Manufacturing solar cells

In general, solar cells are made from thin flat layers of semiconductor that include a p-n junction across the whole area of the cell.

A photon hitting the structure creates an electron-hole carrier pair (an exciton) which is separated by the junction. This develops a potential difference between the front and the back of the cell which can produce a current externally.

There is a relationship between the energy in the photon and the bandgap of the structure which governs the success rate of photon to exciton conversion.
Essentially: a particular semiconductor will only efficiently convert one colour of light - or more accurately, light between two wavelengths with sensitivity peaking somewhere near the middle.

For example, 'single junction' silicon solar cells can only absorb the near-infra red part of the sun's spectrum and have a light to electricity conversion efficiency somewhere around 20 per cent.
GaAs and other compound semiconductors can be used to form junctions with other bandgaps, and these junctions can be stacked to harvest a broader range of wavelengths - or photon energies - depending on whether you are thinking classically or in quantum terms.

Providing they are stacked in the right order so the top layers pass unused light through to lower layers, efficiencies of around 40 per cent can be achieved. The current record is almost 50 per cent.

Double and triple junction cells very expensive, and are found on satellites and transcontinental solar racing cars.

Single crystal junctions achieve the efficiencies mentioned above, but are not the cheapest way to produce solar cells.

Amorphous silicon deposited on glass offers around six per cent efficiency at far lower cost than single crystal silicon, and is frequently seen on solar-powered calculators.
Polysilicon on glass is between amorphous and single crystal silicon in both cost and efficiency.

It is widely believed that, depending on latitude, a minimum efficiency of 10 per cent is required to make cost-effective large-scale solar power installations - and the raw materials will have to be cheap.

Researchers are looking at alternative materials to achieve these aims.

For example: single crystal solar cells are made from IC-grade silicon wafers, whereas less pure silicon could be used with negligible loss in efficiency.

The question of whether an industry will form to produce these less pure wafers remains to be answered.

Organic semiconductors can be used to make solar cells, dopes with materials including carbon nanotubes.

Currently efficiency is a few per cent at most and the cells degrade rapidly in sunlight, but some predict organic solar technology will be the one to take off. In the mean time, these cells are likely to develop enough to be used in solar-powered portable electronics.

Out in the far field are solar cells based on structures that mimic photosynthesis, and various liquid and solid bulk technologies, such as the dye-sensitised solar cells in which the excitons form in dye; titanium dioxide pulls away the electrons; and an electrolyte takes away the holes.

With some forms of organic solar cell, as well as solid dye-sensitised cells, it may be possible to make large areas very cheaply on printing presses.

As the manufacture of nano-scale material powders becomes possible, researchers are not only reviewing existing solar cell types, but looking at schemes in which the light is absorbed by particles of similar size to its wavelength.