Nanostructured Photovoltaics for Space Power

The primary goals for this fellowship include the refinement of the fabrication process, improved understanding of the design tolerances via simulation, characterization and a thorough understanding of the nipi devices radiation tolerance, the fabrication of quantum dot-nipi (QD-nipi) devices, and testing the QD nipi devices to determine if an intermediate band solar cell (IBSC) can be realized. The work completed for my thesis and this fellowship explored the possibility of using a nipi doping SL device in photovoltaic applications. Prior to the beginning of this work a significant amount of theoretical work had gone into understanding the physics of nipi superlattices (SLs); however, very little work had been completed towards the development of photovoltaic devices with nipi SLs. The work contained here continued the theoretical development, and evaluated multiple methods for fabricating nipi solar cells. This provides a path forwards towards the development of high efficiency solar cells that are tolerant to materials with reduced minority carrier lifetimes. Work related to the theoretical understanding of nipi solar cells was the development of the modified dual diode model. The model adapted the conventional dual diode model that describes a single junction solar cell to allow n diodes to be placed in parallel in a nipi parallel connected diode configuration. Additionally the regrown contact could be modeled in parallel to the nipi diodes to account for injection from the carrier-selective contacts. The diode model was paired with calculations for minority carrier lifetimes and dark saturation currents, which were modified by photogenerated carrier concentrations. The coupling of these calculations with the diode model provided a greater understanding of trap state filling for devices measured under increased concentrations. As solar concentration was increased trap states were filled and no longer active, allowing the dark saturation current and VOC of the solar cell to improve. The modeling that was completed allowed a more detailed understanding of defects and traps at the epitaxial regrowth interface. The model was used to characterize experimental measurements of nipi solar cells under concentration. Improvements in VOC were seen under concentration that were greater than expected by the logarithmic increase in VOC expected by theory. The results were explained by the model with a trap density of 4x1014 /cm3 at mid-gap which were filled to allow VOC to fully recover at a concentration of 7 suns. Sensitivity modeling was also completed to demonstrate that a trap density of 5x1013 /cm3 is required in order to ensure that all the trap states were filled at a concentration of one sun, effectively demonstrating an ideal solar cell. In addition to the modeling work, simulation work with Synopsys Sentaurus was also completed to gain a complementary understanding of the devices. The two-dimensional simulation software was used to gain an understanding of expected results of an ideal nipi solar cell, and devices with traps at the epitaxial regrowth interface. Ideal simulations showed an achievable efficiency of 14.4% without an ARC, and simulations with traps allowed a close match to experimental for VOC. Further simulation work was also completed to provide a detailed understanding of the operation of nipi solar cells in a radiation environment. Simulations with 1 MeV electrons showed that nipi solar cells could have beginning of life (BOL) efficiencies close to conventional pin designs and a better absolute efficiency at doses higher than 1x1014 e/cm2 assuming the correct design was used. The majority of the work completed was focused on the fabrication and testing of nipi solar cells, demonstrating a significant improvement in the understanding of how to fabricate a device that requires carrier-selective lateral contacts. Initial characterization of the fabrication method for forming the carrier selective contacts with both epitaxial regrowth and spin-on dopant diffusion processes was completed. Numerous solar cells have been fabricated with the processes developed, which provided a significant amount of information about the fabrication processes required to improve device results. Initial steps were able to determine that using a back-side contact a parasitic shunt path could be eliminated between the anode and cathode. This change resulted in a significant efficiency increase from 6.1% to 9.0% which corresponded to an increase in VOC and fill factor. Further improvements were made by fine tuning the fabrication process to reach an efficiency of 9.14% prior to anti-reflection (AR) coating and 12.5% following AR coating. Evaluation of alternate fabrication routines was also completed by changing the epitaxial regrowth process from a homoepitaxy growth of GaAs on GaAs to heteroexpitaxy using either InGaP or AlGaAs as the regrowth material. The use of InGaP as a regrowth material did show some promise with an increase in VOC from 0.58 V for GaAs to 0.659 V for InGaP. A final alternate process was developed to use a diffused junction approach. Although the device results fabricated with this process were degraded when compared to the epitaxial regrowth process, it was a demonstration of the viability of fabricating nipi devices through various means. With additional process development it is believed that a diffused junction process could yield results similar to the epitaxial regrowth process. Given the broader objective of an IBSC device, work evaluating a QD-nipi SL was completed. Multiple nipi and QD-nipi test structures were grown and evaluated spectroscopically. The relationship between absorption of a nipi SL was compared to theory where the electric field strength was directly proportional to the sub-band absorption. Absorption from QD-nipi test structures showed that putting QD within the nipi layers resulted in a red shift for absorption from the QD states due to the quantum confined Stark effect. This resulted in an increase in absorption for the QD-nipi structure between 920 and 1050 nm when compared to the QD only structure. Photoluminescence (PL) analysis was able to show the same peak shift in the QD states, and temperature dependent PL (TDPL) showed a decrease in the activation energy for the QD-nipi structure to 30 meV from 113 meV for the QD structure. Time-resolved photoluminescence showed an increase in the lifetime measured from the QD-nipi test structure by three orders of magnitude; however, this was theorized to be due to the non-equilibrium carrier population in the nipi layers recombining through the QD states. This was confirmed through time resolved PL (TRPL) measurements of the conventional nipi structures as 900 nm which showed a similar lifetime as the QD nipi device, indicating the enhanced lifetime was due to the incresed lifetime expected in the nipi SL. Further confirmation came from transient absorption analysis which showed the lifetime of the excited states in the QD states when resonantly pumped is on the order of half a picosecond. Finally, analysis of the radiation tolerance of nipi solar cells was completed. Some initial simulations pointed to the enhanced radiation tolerance possible with the nipi device having a higher overall efficiency than the pin solar cell at 1 MeV electron radiation doses greater than 1x1013 /cm2. Experimental results were able to show that nipi solar cells had a remaining factor for efficiency that remained at 100% for a dose up to 2x1015 e/cm2, which was much higher than expected, as compared to the pin control device that had a remaining factor of 68.8% at a dose of 1x1015 e/cm2. The result was explained by simulations that showed trap states at the regrowth interface dominate the dark current and VOC if the trap density is higher than the defect density created by electron radiation. Simulations also showed the possibility for improving the BOL efficiency for nipi solar cells by reducing the trap density; however, this had a minimal effect on end of life (EOL) efficiency.