Kurt L. Barth

Advances in continuous, in-line processing of stable CdS/CdTe devices

A continuous, in-line process suitable for high throughput manufacturing of CdS/CdTe photovoltaic devices has been demonstrated. Utilizing this process, devices with efficiencies of 13% has been fabricated with a low iron soda lime glass (3″×3″) with ant-reflection coatings. The process has been extended to large area devices (16″ ×16″ substrate size). After CdCl<inf>2</inf> treatment, devices showed V<inf>oc</inf> ≫ 700 mV and J<inf>sc</inf> ≫ 20 mA/cm². This performance is similar to the performance of small area devices which showed good stability. Also we have employed Spectroscopic Ellipsometry (SE) as a nondestructive tool to characterize CdS/CdTe heterojunction specifically studying the effects of chemical treatment on the optical properties of the thin-film layers.

Effect of back-contact copper concentration on CdTe cell operation

CdTe solar cells were fabricated with five different concentrations of copper, including zero, used in back-contact formation. Room-temperature J-V curves showed progressive deterioration in fill factor with reduced copper. J/sub SC/ and QE were similar for all Cu-levels. Capacitance measurement suggested enhanced intermixing at the back contact with copper present. Photocurrent mapping was much less uniform for reduced-Cu cells. Elevated-temperature stress induced very little change in J-V when sufficient Cu was used in the contact.

Single vacuum chamber with multiple close space sublimation sources to fabricate CdTe solar cells

Photovoltaic technologies have shown efficiencies of over 40%, however, manufacturing costs have prevented a more significant energy market penetration. To bridge the gap between the high efficiency technology and low cost manufacturing, a research and development tool and process was built and tested. This fully automated single vacuum photovoltaic manufacturing tool utilizes multiple inline close space sublimation (CSS) sources with automated substrate control. This maintains the proven scalability of the CSS technology and CSS source design but with the added versatility of independent substrate motion. This combination of a scalable deposition technology with increased cell fabrication flexibility has allowed for high efficiency cells to be manufactured and studied. The single vacuum system is capable of fabricating a 3.1 × 3.6 in. substrate every 45 min with a cell efficiency of 12% with a standard deviation of 0.6% as measured over 36 months. The substrate is generally scribed into 25 small area dev...

Transient Ion Drift Measurements of Polycrystalline CdTe PV Devices

The well known transient ion drift (TID) method is used to quantify the density of mobile Cu interstitial ions in polycrystalline CdTe PV cells. Average Cu_i ⁺ ion densities in optimally processed cells are 20% of the background ionized acceptor doping level. A preliminary estimate of the diffusion coefficient for Cu _i ⁺ ions in the polycrystalline CdTe absorber is D(Cu_i)=1.3E-6 [cm²/sec] times exp[-0.29 eV/(KB*T)] in the temperature range of 25degC to 55degC from TID measurements. Aspects of the TID method as pertains to practical thin film polycrystalline devices are discussed

Stable Cu-Based Back Contacts for CdTe Thin Film Photovoltaic Devices

Cadmium telluride (CdTe) thin film photovoltaic devices fabricated in a-line process developed at Colorado State University (CSU) have shown stability during long-term (over a 5 year period) accelerated stress testing. These devices have a copper (Cu) containing back contact. The Cu profile as measured by secondary ion mass spectrometry characterization shows, for the maximum stressed device (23,399 h), that there is a significant (two times) change in the concentration of secondary Cu ions in the bulk of the material; however, the Cu concentration gradient at the back of the device has no significant change, and the CdS layer has no significant Cu concentration increase at open-circuit bias and 65°C temperature conditions. This indicates that with a proper CdCl 2 treatment, Cu can be used to form the back contact for CdTe devices with acceptable stability. These devices have a projected field lifetime of greater than 60 years.

Consistent processing and long term stability of CdTe devices

A technology for processing of thin film CdS/CdTe devices has been developed in our laboratory. This in-line, continuous, pilot system enables unique processing steps and conditions not available with batch processing and allows the fabrication of a large number of devices. Results from the pilot scale system are applicable to systems processing larger areas. Utilizing the pilot system, significant progress has been made towards demonstrating consistent stability (resistance to degradation) for thin film CdTe photovoltaics. We have repeatedly shown that devices with good stability can be produced if processed at the optimum set of conditions. Small changes in processes can lead to significant differences in device stability. Among the processing steps for fabrication of CdTe devices, the CdCl/sub 2/ treatment has a significant effect on performance and stability. A metric has been developed to predict the stability of devices at the time of device fabrication. Accelerated stress testing is ongoing. Extremely long duration stress testing (65/spl deg/C, open circuit conditions for /spl sim/30,000 hours with 5 hours of illumination out of 8 hour cycle) has demonstrated that the rate of efficiency loss levels out with final efficiencies in the range of 8.5% /spl sim/ 9.5%. A production prototype system for processing nominally 2 MW/yr. is currently under construction. This system utilizes the process definition developed in the pilot system.

Cadmium chloride assisted re-crystallization of CdTe: The effect of annealing over-treatment

Although the cadmium chloride treatment is an essential process for high efficiency thin film cadmium telluride photovoltaic devices, the precise mechanisms involved that improve the cadmium telluride layer are not fully understood. The treatment parameters have a narrow window, deviating from these even slightly can be detrimental to cell performance. In this investigation we apply advanced microstructural characterization techniques to study the effects of varying two parameters: the temperature of the substrate during the cadmium chloride treatment and the length of time of the treatment. In both cases, the devices have been deliberately over-treated. The effect of the over-treatment on the microstructure of cadmium telluride solar cells, deposited by close spaced sublimation is investigated and related to cell performance. A range of techniques has been used to observe the changes to the microstructure as well as the chemical and crystallographic changes as a function of treatment parameters. Electrical tests that link the device performance with the microstructural properties of the cells have also been undertaken. Techniques used include Transmission Electron Microscopy (TEM) for sub-grain analysis, EDX for chemical analysis and XPS for composition-depth profiling.

All-CSS processing of CdS/CdTe thin-film solar cells with thin CdS layers

Cadmium Sulfide/Cadmium Telluride (CdS/CdTe) thin-film solar cells were fabricated by an in-line, close-space-sublimation (CSS) process at Colorado State University. Source temperature control was used to reduce the deposited CdS thickness. Quantum efficiency (QE) showed CdS thicknesses that varied over a range from 250 to 10 nm. Current-Voltage (J-V) measurements showed increased J sc as CdS was thinned. Thin CdS resulted in reduced voltage (800 mV to 350 mV) and fill factor, which offset gains in current, and caused efficiencies to drop from 12.6% for thick CdS layers to 4.5% for devices with the thinnest CdS. These performance trends are consistent with calculations assuming parallel junctions of CdS/CdTe and SnO 2 /CdTe. Localized weak-junction formation was characterized by high-resolution laser-beam-induced current (LBIC) mapping. Greater incidence of spatial non-uniformities in photocurrent response accompanied thinning of the CdS layer, with 638-nm spectral response varying spatially by 4.5% for thin CdS devices compared to variations less than 1% for devices with thicker CdS. Non-uniformities of cells with thin CdS are highly sensitive to voltage bias and are likely indicative of parallel p-n and Schottky-type junctions.

Cadmium chloride assisted re-crystallization of CdTe: The effect of the annealing temperature

The aim of this investigation is to apply advanced microstructural characterization techniques to study the effects of varying the cadmium chloride annealing temperature on the microstructure of cadmium telluride solar cells deposited by close spaced sublimation (CSS) and relate this to cell performance. A range of techniques have been used to observe the morphological changes to the microstructure as well as the chemical and crystallographic changes as a function of treatment parameters. Electrical tests that link the device performance with the microstructural properties of the cells have also been undertaken. Techniques used include Transmission Electron Microscopy (TEM) for sub-grain analysis, X-ray Photoelectron Spectroscopy (XPS) depth profiles to show the effect of temperature on the diffusion of chlorine into the CdTe. Grain orientation data as well as grain size change has been obtained using Electron Backscatter Diffraction (EBSD) on Focused Ion Beam (FIB) prepared planar sections.

Effect of Chemical Treatment on the Optical Properties of a Cadmium Telluride Photovoltaic Device Investigated by Spectroscopic Ellipsometry

Spectroscopic ellipsometry has been successfully used to characterize the CdS/CdTe heterojunction solar cell deposited on TEC15 glass. The effects of copper treatment on the optical properties of a cadmium chloride treated photovoltaic device were investigated using ellipsometry. No changes in either the band gaps or critical points of CdTe layer were noticed as a result of copper treatment. The copper treated CdTe layer exhibited a higher refractive index in the visible and longer wavelengths (≤3 eV), as compared with the untreated layer. This was attributed to the increased disorder in the case of copper treated layers.