ABSTRACT
Bandgap gradient is a proven approach for improving the open-circuit voltages (VOCs) in Cu(In,Ga)Se2 and Cu(Zn,Sn)Se2 thin-film solar cells, but has not been realized in Cd(Se,Te) thin-film solar cells, a leading thin-film solar cell technology in the photovoltaic market. Here, we demonstrate the realization of a bandgap gradient in Cd(Se,Te) thin-film solar cells by introducing a Cd(O,S,Se,Te) region with the same crystal structure of the absorber near the front junction. The formation of such a region is enabled by incorporating oxygenated CdS and CdSe layers. We show that the introduction of the bandgap gradient reduces the hole density in the front junction region and introduces a small spike in the band alignment between this and the absorber regions, effectively suppressing the nonradiative recombination therein and leading to improved VOCs in Cd(Se,Te) solar cells using commercial SnO2 buffers. A champion device achieves an efficiency of 20.03% with a VOC of 0.863 V.
ABSTRACT
The performance of CdTe solar cells has advanced impressively in recent years with the incorporation of Se. Instabilities associated with light soaking and copper reorganization have been extensively examined for the previous generation of CdS/CdTe solar cells, but instabilities in Cu-doped Se-alloyed CdTe devices remain relatively unexplored. In this work, we fabricated a range of CdSe/CdTe solar cells by sputtering CdSe layers with thicknesses of 100, 120, 150, 180, and 200 nm on transparent oxide-coated glass and then depositing CdTe by close-spaced sublimation. After CdCl2 annealing, Cu-doping, and back metal deposition, a variety of analyses were performed both before and after light soaking to understand the changes in device performance. The device efficiency was degraded with light soaking in most cases, but devices fabricated with a CdSe layer thickness of 120 nm showed reasonably good efficiency initially (13.5%) and a dramatic improvement with light soaking (16.5%). The efficiency improvement is examined within the context of Cu ion reorganization that is well known for CdS/CdTe devices. Low-temperature photoluminescence data and Voc versus temperature measurements indicate a reduction in nonradiative recombination due to the passivation of defects and defect complexes in the graded CdSexTe1-x layer.
ABSTRACT
Motivated by their utility in CdTe-based thin film photovoltaics (PV) devices, an investigation of thin films of the magnesium-zinc oxide (MgxZn1-xO or MZO) alloy system was undertaken applying spectroscopic ellipsometry (SE). Dominant wurtzite phase MZO thin films with Mg contents in the range 0 ≤ x ≤ 0.42 were deposited on room temperature soda lime glass (SLG) substrates by magnetron co-sputtering of MgO and ZnO targets followed by annealing. The complex dielectric functions ε of these films were determined and parameterized over the photon energy range from 0.73 to 6.5 eV using an analytical model consisting of two critical point (CP) oscillators. The CP parameters in this model are expressed as polynomial functions of the best fitting lowest CP energy or bandgap E0 = Eg, which in turn is a quadratic function of x. As functions of x, both the lowest energy CP broadening and the Urbach parameter show minima for x ~ 0.3, which corresponds to a bandgap of 3.65 eV. As a result, it is concluded that for this composition and bandgap, the MZO exhibits either a minimum concentration of defects in the bulk of the crystallites or a maximum in the grain size, an observation consistent with measured X-ray diffraction line broadenings. The parametric expression for ε developed here is expected to be useful in future mapping and through-the-glass SE analyses of partial and complete PV device structures incorporating MZO.
ABSTRACT
Although back-surface passivation plays an important role in high-efficiency photovoltaics, it has not yet been definitively demonstrated for CdTe. Here, we present a solution-based process, which achieves passivation and improved electrical performance when very small amounts of oxidized Al3+ species are deposited at the back surface of CdTe devices. The open circuit voltage (Voc) is increased and the fill factor (FF) and photoconversion efficiency (PCE) are optimized when the total amount added corresponds to â¼1 monolayer, suggesting that the passivation is surface specific. Addition of further Al3+ species, present in a sparse alumina-like layer, causes the FF and PCE to drop as the interface layer becomes blocking to current flow. The optimized deposit increases the average baseline PCE for both Cu-free devices and devices where Cu is present as a dopant. The greatest improvement is found when the Al3+ species are deposited prior to the CdCl2 activation step and Cu is employed. In this case, the best-cell efficiency was improved from 12.6 to 14.4%. Time-resolved photoluminescence measurements at the back surface and quantum efficiency measurements performed at the maximum power point indicate that the performance enhancement is due to a reduction in the interface recombination current at the back surface.
ABSTRACT
There has been a substantial growth in the deployment of solar photovoltaic (PV) panels in the past couple decades. Solar PVs have a life span of about 25 years and much of the deployed PVs will soon reach their end of life (EoL). It is now timely to plan for the EoL of PVs to recover valuable materials and recycle PV modules sustainably. The goal of this study was to analyze the environmental impacts of different recycling methods for crystalline silicon (c-Si) and CdTe panels. A life cycle assessment (LCA) was performed for delamination and material separation phases of recycling solar panels. The LCA results showed that the recycling of c-Si and CdTe PVs contribute 13-25% and 3-4%, respectively to the entire PV lifecycle impacts. Also, for both c-Si and CdTe PVs, the thermal-based recycling methods resulted in lower environmental impacts than chemical and mechanical methods, except for pyrolysis. Nitric acid dissolution used for c-Si PV recycling had the highest impacts among all methods since the material consumption for this method has not been optimized for industrial use. Results from this study suggested that current techniques used in recycling of PVs, produce higher impacts than extraction of Al, Si and glass for c-Si and extraction of glass for CdTe. Lastly, this study identified which materials to prioritize for highest economic and environmentals benefits from recycling. These will be Ag, Al, Si, and glass in c-Si modules, and Te, Cu, and glass in CdTe modules.
ABSTRACT
The replacement of traditional CdS with zinc magnesium oxide (ZMO) has been demonstrated as being helpful to boost power conversion efficiency of cadmium telluride (CdTe) solar cells to over 18%, due to the reduced interface recombination and parasitic light absorption by the buffer layer. However, due to the atmosphere sensitivity of ZMO film, the post treatments of ZMO/CdTe stacks, including CdCl2 treatment, back contact deposition, etc., which are critical for high-performance CdTe solar cells became crucial challenges. To realize the full potential of the ZMO buffer layer, plenty of investigations need to be accomplished. Here, copper thiocyanate (CuSCN) is demonstrated to be a suitable back-contact material with multi-advantages for ZMO/CdTe solar cells. Particularly, ammonium hydroxide as the solvent for CuSCN deposition shows no detrimental impact on the ZMO layer during the post heat treatment. The post annealing temperature as well as the thickness of CuSCN films are investigated. Finally, a champion power conversion efficiency of 16.7% is achieved with an open-circuit voltage of 0.857 V, a short-circuit current density of 26.2 mA/cm2, and a fill factor of 74.0%.
ABSTRACT
Organic-inorganic metal halide perovskites are notoriously unstable in humid environments. While many studies have revealed the morphology and crystal structure changes that accompany exposure to humidity, little is known about changes to the photophysics that accompany the degradation process. By combining in situ steady-state and time-resolved photoluminescence with Hall effect measurements, we examined the changes in the photoexcited carrier dynamics for methylammonium lead iodide (MAPbI3) and bromide (MAPbBr3) films exposed to nitrogen gas containing water vapor at 80% relative humidity. The changes in the photophysics of MAPbI3 interacting with water follow a four-stage process, consisting of surface passivation, free electron doping, interfacial hydration, and bulk hydration. In contrast, MAPbBr3 exhibits only features associated with the first two stages, which occur at a faster rate. Our results elucidate the degradation mechanisms of perovskite films in high humidity from the perspective of the photophysics, providing insights for how humidity affects the stability of the perovskite materials.
ABSTRACT
Real time spectroscopic ellipsometry (RTSE) has been applied for in-situ monitoring of the first stage of copper indium-gallium diselenide (CIGS) thin film deposition by the three-stage co-evaporation process used for fabrication of high efficiency thin film photovoltaic (PV) devices. The first stage entails the growth of indium-gallium selenide (In1-xGax)2Se3 (IGS) on a substrate of Mo-coated soda lime glass maintained at a temperature of 400 °C. This is a critical stage of CIGS deposition because a large fraction of the final film thickness is deposited, and as a result precise compositional control is desired in order to achieve the optimum performance of the resulting CIGS solar cell. RTSE is sensitive to monolayer level film growth processes and can provide accurate measurements of bulk and surface roughness layer thicknesses. These in turn enable accurate measurements of the bulk layer optical response in the form of the complex dielectric function ε = ε1 - iε2, spectra. Here, RTSE has been used to obtain the (ε1, ε2) spectra at the measurement temperature of 400 °C for IGS thin films of different Ga contents (x) deduced from different ranges of accumulated bulk layer thickness during the deposition process. Applying an analytical expression in common for each of the (ε1, ε2) spectra of these IGS films, oscillator parameters have been obtained in the best fits and these parameters in turn have been fitted with polynomials in x. From the resulting database of polynomial coefficients, the (ε1, ε2) spectra can be generated for any composition of IGS from the single parameter, x. The results have served as an RTSE fingerprint for IGS composition and have provided further structural information beyond simply thicknesses, for example information related to film density and grain size. The deduced IGS structural evolution and the (ε1, ε2) spectra have been interpreted as well in relation to observations from scanning electron microscopy, X-ray diffractometry and energy-dispersive X-ray spectroscopy profiling analyses. Overall the structural, optical and compositional analysis possible by RTSE has assisted in understanding the growth and properties of three stage CIGS absorbers for solar cells and shows future promise for enhancing cell performance through monitoring and control.
ABSTRACT
An ex-ante life cycle inventory was developed for single walled carbon nanotube (SWCNT) PV cells, including a laboratory-made 1% efficient device and an aspirational 28% efficient four-cell tandem device. The environmental impact of unit energy generation from the mono-Si PV technology was used as a reference point. Compared to monocrystalline Si (mono-Si), the environmental impacts from 1% SWCNT was â¼18 times higher due mainly to the short lifetime of three years. However, even with the same short lifetime, the 28% cell had lower environmental impacts than mono-Si. The effects of lifetime and efficiency on the environmental impacts were further examined. This analysis showed that if the SWCNT device efficiency had the same value as the best efficiency of the material under comparison, to match the total normalized impacts of the mono- and poly-Si, CIGS, CdTe, and a-Si devices, the SWCNT devices would need a lifetime of 2.8, 3.5, 5.3, 5.1, and 10.8 years, respectively. It was also found that if the SWCNT PV has an efficiency of 4.5% or higher, its energy payback time would be lower than other existing and emerging PV technologies. The major impacts of SWCNT PV came from the cell's materials synthesis.
Subject(s)
Nanotubes, Carbon/chemistry , Electric Power Supplies , Environment , Solar EnergyABSTRACT
Control over grain size and crystallinity is important for preparation of methylammonium lead iodide (MAPbI3) solar cells. We explore the effects of using small concentrations of Cd2+ and unusually high concentrations of methylammonium iodide during the growth of MAPbI3 in the two-step solution process. In addition to improved crystallinity and an enhancement in the size of the grains, time-resolved photoluminescence measurements indicated a dramatic increase in the carrier lifetime. As a result, devices constructed with the Cd-modified perovskites showed nearly a factor of 2 improvement in the power conversion efficiency (PCE) relative to similar devices prepared without Cd addition. The grains also showed a higher degree of orientation in the ⟨110⟩ direction, indicating a change in the growth mechanism, and the films were compact and smooth. We propose a Cd-modified film growth mechanism that invokes a critical role for low-dimensional Cd perovskites to explain the experimental observations.
ABSTRACT
Fatty nitriles are widely used as intermediate molecules in the pharmaceutical and polymer industries. In addition, hydrogenation of fatty nitriles produces fatty amines that are common surfactants. In the conventional fatty nitrile process, triglycerides are first hydrolyzed and the resulting fatty acids are catalytically reacted with NH3 in a liquid-phase reaction. In this study, we report a simpler one-step fatty nitrile production method that involves a direct vapor-phase reaction of triglycerides with NH3 in the presence of heterogeneous solid acid catalysts. The reactions were performed in a tubular reactor maintained at 400 °C into which triglycerides were injected through an atomizer to allow rapid volatilization and reaction; NH3 was fed as a gas. Several metal oxide catalysts were tested, and reactions in the presence of V2O5 resulted in near-theoretical fatty nitrile yields (84 wt % relative to the feed mass). In general, catalysts with higher acidity such as V2O5, Fe2O3, and ZnO showed higher fatty nitrile yields compared to low acidity catalysts such as ZrO, Al2O3, and CuO. Energy balance calculations indicate that the one-step reaction described here would require significantly lower energy than the conventional process primarily because of the elimination of the energy-intense triglyceride hydrolysis.
ABSTRACT
Perovskite/silicon tandem solar cells with high power conversion efficiencies have the potential to become a commercially viable photovoltaic option in the near future. However, device design and optimization is challenging because conventional characterization methods do not give clear feedback on the localized chemical and physical factors that limit performance within individual subcells, especially when stability and degradation is a concern. In this study, we use light beam induced current (LBIC) to probe photocurrent collection nonuniformities in the individual subcells of perovskite/silicon tandems. The choices of lasers and light biasing conditions allow efficiency-limiting effects relating to processing defects, optical interference within the individual cells, and the evolution of water-induced device degradation to be spatially resolved. The results reveal several types of microscopic defects and demonstrate that eliminating these and managing the optical properties within the multilayer structures will be important for future optimization of perovskite/silicon tandem solar cells.
ABSTRACT
We have developed a laser beam induced current imaging tool for photovoltaic devices and modules that utilizes diode pumped Q-switched lasers. Power densities on the order of one sun (100 mW/cm2) can be produced in a â¼40 µm spot size by operating the lasers at low diode current and high repetition rate. Using galvanostatically controlled mirrors in an overhead configuration and high speed data acquisition, large areas can be scanned in short times. As the beam is rastered, focus is maintained on a flat plane with an electronically controlled lens that is positioned in a coordinated fashion with the movements of the mirrors. The system can also be used in a scribing mode by increasing the diode current and decreasing the repetition rate. In either mode, the instrument can accommodate samples ranging in size from laboratory scale (few cm2) to full modules (1 m2). Customized LabVIEW programs were developed to control the components and acquire, display, and manipulate the data in imaging mode.
ABSTRACT
We have discovered that films of carbon single wall nanotubes (SWNTs) make excellent back contacts to CdTe devices without any modification to the CdTe surface. Efficiencies of SWNT-contacted devices are slightly higher than otherwise identical devices formed with standard Au/Cu back contacts. The SWNT layer is thermally stable and easily applied with a spray process, and SWNT-contacted devices show no signs of degradation during accelerated life testing.
ABSTRACT
In this work, we investigate the impact of the solvation environment on single-walled carbon nanotube (SWCNT) photoluminescence quantum yield and optical transition energies (E(ii)) using a highly charged aryleneethynylene polymer. This novel surfactant produces dispersions in a variety of polar solvents having a wide range of dielectric constants (methanol, dimethyl sulfoxide, aqueous dimethylformamide, and deuterium oxide). Because a common surfactant can be used while maintaining a constant SWCNT-surfactant morphology, we are able to straightforwardly evaluate the impact of the solvation environment upon SWCNT optical properties. We find that (i) the SWCNT quantum yield is strongly dependent on both the polarity and electrophilicity of the solvent and (ii) solvatochromic shifts correlate with the extent of SWCNT solvation. These findings provide a deeper understanding of the environmental dependence of SWCNT excitonic properties and underscore that the solvent provides a tool with which to modulate SWCNT electronic and optical properties.
ABSTRACT
We studied the electrocatalytic activity of an [FeFe]-hydrogenase from Clostridium acetobutylicum (CaH2ase) immobilized on single-wall carbon nanotube (SWNT) networks. SWNT networks were prepared on carbon cloth by ultrasonic spraying of suspensions with predetermined ratios of metallic and semiconducting nanotubes. Current densities for both proton reduction and hydrogen oxidation electrocatalytic activities were at least 1 order of magnitude higher when hydrogenase was immobilized onto SWNT networks with high metallic tube (m-SWNT) content in comparison to hydrogenase supported on networks with low metallic tube content or when SWNTs were absent. We conclude that the increase in electrocatalytic activities in the presence of SWNTs was mainly due to the m-SWNT fraction and can be attributed to (i) substantial increases in the active electrode surface area, and (ii) improved electronic coupling between CaH2ase redox-active sites and the electrode surface.
Subject(s)
Hydrogen/chemistry , Hydrogenase/metabolism , Iron-Sulfur Proteins/metabolism , Nanotubes, Carbon/chemistry , Biocatalysis , Clostridium acetobutylicum/enzymology , Electrochemistry , Electrodes , Hydrogenase/chemistry , Iron-Sulfur Proteins/chemistry , Models, Molecular , Oxidation-Reduction , Particle Size , Surface PropertiesABSTRACT
Solid-state (13)C NMR spectroscopy was used to investigate the chemical shift of nanotube carbons on m- and s-SWNTs (metallic and semiconducting single-walled nanotubes) for samples with widely varying s-SWNT content, including samples highly enriched with nearly 100% m- and s-SWNTs. High-resolution (13)C NMR was found to be a sensitive probe for m- and s-SWNTs in mixed SWNT samples with diameters of approximately 1.3 nm. The two highly enriched m- and s-SWNT samples clearly exhibited features for m- and s-SNWT (13)C nuclei (approximately 123 and 122 ppm, respectively) and were successfully fit with a single Gaussian, while five mixed samples required two Gaussians for a satisfactory fit.
ABSTRACT
Single-walled carbon nanotube (SWCNT) fibers were engineered to become a scaffold for the storage of hydrogen. Carbon nanotube fibers were swollen in oleum (fuming sulfuric acid), and organic spacer groups were covalently linked between the nanotubes using diazonium functionalization chemistry to provide 3-dimensional (3-D) frameworks for the adsorption of hydrogen molecules. These 3-D nanoengineered fibers physisorb twice as much hydrogen per unit surface area as do typical macroporous carbon materials. These fiber-based systems can have high density, and combined with the outstanding thermal conductivity of carbon nanotubes, this points a way toward solving the volumetric and heat-transfer constraints that limit some other hydrogen-storage supports.
ABSTRACT
We report a Raman spectroscopy study of charge transfer interactions in complexes formed by single-walled carbon nanotubes (SWNTs) and [FeFe] hydrogenase I (CaHydI) from Clostridium acetobutylicum. The choice of Raman excitation wavelength and sample preparation conditions allows differences to be observed for complexes involving metallic (m) and semiconducting (s) species. Adsorbed CaHydI can reversibly inject electronic charge into the LUMOs of s-SWNTs, while charge can be injected and removed from m-SWNTs at lower potentials just above the Fermi energy. Time-dependent enzymatic assays demonstrated that the reduced and oxidized forms of CaHydI are deactivated by oxygen, but at rates that varied by an order of magnitude. The time evolution of the oxidative decay of the CaHydI activity reveals different time constants when complexed with m-SWNTs and s-SWNTs. The correlation of enzymatic assays with time-dependent Raman spectroscopy provides a novel method by which the charge transfer interactions may be investigated in the various SWNT-CaHydI complexes. Surprisingly, an oxidized form of CaHydI is apparently more resistant to oxygen deactivation when complexed to m-SWNTs rather than s-SWNTs.
Subject(s)
Hydrogenase/chemistry , Hydrogenase/metabolism , Iron-Sulfur Proteins/chemistry , Iron-Sulfur Proteins/metabolism , Nanotubes, Carbon/chemistry , Clostridium/enzymology , Ions/chemistry , Kinetics , Spectrum Analysis, RamanABSTRACT
The ensemble PL quantum yield for raw single-walled carbon nanotubes (SWNTs) dispersed in sodium cholate (SC) is approximately 5 times greater than that for the same raw SWNTs dispersed in sodium dodecyl sulfate (SDS) and approximately 10 times greater than the quantum yield of purified SWNTs dispersed in SC. Absorbance and Raman spectra indicate that purified SC-dispersed SWNTs and raw SDS-dispersed SWNTs are hole-doped by protonation. Experiments comparing PL emission efficiency using E2 and E1 excitation show that protonation significantly affects the E2 --> E1 relaxation process, which has typically been assumed to occur with unit efficiency. The E2 --> E 1 relaxation is 5 times more efficient in producing E 1 PL when SWNTs are unprotonated and protected by the SC surfactant. The results provide clear evidence that extrinsic factors, such as residual acids and the specific nature of SWNT-surfactant and SWNT-solvent interactions, can significantly affect measured SWNT luminescence quantum yields.