CuInGaSe 2 films on 6-inch wide Mo coated polyimide web substrate in a roll-to-roll vapor deposition system from elemental sources is described. Material transport from the sources to the moving web substrate has been modeled by combining an evaporative effusion model and the gas flow kinetics and by experimentally determining the flux intensity profile. The model gives a reasonably good approximation of the Ga profile in the films. Poor adhesion of the selenide film to molybdenum has been resolved by depositing a thin layer of b-(Ga 0Á8 In 0Á2 ) 2 Se 3 precursor layer. Two-dimensional compositional mapping by energy dispersive spectroscopy of 5feet-long web gave 0Á88 AE 1Á9% and 0Á28 AE 2Á5% respectively for Cu/(Ga þ In) and Ga/(Ga þ In) ratios, indicating highly uniform film composition. This compositional uniformity translated to the uniformity in the devices fabricated on the web. The open-circuit voltages of the devices from the centerline of the 5-feet web were measured to be 0Á529 AE 0Á86%. A two-dimensional device efficiency survey gave 9Á2 AE 1%. The process was able to produce high-quality material, as defined by the best device efficiency, for a wide range of Cu/(Ga þ In) ratios. The best efficiency achieved so far was 12Á1% with MgF 2 anti-reflection coating.

A vapor transport process for continuous deposition of elemental and compound thin film materials is presented. The process saturates a carrier gas with a vapor from a subliming source. The saturated mixture is directed over a substrate at lower temperature, resulting in a supersaturation condition and subsequent film growth. The process geometry, comprising the dimensions of the saturation and deposition zones, carrier gas pressure and flow rate, and saturation zone temperature are determined by calculating worst-case characteristic times and simply insuring that the residence time of the carrier gas sufficiently exceeds these times. A model was used to design a system, which is currently being used to deposit 1–10μm thick CdTe films on a 10×10cm2 translating substrate. The process produces film thickness uniformity to within ±5% in the translation direction and across the deposition zone, with a material utilization of 50%. Linear translation speed of 12.5cm∕min has been demonstrated in depositing a 4.5μm CdTe film. The vapor transport process has also been used to deposit CdxZn1−xTe alloy films over a wide range of compositions by addition of ZnTe to the source. Photovoltaic conversion efficiencies of >13% for CdTe and >12% for CdxZn1−xTe have been achieved by devices fabricated from vapor transport deposited films deposited on to moving CdS coated substrates. Refinements are suggested for commercial-scale deposition.

Ag,Cu)(In,Ga)Se 2 (ACIGS) solar cells are optimized at bandgaps greater than 1.2 eV by varying composition profile of the absorber layer using a three-stage evaporation process. Numerical modeling and cumulative process data provides insight into the process. Silver alloying CIGS changes the optimized bandgap profile by reducing carrier concentration, and reducing bandgap gradients. The minimum bandgap position is controlled by the point when the film reaches I/III stoichiometry during the second stage of the three-stage process. We achieved a 19.9% efficient solar cell with V OC = 732 mV at a bandgap of 1.2 eV based on quantum efficiency.

No abstract

To meet the stringent requirements of interconnect metallization for sub-32 nm technologies, an unprecedented level of flux and energy control of film forming species has become necessary to further advance ionized physical vapor deposition technology. Such technology development mandates improvements in methods to quantify the metal ion fraction, the gas∕metal ion ratio, and the associated ion energies in the total ion flux to the substrate. In this work, a novel method combining planar Langmuir probes, quartz crystal microbalance (QCM), and gridded energy analyzer (GEA) custom instrumentation is developed to estimate the plasma density and temperature as well as to measure the metal ion fraction and ion energy. The measurements were conducted in a Novellus Systems, Inc. Hollow Cathode Magnetron (HCM(TM)) physical vapor deposition source used for deposition of Cu seed layer for 65-130 nm technology nodes. The gridded energy analyzer was employed to measure ion flux and ion energy, which was compared to the collocated planar Langmuir probe data. The total ion-to-metal neutral ratio was determined by the QCM combined with GEA. The data collection technique and the corresponding analysis are discussed. The effect of concurrent resputtering during the deposition process on film thickness profile is also discussed.

Even though rapid advances have been made in improving Cu(InGa)Se 2 thin-film-based solar cell efficiencies, the breakthroughs have been limited to the laboratory scale. Most commercially viable thin-film technologies reside at the premanufacturing development stage and scaleup has proven to be much more difficult than expected. Elemental in-line evaporation on flexible substrates in a roll-to-roll configuration is a commercially attractive process for the manufacture of large-area CuInSe 2 -based photovoltaics. At the University of Delaware's Institute of Energy Conversion (IEC), such a process is being investigated, at the pilot scale, for a polyimide web substrate. The process works well for 6-in.-wide substrates and for short deposition runtimes. However, a commercially viable process is required to produce large-area, high-quality films at a muchhigher throughput. Specifically, the desired film thickness (∼2 µm) and composition uniformity must be achieved continuously and reproducibly on large-area (12-in.-wide) substrates at translation speeds (∼1 ft/ min) much higher than those currently used in pilot-scale processes. Although achievement of the desired film thickness and composition setpoints is best addressed by proper control system design, the film thickness uniformity is determined by the source design and individual nozzle effusion rates. The nozzle effusion rates, in turn, are dependent on the melt surface temperature profile and, thus, on the source design itself. Therefore, proper source design is critical in achieving film thickness uniformity. We have identified two modeling requirements for effective thermal evaporation source design: (i) a detailed three-dimensional thermal model of the evaporation source, to predict the melt surface temperature accurately, and (ii) a nozzle effusion model, to predict the effusion rate and the vapor flux distribution for a given nozzle geometry (length and diameter), melt surface temperature, and evaporant. To meet the first requirement, we have developed a first-principles three-dimensional electrothermal model using COMSOL Multiphysics software; the Direct Simulation Monte Carlo (DSMC) method has been used for effusion modeling. In this paper, which is the first of a two-part series, we present the details of the two models and their experimental validation.

We recently presented the development and experimental validation of two models that are essential for effective commercial-scale source design for Cu(InGa)Se2 thin film coevaporative physical vapor deposition processes: a three-dimensional thermal model of the evaporation source, and a Direct Simulation Monte Carlo (DSMC)-method-based effusion model. We showed that these models can be used to obtain reasonably accurate melt temperature dynamics and nozzle effusion flow properties. We now present how these simulation tools are used to develop a scale-up methodology for the effective commercialization of the pilot-scale physical vapor deposition (PVD) process at the University of Delaware’s Institute of Energy Conversion (IEC). We illustrate the methodology using two commercial-scale source design studies: a three-nozzle single source and a four-nozzle modular source. We also show that the proposed source designs are robust to modeling errors and the important process parameter of the source-to-substrate distance.

No abstract