R&D issues in scale-up and manufacturing of amorphous silicon tandem modules

R & D on amorphous silicon based tandem junction devices has improved the throughtput, the material utilization, and the performance of devices on commercial tin oxide coated glass. The tandem junction technology has been scaled-up to produce 8.6
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Technology transfer challenges in the manufacturing of a-Si tandem solar cells D. E. Carlson, G. Ganguly, G. Lin, M. Gleaton, M. Bennett and R. R. Arya BP Solar, Toano, VA 23168 ABSTRACT BP Solar started commercial production of amorphous silicon tandem solar cells in Toano, Virginia in 1997. The scale-up process has involved overcoming technology challenges in several areas. It was necessary to develop high speed, multiple-beam laser scribing systems that could pattern photovoltaic modules over an area of 0.8 square meters with close dimensional tolerances. In addition, the deposition rate of amorphous silicon was doubled while production was being ramped up, and the deposition system was modified to meet capacity requirements. The utilization of the germane feedstock gas was increased by about 25% in order to reduce material costs. In addition, the dimensional tolerances of the deposition system geometry were determined and controlled in order to assure uniform deposition of the amorphous silicon alloys. Effects of contaminants (such as pump oils and residual dopants) on device performance were quantified. Sources of debris that could cause shunts and shorts in the devices were identified and minimized. Current research efforts are focused on further increases in the amorphous silicon deposition rate, improvements in device performance and the development of in-situ diagnostic tools to monitor and control the manufacturing process. INTRODUCTION The amorphous silicon (a-Si) program at BP Solar has its srcin in the early work at RCA Laboratories performed by Carlson and Wronksi [1]. The RCA program started with the invention of the a-Si solar cell in 1974 [2] and continued through 1983 when the technology was sold to Solarex (which was wholly owned by Amoco Corp. at the time). Solarex started commercial production of small-area (few cm 2 ), single-junction a-Si p-i-n solar cells for low-light level applications in 1984 and then focused on larger monolithic modules (up to 0.1 m 2  in size) for terrestrial applications in 1986. A highly automated production line capable of producing 1 MW P  of 0.1 m 2  single-junction a-Si modules per year was started in 1990 [3]. The line was initially configured so that glass plates were automatically loaded at one end, and about 4 hours later, an encapsulated, monolithic, series-connected module came out the other end without any manual handling. Solarex started research on a-Si multijunction structures in the late 1980’s, but it was not until 1995 that the company decided to commercialize the multijunction technology using an a-Si/a-SiGe tandem structure. This decision followed the successful demonstration of an average 8% stable conversion efficiency for 0.4 m 2  tandem modules produced on a pilot line [4]. Solarex started production of tandem modules (total module area = 0.805 m 2 ) at the TF1 facility in Toano, Virginia in 1997. At this time, Solarex was part of a new joint venture (Amoco/Enron Solar) between Amoco and Enron Corporations. In 1998 BP acquired Amoco, and subsequently Solarex and BP Solar were combined into one company, which was briefly called BP Solarex before being renamed BP Solar in 2000. The TF1 facility was designed to produce 10 MW P  per year of monolithic, high-voltage (V MP   ≈  200 V), tandem modules primarily for large solar farms. The plant was subsequently modified to produce a variety of products ranging from 5 to 50 W P . Mat. Res. Soc. Symp. Proc. Vol. 664 © 2001 Materials Research Society A11.4.1  DESCRIPTION OF THE MANUFACTURING PROGRESS The a-Si tandem structure used in the TF1 manufacturing plant is shown schematically in figure 1. The substrate is 3 mm thick window glass that has been coated with ~ 700 nm of textured tin oxide by a commercial vendor. All of the a-Si alloy layers are deposited using DC plasma-enhanced chemical vapor deposition (PECVD) at substrate temperatures in the range of 190 –  230 ° C. The first deposited layer is p-type a-SiC (which is an amorphous silicon carbon alloy doped with boron). Then ~ 200 nm of undoped a-Si is deposited from a discharge in a mixture of silane and hydrogen followed by ~ 10 nm of a microcrystalline, phosphorus-doped n-layer. A tunnel or recombination junction is then formed by depositing ~ 10 nm of another p-type a-SiC layer. This is followed by ~ 200 nm of an a-SiGe i-layer from a discharge in a mixture of silane, germane and hydrogen and then by ~ 20 nm of a phosphorus-doped a-Si n-layer. The back contact is formed by depositing ~ 100 nm of zinc oxide by low-pressure CVD (LPCVD) and then ~ 200 nm of aluminum by magnetron sputtering. The manufacturing process used in the TF1 facility can be best described by considering the plant layout shown in figure 2. The tin oxide coated glass is purchased in large crates and is automatically loaded from stacks into a glass seamer, which uses diamond-coated abrasion wheels to define the dimensions of the plates and to condition the edges. The plates are then washed, and electrical buss bars are formed on the tin oxide by dispensing a silver glass frit in the appropriate pattern and curing the frit in a belt furnace at ~ 520 ° C. The tin oxide is then scribed by a frequency-doubled Nd-YAG laser to form an array of tin oxide strips, which define the individual solar cells and also predetermine the output voltage and current of the modules. After another washing, the plates are loaded into a multi-chamber deposition system where all the a-Si alloy layers shown in figure 1 are deposited by DC PECVD and the zinc oxide layer is deposited by LPCVD. At this stage, another Nd-YAG laser is used to scribe through the a-Si tandem structure down to the tin oxide layer so that these new scribes are in close proximity to the scribes made earlier in the tin oxide. The plates are subsequently coated with aluminum by magnetron sputtering and are then subjected to a laser-scribing step that scribes the metal and completes the series interconnections [4]. The individual strip cells are typically about 9 mm wide while the laser scribed interconnection regions are about 0.3 mm wide. Thus, the inactive area due to the interconnections constitutes only about 3% of the total area. Figure 1.  A schematic of the a-Si/a-SiGe tandem structure. Glass   Silicon Dioxide   Tin Oxide   a-Si : p-i-n   Tunnel Junction   a-SiGe: p-i-n   ZnO   Aluminum   LIGHT   Front Contact   Amorphous   Alloys   Rear Contact   A11.4.2    Figure 2. The layout of the BP Solar TF1 facility in Toano, VA. A fourth laser scribe is performed around the perimeter of the module to electrically isolate the active region. The module is then cleaned using an ultrasonic bath before undergoing an reverse bias electrical curing treatment to remove shorts and shunts [5] using a “ bed of nails ”  cure station. The photovoltaic performance of the modules is then determined using a solar simulator. A strip of material near the outer perimeter of the module is removed by abrasion prior to lamination to assure good isolation from the environment. The modules are encapsulated by laminating the processed plates to a back plate of glass with ethyl vinyl acetate (EVA). The processing is completed by attaching power leads to the buss bars through holes in the back plate, filing the holes with a sealant, and then framing the modules and attaching a  junction box. RECENT PROGRESS In the first few years of operation, the output of the TF1 facility was limited by a number of problems associated mainly with the scaling up of the processing equipment. Difficulties were encountered with the reproducibility and reliability of the laser scribing systems, with the transport of plates within the amorphous silicon chambers, and with the robustness of the magnetron sputtering system for depositing the Al rear contact. These difficulties were overcome after some effort, and as shown in figure 3, the electrical yields have recently been averaging over 90% and the throughput in excess of 7 MW P  per year. A typical distribution of the initial module output power is shown in figure 4. There is generally an improvement in the module output power of a few percent after lamination and annealing, but the output power drops by about 15 –  17% after light soaking for several months in sunlight before stabilizing. In some recent production runs, the stabilized power of the 0.8 m 2  tandem modules have averaged over 50 W P . Tin Oxide Washer Wiring, Finishing Back Plate Alignment Encapsulation EVA Setup Module Load Station Bed of Nails Solar Simulator Ultrasonic Bath Isolation Laser Metal Laser Sputtering System a-Si Laser Vertical a-Si / ZnO Deposition System Glass Seamer Cure Furnace Frit Dispenser Washer A11.4.3  0%20%40%60%80%100%120% F  e b  - 9  7  N  ov - 9  7   S  e  p- 9   8   J   an- 9   9  M ar  - 9   9  D  e c - 9   9  M ar  - 0   0   S  e  p- 0   0  D  e c - 0   0   J   ul    - 0  1  D  e c - 0  1   Month / Year    P  e  r  c  e  n   t  a  g  e  o   f   C  a  p  a  c   i   t  y 0.0%10.0%20.0%30.0%40.0%50.0%60.0%70.0%80.0%90.0%100.0%    Y   i  e   l   d   Figure 3.  Module throughput and electrical yield as a function of time for the TF1 plant in Toano, Virginia. Figure 4. Distribution of the initial performance for 0.8 m 2  tandem modules produced at the BP Solar TF1 facility. 0 2004006008001000120014001600 35   36   37   38   39   40   41   42   43   44   45   46   47   48   49   50   51   52   53   54   55   56   57   58   59   60   61   62   63   64   65   Initial Power (W)    C  o  u  n   t A11.4.4  TECHNOLOGY TRANSFER CHALLENGES There were a number of technology transfer challenges that had to be overcome to make the TF1 plant operate at a profitable level. The areas that presented significant challenges were: (1) laser scribing, (2) a-Si deposition rate and multi-chamber operation, (3) gas feedstock utilization, (4) dimensional tolerances of the deposition system geometry, (5) determination of the effect of contaminants on device performance and (6) identification and reduction of sources of shorting and shunting. As mentioned in the previous section, the laser scribing systems required some modifications to make the systems more reproducible and reliable. Since the laser scribes in the interconnect region are only about 0.13 mm apart, the mechanical tolerances for a 0.8 m 2  high-voltage module with 132 cells in series are on the order of 0.01%. This tolerance has been met by moving the substrates with a high precision X-Y table under stationary laser beams. A Nd-YAG laser was frequency doubled to operate at a wavelength of 540 nm, and the laser light was split into four beams that simultaneously scribe the substrates. A computer-controlled machine vision system is used to assure accurate registration of the substrates so that all scribes are positioned with 0.13 mm of the previous scribes. When the TF1 facility started operation in 1997, the deposition rate of the a-Si i-layer (in the front junction) was about 1 Å  /s. By re-optimizing the DC PECVD process, the deposition rate has been recently increased to about 2 Å  /s without a significant loss in stabilized performance. The behavior of the light-induced degradation for tandem cells made with i-layers at 1.4 and 2 Å  /s is shown in figure 5 where the stabilized efficiency was about 7% for both types of cells on commercial tin oxide coated glass. We found that the light-induced degradation of the tandem devices was in the range of about 15 –  17% for a-Si i-layer deposition rates ranging from 0.7 to 2.0 Å  /s. The overall throughput of the plant has also been increased by decreasing gas-flushing times and increasing the speed of the substrate transport system. 5   5.5   6   6.5   7   7.5   8   8.5   9   1   10   100   1000   1.4 Å/s   1.4 Å/s   2.0 Å/s   2.0 Å/s   Figure 5.  The efficiency of a-Si/a-SiGe tandem cells as a function of exposure time to simulated sunlight for cells with a-Si i-layers deposited at 1.4 and 2.0 Å  /s. Time (hours)    E   f   f   i  c   i  e  n  c  y   (   %   ) A11.4.5
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks