WO2010082642A1

WO2010082642A1 - Photoelectric conversion device and solar cell using the same

Info

Publication number: WO2010082642A1
Application number: PCT/JP2010/050454
Authority: WO
Inventors: Hiroshi Kawakami
Original assignee: Fujifilm Corp
Current assignee: Fujifilm Corp
Priority date: 2009-01-16
Filing date: 2010-01-08
Publication date: 2010-07-22
Anticipated expiration: 2011-07-16
Also published as: JP4550928B2; JP2010165878A

Abstract

Stably providing photoelectric conversion devices having a high withstand voltage and excellent photoelectric conversion efficiency with a high yield rate. Photoelectric conversion device (1) includes substrate (10) of an Al based metal base having an anodized film on at least one surface side on which photoelectric conversion layer (30) which includes a compound semiconductor formed of a group Ib element, a group IIIb element, and a group VIb element, and generates a current by absorbing light, and electrodes (20, 50) for drawing out the current are provided, in which the metal base has a Fe content of 0.05 to 1.0% by mass, and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5μm, in a cross-section of metal base is 1,500 to 40,000/mm².

Description

DESCRIPTION

PHOTOELECTRIC CONVERSION DEVICE AND SOLAR CELL USING THE SAME

Technical Field

The present invention relates to a photoelectric conversion device having a flexible substrate on which a photoelectric conversion layer, which includes a compound semiconductor formed of a group Ib element, a group Ilib element, and a group VI element, is provided, and a solar cell using the same.

Background Art

Photoelectric conversion devices having a photoelectric conversion layer that generates a current by absorbing light and electrodes for drawing out the current generated in the photoelectric conversion layer are used in various applications, such as solar cells and the like. Most of the conventional solar cells are Si cells using bulk monocrystalline Si, polycrystalline Si, or thin film amorphous Si. Recently, however, research and development of compound semiconductor solar cells that do not depend on Si has been carried out. Two types of compound semiconductor solar cells are known, one of which is a bulk system, such as GaAs system and the like, and the other of which is a thin film system, such as CIS (Cu-In-Se) system formed of a group Ib element, a group IHb element, and a group VIb element, CIGS (Cu-In-Ga-Se) , or the like. CIS systems or CIGS systems has high light absorption rates and high energy conversion efficiency is reported.

Currently, glass substrates are mainly used as the substrates for solar cells, but they lack flexibility and are likely to be damaged, thereby making it difficult to reduce the thickness and weight of solar cells. Resin substrates are flexible and allow thinner and more lightweight solar cells, but have low upper temperature limits in comparison with inorganic substrates, limiting the processing temperature and making it difficult to form a photoelectric conversion layer having high photoelectric conversion efficiency. Use of flexible metal substrates as solar cell substrates has been under study. When a metal substrate is used, it is essential to provide an insulation film on a surface of the substrate in order to prevent a short circuit between the substrate and electrodes or a photoelectric conversion layer formed on the substrate. Here, it is important that the insulation film has a high withstand voltage, i.e., insulation film is not broken down under operation in which a voltage is applied across the film. If a breakdown occurs in the insulation film, leakage current flows and the photoelectric conversion efficiency is degraded. Accordingly, it is necessary to stably form insulation films having a high withstand voltage in order to improve the yield rate.

In order to prevent warpage of the substrate due to thermal stress, it is preferable that the difference in thermal expansion coefficient between the substrate and each layer formed on the substrate is small. As for the metal substrate, an Al based metal base is preferably used from the viewpoints of the difference in thermal expansion coefficient with the photoelectric conversion layer or lower electrode (rear electrode) , cost, and characteristics required of solar cells.

Japanese Unexamined Patent Publication No. 2000-349320 proposes the use of a substrate having an Al base on which a porous anodized film (AI2O3 film) is formed or a substrate providedby filling vacant holes of the anodized film with an oxide as a solar cell substrate. Such method allows an insulation film to be formed easily over the entire surface of a substrate without any pinhole even if the substrate has a large area.

As described above, for a metal substrate of a photoelectric conversion device, it is important that an insulation film formed on the surface of the substrate has excellent insulation properties, and the substrate has a high withstand voltage and heat resistance sufficient for preventing softening and deformation of the substrate in a high temperature film forming process. It is conceivable that such properties are influenced by an impurity in the substrate, but Japanese Unexamined Patent Publication No. 2000-349320 does not mention on this point at all.

The present invention has been developed in view of the circumstances described above, and it is an object of the present invention to stably provide a photoelectric conversion device having a high withstand voltage and excellent photoelectric conversion efficiency with a satisfactory yield rate, the device including an Al based metal base having an anodized film on at least one surface on which a photoelectric conversion layer, which includes a compound semiconductor formed of a group Ib element, a group IIIb element, and a group VIb element, is provided.

Disclosure of Invention

A photoelectric conversion device of the present invention is a device, including a substrate of Al based metal base having an anodized film on at least one surface side on which a photoelectric conversion layer which includes a compound semiconductor formed of a group Ib element, a group 11Ib element, and a group VIb element and generates a current by absorbing light, and electrodes for drawing out the current are provided, wherein: the metal base has a Fe content of 0.05 to 1.0% by mass; and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3um and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5um, in a cross-section of the metal base is 1,500 to 40,000/mm². The major component of the metal base is defined as a component occupying 98% by mass of the base. The metal base may be a substrate that includes a minor component, a pure Al substrate, or an alloy substrate of Al and the other metal element.

Element group representation herein is based on the short period periodic table. A compound semiconductor formed of a group Ib element, a group IHb element, and a group VIb element is sometimes represented herein as "group I-III-VI semiconductor". Each of the group Ib element, group IHb element, and group VIb element, which are constituent elements of group I-III-VI semiconductor, may be one type or two or more types of elements. Further, the I-III-VI semiconductor included in the photoelectric conversion layer may be one type or two or more types of semiconductors.

Preferably, in the photoelectric conversion device of the present invention, the metal base has an Al content of not less than 98.0% by mass, a Si content of not greater than 0.25% by mass, and a Cu content of not greater than 0.20% by mass.

"Determination of the amount of a minor component in the metal base" may be made by plasma emission spectrometry specified in JIS H 1305-1307 or by atomic absorption spectrometry. A method for mass-spectrometering a sputtered metal base may also be used in combination of the method described above.

A minor component in the metal base is solidified substantially uniform over the entire substrate if the amount is minuscule, but the minor component exceeding a solidification limit is unevenly distributed as a cluster-like domain. A fine cluster-like domain in which Fe is eccentrically located is herein referred to as "a Fe containing cluster".

The measurement of "the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5μm" may be performed by observing a cross-section of the metal base with a surface analyzer having a scanning electron microscope

(SEM) and an electron probe microanalyzer (EPMA) . The visibility of Fe containing clusters may vary depending on the orientation of the cross-section with respect to the rolling direction. Therefore, each cross-section is taken perpendicular to the surface of the metal base and cross-sections in six directions taken at an interval of 30 degrees with an arbitrary direction as the origin are observed. The visual field of observation is divided such that the total area of the visual field corresponding to each direction becomes substantially equal to each other, and observation of an area not less than 10^"3 mm² in one direction is sufficient. Where the number of target clusters is great, the visual field may be an area where more than 250 clusters are observed in accumulation. The size of each cluster is evaluated by the size of a whitish area (area of large reflection electron emission) in a SEM image, and a determination whether or not the cluster is a cluster in which Fe is eccentrically located is made by determining whether or not a Fe-derived component is included in a characteristic X-ray response signal in EPMA..

When the EPMA. is an energy dispersive type (EDX) , the Fe-derived response signal is observable at X-ray photon energies near 0.69 keV, near 6.4 keV, and near 7.1 keV. When a cross-section of the Al based metal base is observed, few signal is observed from an area other than the whitish area in a SEM image. Consequently, the determination as to whether or not a whitish area in a SEM image is a Fe containing cluster may be made by the method describe above .

It is difficult to evaluate the size of a Fe containing cluster from a spatial distribution image of response signal in EPMA because of insufficient spatial resolution of EPMA. Further, if the sum of minimum and maximum diameters divided by 2 falls below 0.5um, the Fe-derived response signal in EPMA becomes weak and is not detected clearly. Consequently, in the present invention, the lower limit of the value obtained by dividing the sum of minimum and maximum diameters of Fe containing clusters by 2 is set to 0.5um. Further, in actuality, a Fe containing cluster, in which the sum of minimum and maximum diameters divided by 2 exceeds 2.5um, is rarely formed. What is more, the presence of even one Fe containing cluster, in which the sum of minimum and maximum diameters divided by 2 exceeds 2.5um, causes a dielectric breakdown and becomes unusable. Therefore, in the present invention, the upper limit of the sum of minimum and maximum diameters of Fe containing clusters divided by 2 is set to 2.5μm.

Preferably, in the photoelectric conversion device of the present invention, the photoelectric conversion layer includes a compound semiconductor formed of: at least one type of group Ib element selected from the group consisting of Cu and Ag; at least one type of IHb element selected from the group consisting of Al, Ga, and In; and at least one type of VIb element selected from the group consisting of S, Se, and Te.

A solar cell of the present invention is a cell, including the photoelectric conversion device described above. According to the present invention, photoelectric conversion devices having a high withstand voltage and excellent photoelectric conversion efficiency are stably provided with a high yield rate, each device including a substrate of Al based metal base having an anodized film on at least one surface side on which a photoelectric conversion layer which includes a compound semiconductor formed of a group Ib element, a group IHb element, and a group VI element, are provided is provided.

Brief Description of Drawings Figure IA is a schematic sectional view of a photoelectric conversion device according to an embodiment of the present invention taken along a lateral direction.

Figure IB is a schematic sectional view of a photoelectric conversion device according to an embodiment of the present invention taken along a longitudinal direction.

Figure 2 is a schematic sectional view of a substrate, illustrating the structure thereof.

Figure 3 is a perspective view of a substrate, illustrating a manufacturing method thereof. Figure 4 illustrates the relationship between the lattice constant and band gap of I-III-VI compound semiconductors.

Figure 5A shows an example of SEM cross-section photo of a substrate.

Figure 5B shows examples of EDX chart.

Best Mode for Carrying Out the Invention [Photoelectric Conversion Device]

A structure of a photoelectric conversion device according to an embodiment of the present invention will be described with reference to the accompanying drawings. Figure IA is a schematic sectional view of the photoelectric conversion device in a lateral direction, and Figure IB is a schematic sectional view of the photoelectric conversion device in a longitudinal direction. Figure 2 is a schematic sectional view of a substrate, illustrating the structure thereof, and Figure 3 is a perspective view of a substrate, illustrating a manufacturing method thereof. In the drawings, each component is not drawn to scale in order to facilitate visual recognition.

Photoelectric conversion device lisa device having substrate 10 on which lower electrode (rear electrode) 20, photoelectric conversion layer 30, buffer layer 40, and upper electrode 50 are stacked in this order. Photoelectric conversion device 1 has first separation grooves 61 that run through only lower electrode 20, second separation grooves 62 that run through photoelectric conversion layer 30 and buffer layer 40, and third separation grooves 63 that run through only upper electrode layer 50 in a lateral sectional view and fourth separation grooves 64 that run through photoelectric conversion layer 30, buffer layer 40, and upper electrode layer 50 in a longitudinal sectional view. The above configuration may provide a structure in which the device is divided into many cells C by first to fourth separation grooves 61 to 64. Further, upper electrode 50 is filled in second separation grooves 62, whereby a structure in which upper electrode 50 of a certain cell C is serially connected to lower electrode 20 of adjacent cell C may be obtained. (Substrate)

In the present embodiment, substrate 10 is a substrate obtained by anodizing at least one side of Al based metal base 11. Substrate 10 may be a substrate of metal base 11 having anodized film 12 on each side as illustrated on the left of Figure 2 or a substrate of metal base 11 having anodized film 12 on either one of the sides as illustrated on the right of Figure 2. Here, anodized film 12 is an Al₂Cb based film.

Preferably, substratelO is a substrate of metal base 11 having anodized film 12 on each side as illustrated on the left of Figure 2 in order to prevent warpage of the substrate due to the difference in thermal expansion coefficient between Al andAl₂Cb, and detachment of the film due to the warpage during the device manufacturing process . The anodizing method for both sides may include, for example, a method in which anodization is performed on a side-by-side basis by applying an insulation material and a method in which both sides are anodized at the same time.

When anodized film 12 is formed on each side of substrate 10, it is preferable that two anodized films are formed to have substantially the same film thickness or anodized film 12 on which a photoelectric conversion layer and some other layers are not provided is formed to have a slightly thicker film thickness than that of the anodized film 12 on the other side in consideration of heat stress balance between each side. Metal base 11 maybe Japanese Industrial Standards (JIS) 1000 pure Al or an alloy of Al with another metal element, such as Al-Mn alloy, Al-Mg alloy, Al-Mn-Mg alloy, Al-Zr alloy, Al-Si alloy, Al-Mg-Si, or the like (Aluminum Handbook, Fourth Edition, published by Japan Light Metal Association, 1990) . Metal base 11 may include traces of various metal elements, such as Fe, Si, Mn, Cu, Mg, Cr, Zn, Bi, Ni, Ti, and the like.

Anodization may be performed by immersing metal base 11, which is cleaned, smoothed by polishing, and the like as required, as an anode with a cathode in an electrolyte, and applying a voltage between the anode and cathode. As for the cathode, carbon, aluminum, or the like is used. There is not any specific restriction on the electrolyte, and an acid electrolyte including one type or more types of acids, such as sulfuric acid, phosphoric acid, chromic acid, oxalic acid, sulfamic acid, benzenesulfonic acid, amido-sulfonic acid, and the like, is preferably used.

There is not any specific restriction on the anodizing conditions and dependent on the electrolyte used. As for the anodizing conditions, for example, the following are appropriate: electrolyte concentration of 1 to 80% by mass, solution temperature of 5 to 70⁰C, current density of 0.005 to 0.60 A/cm², voltage of 1 to 200 V, and electrolyzing time of 3 to 500 minutes.

As for the electrolyte, a sulfuric acid, a phosphoric acid, an oxalic acid, or a mixture thereof may preferably be used. When such an electrolyte is used, the following conditions are preferable : electrolyte concentration of 4 to 30% by mass, solution temperature of 10 to 30⁰C, current density of 0.05 to 0.30 A/cm², and voltage of 30 to 150 V.

It is preferable to control the current applied to metal base 11 so as to have a profile in which the amount of the current is small just after the start of anodization and then the amount is gradually increased to a desired value, because this may reduce the number of spots locally produced during the anodization.

As shown in Figure 3, when Al based metal base 11 is anodized, an oxidization reaction proceeds from surface 11s in a direction substantially perpendicular to surface 11s, and AI2O3 based anodized film 12 is formed. Anodized film 12 generated by the anodization has a structure in which multiple fine columnar bodies, each having a substantially regular hexagonal shape in plan view, are tightly arranged. Each fine columnar body 12a has a fine pore 12b, in substantially the center, extending substantially linearly in a depth direction from surface 11s, and the bottom surface of each fine columnar body 12a has a rounded shape. Normally, a barrier layer without any fine pore 12b is formed (generally, with a thickness of 0.01 to 0.4um) at a bottom area of fine columnar bodies 12a. Anodized film 12 without any fine pore 12b may also be formed by appropriately arranging the anodizing conditions.

There is not any specific restriction on the diameter of fine pore 12b of anodized film 12. Preferably the diameter of fine pore 12b is 200nm or less, and more preferably lOOnm or less from the viewpoints of surface smoothness and insulation properties. It is possible to reduce the diameter of fine pore 12b to about lOnm.

There is not any specific restriction of the pore density of fine pores 12b of anodized film 12. Preferably, the pore density of fine pores 12b is 100 to 10000/um², and more preferably 100 to 5000/μm², and particularly preferably 100 to 1000/um² from the viewpoint of insulation properties .

There is not any specific restriction on the surface roughness

Ra. From the viewpoint of uniformly forming the upper layer of photoelectric conversion layer 30, high surface smoothness is desirable. Preferably, the surface roughness Ra is 0.3μm or less, and more preferably O.lum or less.

There is not any specific restriction on the thicknesses of metal base 11 and anodized film 12. Preferably, the thickness of metal base 11 prior to anodization is, for example, 0.05 to 0.6mm, and more preferably 0.1 to 0.3mm in consideration of the mechanical strength of substrate 10, and reduction in the thickness and weight. When the insulation properties, mechanical strength, and reduction in the thickness and weight are taken into account, a preferable range of the thickness of anodized film 12 is 0.1 to lOOum. In order to prevent variations in the arrangement and diameter of fine pores 12b, the anodization may be performed after forming depressions of starting points of fine pores 12. Further, fine pores 12b of anodized film 12 may be sealed by any known sealing method as required. This sealing of fine pores may improve the withstanding voltage and insulation properties. Further, if the sealing is performed using a material which includes alkali metal ions, the alkali metal ions, preferably Na ions, are diffused into photoelectric conversion layer 30 of CIGS or the like, which may sometimes improve the crystallization of photoelectric conversion layer 30 and hence the photoelectric conversion efficiency of layer 30.

It is important for the substrate of a photoelectric conversion device to include an insulation film, formed on a surface, having excellent insulation properties, and has a high withstand voltage and heat resistance sufficient for preventing softening and deformation of the substrate in a high temperature film forming process. It is conceivable that such properties are influenced by an impurity in the substrate, but no such specific study report has been found. In Al based metal base 11, increase in the content of Fe tends _

to improve the heat resistance. An excessive content of Fe, however, tends to reduce the withstanding voltage of anodized film 12. The inventor of the present invention has paid attention not only to Fe content but also to the existence form of Fe in metal base 11. Preferably, Fe is dispersed in metal base 11 in particles as small as possible because it increases the heat resistance of metal base 11. Further, if Fe is present in the form of a cluster of large particles, a dielectric breakdown of anodized film 12 is likely to occur. The inventor of the present invention has found that photoelectric conversion devices 1 having excellent insulation properties and photoelectric conversion efficiency may be stably provided with a high yield rate when the metal base 11 has a Fe content if 0.05 to 1.0% by mass, and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5μm, in a cross-section of metal base 11 is 1,500 to 40,000/mm². The inventor of the present invention has also found that photoelectric conversion devices 1 having excellent insulation properties and photoelectric conversion efficiency may be stably provided with a high yield rate in a high temperature process of 470 to 550⁰C if metal base 11 satisfies such requirements.

Preferably, the metal base 11 has a Fe content of 0.1 to 0.7% by mass. Preferably, the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3um and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to

2.5um, in a cross-section of metal base 11 is 3,000 to 24,000/mm².

The Fe content and the number of Fe-containing clusters may be controlled within the ranges described above by adjusting the amount of a minor component and a manufacturing condition when manufacturing metal base 11.

Al based metal base 11 may be manufactured by casting a molten metal, prepared according to the composition and cleaned up as required, and rolling the cast metal . The molten metal cleanup process may include a degassing process for removing an unnecessary gas, such as hydrogen, from the molten metal, a filtering process for removing a foreign substance, and the combination of these processes. The casting method may include DC casting, continuous casting, and the like. In DC casting, solidification occurs at a cooling speed in the range from 0.5 to 30°C/sec. Facing is performed, as required, on the obtained ingot in which 1 to 30mm of the surface layer is cut. Before and after the facing, temperature equalization is effected. In the temperature equalization, heat treatment is performed for 1 to 48 hours at 450 to 620⁰C to prevent coarsening of intermetallic compounds.

After the DC casting, hot rolling and cold rolling are performed to obtain a rolled metal plate. The appropriate hot rolling initial temperature is 350 to 500⁰C. Process annealing may be performed before or after the hot rolling, or otherwise in the middle of the hot rolling. Conditions of process annealing are to heat the metal for 2 to 20 hours at 280 to 600⁰C using a batch annealing furnace or to heat the metal for 6 minutes or less at 400 to 600⁰C using a continuous annealing furnace. A finer crystalline structure may also be obtained by heating the metal at a temperature increase rate of 10 to 200°C/sec using a continuous annealing furnace.

Among continuous casting methods, methods that use cooling rolls typified by twin-roll process (Hunter method) and 3C process, or methods that uses a cooling belt or a cooing block typified by twin belt process (Hazelett process) and Alusuisse caster II type are industrially used. In the continuous casting, solidification occurs at a cooling speed in the range from 100 to l,000°C/sec.

Where continuous casting is performed, if, for example, a method that uses cooling rolls, such as Hunter method, is used, a cast plate with a thickness of 1 to 10mm may be cast directly and continuously, whereby the hot rolling process maybe omitted. Where, a method that uses a cooling belt, such as Hazelett process, a cast plate with a thickness of 10 to 50mm may be cast, which is generally rolled continuously by hot rolling rolls provided just after the casting process, whereby a continuous cast rolled plate with a thickness of 1 to lOinmmay be obtained. Thereafter, in either method, cold rolling is performed to obtain a rolled plate. In the continuous casting also, the process annealing may be performed.

The rolled pate obtained by either the DC casting or continuous casting is subjected to post treatment, such as surface smoothing, as required.

By appropriately selecting the amount of minor component and conditions of casting, rolling, and process annealing, the composition and state of crystalline structure of metal base 11 may be controlled, whereby the Fe content and the number of Fe-containing clusters may be controlled within the defined ranges of the present invention.

Qualitatively, the average volume of Fe-containing clusters increases as the high temperature period is increased and as the temperature decrease rate is reduced in each process of casting, rolling, and annealing. Here, the total number of Fe-containing clusters decreases in inversely proportional to the average volume. In the casting and rolling process, the process of DC casting, hot rolling, and cold rolling has a longer high temperature period and a smaller temperature decrease rate than the process of continuous casting and cold rolling, so that the average volume of the Fe-containing clusters tends to increase and the number tends to decrease. Therefore, the number of Fe-containing clusters in a specific size range may be controlled by using the nature described above even in a case where metal bases 11 have the same Fe content.

There is not any specific restriction on the minor component in metal base 11 other than Fe. Preferably, metal base 11 has an

Al content of not less than 98.0% by mass, a Si content of not greater than 0.25% by mass, and a Cu content of not greater than 0.20% by mass . A more preferable Al content is not smaller than 99.0% by mass . A more preferable Si content is not greater than 0.15% by mass, and a further preferable Si content is in the range from 0.03 to 0.15% by mass. A more preferable Cu content is not greater than 0.15% by mass, and a further preferable Cu content is in the range from 0.02 to 0.15% by mass. Where the amounts of minor components other tha Fefall within these ranges, photoelectric conversion device substrate 10 having a high withstand voltage may be obtained, whereby photoelectric conversion devices 1 having high photoelectric conversion efficiency may be stably obtained with a high yield rate. (Photoelectric Conversion Layer)

Photoelectric conversion layer 30 includes one or more types of compound semiconductors of at least one type of group Ib element, at least one type of group IHb element, and at least one type of group VIb element (group I-III-VI semiconductors) and generates a current by absorbing light.

Preferably, photoelectric conversion layer 30 is a layer that includes one or more types of compound semiconductors of at least one type of group Ib element selected from the group consisting of Cu and Ag, at least one type of group IHb element selected from the group consisting of B, Al, Ga, and In, and at least one type of group VIb element selected from the group consisting of O, S, Se, and Te.

In view of a high light absorption rate and high photoelectric conversion efficiency, it is preferable that photoelectric conversion layer 30 includes one or more types of compound semiconductors of at least one type of group Ib element selected from the group consisting of Cu and Ag, at least one type of group IHb element selected from the group consisting of Al, Ga, and In, and at least one type of group VIb element selected from the group consisting of S, Se, and Te.

The compound semiconductors described above may include but not limited to CuAlS₂, CuGaS₂, CuInS₂, CuAlSe₂, CuGaSe₂, CuInSe₂ (CIS), AgAlS₂, AgGaS₂, AgInS₂, AgAlSe₂, AgGaSe₂, AgInSe₂, AgAlTe₂, AgGaTe₂, AgInTe₂, Cu (In^_xGa_x) Se₂ (CIGS), Cu(Ini__xAl_x) Se₂, Cu(Ini-_xGa_x) (S, Se)₂, Ag(Ini__xGa_x)Se₂, and Ag(In^_xGa_x) (S, Se)₂.

It is particularly preferable that photoelectric conversion layer 30 includes CuInSe₂ (CIS) and/or CuInSe₂ solidified with Ga, i.e., Cu(In, Ga)Se₂ (CIGS) . CIS and CIGS are semiconductors having a chalcopyrite crystal structure, and high light absorption rates and high energy conversion efficiency are reported. They also have excellent durability and have less deterioration in the efficiency due to light exposure.

Photoelectric conversion layer 30 includes an impurity for obtaining an intended semiconductor conductivity type. The impurity may be included in photoelectric conversion layer 30 by diffusing from an adjacent layer or by active doping.

Photoelectric conversion layer 30 may have a density distribution of constituent elements of group I-III-VI semiconductors and/or impurities, and may have a plurality of layer regions of different semiconductivities, such as n-type, p-type, i-type, and the like. For example, in the CIGS system, the band gap width/carrier mobility and the like may be controlled by providing a distribution of the amount of Ga in photoelectric conversion layer 30 in the thickness direction, whereby high conversion efficiency may be designed.

Photoelectric conversion layer 30 may include one or more types of semiconductors other than the group I-III-VI semiconductor. Semiconductors other than the group I-III-VI semiconductor may include but not limited to a semiconductor of group IVb element, such as Si (group IV semiconductor) , a semiconductor of group IHb element and group Vb element such as GaAs (group III-V semiconductor) , and a semiconductor of group lib element and group VIb element, such as CdTe (group II-VI semiconductor) .

Photoelectric conversion layer 30 may include any arbitrary component other than semiconductors and an impurity for causing the semiconductors to become an intended conductivity type within a limit that does not affect the properties. There is not any specific restriction on the content of group I-III-VI semiconductors in photoelectric conversion layer 30, in which not less than 75% by mass is preferable, not less than 95% by mass is more preferable, and not less than 99% by mass is particularly preferable.

As for the film forming method of CIGS layer, the following are known: 1) multi source simultaneous deposition, 2) selenization process, 3) sputtering, 4) hybrid sputtering, and 5) mechanochemical processing. 1) Multi Source Simultaneous Deposition

As for the multi source simultaneous deposition, a three-stage process ("The Performance of Cu(In, Ga) Se₂-Based Solar Cells in Conventional and Concentrator Applications", J.R. Tuttle et al., Mat. Res. Soc. Symp. Proc, Vol. 426, pp. 143-151, 1996, and the like) , and a simultaneous deposition of EC group ("THIN FILM SOLAR CELL MODULES BASED ON CU (IN, GA.) SE₂ PREPARED BY THE COEVAPORATION METHOD", L. Stolt et al., 13™ EUROPEAN PHOTOVOLTAIC SOLAR ENERGY CONFERENCE, PP.1451-1455, 1995, and the like) are known. The former three-stage process in a process in which In, Ga, and Se is deposited simultaneously with a substrate temperature of 300°C in a high vacuum, then Cu and Se are deposited simultaneously after increasing the temperature to 500 to 56O⁰C, and finally In, Ga, and Se are deposited simultaneously. The latter EC group simultaneous deposition is a method in which Cu-excess CIGS is deposited in the first half and In-excess CIGS is deposited in the latter half.

As the modified method for improving the crystallization of a CIGS film, the following are known. a) A method that uses ionized Ga ("Growth of high-quality CuGaSe₂ thin films using ionized Ga precursor", H. Miyazaki et al., phys.

Stat. sol. (a), Vol.203, No. 11, pp.2603-2608, 2006, and the like) . b) A method that uses cracked Se ("Growth of Cu(Ini__xGa_x) Se₂ thin films using cracked selenium", M. Kawamura et al., Proc. of 68^th Meeting of Japan Society of Applied Physics (JSAP), 7p-L-6, 2007 autumn, Hokkaido Institute of Technology, and the like) . c) A method that uses radial Se ("Preparation of Cu(Ini-_xGa_x)Se₂ thin films using a Se-radical beam source and solar cell performance", S. Ishizuka et al., Proc. of 54^th Meeting of JSAP, 29p-ZW-10, 2007 spring, Aoyama Gakuin University, and the like) . d) A method that uses an optical excitation process ("High Quality CIGS Thin Films and Devices by Photo-Excited Deposition Process", Y. Ishiietal., Proc. of 54^th Meeting of JSAP, 29p-ZW-14, 2007 spring, Aoyama Gakuin University, and the like) .

2) Selenization Process The selenization process is also called a two-stage process, in which a metal precursor of a stacking film, such as Cu layer/In layer or (Cu Ga) layer/In layer is formed by sputtering, deposition method, or electrodeposition method, which are then heated at about 450 to 550°C in a selenium vapor or in a hydrogen selenide to induce thermal diffusion, whereby a selenium compound, such as Cu(Ini-_xGa_x)Se2 or the like, is formed. This process is called vapor phase selenization process. In addition, a solid phase selenization process is also known, in which solid phase selenium is stacked on a metal precursor film to induce solid phase diffusion reaction with the stacked solid phase selenium as the selenium source, thereby effecting selenization.

In the selenization process, in order to avoid a rapid volume expansion that occurs at the time of selenization, the following two methods are known. That is, a method in which selenium is mixed in a metal precursor film at a certain rate ("CuInSe₂-based solar cells by Se-vapor selenization from Se-containing precursors", T. Nakada et al., Solar Energy Materials and Solar Cells, Vol. 35, pp. 209-214, 1994, and the like), and a method in which multi-layer precursor film is formed by sandwiching selenium between thin metal films (e.g., Cu layer/In layer/Se layer Cu layer/In layer/Se layer) ("THIN FILMS OF CuInSe₂ PRODUCED BY THERMAL ANNEALING OF MULTILAYERS WITH ULTRA-THIN STACKED ELEMENTAL LAYERS", T. Nakada et al., 10™ EUROPEAN PNOTOVOLTAIC SOLAR ENERGY CONFERENCE, pp. 887-890, 1991, and the like) . As a film forming method for forming a graded band-gap CIGS film, a method in which a Cu-Ga alloy film is stacked first, then an In film is stacked thereon, and Ga density is inclined in the thickness direction using natural thermal diffusion when effecting selenization (K. Kushiya et al., Tech. Digest 9^th Photovoltaic Science and Engineering Conf. Miyazaki, 1996 (Intn. PVSEC-9, Tokyo, 1996), p. 149, and the like). 3) Sputtering

As the sputtering method, the following three methods are known. The first is a method that uses polycrystal CuInSe₂ as the target. The second is a two source sputtering in which Cu₂Se and In₂Se₃ are used as the target, and a H₂Se/Ar mixed gas is used as the sputtering gas PCdS/CuInSe₂ JUNCTIONS FABRICATED BY DC MAGNETRON SPUTTERING OF Cu₂Se AND In₂Se₃", J. H. Ermer et al., Proc. of 18^th IEEE Photovoltaic Specialists Conf., pp. 1655-1658, 1985, and the like) . The third is a three source sputtering in which Cu target, In target, and Se or CuSe target are sputtered in Ar gas ("Polycrystalline CuInSe₂ Thin Films for Solar Cells by Three-Source Magnetron Sputtering", T. Nakada et al., Jpn. J. Appl. Phys., Vol. 32, pp. L1169-L1172, 1993, and the like) . 4) Hybrid Sputtering

A hybrid sputtering in which Cu and In metals are DC sputtered and only Se is deposited is known ("Microstructural Characterization for Sputter-Deposited CuInSe₂ Films and Photovoltaic Devices", T, Nakada et al., Jpn. J. Appl. Phys., Vol. 34, pp. 4715-4721, 1995, and the like) .

5) Mechanochemical Processing

The mechanochemical processing is a method in which materials according to the CIGS composition are put in a planetary ball mill and mixed by mechanical energy to obtain CIGS powder, then the power is applied to a substrate by screen printing and annealed, whereby a CIGS film is obtained ("Fabrication of Cu(In, Ga) Se₂ thin films by a combination of mechanochemical and screen-printing/sintering processes", T. Wada et al., Phys. Stat. sol. (a), Vol. 203, No. 11, pp. 2593-1597, 2006, and the like) . 6) Other CIGS Film Forming Methods

Other CIGS film forming methods include but not limited to screen printing method, close-space sublimation technique, MOCVD method, and spraying method. For example, a fine particle film, including a group Ib element, a group IHb element, and a group VIb element, may be formed on a substrate by the screen printing method or spraying method, and the film is subjected to pyrolytic treatment or the like, whereby a crystal having a desired composition may be obtained as described, for example, in Japanese Unexamined Patent Publication Nos. 9 (1997) -074065 and 9 (1997) -074213. Figure 4 illustrates the relationship between the lattice constant and band gap of major I-III-VI compound semiconductors. Various band gaps may be obtained by changing the composition ratio. When a photon having energy greater than the band gap is incident on a semiconductor, the energy exceeding the band gap becomes heat loss . It is known by theoretical calculation that maximum conversion efficiency is ^"obtained at about 1.4 to 1.5 eV in the combination of the solar light spectrum and band gap.

A high conversion efficiency band gap may be obtained by increasing the band gap, for example, by increasing the Ga density of Cu(In, Ga) Se₂ (CIGS) , by increasing the Al density of Cu(In, Al)

Se₂, or by S density of Cu(In, Ga) (S, Se)₂. In the case of CIGS, the band gap may be controlled within the range from 1.04 to 1.68 eV.

The band structure may be inclined by changing the composition ratio in the film thickness direction. Two types of inclined band structures are known. One of which is a single graded band gap in which the band gap is increased from the light input window side toward the electrode on the opposite side, and the other of which is a double graded band gap in which the band gap is decreased from the light input window side toward the PN junction and then increased after the PN junction ("Anew approach to high-efficiency solar cells by band gap grading in Cu(In, Ga)Se₂ chalcopyrite semiconductors", T. Dullweber et al., Solar Energy Materials & Solar Cells, Vol. 67, pp. 145-150, 2001, and the like) . In either case, carriers induced by light are accelerated by an electric field internally generated by the inclination of the band structure and likely to reach the electrode easily, and the combining probability with recombination centers so that the electricity generation efficiency is improved as described, for example, in International Patent Publication No. WO04/090995. When a plurality of semiconductors having different band gaps is used with respect to each spectral range, the electricity generation efficiency may be improved by reducing heat loss due to the difference between photon energy and band gap. Use of a plurality of such photoelectric conversion layers stacked on top of another is referred to as tandem type. In the case of a two-layer tandem, electricity generation efficiency may be improved, for example, by using a combination of 1.IeV and 1.7eV. (Electrodes, Buffer Layer)

Each of lower electrode 20 and upper electrode 50 is made of a conductive material. Upper electrode 50 on the light input side needs to be transparent. In view of effective use of light, it is preferable that lower electrode 20 on the substrate side has light reflectivity.

When a main layer of photoelectrical conversion layer 30 excluding a region adjacent to buffer layer 40 is a p-type semiconductor, lower electrode 20 is used as a positive electrode and upper electrode 50 is used as a negative electrode. If the main layer of photoelectric conversion layer 30 is an n-type semiconductor, the polarity of lower electrode 20 and upper electrode 50 is reversed. As for the major component of lower electrode 20, Mo, Cr, W, or a combination thereof is preferably used. As for the major component of upper electrode 50, ZnO, ITO (indium tin oxide) , SnO₂, or a combination thereof is preferably used. Lower electrode 20 and/or upper electrode 50 may have a single layer structure or a laminated structure, such as a two-layer structure. As for buffer layer 40, CdS, ZnS, ZnO, ZnMgO, ZnS(O, OH), or a combination thereof is preferably used. The major component of the electrodes and buffer layer is defined as a component occupying 50% by mass or more. A preferable combination of the compositions is, for example, Mo lower electrode/CdS buffer layer/CIGS photoelectric conversion layer/ZnO upper electrode.

It is reported that, in a photoelectric conversion device using a soda lime glass substrate, an alkali metal element (Na element) in the substrate is diffused into the CIGS film, thereby improving energy conversion efficiency. In the present embodiment, it is also preferable to diffuse an alkali metal into the CIGS film. As for the alkali metal diffusion method, a method in which a layer including an alkali metal element is formed on a Mo lower electrode by deposition or sputtering as described, for example, in Japanese Unexamined Patent Publication No. 8 (1996) -222750, a method in which an alkali layer of Na₂S or the like is formed on a Mo lower electrode by soaking process as described, for example, in International Patent Publication No. WO03/069684, a method in which a precursor of In, Cu, and Ga metal elements is formed on a Mo lower electrode and then, for example, a water solution including sodiummolybdate is deposited on the precursor, or the like may be cited.

It is also preferable that lower electrode 20 is designed to have a laminated structure and a layer, including one or more types of alkali metal compounds, such as Na₂S, Na₂Se, NaCl, NaF, and sodium molybdate, is provided between the laminations of lower electrode 20. The layer may include a material that does not include an alkali metal, such as an aluminum oxide.

There is not any specific restriction on the conductivity type of photoelectric conversion layer 30 to upper electrode 50. Generally, photoelectric conversion layer is a p-layer, buffer layer 40 is an n-layer (n-CdS, or the like) , and upper electrode 50 is an n-layer (n-ZnO layer, or the like) or has a laminated structure of i-layer and n-layer (i-ZnO layer and n-ZnO, or the like) . It is believed that such conductivity types form a p-n junction or a p-i-n junction between photoelectric conversion layer 30 and upper electrode 50. Further, it is thought that provision of CdS buffer layer 40 on photoelectric conversion layer 30 results in an n-layer to be formed in a surface layer of photoelectric conversion layer 30 by Cd diffusion, whereby a p-n junction is formed inside of photoelectric conversion layer 30. It is also conceivable that an i-layer may be provided below the n-layer inside of photoelectric conversion layer 30 to form a p-i-n junction inside of photoelectric conversion layer 30. (Other Layers) Photoelectric conversion device 1 may have any other layer as required in addition to those described above. For example, a contact layer (buffer layer) for enhancing the adhesion of layers may be provided, as required, between substrate 10 and lower electrode 20, and/or between lower electrode 20 and photoelectric conversion layer 30. Further, an alkali barrier layer for preventing diffusion of alkali ions may be provided, as required, between substrate 10 and lower electrode 20. A reference is directed to Japanese Unexamined Patent Publication No. 8 (1996) -222750 for details of the alkali barrier layer. Photoelectric conversion device 1 of the present embodiment is structured in the manner as described above. Photoelectric conversion device 1 of the present embodiment is a device which includes a substrate 10 of Al based metal base 11 having anodized film 12 on at least one side on which photoelectric conversion layer 30 which includes a compound semiconductor of a group Ib element, a group IIIb element, and a group VIb element is provided. Here, Fe content in metal base 11 is 0.05 to 1.0% by mass and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3um and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5μm, in a cross-section of metal base 11 is 1,500 to 40,000/mm². According to the present embodiment, photoelectric conversion devices 1 having a high withstand voltage and excellent photoelectric conversion efficiency may be provided stably with a high yield rate. Photoelectric conversion device 1 is preferably applicable to a solar cell and the like. Photoelectric conversion device 1 may be turned into a solar cell by attaching, as required, a cover glass, a protection film, and the like. (Design Changes) The present invention is not limited to the embodiment described above, and design changes may be made as appropriate without departing from the sprit of the present invention. The present invention is applicable to any process of CIGS system photoelectric conversion devices. For example, research work has been conducted on CIGS system photoelectric conversion device using a substrate of resin, such as polyimide. Where a resin substrate is used, it is necessary to form a photoelectric conversion layer at a temperature lower than the upper temperature limit of the resin, and a temperature of up to about 400⁰C is the limit of the process. It is difficult to form a photoelectric conversion layer of high properties at this temperature, so that various efforts are made such as providing an energy assisting layer and the like. The present invention is applicable to such a low temperature process photoelectric conversion device. But, the present invention is more advantageously applicable to a high temperature process in which a high heat tolerance is required, more specifically, to a process of not less than 470⁰C. [Examples]

Examples and Comparative Examples according to the present invention will now be described.

(Manufacture of Photoelectric Conversion Device Substrate, SUB 1 to SUB 15) <Manufacture of Metal Base>

A total of 15 types of Al rolled plates (each with an Al purity of not less than 98.0% by mass) were obtained by changing the amount of minor components and conditions of casting, rolling, and process annealing. Each of the obtained Al substrates was subjected to rolling oil removal and desmutting in a water solution of 30% by mass of H2SO4, and the surface is polished in the following three steps. Each Al substrate is cut into a 5cm square sample and subjected to polishing by sticking to a mirror finished metal block by a double sided tape.

1) Mechanical Polishing by Sandpaper

Polishing Machine: Marumoto Struers K.K, Trade Name: LAPO5 Sandpaper: Marumoto Struers K.K, Waterproof Sand Paper

The sandpaper is attached to the polishing machine and rotated, and each Al substrate (5cm square sample) is brought into contact with the sandpaper to polish the surface. The polishing was performed by gradually increasing the sandpaper count, like #80 → #240 → #500 → #1000 → #1200 → #1500 until surface unevenness is not visually recognizable.

2) Mechanical Polishing by Diamond Slurry Polishing Machine: Ditto

Abrasive Cloth: Marumoto Struers K.K, Polishing Cloth No. 773 (abrasive powder diameter of not smaller than lOum) , No. 751 (abrasive powder diameter of less than lOμm) Abrasive Powder: Marumoto Struers K.K, Diamond Abrasive DP-Spray P

The polishing cloth is attached to the polishing machine and rotated while the abrasive powder is supplied thereto, and each

Al substrate (5cm square sample) subjected to the mechanical polishing by sandpaper was brought into contact with the cloth to polish the surface. The polishing was performed by changing the polishing power in the following manner until surface unevenness is not visually recognizable. The Polishing cloth was replaced each time the polishing power was changed.

SPRIR (particle diameter of 45um) — SPRAM (particle diameter of

25um) → SPRUF (particle diameter of 15um) → SPRAC (particle diameter of 9um) — SPRIX (particle diameter of 6um) — SPRRET (particle diameter of 3um) → SPRON (particle diameter of lum)

→ SPRYT (particle diameter of 0.25um) 3) Electropolishing

Electrolyte: Mixed water solution of phosphoric acid, sulfuric acid, ethylene glycol, monoethyl ether, and water. Temperature: 50⁰C Duration: 5 min Energization Condition: DC 15V

Through the three-step polishing described above, the surface of each Al substrate was finished with a residual swell density of 0/dm², a surface roughness Ra of O.lum, an average glossiness of 75%.

The surface roughness Ra was assessed in the middle of the polishing process with the following means.

First, surface roughness Ra was measured by a JIS-B601-1994 compliant stylus roughness meter and by an AFM when the surface roughness Ra was less than O.lum.

The Ra measurement conditions by the roughness meter are listed below.

Model: Surfcom 575A, Tokyo Seimitsu Co., Ltd Measurement Conditions : cutoff, 0.8mm; tilt correction, FLAT-ML; measurement length, 2.5mm; T-speed, 0.3mm/s; polarity, positive Measuring stylus: sapphire stylus with a tip diameter of lOμm A value (μm) obtained by folding a roughness curve acquired by the measurement at the center line and dividing the area obtained by the roughness curve and center line by the length L is Ra. The Ra measurement by AFM was performed in DFM, cyclic contact mode under the following conditions. Scanning Area: 3000nm Scanning Frequency: 0.5Hz Amplitude Attenuation Rate: -0.16 I-Gain: 0.0749/P-Gain: 0.0488 Q-Curve Gain: 2.00 Vibration Voltage: 0.044V Resonance Frequency: 318.5 KHz Measuring Frequency: 318.2 KHz Vibration Amplitude: 0.995V Q-Value: near 460

Measuring Stylus: Si stylus with a tip diameter of lOnm (Seiko Instruments Inc., Trade Name: Cantilever SI DF40P)

Substrate manufacturing conditions, Al purities, minor component contents, and numbers of Fe-containing clusters in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to

2.5um in a cross-section of substrates are shown in Table 1. In the table, values of Fe content and the number of Fe-containing clusters outside of the defined ranges of the present invention are denoted by the mark "x".

Minor component contents were measured by the photoelectric emission spectrochemical analysis method defined in JIS H 1305. Numbers of Fe-containing clusters were obtainedby an analysis device of integrated SEM/EDX with an acceleration voltage of 15 kV. Measured examples of SEM cross-section photo and EDX chart are shown in Figures 5A and 5B respectively. The EDX charts show data of Fe rich portion 004 and Fe poor portion 007 in the SEM cross-section photo respectively. <Anodization> An anodized film was formed on each side of each Al substrate (0.30 mm thick) obtained in the manner as described above by DC anodization in an electrolyte of 10% by mass of sulfuric acid with a temperature kept at 23⁰C. The amount of current applied to the Al substrate is control so as to have a profile in which the amount of current just after the start of the anodization is 0.02 A/cm² and then gradually increased up to 0.20 A/cm². The anodized film was formedwith a thickness of 9.0um (including 0.38μmof a barrier layer) and a pore diameter of about lOOnm. In this way, photoelectric conversion device substrates SUB 1 to SUB 15 were obtained.

A terminal of a voltage regulator is connected to a point of a test piece of each of photoelectric conversion device substrates SUB 1 to SUB 15, then a mercury grain was placed on the insulation film (anodized film) , and a cupper wire end with the other end connected to the voltage regulator was put in the mercury grain. In this configuration, application voltage is increased to induce a dielectric breakdown, whereby the withstand voltage was measured. The results are shown in Table 1.

(Examples 1-1 to 1-8, Comparative Examples 1-1 to 1-8)

In each example, a photoelectric conversion device was produced under the same conditions other than using a different type of photoelectric conversion device substrate. Substrates used in the respective examples are shown in Table 2. As the substrates, a 0.50mm thick soda lime glass substrate generally used as a solar cell substrate is used in addition to SUB 1 to SUB 15 obtained in the above.

Mo, NaF, and Mo layers were sequentially formed on each photoelectric conversion device substrate by RF sputtering (radio-frequency sputtering) , whereby a lower electrode having such laminated structure was formed. The overall thickness of the lower electrode was l.Oμm. Note that the NaF layer was not formed on the substrate of soda lime glass, because Na is included in the substrate. After forming the lower electrode, scribing was performed to form first separation grooves.

Next, as a photoelectric conversion layer, a two-layer structure Cu(Ini-_xGaχ)Se₂ thin film was formed on the lower electrode by multi source simultaneous deposition. The deposition of the Cu (Ini-_xGa_x) Se₂ thin film was performed under a vacuum degree of about 10^~4 Pa (10^~7 Torr) by providing Cu, In, Ga, and Se deposition sources in a vacuum vessel. Here, the temperature of the deposition crucible was controlled appropriately.

The first layer was formed such that the Cu atomic composition becomes excessive with respect to the total atomic composition of In and Ga, and the second layer was formed such that the total atomic composition of In and Ga becomes excessive with respect to the Cu atomic composition. The substrate temperature was maintained at 530⁰C. The thickness of the first layer was 2μm. The composition ratio (molar ratio) of the first layer was Cu/ (In + Ga) = about 1.0 to 1.2. Then, the second layer was formed to a thickness of lμm such that the final composition ratio (molar ratio) becomes Cu/ (In + Ga) = 0.8 to 0.9.

Then, as a buffer layer, a semiconductor film of laminated structure was formed. First, a 50nm thick CdS film was deposited by chemical deposition. The chemical deposition was performed by heating a water solution including cadmium nitrate, thiourea, and ammonia to about 80⁰C and immersing the photoelectric conversion layer in the solution. Further, a ZnO film of about 80nm thickness was formed on the CdS film by MOCVD method. After forming the buffer layer, scribing was performed on the layer stack of the photoelectric conversion layer and buffer layer to form second separation grooves .

Next, as an upper electrode, an Al doped ZnO film of about

500nm thickness was deposited by MOCVD method. After forming the upper electrode, scribing was performed on the upper electrode to form third separation grooves. Further, fourth separation grooves were formed by performing scribing on the layer stack of the photoelectric conversion layer, buffer layer, and upper electrode.

Thereafter, as drawing-out external electrodes, Al was deposited, whereby a photoelectric conversion device was obtained. Finally, a transparent resin for sealing was laminated, whereby a solar cell module was obtained. A total of 20 solar cell modules were produced under the same condition. Each module has a structure in which three cell units, each having 24 cells connected in series, are connected in parallel.

<Photoelectric Conversion Efficiency and Yield Rate Evaluations> Photoelectric conversion efficiency was evaluated for each produced solar cell module using pseudo sunlight of Air Mass (AM) = 1.5, lOOmW/cm². Photoelectric conversion efficiency was measured for 20 samples, and those having photoelectric conversion efficiency of 80% or more of a maximum value among them were evaluated as acceptable products and those other than the acceptable products were evaluated as unacceptable products. Then, an average value of photoelectric conversion efficiency of the acceptable products was obtained as the photoelectric conversion efficiency. Further, the yield rate was obtained by the formula below.

Yield Rate = number of acceptable products/total number of evaluated samples (%) . (Examples 2-1 to 2-8, Comparative Examples 2-1 to 2-8) Photoelectric conversion devices were produced and evaluated in the similar manner to that of Examples 1-1 to 1-8 and Comparative Examples 1-1 to 1-8, other than forming the photoelectric conversion layers by selenization. Substrates used and evaluation results are summarized in Table 3. Formation of each photoelectric conversion layer was performed in the following manner. That is, a stacked film of (Cu-Ga) layer/In layer was formed by sputtering such that the composition rate (molar rate) of total Cu/ (In + Ga) becomes about 0.9. Then, the substrate is heated to 470 to 480°C under a selenium vapor atmosphere to induce thermal diffusion, whereby a photoelectrical conversion layer with a composition substantially corresponding to Cu(Ini__xGa_x)Se2 was obtained. (Evaluation Results)

As Table 1 clearly shows, a photoelectrical conversion device substrate using an Al substrate of a high Al purity with small amounts of minor components has a high withstand voltage, and the withstand voltage tends to decrease as the amounts of minor components are increased. From the viewpoint of only the withstand voltage, it is preferable to increase the purity of Al substrate as much as possible. But, as shown in Tables 2 and 3, if a photoelectric conversion device using an Al substrate produced by putting priority on high purity is exposed to a high temperature in the manufacturing process, the photoelectric conversion efficiency of the device is degraded with a low yield rate, that is, a quality stable photoelectric conversion device is not produced. More specifically, Comparative Examples 1-1 to 1-3 and 2-1 to 2-3 using photoelectric conversion device substrates SUB 1 to SUB 3 of Al substrate with an Al purity of 99.99% by mass have low photoelectric conversion efficiency and a low yield rate.

Also, Comparative Examples 1-4 to 1-7 and 2-4 to 2-7 using photoelectric conversion device substrates SUB 12 to SUB 15 of Al substrate with excess Fe content and number of Fe-containing clusters have low photoelectric conversion efficiency and a low yield rate.

Photoelectric conversion device substrates SUB 4 to SUB 11 using an Al substrate having a Fe content of 0.05 to 1.0% by mass and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5um, in a cross-section of the Al substrate of 1,500 to 40,000/mm² have relatively high withstand voltages. Further, Examples 1-1 to 1-8, and 2-1 to 2-8 in which photoelectric conversion devices were manufactured using SUB 4 to SUB 11 respectively have high photoelectric conversion efficiencies even exposed to a high temperature in the manufacturing process. Examples 1-1 to 1-8 and 2-1 to 2-8 may provide high photoelectric conversion efficiencies of 12 to 16%.

Examples 1-1 to 1-8 in which photoelectric conversion layer were formed by multi source simultaneous deposition may provide high photoelectric conversion efficiencies of not less than 14%. This result is obtained by forming the photoelectric conversion layers with a substrate temperature of 530⁰C, and advantageous effects of the present invention over the comparative examples became more significant when the substrate temperature was increased to 55O⁰C. Examples 2-1 to 2-8 in which photoelectric conversion layers were formed by selenization may also provide high photoelectric conversion efficiencies of not less than 12%. This result is obtained by forming the photoelectric conversion layers with a substrate temperature of 470 to 48O⁰C when heating in the presence of the group VI element, and advantageous effects of the present invention over the comparative examples became more significant when the substrate temperature was increased to 500 to 51O⁰C. From these results, it may well be said that the high photoelectric conversion efficiencies and yield rates of the present invention are a result of improved stability of the substrates in a high temperature environment.

A weight comparison between solar cells formed of example photoelectric conversion devices and solar cells using soda lime glass as the substrates resulted in that the former was 390 g/m² while the latter was 1.3 kg/m², showing that a significant weight reduction has been achieved. Further, example photoelectric conversion devices are solar cells using Al substrates with a thickness of about 300um, so that they are also superior, in flexibility and robustness, to solar cells with a general glass substrate.

O O

O CS

H U Sh

C5

CS

<o

CS

30

O

O CS

O

[Table 2]

Multi Source Simultaneous Deposition (53O⁰C)

Substrate P/E Conv. Eff. (%) Y/Rate (%)

C/E 1-1 SUBl 5.8 40

C/E 1-2 SUB2 6.2 45

C/E 1-3 SUB3 8.3 60

EG 1-1 SUB4 14.2 75

EG 1-2 SUB5 15.8 85

EG 1-3 SUB6 15.3 85

EG 1-4 SUB7 14.9 90

EG 1-5 SUB8 15.5 95

EG 1-6 SUB9 15.0 90

EG 1-7 SUBlO 14.4 85

EG 1-8 SUBIl 14.1 80

C/E 1-4 SUB12 9.5 75

C/E 1-5 SUBl3 9.0 75

C/E 1-6 SUB14 10.0 70

C/E 1-7 SUBl5 9.0 55

C/E 1-8 Soda/G 14.6 90

[Table 3] Selenization (470°O480°C)

Claims

1. A photoelectric conversion device, comprising a substrate of Al based metal base having an anodized film on at least one surface side on which a photoelectric conversion layer which includes a compound semiconductor formed of a group Ib element, a group IHb element, and a group VIb element and generates a current by absorbing light, and electrodes for drawing out the current are provided, wherein: the metal base has a Fe content of 0.05 to 1.0% by mass; and the number of Fe-containing clusters, in which a minimum diameter is not less than 0.3μm and a sum of minimum and maximum diameters divided by 2 falls within the range from 0.5 to 2.5um, in a cross-section of the metal base is 1,500 to 40,000/mm².

2. The photoelectric conversion device of claim 1, wherein the metal base has an Al content of not less than 98.0% by mass, a Si content of not greater than 0.25% by mass, and a Cu content of not greater than 0.20% by mass.

3. The photoelectric conversion device of claim 1 or 2, wherein the photoelectric conversion layer includes a compound semiconductor formed of: at least one type of group Ib element selected from the group consisting of Cu and Ag; at least one type of Ilib element selected from the group consisting of Al, Ga, and In; and at least one type of VIb element selected from the group consisting of S, Se, and Te.

4. A solar cell, comprising the photoelectric conversion device of any of claims 1 to 3.