POLYCRYSTALLINE SILICON ROD, PRODUCTION METHOD THEREFOR, AND FZ SILICON SINGLE CRYSTAL

Abstract

A plate-shaped sample with a cross-section perpendicular to a radial direction of a polycrystalline silicon rod as a principal surface is sampled from a region from a center (r=0) of the polycrystalline silicon rod to R/3. Then, the sample is disposed at a position at which a Bragg reflection from a (111) Miller index plane is detected. In-plane rotation with a rotational angle φ on the sample is performed with a center of the sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the sample. A ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart is calculated.

Claims

1.-6. (canceled)

7. A production method for a polycrystalline silicon rod having a radius R of 65 mm or more by a chemical vapor deposition method, the production method comprising causing the polycrystalline silicon rod to grow under a condition set such that a feed gas amount is reduced by 2% or more in a second region from r=R/3 to r=R/2, and furthermore, the feed gas amount is reduced by 5% or more in a third region from r=R/2 to r=R in a deposition step of the polycrystalline silicon rod, where the feed gas amount at the start of deposition is set to be 100, and r=0 at a center of the polycrystalline silicon rod.

8. A polycrystalline silicon rod produced by the method according to claim 7, wherein an area ratio S.sub.p/S.sub.t which is evaluated through the following procedure by an X-ray diffraction method being 2% or less, the procedure including: (1a) sampling a plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from a region from a center (r=0) of the polycrystalline silicon rod to R/3; (1b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected; (1c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample; and (1d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart.

9. The polycrystalline silicon rod according to claim 8, wherein the peak part is defined to have an S/N ratio of 3 or more.

10. An FZ silicon single crystal obtained by being caused to grow with the polycrystalline silicon rod according to claim 8 as a raw material.

11. A polycrystalline silicon rod produced by the method according to claim 7, wherein an average value of area ratios S.sub.p/S.sub.t which are evaluated through the following procedure by an X-ray diffraction method being 0.5% or less, the procedure including: (2a) sampling at least one plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from each region of a first region from a center (r=0) of the polycrystalline silicon rod to r=R/3, a second region from r=R/3 to r=R/2, and a third region from r=R/2 to r=R; (2b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected; (2c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample; (2d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart; and (2e) calculating an average value of the area ratios S.sub.p/S.sub.t for the plurality of plate-shaped samples.

12. The polycrystalline silicon rod according to claim 11, wherein the peak part is defined to have an S/N ratio of 3 or more.

13. An FZ silicon single crystal obtained by being caused to grow with the polycrystalline silicon rod according to claim 11 as a raw material.

14. The production method for a polycrystalline silicon rod according to claim 7, wherein the condition is set such that the feed gas amount is reduced by 2% to 5% in the second region from r=R/3 to r=R/2, and furthermore, the feed gas amount is reduced by 5% to 8% in the third region from r=R/2 to r=R.

15. A polycrystalline silicon rod produced by the method according to claim 14, wherein an area ratio S.sub.p/S.sub.t which is evaluated through the following procedure by an X-ray diffraction method being 2% or less, the procedure including: (1a) sampling a plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from a region from a center (r=0) of the polycrystalline silicon rod to R/3; (1b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected; (1c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample; and (1d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart.

16. The polycrystalline silicon rod according to claim 15, wherein the peak part is defined to have an S/N ratio of 3 or more.

17. An FZ silicon single crystal obtained by being caused to grow with the polycrystalline silicon rod according to claim 15 as a raw material.

18. A polycrystalline silicon rod produced by the method according to claim 14, wherein an average value of area ratios S.sub.p/S.sub.t which are evaluated through the following procedure by an X-ray diffraction method being 0.5% or less, the procedure including: (2a) sampling at least one plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from each region of a first region from a center (r=0) of the polycrystalline silicon rod to r=R/3, a second region from r=R/3 to r=R/2, and a third region from r=R/2 to r=R; (2b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected; (2c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample; (2d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart; and (2e) calculating an average value of the area ratios S.sub.p/S.sub.t for the plurality of plate-shaped samples.

19. The polycrystalline silicon rod according to claim 18, wherein the peak part is defined to have an S/N ratio of 3 or more.

20. An FZ silicon single crystal obtained by being caused to grow with the polycrystalline silicon rod according to claim 18 as a raw material.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0030] FIG. 1A is a diagram for explaining an example of sampling plate-shaped samples for X-ray diffraction measurement from a polycrystalline silicon rod obtained by being caused to grow through deposition by a chemical vapor deposition method.

[0031] FIG. 1B is a diagram for explaining an example of sampling plate-shaped samples for X-ray diffraction measurement from a polycrystalline silicon rod obtained by being caused to grow through deposition by a chemical vapor deposition method.

[0032] FIG. 2 is a diagram for schematically explaining an example of a measurement system when obtaining an X-ray diffraction profile from the plate-shaped sample by a so-called θ-2θ method.

[0033] FIG. 3 is an example of an X-ray diffraction chart of θ-2θ.

[0034] FIG. 4 is a diagram for schematically explaining an example of a measurement system when obtaining an X-ray diffraction profile from the plate-shaped sample by a so-called φ-scanning method.

[0035] FIG. 5A is an example of a φ-scanning diffraction chart obtained from a plate-shaped sample sampled from a polycrystalline silicon rod in which disappearance of the crystal habit line arises when using the polycrystalline silicon rod as a production raw material of an FZ silicon single crystal, and FIG. 5A shows one for a sample sampled from a region A (substantial center (r≈0) of the polycrystalline silicon rod).

[0036] FIG. 5B is an example of a φ-scanning diffraction chart obtained from a plate-shaped sample sampled from the polycrystalline silicon rod in which disappearance of the crystal habit line arises when using the polycrystalline silicon rod as the production raw material of the FZ silicon single crystal, and FIG. 5B shows one for a sample sampled from a region B (r≈R/2).

[0037] FIG. 5C is an example of a φ-scanning diffraction chart obtained from a plate-shaped sample sampled from the polycrystalline silicon rod in which disappearance of the crystal habit line arises when using the polycrystalline silicon rod as the production raw material of the FZ silicon single crystal, and FIG. 5C shows one for a sample sampled from a region C (r≈R).

[0038] FIG. 6 is a diagram obtained by collecting data that were obtained by using 40 polycrystalline silicon rods for relation between FZ yields (%) and area ratios (S.sub.p/S.sub.t) of a (111) diffraction peak part.

DESCRIPTION OF EMBODIMENTS

[0039] Hereafter, an embodiment of the present invention is described with reference to the drawings.

[0040] During an advance of the study on crystal quality improvement of polycrystalline silicon rods for the purpose of stable production of FZ single crystal silicon having a large diameter of 130 mm or more, the inventors have found the knowledge that a difference in the degree of crystal orientation arises among polycrystalline silicon rods depending on conditions in deposition of polycrystalline silicon. As disclosed in Patent Literature 3, differently from single crystal silicon, a polycrystalline silicon block contains many crystal grains, and it tends to be thought that these many crystal grains are in random orientation. Nevertheless, according to inventors' study, the crystal grains contained in the polycrystalline silicon block are not necessarily in complete random orientation.

[0041] Further, as apparent from there being a case where, as mentioned above, disappearance of the crystal line arises in producing single crystal silicon even when the number of peaks from the (111) Miller index plane is 24 quantity/cm.sup.2 or less in the case of a polycrystalline silicon rod, for example, having a diameter of 150 mm, more highly precise evaluation of the degree of crystal orientation is requested for a larger-diameter polycrystalline silicon rod.

[0042] According to inventors' experiments in use of polycrystalline silicon rods having diameters of 150 to 200 mm as FZ silicon single crystal formation raw materials, there has been found relation between the presence or absence of disappearance of the crystal habit line during the melting zone process and the crystal orientation of the polycrystalline silicon rod.

[0043] While in the method disclosed in Patent Literature 3, the number of peaks appearing in a φ-scanning chart is employed as the index of crystal orientation, in place of this in the present invention, a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart is used as the index of crystal orientation, which thereby enables more highly precise evaluation of the degree of crystal orientation.

[0044] Specifically, targeting polycrystalline silicon rods obtained by being caused to grow through deposition by a chemical vapor deposition method and having a radius R of 65 mm or more (that is, a diameter of 130 mm or more), a polycrystalline silicon rod with an area ratio S.sub.p/S.sub.t which is evaluated through the following procedure by an X-ray diffraction method being 2% or less is selected as a single crystal silicon production raw material, the procedure including:

[0045] (1a) sampling a plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from a region from a center (r=0) of the polycrystalline silicon rod to R/3;

[0046] (1b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected;

[0047] (1c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample; and

[0048] (1d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart.

[0049] For more highly precise evaluation, a polycrystalline silicon rod with an average value of area ratios S.sub.p/S.sub.t which are evaluated through the following procedure by an X-ray diffraction method being 0.5% or less is selected as a single crystal silicon production raw material, the procedure including:

[0050] (2a) sampling at least one plate-shaped sample with a cross-section perpendicular to a radial direction of the polycrystalline silicon rod as a principal surface from each region of a first region from a center (r=0) of the polycrystalline silicon rod to r=R/3, a second region from r=R/3 to r=R/2, and a third region from r=R/2 to r=R;

[0051] (2b) disposing the plate-shaped sample at a position at which a Bragg reflection from a (111) Miller index plane is detected;

[0052] (2c) performing in-plane rotation with a rotational angle φ on the plate-shaped sample with a center of the plate-shaped sample as a rotational center such that an X-ray irradiation region defined by a slit performs φ-scanning on the principal surface of the plate-shaped sample to obtain a diffraction chart indicating dependency of a Bragg reflection intensity from the (111) Miller index plane on a rotational angle of the plate-shaped sample;

[0053] (2d) calculating a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a peak part appearing in the diffraction chart and a total area S.sub.t of the diffraction chart; and

[0054] (2e) calculating an average value of the area ratios S.sub.p/S.sub.t for the plurality of plate-shaped samples.

[0055] According to inventors' study, the aforementioned area ratio (S.sub.p/S.sub.t) shows dependency on a site where the plate-shaped sample is sampled (distance from the center of the polycrystalline silicon rod). As general tendency of polycrystalline silicon rods having diameters of 130 mm or more, it shows the highest value immediately in the vicinity of the center part (silicon core wire), which value decreases as going more toward the surface side. Moreover, it has been confirmed that when using a polycrystalline silicon rod as an FZ single crystal formation raw material, the aforementioned area ratio (S.sub.p/S.sub.t) of the polycrystalline silicon rod that is used as the raw material when the crystal habit line disappears during the FZ process is relatively high.

[0056] Based on such results of the study, for the present invention, it has been contrived to select, as a single crystal silicon production raw material, a polycrystalline silicon rod with the area ratio S.sub.p/S.sub.t being 2% or less, or with the average value of the area ratios S.sub.p/S.sub.t being 0.5% or less.

[0057] Notably, the aforementioned peak part is preferably defined to have an S/N ratio of 3 or more.

[0058] The inventors think out the reason that the area S.sub.p(specifically, the area ratio S.sub.p/S.sub.t) of the peak part appearing in the diffraction chart indicating dependency of the Bragg reflection intensity from the (111) Miller index plane on the rotational angle of the plate-shaped sample can be the index of crystal orientation as follows.

[0059] A chemical vapor deposition method such as a Siemens method is a method of feeding hydrogen gas and trichlorosilane gas onto the surface of a silicon core wire which is red-heated at high temperature to deposit polycrystals thereon through the CVD reaction and to enlarge its diameter.

[0060] In this deposition process, a difference (ΔT) in temperature between the center part and the surface is almost negligible in the stage of its small diameter. Nevertheless, as the diameter becomes larger, ΔT gradually increases. Control of the surface temperature of the polycrystalline silicon rod is performed while controlling applied voltage and supplied current to the silicon core wire in order to attain the red-heated state at high temperature. In the stage of a large diameter, a difference (ΔI=I.sub.c−I.sub.s>0) between an amount (I.sub.c) of the current flowing in the center part and an amount (I.sub.s) of the current flowing in the surface region becomes larger.

[0061] The reason is that the surface of the polycrystalline silicon rod is always in the state where heat is removed by a hydrogen gas flow and a trichlorosilane gas flow. This means that in order to maintain the surface temperature of the polycrystalline silicon rod during the deposition process to be constant, the amount (I.sub.s) of the current flowing in the surface region is needed to be gradually increased along with enlargement of the diameter. Hence, the amount (I.sub.c) of the current flowing in the center part, which amount is in the relation of I.sub.c>I.sub.s, has to become higher.

[0062] According to inventors' study, along with a larger-diameter polycrystalline silicon rod, the phenomenon becomes more significant that during the CVD process, a region that has already crystallized partially melts, and after that, recrystallizes. Such recrystallization makes the degree of crystal orientation of the polycrystalline silicon rod finally obtained relatively high (property of random orientation relatively low). The aforementioned recrystallization is particularly significant for a polycrystalline silicon rod having a diameter 2R=130 mm or more, and it is found that the aforementioned partial melting and recrystallization due to the excess current tend to arise in a region from the center (r=0) of the polycrystalline silicon rod to R/3, which causes tendency that the degree of crystal orientation in the relevant region becomes high.

[0063] Therefore, for the present invention, it is contrived to sample a plate-shaped sample from a region from the center (r=0) of a polycrystalline silicon rod that is obtained by being caused to grow through deposition by a chemical vapor deposition method and has a radius R of 65 mm or more (diameter of 130 mm or more) to R/3 with a cross-section perpendicular to the radial direction of the polycrystalline silicon rod as a principal surface, and to evaluate its degree of crystal orientation by an X-ray diffraction method. As mentioned above, as the index of this degree of crystal orientation, the area S.sub.p (specifically, the area ratio S.sub.p/S.sub.t) of the peak part appearing in the diffraction chart indicating dependency of the Bragg reflection intensity from the (111) Miller index plane on the rotational angle of the plate-shaped sample is used.

[0064] The inventors have confirmed that when using, as a production raw material of an FZ silicon single crystal, a polycrystalline silicon rod in which a diffraction peak is found at a position corresponding to the (111) Miller index plane in a diffraction chart indicating dependency on the rotational angle of a plate-shaped sample, this tends to result in disappearance of the crystal habit line. The reason is thought to be that a polycrystalline silicon rod relatively high in the degree of crystal orientation tends to afford partial unmolten portions during the zone melting process, which disturbs the equilibrium state at the solid-liquid interface.

[0065] FIG. 1A and FIG. 1B are diagrams for explaining an example of sampling plate-shaped samples 20 for X-ray diffraction profile measurement from a polycrystalline silicon rod 10 obtained by being caused to grow through deposition by a chemical vapor deposition method such as a Siemens method. In the figure, reference numeral 1 designates a silicon core wire for depositing polycrystalline silicon on its surface to afford the silicon rod. Notably, in this example, in order to examine the presence or absence of radial directional dependency of the degree of crystal orientation in the polycrystalline silicon rod, the plate-shaped samples 20 are sampled from three regions (a region A: a first region from the center (r=0) of the polycrystalline silicon rod to r=R/3, a region B: a second region from r=R/3 to r=R/2, and a region C: a third region from r=R/2 to r=R).

[0066] The diameter of the polycrystalline silicon rod 10 exemplarily shown in FIG. 1A is approximately 150 mm. This polycrystalline silicon rod 10 is hollowed out from its lateral surface side perpendicularly to the longitudinal direction of the silicon core wire 1 to afford a rod 11 with a diameter of approximately 20 mm and a length of approximately 75 mm.

[0067] Then, as shown in FIG. 1B, the plate-shaped samples (20.sub.A, 20.sub.B and 20.sub.C) the principal surface of each of which is a cross-section perpendicular to the radial direction of the polycrystalline silicon rod 10 and the thickness of each of which is approximately 2 mm are respectively sampled from a site (region A), in the rod 11, close to the silicon core wire 1, a region (region B) from r=R/3 to r=R/2 of the polycrystalline silicon rod, and a site (region C) close to the lateral surface of the polycrystalline silicon rod 10.

[0068] Notably, the site from which the rod 11 is sampled, the length of the rod 11, and the number of the rods 11 only have to be properly determined in accordance with the diameter of the silicon rod 10 and the diameter of the rod 11 obtained by hollowing-out. While the plate-shaped samples 20 may be sampled from any sites in the rod 11 obtained by hollowing-out, the sites are preferably positions where the nature of the whole silicon rod 10 can be rationally estimated. In the present invention, at least from the region from the center (r=0) of the polycrystalline silicon rod to R/3, a plate-shaped sample is sampled.

[0069] Moreover, while in the aforementioned example, one plate-shaped sample 20 is sampled from each of the three regions (the region A: the first region from the center (r=0) of the polycrystalline silicon rod to r=R/3, the region B: the second region from r=R/3 to r=R/2, and the region C: the third region from r=R/2 to r=R), if a plurality of plate-shaped samples 20 are sampled, the degree of crystal orientation can be more precisely evaluated. Notably, as mentioned above, in the present invention, at least from the aforementioned region A (first region from the center (r=0) of the polycrystalline silicon rod to r=R/3), a plate-shaped sample is definitely sampled.

[0070] Furthermore, to set the diameter of the plate-shaped sample 20 to be approximately 20 mm is merely exemplary, and the diameter thereof only has to be properly determined so as not to cause a problem in X-ray diffraction measurement.

[0071] FIG. 2 is a diagram for schematically explaining an example of a measurement system when obtaining an X-ray diffraction profile from the plate-shaped sample 20 by a so-called θ-2θ method. An X-ray beam 40 (Cu-Kα line: 1.54 Å of wavelength) which is emitted from a slit 30 and collimated enters the plate-shaped sample 20. The intensity of the diffracted X-ray beam is detected by a detector (not shown) for each sample rotational angle (θ) while rotating the plate-shaped sample 20 in the XY-plane to obtain an X-ray diffraction chart of θ-2θ.

[0072] FIG. 3 is an example of the X-ray diffraction chart of θ-2θ obtained above. In the chart, strong Bragg reflections from <111>, <220>, <311> and <400> Miller index planes appear as peaks at the positions with 2θ=28.40°, 47.24°, 55.98° and 68.98°, respectively.

[0073] FIG. 4 is a diagram for schematically explaining a measurement system when obtaining an X-ray diffraction profile from the plate-shaped sample 20 by a so-called φ-scanning method. For example, θ of the plate-shaped sample 20 above is set to be the angle at which the Bragg reflection from the (111) Miller index plane is detected, in this state, an X-ray is radiated onto a thin rectangular region defined by a slit in a region across from the center to the circumferential end of the plate-shaped sample 20, and the plate-shaped sample 20 is rotated (φ=0° to 360°) in the YZ-plane with the center of the plate-shaped sample 20 as the rotational center such that this X-ray irradiation region scans the whole surface of the plate-shaped sample 20.

[0074] FIGS. 5A to 5C are φ-scanning diffraction charts obtained from plate-shaped samples sampled from a polycrystalline silicon rod in which disappearance of the crystal habit line arises when using the polycrystalline silicon rod as a production raw material of an FZ silicon single crystal. A sample A shown in FIG. 5A, a sample B shown in FIG. 5B and a sample C shown in FIG. 5C are respectively sampled from the region A (substantial center (r≈0) of the polycrystalline silicon rod), the region B (r≈R/2) and the region C (r≈R). From each of these diffraction charts, a ratio (S.sub.p/S.sub.t) between an area S.sub.p of a (111) diffraction peak part and a total area S.sub.t of the diffraction chart is calculated for each of the plate-shaped samples.

[0075] As a result, S.sub.p/S.sub.t=11.2% for the sample A, S.sub.p/S.sub.t=1% for the sample B, and S.sub.p/S.sub.t=0% for the sample C. The S.sub.p/S.sub.t ratio increases as the sample is sampled from the region closer to center of the polycrystalline silicon rod, and the S.sub.p/S.sub.t ratio decreases as the sample is sampled from the region closer to the surface thereof.

[0076] FIG. 6 is a diagram obtained by collecting data that were obtained by using 40 polycrystalline silicon rods for relation between the aforementioned FZ yields (%) and the area ratios (S.sub.p/S.sub.t) of the (111) diffraction peak part. Specifically, correlation between the yields of FZ silicon single crystal formation (total length ratios (%) of the length of the region in which disappearance of the crystal habit line was not found) and the aforementioned area ratios (S.sub.p/S.sub.t) with these polycrystalline silicon rods used as production raw materials of FZ silicon single crystals was studied. Statistically processing these results, a correlation coefficient r0 is 0.862. Statistically examining the significance of the correlation coefficient,

r=0.862**>>r0(38,0.01)=0.402

and with the decision condition of 1% of critical rate, it is determined “highly significant correlation is found”. In the case of 5% of critical rate, r0(38,0.05)=0.314. The reference value r0 of the decision is a value determined based on the degree of freedom and the critical rate, and the degree of freedom is the data number-2 which is 40−2=38 in this case.

[0077] From the results shown in this figure, it is clearly understood that the degree of disappearance of the crystal habit line in the case of using these polycrystalline silicon rods as the production raw materials of the FZ silicon single crystals is higher as the aforementioned S.sub.p/S.sub.t ratio is larger (as the peak area ratio from the (111) Miller index plane is higher).

[0078] Notably, when calculating the aforementioned S.sub.p/S.sub.t ratio, intensity calculation (area calculation) of the peak from the (111) Miller index plane is needed. Therefore, first, an average value I.sub.a of all the diffraction intensities appearing in the diffraction chart was calculated, a baseline I.sub.b of the chart was read from the chart, and the value of (I.sub.a−I.sub.b)/I.sub.b×100 was employed.

[0079] The inventors accumulated evaluation results of the degrees of crystal orientation of polycrystalline silicon rods by the aforementioned technique, and advanced the study on the growth conditions, of a polycrystalline silicon rod, under which the aforementioned (111) diffraction peak was caused not to arise.

[0080] In order not to cause the aforementioned (111) diffraction peak to arise, it is important to limit the current flowing in the near-center region of a polycrystalline silicon rod, and meanwhile, to control the temperature of the outer surface to be constant during the CVD process.

[0081] The inventors thought, to this end, that it was important to reduce the amount of heat removed from the surface along with a gas flow, and contrived to set the total amount of feed gas (sum total of the hydrogen gas amount and the trichlorosilane gas amount) to be lower than conventional one without changing the concentration of trichlorosilane in the feed gas.

[0082] With such raw material gas feed, since the amount of fed trichlorosilane decreases, the growth speed decreases by the amount of this decrease. Nevertheless, from repeated experiments by the inventors, it was found that productivity itself of silicon single crystals did not deteriorate by the (111) diffraction peak caused not to arise since the yields of FZ silicon single crystal formation in the case of using such polycrystalline silicon rods as raw materials were dramatically improved.

Examples

[0083] Six polycrystalline silicon rods (A to F) which were caused to grow under different deposition conditions and had a diameter of approximately 140 mm were prepared. For each of these polycrystalline silicon rods, the plate-shaped samples (20.sub.A, 20.sub.B and 20.sub.C) whose thickness was approximately 2 mm were sampled from the three sites shown in FIGS. 1A and 1B to obtain φ-scanning charts of their (111) Miller index planes with the measurement system shown in FIG. 4. Notably, the diameter of the plate-shaped sample 20 was approximately 20 mm.

[0084] The polycrystalline silicon rods A to C were caused to grow under the conditions set such that the feed gas amount was maintained to be still 100 in the first region from the center (r=0) of the polycrystalline silicon rod to r=R/3, reduced by 2% to 5% in the second region from r=R/3 to r=R/2, and furthermore, reduced by 5% to 8% in the third region from r=R/2 to r=R, where the feed gas amount at the start of deposition was set to be 100.

[0085] On the contrary, the polycrystalline silicon rods D to F were caused to grow under the conditions set such that the feed gas amount was maintained to be still 100 in the first region from the center (r=0) of the polycrystalline silicon rod to r=R/3, reduced by 0.2% to 1% in the second region from r=R/3 to r=R/2, and furthermore, reduced by 0.5% to 3% in the third region from r=R/2 to r=R, where the feed gas amount at the start of deposition was set to be 100.

[0086] There are collected in Table 1 a (111) diffraction peak area ratio (%) of the sample sampled from the first region, an average (%) of the (111) diffraction peak area ratios of the samples sampled from the first to third regions, reducing rates (%) of the feed gas amount, the presence or absence of residual stress in polycrystalline silicon, and a yield of FZ silicon single crystal formation in the case of using each of these polycrystalline silicon rods as a raw material (total length ratio (%) of the region length in which disappearance of the crystal habit line was not found), these measured using the aforementioned plate-shaped samples.

TABLE-US-00001 TABLE 1 Examples Comparative Examples Polycrystalline Silicon Rod A B C D E F Diameter (mm) 140 142 140 141 142 140 (111) Peak r = 0 to R/3 0 0 1.9 2.8 5.9 7.9 Area Ratio Average for 0 0 0.5 0.9 1.4 2.7 (%) Three Regions Feed Gas r = R/3 to R/2 5 3 2 1 0.5 0.2 Reducing r = R/2 to R 8 6 5 3 1 0.5 Rate (%) Residual Compressive Present Present Present Present Present Present Stress Tensile Absent Absent Absent Present Present Present FZ Yield (%) 100 100 100 68 22 12

[0087] As presented in Table 1, as compared with the polycrystalline silicon rods (D to F), the polycrystalline silicon rods (A to C) with which disappearance of the crystal habit line was not found have smaller (111) diffraction peak area ratios. Moreover, while in the polycrystalline silicon rods (A to C), tensile stress is not found although compressive stress is found, in the polycrystalline silicon rods (D to F), both compressive stress and tensile stress are found.

[0088] According to these results, it is preferable to cause a polycrystalline silicon rod to grow under the conditions set such that the feed gas amount is reduced by 2% or more in the second region from r=R/3 to r=R/2, and furthermore, reduced by 5% or more in the third region from r=R/2 to r=R in the deposition step of the polycrystalline silicon rod, where the feed gas amount at the start of deposition is set to be 100. For example, a polycrystalline silicon rod is caused to grow under the conditions set such that the feed gas amount is reduced by 2% to 5% in the second region from r=R/3 to r=R/2, and furthermore, reduced by 5% to 8% in the third region from r=R/2 to r=R, where the feed gas amount at the start of deposition is set to be 100.

INDUSTRIAL APPLICABILITY

[0089] The present invention proposes a novel technique to evaluate crystal orientation of a polycrystalline silicon rod having a diameter of 130 mm or more. Further, a large-diameter polycrystalline silicon rod suitable for a single crystal silicon production raw material is selected with this technique, which thereby contributes stable production of single crystal silicon.

REFERENCE SIGNS LIST

[0090] 1 Silicon core wire [0091] 10 Polycrystalline silicon rod [0092] 11 Rod [0093] 20 Plate-shaped sample [0094] 30 Slit [0095] 40 X-ray beam