JP2013172072A - Two-junction solar cell - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable constituting a two-junction solar cell by combination of a solar cell using a group III-V compound semiconductor and a solar cell using Si, without including As.SOLUTION: A two-junction solar cell includes a first solar cell 121 including a p-type silicon substrate 101 and an n-type silicon layer 102 provided on the silicon substrate 101 and composed of n-type silicon. Moreover, the two-junction solar cell includes a second solar cell 123 including a first semiconductor layer 103 composed of a p-type group III-V compound semiconductor configured without including As, a multiquantum well structure 122 including a barrier layer 104 composed of InGaP and a quantum well layer 105 composed of InGaP, and a second semiconductor layer 106 composed of an n-type group III-V compound semiconductor configured without including As.

Description

  The present invention relates to a two-junction solar cell manufactured using a III-V compound semiconductor and silicon.

  In order to avoid serious energy problems, solar cells, which are clean energy sources without emissions such as carbon dioxide, are drawing attention. Since the solar cell is also highly safe, it is required to increase the ratio of electric power generated by the solar cell. For this reason, studies are underway to improve the photoelectric conversion efficiency of solar cells, that is, the conversion efficiency from sunlight to electricity.

  Among solar cells currently in commercial use, the amount of power generation is the largest, and the manufacturing technology is advanced in solar cells using Si. However, a high photoelectric conversion efficiency exceeding 30% cannot be obtained in a solar cell (single-junction solar cell) configured by only one pn junction using Si. This is due to the relatively small band gap of Si, about 1.1 eV. FIG. 10 is a distribution diagram showing the distribution of sunlight on the ground surface. The spectrum of sunlight is distributed around a wavelength of about 0.6 μm (2.05 eV in terms of energy).

  On the other hand, the band gap of Si is about 1.1 eV as described above, and can absorb light having a wavelength shorter than 1.1 μm, so that a considerable proportion of sunlight can be absorbed. However, although sunlight absorbed by Si has energy larger than the band gap of Si, carriers (electrons, holes) generated by light absorption of sunlight in Si are within the crystal. Since energy is lost due to heat or the like, finally, only a voltage corresponding to the band gap of Si can be generated. This is a major factor that a high photoelectric conversion efficiency cannot be obtained with a solar cell using Si as described above.

  On the other hand, solar cells using III-V group compound semiconductors have achieved more than 30%, and have been put to practical use for satellite mounting. In a solar cell using a group III-V compound semiconductor, by making a structure (multi-junction solar cell) in which two or more pn junctions (solar cell) made of different materials are laminated, the energy of the solar spectrum is increased. Accordingly, light can be absorbed in multiple stages. In the multi-junction solar cell using the group III-V compound semiconductor, a material having a larger band gap than Si can be used, and the energy of light can be efficiently converted into electric energy.

  A solar cell using a group III-V compound semiconductor is considered to be applied not only to a satellite but also to the ground. In the application to the ground, the manufacturing cost of the solar cell module per power generation amount becomes a problem. In general, it is known that a solar cell can improve photoelectric conversion efficiency by collecting sunlight. For this reason, in this concentrating solar cell, the material cost of the semiconductor to be used can be greatly reduced. It is known that the solar cell using a III-V group compound semiconductor is larger in the photoelectric conversion efficiency by condensing the solar cell than the solar cell using Si. For this reason, by using a concentrating solar cell module, the production cost per power generation amount is about the same as that of a solar cell using Si, even if it uses a III-V group compound semiconductor. It is thought to be possible.

H. Wada et al., "Effects of heat treatment on bonding properties in InP-to-Si direct wafer bonding", Japanese Jourrnal of Applied Physics, Vol. 33, Part 1, No.9A, pp.4878-4879, 1994 . M. Howlader et al., "Characterization of the bonding strength and interface current of p-Si / n-InP wafers bonded by surface activated bonding method at room temperature", Jourrnal of Applied Physics, Vol.91, No.5, pp .3062-3066, 2002. Tanabe et al., “AlGaAs / Si two-junction solar cell by wafer fusion”, 71st Japan Society of Applied Physics (2010, Nagasaki University), 15p-NC-4. S. Kurtz et al., "Modeling of two-junction series-connected tandem solar cells using top-cell thickness as an adjustable parameter", Jourrnal of Applied Physics, Vol.68, No.4, pp.1890-1895, 1990 . A. Bett et al., "III-V compounds for solar cell applications", Applied Physics A, Vol.69, pp.119-129, 1999. D. Bushnell et al., "Effect of well number on the performance of quantum-well solar cells", Jourrnal of Applied Physics, Vol.97, No.12, 124908, 2005. N. Tansu et al., "High-performance strain-compensated InGaAs-GaAsP-GaAs (l = 1.17μm) quantum-well diode lasers", IEEE Photonics Technology Letters, Vol.13, No.3, pp.179-181 , 2001. H. Asai et al., "Energy band-gap shift with elastic strain in GaxIn1-xP epitaxial layers on (001) GaAs substrates", Jourrnal of Applied Physics, Vol.54, No.4, pp.2052-2056, 1983 . C. Silfvenius et al., "Design, growth and performance of different QW structures for improved 1300 nm InGaAsP lasers", Journal of Crystal Growth, Vol.195, pp.700-705, 1998.

  As described above, a solar cell using Si has a band gap that can absorb a large amount of sunlight, but it is difficult to increase the photoelectric conversion efficiency because the band gap of Si is small. On the other hand, in the multijunction solar cell using the III-V group compound semiconductor as described above, high photoelectric conversion efficiency is realized by combining cells having different band gaps.

  In the two-junction solar cell, the top cell and the bottom cell can be connected in series with a small electric resistance via a tunnel junction. In the two-junction solar cell, there are a two-terminal type in which electrodes are provided on the upper part of the top cell and a lower part of the bottom cell, and a four-terminal type in which electrodes are provided above and below each of the top cell and the bottom cell. Compared to the two-terminal type, the four-terminal type has a complicated manufacturing process and increases the manufacturing cost. For this reason, a two-terminal solar cell is often used in a two-junction solar cell.

  FIG. 11 is a cross-sectional view schematically showing the structure of a two-junction solar cell that is a multi-junction solar cell. FIG. 11 shows a two-terminal type two-junction solar cell, which includes a bottom cell 1101 and a top cell 1103 joined to the surface of the bottom cell 1101 through a tunnel junction 1102. On the top cell 1103, an antireflection film 1104 for making light incident efficiently is formed. A surface finger electrode 1106 is formed on a part of the top cells 1103 via a contact layer 1105. On the other hand, a back electrode 1107 is formed on the back side of the bottom cell 1101.

  In two-junction solar cells using III-V group compound semiconductors that have been studied so far, most of the bottom cell light absorption layer is made of GaAs. In a multi-junction solar cell having three or more junctions, investigation using Ge as a bottom cell is also in progress. In this case, GaAs or InGaAs containing a small amount of In is used as a material for a middle cell. In a multi-junction solar cell using a III-V group compound semiconductor realized so far, a material mainly containing As as a group V element such as GaAs, InGaAs, or AlGaAs is used as a light absorption layer in any of the constituent cells. It is used.

  In increasing the ratio of electric power energy by solar cells, it is necessary to consider not only the improvement of photoelectric conversion efficiency but also safety including disposal of solar cells after use. As described above, in a solar cell using a III-V group compound semiconductor, although high photoelectric conversion efficiency can be obtained, a material having As as an environmental load substance as a main group V element is used.

  In a solar cell using a group III-V compound semiconductor, the area of the solar cell itself can be reduced by making it a concentrating type, but even in this case, in order to increase the amount of power generation, 1 mm per module. An element to be a solar cell having a size of about × 1 mm to 10 mm × 10 mm is required. This is an order of magnitude larger than semiconductor lasers and electronic devices that are currently in practical use. For this reason, a solar cell made of a material having As as a main group V element contains an amount of As that cannot be compared with a semiconductor laser or an electronic device.

  As described above, in consideration of safety including the disposal of solar cells, from the viewpoint of preventing the diffusion of environmentally hazardous substances, a solar cell composed of a material that does not contain As or has a low As content. Realization is desired. However, as described above, in a multi-junction solar cell using a group III-V compound semiconductor, a material containing As as a main group V element such as GaAs, InGaAs, or AlGaAs is used in any of a plurality of cells. .

  In the two-junction solar cell, as the layer structure in which the material that does not contain As in either the top cell or the bottom cell is used as the absorption layer, Si is used for the bottom cell, and the top cell does not contain As such as InGaP, InGaAlP, InAlP. A configuration using a group V compound semiconductor is conceivable. Conventionally, it has been difficult to realize a semiconductor device in which Si and a III-V group compound semiconductor are combined. However, due to recent progress in semiconductor process technology, Si and a III-V group compound semiconductor are not connected via a metal or a dielectric. It is possible to join directly.

  This bonding technique is also referred to as “Direct Wafer Bonding”, “Wafer Fusion”, or the like, and can directly bond Si and various III-V compound semiconductors. As a method of direct bonding, first, a hydroxyl group (expressed as —OH, also referred to as a hydroxyl group) is attached to the wafer surface, two wafers are adsorbed through this, and then bonded by thermal annealing. is there. Also, there is a method of bonding wafers in a vacuum after activating each wafer surface with Ar plasma.

  In any method, it is known that the joint surface between Si and the III-V compound semiconductor can be electrically connected, and current can be passed between Si and the III-V compound semiconductor (for example, Non-Patent Document 1). Non-Patent Document 2). This direct bonding technology of Si and III-V compound semiconductor is known to be useful in the production of solar cells, and is also applied to the bonding of top cells using AlGaAs and bottom cells using Si. (See Non-Patent Document 3).

  A method of manufacturing a two-junction solar cell that does not include As using this direct bonding will be described with reference to FIGS. 1212A to 12D. 12A to 12D are explanatory diagrams for explaining a process of manufacturing a two-junction solar cell by directly bonding a semiconductor solar cell using a Si solar cell and a III-V group compound semiconductor. First, as shown to FIG. 12A, the Si photovoltaic cell 1201 used as a bottom cell is produced. Further, as shown in FIG. 12B, a solar battery cell 1203 made of a III-V group semiconductor grown on a GaAs substrate 1202 serving as a top cell is manufactured.

  Next, as shown to FIG. 12C, the photovoltaic cell 1203 and the Si photovoltaic cell 1201 are bonded together and joined. Next, in order to electrically connect Si and the III-V group compound semiconductor, after performing a treatment such as heating as necessary, for example, using selective etching or the like, as shown in FIG. The GaAs substrate 1202 is removed from 1203. Through the above steps, a two-junction solar cell can be manufactured. In addition, when the solar battery cell 1203 is made of a III-V group compound semiconductor that does not contain As, a two-junction solar battery can be manufactured without containing arsenic.

In the two-junction solar cell, it is known that the photoelectric conversion efficiency varies greatly depending on the combination of the band gap of the light absorption layer of the bottom cell and the band gap of the light absorption layer of the top cell. A change in photoelectric conversion efficiency due to a combination of band gaps can be estimated by calculation (see, for example, Non-Patent Document 4 and Non-Patent Document 5). FIG. 13 is a characteristic diagram showing a change in photoelectric conversion efficiency due to the band gap between the bottom cell and the top cell in a two-junction solar cell obtained by calculation. As incident light, it is assumed that air mass 1.5 direct (AM-1.5Direct, incident light intensity 768 W / m ² ) is condensed 500 times.

  As described above, the band gap of Si serving as the bottom cell is about 1.1 eV. From FIG. 13, when Si is used for the bottom cell, the band gap of the top cell that maximizes the photoelectric conversion efficiency is around 1.65 eV, and if this band gap is set in the range of 1.65 eV to 1.85 eV, It can be seen that a photoelectric conversion efficiency of 36% or more can be obtained under 500 times light collection. The absolute value of the photoelectric conversion efficiency obtained by calculation is the physical property parameters such as the intensity of the incident sunlight, the absorption spectrum in the light absorption layer of the cell, the mobility of carriers (electrons and holes), lifetime, and effective mass. It will change greatly. However, the band gap range of the top cell where high photoelectric conversion efficiency can be expected does not change greatly even if these physical property parameters change.

  A structure using InGaP, InGaAlP, InAlP, or the like is conceivable as a top cell that does not contain As in the above-described two-junction solar cell. Since these materials can be lattice-matched with GaAs and a high-quality substrate such as a GaAs substrate can be used, crystal growth is easy. Among InGaP, InGaAlP, and InAlP, the material with the smallest band gap is InGaP, but the band gap when InGaP is lattice-matched with GaAs is about 1.9 eV. On the other hand, in order to obtain a large photoelectric conversion efficiency in a two-junction solar cell using Si as the bottom cell as described above, it is necessary to use a material having a band gap of 1.85 eV or less as the top cell. There are the following two methods for reducing the band gap of InGaP for crystal growth on GaAs.

  One method is a method of reducing the band gap by using a phenomenon called ordering in which In and Ga, which are Group III elements constituting InGaP, are regularly arranged. This ordering phenomenon can be caused by changing the crystal growth conditions of InGaP. However, the band gap of InGaP can only be reduced to about 1.85 eV even if ordering is used, and it is difficult to make it less than this.

  Another method is to apply compressive strain to InGaP. FIG. 14 is a characteristic diagram of the relationship between band gap and lattice strain obtained by calculation for InGaP grown on GaAs. In FIG. 14, the horizontal axis indicates that compressive strain is applied to InGaP when the lattice strain for GaAs is positive (+), and tensile strain is applied when the strain is negative (−). As shown in FIG. 14, it can be seen that the band gap can be made smaller than 1.85 eV by applying a compressive strain of 0.4% or more to InGaP.

  However, in InGaP to which compressive strain is applied, it becomes difficult to increase the film thickness as the compressive strain increases. This is because strain stress applied to InGaP accumulates as the film thickness increases, and crystal strain due to misfit transition occurs due to this strain stress. The maximum film thickness at which this misfit transition does not occur is generally called a critical film thickness and can be estimated by calculation. Although this critical film thickness is different from the model used for calculation, in InGaP to which 0.4% of compressive strain is applied, it is 100 nm or less at most.

  On the other hand, in order to obtain high photoelectric conversion efficiency in a two-junction solar cell in which Si is the bottom cell and InGaP is the top cell, the current value generated in the top cell is matched with the current value generated in the Si bottom cell, that is, It is necessary to match the current. Since the current generated in the top cell increases as the thickness of the light absorption layer increases, the film thickness of InGaP, which is the light absorption layer of the top cell, is 100 nm even if it is thin to match the current with the Si bottom cell. The above is necessary.

  However, as described above, it is difficult to grow InGaP to which a compressive strain of 0.4% or more is applied to a film thickness of 100 nm or more without generating crystal defects. As described above, the configuration using InGaP formed on the GaAs substrate by applying compressive strain is effective in reducing the band gap, but has a problem that it cannot be applied to an actual solar cell. .

  This invention is made | formed in order to eliminate the above problems, and the 2 junction solar cell by the combination of the photovoltaic cell using a III-V group compound semiconductor and the photovoltaic cell using Si is provided. It is an object to enable configuration without including As.

  A two-junction solar cell according to the present invention is composed of a first solar cell composed of a p-type silicon layer and an n-type silicon layer, and a p-type III-V compound semiconductor composed of no As. A first semiconductor layer, a multiple quantum well structure including a barrier layer made of InGaP and a quantum well layer made of InGaP, and a second semiconductor layer made of an n-type III-V group compound semiconductor that does not contain As At least a second solar battery cell, and a bonding layer disposed between the first solar battery cell and the second solar battery cell.

  In the above-described two-junction solar cell, the band gap energy of the quantum well layer may be in the range of 1.65 eV to 1.85 eV. The quantum well layer may be in a state having a compressive strain in the range of 0.5 to 2% with respect to GaAs. Moreover, the barrier layer should just have a tensile strain with respect to GaAs.

  As described above, according to the present invention, since a multiple quantum well structure made of InGaP is used, the combination of a solar battery cell using a III-V group compound semiconductor and a solar battery cell using Si is used. An excellent effect that the two-junction solar cell can be configured without containing As is obtained.

FIG. 1 is a configuration diagram showing the configuration of the two-junction solar cell in the first embodiment of the present invention. FIG. 2 is a characteristic diagram showing changes in the band gap due to lattice strain applied to the quantum well layer 105. FIG. 3 is a band diagram schematically showing conduction band discontinuity in a strain compensated quantum well structure in which compressive strain is applied to the quantum well layer 105 and tensile strain is applied to the barrier layer 104. FIG. 4 is a characteristic diagram showing changes in band discontinuity due to lattice strain applied to the quantum well layer. FIG. 5 is a characteristic diagram showing changes in band discontinuity due to lattice strain applied to the quantum well layer. FIG. 6 shows a photoluminescence emission spectrum when the number of quantum well layers in the multiple quantum well is changed. FIG. 7 is a characteristic diagram showing photoluminescence emission spectra of Sample 1 and Sample 2 having different effective strains at room temperature at room temperature. FIG. 8 is a configuration diagram showing a partial configuration of the two-junction solar cell in the third embodiment of the present invention. FIG. 9 is a configuration diagram showing a configuration of a solar cell manufactured for the result of a characteristic investigation of the second solar cell constituting the two-junction solar cell in the third embodiment. FIG. 10 is a distribution diagram showing the distribution of sunlight on the ground surface. FIG. 11 is a cross-sectional view schematically showing the structure of a two-junction solar cell that is a multi-junction solar cell. FIG. 12A is an explanatory diagram for explaining a process of manufacturing a two-junction solar cell by directly bonding a Si solar cell and a semiconductor solar cell using a group III-V compound semiconductor. FIG. 12B is an explanatory diagram for explaining a process of manufacturing a two-junction solar cell by directly bonding a semiconductor solar cell using a Si solar cell and a III-V group compound semiconductor. FIG. 12C is an explanatory diagram for explaining a process of manufacturing a two-junction solar cell by directly bonding a semiconductor solar cell using a Si solar cell and a III-V group compound semiconductor. FIG. 12D is an explanatory diagram for explaining a process of manufacturing a two-junction solar cell by directly bonding a semiconductor solar cell using a Si solar cell and a III-V group compound semiconductor. FIG. 13 is a characteristic diagram showing a change in photoelectric conversion efficiency due to the band gap between the bottom cell and the top cell in a two-junction solar cell obtained by calculation. FIG. 14 is a characteristic diagram of the relationship between band gap and lattice strain obtained by calculation for InGaP grown on GaAs grown on GaAs.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[Embodiment 1]
FIG. 1 is a configuration diagram showing the configuration of the two-junction solar cell in the first embodiment of the present invention. FIG. 1 schematically shows a cross section. The two-junction solar cell includes a p-type silicon substrate (p-type silicon layer) 101 and an n-type silicon layer 102 made of n-type silicon formed on the silicon substrate 101. A cell 121 is provided. Each silicon layer has a conductivity type by introducing a predetermined impurity by ion implantation or the like.

  In addition, a multiple quantum well structure 122 including a first semiconductor layer 103 made of a p-type III-V group compound semiconductor configured without containing As, a barrier layer 104 made of InGaP, and a quantum well layer 105 made of InGaP, And a second solar cell 123 made of the second semiconductor layer 106 made of an n-type III-V compound semiconductor and not containing As.

  Further, formed on the first junction layer 107 and the first junction layer 107 made of p-type III-V group compound semiconductor which is formed on the first semiconductor layer 103 and does not contain As. And a second bonding layer 108 made of an n-type III-V group compound semiconductor that does not contain As. The first bonding layer 107 and the second bonding layer 108 are doped with p-type and n-type impurities at a higher concentration than other layers, and form a known tunnel junction.

  In the above-described configuration, the n-type silicon layer 102 and the second bonding layer 108 are bonded to integrate the first solar cell 121 and the second solar cell 123, and the second junction is formed on the n-type silicon layer 102. The layer 108, the first junction layer 107, the first semiconductor layer 103, the multiple quantum well structure 122, and the second semiconductor layer 106 are stacked in this order to form a two-junction solar cell. The 1st photovoltaic cell 121 is a bottom cell, and the 2nd photovoltaic cell 123 is a top cell.

  Here, the band gap energy of the quantum well layer 105 may be in the range of 1.65 eV to 1.85 eV. The quantum well layer 105 preferably has a compressive strain in the range of 0.5 to 2% with respect to GaAs, and the barrier layer 104 preferably has a tensile strain with respect to GaAs.

  An electrode 113 is formed on a part of the second semiconductor layer 106 via a contact layer 112 made of an n-type III-V group compound semiconductor that does not contain As, and a silicon substrate. An electrode 114 is also formed on the back surface of 101. In addition, an antireflection film 111 is formed on the second semiconductor layer 106 where the electrode 113 serving as the sunlight incident portion is not formed.

  As described above, in the first embodiment, by using the multiple quantum well structure 122 made of InGaP, a band gap of 1.65 eV to 1.85 eV that is difficult to realize with a bulk InGaP layer formed on GaAs is realized. There is a feature.

  Hereinafter, it will be described that a band gap in a range of 1.65 eV to 1.85 eV can be realized in a strained multiple quantum well structure using InGaP as a well layer. Below, the point mentioned above is demonstrated with description of the manufacturing method example of the 2nd photovoltaic cell 123 mentioned above.

  First, on a GaAs substrate, a second semiconductor layer 106 made of n-InGaP (or p-InGaP) lattice-matched to GaAs, a multiple quantum well structure 122 made of a barrier layer 104 and a quantum well layer 105, p-InGaP ( Alternatively, a solar cell (second solar cell) in which the first semiconductor layer 103 made of n-InGaP is crystal-grown is manufactured. These can be produced by a known organometallic molecular beam epitaxy method.

  About the solar cell mentioned above, the band gap of the quantum well layer 105 was estimated by calculation. FIG. 2 is a characteristic diagram showing changes in the band gap due to lattice strain applied to the quantum well layer 105 when the quantum well layer 105 has three layer thicknesses of 5 nm, 7 nm, and 9 nm. The numbers in FIG. 2 indicate the layer thickness of the quantum well layer 105. The band gap of the quantum well layer 105 means the energy between the first levels of electrons and holes (heavy holes) quantized by the quantum size effect. The layer thickness of the barrier layer 104 is 12 nm, and the lattice strain is such that the effective strain ε represented by the following formula (1) becomes 0 for the entire multi-quantum well structure 122.

Here, ε _w is the lattice strain of the quantum well layer 105, ε _b is the lattice strain of the barrier layer 104, L _w is the layer thickness of the quantum well layer 105, and L _b is the layer thickness of the barrier layer 104. The effective strain indicates the lattice strain per cycle of the multiple quantum well structure 122 composed of the quantum well layer 105 and the barrier layer 104. If this effective strain is 0%, the number of quantum well layers 105 can be increased freely without causing misfit transition. As shown in FIG. 2, it is understood that the band gap of the quantum well layer 105 can be reduced to 1.85 eV or less by applying a compressive strain of 0.8% when the layer thickness of the quantum well layer 105 is 5 nm. . Even if the compressive strain of the quantum well layer 105 is constant at 0.8%, the band gap can be reduced to nearly 1.8 eV by increasing the layer thickness of the quantum well layer 105 to 5 nm, 7 nm, and 9 nm.

  Furthermore, as shown in FIG. 2, it can be seen that the band gap can be drastically reduced by increasing the compressive strain applied to the quantum well layer 105. The upper limit of the compressive strain that can be applied to the quantum well layer 105 depends on the crystal growth conditions, and thus it is difficult to define an exact value. However, in the strained quantum well structure on GaAs, the compressive strain is 2. A structure having a 5% quantum well layer 105 is also applied to a laser, and can grow if the compressive strain of the quantum well layer 105 is about 2% (see, for example, Non-Patent Document 7). In this case, the band gap can be reduced to close to 1.7 eV.

  The relationship shown in FIG. 2 is a result when a large tensile strain is applied to the barrier layer 104 in order to reduce the effective strain represented by the equation (1) to 0%. The tensile strain applied to 104 may be smaller. By reducing the tensile strain of the barrier layer 104, the band gap of the quantum well layer 105 can be made smaller than in the case of FIG. 2 even if the compressive strain applied to the quantum well layer 105 is the same. As a result, the band gap of the quantum well layer 105 can be reduced to 1.65 eV.

  As described above, the use of the multiple quantum well structure 122 in which the compressive strain InGaP is the quantum well layer 105 and the tensile strain InGaP is the barrier layer 104 is necessary to obtain a solar cell with high photoelectric conversion efficiency. A structure having a band gap of .65 eV or more and 1.85 eV or less and containing no As can be formed on GaAs.

  So far, a multiple quantum well structure using InGaP as a quantum well layer has been studied as an active layer for a laser used for reading a DVD (Digital Versatile Disc), but there is no idea of applying it to a solar cell. It was. Further, in the multiple quantum well structure using InGaP as a quantum well layer used for an active layer for a DVD laser, the barrier layer is InGaAlP, and InGaP is not used as the barrier layer. None of the layers was studied for a quantum well structure of InGaP itself.

  This is because the barrier layer used in the active layer for laser needs to have a sufficiently larger band gap than the quantum well layer in order to efficiently confine electrons in the quantum well layer. This is because the structure using InGaP has a small electron confinement in the quantum well layer and is not suitable for a laser active layer.

  On the other hand, in a solar cell, it is necessary to efficiently extract carriers (electrons and holes) generated by light absorption from the quantum well layer, and a barrier layer having a large band gap makes it difficult to extract carriers from the quantum well layer. . In the present invention, by using InGaP having a smaller band gap than InGaAlP for the barrier layer, carrier confinement in the quantum well layer is reduced, and carriers generated in the quantum well layer by photoexcitation of sunlight are efficiently It is easy to pull out. The small carrier confinement means that the band discontinuity of the conduction band and the valence band is small in the quantum well structure. Next, band discontinuity in this quantum well structure will be described.

  In a semiconductor layer to which lattice strain is added, it is necessary to consider changes in the band structure of the conduction band and the valence band. In particular, in the valence band, attention is necessary because energy at the top of each band is split between heavy and light holes due to lattice distortion (see Non-Patent Document 8). FIG. 3 shows a band discontinuity 301 in a conduction band and a band discontinuity 302 in a valence band (heavy hole) in a strain compensated quantum well structure in which a compressive strain is applied to the quantum well layer 105 and a tensile strain is applied to the barrier layer 104. FIG. 4 is a band diagram schematically showing a band discontinuity 303 of a valence band (light holes). In the quantum well layer 105 to which compressive strain is applied, the energy at the top of the heavy hole band in the valence band is larger than that of the light hole. On the contrary, in the barrier layer 104 to which tensile strain is applied, the band superposition energy of light holes becomes larger than that of heavy holes. FIG. 3 considers the change in the band structure due to the lattice strain.

  FIG. 4 shows a strain-compensated quantum well structure in which a quantum well layer made of InGaP having a layer thickness of 7 nm and a barrier layer made of InGaP having a thickness of 12 nm to which tensile strain is applied are set so that the effective strain becomes 0%. FIG. 5 is a characteristic diagram showing changes in band discontinuity due to lattice strain applied to the quantum well layer. As shown in FIG. 4, it can be seen that both the band discontinuity of the conduction band and the valence band is 100 meV or less when the compressive strain of the quantum well layer is 1%, and 200 meV or less even when the compressive strain is 2%. .

  In the quantum well structure, the upper limit of the band discontinuity at which carriers can move is largely dependent on the number of wells and the physical property parameters, so that it is difficult to limit numerically. In a laser structure in which an internal electric field is not applied to the quantum well structure, it is known that when the band discontinuity is about 130 meV, carriers can easily move between the quantum wells, and as a result, the laser characteristics do not deteriorate. (Refer nonpatent literature 9). On the other hand, in the solar cell, carriers can be moved more easily than the laser structure. This is because the solar cell has a PIN structure in which an undoped multiple quantum well structure 122 is sandwiched between a p-type first semiconductor layer 103 and an n-type second semiconductor layer 106 as described with reference to FIG. This is because an electric field is applied to the multiple quantum well structure 122. For this reason, in a solar cell, it is thought that a carrier can be extracted even if the band discontinuity of about 200 meV is larger than that of a laser.

FIG. 4 shows a band discontinuity in a strain compensated quantum well in which a tensile strain is applied to the barrier layer to reduce the effective strain to 0%, but a layer (In _0.48 Ga _0.52 P) having a composition in which the barrier layer is lattice-matched to GaAs. This band discontinuity can be further reduced. FIG. 5 is a characteristic diagram showing a band discontinuity change due to lattice strain applied to the quantum well layer when InGaP lattice-matched with GaAs is used for the barrier layer. In this case, even if the compressive strain of the quantum well layer is 2%, the band discontinuity is 120 meV, and carriers can be moved more easily than the strain compensation structure.

  As will be described later, in an actual multiple quantum well structure, the effective strain does not need to be 0%, and the barrier layer can be brought close to lattice matching, so that the band discontinuity as shown in FIG. Is possible. The strain amount of the InGaP quantum well layer and the InGaP barrier layer can be easily adjusted by changing the composition ratio of In and Ga in each layer, and can be easily adjusted by adjusting the raw material supply amount during crystal growth. Can be controlled.

Next, the result of actually producing the quantum well structure used in the two-junction solar cell in the first embodiment described with reference to FIG. 1 and confirming the band gap of the produced quantum well structure will be described. For the crystal growth, an organometallic molecular beam epitaxy method using trimethylindium (TMIn), triethylgallium (TEGa) as a group III source gas, and phosphine (PH ₃ ) or arsine (AsH ₃ ) as a group V source material is used.

  In the quantum well structure, n-GaAs having a layer thickness of 0.2 μm and n-InGaP having a layer thickness of 0.3 μm are grown on an n-type GaAs substrate, and subsequently a multi-quantum well structure including an InGaP barrier layer and an InGaP quantum well layer. Finally, p-InGaP having a layer thickness of 0.05 μm is grown at a growth temperature of 510 ° C. The quantum well layer is InGaP with a compressive strain of 1.1% and a layer thickness of 7.5 nm, and the barrier layer is InGaP with a tensile strain of 0.6% and a layer thickness of 12.5 nm.

  In this case, the effective strain represented by the equation (1) is + 0.04%, not 0%. FIG. 6 shows the case where the number of quantum well layers in the multiple quantum well is changed to 6, 10, 20 (the thickness of the multiple quantum well structure is about 0.13 μm, 0.21 μm, 0.41 μm, respectively). Fig. 2 shows a photoluminescence emission spectrum. The measurement was performed at room temperature (about 23 ° C.) using a laser having a wavelength of 532 nm. In this photoluminescence emission spectrum, the emission peak corresponds to the band gap of the InGaP quantum well layer.

  As shown in FIG. 6, it can be seen that a band gap of 1.76 eV can be obtained with any multiple quantum well structure. On the other hand, the intensity of the emission peak increases with the number of quantum well layers. This indicates that even if the effective strain is not completely 0%, the number of quantum well layers can be increased while suppressing the generation of crystal defects. In the second solar cell using this multiple quantum well structure, the layer thickness of the entire second solar cell required for matching the current with the first solar cell using silicon is 0.5 to 1 μm. Estimated with degree. The layer thickness of the entire second solar battery cell substantially matches the total value of the multiple quantum well structure and the upper and lower first semiconductor layers and second semiconductor layers.

  In the example described above, the total thickness of the upper and lower first semiconductor layers and the second semiconductor layers is 0.35 μm, and the layer thickness per cycle of the quantum well structure is 20 nm. In the case of 20, the thickness of the entire top cell is 0.5 to 0.8 μm. Thus, the multiple quantum well structure mentioned above is a structure applicable to the absorption layer of the 2nd photovoltaic cell of the 2 junction solar cell in this invention.

  In addition, although the case where the organometallic molecular beam epitaxy method was used as a manufacturing method of a multiple quantum well structure etc. was demonstrated above, it is not restricted to this. Needless to say, the applicable growth technique is that each compound semiconductor layer can be formed, and techniques such as metal organic vapor phase epitaxy, gas source molecular beam epitaxy, and molecular beam epitaxy may be used.

  Further, in the above description, the structure is exemplified only by forming a III-V group compound semiconductor layer in which an n-type impurity is introduced at a high concentration on a III-V group compound semiconductor layer in which a p-type impurity is introduced at a high concentration as a tunnel junction. However, it is not limited to this. For example, as shown in Non-Patent Document 3, a tunnel is formed by directly bonding a silicon layer in which an n-type impurity is introduced to an effect and a III-V group compound semiconductor layer in which a p-type impurity is introduced at a high concentration. You may make it form a junction and combine a 1st photovoltaic cell and a 2nd photovoltaic cell.

  Further, as shown in Non-Patent Document 4, the stacking order of the p-type layer and the n-type layer is reversed, and the n-type impurity is introduced at a high concentration on the III-V group compound semiconductor layer. A tunnel junction may be formed by stacking III-V compound semiconductor layers into which a p-type impurity having a concentration is introduced. What is necessary is just to provide and join the junction layer which comprises the tunnel junction as mentioned above between the 1st photovoltaic cell and the 2nd photovoltaic cell, and this junction has what is normal knowledge in this field, Many variations and combinations are possible.

[Embodiment 2]
Next, a second embodiment of the present invention will be described. In the first embodiment described above, the case where a multiple quantum well structure with an effective strain of + 0.04% is mainly exemplified, but if the number of quantum well layers is about 20, the effective strain is from 0%. In this case, the tensile strain applied to the barrier layer can be further reduced. A suitable example is shown below.

  Multiple quantum well structure sample 1 having 20 quantum well layers and an effective strain of + 0.24%, and multiple quantum well structure sample 2 having 20 quantum well layers and an effective strain of + 0.14% Are respectively prepared, and the photoluminescence emission spectra are compared. As described above, the metal organic molecular beam epitaxy method is used for the preparation of each multi-quantum well structure sample. In Sample 1, the barrier layer is made of InGaP having a tensile strain of 0.35% and a layer thickness of 13 nm. In Sample 2, the barrier layer is made of InGaP having a tensile strain of 0.44% and a layer thickness of 13 nm.

  In sample 1 and sample 2, in order to match the emission peak with that in the first embodiment described above, the compressive strain and layer thickness of the quantum well layer made of InGaP are 1.4% and 7 nm in sample 1, respectively. Sample 2 has 1.2% and 7.5 nm.

  FIG. 7 is a characteristic diagram showing photoluminescence emission spectra at room temperature of Sample 1 and Sample 2 made of multiple quantum well structures with different effective strains produced as described above. There is no significant difference in the intensity of the emission peak between the two samples. Thus, in the multiple quantum well structure using InGaP, even if the tensile strain of the barrier layer is small and the effective strain in the quantum well structure is larger than 0%, the emission intensity caused by crystal defects is reduced. It is possible to suppress. For this reason, the barrier layer is required to have a strain in the opposite direction to the quantum well layer, i.e., a tensile strain, but the tensile strain in the barrier layer depends on the compressive strain of the quantum well layer and the quantum well layer. It can be seen that this parameter can be determined as appropriate according to the number of

[Embodiment 3]
Next, Embodiment 3 of the present invention will be described with reference to FIG. FIG. 8 is a configuration diagram showing a partial configuration of the two-junction solar cell in the third embodiment of the present invention. In FIG. 8, the structure in the middle of manufacture of a 2nd photovoltaic cell is shown with typical sectional drawing.

  In the second solar cell, first, on a substrate 801 made of n-type GaAs, a buffer layer 802 made of n-type GaAs and having a thickness of 0.2 μm, a contact layer 803 made of n-type InGaP, and an n-type contact layer 803. A field layer 804 having a layer thickness of 0.03 μm made of InGaAlP and a second semiconductor layer 805 having a layer thickness of 0.4 μm made of n-type InGaP are stacked. Further, a multiple quantum well structure 806 including a barrier layer made of InGaP and a quantum well layer made of InGaP is formed on the second semiconductor layer 805. The multiple quantum well structure 806 is composed of a quantum well layer made of InGaP having a compressive strain of 1.5% and a thickness of 6 nm, and a barrier layer made of InGaP having a tensile strain of 0.3% and a thickness of 13 nm. The number of quantum well layers is 12. The band gap of the quantum well layer is 1.75 eV.

A first semiconductor layer 807 made of p-type InGaP and having a layer thickness of 0.05 μm is formed on the multiple quantum well structure 806, and a layer made of p-type InGaAlP is formed on the first semiconductor layer 807. An electric field layer 808 having a thickness of 0.03 μm is formed. Note that the electric field layer 804 is formed to suppress diffusion of electrons as carriers from the multiple quantum well structure 806 side to the electric field layer 804 side. The electric field layer 808 is formed in order to suppress diffusion of holes as carriers from the multiple quantum well structure 806 side to the electric field layer 808 side. Further, on the field layer 808, the second bonding layer 810 made of InGaP of the first bonding layer 809, and n ⁺ type consisting p ⁺ -type InGaP is formed. When the n-type silicon layer of the first solar battery cell made of Si is directly bonded to the second bonding layer 810 and the substrate 801 and the buffer layer 802 are removed, a two-junction solar cell can be formed.

  Here, the characteristic investigation result of the 2nd photovoltaic cell mentioned above is demonstrated. In this investigation, a solar cell having a layer structure shown in FIG. 9 was produced. First, on a substrate 901 made of n-type GaAs, a buffer layer 902 made of n-type GaAs and having a thickness of 0.2 μm, an electric field layer 903 made of n-type InGaAlP and having a thickness of 0.03 μm, and an n-type InGaP. A second semiconductor layer 904 having a thickness of 0.4 μm is stacked. Further, a multiple quantum well structure 905 including a barrier layer made of InGaP and a quantum well layer made of InGaP is formed on the second semiconductor layer 904. The multi-quantum well structure 905 includes a quantum well layer made of InGaP having a compressive strain of 1.5% and a thickness of 6 nm and a barrier layer made of InGaP having a tensile strain of 0.3% and a thickness of 13 nm. The number of quantum well layers is 12. The band gap of the quantum well layer is 1.75 eV.

  A first semiconductor layer 906 having a layer thickness of 0.05 μm made of p-type InGaP is formed on the multiple quantum well structure 905, and a layer made of p-type InGaAlP is formed on the first semiconductor layer 906. An electric field layer 907 having a thickness of 0.03 μm is formed. On the electric field layer 907, a window layer 908 made of p-type InAlP and having a layer thickness of 0.03 μm is formed. A surface finger electrode 910 is formed on a part of the window layer 908 via a contact layer 909 made of p-type InGaP. On the other hand, a back surface electrode 911 is formed on the back surface of the substrate 901. Further, an antireflection film 912 is formed outside the electrode formation region of the window layer 908.

Each semiconductor layer is formed by epitaxial growth by a gas source molecular beam epitaxy method using In, Ga, Al as a group III material, and phosphine (PH ₃ ) and arsine (AsH ₃ ) as a group V material. The growth temperature is 530 ° C. In addition, the contact layer 909 is formed by using a resist pattern produced by a known lithography technique as a mask and removing it so as to leave a part by selective etching by dry etching using BBr ₃ . A surface finger electrode 910 can be formed by forming a metal film on the remaining contact layer 909 by vapor deposition and performing heat treatment. The back electrode 911 can be formed by forming a metal film on the back surface of the substrate 901 by vapor deposition and performing heat treatment. The formation side of the antireflection film 912 is the incident side of sunlight.

The solar cell described above is measured for current-voltage characteristics under conditions of air mass 1.5 global (AM-1.5 Global, incident light intensity 1000 W / m ² ), non-condensing. In this case, the release voltage is 1.32 V, the short-circuit current density is 11.5 mA / cm ² , the fill factor is 0.78, and the photoelectric conversion efficiency is 11.8%. In a solar cell using only an InGaP PN junction that does not use a multiple quantum well structure as a light absorption layer, the release voltage is 1.37 V, the short-circuit current density is 9.2 mA / cm ² , and the fill factor is 0.83. The conversion efficiency is 10.5%. In a solar cell using a multiple quantum well structure, the band gap is smaller than that of a solar cell using only an InGaP PN junction as a light absorption layer. The reason why the solar cell using the multiple quantum well structure has higher photoelectric conversion efficiency is that the light having a small energy that cannot be absorbed by InGaP can be absorbed by the multiple quantum well structure.

  In addition, the short circuit current density of the solar cell using a multiple quantum well structure can be increased by increasing a quantum well layer. For this reason, the number of quantum well layers is not limited to the number described above, and may be increased or decreased so that the current matches the Si bottom cell when used for the top cell of a two-junction solar cell.

  As described above, in the present invention, since a multiple quantum well structure using InGaP is used, a combination of a solar battery cell using a III-V group compound semiconductor and a solar battery cell using Si is used. A two-junction solar cell can be configured without including As. The present invention is suitable as a solar cell that efficiently converts sunlight from light to electricity and further contains less environmentally hazardous substances. This makes it possible to meet the social demand for increasing the ratio of power generated by solar cells among the currently required power energy.

  The present invention is not limited to the embodiment described above, and many modifications and combinations can be implemented by those having ordinary knowledge in the art within the technical idea of the present invention. It is obvious. For example, an n-type III-V group compound semiconductor layer thinner than the second bonding layer 108 may be provided between the second bonding layer 108 and the n-type silicon layer 102. The first solar cell need only be composed of a p-type silicon layer and an n-type silicon layer, and it goes without saying that a p-type silicon layer may be provided on an n-type silicon substrate. Yes.

  DESCRIPTION OF SYMBOLS 101 ... Silicon substrate (p-type silicon layer), 102 ... n-type silicon layer, 103 ... 1st semiconductor layer, 104 ... Barrier layer, 105 ... Quantum well layer, 106 ... 2nd semiconductor layer, 107 ... 1st junction layer 108, second bonding layer, 111, antireflection film, 112, contact layer, 113, electrode, 114, electrode.

Claims (4)

a first solar cell composed of a p-type silicon layer and an n-type silicon layer;
A first semiconductor layer made of a p-type III-V group compound semiconductor configured without containing As, a multiple quantum well structure including a barrier layer made of InGaP and a quantum well layer made of InGaP, and without containing As A second solar cell comprising a second semiconductor layer comprising an n-type III-V group compound semiconductor,
A two-junction solar cell comprising at least a bonding layer disposed between the first solar cell and the second solar cell. The two-junction solar cell according to claim 1,
The two-junction solar cell according to claim 1, wherein a band gap energy of the quantum well layer is in a range of 1.65 eV to 1.85 eV. The two-junction solar cell according to claim 1 or 2,
The two-junction solar cell, wherein the quantum well layer has a compressive strain in a range of 0.5 to 2% with respect to GaAs. In the 2 junction solar cell of any one of Claims 1-3,
The two-junction solar cell, wherein the barrier layer has tensile strain with respect to GaAs.