JP2004006787A - Liquid phase growing method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide, at a lower cost, a large amount of a polycrystal which has a characteristic equivalent to a polycrystal which has been used for a conventional solar battery, and has few possibility of the disconnection of a grid electrode or the like even if used in a solar battery, and to provide a manufacturing device preferred for embodying this method. <P>SOLUTION: In a liquid phase growing method, a polycrystalline substrate is dipped in a melt in a crucible in which a crystalline raw material is fused, and the polycrystal is grown on the substrate. This method comprises a first step of growing the polycrystal having a predetermined thickness and a second step of fusing a part of the polycrystal grown in the first step in the melt. A relative position between the substrate and the melt is altered in the first step and the second step, and the melts at different temperatures are brought into contact with the surface of the polycrystal. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a liquid phase growth method for a polycrystalline semiconductor such as silicon that can be suitably used as a semiconductor for a solar cell, and a liquid phase growth apparatus suitable for performing the method.
[0002]
[Prior art]
With increasing awareness of the environment, solar cells have been widely used in consumer applications. For a solar cell for consumer use, a monocrystalline or polycrystalline silicon wafer is mainly used.
[0003]
Conventionally, as a single-crystal silicon wafer, a silicon wafer which has conventionally been out of the standard in the IC industry or the like, or a silicon wafer which has been re-pulled from remaining silicon has been often used, and the supply amount has been limited.
[0004]
On the other hand, polycrystalline silicon wafers have a higher throughput in the crystallization process than single crystals, but when high-purity silicon for semiconductors is used as a raw material, the supply is still limited, and the price is so low. Can not.
[0005]
Therefore, a method of growing a polycrystalline silicon film on a substrate of ceramics such as glassy carbon or mullite has been attempted. However, the growth must be usually performed at a high temperature of 1000 to 1500 ° C. From the viewpoint of matching thermal expansion coefficients, it is difficult to use inexpensive materials such as metal plates and glass. Furthermore, a polycrystalline silicon film grown on such a substrate has not been put to practical use because the crystal grains are small and the surface flatness tends to be poor.
[0006]
One of the attempts to solve such a problem is to use a polycrystalline silicon (metal-grade silicon) manufactured using inexpensive unpurified silicon as a substrate, and grow a polycrystalline silicon on the substrate using high-purity silicon. There is a method of forming a solar cell using crystalline silicon as an active layer.
[0007]
For example, in Non-Patent Document 1, SiH is placed on a metal-grade silicon substrate. ₂ Cl ₂ Are used to grow silicon polycrystals by the CVD method and prototype solar cells. In this case, the substrate is silicon, and there is no problem such as heat resistance and thermal expansion mismatch. In addition, since the raw materials are inexpensive, there is no restriction on resources, and there is a possibility that they can be provided at low cost. Furthermore, since the grown polycrystalline silicon film inherits the crystallinity of the substrate, it is easier to grow a high-quality polycrystal than when glassy carbon, ceramics, or the like is used as the substrate.
[0008]
However, when the growth is performed by the above-described CVD method, the source gas used cannot be used 100% effectively for growth, and when a silicon having a thickness of several tens μm required for a solar cell is grown. There are many manufacturing problems, such as the need to frequently maintain the growth apparatus.
[0009]
For the growth of polycrystalline silicon, a liquid phase growth method can be used. In the liquid phase growth method, a raw material such as silicon is dissolved in a low melting point metal such as indium, gallium, tin, copper, and aluminum to adjust a melt, and then supersaturated by a method such as cooling, and a substrate immersed in the substrate Grow crystals on top. Since silicon does not flow out of the melt, the input materials are not wasted, and the maintenance of the apparatus is easy.
[0010]
Patent Literature 1 discloses a method of growing high-quality polycrystalline silicon on an inexpensive metal-grade silicon polycrystalline substrate by the liquid phase growth method and forming a solar cell using the polycrystalline silicon. ing. According to this method, since a polycrystalline substrate of metal-grade silicon can be filled in a mold, melted, and directly formed, there is no need to take time for ingot formation or slicing, which is further advantageous in terms of cost. However, of course, a substrate obtained by forming an ingot using inexpensive metal-grade silicon and slicing the ingot can also be suitably used.
[0011]
However, as shown in FIG. 6, when a solar cell is manufactured using not only metal-grade silicon but also a polycrystalline growth layer 202 grown on a polycrystalline substrate 200 by a liquid phase growth method, the surface thereof may be In many cases, disconnection occurs in the formed grid electrode 206. The grid electrode 206 plays a role of collecting the carriers generated by the incident light. If the grid electrode 206 is broken, the conversion efficiency of the solar cell is significantly impaired. Further, if the grid electrode 206 is made thicker, disconnection is less likely to occur, but the loss of incident light due to the shadow of the grid electrode 206 increases, which is also undesirable.
[0012]
The cause of such a problem is considered as follows. That is, the grain boundary 201 exists in the polycrystalline substrate 200, and the crystal orientation is different inside the crystal grains on both sides of the grain boundary. When the polycrystalline growth layer 202 is grown on the surface, microscopically, the crystal grows epitaxially in each region of the crystal grains. Therefore, the polycrystalline growth layer 202 also has a grain boundary at the same position as the grain boundary 201 of the substrate. 203 is formed. Although the groove 205 generally occurs on the grain boundary 203, as described in Non-Patent Document 2, the polycrystalline grown by the liquid phase method tends to have a deeper groove than the polycrystalline grown by the CVD method. . Therefore, although the polycrystal grown by the liquid phase method has excellent characteristics, the solar cell using the polycrystal is liable to cause disconnection of the grid electrode 206.
[0013]
On the other hand, if the thickness of the polycrystal serving as the active layer of the solar cell increases, the absorption of incident light increases and the output current increases, which is preferable in terms of characteristics. However, the groove is also deepened at the same time, and disconnection of the grid electrode 206 is likely to occur.
[0014]
Therefore, the polycrystalline layer shown in FIG. 6 has an average thickness W as large as possible, while an average depth G2 of the groove is as small as possible, in other words, G2 / W is as small as possible. It is an indication of suitability for battery applications. In fact, as shown in FIG. 2, which will be described later in the description of the present invention, when G1 / W is small, disconnection of the grid electrode can be prevented, output current increases, and high conversion efficiency can be obtained.
[0015]
Non-Patent Document 3 discloses a method of reducing the depth of a groove generated at a grain boundary when polycrystalline silicon is grown in a liquid phase on a polycrystalline silicon substrate. In this method, as shown in FIG. 7, after the substrate is immersed in the melt (t = t0), the melt is monotonously decreased from the temperature Th to Tl until the substrate is pulled up from the melt (t = tf). Instead, the groove can be made shallower by providing a period in which the melt is heated by ΔTh in the middle. They argue that growth is particularly close to thermal equilibrium because grain boundaries are structurally unstable and growing on top of them requires more energy to grow inside the grains. In the liquid phase growth method, the growth rate is relatively lowered to generate grooves. However, when heated, silicon melts out preferentially (melt back) in the region inside the crystal grains and outside the grain boundaries. , The groove is relatively shallow.
[0016]
However, a large-sized liquid phase growth apparatus needs to handle a large amount of melt such as several tens kg to several hundred kg. At this time, if the cooling and heating of the melt are repeated as shown in FIG. 7, the time required for growth is increased, the production throughput is reduced, and unnecessary energy is required for production, which is not desirable.
[Patent Document 1]
JP-A-10-098205
[Non-patent document 1]
Haruo ITO, Tadashi SAITOH, Noboru NAKAMURA, Sunao Matsubara, Terunori WARABISAKO, Takashi Tokuyama, J. Crys. Growth 45 (1978) 446-453
[Non-patent document 2]
K. J. Weber, A .; W. Blakers, J.M. Crys. Growth 154 (1995) 54-59
[Non-Patent Document 3]
G. FIG. Ballhorn, K .; J. Weber, S.M. Armand, M .; J. Stokes, A. W. Blakers, Solar Energy Materials & Solar Cells 52 (1998) 61-68.
[0017]
[Problems to be solved by the invention]
The present invention has been made in view of the above problems, and has an object to have properties comparable to polycrystals used in conventional solar cells, and to provide a grid electrode even when used in solar cells. An object of the present invention is to provide a liquid layer growth method capable of obtaining a large amount of polycrystals at a low cost with a low risk of causing troubles such as disconnection, and a liquid phase growth apparatus suitable for performing the method.
[0018]
[Means for Solving the Problems]
In order to achieve the above object, the liquid phase growth method of the present invention is a liquid phase growth method in which a polycrystalline substrate is immersed in a melt in a crucible in which a crystal raw material is melted, and a polycrystal is grown on the substrate. hand,
A first step of growing a polycrystal of a predetermined thickness;
A second step of melting a part of the polycrystal grown in the first step in a melt,
The relative position between the substrate and the melt is changed between the first step and the second step, and the melt at different temperatures is brought into contact with the polycrystalline surface.
.
[0019]
In the above liquid phase growth method, it is preferable that the first step and the second step are alternately repeated a plurality of times.
[0020]
In this case, in the second step, it is preferable that the polycrystal grown in the first step is melted in a melt in a region of 5% or more and 30% or less with respect to the predetermined thickness. The thickness of the polycrystal grown in the first step is preferably 1 to 20 μm.
[0021]
Also, in the first step, the substrate is brought into contact with the melt in a relatively high temperature region, and in the second step, the substrate is brought into contact with the melt in a relatively low temperature region,
Further, it is preferable to lower the temperature of each region of the melt.
[0022]
The first step and the second step are preferably performed in the same crucible. In this case, the substrate is reciprocated up and down in the crucible, so that the substrate is relatively free from the melt in a relatively high temperature region. It is good to alternately contact the melt in an area as low as possible.
[0023]
Further, the first step and the second step may be performed in different crucibles.
[0024]
In addition, it is preferable that the content of the crystal raw material per unit volume in the melt in contact with the polycrystalline surface differs between the first step and the second step.
[0025]
On the other hand, the liquid phase growth apparatus of the present invention comprises at least a melt for dissolving the crystal raw material,
A crucible containing the melt,
Heating means capable of forming a temperature distribution in the melt;
A support frame for holding the substrate,
Support frame driving means for changing the relative position between the substrate and the melt between a relatively high temperature region and a relatively low temperature region of the melt,
It is characterized by having.
[0026]
In the above-mentioned liquid phase growth apparatus, it is preferable that the heating means capable of forming the temperature distribution of the melt is a heater which is vertically divided and which can be controlled independently.
[0027]
In this case, it is preferable that the means for changing the relative position between the substrate and the melt is means for moving the substrate up and down.
[0028]
It is also preferable that the heating means capable of forming the temperature distribution of the melt is a heater which is divided in the circumferential direction and can be controlled independently.
[0029]
In this case, the means for changing the relative position between the substrate and the melt is preferably means for rotating the substrate around the central axis, and the means for changing the relative position between the substrate and the melt rotates the crucible. It is preferred to include means.
[0030]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to the embodiments.
[0031]
The liquid phase growth method of the present invention has a second step of melt-backing the polycrystal after the first step of growing the polycrystal. The present invention is characterized in that a polycrystal is intentionally grown on a substrate and melted back by controlling the temperature of the melt and the like, and in this regard, the liquid is simply immersed in the melt and pulled up. Different from the phase growth method. These steps are preferably performed a plurality of times, for example, first → second → first → second →. To achieve high efficiency in a solar cell, a certain thickness or more is necessary to obtain a photocurrent.However, if the growth is continued for a long time, a melt having a high silicon concentration becomes deeper in the groove as the groove becomes deeper. Since it is difficult to supply the gas, the growth is further delayed from the surroundings, and the groove is extremely developed. It is also possible to selectively melt-back portions other than the groove later, but it is preferable to perform the melt-back before the groove is extremely developed and grow it a plurality of times to obtain a necessary film thickness. .
[0032]
The thickness of the polycrystal grown in one single first step is preferably 1 to 20 μm. Below this range, it is difficult to control the film thickness, and the number of steps increases, resulting in poor productivity. Above this range, the irregularities become too large and the flattening effect becomes extremely small.
[0033]
Further, the amount of polycrystal to be melted back in the second step is preferably 5% to 30% of the polycrystal grown in the first step. Below this range, it is almost impossible to control the unevenness, it is difficult to control the film thickness, and the number of steps increases, resulting in poor productivity. If it exceeds this range, triangular or square depressions are likely to be formed on the surface after the growth, and more of the grown polycrystal is wasted, resulting in poor productivity.
[0034]
FIG. 1 shows an example of a liquid phase growth apparatus suitable for carrying out the liquid phase growth method of the present invention, and a preferred embodiment of the present invention will be described with reference to FIG. Hereinafter, the case where polycrystalline silicon is grown on a polycrystalline silicon substrate will be specifically described as an example, but the present invention is similarly applied to the growth of polycrystalline such as germanium or gallium phosphide suitable for manufacturing devices such as solar cells. Can be applied to
[0035]
In FIG. 1, reference numeral 300 denotes a crucible made of quartz glass or carbon for containing the melt 301. The melt 301 is provided with two heaters 302a and 302b, which can be independently controlled, divided into upper and lower portions as heating means capable of forming a temperature distribution of the melt 301 in the vertical direction of the crucible 300. These heaters 302a and 302b control the temperature and the vertical temperature distribution. The crucible 300 is held in an airtight chamber (not shown). If necessary, the inside of the crucible 300 can be made into a hydrogen atmosphere, so that high-quality polycrystal can be grown.
[0036]
First, a low melting point metal such as indium, gallium, tin, copper, or aluminum is charged into the crucible 300 and heated and melted by the heaters 302a and 302b. Silicon as a raw material is dissolved therein and saturated to obtain a melt 301. The raw material silicon may be in a wafer shape, a powder shape, or a nugget shape.
[0037]
After that, the heaters 302a and 302b are controlled to start cooling the melt as shown in FIG. When cooled down by several degrees C. to several tens degrees C. (t = t0, Ta = Tb = Th), the substrate 303 is immersed in the melt 301 and held at the position 303a. At the same time, the heater 302b is controlled independently of the heater 302a, and the temperature (Tb) near the bottom of the crucible 300 is gradually lowered from the temperature (Ta) near the surface.
[0038]
Further, at time t = t1, the substrate 301 is lowered from the position 302a to a position 303b near the bottom of the melt by a substrate vertical mechanism (not shown). At that time, the temperature of the melt around the substrate increases by ΔTh. Then, at time t = t2, the substrate 301 is raised again from the position 303b to the position 303a. At that time, the temperature of the melt around the substrate decreases by ΔTc. Similarly, the substrate 303 is reciprocated between the positions 303a and 303b. Due to this reciprocation, the growth temperature of the substrate changes between Ta and Tb at each point in time, and substantially follows the change indicated by the thick line in FIG. That is, a region having a relatively high temperature and a region having a relatively low temperature are formed inside the melt, and the substrate is alternately moved between them. The melt around the substrate undergoes a temperature increase of ΔTh and a temperature decrease of ΔTc. repeat.
[0039]
By the way, there is a melt connection between a high temperature region and a low temperature region, and silicon can diffuse mutually, so that the concentration of silicon as a whole tends to be uniform. As a result, in a low temperature region, the silicon concentration is high and supersaturated at a low temperature, and in a high temperature region, the silicon concentration is low and unsaturated at a high temperature. Therefore, the substrate 303 moves back and forth between them, so that growth and meltback occur alternately on the surface. It is considered that the growth amount and the meltback amount in a predetermined period are almost proportional to the temperature difference unless the difference between the initial melt temperature and the melt temperature at the end of the period is extremely large. Therefore, the ratio of the thickness at which the polycrystal grows and the thickness at which the polycrystal grows is approximately | ΔTh |: | ΔTc |. In this manner, the development of the grooves is suppressed, as in the case of cooling and heating the entire melt shown in the case of FIG. 7, and as shown in FIG. W becomes smaller, and the possibility that the grid electrode formed on the surface is disconnected is reduced.
[0040]
Moreover, according to the liquid phase growth method of the present invention, even when a large amount of melt is used, it is not necessary to repeatedly heat and cool the entire melt, so that the growth can be completed in a short time and the production throughput can be reduced. improves. In addition, high-quality polycrystals can be mass-produced without wasting energy due to repeated heating and cooling.
[0041]
The following forms can be considered as modified examples of the above-described embodiment.
[0042]
FIGS. 3A and 3B show the structure of another liquid phase growth apparatus suitable for carrying out the liquid phase growth method of the present invention, wherein FIG. 3A is a schematic view of a side sectional structure, and FIG. It is.
[0043]
In this liquid-phase growth apparatus, heaters 402a and 402b divided into two parts in the circumferential direction are provided as heating means capable of forming a temperature distribution of the melt 401 around the central axis of the crucible 400. These heaters 402a and 402b are provided to face the side of the crucible 400, and can be controlled independently. Further, the substrate 403 is configured such that two rows holding a plurality of substrates are arranged at predetermined intervals in the vertical direction and supported by a support rack (not shown). Further, with the rotation of the support frame, each substrate can alternately pass through a position 403a near the heater 402a and a position 403b near the heater 402b. Further, by controlling the heater 402b independently of the heater 402a, the temperature of the melt at the position 403b can be set lower than the temperature of the melt at the position 403a.
[0044]
Even if this apparatus is used, the same liquid phase growth method as that for moving the substrate in the vertical direction can be realized.
[0045]
Further, for example, in the above-described embodiment, the substrate is moved between a low-temperature region and a high-temperature region in the same crucible, but the melt is held in a plurality of crucibles and the temperature of each melt is reduced. You may make it different.
[0046]
Instead of providing a temperature distribution as described above, the substrate may be moved between regions having different densities even at a constant temperature. In this case, the melt is in contact with the polycrystalline surface (more strictly, the polycrystalline substrate surface in the first step and the polycrystalline surface formed immediately before in the other steps) in the first step and the second step. In this case, the content of the crystal raw material per unit volume is different. Examples of a method for providing a concentration distribution include a method using a plurality of crucibles and a method using a specific gravity difference between a solute and a solvent.
[0047]
Furthermore, in a preferred embodiment of the present invention, it is preferable that the positional relationship between the melt and the substrate is relatively changed. Therefore, instead of moving the substrate as described above, the crucible may be moved. In order to simplify the moving mechanism, a mode in which the substrate is moved is more preferable.
[0048]
【Example】
[Example 1]
Metal-grade silicon having a purity of 98.5% was melted and gradually cooled in a vacuum melting furnace to form a polycrystalline silicon ingot. This was sliced, molded into a polycrystalline silicon wafer having a side of 5 inches, and the surface was etched with a mixed solution of hydrofluoric acid, nitric acid, and acetic acid to remove the surface roughness accompanying the slicing.
[0049]
This polycrystalline silicon consisted of amorphous crystal grains having a diameter of about several mm to 1 cm. The specific resistance was about 0.05 Ωcm, and the conductivity type was p-type.
[0050]
Using the liquid phase growth apparatus shown in FIG. 1, indium was charged into crucible 300, and the temperature was set to 950 ° C. in a hydrogen atmosphere. A solar cell grade polycrystalline silicon wafer was set on a substrate up / down mechanism (not shown) and immersed in indium while maintaining a hydrogen atmosphere to form a melt 301 saturated with silicon.
[0051]
Next, the polycrystalline silicon wafer was removed from the substrate up-down mechanism, and five substrates 303 made of metal-grade silicon were set, and then cooling of the melt was started. When Ta = Tb = Th = 950 ° C. (t = t0), the substrate 303 was immersed in the melt and held at the position of 303a. When Ta = 900 ° C. (ΔTc = −50 ° C., t = t1), the substrate 303 was slowly moved to the position of 303b. At the end of the transfer, Tb was 915 ° C. (ΔTh = 15 ° C.). Then, the substrate 303 was slowly moved to the position 303a. At the end of the movement, Ta = 865 ° C. (ΔTc = −50 ° C.).
[0052]
Similarly, the substrate 303 is repeatedly moved up and down so that ΔTh = 15 ° C. and ΔTc = −50 ° C. in FIG. 4. Finally, the substrate 303 is held at the position 303b, and when Tb = 800 ° C. (t = tf) The substrate 303 was pulled out of the melt. That is, the temperature of the melt near the substrate changed according to the profile of the thick line in FIG. It took 70 minutes during this time. When this temperature profile is followed, the thickness at which the polycrystal melts back is considered to be 30% of the grown thickness.
[0053]
Next, a phosphorus diffusing agent (OCD) was applied to the surface of the polycrystalline silicon, and phosphorus was diffused at a temperature of 900 ° C. for 30 minutes. As a result, phosphorus diffuses from the surface of the polycrystalline silicon to a depth of about 0.2 μm, and n ⁺ Layer 104 was formed. After removing the diffusing agent, a pattern of the grid electrode 106 was printed on the surface using a silver paste. The grid electrode 106 was adjusted to have a width of 150 μm and a thickness of about 20 μm. Titanium oxide was deposited thereon by sputtering to a thickness of about 550 angstroms to form a solar cell as an anti-reflection film (not shown).
[0054]
When the solar cell was measured by a solar simulator of AM1.5, a conversion efficiency of 15.0% was obtained. Further, when observed with a microscope, no disconnection of the grid electrode 106 was found on the groove 105. Furthermore, when this solar cell was cut and its cross section was observed with a metallographic microscope, a polycrystalline silicon film having an average thickness of 50 μm was grown on the surface of the substrate 303. Grooves were formed along the grain boundaries in the polycrystalline silicon film, and the depth was 10 μm on average. That is, G1 / W = 20% in the definition of FIG.
[0055]
Also, except that ΔTh = 2.5 ° C. and ΔTc = −50 ° C., liquid phase growth was performed in the same procedure as above, and a solar cell having a similar shape was produced. When this temperature profile is followed, the thickness at which the polycrystal melts back is considered to be 5% of the grown thickness.
[0056]
With this solar cell, a conversion efficiency of 14.5% was obtained. According to microscopic observation, the thickness of the polycrystalline growth layer was about 50 μm, the depth of the groove was about 20 μm, and G1 / W = 40% as defined in FIG. Furthermore, when the shape of the grid electrode of this solar cell was observed with a metallurgical microscope, disconnections were observed in some places. The reason that the conversion efficiency was slightly lower than that of the solar cell with ΔTh = 15 ° C. and ΔTc = −50 ° C. is presumably due to the disconnection of the grid electrode 206 caused by the slightly deep groove.
[0057]
(Reference example)
For comparison, on the polycrystalline silicon substrate made of the same metal-grade silicon as used in Example 1, ΔTh and ΔTc were changed by adjusting the control parameters of the heater divided into two for heating the melt. Polycrystalline silicon was grown in the same manner except for the above.
[0058]
After melting the polycrystalline silicon wafer into indium at a temperature of 950 ° C. in the same manner as in Example 1, the cooling of the melt was started, and when the temperature of the melt reached 940 ° C., the metal-grade silicon substrate 303 was removed. It was immersed and cooled monotonously, that is, ΔTc = 0 ° C., and the temperature was raised when the melt reached a temperature of 800 ° C. During this time, 70 minutes were required.
[0059]
As in Example 1, polycrystalline silicon having a thickness of about 50 μm was grown on the substrate. Then, as in the first embodiment, n ⁺ A layer 104 and an antireflection film (not shown) were formed.
[0060]
The conversion efficiency of this solar cell was 11.5%. When the solar cell was observed with a microscope, the grid electrode 206 was broken everywhere on the groove as shown in FIG. 6, but the average depth of the groove was 30 μm, and G2 / W = 80% in FIG. Had become. The reason that the solar cell of this comparative example is lower in efficiency than the solar cell of Example 1 is because the liquid crystal growth method of the present invention was not used, so that a deep groove was generated, the grid electrode was disconnected, and the current generated by light incidence was low. It turned out to be due to inefficient collection.
[0061]
Similarly, polycrystalline silicon was grown at ΔTc = −50 ° C. and ΔTh = 2 ° C. When this temperature profile is followed, the thickness at which the polycrystal melts back is considered to be 4% of the grown thickness. G2 / W = 60% when the groove depth was 30 μm, and the conversion efficiency of the solar cell thus produced was 12.3%.
[0062]
Similarly, polycrystalline silicon was grown at ΔTc = −50 ° C. and ΔTh = 20 ° C. When this temperature profile is followed, the thickness at which the polycrystal melts back is considered to be 40% of the grown thickness. Although the groove depth was improved to G2 / W = 14% at 7 μm, the conversion efficiency of the manufactured solar cell was rather reduced to 13.7%. When the surface of the other crystal was observed again, an inverted pyramid-shaped depression was partially observed, and it was found that the characteristics were impaired. It is presumed that this dent had a large melt-back amount, and thus a large melt-back occurred locally around a small defect on the polycrystalline surface.
[0063]
From the above experiment, it was found that the thickness of the polycrystal to be melted back is appropriately in the range of 5 to 30% of the grown thickness.
[0064]
(Comparative example)
Polycrystalline silicon was grown on a polycrystalline silicon substrate made of the same metal-grade silicon used in Example 1 using the same liquid phase growth apparatus used in Example 1. However, in this comparative example, an attempt was made to control the heater and control the temperature of the entire melt in accordance with the change in the thick line in FIG.
[0065]
However, the melt has a weight of 10 kg, and even when heating the melt in the course of cooling or cooling the melt in the middle of heating, the temperature of the melt does not rapidly follow, so the capacity of the heater is increased, In order to perform cooling, a device such as flowing hydrogen gas for cooling around the crucible was devised. However, in order to change the temperature from 940 ° C. to 800 ° C. while applying ΔTh = 20 ° C. and ΔTc = −60 ° C. It took a minute.
[0066]
In the solar cell using polycrystalline silicon prepared in this comparative example, no disconnection of the grid electrode was observed, and the conversion efficiency was 14.8%, which was almost equivalent to that of Example 1. However, according to the results of this comparative example, when a large amount of melt is used, it takes time to control the temperature of the entire melt and apply ΔTh and ΔTc, and it is more difficult than in Example 1 in terms of production efficiency. It was confirmed that it was inferior.
[0067]
[Example 2]
In this embodiment, an example using the liquid phase growth apparatus shown in FIG. 3 which is optimal for growing a larger amount of substrates per batch will be described.
[0068]
After melting and slowly cooling metal grade silicon having a purity of 98.5% as a substrate in a vacuum melting furnace, then pouring into a carbon mold and slowly cooling to directly form a polycrystalline silicon substrate having a side of 5 inches. Then, the surface is etched with a mixed solution of hydrofluoric acid and nitric acid to remove the surface roughened in contact with the mold. This polycrystalline silicon consisted of amorphous crystal grains having a diameter of about several mm to 1 cm. The specific resistance was about 0.05 Ωcm, and the conductivity type was p-type.
[0069]
In this liquid-phase growth apparatus, heaters 402a and 402b divided into two parts in the circumferential direction are provided as heating means capable of forming a temperature distribution of the melt 401 around the central axis of the crucible 400. These heaters 402a and 402b are provided to face the side of the crucible 400, and can be controlled independently. In addition, a large number of substrates 403 are arranged such that two rows holding 50 substrates at predetermined intervals in the vertical direction are arranged, and a total of 100 substrates 403 are supported by a support frame (not shown). Further, with the rotation of the support frame, each substrate can alternately pass through a position 403a near the heater 402a and a position 403b near the heater 402b. Further, by controlling the heater 402b independently of the heater 402a, the temperature of the melt at the position 403b can be set lower than the temperature of the melt at the position 403a.
[0070]
Indium was charged into crucible 400, and the entire melt was heated to 950 ° C. in a hydrogen atmosphere. A solar cell grade polycrystalline silicon nugget was set in a quartz glass dedicated holder (not shown) and immersed in indium to form a melt 401 saturated with silicon.
[0071]
Next, a substrate 403 formed using metal-grade silicon was set on a support frame (not shown). Then, cooling of the melt was started so that the temperature of the melt near the heater 402a was Ta in FIG. 5 and the temperature of the melt near the heater 402b was Tb in FIG. When Ta = Tb = Th = 940 ° C. (t = t0), the substrate 403 was immersed in the melt. When Ta = 900 ° C. (t = t1), the support was slowly rotated by 180 degrees. The substrate which was initially at the position 403a was moved to the position 403b, and the temperature of the substrate reached 920 ° C. On the other hand, the temperature of the substrate at the position of 403b was changed from about 930 ° C. to about 900 ° C. Hereinafter, at times t = t2, t = t3,..., The support rack was slowly rotated by 180 degrees.
[0072]
As a result, the temperature of the substrate in the row at the initial position 403a changed according to the thick solid line in FIG. The temperature of the substrate in the row that was initially at the position 403b changed according to the thick dotted line in FIG. It took 70 minutes from t = t0 to t = tf. The liquid phase growth apparatus of this embodiment weighs about 200 kg, which is about 20 times that of the liquid phase growth apparatus of Example 1. However, since the amount of the melt is large, there is no difficulty in controlling the temperature. The effects of the invention have been demonstrated.
[0073]
Hereinafter, a solar cell was prepared in the same procedure as in Example 1. The conversion efficiency was 14.7% on average. According to microscopic observation of the cross section, G1 / W = 25% on average. The reason why the conversion efficiency was slightly lower and G1 / W was larger than in the case of Example 1 is presumably because silicon was insufficiently diffused between the region where the melt temperature was high and the region where the melt temperature was low.
[0074]
[Example 3]
In this embodiment, a method for further improving the result of the second embodiment will be described. In this embodiment, a liquid phase growth apparatus similar to the liquid phase growth apparatus of FIG. 3 is used, but the crucible can be rotated slowly around its central axis.
[0075]
As the substrate, a substrate manufactured by the same method as in Example 2 was used. As in the second embodiment, the melt temperature near the heater 402a in FIG. 3 and the melt temperature near the heater 402b were controlled to change according to Ta and Tb in FIG. Further, from t0 to tf, the support frame of the substrate was rotated at the same timing as in the comparative example. Meanwhile, the crucible was rotated once every 5 minutes. Using the grown polycrystal, a solar cell was produced in the same manner as in Example 2.
[0076]
The conversion efficiency was 15.2% on average. According to microscopic observation of the cross section, G1 / W = 18% on average. The reason why the conversion efficiency was higher and G1 / W was smaller than in the case of Example 2 is that the silicon melted back from the substrate in the high temperature region was transported to the low temperature region because the crucible rotated slowly, Concentration of silicon increased, and the precipitation of silicon on the substrate became active.On the other hand, in the high temperature region, the concentration of silicon in the melt decreased, and the melt back of silicon from the substrate became active. It appears to be.
[0077]
【The invention's effect】
As described above, according to the present invention, even when a polycrystal is grown on a polycrystal substrate by the liquid phase growth method, the groove generated in the polycrystal does not become deeper than the thickness of the polycrystal, so Disconnection of the grid, which tends to occur when a solar cell is manufactured using crystals, can be prevented. In particular, even when a large amount of melt is used, it can be easily applied without lowering the throughput, so that it is suitable for mass production of solar cells.
[Brief description of the drawings]
FIG. 1 shows a structure of a liquid phase growth apparatus suitable for carrying out a liquid phase growth method of the present invention, wherein (a) is a schematic diagram of a side sectional structure, and (b) is a plan view thereof. .
FIG. 2 is an explanatory diagram showing a cross-sectional state of a solar cell grown on a polycrystalline silicon substrate by the liquid phase growth method of the present invention.
FIGS. 3A and 3B show the structure of another liquid phase growth apparatus suitable for carrying out the liquid phase growth method of the present invention, wherein FIG. 3A is a schematic view of a side sectional structure, and FIG. It is.
FIG. 4 is an explanatory diagram showing an example of a cooling profile of a melt in the liquid phase growth method of the present invention.
FIG. 5 is an explanatory view showing another example of a cooling profile of a melt in the liquid phase growth method of the present invention.
FIG. 6 is an explanatory diagram showing a cross-sectional state of a solar cell in which a grid electrode has been broken because it has been grown on a polycrystalline silicon substrate by a conventional liquid phase growth method.
FIG. 7 is an explanatory view showing an example of a cooling profile of a melt recommended for suppressing a polycrystalline groove in liquid phase growth.
[Explanation of symbols]
100 polycrystalline substrate
101 Substrate grain boundary
102 Polycrystalline growth layer
103 Grain boundary of polycrystalline growth layer
104 n ⁺ layer
105 grooves
106 grid electrode
300 crucible
301 Melt
302a first heater
302b Second heater
303a Substrate in first position
303b Substrate in second position
303 substrate
400 crucible
401 Melt
402a first heater
402b Second heater
403a Substrate in first position
303b Substrate in second position
403a substrate
403b substrate

Claims (16)

A liquid phase growth method in which a polycrystalline substrate is immersed in a melt in a crucible in which a crystal raw material is melted, and a polycrystal is grown on the substrate.
A first step of growing a polycrystal of a predetermined thickness;
A second step of melting a part of the polycrystal grown in the first step in a melt,
A liquid phase growth method, wherein the relative positions of a substrate and a melt are changed between a first step and a second step, and melts having different temperatures are brought into contact with a polycrystalline surface. The liquid phase growth method according to claim 1, wherein the first step and the second step are alternately repeated a plurality of times. 3. The method according to claim 1, wherein in the second step, the polycrystal grown in the first step is melted in a melt in a region of 5% or more and 30% or less with respect to the predetermined thickness. 3. The liquid phase growth method described in 1. above. 4. The liquid phase growth method according to claim 1, wherein the thickness of the polycrystal grown in one first step is 1 to 20 [mu] m. Contacting the substrate with the melt in a relatively high temperature region in a first step, and contacting the substrate with the melt in a relatively low temperature region in a second step;
5. The liquid phase growth method according to claim 1, wherein the temperature of each region of the melt is further reduced. 6. The liquid phase growth method according to claim 1, wherein the first step and the second step are performed in the same crucible. In the first step and the second step, the substrate is alternately brought into contact with a melt in a relatively high temperature region and a melt in a relatively low temperature region by reciprocating the substrate up and down in a crucible. The liquid phase growth method according to claim 5, wherein: In the first step and the second step, the substrate is disposed at a position deviated from the central axis of the crucible, and is rotated around the central axis, so that the substrate is melted in a region having a relatively high temperature. 7. The liquid phase growth method according to claim 5, wherein the melt is alternately brought into contact with the melt in a relatively low region. 6. The liquid phase growth method according to claim 1, wherein the first step and the second step are performed in different crucibles. 10. The liquid phase growth according to claim 1, wherein the content of the crystal raw material per unit volume in the melt in contact with the polycrystalline surface differs between the first step and the second step. Method. At least a melt for dissolving the crystal raw material,
A crucible containing the melt,
Heating means capable of forming a temperature distribution in the melt;
A support frame for holding the substrate,
Support frame driving means for changing the relative position between the substrate and the melt between a relatively high temperature region and a relatively low temperature region of the melt,
A liquid phase growth apparatus comprising: The liquid phase growth apparatus according to claim 11, wherein the heating means capable of forming the temperature distribution of the melt is a vertically divided heater which can be controlled independently. 13. The liquid phase growth apparatus according to claim 11, wherein the means for changing the relative position between the substrate and the melt is means for moving the substrate up and down. The liquid phase growth apparatus according to claim 11, wherein the heating means capable of forming the temperature distribution of the melt is a heater which is divided in a circumferential direction and can be controlled independently. 15. The liquid phase growth apparatus according to claim 11, wherein the means for changing the relative position between the substrate and the melt is a means for rotating the substrate around the central axis. 16. The liquid phase growth apparatus according to claim 11, wherein the means for changing the relative position between the substrate and the melt includes means for rotating a crucible.

Images