Wired: More efficient solar cells

In spite of its intuitive appeal, harnessing the sun's energy using solar cells has proved to be one of the more costly ways of generating renewable energy.

Traditional photovoltaic cells are made from wafers of semiconducting material such as crystalline silicon. The wafers are specially treated to produce an electric field with one surface positively charged and the other negatively charged, forming a so-called p-n junction. The two sides of the wafer can then be wired together into a circuit. When photons hit the cell, electrons are knocked out of the atoms making up the silicon crystal, creating direct current in the circuit.

Different semiconductors require different amounts of energy to knock out an electron, a property known as the band-gap. This means that a given wafer only generates electricity from the portion of sunlight above a certain frequency. The rest is wasted as heat.

Solar cells are an expensive way of generating energy for two reasons. Growing silicon crystals to make the wafers is costly and the efficiency of solar cells is relatively low (they only convert up to about 25% of solar energy into electricity). Efforts to develop more cost-effective solar cells have generally focused either on alternative materials (which tend to be cheaper to produce but less efficient) or improving the efficiency of crystalline cells (typically by stacking wafers made from semiconductors with different band gaps).

However, Harry Atwater, Michael Kelzenberg and their colleagues at Caltech have made progress towards an alternative solution which uses less silicon without sacrificing efficiency. They created silicon nanowires, each containing a p-n junction, measuring 30 – 100 microns in length and just 1 micron in diameter (about 1/100th the width of a human hair). By embedding an array of these wires in polymer, Atwater's team created a solar cell that can absorb up to 85% of sunlight. Although slightly lower than the 87% achieved by a crystalline cell, the wire array uses just 1% of the amount of silicon. In the cells tested, as little as 2% of the array's volume is taken up by silicon.

This came as a surprise to Kelzenberg, who had initially tried growing the wires close together expecting light falling on the spaces between them to be wasted. However, the team managed to achieve high absorption rates with sparsely spaced wires by trapping light in the array. In a recent issue of Nature Materials, they describe three ways of improving absorption by trapping light: using a back-reflector on the base of the array, coating anti-reflective material on the wires and embedding light-scattering particles in the polymer.

Another advantage of the wire arrays is that they can be embedded in thin, flexible sheets of polymer, which could make them cheaper to install and more versatile than traditional crystalline cells. Although the technique has the potential to produce cheaper solar cells by reducing the amount of silicon needed, it has so far been tested only in tiny cells of a few square centimetres. Atwater and his team are now scaling the arrays up to match the size of traditional solar cells. If they can do so successfully, more affordable solar power could be just on the horizon.

Kelzenberg et al. (2010). Enhanced absorption and carrier collection in Si wire arrays for photovoltaic applications. Nature Materials, 9, 239-244.