J. A. Cape

26.1% solar cell efficiency for Ge mechanically stacked under GaAs

We have processed a diffused Ge wafer into a Ge concentrator solar cell and mechanically stacked it under a GaAs cell fabricated by Varian. We measured this stack’s efficiency to be 26.1% for terrestrial air mass 1.5 direct (AM1.5D) conditions at a 285× concentration ratio. We showed that this efficiency is limited by optical absorption in the Varian GaAs cell caused by high 2–4 (1018) cm−3 substrate doping. We used a 2×1017 cm−3 doped GaAs filter to estimate the stack efficiency as 27.4%, which would be achieved with the same Varian GaAs cell formed on a lower doped substrate. We project efficiencies assuming the best properties reported for a GaAs device. This gives a 29.6% efficiency for an improved, planar Ge cell and 31.6% efficiency for a proposed point contact geometry for the Ge cell. The corresponding space (AM0) efficiencies at a 159× concentration ratio range from the 23.4% value we measured on the stack up to 28.4% projected for the point contact Ge place under the best GaAs cell. We showed th...

Epitaxial films grown by vacuum MOCVD

Abstract We are developing high efficiency multicolor solar cells for terrestrial applications using a novel MOCVD growth technique, vacuum MOCVD, for growing the sequential epitaxial layers. We believe the vacuum MOCVD growth technology offers several advantages for production scale-up including the more efficient use of the metal alkyl source materials. In addition, the use of stainless steel rather than glass offers important safety advantages. For two color cell applications, we are currently developing the GaAs1−xPx and GaAs1−ySby ternary alloys. In this paper, we first describe our vacuum MOCVD system and then the current status of our vacuum MOCVD grown GaAs1−ySby materials. Triethyl-Sb and trimethyl-Sb are compared as sources of Sb, and dicyclopentadienyl-Mg and diethyl-Zn are compared as p-type dopant sources

GaAs films grown by vacuum chemical epitaxy using thermally precracked trimethyl‐arsenic

Trimethyl‐arsenic (TMAs) is used as a source of arsenic for GaAs film growth. In the process used, vacuum chemical epitaxy, TMAs is thermally decomposed into arsenic upstream in a hot cracker furnace. The arsenic and stable hydrocarbons are then transported in vacuum without condensation to the epitaxial growth zone. The hole carrier concentration and carbon content in grown films are studied via Hall, electrochemical profile, and secondary ion mass spectroscopy as a function of cracker furnace design. It is shown that when the TMAs decomposition efficiency is poor, the carbon content can be as high as 1019/cm3 but for a more efficient cracker, the carbon content can be reduced into the 1016/cm3 range. Toxic injury hazards can be reduced substantially by substituting TMAs for the more widely used arsine in GaAs growth systems.

A new sequentially etched quantum‐yield technique for measuring surface recombination velocity and diffusion lengths of solar cells

We have developed a new technique to characterize the individual layers of high‐efficiency solar cells. In general, the technique allows one to set lower bounds for diffusion lengths and upper and lower bounds for interface recombination velocity. This is sufficient to determine which parameter limits performance, and often the actual parameter values are also determined accurately. We obtain this information by fitting a theoretical model to quantum‐yield spectra measured on a sample in its initial state, and after its window passivation and top active layers are sequentially etched away. With such data on two p on n GaAs solar cells with AlxGa1−xAs passivation, we determined minority‐carrier hole diffusion lengths of 1.0±0.2 and 0.2±0.05 μ in the Te‐doped n layers for first and second samples, respectively. We found lower limits for the minority‐carrier electron diffusion lengths in the top p layers of 2.0 μ in the carbon‐doped first sample and 4.0 μ in the Mg‐doped second sample. We determined interfac...

Vacuum chemical epitaxy utilizing organometallic sources

Herein, we describe a process, Vacuum Chemical Epitaxy (VCE), which incorporates some of the principle advantages of molecular beam epitaxy (MBE) and metal-organic chemical vapor deposition (MOCVD) systems while rejecting their disadvantages and limitations. In this process, multiple group(III)-alkyl molecular beams are directed through a water cooled gas distribution block onto wafers providing for the growth of uniform films over large areas with high group(III)-alkyl utilization efficiency. The group(V) source, on the other hand, is injected at a single point on one side of the deposition zone. The group(V) molecules are confined and undergo molecular flow across the deposition zone. A variety of group(V) source molecules are used including the group(V) hydrides (AsH3 and PH3) and elementary group(V) molecules (As2 and P2). In the work presented here, the elemental group(V) molecules are generated by thermally cracking the hydrides. However, the use of conventional MBE elemental group(V) evaporative sources is also possible thereby eliminating the safety issues associated with the hydride source gases. In this paper, our VCE reactor is described in some detail along with the properties of III-V films grown with this equipment. The fabrication of a GaAsSb solar cell with an active area energy conversion efficiency of 26.7% demonstrates that Vacuum Chemical Epitaxy has the capability of producing high performance devices.

GaSb films grown by vacuum chemical epitaxy using triethyl antimony and triethyl gallium sources

GaSb films have been grown using triethyl‐Ga and triethyl‐Sb sources. In a hot‐wall reaction chamber located within a high‐vacuum chamber, multiple group‐III alkyl molecular beams are directed into the reaction chamber onto wafers. The group‐V molecules are injected from the perimeter of the reaction chamber and undergo molecular flow across the deposition zone. The utilization efficiency of the group‐V source material is enchanced by the use of a thermal cracker located at the point of group‐V gas injection and by the use of the hot‐wall chamber. Both unintentionally doped p‐type and Te doped n‐type GaSb films are grown and characterized. GaSb p‐n junction photodiodes are also reported with internal quantum yields as high as 85%. Unintentionally doped films were shown to have background carrier concentrations of 4×1016 cm3 by capacitance versus voltage measurement.

Measurement of a long diffusion length in a GaAs film improved by metalorganic-chemical-vapor-deposition source purifications

The vacuum metalorganic‐chemical‐vapor‐deposition (Vacuum MOCVD) process was combined with two source purifications to grow p‐GaAs epitaxial films of high quality. Theoretical modeling of quantum yield spectra measured on a specially configured n+‐p sample determined the minority‐carrier electron diffusion length to be 10 μm to within a factor of 2 in the p layer. The p doping was reduced to the 5×1017 cm−3 level to avoid suppression of the diffusion length by Auger recombination. Multiple vacuum sublimations of dicyclopentadienyl magnesium (CP2Mg), the source of Mg for p doping, reduced the contamination by air and by cyclopentadiene (CP) by an order of magnitude. A dry ice/acetone cold trap was operated at slightly below 180‐Torr pressure to reduce the water vapor content of arsine, used as the As source, from the hundreds of ppm down level down to the 2 ppm range. The vacuum growth process reduced residual gas contamination. These techniques were combined to grow a p on n GaAs solar cell with an effici...

High‐efficiency GaAs0.7P0.3 solar cell on a transparent GaP wafer

The fabrication of a high performance GaAs0.7P0.3 solar cell on a transparent GaP substrate with an AM1.5 efficiency of 15.4% for a concentration ratio of 30× is reported for the first time. The measured transparency of the GaP substrate allows these cells to be mechanically stacked on silicon solar cells in a manner that should yield combined conversion efficiencies well over 25%. This mechanically stacked two‐band‐gap cell design is particularly attractive because it utilizes the already well‐developed Si solar cell and because the materials foundation for the GaAs0.7P0.3 on GaP cell has been laid by work on light emitting diodes.

Near-term higher efficiencies with mechanically stacked two-color solar batteries

Abstract Two-band-gap (or two-color) solar batteries have the potential for much higher efficiencies than can be achieved with single-junction devices. While two-terminal monolithic stacks with two active junctions and a series interconnect junction grown sequentially onto a single substrate are conceptually very elegant, they require the development of three new components in a complex multi-element III–V alloy system. Although such devices have been demonstrated, each of the three novel components (the two cells and the interconnect) must be optimized so that it will perform near its theoretical limit in order for the two-color device to outperform the more developed single-junction devices. Tandem mechanically stacked two-color solar batteries offer a shorter path to commercialization, primarily because one of the two active cells can be an already-developed Si or GaAs cell and because the interconnect problem can be easily solved with conventional wire-bond technology. Thus, the problem of developing three new components for the monolithic-device option is replaced by the problem of developing one new component for the mechanical-stack option. We previously observed that, if the Si cell is chosen as the well-developed cell, a GaAs 0.7 P 0.3 cell on a GaP substrate would be a logical choice for the novel component in a solar battery. In studying this particular option, we recently noted that the four terminals available in a mechanical-stack battery can be used in a series-parallel interconnect scheme, allowing voltage matching at the module level. This interconnect scheme makes the module energy-conversion efficiency quite insensitive to variations in the solar spectrum and opens up a broader range of materials that are usable in mechanical stacks. In particular, it may be desirable to choose a GaAs cell with a band gap of 1.42 eV as the well-developed cell and a GaSb cell with a band gap of 0.72 eV as the novel cell in a two-color solar battery. The advantage of this option is that the high-band-gap GaAs cell currently holds the world record for energy-conversion efficiency, and the GaSb cell would simply boost this record. Moreover, GaSb is advantageous in comparison to GaAsP because it is a binary compound rather than an alloy and would not require compositionally graded transition layers between the active junction and the substrate. Both of these considerations should greatly improve the chances for higher performance. GaSb photodiodes with high quantum efficiencies have already been described in the literature, and available data allow the projection of stack efficiencies over 30% (AM 1.5) with this option.

GaP and GaAs films grown by vacuum chemical epitaxy using TEGa and TMGa sources

Abstract We have previously described a hybrid MBE/MOCVD growth technique in which organometallic vapors are injected into a high vacuum system to grow III–V epitaxial films. In our process, called vacuum-chemical-epitaxy (VCE), multiple group(III)-alkyl molecular beams are directed onto wafers from a water cooled gas distribution manifold providing for the growth of uniform films over large areas with high group(III)-alkyl utilization efficiency. The group(V) source, on the other hand, is injected at a single point on one side of the deposition zone. The group(V) molecules are confined and undergo molecular flow across the deposition zone. The utilization efficiency of the Group(V) source material can be enhanced by the use of a thermal cracker at the point of group(V) gas injection. We have previously described the growth of high purity epitaxial GaAs films using triethygallium (TEGa) in a VCE reactor. In this paper, we describe the growth of GaP films using TEGa and trimethylgallium (TMGa) in VCE. Data are also presented on the effect of the thermal cracker temperature on PH 3 cracking efficiency and on GaAsP film composition. Finally, the properties of GaAs films grown using TMGa will be described. In particular, it is noted that TMGa can be used as a source of carbon doping for p-type GaAs films and that these carbon doped films can be incorporated in solar cells yielding devices with high quantum yields.