In Depth: How space-age solar is coming down to Earth

A concentrating PV (CPV) concept based on ultra-high-efficiency solar cells used in satellites is about to make the leap to utility scale at select sites around the world, after trials at a grid-­connected 1MW pilot project in New ­Mexico.

The US site marks the latest test of lens-enhanced “III-V” CPV technology being developed by Concentrix, a spin-off from German solar research institute Fraunhofer ISE, that has demonstrated in-field ­efficiencies of over 28%, roughly twice as high as conventional silicon PV ­installations.

At the heart of the design is a “space standard” triple-junction PV cell made up of stacked ­layers of gallium indium phosphide (GaInP), gallium arsenide (GaAs) and germanium.

Each semiconductor layer captures a different slice of the solar spectrum — short-wave radiation, medium-wave radiation and infrared — that together can boost cell-level efficiency to more than 40%.

“The idea of the III-V solar cell was around even before we started Fraunhofer ISE 30 years ago but only as a research interest,” states Andreas Bett, director of Fraunhofer’s solar cells and technology division.

“It was really in the last ten to 20 years that we began to look at it for different applications, starting with space because of the relatively high cost of the ­technology [which made ­Earth-based systems too expensive] but also because GaAs [the cell’s main semiconductor] is much more favourable than silicon in resisting the damage caused by energy particles in space.”

To bring the price of the satellite solar cell — which is about 100 times more expensive than traditional silicon versions — down for terrestrial applications, Concentrix began work with Germany’s Azur Space Solar Power to create a mass-producible cell that could be married to a low-cost, high-magnification collector lens.

The panel design, finalised after Concentrix was acquired by French semi­conductor electronics giant Soitec in 2009, consists of a rectangle of 5.7sq cm embossed-silicon Fresnel lenses set over an arrangement of 96 micro-scale cells.

Individual cells ­consist of a germanium substrate on which are “grown” multiple layers of GaAs and GaInP, joined together by tunnel diodes — ultra-fast, optically transparent electrical contacts.

“The process is to grow these layers almost atom by atom to create a solar cell that looks like a conventional silicon cell but internally has three junctions stacked on each other and so has a more complex range of absorption [of solar radiation],” says Bett.

This architecture, built using Soitec’s Smart Cut and Smart Stacking “atomic scalpel” technology, allows the cell’s band-gap — which determines the range of the solar spectrum a PV cell absorbs — to be tuned for different environments.

“These are proven technologies with heritage for our work in space solar cells,” notes Klaus-­Dieter Rasch, managing director of Azur Space, which has developed PV cells for more than 300 satellite applications since 1964.

“It is one of the reasons that we have been able to have this quick success [at the pilot project], moving from cell development to system installation.”

Proof has meant performing 30,000km above the Earth. “If these cells can survive from -150°C to 150°C, it is a good starting point,” he says. “It made our CPV concept the most straightforward concept you could think of because, in space, reliability is so important.”

Although Concentrix’s terrestrial module costs are being reduced “considerably” with the help of fabrication techniques such as blue-tape, sawing, ­dicing and pick-and-place, and the 30MW-capacity manufacturing line set up in Freiburg two years ago, the III-V cells remain pricey.

By using inexpensive Fresnel lenses with a concentration factor of 500 to focus sunlight onto these multi-junction cells, however, researchers have been able to slash the amount of semiconductor material needed to a small fraction, with first-­generation circular commercial cells just 3mm in diameter.

Rasch describes the difference. “For traditional flat-plate silicon panels you would get 1MW from a football-pitch-size installation, but with this CPV concept you would have the equivalent of a football pitch of area covered in Fresnel lenses but the active cells would only amount to the area of a medium-size car, about eight square metres.”

The lenses are fabricated by overlaying a 0.2-micrometre-thin film of silicon on a roughly 250sq cm glass sheet that is “stamped” with the rows of ­Fresnel structures and allowed to harden. The resulting topsheet — “lenses under glass” — translates into a number of operational advantages.

“A glass superstrate is easy to clean and performs very favourably in terms of durability under harsh conditions, and the Fresnel structure is protected underneath,” says Bett.

“To keep the module costs low we had to look to every manufacturing option. This embossing technology allows us high-­volume production of high-precision, robust collector lenses.”

The completed modules, which draw on design elements from the circuit-board industry and insulating-glass technology, use front and back plates made of glass similar to structural ­glazing found on skyscrapers.

They are fitted to a two-axis tracking system that follows the Sun to within 0.1 of a degree, driven by a control system that works on astronomical positioning and direct-current power output to calculate the next optimum angle for maximum power ­generation.

The New Mexico project, running commercially since early this year, follows a first outing for the III-V CPV technology in 2007, at 100kW-scale, at the ­ Instituto de Sistemas Fotovoltaicos de Concentración in Spain’s Castilla-La Mancha region.

Since then, prototype modules have been trialled in desert conditions at sites including the solar panel test field at Masdar in Abu Dhabi.

The 1MW demonstrator, outside a village called Questa, near Taos, is operated by oil major Chevron.

The brownfield project, located on the site of a decommissioned ­molybdenum mine, consists of 173 trackers positioning a total of 15,000 panels and is producing electricity under a power-purchase agreement (PPA) with utility Kit Carson Electric.

“We have taken what was a laboratory or module concept prototype and turned it into a real, industrial product,” states Soitec senior vice-president Hansjörg Lerchenmüller, who was at Fraunhofer when Concentrix was spun off.

“The Questa installation is at a relatively high altitude, where it benefits from a terrific solar resource and has been performing at very promising efficiencies — on most days, at a power-plant level of greater than 25% efficiency, and that includes all inverter and cabling losses.

“It has also shown the reliability of the technology. ­Degradation on the modules was ‘below measurement accuracy’ over the lifetime of testing to date. The end-of-life power of a satellite is much more ­important than that of the first days. This will become one of the main ­advantages of our CPV as well.”

The space-saving potential is expected to open up the option of agricultural grazing around future solar installations, which would be built with open rows between the mast-mounted modules. No permanent shading also means minimal impact on a site’s flora and fauna.

“Grass is beginning to grow in the shadows of the panels,” says Rasch. “So it is possible that you could recultivate the desert areas within the bounds of these solar farms.”

The technology’s environmental credentials are rounded out by making the modules from fully recyclable materials — chiefly glass, steel and copper — with an “extremely small amount” of semiconductor material.

Industrial-scale economics are borne out by a study by the Energy Research Centre of the Netherlands, which suggested that a III-V CPV system operating “in a good solar resource” could produce enough energy to offset the amount used during construction in just nine months.

The installation at Questa is paving the way for the roll-out of hundreds of megawatts of III-V CPV solar farms. PPAs are in the pipeline for one 150MW project and a series of installations totalling 155MW for utilities in California. Negotiations are also ongoing for developments in ­Europe and Africa.

Soitec is betting big on III-V CPV. It plans to expand ­capacity at its German manufacturing plant by 50MW in the next 18 months, and is close to ­selecting a factory site near San ­Diego, California. Azur Space, for its part, expects to expand its “short-range” production capacity for the cells, which currently stands at 200MW.

Demonstration projects are under way in countries including South Africa (70kW, with project blueprints for a 50MW buildout) and France (50kW grid-­connected, and proposals for a development of up to 50MW), with commissioning of the lead-off phase of a 10MW ­project in ­Morocco slated for early next year.

“The first of these small-scale projects greatly reduce the risk-perception when entering final negotiations with the banks and equity investors because it gives us months or even a year of experience — it is not guessing any more [about performance] when transferring from region to region,” states Lerchenmüller.

“For CPV, you have to go for solar-power stations of a certain size for it to be economic. Bigger is better,” says Bett.

“Take the Western US. Depending on the financial conditions, the sun resource there means that with this technology you could get [electricity production costs] down into the range of $0.13 per kWh, which is highly competitive.”

Nor has the triple-junction technology at the core of the III-V modules reached the peak of its development. Five- and six-junction cells with better than 50% efficiency are being worked up in the Fraunhofer ISE labs, with a view to creating CPV systems with 35% efficiency.

“This is achievable,” says Bett. “It will just take two or three more years.”