Moore's Law doesn't work for solar cells

In pushing the US to produce all of its electricity from environmentally friendly sources in 10 years, former Vice President Al Gore points to the accomplishments of the technology industry, especially the dramatic gains in computing capability, as an example of what tech innovators can do.

"Think about what happened in the computer revolution," Gore said on NBC's Meet the Press programme recenty. "We saw cost reductions for silicon computer chips of 50% for every year and a half for the last 40 years," he said. "We're now beginning to see the same kind of sharp cost reductions as the demand grows for solar cells — they build new, more efficient facilities to build these solar cells."

Gore, who has formed a group, The Alliance for Climate Protection, for solar cell creation, was referring to Moore's Law, which explains the dramatic gains in compute performance. It stems from a 1965 paper written by Intel co-founder Gordon Moore, which found that the number of transistors put on a chip doubles every 18 months.

But does Moore's Law also apply to the solar energy industry? The short answer is no. As with microprocessor technology, the price and performance of photovoltaic solar electric cell is improving. And Gore can clearly point to price drops of solar cells to make his case. But the efficiency of those solar cells — their ability to convert sunlight into electric energy — is not increasing as rapidly.

In 1980, a solar cell cost about US$18 a watt; today that cost ranges roughly from US$3-US$4 (NZ$3.9-NZ$5.2) a watt, although some manufacturers say they can get the price much lower. For instance, California-based XsunX, has said its technology can produce energy for US$1.58 a watt and is aiming for improvements that would take that price below US$1 a watt.

But solar cells are just one aspect of the cost of solar power. Solar modules, the panels that comprise solar cells, have to be constructed with equipment such as inverters that convert DC into AC power, batteries and increasingly expensive metals, such as copper. Those necessities can double the per-watt cost of what someone pays for electricity from their utility. That's one reason installations are heavily subsidised by the US government.

Microprocessors use silicon, as so do solar electric cells. But while microprocessor producers can use improved design and manufacturing techniques to improve chip efficiency and performance, improvements to solar cells means changing materials. "Solar cells are dealing with the material science...," says Bill Johnson, senior research engineer at the Florida Solar Energy Centre. "It's really hard to change the materials."

Although the cost of producing solar cells has declined, their ability to convert sunlight into electricity remains low. Sunlight delivers about 1 kilowatt of energy per square metre, and only around 10% of it is converted into electricity, though some laboratories and companies report higher efficiency results.

Gore is half right when he points to decreasing production costs, says Nicholas Lenssen, an analyst at Energy Insights, a market research firm that's part of IDC. But the efficiency of solar cells isn't increasing at anywhere near the pace of production. "The efficiency gains are really slow," Lenssen says.

Even so, over time the cost-per-watt, especially in large solar farm deployments that could become alternatives to fossil fuel generations, could decline enough to be competitive.

Interest in reducing the price of solar generation is coming from a broad range of companies, including those in the IT sector. In May, for instance, IBM said it developed a technology for cooling solar cells used in solar farms, a thin layer of liquid metal to reduce temperatures of 1,600 degrees Celsius. This allows more light to be concentrated on the solar cell, improving the efficiency of solar cells.

Al Compaan, a physics professor at the University of Toledo in Ohio who studies solar energy, says Gore's effort "is a great stimulus to getting people to think about renewable energy." While the improvements in solar energy won't change as rapidly as Gore might like, there is, nonetheless, enormous potential, he says.