Search NASA⌕ Search

Engineering topics

Good, Brian

Publications and source records attributed to Good, Brian.

At least 19 records

SnO2 Buffer Layers for High Efficiency CdSeTe/CdTe Devices

SnO2 buffer layers of different thickness were deposited onto TEC 15 Fluorine doped tin oxide coated glass substrates using rf magnetron sputtering. The buffer layers were then incorporated into Cu-doped CdSeTe/CdTe devices using a range of CdCl 2 activation treatments and CuCl2 annealing temperatures to determine the effects of buffer layer thickness on device performance. Results show that all devices fabricated with thinner buffer layers resulted in much better J - V characteristics than their thicker counterparts. This was mainly due to a reduced open-circuit voltage (Voc) when using thicker buffer layers. The best device produced a conversion efficiency of 16.59%, fill factor of 71.62%, Jsc of 28.44 mA/cm 2 and Voc of 814.23 mV.

buffer layers↗

Effect of Near-Interface Compensation of CdSeTe Absorber Layers on Solar Cell Performance

Arsenic has been shown to be an effective p-type dopant of CdTe. However, challenges remain in the fabrication of high efficiency CdTe solar cells using As. As-doped CdTe is prone to self-compensation and observed accumulation of dopant atoms in CdTe near the interface with MgZnO (MZO) suggests that this could be occurring. In this study, we use SCAPS 1-D modeling software to investigate the effect of near-interface compensation. We consider three cases: shallow donors, deep recombination centers, and a thin layer of excess positive charge accumulation. All of these cases are shown to have significant effects on the current-voltage characteristics, while the thin charge layer also affects capacitance-voltage measurements.

arsenic↗

Effects of Absorber Near-Interface Compensation on Cd(Se,Te) Solar Cell Performance

Arsenic (As) has been shown to be an effective p-type dopant for CdTe, although high performance in As-doped devices remains difficult to achieve. Arsenic is prone to self-compensation in CdTe, as evidenced by the accumulation of dopant atoms in CdTe/Cd(Se,Te) near the interface with MgxZn1-xO (MZO). In this study, we use SCAPS 1D modeling software to investigate the effect of near-interface compensation, helping elucidate loss pathways in present-day As-doped devices and informing future growth directions. We consider three possible results of As accumulation: shallow donors, deep recombination centers, and a thin layer of excess acceptor accumulation. The reduction in near-interface carrier concentration caused by shallow donors is shown to improve open-circuit voltages (VOC), whereas deep levels are detrimental to all performance parameters. The thin charge layer affects capacitance-voltage (CV) measurements by reducing the depletion width while maintaining the same carrier concentration, replicating CV behavior that has been observed in actual devices. These results illustrate the importance of monitoring dopant accumulation within 100 nm of the interface, and suggest that reducing or eliminating the As concentration in this region would be beneficial. An undoped Cd(Se,Te) layer at the interface is suggested as a possible device structure to boost performance.

absorber↗

Investigating the Role of Copper in Arsenic Doped CdSeTe Photovoltaic Devices

As part of NREL's development of arsenic doped CdSeTe devices, co-doping with copper has become a common practice and, despite little difference in carrier concentration (often ~1016 cm-3), co-doped devices regularly show improved Voc. Given the critical importance of improving Voc, we are investigating this trend with a wide variety of characterization techniques including SEM, EBSD, SIMS, PL, CL, XPS, TRPL, JVT, KPFM, and DLTS. Together these indicate that Cu facilitates improved absorber-buffer interface properties and potentially improved bulk absorber characteristics, though specific mechanisms have not yet been determined. Despite the improved performance of co-doped devices, Voc is still far below its potential, and we explore the possibility that this limitation is the result of low buffer doping in conjunction with high absorber doping.

arsenic doping↗

Codesigning Alloy Compositions of CdSe y Te 1− y Absorbers and Mg x Zn 1− x O Contacts to Increase Solar Cell Efficiency

Thin‐film solar cells such as CdTe are a major commercial photovoltaic technology, with more than 25 GW installed worldwide and levelized costs of electricity competitive with fossil fuels. Further progress may result from integrating CdSe y Te 1− y absorbers with Mg x Zn 1− x O contacts, but the device efficiency is difficult to maximize due to coupled dependence on chemical composition of both alloys. Herein, a high‐throughput approach is demonstrated to codesign chemical compositions in alloyed Mg x Zn 1− x O/CdSe y Te 1− y thin‐film solar cells, using combinatorial libraries of PV devices with orthogonal composition gradients in CdSe y Te 1− y absorbers and Mg x Zn 1− x O contacts. It is found that the solar cell performance is a strong and coupled function of both elemental compositions, with efficiency up to 17.7% ( V OC = 836 mV, fill factor = 69%, J SC = 30.6 mA cm −2 ) at atomic compositions of Mg/(Mg + Zn) ≈18% and average Se/(Se + Te) ≈4%. These performance trends among >100 devices are explained by >100 ns lifetime of photoexcited charge carriers at the Mg x Zn 1− x O/CdSe y Te 1− y interface where strong Se accumulation is also observed. This study reports the optimal compositions of the commercially relevant Mg x Zn 1− x O/CdSe y Te 1− y solar cells and demonstrates a general approach to codesigning performance of alloyed thin‐film solar cells and other optoelectronic devices.

14 SOLAR ENERGY↗

Investigating the Role of Copper in Arsenic Doped Cd(Se,Te) Photovoltaics

The open circuit voltage (VOC) deficit in Cd(Se,Te)-based photovoltaics remains a critical obstacle for pushing the technology closer to theoretical performance limits. Arsenic doping has become a dominant and promising route to achieve the higher p-type carrier concentrations necessary for higher VOC, but challenges associated with this alternate defect chemistry and higher doping density have hindered progress. Here we show that while arsenic doping enables high carrier concentrations (>1016 cm-3), co-doping with copper can provide a boost to VOC without a significant change to carrier concentration. A large data set is initially used to explore current-voltage and capacitance-voltage trends associated with arsenic doped devices with and without copper. A smaller subset is then used to probe these trends using a wide variety of characterization techniques. Copper is found to facilitate reduced interface recombination and potentially improved bulk absorber characteristics, though the mechanisms for these improvements are not yet clear. Despite the improved performance of co-doped devices, VOC is still far below its potential especially for highly doped devices. Low emitter doping in conjunction with high absorber doping seems to be a plausible cause for this significant deficit, though other device properties may exacerbate this problem.

CdSeTe↗

Dual-Wavelength Time-Resolved Photoluminescence Study of CdSe x Te 1-x Surface Passivation via Mg y Zn 1-y O and Al 2 O 3

Voltage loss is currently one of the biggest challenges facing cadmium telluride (CdTe) based photovoltaics. Determining the location(s) of major voltage loss within the device stack (e.g., front/back interface, grain boundaries) is therefore of primary interest. Here, we present a custom-built time-resolved photoluminescence system with two excitation wavelengths - 670 (standard) and 405 nm - to probe the device stack at depths of approximately 130 and 35 nm, respectively; their comparison helps differentiate interface and bulk contributions to carrier lifetime. We apply this system to examine the passivation effect of two significant recent advances in CdTe: the incorporation of Se to form graded CdSe x Te 1-x and the replacement of CdS with Mg y Zn 1-y O. It is found that x = 0.2 Se is required to obtain lifetime improvements, primarily in the bulk. Additionally, evidence for trapping at the Mg y Zn 1-y O/CdSe x Te 1-x interface was observed. Furthermore, this indicates further work is required to sufficiently passivate the front interface.

CdSeTe↗

Carrier Lifetime as a Function of Se Content for CdSexTe1-x Films Grown on Al2O3 and MgZnO: Preprint

Time-resolved photoluminescence (TRPL) with two excitation wavelengths – 670 (standard) and 405 nm – was used to examine the effect on carrier lifetime of two significant recent advances in CdTe: the incorporation of Se to form graded CdSe x Te 1-x and the use of Mg y Zn 1-y O buffers. The two excitation wavelengths probe depths of approximately 130 and 35 nm, respectively, and their comparison helps differentiate interface and bulk contributions to carrier lifetime. It was found that x = 0.2 Se was required to obtain lifetime improvements, primarily in the bulk. Additionally, TRPL traces for Mg y Zn 1-y O/CdSe x Te 1-x samples showed fast initial decay followed by a long-lived tail, which may be indicative of trap-dominated recombination. This behavior was not present for CdSe x Te 1-x films grown on Al 2 O 3 , which is currently state-of-the-art for surface passivation in CdTe. This indicates that further work is required to sufficiently passivate the front interface.

cadmium telluride↗

Carrier Lifetime as a Function of Se Content for CdSexTe1-x Films Grown on Al2O3 and MgZnO

Time-resolved photoluminescence (TRPL) with two excitation wavelengths – 670 (standard) and 405 nm – was used to examine the effect on carrier lifetime of two significant recent advances in CdTe: the incorporation of Se to form graded CdSe x Te 1-x and the use of Mg y Zn 1-y O buffers. The two excitation wavelengths probe depths of approximately 130 and 35 nm, respectively, and their comparison helps differentiate interface and bulk contributions to carrier lifetime. It was found that x = 0.2 Se was required to obtain lifetime improvements, primarily in the bulk. Additionally, TRPL traces for Mg y Zn 1-y O/CdSe x Te 1-x samples showed fast initial decay followed by a long-lived tail, which may be indicative of trap-dominated recombination. This behavior was not present for CdSe x Te 1-x films grown on Al 2 O 3 , which is currently state-of-the-art for surface passivation in CdTe. This indicates that further work is required to sufficiently passivate the front interface.

cadmium telluride↗

Electro-Optical Stability in Gallium Magnesium Zinc Oxide Layers for CdTe Solar Cells

In CdTe solar cells, the emitter electron density, bandgap, and band alignment can be tuned by adjusting alloy composition and stoichiometry in gallium magnesium zinc oxide (GMZO) layers, thereby allowing for critical front interface engineering. For example, computational modeling indicates efficiency can change from 7% to 25% by adjusting these buffer properties even with fixed CdSeTe and CdTe material properties. However, the GMZO as-deposited electro-optical properties can shift during subsequent CdTe deposition, CdCl 2 annealing, and solar cell operation. Here, we examine these changes by sputtering GMZO films on glass, and measuring the electron density before and after post-deposition treatments over a range of temperatures and different ambients. We then compare GMZO to SnO 2 and MZO. The addition of Ga to MgZnO is found to have profound effects on both the electron density and electro-optical stability that should be advantageous for CdTe solar cells.

41 EE - Solar Energy Technologies Office (EE-4S)↗

Tailoring SnO 2 , (Mg,Zn)O, and Ga:(Mg,Zn)O electro-optical properties and stability for solar cells

The electron density, mobility, bandgap, and band alignment of transparent conducting oxides (TCOs) can be tailored by adjusting composition and stoichiometry, thereby enabling interface engineering for diverse semiconductor applications. For example, solar cell efficiency can change enormously by adjusting TCO properties. At the same time, these TCO properties can shift during the deposition of other layers, anneals, and device operation. An ideal TCO should have tunable but stable electro-optical properties. Here, we deposit SnO 2 , (Mg,Zn)O (MZO), and Ga:(Mg,Zn)O (GMZO) films on glass and measure electro-optical characteristics before and after reducing, inert, oxidizing, and CdCl2 anneals over a range of temperatures. Electron density generally increases in the progression from oxidizing to inert and reducing ambients. SnO 2 is relatively stable compared to MZO but has less flexibility for interface engineering. We investigate GMZO as a similar but more stable alternative to MZO.The addition of Ga to MZO has significant effects on electron density and improves electro-optical stability, which can be advantageous for semiconductor applications. Furthermore, we demonstrate that GMZO can be readily incorporated into solar cells.

14 SOLAR ENERGY↗

Development of Advanced Environmental Barrier Coatings for SiC/SiC Ceramic Matrix Composites: Path Toward 2700 F Temperature Capability and Beyond

Advanced environmental barrier coating systems for SiC-SiC Ceramic Matrix Composite (CMC) turbine and combustor hot section components are currently being developed to meet future turbine engine emission and performance goals. One of the significant coating development challenges is to achieve prime-reliant environmental barrier coating systems to meet the future 2700F EBC-CMC temperature stability and environmental durability requirements. This presentation will emphasize recent NASA environmental barrier coating system testing and down-selects, particularly the development path and properties towards 2700-3000F durability goals by using NASA hafnium-hafnia-rare earth-silicon-silicate composition EBC systems for the SiC-SiC CMC turbine component applications. Advanced hafnium-based compositions for enabling next generation EBC and CMCs capabilities towards ultra-high temperature ceramic coating systems will also be briefly mentioned.

Ceramic Matrix Composite↗

Kinetic Monte Carlo Simulations of Diffusion in Environmental Barrier Coating Materials

Ceramic Matrix Components (CMC) components for use in turbine engines offer a number of advantages compared with current practice. However, such components are subject to degradation through a variety of mechanisms. In particular, in the hot environment inside a turbine in operation a considerable amount of water vapor is present, and this can lead to corrosion and recession. Environmental Barrier Coating (EBC) systems that limit the amount of oxygen and water reaching the component are required to reduce this degradation and extend component life. A number of silicate-based materials are under consideration for use in such coating systems, including Yttterbium and Yttrium di- and monosilicates. In this work, we present results of kinetic Monte Carlo computer simulations of oxygen diffusion in Yttrium disilicate, and compare with previous work on Yttterbium disilicate. Coatings may also exhibit cracking, and the cracks can provide a direct path for oxygen to reach the component. There is typically a bond coat between the coating and component surface, but the bond coat material is generally chosen for properties other than low oxygen diffusivity. Nevertheless, the degree to which the bond coat can inhibit oxygen diffusion is of interest, as it may form the final defense against oxygen impingement on the component. We have therefore performed similar simulations of oxygen diffusion through HfSiO4, a proposed bond coat material.

Computer Simulation↗

Kinetic Monte Carlo Simulation of Cation Diffusion in Low-K Ceramics

Low thermal conductivity (low-K) ceramic materials are of interest to the aerospace community for use as the thermal barrier component of coating systems for turbine engine components. In particular, zirconia-based materials exhibit both low thermal conductivity and structural stability at high temperature, making them suitable for such applications. Because creep is one of the potential failure modes, and because diffusion is a mechanism by which creep takes place, we have performed computer simulations of cation diffusion in a variety of zirconia-based low-K materials. The kinetic Monte Carlo simulation method is an alternative to the more widely known molecular dynamics (MD) method. It is designed to study "infrequent-event" processes, such as diffusion, for which MD simulation can be highly inefficient. We describe the results of kinetic Monte Carlo computer simulations of cation diffusion in several zirconia-based materials, specifically, zirconia doped with Y, Gd, Nb and Yb. Diffusion paths are identified, and migration energy barriers are obtained from density functional calculations and from the literature. We present results on the temperature dependence of the diffusivity, and on the effects of the presence of oxygen vacancies in cation diffusion barrier complexes as well.

Good, Brian↗

Kinetic Monte Carlo Simulation of Oxygen and Cation Diffusion in Yttria-Stabilized Zirconia

Yttria-stabilized zirconia (YSZ) is of interest to the aerospace community, notably for its application as a thermal barrier coating for turbine engine components. In such an application, diffusion of both oxygen ions and cations is of concern. Oxygen diffusion can lead to deterioration of a coated part, and often necessitates an environmental barrier coating. Cation diffusion in YSZ is much slower than oxygen diffusion. However, such diffusion is a mechanism by which creep takes place, potentially affecting the mechanical integrity and phase stability of the coating. In other applications, the high oxygen diffusivity of YSZ is useful, and makes the material of interest for use as a solid-state electrolyte in fuel cells. The kinetic Monte Carlo (kMC) method offers a number of advantages compared with the more widely known molecular dynamics simulation method. In particular, kMC is much more efficient for the study of processes, such as diffusion, that involve infrequent events. We describe the results of kinetic Monte Carlo computer simulations of oxygen and cation diffusion in YSZ. Using diffusive energy barriers from ab initio calculations and from the literature, we present results on the temperature dependence of oxygen and cation diffusivity, and on the dependence of the diffusivities on yttria concentration and oxygen sublattice vacancy concentration. We also present results of the effect on diffusivity of oxygen vacancies in the vicinity of the barrier cations that determine the oxygen diffusion energy barriers.

Good, Brian↗

Multiscale Computer Simulation of Tensile and Compressive Strain in Polymer- Coated Silica Aerogels

While the low thermal conductivities of silica aerogels have made them of interest to the aerospace community as lightweight thermal insulation, the application of conformal polymer coatings to these gels increases their strength significantly, making them potentially useful as structural materials as well. In this work we perform multiscale computer simulations to investigate the tensile and compressive strain behavior of silica and polymer-coated silica aerogels. Aerogels are made up of clusters of interconnected particles of amorphous silica of less than bulk density. We simulate gel nanostructure using a Diffusion Limited Cluster Aggregation (DLCA) procedure, which produces aggregates that exhibit fractal dimensions similar to those observed in real aerogels. We have previously found that model gels obtained via DLCA exhibited stress-strain curves characteristic of the experimentally observed brittle failure. However, the strain energetics near the expected point of failure were not consistent with such failure. This shortcoming may be due to the fact that the DLCA process produces model gels that are lacking in closed-loop substructures, compared with real gels. Our model gels therefore contain an excess of dangling strands, which tend to unravel under tensile strain, producing non-brittle failure. To address this problem, we have incorporated a modification to the DLCA algorithm that specifically produces closed loops in the model gels. We obtain the strain energetics of interparticle connections via atomistic molecular statics, and abstract the collective energy of the atomic bonds into a Morse potential scaled to describe gel particle interactions. Polymer coatings are similarly described. We apply repeated small uniaxial strains to DLCA clusters, and allow relaxation of the center eighty percent of the cluster between strains. The simulations produce energetics and stress-strain curves for looped and nonlooped clusters, for a variety of densities and interaction parameters.

Good, Brian↗

Heats of Segregation of BCC Metals Using Ab Initio and Quantum Approximate Methods

Many multicomponent alloys exhibit surface segregation, in which the composition at or near a surface may be substantially different from that of the bulk. A number of phenomenological explanations for this tendency have been suggested, involving, among other things, differences among the components' surface energies, molar volumes, and heats of solution. From a theoretical standpoint, the complexity of the problem has precluded a simple, unified explanation, thus preventing the development of computational tools that would enable the identification of the driving mechanisms for segregation. In that context, we investigate the problem of surface segregation in a variety of bcc metal alloys by computing dilute-limit heats of segregation using both the quantum-approximate energy method of Bozzolo, Ferrante and Smith (BFS), and all-electron density functional theory. In addition, the composition dependence of the heats of segregation is investigated using a BFS-based Monte Carlo procedure, and, for selected cases of interest, density functional calculations. Results are discussed in the context of a simple picture that describes segregation behavior as the result of a competition between size mismatch and alloying effects

Good, Brian↗

Surface Segregation in Ternary Alloys

Surface segregation profiles of binary (Cu-Ni, Au-Ni, Cu-Au) and ternary (Cu-Au-Ni) alloys are determined via Monte Carlo-Metropolis computer simulations using the BFS method for alloys for the calculation of the energetics. The behavior of Cu or Au in Ni is contrasted with their behavior when both are present. The interaction between Cu and Au and its effect on the segregation profiles for Cu-Au-Ni alloys is discussed.

Good, Brian↗