While delving into the intricacies of how photovoltaic (PV) cells are produced and the vital role that silicon (Si) plays in the technology, it also becomes clear how the introduction of minute quantities of impurities like phosphorous and boron transform this electrically neutral, tetravalent metalloid, into versatile semiconductor materials that form the basis of most PV cells developed to date.
First noted by a French Physicist, Edmund Bequerel in 1839, and explained by Albert Einstein in 1905, the photoelectric effect is the phenomenon upon which PV technology is based.
Electromagnetic radiation transmits energy through space; the spectrum of visible light occurs in the wavelength range from 400-700 nm while the shorter wavelength (higher frequency) range is identified as ultraviolet and longer wavelength (lower frequency) characterises the infrared range.
PV cells are made from semiconductor materials and designed to generate an output voltage when exposed to electromagnetic radiation in the ultra violet and visible light spectrum. When connected together in sufficient numbers to form PV panels, PV modules and PV arrays, they are capable of providing electric power either for driving off-grid equipment or to supplement grid power distribution systems.
Bell Laboratories built the first PV module in 1954 and during the 1960s space-race these novel devices were chosen to provide electric power onboard space vehicles. This application drove further development until reliability was established and production costs began to fall.
Useful properties of Si
The particular crystal structure of pure Si is determined by its molecular structure having 14 electrons, four of which exactly half-fill the outer shell.
The controlled addition of carefully selected impurities to pure Si is commonly referred to as doping and was perfected in the development of solid-state electronic devices such as diodes, transistors and integrated circuits.
Doping with phosphorous that has five outer shell electrons creates a prevalence of free electrons and material takes on a negative electrostatic charge. This is known as N-type Si and contains a phosphorous atom for every million or so Si atoms. In contrast, doping with boron that has three outer-shell electrons creates a prevalence of free openings where electrons are ‘missing’ from the crystal structure of pure Si and this material is known as P-type Si.
These modified forms of Si are referred to as semiconductors because they are no longer electrically neutral and are far more conductive than the pure Si prior to doping. Historically, the solar industry has relied on supplies of off-specification Si material rejected by the semiconductor industry.
Recently the availability of scrap has been inadequate to meet the demand of the rapidly growing PV industry and more expensive prime-grade Si is now also being used.
Types and sources of Si
Mono-crystalline silicon (c-Si): This material is made as a single large crystal in which the lattice structure is homogeneous, free of grain boundaries and recognisable by an even external colouring.
Using the Czochralski process, these are pulled from molten Si held in a large quartz crucible heated to about 1600°C. Known as ‘solar grade silicon’; c-Si was used in developing the first solar cells and remains the prevalent raw material in PV manufacturing.
Most c-Si PV panels have uncovered gaps at the four corners of the cells because they are cut from cylindrical ingots resulting in a substantial waste of refined silicon.
Poly or multi-crystalline silicon (poly-Si or mc-Si): This material contains multiple small silicon crystals and is made by casting pure Si into cubic ingots that are carefully cooled and solidified. Poly-Si cells can be recognised by a visible grain or a ‘metal flake effect’. They are less expensive to produce than c-Si cells, but are less efficient.
Ribbon silicon is a type of multi-crystalline silicon: This material is formed by drawing flat thin films from molten silicon and results in a multi-crystalline structure. These cells have lower efficiencies than poly-Si, but save on production costs due to a great reduction in silicon waste, as this approach does not require sawing from ingots.
Amorphous silicon (a-Si or α-Si) is a non-crystalline allotropic form of silicon: The atoms within a-Si form a continuous random network. This disordered nature of the material results in some atoms having a dangling bonds and these are physical defects that may cause anomalous electrical behaviour.
PV cell manufacturing processes
Most PV solar cells are made from bulk Si materials cut into wafers of between 180 to 240 micro-metres thick that form the substrate material of the cell.
When it comes to crystalline silicon solar cells, the Si wafer is usually lightly p-doped and conductive for positive charge carriers. On the front side, a thin heavily n-doped layer is formed which is conductive for negative charge carriers or electrons. This forms a photo-active p/n-junction which separates the charge carrier pairs generated by the absorption of sunlight.
Next, metallic contacts are formed to drain the solar-induced electric current. A conductive aluminium layer is deposited on the underside, while silver contact fingers are generated on the top side which allow most of the sunlight to pass into the cell. Finally an anti-reflective silicon nitride coating is deposited to minimise reflective loss of incident light.
Let’s move from these type of cells, to thin-film solar cells. The electric carriers generated in a solar cell by light radiation have a limited lifespan and diffusion length in the semiconductor material. Reducing the thickness of deposited layers increases the potential efficiency of photovoltaic cells by shortening the path for charge carriers to reach the electrodes. Hence, the rise of thin-film solar cells.
But, despite claims that thin-film PV cells are less expensive to produce and there are continuing advances in their conversion efficiency, the market-share of thin film modules in 2010 actually declined by 30%.
There are three main types of thin-film solar cells that are produced:
Thin-film solar cells based on wafers of hydrogenated amorphous silicon (a-Si:H).Because hydrogen bonds to the dangling bonds, it is sufficiently defect-free to be used within PV devices.
Unfortunately hydrogen does cause light induced degradation of the material, termed the Staebler-Wronski Effect. The solar cell consists of a transparent, conductive front layer, followed by a p-doped a-SiC:H-layer, the absorber layer of intrinsic a-Si:H or a-SiGe:H and an n-doped a-Si:H-layer, forming a p-i-n-structure.
For the intrinsic absorber layer, a compromise has to be found between light absorption and electrode density. One solution is the preparation of stacked multi-junction solar cells, so-called tandem, or dual, or triple junction cells.
Thin-film solar cells made up of thin-film layers applied by Plasma Enhanced Chemical Vapour Deposition onto supporting substrates such as glass or metal. The materials used include cadmium telluride (CdTe), cadmium selenide (CdSe), cadmium sulphide (CdS), copper indium selenide (CIS), copper indium gallium selenide or sulphide (CIGS) which are better light absorbers than crystalline silicon allowing reduced material consumption and lower production costs.
Dye-sensitised solar cells (DSSC’s) are made of low-cost materials and do not need elaborate equipment to manufacture, so they can be made by screen printing and/or the use of Ultrasonic Nozzles. DSSC’s can be engineered into flexible sheets, and although conversion efficiency is less than the best thin-film cells, they should be significantly less expensive in bulk. The potentially low price/performance ratio makes this is a popular emerging technology.
Demand for electronic specialty gases
Dave Tavianini, Photovoltaics Business Development Manager for US-based industrial gas supplier Air Products, claims that, “Solar capacity is growing at more than 30% p.a. and with photovoltaics using many of the same raw materials as do semiconductor manufacturers, so we would expect to see strong growth.”
The three major gases used in semiconductors, liquid crystal displays (LCD’s) and photovoltaics are nitrogen trifluoride (NF3), silane (SiH4) and ammonia (NH3).
“We continue to supply our customers as effectively as we have in the past, but supply remains tight given the demand for polysilicon in semiconductors and photovoltaics,” adds Mike Hilton, Air Products’ Senior Vice-President and General Manager for Electronics and Performance Materials.
Air Products has been routinely expanding its NF3 capacity since the late 1990s to meet demand, sometimes by as much as 50% at a time.
Cell manufacture also uses silane (SiH4) as a precursor with nitride and oxide before deposition. Above 420°C, silane decomposes into silicon and hydrogen and is used in the chemical vapour deposition of silicon, silicon oxide and silicon nitride on glass substrates. SiH4 is also a key precursor in the production of polysilicon for some polysilicon suppliers.
Mike Corbett, Managing Partner of Linx Consulting, based in Boston, Massachusetts, US warns, “There is the potential for an SiH4 shortage to hit the market eventually and this would ripple through photovoltaics and LCD production in a bad way.”