Integrated Solar Collectors Using Epitaxial Lift Off and Cold Weld Bonded Semiconductor Solar Cells

Abstract

There is disclosed ultrahigh-efficiency single- and multi-junction thin-film solar cells. This disclosure is also directed to a substrate-damage-free epitaxial lift-off (“ELO”) process that employs adhesive-free, reliable and lightweight cold-weld bonding to a substrate, such as bonding to plastic or metal foils shaped into compound parabolic metal foil concentrators. By combining low-cost solar cell production and ultrahigh-efficiency of solar intensity-concentrated thin-film solar cells on foil substrates shaped into an integrated collector, as described herein, both lower cost of the module as well as significant cost reductions in the infrastructure is achieved.

Claims

1-17. (canceled)

18. A photovoltaic device comprising: a substrate comprising a reflective surface; and a solar cell comprising an active photovoltaic region; wherein the substrate is shaped as a cylindrical compound parabolic concentrator to concentrate the incident light onto the active photovoltaic region of the solar cell, and wherein the solar cell is shaped as a strip, the strip being disposed along a center of the cylindrical compound parabolic concentrator.

19. The photovoltaic device of claim 18, wherein the solar cell comprises a thin-film solar cell.

20. The photovoltaic device of claim 18, wherein the solar cell comprises a reflective, full-coverage ohmic contact such that the incident light has two passes through the active photovoltaic region.

21. The photovoltaic device of claim 20, wherein the reflective, full-coverage ohmic contact is bonded to the reflective surface of the substrate.

22. The photovoltaic device of claim 18, wherein the substrate comprises a metal foil.

23. The photovoltaic device of claim 22, wherein the solar cell is bonded to the metal foil.

24. The photovoltaic device of claim 18, wherein the substrate comprises a metallized plastic foil.

25. The photovoltaic device of claim 24, wherein the solar cell is bonded to the metallized plastic foil.

26. The photovoltaic device of claim 18, wherein the cylindrical compound parabolic concentrator is configured to provide a concentration that falls in a range from 4 to 10.

27. The photovoltaic device of claim 18, wherein the solar cell comprises a multi-junction cell.

28. The photovoltaic device of claim 18, wherein the solar cell comprises a III-V semiconductor solar cell.

29. The photovoltaic device of claim 18, wherein the strip is attached to the center of the cylindrical compound parabolic concentrator.

30. A photovoltaic device comprising: a substrate comprising a reflective surface; and a solar cell comprising a metal contact and an active photovoltaic region bonded to the metal contact; wherein: the metal contact of the solar cell is bonded to the reflective surface of the substrate, and the substrate is shaped as a compound parabolic concentrator to concentrate the incident light onto the active photovoltaic region of the solar cell.

31. The photovoltaic device of claim 30, wherein the solar cell comprises a thin-film solar cell.

32. The photovoltaic device of claim 30, wherein the metal contact is configured as a reflective, full-coverage ohmic contact such that the incident light has two passes through the active photovoltaic region.

33. The photovoltaic device of claim 30, wherein the substrate comprises a metal foil.

34. The photovoltaic device of claim 30, wherein the substrate comprises a metallized plastic foil.

35. The photovoltaic device of claim 30, wherein the compound parabolic concentrator is configured to provide a concentration that falls in a range from 4 to 10.

36. The photovoltaic device of claim 30, wherein: the compound parabolic concentrator is cylindrically shaped; and the solar cell is shaped as a strip attached to a center of the compound parabolic concentrator.

37. The photovoltaic device of claim 30, wherein the solar cell comprises a multi-junction cell.

Description

[0041] The accompanying figures are incorporated in, and constitute a part of, this specification.

[0042] FIG. 1. Is a schematic showing the ELO process according to the present disclosure for InP based solar cells.

[0043] FIG. 2. Is a photograph of a two inch InP epitaxial layer lifted off and bonded to a Au-coated Kaption sheet. ITO contacts form the Schotty solar cells.

[0044] FIG. 3. Is an atomic force microscope image of the original epi-ready InP substrate and recovered surfaces after the first and second ELO processes, with and without the use of protection layers.

[0045] FIG. 4. Is test data and a representative GaAs PV cell layer structure showing cell parameters.

[0046] FIG. 5. Is test data showing fourth quadrant current voltage and external quantum efficiency (inset) of a 23.9% efficient first-growth cell and a 22.8% efficient cell grown on a reused wafer.

[0047] FIG. 6. Is a schematic showing the ELO process as applied to an InP material according to the present disclosure.

[0048] FIG. 7. Is a schematic of a trilayer protection scheme with AlAs layer and AlAs lift-off layer.

[0049] FIG. 8. Is a schematic of a proposed multi-junction cell structure according to the present disclosure.

[0050] FIG. 9. Is a schematic of (a) conventional N/P tunnel junctions, and (b) N/ErP/P junction showing the reduced tunneling barriers.

[0051] FIG. 10. Is a schematic of an integrated reflector with cold-welded bonded ELO multi-junction cell.

[0052] One embodiment of the ELO process is shown schematically in FIG. 1. It begins with the epitaxial growth of the chemically distinct, thin “protection layers” consisting of InGaAs and InP, a sacrificial layer of AlAs, a second set of protection layers of InP and InGaAs, and finally the active photovoltaic cell layers. Next, the top epitaxial layer is coated with Au, as is a very thin plastic (e.g. Kapton™, a polyimide film marked by DuPont) host substrate. By pressing the two clean Au surfaces together at only a few kPa pressure, they form an electronically continuous and permanent, adhesive-free cold-weld bond whose properties are indistinguishable from a single, bulk Au film.

[0053] Once bonded to the plastic handle, the wafer is ready for ELO. The cold-weld bond is used not only for the ELO process (the epi-layer is attached permanently to the foil substrate prior to the liftoff, peeling away the parent substrate for eventual reuse) but also as the adhesive to the new host substrate on which the solar cells are eventually fabricated.

[0054] Replacement of adhesives conventionally used in lift-off by the cold-weld has several benefits: (i) attachment to the foil substrate is simple and is an integral part of the fabrication sequence, (ii) it is lightweight as it completely eliminates an adhesive layer, (iii) it is thermally and electrical “transparent” since the cold-weld interface is indistinguishable from the bulk of the film, and (iv) it is robust and resistant to failure. A selective chemical etchant, such as HF:H.sub.2O, 1:10, is used to remove the 4 nm to 10 nm-thick AlAs sacrificial ELO layer, parting the entire wafer from the photovoltaic epitaxial layers, leaving the protection layers exposed. The purpose of the protection layer nearest the AlAs ELO layer (InP in this case) is to provide an etch selectivity>108:1 and is removed from both the substrate and the parted epitaxial layers with a second wet etch (HCl:H.sub.3PO.sub.4, 3:1) that stops at the InGaAs protection layer surface. The requirements of the second protection layer are that it can be removed with a wet etchant that stops abruptly at the InP substrate. The InGaAs layer is removed from the wafer using H.sub.2SO.sub.4:H.sub.2O.sub.2:H.sub.2O (1:1:10), followed by C.sub.6H.sub.8O.sub.7:H.sub.2O.sub.2 (20:1), both of which have high selectivity to the InP substrate, InP buffer, and epitaxial layers, and assist in the removal of any debris or asperities remaining after the previous etch. Solar cells are fabricated on the epitaxial layers that are attached to the Kapton™ handle by sputtering indium tin oxide (ITO) Schottky contacts. The resulting flexible InP-ITO Schottky solar cells with efficiencies of ˜15% under 1 sun AM1.5G illumination are shown in FIG. 2. These bonded epitaxial sheets have been repeatedly cycled to >200° C. without delamination.

[0055] Previous to subsequent growth, the substrate is solvent cleaned, an intentional oxide is grown via exposure to UV/Ozone, and then returned to the growth chamber. The process has been employed multiple times with a single substrate to demonstrate degradation-free reuse of InP wafers, and as shown in FIG. 3, the smoothness of the surface can be improved over that of the commercial epi-ready wafers that are initially used, in principle allowing for indefinite reuse.

[0056] The Inventors have recently extended this damage-free regrowth process to GaAs-based single p-n junction photovoltaic cells fabricated on a parent wafer, resulting in efficiencies of 23.9%. FIG. 4 is a schematic representation of such a cell. The lift-off process is similar to that used for the InP cells, although the two-protection-layer scheme used for InP is replaced by a three-layer (InGaP/GaAs/InGaP), fully lattice-matched (to the AlAs sacrificial layer) system. This allows for improved etch-selectivity between layers while eliminating debris or surface roughening incurred in the ELO process. The AlAs layer is removed in HF, followed by removal of the InGaP and GaAs protection layers with HCl:H.sub.3PO.sub.4 (1:1) and H.sub.3PO.sub.4:H.sub.2O.sub.2:H.sub.2O (3:1:25), respectively.

[0057] Following this process, a second cell is grown on the parent wafer, reaching an efficiency of 22.8%. The slight (1%) reduction in power conversion efficiency between the first and second growths is due to the choice of the dry mesa-isolation etch recipe, resulting in a slight reduction in fill factor (see FIG. 4). Furthermore, the anti-reflection coating thickness was not optimal, reducing the external quantum efficiency and short circuit current as shown in FIG. 5. However, even higher efficiencies, for example greater than 25%, are expected when the coating thickness is optimized.

[0058] In one embodiment, a protection layer scheme based on the fully lattice-matched InGaP/GaAs/InGaP trilayer can be used. This tri-layer affords etch chemistries with sufficient rate selectivity between layers required to reproducibly remove the protection layers and to expose a pristine (physically and chemically undamaged) surface. In one embodiment, regrown thin-film cells are bonded via cold-welding to Au-coated plastic (Kapton™) substrates. It has been shown that a PCE=23.9% for a first growth wafer, and PCE=22.8% for a reused wafer can be achieved, which exceeds the Next Generation Photovoltaics II metric of 20% (see FIG. 5). A depiction of the actual ELO process apparatus and method are shown in FIG. 6.

[0059] Following each reuse, both the parent wafer and the lifted off epitaxial layers can be thoroughly studied for damage or subtle degradation. These methods include x-ray photoelectron spectroscopy (XPS) to determine chemical changes to the growth and regrowth surfaces, atomic force microscopy, scanning electron microscopy, and surface profilometry to determine surface morphological changes, cross-sectional transmission electron microscopy to examine defects that are incurred within the bulk of the epitaxy, and compositional depth profiling using secondary ion mass spectroscopy (SIMS).

[0060] Completed cells, including anti-reflection coating, can also be electrically tested using standard illumination conditions (1 sun, AM1.5G spectrum). Parameters to be measured include PCE, fill factor (FF), open circuit voltage (Voc), short circuit current (Jsc), series and parallel resistance.

[0061] It has been found that extended exposure (>2 days) of Ga-containing compounds (i.e. GaAs, and to a lesser degree InGaP) to HF results in surface contamination that is difficult to remove. This reaction, however, is absent for InP surfaces exposed to HF for over 7 days. In one embodiment a thin layer of strained InP placed immediately below the AlAs sacrificial layer will improve the fidelity of the surface, as shown in FIG. 7.

[0062] The thickness of the InP is limited to prevent strain relaxation, which can degrade the subsequently grown PV layer quality. The critical thickness of InP on GaAs is between 5 and 6 monolayers, corresponding to ˜1.7 nm. In this case the protection layer scheme would comprise InGaP/GaAs/InP or InGaP/GaAs/InGaP/InP, where the additional InGaP layer in the latter structure provides improved protection above the GaAs.

[0063] In another embodiment, the etch selectivity and preservation of the as-purchased wafer quality is carried out by using additional materials combinations, for example by replacing the InGaP layer adjacent to the InAlP. An InAlP/InGaP/GaAs/InAlP structure may be advantageous since InAlP can be etched with HCl:H.sub.2O (1:5), which stops abruptly at GaAs (>400:1 etch ratio), whereas HCl:H.sub.3PO.sub.4(1:1) used to etch InGaP slowly attacks the GaAs which results in roughening. By placing the InAlP adjacent to the AlAs layer, the InAlP is attacked by the HF and reduces the buildup of arsenic oxide which can slow the liftoff process. Also, InGaP may be used as an etch stop for the GaAs etch (H.sub.3PO.sub.4:H.sub.2O.sub.2:H.sub.2O, 3:1:25) to ensure that the lower InAlP layer is only removed in the final etch step.

[0064] Additional cost reduction may be possible by bonding to metal-foil substrates such as Au-coated Cu foils, use of less expensive metals for cold-welding (e.g. Ag instead of Au), reduced consumption of HF, reduced protection layer thicknesses, and accelerating the lift-off process. The extended exposure to HF used to dissolve the AlAs sacrificial layer limits the choice of metal host substrates that can be employed. In one embodiment, Cu foils, which can be used for cold-welding, are used to increase resistance upon exposure to HF, as their use may be simpler than coating the foil with a noble metal such as Au. An additional benefit to using Cu foil is its high thermal conductivity (˜4 W cm.sup.−1*C.sup.−1) that can be exploited to extract heat from the concentrated cells.

[0065] There is also disclosed very high-efficiency multi-junction (GaAs/InGaP) solar cells following the two cell example structure shown in FIG. 8.

[0066] The design is inverted relative to a conventional multi-junction cell growth sequence to accommodate the “upside down” bonding geometry used in the adhesive-free cold-weld process; the structure includes a 25% GaAs cell architecture. In this case, the GaAs cell thickness is reduced to 2 μm (50% of the conventional substrate-based cell) since the reflective, full-coverage ohmic contact allows for two passes of the incident light through the device active region. The primary focus will be on optimizing the tandem PV structure for maximum efficiency, including InGaP cell design (layer thickness, window layer, layer composition, etc.), improving the wide-gap tunnel junctions (TJ) between elements in the stack, and perfecting the multiple lift-off process over large areas for this multi-junction cell.

[0067] Solar cells will be grown with n-type material on top of p-type layers, whereas the tunnel junctions must be grown with the opposite polarity. The cells may employ carbon-doping in all or several of the p-type layers, since carbon does not readily migrate to the growth surface as does the conventional p-dopant, Be. As the tandem cells are generally limited by the current in the GaAs cell, the InGaP cell thickness needs to be adjusted to current-match the InGaP and GaAs cells; the thickness of the InGaP layer is expected to range from 0.55 to 0.80 μm.

[0068] Efficient tunnel junctions (TJ) are essential for high performance tandem cells. They need to be nearly loss-less in both voltage and absorption. It is advantageous to use an InGaP TJ in MJ cells to avoid GaAs TJ absorption that may be as high as 3%. A conventional TJ is an abrupt P+/N+ junction where the electron can tunnel directly from the conduction band on the n-type side to the valence band on the p-type side (FIG. 9(a)). Little work has been performed on MBE grown wide gap TJs, although doping levels that are sufficiently high to transport currents generated at 1 sun illumination have been reported using MBE.

[0069] One embodiment is directed to InGaP tunnel junctions that have a voltage drop of several tens of mV at 1 sun. Research suggests that Be and Si are suitable dopants (attaining densities of 3.7×10.sup.19 and 1.8×10.sup.19 cm.sup.3, respectively). However, if a reduced tunneling resistance is required, the use of engineered defects at the P+/N+ interface, can be done, such as by adding ErAs to a GaAs tunnel junction. In this case, ErP or LuP may be used as shown in FIG. 9b. The ErP or LuP form epitaxial islands on the semiconductor surface that are ˜4 monolayers thick, are metallic, and split the tunneling process into two steps with significantly higher tunneling probabilities. By employing ErP in the TJ, several orders of magnitude increase in the tunneling current may result, and lead to voltage drops in the sub-mV range for the currents anticipated in the fabricated PV cells.

[0070] As in the case of the single junction cells, the multi-junction cell can be microscopically and chemically examined after each iteration of the growth-ELO-reuse cycle. Completed cells, including anti-reflection coating, can be electrically tested using standard illumination conditions (AM1.5G spectrum), but over a range of intensities up to 10 suns. Parameters to be measured include PCE, fill factor, open circuit voltage, short circuit current, series and parallel resistance, as in the case of the single junction cells.

[0071] Thin-film multi-junction cells bonded onto reflective and flexible substrates provide a unique opportunity to integrate the solar collector with the thin-film cell without introducing significant additional costs. FIG. 10 shows that a strip consisting of the ELO multi-junction cell is bonded to the center of a larger, flexible, reflective film. The film is then molded (by placement in a thermally conductive or actively cooled preform) into the shape of a compound parabolic collector (e.g. a CPC, or Winston collector). This geometry concentrates parallel solar rays onto the cell strip at its focus, as well as collects diffuse light within the acceptance cone.

[0072] The small levels of concentration (4-10×) generally used in the cylindrical Winston-type collectors allow the concentrators to be highly efficient, and to direct a significant amount to diffuse light into the cell. The efficiency of collection is given by CEff=T.sub.CPCγ, where T.sub.CPC is the effective transmittance of the CPC, including losses of multiple bounces that are ˜2% for common reflector materials. The correction for diffuse light is γ=1−(1−1/C)Gdiff/Gdir, where C is the intended concentration, and Gdiff/Gdir is the fraction of diffuse to total incident light. Typically, Gdiff/Gdir˜0.11 for a low-cloudiness day. Then for C=4, γ=90% at AM1.5G, which is comparable to the power available at AM1.5D.

[0073] For a 4× CPC, and assuming a solar cell strip width of 1 cm, the aperture is then 4 cm wide×10 cm deep, providing a practical form factor compatible with panels used in single family dwellings. At higher concentrations, the size of the concentrator increases considerably. For example, a 10× concentration used with the same 1 cm wide cell strip requires an aperture of 10 cm with a depth of ˜55 cm. This can be reduced to ˜40 cm with negligible effect on concentration efficiency. [25] The amount of reflective material needed is 4-5 times larger for 4× concentration, and 8-11 times for 10× concentration.

[0074] Additional benefits to the small concentrations used include the allowed use of single-axis tracking (daily or seasonally, depending on orientation of the collector), and simplified passive cooling than are needed for higher concentrations. Indeed, the very thin substrates used greatly simplify heat transfer: calculations indicate that at 10× concentration and a 25 mm thick Kapton™ substrate placed against a passively cooled Cu heat sink results in a temperature rise of only 5-20° C., obviating the need for more aggressive cooling methods.

[0075] Note that the ELO cell technology can also be applied to systems with large concentration factors; however, here the present disclosure focused only on smaller concentrations that lead to simple and economical designs that are applicable to residential systems. The cost reductions in this integrated solar collector+ELO multi-junction concentrator cell assembly is expected to radically reduce the cost of concentrated systems, as well as their footprint (owing to the high PCE).

[0076] Unless otherwise indicated, all numbers expressing quantities of ingredients, reaction conditions, and so forth used in the specification and claims are to be understood as being modified in all instances by the term “about.” Accordingly, unless indicated to the contrary, the numerical parameters set forth in the following specification and attached claims are approximations that may vary depending upon the desired properties sought to be obtained by the present invention.

[0077] Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope of the invention being indicated by the following claims.