Patent Grant
9546436
The present invention relates to polycrystalline silicon and a method of casting the same, and in particular relates to polycrystalline silicon and a method of casting the same, suitable for increasing the conversion efficiency of solar cells.
Currently, as substrates for solar cells, silicon crystal is chiefly used. Such silicon crystal is divided into single crystals and polycrystals, and solar cells using single crystal silicon as a substrate have higher conversion efficiency for converting energy of incident light into electrical energy than ones using polycrystalline silicon as a substrate.
Single crystal silicon is produced as a dislocation-free high quality crystal for example by Czochralski process. However, when single crystal silicon is produced by Czochralski process, the production cost is high and such single crystal silicon is not practical being used as a substrate for solar cells. It is therefore common to produce solar cells using polycrystalline silicon which can be cast from inexpensive materials.
One of the methods of casting such polycrystalline silicon is electromagnetic casting (for example, see JP 2007-051026 A (PTL 1)). Electromagnetic casting is a method in which a silicon raw material in a crucible is heated and melted by high frequency induction, meanwhile molten silicon is suspended by the action of strong electromagnetic force, thereby continuously growing an ingot. On this occasion, the molten silicon is not contact with the crucible, so that a high quality ingot can be cast. A polycrystalline silicon wafer obtained from a thus cast polycrystalline silicon ingot is characterized by enabling the production of solar cells that have high uniformity in quality in the casting direction, less variation in the conversion efficiency, and stable characteristics.
In order to increase the conversion efficiency of solar cells, it is important to appropriately control impurities such as oxygen, carbon, and nitrogen that are contained in a polycrystalline silicon wafer. Among others, carbon promotes the precipitation of oxygen and the precipitated oxygen serves as dislocation multiplication sources. The formed dislocations serve as recombination centers of carriers, which results in reduced conversion efficiency of solar cells; thus, it has been considered that the lower the carbon content is in the wafer, the better. For example, WO 2006/104107 (PTL 2) describes a technique of controlling the carbon concentration to 1×1017 atoms/cm3 or less (the interstitial oxygen concentration is 2×1017 atoms/cm3 or more), thereby realizing high conversion efficiency of solar cells.
PTL 1: JP 2007-051026 A
PTL 2: WO 2006/104107
However, when polycrystalline silicon having an extremely low carbon concentration satisfying the requirement for the carbon concentration prescribed in PTL 2 was actually cast to fabricate solar cells, the conversion efficiency was found to be reduced instead.
An object of the present invention therefore is to provide polycrystalline silicon and a method of casting the same that are suitable for increasing the conversion efficiency of solar cells.
The inventors of the present invention made various studies on ways to solve the above problems. After the inventors actually cast polycrystalline silicon having an extremely low carbon concentration to fabricate solar cells and measured the conversion efficiency to find that the conversion efficiency was reduced instead as described above, they cast polycrystalline silicon of different carbon concentrations and examined the conversion efficiency of the solar cells. As a result, they found that instead of excessively reducing the carbon concentration, controlling the concentrations of oxygen and nitrogen in polycrystalline silicon and adjusting them by adding an appropriate amount of carbon to achieve a predetermined concentration range can improve the conversion efficiency. Thus, they have completed the present invention.
Specifically, the present invention primarily includes the following features.
placing a bottomless cooling crucible divided at least partially in the axis direction into a plurality of parts in the peripheral direction and having an inner surface coated with a release agent containing nitrogen, in an induction coil of a chamber charged with an inert gas;
melting a raw material of polycrystalline silicon introduced into the bottomless cooling crucible by electromagnetic induction heating using the induction coil; and
pulling out the molten silicon downward while sequentially cooling and solidifying it,
wherein the carbon concentration of the molten silicon is adjusted to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less, the oxygen concentration thereof to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less, and the nitrogen concentration to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less.
According to the present invention, the crystallinity can be improved by controlling the concentrations of carbon, oxygen, and nitrogen to predetermined concentration ranges, and polycrystalline silicon suitable for increasing the conversion efficiency of solar cells can be cast.
Embodiments of the present invention will be described below with reference to the drawings.
It is important that the carbon concentration of polycrystalline silicon of the present invention is adjusted to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less and the oxygen concentration thereof to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less, and the nitrogen concentration thereof to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less.
As described above, it has conventionally been considered that as the carbon concentration is lower, the precipitation of oxygen precipitate serving as dislocation multiplication sources is suppressed, thereby improving the conversion efficiency of solar cells. However, the inventors cast polycrystalline silicon having an extremely low carbon concentration adjusted to within the carbon concentration range prescribed in PTL 2 to fabricate solar cells, to find that the conversion efficiency was reduced instead. In view of the above, departing from the conventional concepts, the inventors cast polycrystalline silicon of different carbon concentrations to fabricate polycrystalline silicon wafers; fabricated solar cells using the wafers; and measured the conversion efficiency of the solar cells. As a result, it was revealed that the conversion efficiency was improved more by adding an appropriate amount of carbon to control the concentration to within a predetermined concentration range than by excessively reducing the concentration of carbon in polycrystalline silicon.
Although the cause of this is not perfectly clear, it is assumed as follows. That is, as described above, carbon promotes the precipitation of oxygen and the oxygen precipitate formed becomes a dislocation multiplication source. The dislocations formed become recombination centers to deteriorate the conversion efficiency of solar cells. On the other hand, as described in K. Sumino “Impurity Reaction with Dislocations in Semiconductors”, Phys. Stat. Sol. (a)171, 111 (1999), the an oxygen precipitate in silicon has the effect of suppressing the propagation of dislocations. Thus, the crystallinity of polycrystalline silicon is improved and the conversion efficiency is increased. Specifically, it is conceivable that although the addition of an appropriate amount of carbon and nitrogen promoted the precipitation of oxygen to increase the formation of the locations, the propagation of the dislocations formed was suppressed by the oxygen precipitate, resulting in improved crystallinity than the case of suppressing the formation of oxygen precipitate by excessively reducing the concentration of carbon and nitrogen contained in a polycrystalline silicon wafer, which consequently improved the conversion efficiency. The requirements for the concentrations of carbon, oxygen, and nitrogen to be satisfied with respect to polycrystalline silicon of the present invention now will be described.
First, the carbon concentration of polycrystalline silicon is set to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less. Here, the concentration is set to 4.0×1017 atoms/cm3 or more, since a carbon concentration of less than 4.0×1017 atoms/cm3 is too low to propagate dislocations and accordingly prevents the crystallinity from being improved. Further, the concentration is set to 6.0×1017 atoms/cm3 or less because the addition of carbon at more than 6.0×1017 atoms/cm3 causes the precipitation of foreign matter including SiC in a polycrystalline silicon ingot to increase crystal defects, which leads to a reduced conversion efficiency.
Further, the oxygen concentration is set to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less. Here, the concentration is set to 0.3×1017 atoms/cm3 or more, since a concentration of less than 0.3×1017 atoms/cm3 is too low to cause oxygen precipitation. Further, the concentration is set to 5.0×1017 atoms/cm3 or less because a concentration exceeding 5.0×1017 atoms/cm3 results in the formation of a complex of oxygen and boron.
Further, the concentration of nitrogen in the polycrystalline silicon is set to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less. Here, the concentration is set to 8.0×1013 atoms/cm3 or more, since a nitrogen concentration of less than 8.0×1013 atoms/cm3 is too low to propagate dislocations and accordingly prevents the crystallinity from being improved. Further, the concentration is set to 1.0×1018 atoms/cm3 or less because a concentration exceeding 1.0×1018 atoms/cm3 results in the precipitation of foreign matter of silicon nitride (Si3N4) to reduce the conversion efficiency.
As such, adjustments are made to achieve the predetermined concentration ranges by adding an appropriate amount of carbon, oxygen, and nitrogen, thereby obtaining polycrystalline silicon having improved crystallinity, which is suitable for solar cells having a high conversion efficiency of 16% or more.
Next, a method of casting polycrystalline silicon according to the present invention will be described.
In the center of the chamber 1, a bottomless cooling crucible 7, an induction coil 8, after heaters 9, and soaking tubes 10 are disposed.
The bottomless cooling crucible 7 is a square cylindrical body made of a metal material such as copper, which is excellent in thermal conductivity and electrical conductivity, serves not only as a melting vessel for melting a silicon raw material 12 introduced but also as a mold, and is suspended in the chamber 1. This bottomless cooling crucible 7 is configured, except its upper part, to be divided in the circumferential direction into a plurality of pieces of short strips such that it is subjected to forced cooling with cooling water flowing inside.
The induction coil 8 is provided concentrically around the bottomless cooling crucible 7 to surround the bottomless cooling crucible 7 and is connected to a power supply device (not shown).
A plurality of after heaters 9 are placed below the bottomless cooling crucible 7 to be concentric to the crucible 7, and include a heater (not shown) or a heat insulator (not shown). The after heaters 9 heat an ingot 3 lowered from the bottomless cooling crucible 7 to give a predetermined temperature gradient in the lowering direction of the ingot 3, thereby preventing crystal defects to be formed in the ingot 3.
In order to prevent cracks from being formed in the ingot 3 due to the residual stress caused by the cooling, the ingot 3 is soaked in the soaking tubes 10 by being maintained at a predetermined temperature for a predetermined time.
A raw material inlet pipe 11 is attached to the upper wall of the chamber 1, and the bottomless cooling crucible 7 is charged with the silicon raw material 12 in particulate or aggregate form through the raw material inlet tube 11 from a raw material supplying unit (not shown) via a shutter 2.
Further, above the bottomless cooling crucible 7, a plasma torch 14 for melting the silicon raw material 12 is liftably provided. One of the poles of a plasma power supply device (not shown) is connected to the plasma torch 14, whereas the other pole is connected to the ingot 3 side. The plasma torch 14 is lowered to be introduced into the bottomless cooling crucible 7.
A side wall of the chamber 1 is provided with a gas inlet 5 for introducing into the chamber 1, an inert gas; or a carbon monoxide gas, an oxygen gas, or a carbon dioxide gas for controlling the oxygen or carbon concentrations of the ingot 3. The gas introduced into the chamber 1 circulates inside the chamber 1 through a circulation pipe 17. The gas introduced into the chamber 1 is exhausted through an exhaust port 6 provided at a lower part of a side wall of the chamber 1.
Further, the bottom wall of the chamber 1 is provided with an outlet 4, configured such that the ingot 3 placed on a support base 16, which has been subjected to a soaking process using the soaking tubes 10 is drawn out from the outlet 4.
Here, a release agent is applied to the top surface of the support base 16 (that is, the surface in contact with the ingot 3) for the purpose of preventing the ingot 3 from being fused to the support base 16. The slurry-formed release agent can be obtained by mixing release agent powder such as silicon nitride, silicon carbide, or silicon oxide into a solution composed of a binder and a solvent. After applying the release agent to the top surface of the support base 16 with a brush or spray, a debinding treatment for removing the solvent and the binder is performed, thereby forming a release agent layer on the support base 16. This debinding treatment is specifically a treatment for removing the binder and the solvent in the release agent by heat treating the support base 16 in the atmosphere, for example, at a temperature of 120° C. for one hour.
In the present invention, the release agent is used not only for preventing the fusion between the ingot 3 and the support base 16 mounted on a support shaft 15 as described above but also for controlling the nitrogen concentration of the ingot 3. Therefore, a release agent containing nitrogen is used. The release agent layer obtained from such a release agent is brought into contact with molten silicon 13 at an initial stage of casting; accordingly, when nitrogen is contained in the release agent layer, nitrogen diffuses into the molten silicon 13 to be added to the ingot 3 (that is, polycrystalline silicon).
Here, as the release agent powder containing nitrogen, silicon nitride can be used. As the binder, polyvinyl alcohol (PVA), ethyl silicate, or the like can be used, whereas as the solvent, purified water or alcohol can be used. Further, when ethyl silicate is used as the binder, hydrochloric acid can be used as an additive in order to promote the hydrolysis. They are mixed to form a slurry to be applied to the top surface of the support base 16, thereby performing the debinding treatment; thus, the release agent layer can be formed on the top surface of the support base 16.
The nitrogen concentration of the ingot 3 can be controlled by adjusting the area to be coated with the above release agent. For example, when 100 g of silicon nitride powder, 100 ml of ethyl silicate as a binder, 400 ml of purified water as a solvent, and 0.5 ml of hydrochloric acid as an additive are mixed to form a slurry and the area of the top surface of the support base 16 to be coated with the resultant release agent is changed by 200 cm2, the nitrogen concentration of the ingot 3 can be changed by 1.6×1017 atoms/cm3. The nitrogen concentration of the ingot 3 in the case where the release agent is not applied varies between each ingot 3 depending on the purity of the silicon raw material 12 and the like; the nitrogen concentration is adjusted to the range prescribed above by controlling the area to be coated with the release agent in accordance with the nitrogen concentration in the case where the release agent is not applied. For example, when the nitrogen concentration of the ingot 3 in the case where the release agent is not applied is 7.0×1013 atoms/cm3 and the nitrogen concentration of the release agent is 17 mass %, the nitrogen concentration can be adjusted to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less with an application area of 1 cm2 or more to 1300 cm2 or less.
The various studies having been conducted by the inventors revealed that although the nitrogen concentration of a release agent is changed, the amount of nitrogen diffusing into the ingot 3 does not change and accordingly the nitrogen concentration changes depending on the application area of the release agent.
Further, the nitrogen concentration of the above-described ingot 3 can be measured, for example, by secondary ion mass spectrometry (SIMS).
Subsequently, the shutter 2 is opened to introduce the silicon raw material 12 of polycrystalline silicon or the like into the crucible 7 from the raw material supplying unit (not shown) through the raw material inlet tube 11.
Next, the lowered plasma torch 14 is energized while applying alternating current to the induction coil 8. Here, the pieces of the short strips forming the bottomless cooling crucible 7 are electrically isolated, which results in the generation of eddy current in each piece with the electromagnetic induction caused by the induction coil 8; thus, the eddy current on the inner wall side of the bottomless cooling crucible 7 generates a magnetic field in the crucible. Consequently, the silicon raw material 12 in the bottomless cooling crucible 7 is melted by electromagnetic induction heating to obtain the molten silicon 13.
Here, the plasma torch 14 is used, so that plasma arc is generated between the plasma torch 14, and the silicon raw material 12 and the molten silicon 13, and the silicon raw material 12 is also heated by the resultant Joule heat to be melted. Therefore, the burden on the electromagnetic induction heating is reduced, so that the molten silicon 13 can be obtained efficiently.
Here, the molten silicon 13 in the bottomless cooling crucible 7 is subjected to the interaction between the magnetic field generated with the eddy current on the inner walls of the crucible 7, and the current generated on the surface of the molten silicon 13 causes a force in a direction from the bottomless cooling crucible 7 to the molten silicon (pinch force); thus, the molten silicon is kept without being in contact with the bottomless cooling crucible 7. Thus, contamination can be prevented by controlling mixing of impurities from the bottomless cooling crucible 7 into the molten silicon 13, and it is made easier to lower the ingot 3 mounted on the support shaft 15.
Normally, when the above silicon raw material 12 is melted, an inert gas such as argon is introduced from the gas inlet 5 into the chamber 1; however, in the present invention, in order to control the oxygen and carbon concentrations of the ingot 3, a carbon monoxide gas and an oxygen gas are introduced with the inert gas. As described above, in order to obtain single crystal silicon suitable for solar cells having high conversion efficiency, in addition to the control of the nitrogen concentration using the above described release agent, it is required to control the oxygen concentration to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less as well as to add carbon at an appropriate amount to control the carbon concentration to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less.
The inventors made various studies on the oxygen and carbon concentrations of the ingot 3 to find that the partial pressures of the oxygen gas and the carbon monoxide gas with respect to the inert gas have effects on the oxygen concentration and the carbon concentration of the ingot 3, respectively. Further, it was found that in order to increase the carbon concentration of the ingot 3 by 1.0×1017 atoms/cm3, for example, when argon gas is used as an inert gas at a flow rate of 200 l/min and an internal reactor pressure of 0.03 kg/cm2, a carbon monoxide gas having a partial pressure 1.2×10−4 times that of the argon gas may be supplied. It was found that also when controlling the oxygen concentration, in order to control the oxygen concentration by 1.0×1017 atoms/cm3, an oxygen gas having a partial pressure 4.6×10−5 times that of the argon gas may be supplied.
Note that as with the nitrogen concentration, the carbon concentration and the oxygen concentration of the ingot 3 where a carbon monoxide gas and an oxygen gas are not mixed into the inert gas, vary depending on each ingot 3. Accordingly, in accordance with the carbon concentration and the oxygen concentration of the ingot 3 where a carbon monoxide gas and an oxygen gas are not mixed into the inert gas, the partial pressures of the carbon monoxide gas and the oxygen gas with respect to the inert gas are controlled to be adjusted to the ranges of the carbon concentration and the oxygen concentration described above.
For example, when the carbon concentration of the ingot 3 in the case where a carbon monoxide gas and an oxygen gas are not mixed into the inert gas is 2.0×1017 atoms/cm3, whereas the relevant oxygen concentration is 0.1×1017 atoms/cm3, the oxygen concentration of the ingot 3 can be adjusted to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less, and the carbon concentration can be adjusted to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less by controlling the partial pressure of the carbon monoxide gas to 2.5×10−4 times or more to 5.0×10−4 times or less with respect to the inert gas and controlling the partial pressure of the oxygen gas to 1.2×10−5 times or more to 1.9×10−4 times or less.
Note that the carbon concentration and the oxygen concentration of the ingot 3 can be measured, for example, by Fourier transform infrared spectroscopy (TF-IR), and the oxygen concentration is a concentration defined based on ASTM F121-1979, whereas the carbon concentration is a concentration defined based on ASTM F123-1981 (ASTM: American Society for Testing and Materials)).
Instead of mixing of the carbon monoxide gas as described above, carbon powder can be added to the silicon raw material 12 to control the carbon concentration. For example, when the carbon concentration of the ingot 3 in the case where carbon powder is not added to the ingot 3 is 2.0×1017 atoms/cm3, carbon powder of 1.7×10−3 times or more to 3.4×10−3 times or less with respect to the total weight of the raw material subtracting the carbon powder weight is added, thereby satisfying the requirements relating to the above carbon concentration. Here, graphite powder is preferably used as the carbon powder, since it easily melts in molten silicon.
Further, instead of controlling the carbon and oxygen concentrations by the mixing of an oxygen gas and a carbon monoxide gas as described above, the oxygen and carbon concentrations can be simultaneously controlled by mixing carbon dioxide. Also in this case, the partial pressure of the carbon dioxide gas with respect to the inert gas has effects on the oxygen concentration and the carbon concentration of the ingot 3. For example, when the carbon concentration of the ingot 3 when a carbon dioxide is not mixed into the inert gas is 2.0×1017 atoms/cm3 and the oxygen concentration is 0.1×1017 atoms/cm3, the partial pressure of the carbon dioxide can be adjusted to 2.5×10−4 times or more to 5.0×10−4 times or less with respect to the inert gas, thereby adjusting the oxygen concentration to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less and the carbon concentration to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less.
As described above, for the molten silicon 13 having the adjusted carbon, oxygen, and nitrogen concentrations, the shaft 15 supporting the molten silicon 13 is gradually lowered to solidify the molten silicon 13. Specifically, the molten silicon 13 inside the crucible 7 is moved away from the bottom end of the induction coil 8 by lowering the support shaft 15, so that the induced magnetic field is reduced and the heating value and the pinch force are reduced; accordingly, the molten silicon 13 is cooled by the bottomless cooling crucible 7 to be solidified from the periphery.
Thus, while the above support shaft 15 is lowered, the silicon raw material 12 is continuously and additionally introduced to the bottomless cooling crucible 7 through the raw material inlet tube 11, thereby successively melting and solidifying the silicon raw material 12, which allows polycrystalline silicon to be continuously cast.
The ingot 3 obtained through the solidification of the molten silicon 13 is cooled over long hours to room temperature using the after heaters 9 and the soaking tubes 10. The ingot 3 is heated using the after heaters 9 to impart an appropriate temperature gradient in the lowering direction, thereby preventing the formation of crystal defects in the ingot 3 while cooling.
Further, in order to prevent the formation of cracks in the ingot 3 due to the residual stress caused by the cooling, the ingot 3 is subjected to soaking by being maintained at a predetermined temperature for a predetermined time using a heater (not shown) of the soaking tubes 10. The maintained temperature of the ingot 3 in soaking is, in general, preferably about 1100° C., since at a temperature exceeding 1200° C., the speed of dislocation growth in crystals is high and crystal defects are easily formed.
After soaking, the output of the heater (not shown) of the soaking tubes 10 is lowered to cool the ingot 3 to room temperature. Thus, the polycrystalline silicon ingot 3 can be cast through adjusting the oxygen concentration to 0.3×1017 atoms/cm3 or more to 5.0×1017 atoms/cm3 or less, the carbon concentration to 4.0×1017 atoms/cm3 or more to 6.0×1017 atoms/cm3 or less, and the nitrogen concentration to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less.
Note that in the above method, the nitrogen concentration of the ingot 3 is controlled by controlling the area of the release agent containing nitrogen applied to the top surface of the support base 16; alternatively, a nitrogen gas can be mixed into the inert gas as in the case of controlling the carbon concentration and the oxygen concentration. Also in this case, the partial pressure of the nitrogen gas with respect to the inert gas has an effect on the nitrogen concentration of the ingot 3. For example, when the nitrogen concentration of the ingot 3 in the case where a nitrogen gas is not mixed into the inert gas is 1.0×1013 atoms/cm3, the partial pressure of the nitrogen gas is controlled to 7.2×10−7 times or more to 1.0×10−2 times or less with respect to the inert gas, thereby adjusting the nitrogen concentration to 8.0×1013 atoms/cm3 or more to 1.0×1018 atoms/cm3 or less.
Here, in the case where the nitrogen concentration is adjusted by mixing of a nitrogen gas, a release agent free of nitrogen is used.
Further, the conductivity of the polycrystalline silicon can be controlled by introducing a silicon raw material 12 doped with a dopant. Specifically, when casting p-type polycrystalline silicon, boron, gallium, aluminum, or the like can be used as the dopant. On the other hand, when casting n-type polycrystalline silicon, phosphorus, arsenic, antimony, or the like can be used as the dopant.
Thus, with the oxygen and nitrogen concentrations being adjusted to appropriate ranges, the carbon concentration can be adjusted to an appropriate range by incorporating carbon in the crystal at an appropriate content, thereby improving the crystallinity. Accordingly, polycrystalline silicon suitable for solar cells having high conversion efficiency can be cast.
A polycrystalline silicon ingot obtained by the method of casting polycrystalline silicon according to the present invention is divided into polycrystalline blocks and the polycrystalline blocks are then sliced, thereby producing polycrystalline silicon wafers suitable for solar cells having high conversion efficiency.
Examples of the present invention will now be described.
Polycrystalline silicon ingots were cast in accordance with a method of the present invention. Specifically, first, a release agent obtained by mixing 100 g of silicon nitride as release agent powder, 100 ml of ethyl silicate as a binder, 400 ml of purified water as a solvent, and 0.5 ml of hydrochloric acid as an additive was prepared and applied to the top surface of the support base 16. The nitrogen concentration of the ingots 3 was adjusted by controlling the area of the top surface of the support base 16 coated with the release agent to 1 cm2, 50 cm2, and 1300 cm2, such that the nitrogen concentration is adjusted to 8×1013 atoms/cm3, 1×1015 atoms/cm3, and 1×1018 atoms/cm3, respectively. After that, for debinding, the support base 16 was heat treated and maintained at 120° C. for one hour, and then cooled to room temperature, thus forming a release agent layer on the top surface of the support base 16.
Next, as shown in
Subsequently, the silicon raw material 12 inside the crucible 7 was heated to 1420° C. and melted by an induction coil 8 and a plasma torch 14, thus obtaining molten silicon 13. Further, the support base 16 was lowered while adding the silicon raw material 12 and melting the silicon raw material 12, thus casting each polycrystalline silicon ingot 3 having a length of 7000 mm.
At that time, an oxygen gas and a carbon monoxide gas were supplied with an argon gas from a gas inlet 5 into a chamber 1. The partial pressure of the carbon monoxide gas was 2.5×10−4 times, 3.7×10−4 times, and 5.0×10−4 times with respect to the argon gas, such that the carbon concentration of the ingot 3 was 4.0×1017 atoms/cm3, 5.0×1017 atoms/cm3, and 6.0×1017 atoms/cm3. Note that the carbon concentration of the polycrystal was 2.0×1017 atoms/cm3 on the assumption that the carbon monoxide gas was not added. Further, the partial pressure of the oxygen gas was 1.2×10−5 times, 6.2×10−5 times, 1.9×10−4 times with respect to the argon gas such that the oxygen concentration was 3.0×1016 atoms/cm3, 2.0×1017 atoms/cm3, and 5.0×1017 atoms/cm3, respectively. Note that the oxygen concentration of the polycrystal was 1.0×1016 atoms/cm3 on the assumption that the oxygen gas was not added.
Thus obtained ingots each having a size of 345 mm×510 mm were each divided into six, and 1000 wafers having a square shape with a side length of 156 mm and a thickness of 180 μm were cut out from a position at the center in the longitudinal direction (2000 mm from the end of the ingot) of each polycrystalline silicon ingot of Invention Examples. Ten wafers of them were subjected to simulated acid texturing (HF:HNO3:H2=1:4:5, room temperature, five minutes), and a nitride film was then formed on each wafer by vapor phase epitaxy (CVD).
Polycrystalline silicon was cast as in the invention examples. However, the partial pressure of the carbon monoxide was 1.9×10−4 times and 8.7×10−4 times with respect to the argon gas such that the carbon concentration of the ingots 3 was 3×1017 atoms/cm3 and 7×1017 atoms/cm3, respectively. Note that the carbon concentration of the polycrystal was 2.0×1017 atoms/cm3 on the assumption that the carbon monoxide gas was not added. Further, the partial pressure of oxygen was 2.5×10−4 times with respect to the argon gas such that the oxygen concentration was 6.0×1017 atoms/cm3. Note that the oxygen concentration of the polycrystal was 1.0×1016 atoms/cm3 on the assumption that the oxygen gas was not added. Moreover, the area of the release agent applied to the top surface of the support base 16 was 0 cm2 and 1700 cm2 such that the nitrogen concentration was 7.0×1013 atoms/cm3 and 2.0×1018 atoms/cm3. All the other conditions are the same as those in Invention Example 1.
<Conversion Efficiency>
Ten of the fabricated 1000 wafers were used to fabricate solar cells for evaluation and the conversion efficiency was measured. The relationship between the carbon concentration of the ingots 3 and the conversion efficiency is shown in
In the prevent invention, with the oxygen and nitrogen concentrations being adjusted to appropriate ranges, the carbon concentration can be adjusted to an appropriate range by incorporating carbon in the crystal at an appropriate content, thereby improving the crystallinity. Accordingly, thus obtained polycrystalline silicon is suitable for solar cells having high conversion efficiency.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2012-069469 | Mar 2012 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2013/001261 | 3/1/2013 | WO | 00 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2013/145558 | 10/3/2013 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 4915723 | Kaneko | Apr 1990 | A |
| 5211802 | Kaneko | May 1993 | A |
| 20030150374 | Sasatani | Aug 2003 | A1 |
| 20030205358 | Kimura | Nov 2003 | A1 |
| 20070169684 | Stoddard | Jul 2007 | A1 |
| 20080210156 | Sasatani et al. | Sep 2008 | A1 |
| 20090044926 | Kida | Feb 2009 | A1 |
| 20130247618 | Yoshihara | Sep 2013 | A1 |
| Number | Date | Country |
|---|---|---|
| 101370969 | Feb 2009 | CN |
| 101796226 | Aug 2010 | CN |
| 102312290 | Jan 2012 | CN |
| 2-51493 | Feb 1990 | JP |
| 2007-51026 | Mar 2007 | JP |
| 2009-523694 | Jun 2009 | JP |
| 2010-87327 | Apr 2010 | JP |
| 2012036056 | Feb 2012 | JP |
| 2006104107 | Oct 2006 | WO |
| Entry |
|---|
| Chinese Office Action issued with respect to application No. 201380017179.5, mail date is Mar. 25, 2016. |
| International search report issued with respect to application No. PCT/JP2013/001261, mail date is Apr. 2, 2013. |
| Taiwanese Office Action issued with respect to application No. 10320668100, mail date is May 29, 2014. |
| Number | Date | Country | |
|---|---|---|---|
| 20150082833 A1 | Mar 2015 | US |
