Solar power is getting so cheap it is overtaking fossil fuels – and that’s without next-generation photovoltaic technology and artificial photosynthesis
By Tom Chivers
“WHEN I was a student in the mid-1980s,” says Harry Atwater, “there were only a few megawatts of solar deployed worldwide. Enough to power a supermarket or something. Now, solar technology has a global output of about 500 gigawatts.” That is enough to provide Manhattan with all the electricity it needs about 50 times over. And solar power is only just getting going.
Atwater, who is director of the Joint Center for Artificial Photosynthesis at the California Institute of Technology, has been working on photovoltaic technology for his entire career. He has seen it grow from a novelty found in only a few labs into a giant industry bigger than the flat-screen TV market. Some $131 billion was invested in solar in 2018 alone. But Atwater reckons the advances set to arrive in the next few years could leave that progress in the shade.
Solar power has become so cheap that before long, it will overtake fossil fuels as the world’s preferred source of electricity. And a lot of low-cost solar power will be crucial if we are to have any chance of limiting global warming to 1.5°C.
Financial company Lazard releases an annual report into the “levelised” cost of different energy sources – that is, the cost once you take into account the whole life cycle of a power plant, including manufacture and disposal, and disregard government subsidies. Its 2018 report showed that large-scale solar power and onshore wind were, on average, the cheapest forms of energy. Sometimes, gas could be cheaper, but more often it wasn’t. Lazard found that between 2017 and 2018, the cost of solar energy fell from around $50 per megawatt-hour to around $43, a drop of 14 per cent in a single year. Since 1976, its cost has plummeted by about 99.9 per cent.
This has led to remarkable growth in its use. “I’ve been an analyst for 13 years,” says Chase. “When I started, I thought that solar could only ever be at most 1 per cent of global electricity generation. In 2018, it was around 2.3 per cent. In China, it’s 2.5 per cent, which doesn’t sound like a lot, but China uses a lot of electricity. And in places like California, they already have 8 per cent from solar.” Last year, a Bloomberg NEF report estimated that the cost of photovoltaic cells will fall by a further two-thirds by 2050, at which time photovoltaic solar will make up 24 per cent of global energy generation.
It is astonishing progress – and it can’t come too fast, with soaring carbon emissions heating the world to temperatures not seen for 125,000 years. Solar energy, through the venerable but reliable technology of silicon photovoltaic cells, will be a major part of efforts to avoid the worst of climate change.
So too will new technologies that we can expect in coming years and decades. First will be improvements to photovoltaic technologies, followed by changes in the storage of solar energy that could start to have an impact within the next decade or two. Finally, we may see much more speculative, long-term ideas, such as harvesting sunlight from space and beaming it down to Earth, come to fruition. These latter developments could revolutionise the energy economy.
The near-term technologies rely on new forms of solar panel. At the moment, most photovoltaic cells are built from silicon. The first practical photovoltaic cell, made by Bell Labs in 1954, turned just 6 per cent of the solar energy that struck it into usable electricity. By 1959, a rival firm had achieved 10 per cent. A quarter of a century later, in 1985, an Australian team reached 20 per cent efficiency. And in 2016, a group at the University of New South Wales achieved almost 35 per cent, the current world record. Most commercially available solar cells peak at about 20 per cent efficiency.
But silicon, although widely available, is hard to work with. “Silicon cells are made by taking sand, heating it up to thousands of degrees, and reducing it from silicon oxide to pure silicon,” says Lance Wheeler, a materials scientist at the www.nrel.gov National Renewable Energy Laboratory (NREL) in Colorado. “It’s energy-intensive and a very sensitive process – you have to do it in perfect conditions.” That means it is prohibitively expensive to produce in the West. Most photovoltaic cells are built in China, backed by large amounts of government subsidy and where lower-wage skilled labour also helps the economics. Manufacturing costs have fallen as the industry has grown, but there is a limit to the energy savings possible.
In recent years, however, a new form of photovoltaic cell has become increasingly attractive. Instead of silicon, this type uses various kinds of minerals called perovskites. The first perovskite cell was developed in 2009. It had an unpromising efficiency of 3.8 per cent, lower than that of the first silicon cell in 1954. But whereas it took silicon cells 30 years to get to 20 per cent efficiency even in the lab, perovskite-based photovoltaic cells took less than 10, reaching over 23 per cent last year. “They’re still something like 2 to 4 per cent less efficient than [regular] solar,” says Wheeler. “But it’s gone from mediocre to fantastic in a decade.”
“Solar power will overtake fossil fuels as the world’s preferred form of electricity”
More importantly, perovskite is just as abundant and cheap as silicon, and much easier to work with. “You can start with a jar of liquid and print it, with any kind of printing press you can think of, at room temperature,” says Wheeler. He mentions one company that has bought an old Kodak factory in New York and repurposed its printers to deposit thin layers of perovskite onto plastic film for use in photovoltaic cells. “An old newspaper printing press could be retrofitted too,” says Wheeler.
There is a problem, though. Silicon photovoltaic cells are incredibly durable. If you install some on your roof, you can confidently bet that they will work for 25 years. Until recently, perovskite cells were only stable for a few hours. But NREL now has cells in its lab that work at 90 per cent of their starting efficiency for 250 days. “It turns out that perovskite is not inherently unstable,” says Wheeler. “It’s the materials next to it that have been the problem. If you can engineer them, you get good stability.”
Chase warns against premature excitement. “There’s no commercial product yet,” she says. But she says that longer term, perovskite could be an interesting possibility. She and Wheeler both have hopes for hybrid cells, with a layer of perovskite sitting under one made of silicon, because they absorb different wavelengths and so would push efficiency even higher.
Another fundamental problem with solar energy is that the sun doesn’t shine over an area all the time. You get more sun at noon than at midnight and at higher latitudes you get more in midsummer than in midwinter. But you need energy all day and all year. So other researchers are working to harness the sun’s energy in a more storable way.
The storage problem
“In the coming world of sustainable energy – with high levels of electrification – we’re going to have a serious storage issue,” says Atwater. “Batteries will be part of the solution, but imagine you’re in Norway or Sweden where they have 10 times as much sun in summer as winter. You won’t be able to store energy for six months in a battery.” He contrasts this with fossil fuels, which have stored solar energy for millions of years. “We use them every day because they’re amazingly energy-dense, storing them is inexpensive and they’re portable. Chemicals represent the ultimate form of energy storage,” he says
The goal is to use sunlight to create chemical fuels. Atwater’s lab is working on artificial photosynthesis: harnessing the sun’s energy to turn inert chemicals into energy-filled ones.
The most basic version of this splits water to produce hydrogen. A photovoltaic cell creates a current that passes through water via a positively charged anode and a negative cathode. Oxygen ions gather at the anode and hydrogen ions at the cathode. “A lab prototype here reported a world record of 19.3 per cent efficiency,” says Atwater: that is, nearly 20 per cent of the solar energy hitting the photovoltaic cell was stored in usable hydrogen fuel. “It’s an expensive device, but it’s a demonstration of what’s possible.”
It is relatively easy to make hydrogen from water. “We did it as high school students with a copper wire and a platinum wire,” says Atwater. “Your feedstock is inexpensive.” One complication is that pure water is needed to avoid the cathode becoming fouled by impurities, but it is straightforward to produce very pure hydrogen that can be used in fuel cells.
“Artificial photosynthesis will allow solar energy to be stored through the winter”
A bigger hurdle is making hydrocarbon fuels from atmospheric carbon dioxide, says Atwater. “It’s more difficult for fundamental reasons. The chemistry of carbon is very rich: diamond, graphite, polymers, methane – all are based on carbon.” Carbon dioxide can be turned into many different forms, he says. “The challenge is designing a process where if you want to make one thing, you get exclusively that. If you want to make ethanol, you don’t want to end up with carbon monoxide.” But although it is harder, it has two advantages over water-splitting: it strips a greenhouse gas from the atmosphere and, alongside the fuel, it results in forms of carbon that can be used to make many complex molecules, including plastics.
At the moment, Atwater’s lab is able to turn atmospheric CO2 into a combination of about 70 per cent ethanol, a fuel, and ethene, the basis for many plastics. The other 30 per cent is mainly hydrogen, which is itself useful. It is a far less efficient process than water-splitting, but progress is being made.
The difficulty is that artificial photosynthesis relies on catalysts, but existing ones rely on external energy and are often toxic. So some other groups are looking to nature for alternatives. Erwin Reisner and his team at the University of Cambridge have made progress with “semi-artificial” photosynthesis, using natural enzymes as the catalysts. These proteins are much more effective catalysts and are better at creating precisely the products you want with no extra energy push. The downside is that they break down in a few hours or days. “In organisms, they are constantly repaired and replaced,” says Reisner. He has two hopes: that this semi-artificial technology can be harnessed to improve fully artificial catalysts or, further in the future, that bioengineered algae could create the required hydrocarbons, using natural enzymes that they repair in the traditional way.
Artificial photosynthesis will allow solar energy to be stored through the winter and make it easier to power systems that are hard to electrify, such as heavy vehicles and air transport, says Atwater. And although the technology is still new, he believes it could improve rapidly. When talking to his grad students, he tells them about the photovoltaic sector when he started out. “Now it’s a mature, fully integrated industry,” he says. “I think artificial photosynthesis is at the same stage the photovoltaic industry was at when I was a student.”
As well as artificial photosynthesis, scientists at Chalmers University of Technology in Sweden are working on a liquid that traps solar energy and releases it as heat. It can already store energy about as effectively, in terms of joules per kilogram, as a lithium-ion battery, and for months at a time. It is less efficient than a solar panel, but team leader Kasper Moth-Poulsen is hopeful it will improve. “For it to absorb energy, all you need do is put it in the sun,” he says. Then, when you need to release the heat, you just add a catalyst. An efficient solar heating system that works all year round could have a huge impact on emissions.
With luck, commercial artificial photosynthesis and solar liquids could be with us in a couple of decades. Beyond that, Atwater is imagining the ultimate form of energy from the sun: space-based solar panels (see “Orbiting power plants“).
As exciting as such developments could be, on its own, solar won’t be enough. “We need to be more efficient,” says Chase. “We need to use more public transport, more cycling.” Still, though, the growth of solar is a reason for cautious optimism. “The progress has made me start to think that we can maintain some sort of developed-world lifestyle using solar,” she says.
Orbiting power plants
At New Scientist, writers are steered away from comparing new developments to science fiction because this can become a prop that is reached for too readily. But here, we can’t avoid it. Harry Atwater at the California Institute of Technology is working on a project that is literally the stuff of science fiction: space solar. “This is an idea that was first floated in 1941 in an Isaac Asimov short story,” says Richard Madonna, a consultant on the project. “He envisaged sending sunlight from space to Earth via electromagnetic energy.”
In the 1960s, scientists started to look seriously at the idea. But launching things into space was expensive and the components – solar panels, solar reflectors and radio transmitters – were all heavy, bulky and inflexible. So it was economically infeasible.
But the Caltech group, including Atwater and a colleague called Sergio Pellegrino, recently revisited the idea. The resulting plan was for carbon booms that could be wrapped up and would spring back into shape to provide structure, as well as flexible perovskite photovoltaic cells that would be just a micrometre thick. “The whole thing could be rolled into a cylinder for launch,” says Madonna. He envisages 60-square metre structures that would be linked to form a superstructure in orbit. SpaceX’s Falcon Heavy rocket, “which is now becoming a feasible space vehicle”, could carry about 10 of them in a launch, says Madonna.
Instead of having a single, central radio transmitter, which would have needed lots of copper wire to transport the electricity, each cell would act as its own transmitter, beaming radio frequency energy down to Earth out of the side facing away from the sun. Diffraction effects mean that the larger the structure, the more tightly focused the beam could be, and the more efficient the transmission. These structures would have to be large. “It depends on how efficient it is,” says Madonna. “But to power a city, it would have to be kilometres across.”
This technology isn’t imminent. The Caltech group hopes to have a small-scale test in low orbit in the next 10 years, but it could be decades before it is really possible, if it happens at all. Madonna says it won’t be like nuclear fusion though: always the next big thing but never quite arriving. “It’ll either happen or it won’t,” he says. “It won’t be 20 years away for the next 50 years.”
First published at New Scientist – Wednesday August 7 2019.