R&D Pushes c-Si Solar Cells Through Hurdles
Although crystalline silicon solar cells continue to dominate the photovoltaics market, they face any number of challenges to sustained growth, stemming from increasing pricing pressures. Researchers are working on several novel concepts to meet the challenges.
Els Parton, Philip Pieters and Jef Poortmans, IMEC, Leuven, Belgium -- PV Society, 9/1/2009
Over the past five years, production capacity has been building extremely quickly in the photovoltaics industry, and it has been largely based on turnkey lines. To be ready for the next wave of investments, it is crucial to build up and aggregate new IP, which could be facilitated by partnerships among PV manufacturers, equipment suppliers and research institutes.
 |
|
IMEC has developed i-PERC solar cells that improve rear surface passivation to allow thinner cells with higher efficiency. (Source: IMEC) |
The current PV market is dominated by crystalline silicon (c-Si) solar cells, which hold >90% market share. This dominance is expected to continue for at least the next decade, but there are several economic challenges that must be met to fuel sustained growth. Crystalline silicon will face increased price competition from new thin-film technologies. Although these thin-film technologies are all facing sustainability issues for the long term, the appearance of these technologies will first lead to downward pressure on prices for c-Si cells. Some players are predicting price erosion of 30-40% in the coming three to five years.
In the past, introducing process changes into existing solar cell lines was a relatively slow process, certainly compared with the rapid technology evolution within the microelectronics sector. However, the introduction of new technologies is likely to accelerate for key reasons: the increasing size of existing industrial PV players; and the entrance of new players from the microelectronics sector, with their strong background in semiconductor processing and the rapid implementation of new processes.
The development of PV-dedicated equipment is also becoming more economically viable, a trend that has been reinforced by the entrance in the industry of equipment manufacturers that have traditionally been devoted to the microelectronics sector. These equipment manufacturers are driven by consolidation within the microelectronics sector, which is only leaving room for a few large players, along with the room for expansion within the PV sector.
The strong growth of the past decade and the expected growth for the next decade have been enabled by specific government measures, especially in Japan and Germany. In particular, the creation of the feed-in tariff system in Germany has given a tremendous boost to the sector. However, in the coming years, pressure will likely increase from public authorities to accelerate the reduction of feed-in tariffs from their original levels. As a result, PV companies will be obliged to reduce their target costs faster than originally anticipated.
c-Si cost reductions
There are various strategies that the PV industry can follow to reduce the cost of c-Si solar cells. To offset what was a temporary lack of polysilicon feedstock, and to reduce cost in comparison to thin-film approaches, the c-Si industry has been aggressively pursuing a reduction in grams of pure silicon per watt (Si/Wp). This requires a significant change in c-Si technology in order to use thinner wafers without sacrificing solar cell efficiency. In general, the objective is to reduce Si/Wp by a factor of 2 while increasing efficiency by ~25% (from 16% to >20% for industrial c-Si solar cells).
Other cost-saving changes are also foreseen in metallization, where it is crucial to replace aluminum pastes or silver contacts with lower-cost options. Replacement of the silver contacts is particularly important because silver will face sustainability issues when the PV market reaches a size on the order of several 10 GWp/year.
All of this must go hand in hand with a reduction of solar manufacturing costs. One way to achieve this is to scale wafers up from 156 to 210 mm2, even as substrate thickness is reduced. Equipment scaling is an option, as well as an increase of the areal throughput. Fabs will likely scale by a factor of 10-20 from 50 MW to 1 GW/year. Vertical integration upstream (polysilicon feedstock production, crystallization and wafering) as well as downstream (modules, integration of the inverter in the module) is required in the long run to drive down costs to the desired levels of grid parity and below (grid parity will not be sufficient for the long term, and PV will have to become competitive with electricity costs at the level of the power plant).
Thinning the wafers
The mainstream approach today is to process solar cells in bulk silicon wafers of ~180 μm thick. Most PV fabs are not equipped to handle much thinner cells, but research labs have demonstrated that cells as thin as 50 μm with efficiencies >20% can be processed. By using advanced wafer handling (e.g. temporary carrier wafers), wafer-based processing may be extended down to 40 μm cells. It is also likely that cell processing steps will gradually be integrated into module assembly lines.
To meet efficiency targets in such thin wafers, PERL-style passivation (passivated emitter, rear locally diffused) and inter-digitated back-side contacts (i-BC) will need to be introduced in industrial process flows (i-PERL, i2-BC). It is expected that PERL concepts will enter the market first since these are a logical next step for today's front-side contacted cell manufacturing lines. In the longer term, BC cells will gain an important market share and might overtake front-side cells.
Epitaxial solar cells are a new approach in which a thin (
For mainstream fabs, this approach constitutes the bridge between bulk substrate silicon solar cells and thin-film solar cells, since the epi process can be implemented with limited equipment investment. In the lab, efficiencies of 16% (small area) have already been achieved. Epitaxial-based cells are likely to move into production earlier than silicon thin-film cells in which polycrystalline silicon layers are deposited on large-area non-silicon substrates (such as glass). These crystalline thin-film silicon cells have the ultimate cost potential, but their market introduction still requires major efficiency breakthroughs.
Belgian research institute IMEC has been studying new solar cell technologies and concepts for 25 years. Together with PV partners worldwide, the research institute aims to tackle a variety of limitations to current technologies.
Advanced surface and contact passivation
One such limitation is with passivated emitter and rear cells (PERC), which have a high open-circuit voltage and short-circuit current density related to the very low recombination in the bulk, and at the front and rear surface. The rear ohmic contact areas formed by the intimate aluminum-silicon contact are still areas left relatively poorly passivated. The effect of recombination can be minimized by separating contact areas by a distance larger than the substrate thickness, but this large separation causes relatively low fill factors because of the substrate lateral resistance. Furthermore, the rear aluminum-silicon contact has a relatively high contact resistance, which further constrains cell design.
Although the passivation techniques applied on the highly doped regions in conventional front and rear contacted cells suffice, the demands on passivating the critical regions in cells where the junction is moved to the rear surface â such as the front surface and the extensive surface extension of the space charge regions â is much more critical. The widely used SiN passivation is not expected to fulfill this requirement related to the charges being present in the layer.
New technologies are being researched to create locally diffused regions acting as local front or back-surface fields. This will reduce the effective recombination velocity at base contact regions over what is typically achieved by back-surface fields (BSFs), realized by aluminum screen-printed contacts. Boron BSFs are the proposed route to reach higher cell efficiencies. To date, boron diffusion from gaseous sources such as BBr3 has been the baseline process used.
However, although the process has been studied for many years, it is not widely used in production because of several drawbacks. It requires process temperatures >1000°C (to diffuse the boron into the silicon), which is not preferred with regard to bulk lifetime. The so-called boron-oxide skin formed is too boron-rich to be etched in HF (so needs to be etched and re-oxidized in iterative steps). For this reason, IMEC is studying plasma-implanted BSF, boron-diffused BSF from liquid sources, and epitaxial BSF by selective epitaxy (Fig. 1).
|
|
1. Novel approaches for boron doping include plasma-implanted back-surface field (BSF, left) and epi BSF (right). |
Researchers at IMEC are also studying novel
passivation options. Passivation schemes based on standard dielectrics such as
SiN do not have potential for cells >20%. In addition, longer-term
module-level processing of BC cells requires lower process temperatures.
Therefore, research includes reduced effective recombination velocity in the
inter-contact regions by negative charge dielectrics (e.g. Al2O3
deposited by ALD), PECVD a-Si:H passivation, and alternative dielectrics (e.g.
low-k dielectrics, porous dielectric layer, sulfur passivation, etc.).
Advanced and novel emitters
Most state-of-the-art industrial silicon solar cell processes make use of POCl3 to diffuse phosphorous into the substrate and create the emitter. The exact profile of the emitter has a large impact on the final solar cell performance, and requires a compromise between emitter doping and contact formation. For example, when screen-printed contacts are used, the surface area of the emitter has to be relatively highly doped (>1020 cm-3) to allow for a good ohmic contact. Ideally, the surface concentration should be lower to improve surface passivation, causing a need for more elaborated emitter profiles.
The main objective of the research in this area is to increase the blue response of front-side contacted cells, reduce the emitter saturation current density and increase the open-circuit voltage of front and back contacted cells (Fig. 2).
|
|
2. Epitaxial emitters show excellent UV response. |
Possible routes to achieve this include shallow homo-junction emitters
by POCl3 and/or implantation contacted with optimized screen printed
paste; "selective" front side emitters (shallow n+ emitter
combined with n++ emitter under the contact areas); epitaxial
emitters; hetero-junction emitters a-S:H/c-Si; and MIS-emitters
(metal-insulator-semiconductor inversion layer).
Contacting
The standard industrial process for contacting solar cells has several drawbacks. First, the screen-printed front side metallization with silver paste leads to contacts with a poor aspect ratio at a relatively high width and a low conduction quality because of the composition of a printable paste. Also, the poor contact resistance requires high surface doping concentrations of the emitter.
Therefore, new metallization techniques must be developed with high-throughput potential, guaranteed sustainability (e.g. phasing out silver for front-side metallization), and potential for improvement of low metal coverage, low finger and contact resistivity, and/or reduced material cost. These metallization techniques must be matched with cost-effective via fabrication techniques, especially for back contact cells where pitch requirements will probably be very demanding.
For metallization, the process technologies that show potential benefits are nano-inks, metal aerosol jetting, laser sintering of metal powder, contact to hetero-junctions, plating, and related barrier layers for copper â namely TiN, TiW or TaN by PVD. For via fabrication, (UV) laser ablation (for pattern definition) is being studied as well as inkjet printing of masks for etching dielectric, or inkjet printing for developing blanket resist layers.
Optical enhancement
The optical performance of traditional solar cells is limited. The reduction of the first reflectance and oblique light coupling properties of traditional front structures ranges from moderate to good, but the rear structure shows poor internal reflection and high parasitic absorption. As a result, the path length enhancement factor is ~7 in the best case. This is still considered acceptable if the wafer is thick, but is insufficient for high-efficiency cells on very thin substrates. Also, these techniques consume a substantial amount of silicon, which should be avoided when the wafers are very thin to start with.
Another drawback of traditional optical enhancement methods is that they are essentially two-sided processes, achieved by immersing wafers in a chemical solution. This is not compatible with the high-quality surface passivation required at the rear, which calls for a flat rear surface. Therefore, practical and efficient processes for one-sided texturing need to be developed. The reflector at the rear needs to have a high reflectance (well above 90%), but needs to be cost-effective and fit well with the surface passivation scheme.
Finally, advanced antireflective coatings (ARCs) need to be investigated. SiN does a good job, but to reach very high efficiencies, it is important to squeeze out the maximum of the potential current, which probably implies the use of a multilayer ARC (double, multiple or graded). Better transmission in the UV may be brought about by the introduction of new encapsulants.
The objective of the research community is to enhance the optical path length without sacrificing surface properties. This compensates for the lower absorption in thinner wafers, which is essential for increasing efficiencies while reducing wafer thickness. The techniques that need to be pursued for this purpose are plasma texturing for single-sided texturing (Fig. 3); random pyramid process by wet etching; ARCs with lower reflectance than standard SiN ARCs; and dielectric or pigmented reflectors at the back side of the cell.
Conclusion
The PV sector, with revenues of $15B-$20B today, will become a $100B+ sector by 2020. To achieve and maintain such an enormous growth, the PV industry must invest in R&D to develop cheaper and more efficient solar cells. Although epitaxial and thin-film solar cells are innovative concepts, first innovations in process technology for bulk silicon solar cells will need to guarantee the success of the industry. Joint R&D has proven to be a successful strategy in the IC industry, and will surely also be for the PV industry.
Si and Ge Solar Cell Efficiency Records Set
10/11/2007Schott Solar Joins IMEC Silicon PV Program
06/03/2009
