Requirements for solar cell materials Monocrystalline silicon Multicrystalline silicon Crystalline thin-film silicon
Amorphous silicon a-Si/c-Si heterostructures Copper indium diselenide Cadmium telluride The silicon supply problem
Inorganic solar cells
One important way to convert solar radiation into electricity occurs by the photovoltaic effect which was first observed by Becquerel. It is quite generally defined as the emergence of an electric voltage between two electrodes attached to a solid or liquid system upon shining light onto this system. Practically all photovoltaic devices incorporate a pn-junction in a semiconductor across which the photovoltage is developed. The semiconductor material has to be able to absorb a large part of the solar spectrum. Dependent on the absorption properties of the material the light is absorbed in a region more or less close to the surface. When light quanta are absorbed, electron hole pairs are generated and if their recombination is prevented they can reach the junction where they are separated by an electric field. Even for weakly absorbing semiconductors like silicon most carriers are generated near the surface .
Requirements for solar cell materials
A large part of solar cells are based on silicon although from solid state physics we know that silicon is not the ideal material for photovoltaic conversion. The solar spectrum can be approximated by a black body of 5900 K which results in a very broad spectrum ranging from the ultraviolet to the near infrared. A semiconductor, on the other hand, can only convert photons with the energy of the bandgap with good efficiency. Photons with lower energy are not absorbed and those with higher energy are reduced to gap energy by thermalization of the photogenerated carriers. Therefore, the curve of efficiency versus bandgap goes through a maximum (Fig. 1) .
Fig. 1. Dependency of the conversion efficiency on the semiconductor bandgap.
It can be seen that silicon is not at the maximum but relatively close to it. A much more serious point is that silicon is an indirect semiconductor, meaning that valence band maximum and conduction band minimum are not opposite to each other in k-space. Light absorption is much weaker in an indirect semiconductor than in a direct semiconductor. This has serious consequences from a materials point of view: for a 90% light absorption it takes only 1 μm of GaAs (a direct semiconductor) versus 100 μm of Si. The photogenerated carriers have to reach the pn-junction which is near the front surface. The diffusion length of minority carriers has to be 200 μm or at least twice the silicon thickness. Thus, the material has to be of very high purity and of high crystalline perfection. In view of these physical limitations it is quite surprising that silicon has played such a dominant role in the market. The main reason is that silicon technology has already been highly developed before the advent of photovoltaics and high quality material is being produced in large quantities for the microelectronics market. It is not surprising that a lot of effort has been going and is still going into the search for new materials .
Requirements for the ideal solar cell material are:
- bandgap between 1.1 and 1.7 eV;
- direct band structure;
- consisting of readily available, non-toxic materials;
- easy, reproducible deposition technique, suitable for large area production;
- good photovoltaic conversion efficiency;
- long-term stability.
A material fulfilling all these requirements has not yet been found. Since the most important requirement is a high light absorption coefficient the materials defined above are ‘‘thin-film materials’’ in the sense that only about 1 μm of active material is required. Thus, the amount of material needed is drastically reduced compared to crystalline silicon. An additional advantage of thin-film materials is that they can easily be connected in series in an integral manner on a large area substrate. For the future of solar energy materials three scenarios can be envisioned:
- continued dominance of the present single crystal or cast polycrystal technology;
- new crystalline film Si materials of medium thickness either as ribbons or on foreign substrates;
- breakthrough to mass production of true thin-film materials like a-Si or CIS or CdTe.
In the long-term new concepts or new classes of materials like organic solar or III/V-tandem cells also are a possibility. At this point those scenarios have about equal probability. Even more likely is that two or three of them will coexist for a considerable period and that each technology will find its own market. From an overall point of view it can be considered an advantage that so many avenues exist that potentially lead to a low cost solar cell. In this way the likelihood of achieving this goal is greatly increased .
The first silicon solar cell was developed in 1954 and it had an efficiency of 6% which was rapidly increased to 10%. The main application for many years was in space vehicle power supplies .
Polycrystalline material in the form of fragments obtained from highly purified silicon is placed in a quartz crucible and melted under inert gases by induction heating. A seed crystal is immersed and slowly withdrawn under rotation. The silicon melt reacts with every material to a large extent. Only silica can be used as a crucible mterial, because its product of reaction, silicon monoxide, evaporates easily from the melt. Neverthless grown crystals contain 1017-1018 atoms/cm3 of mainly interstitial oxygen. This lead to degradation effects. Recently, interesting results have been obtained with an advanced technique. A magnetic field interacts with the free electrons of the silicon and retards convective melt flows. The transport of oxygen from the crucible walls is minimized .
For solar cells the crystal rods are separated into wafers of 0.2-0.5 mm thickness by sawing. This is a costly process because silicon is a very hard material which can only be cut with diamond and up to 50% of the silicon is lost in this process .
It has been known that high efficiency solar cells made from monocrystalline silicon undergo a moderate degradation of efficiency when exposed to light. This effect becomes only now important as solar cells in production reach efficiency above 15%. This degradation can be completely reversed by an anneal step of around 200°C in room ambient. The simultaneous presence of boron and oxygen is responsible for the degradation of efficiency. A strategy for the elimination of degradation is the substitution of boron by gallium in p-type Si. Since for all investigated material types no degradation was observed, this approach seems to be very promising. The only disadvantage is a higher resistivity compared to boron doping .
The best laboratory efficiency for single crystal silicon is today 24.5%. This efficiency can only be realized with very elaborate technology. Experience has shown that progress in laboratory efficiency leads to improvement in production with a certain time delay. The best production cells now have an efficiency of 15-16% .
As the cost of silicon is a significant proportion of the cost of a solar cell, great efforts have been made to reduce these costs. One technology which dates back to the seventies is block casting which avoids the costly pulling process. Silicon is melted and poured into a square SiO/SiN-coated graphite crucible. Controlled cooling produces a polycrystalline silicon block with a large crystal grain structure. The grain size is some mm to cm and the silicon blocks are sawn into wafers by wire sawing as previously mentioned. Cast silicon also called polycrystal silicon is only used for solar cells and not for any other semiconductor devices. It is cheaper than single crystal material but yields solar cells with a somewhat lower efficiency. An advantage is that the blocks can be manufactured easily into square solar cells in contrast to pulled crystals which are round. It is much easier to assemble multicrystalline wafers into modules with nearly complete utilization of the module area. Thus, the lower efficiency of cast material tends to disappear at the module level. Because of the contact with the crucible polycrystal silicon has a higher impurity content and, thus, lower carrier lifetime and lower efficiency than monocrystalline silicon .
Crystalline thin-film silicon
Nowadays, the driving force for the development of crystalline silicon thin-film solar cells (c-SiTFC) is the inherent possibility for cost reduction, although this advantage has not yet been converted into commercial products .
A schematic representation of crystalline thin-film silicon solar cell is given in Fig 2.
Fig.2 The basic components of a crystalline thin-film silicon solar cell:
a- antireflection layer; b- silicon layer; c- back side reflector; d- substrate.
The linking feature of all c-SiTFC approaches is the underlying substrate needed as a mechanical support due to the reduced thickness of the active silicon layer of typically 5-50 μm. The substrate consists either of low quality silicon, or of foreign substrates such as glass, ceramics or graphite. The choice of the substrate material determines the maximum allowed temperature for solar cell processing .
As a general trend, the cell efficiency increases with the grain size. This is due to the fact that grain boundaries in their unpassivated state can be very effective recombination centers which reduce the diffusion length of the minority carriers drastically .
Due to the weak absorbance of crystalline Si, the light trapping is one of the crucial measures which have to be applied in order to realize a high degree of total internal reflection. This can be achieved by a back side reflector or by a textured front surface in combination with a reflecting back side .
The high expectancy in this material was curbed by the relatively low efficiency obtained so far and by the initial light induced degradation for this kind of solar cells. Today, a-Si has its fixed place in consumer applications, mainly for indoor use. After understanding and partly solving the problems of light induced degradation amorphous silicon begins to enter the power market. Stabilized cell efficiencies reach 13%. Module efficiencies are in the 6-8% range .
Fig. 3. Amorphous silicon Fig. 4. Schematic structure of an amorphous pin solar cell and corresponding band structure.
Amorphous silicon is an alloy of silicon with hydrogen. The distribution of bond length and bond angles disturbs the long range of the crystalline silicon lattice order and consequently changes the optical and electronic properties. The optical gap increases from 1.12 to about 1.7 eV .
The mobility of charge carriers in amorphous silicon is generally quite low so that collection of photogenerated carriers has to be supported by an internal electrical field. In order to create a high field the cells have to be, of the order of a few hundred nanometers. By applying proper light trapping schemes, cells can reach high efficiencies .
Another strategy to improve the stabilized efficiency of the devices is to use a stack of two cells or even three cells. Introducing cells with different bandgaps in the stack results in a ‘‘real’’ tandem cell which can make better use of the solar spectrum, and at the same time improve the efficiency of the devices. The bandgap can be increased by alloying with carbon .
A very interesting new development is the combination of crystalline and amorphous technologies in heterostructures. Absorption of sunlight still occurs mainly in a wafer of mono or polycrystalline silicon .
This configuration has the following advantages:
- potential for high efficiency;
- low processing temperatures; all processing steps occur below 200 °C;
- Reduction of energy pay back time;
- reduced cost of cell technology.
The best results with this approach were obtained by the Japanese company Sanyo. The latest achievement is a conversion efficiency of 20.7% .
Copper indium diselenide
A very challenging technology is based on the ternary compound semiconductors CuInSe2, CuGaSe2, CuInS2 and their multinary alloy Cu(In,Ga)(S,Se)2 (CIGS). Unlike the basic silicon solar cell, which can be modelled as a simple p-n junction, these cells are best described by a more complex heterojunction model. The best efficiency of a thin-film solar cell was 19.5%. Higher efficiencies (around 30%) can be obtained by using optics to concentrate the incident light. Further developments are directed towards reduction of film thickness and replacement of In and Ga as rare elements .
Thin-film solar cells based on CdTe are the cells with the longest tradition, but they are really not ‘‘of age’’. After a long steeplechase they arrived at cell efficiencies of 16% and large-area module efficiencies of over 10% .
CdTe is a nearly ideal material for thin-film photovoltaics because it combines several advantageous properties. Besides an optical bandgap close to the optimum for solar energy conversion it is very easy to handle in thin-film deposition processes. Therefore, many efforts were and still are directed towards the large scale fabrication of modules with capacities in the multimegawatt range. The thin-film module with the worlds highest power output has been produced based on CdTe.
A non-technical problem associated with CdTe is the acceptance in the market place because Cd and to a lesser extent Te are toxic materials although the compound is quite stable and harmless .
The silicon supply problem
A big question mark for the future is the source of highly purified silicon for solar cells. The 50% of the cost of a module is due to the cost of processed silicon wafers. The PV industry has in the past used reject material from the semiconductor industry that was available at low cost. This created a dependence that is only viable if both sectors grow at the same rate. An additional problem is that the semiconductor market is characterized by violent cycles of boom and depression superimposed on a relatively steep growth curve. In boom times, the materials supply becomes tight and prices increase .