III-V thin films for PV applications

The ability of the III-V materials system to maintain high efficiencies at high concentration levels using very thin films is particularly attractive …

SIMON RUSHWORTH, SAFC Hitech, Bromborough, Wirral, UK

The ability of the III-V materials system to maintain high efficiencies at high concentration levels using very thin films is particularly attractive in relation to reducing the volumes needed for the conversion process.

The rise in global interest in alternative and renewable energy sources over the past couple of years has seen the PV market expand rapidly, with bulk silicon solar cells dominating. But emerging technologies based on thin film devices are also making rapid progress, to the extent that this sector is now one of the largest growth areas being developed to help meet anticipated future PV needs. It is expected that by 2012, the rising production capacity of thin film technologies will account for more than 30% of total installed PV manufacturing (Fig. 1).

Techniques such as chemical bath deposition and closed space sublimation are leading the way in the fabrication of thin film solar cells. But there is also significant interest from the market in CVD (chemical vapor deposition) as a large area process on glass and other substrates. Competing material systems for the active thin films in PV manufacturing are cadmium telluride (CdTe) based, copper indium gallium (di)selenide (CIGS) based, and III-V materials–along with, of course, silicon.

III-V production considerations

For all solar cells, the efficiency of the active region at absorbing sunlight varies. In the III-V area, the conversion of light to electricity is most efficient at very high illumination levels; therefore, concentrator technologies are needed to generate around 500 suns, which is focused on the active area to achieve the best results. Under the highly demanding conditions applied to the active components, the structure of the complete cell must be perfect, and of very high purity with respect to a number of key contaminants to avoid internal losses due to non-radiative centers and the generation of excessive heat.

Significant experience in the area of III-V semiconductor deposition can be accessed via high brightness LED production technologies, since similar requirements for purity and control are needed to obtain the highest device performance. These approaches have led to the Fraunhofer Institute for Solar Energy recently achieving record efficiency levels using arsenide and phosphide materials. In particular, it has been observed that oxygen (O) is an unwanted non-radiative center that detracts from cell operating efficiencies and lifetimes. To minimize oxygen levels in the deposited films, the precursors used must be of the highest quality and, in particular, the organoaluminum source must have contaminant levels less than 1ppm.

In general, strict quality control of proprietary production and purification methodologies must be applied to ensure that the lowest contamination levels are available for PV applications, allowing generation of reliable, high-performance devices.

To achieve the targeted specifications, state of the art analytical capabilities must be employed to characterize the starting chemicals. For accurate oxygen contamination determination, significant problems exist due to the pyrophoric nature of many of the precursors used to deposit the optimum layers. In particular, group III metal alkyls, such as trimethylaluminum (TMA) and trimethylindium (TMI), require careful handling to avoid atmosphere contact due to high reactivity.

Figure 1. Forecasted PV manufacturing capacity for crystalline silicon vs. thin film technologies. (Source: Joint Research Centre Renewable Energy Unit–PV Status Report 2008; www.jrc.ec.europa.eu)

For example, any metal pipework or vessels contacted by the liquid or solid product must be totally O2/H2O-free. Typically, vacuum heat treatment for extended periods is necessary to desorb these species from surfaces and eliminate potential contamination. Figure 2  shows the impact on removal efficiency of residual gases with extended treatment protocols and the significant reduction of impurities observed. As shown in Fig. 3, 48 hour versus 24-hour vessel processing prior to product fill has been demonstrated to give improved performance in the most critical structures.

Figure 2. Residual gas analysis of vessels treated using different process parameters. (Source: Qinetiq–contracted trial with T. Martin group at Malvern, UK [not published])

Assessing the quality of the precursors themselves without actual growth testing requires careful correlation of physical impurity levels against reference values. For oxygen species, it has been found that the alkoxide peak present in the product proton nuclear magnetic resonance (NMR) can be directly used to indicate performance in growth. Figure 4 shows the detected O species levels in a series of trimethylaluminium (Me3Al) samples, and Table 1 highlights the properties of aluminum gallium arsenide (AlGaAs) deposited using these samples. A direct correlation between the two sets of data is clearly seen, with contamination levels of less than 1ppm in the source demonstrated as key to achieving the highest final film quality.

Delivery system developments

The requirement for larger volumes of precursor to allow large area deposition at high throughput and low cost has led to a number of innovative equipment developments targeting bulk delivery of chemicals to the growth tools. Safer handling of increased lot sizes helps to reduce the overall cost of ownership for the deposition processes by minimizing tool downtime and qualification run requirements. The use of larger batches and the above economies of scale without compromising quality are key elements in enabling customers to achieve economically viable processes and meet increasing consumer demands in a robust and sustainable manner.

For liquid precursors, the preferred method of bulk delivery involves pumping product to the tool for vaporization, as is normally employed for smaller batch sizes. The advantage of this approach is that it offers easy retrofitting to existing equipment without the need to change process parameters. In addition, the benefits in cost of ownership are significant for production tools running the same process continuously.

To optimize output, the loading on all deposition tools should be as close to 100% as possible, since downtime replacing or changing out bubblers when a precursor lot has been depleted and requires refilling can be detrimental to the economics of the process. By adding the ability to refill the tool container and immediately resume production on the fly at the touch of a button, time and money can be saved. Obviously, the safe movement of highly hazardous chemicals between two points requires specifically engineered equipment, and precursor suppliers have applied their expertise to the design of such systems as SAFC Hitech’s EpiFill.

Figure 3. Oxygen level in AlGaAs-based structures grown under a variety of conditions, with Me3Al filled into containers treated for 24 hours (orange trace) and 48 hours (blue trace). The lower the O level, the better; hence, 48 hours is better than 24-hour treatment of the bubbler. (Source: Fraunhofer Institute for Solar Energie–contracted trial with F. Dimroth group at Freiburg, Germany [not published])

An alternative approach for solid precursor delivery, or where simplified deposition tools are to be installed, is a bulk vaporization unit; it can supply precursor vapors directly to the deposition chamber via a mixing manifold only, thus removing the necessity for multiple, individually-controlled temperature environments on each system. As for the liquid case, the cost advantages of reduced downtime are attractive and, by moving the evaporation stage away from the growth kit, the control of gases entering the deposition chamber can be addressed without resorting to complex systems, making process control more reliable. Again, chemical suppliers have employed their expertise to develop robust equipment to deliver the desired volumes of precursor vapors in a safe, reliable fashion.

Figure 4. Alkoxide contamination quantification in Me3Al samples using proton NMR. (Source: SAFC Hitech in-house analysis)

New levels of solar cell efficiency

The impact of controlled precursor provision as an enabling technology is illustrated by the advances made in the field of III-V thin film solar cells where, as noted earlier, record efficiencies are being achieved using high purity products. The reliable supply of such chemicals has enabled process and device optimization to be performed with the confidence that the composition of thin film structures can be controlled precisely and reproducibly across the range of test parameters employed to identify device improvements.

Table 1. Correlation of O levels in Me3Al precursor and AlGaAs films. (Source: Fraunhofer Institute for Solar Energie, contracted trial with F. Dimroth group at Freiburg, Germany [1])

Monolithic multi-junction solar cells based on epitaxial III-V semiconductors have evolved over many years to the point at which purity and crystallinity levels are now extremely high. This is reflected in the rise in efficiencies over this period for the most successful designs. The latest three PN-junction combinations with gallium indium phosphide (Ga0.35In0.65P), gallium indium arsenic (Ga0.83In0.17As) and germanium (Ge) absorb sunlight across the ranges 300-780nm, up to 1020nm, and up to 1880nm, respectively, which has been predicted as particularly advantageous for the terrestrial solar spectrum conversion to electricity.

A new record-breaking solar cell from the Fraunhofer ISE has a cell area of 5.09mm2 and an overall efficiency when operated at 454 suns of 41.1%. The ability to operate at even higher concentrations while keeping high efficiency (37.6% @ 1700 suns) is a key advantage of this cell design, but this function is highly dependent on perfect construction of all the individual layers and interfaces to avoid charge trapping and more problematic defect propagation. The degradation of quality by such methods leads to reduced lifetimes, which are not acceptable in a commercial device; hence, the focus on deposition technologies to achieve high-quality epitaxy throughout the multilayer structure.


To achieve record-breaking values for solar cell efficiencies and to continue to improve cell performance in the future, a complex structure with up to 40 individual layers must be prepared perfectly. Therefore, robust growth processes and precursor delivery capabilities that maintain high purity and offer high controllability at high volumes are of critical importance.

Performance maintenance through the scale-up of these processes will lead to production cost reductions. Furthermore, the ability of the III-V materials system to maintain high efficiencies at high concentration levels using very thin films is particularly attractive in relation to reducing the volumes needed for the conversion process. This, in turn, will further reduce electricity generation costs in the future, helping to drive this technology to the forefront of PV solutions for alternative energy supplies on a global scale.


EpiFill and EpiPure are trademarks of SAFC Hitech.


1. S.A. Rushworth, et al., “Correlation of Reduced Oxygen Content in Precursors with Improved MOVPE Layer Quality,” Jour. Of Crystal Growth, 248, pp. 86-90, 2003.

Simon Rushworth received his chemistry degree from Nottingham U., and is technical information manager at SAFC Hitech, Power Road, Bromborough, Wirral, Merseyside, UK, CH62 3QF; +441513342774; Simon.Rushworth@sial.com


More Photovoltaics World Issue Articles

No posts to display