T exceeds 200 C., collapse in a CVD reaction furnace frequently occurs, which is dangerous, and therefore this case was excluded in this Example. Further, the degree of difficulty of fracturing by a hammer changes at T=160 C. When T is 160 C. or more, the polycrystalline silicon rod is easily fractured. When T is less than 160 C., the polycrystalline silicon rod is not easily fractured. When T is 170 C. or more, the polycrystalline silicon rod is extremely brittle to the extent that one hesitates to hold the polycrystalline silicon rod in an FZ furnace. When a polycrystalline silicon raw material obtained from a polycrystalline silicon rod grown with T=165 C. was held in an FZ furnace, fall in the furnace sometimes occurred.

TABLE-US-00002 TABLE 2 T 42 89 140 156 160 165 177 182 195 Compressive stress Tensile stress x x x x Collapse in No No No No No No Yes Yes Yes furnace Degree of Difficult Difficult Difficult Difficult Easy Easy Brittle Brittle Brittle difficulty of fracturing Fall in No No No No No Yes Impossible Impossible Impossible furnace

INDUSTRIAL APPLICABILITY

[0111] The present invention provides a technique for manufacturing a polycrystalline silicon rod based on a new method for controlling with high precision the surface temperature of a polycrystalline silicon rod during a deposition process when manufacturing a polycrystalline silicon rod by the Siemens process.

REFERENCE SIGNS LIST

[0112] 1 silicon core wire [0113] 10 polycrystalline silicon rod [0114] 11 rod [0115] 20 plate-shaped sample [0116] 30 slit