"The design derives its inspiration from skyscrapers," Jud Ready, senior engineer at Georgia Tech Research Institute, US, told optics.org. "A conventional solar cell is like a flat pasture, while our design is more like downtown Manhattan, albeit on the nanoscale."

The Georgia Tech device is made up of 100 µm towers with a base area of 40x40 µm that are packed within 10 µm of each other. A special double-layer coating of cadmium sulphide and cadmium telluride acts as a p--n junction that converts incident light into electricity.

When light falls on the surface of one of the towers, a small portion is absorbed to produce a photoelectric current, while the rest is reflected away. However, the close packing of the towers ensures that a beam reflected from one tower is directed to another one close by, where the absorption/reflection process is repeated.

The CNT array therefore offers multiple absorption opportunities for the incoming light pulse. "In a sense the light is fed back to another tower in 3D array, and so we have called it the 3D photovoltaic cell," said Ready.

The photovoltaic array is made by patterning a thin layer of iron using a photolithographic process. Chemical vapor deposition (CVD) is then used to grow arrays of multi-walled carbon nanotubes on top of the iron patterns. This is followed by a molecular beam epitaxy process to coat the towers with cadmium sulphide and cadmium telluride, and with a protective outer coat of indium tin oxide.

The Georgia Tech team says that its design offers two key advantages over conventional solar cells. First, unlike most flat solar cells, the CNT photovoltaic array achieves high conversion efficiencies even when it is not placed perpendicular to the direction of sunlight. This avoids the need for bulky and expensive mechanical arrays to move the solar panel to face the sun.

Second, making more efficient use of the reflected light means that more photons are available for absorption, which in turn means that the p-n region can be made thinner. In contrast, reflection from the front surface of conventional solar cells reduces the number of photons available for photoabsorption, which dictates that the p-n region be made relatively thick. Space-based satellites, which have huge solar arrays, could therefore benefit from this new, lightweight design.

Early tests show the new design achieves an efficiency of 3.5% when the sun was perpendicular to the cell, and a maximum of 7% at a 45 deg incidence angle. "Looking forward, we see potential for this figure to go up to 40%, because our light-trapping structure allows us to reduce the thickness of the absorbing layers which in turn brings down recombination effects," said Ready.

The team is now studying the effects of dark current on the cells, while also investigating cheaper fabrication methods. "Due to its many interfaces, the current prototype also has a high series resistance, and we intend to study these effects in detail in the coming months," added Ready. "Our highest priority application is currently in aerospace and satellite systems and we would look into commercial power generation units at a later stage."

This work was reported in JOM, the journal of The Minerals, Metals and Materials Society.