MIT Thinks It Has Discovered the Perfect Solar Cell

'Thermophotovoltaics' offer the possibility of "ideal" solar efficiency.

new MIT study offers a way out of one of solar power's most vexing problems: the matter of efficiency, and the bare fact that much of the available sunlight in solar power schemes is wasted. The researchers appear to have found the key to perfect solar energy conversion efficiency—or at least something approaching it. It's a new material that can accept light from an very large number of angles and can withstand the very high temperatures needed for a maximally efficient scheme.

Conventional solar cells, the silicon-based sheets used in most consumer-level applications, are far from perfect. Light from the sun arrives here on Earth's surface in a wide variety of forms. These forms—wavelengths, properly—include the visible light that makes up our everyday reality, but also significant chunks of invisible (to us) ultraviolet and infrared light. The current standard for solar cells targets mostly just a set range of visible light.

That makes sense because visible light is the most intense form of light that reaches the Earth's surface. Many other forms, such as microwaves and x-rays, are mostly blocked by the planet's atmosphere, but the full spectrum reaching Earth still extends outward from what's known as the solar cell "band gap." This is the range of frequencies within which a material is able to convert solar energy into electrical energy.

The band gap is a feature of photovoltaic solar cells in particular. This is the scheme in which photons, the carriers of the electromagnetic force, and what we'd usually call "light," collide with atoms in some material. This collision delivers a bunch of extra force to those atoms, which respond by shedding electrons. All those electrons add up to current—electricity. It's an ingenious way to harvest energy, but it's currently not all it could be.

It's not terribly practical to build materials that can "harvest" photons at all energies. It's just not a generic interaction—any photon doesn't simply blast electrons from just any atom. The material is only tuned for a specific range of incoming photons, and the rest is dissipated as heat instead of electric current. 

We could stack different solar materials together into tight sandwiches, with each layer snagging different sorts of photons, but that quickly becomes too expensive. The band gap is really just a fundamental limit.

Image: stanford.edu

There are other ways of snagging photons than the photovoltaic effect, however, and it's here that that new MIT scheme comes in. It's a promising alternative that's known as solar-thermophotovoltaics. Instead of converting solar energy directly to current, these materials convert it all to heat.

Thermophotovoltaics take advantage of the everyday phenomenon of thermal emission. Basically, any material heated above absolute zero will emit some radiation (photons) because anything above absolute zero will feature some motion of charged particles taking place as increased amounts of energy (as heat) increase the kinetic energy of the material's constituent particles. Get electrons really loaded on caffeine (heat) and they start shaking and sweating off photons as radiation.

The neat thing about a thermophotovoltaic element is that it can take in a bunch of different solar wavelengths and convert them to just one, which can then be converted by a standard photovoltaic element to current. This is illustrated above.

The catch with thermophotovoltaics is that in order to be suitably efficient, they need the addition of sunlight concentrators, e.g. those big arrays of mirrors that focus sunlight at one location. That's fine, but concentration means loads of heat and also the need to aim that light at a certain place. Until a material comes out that can withstand loads of heat energy, thermophotovoltaics can't reasonably beat standard photovoltaics.

The MIT team, led by postdoc researcher Jeffrey Chou, suggests a new "two-dimensional metallic dielectric photonic crystal" as the solution. Their crystal is capable of both absorbing light from a wide variety of angles—meaning, a sunlight concentration system doesn't have to have a sun-tracker component—and can withstand temperatures of up to 1,800 degrees Fahrenheit for up to 24 hours at a time.

The end result is large-scale, low cost, and efficient solar-thermal energy conversion.

To achieve maximal efficiency, a thermophotovoltaic material needs to be able to trap and hold maximal amounts of solar energy. This means a having a very finely-tuned spectrum of sunlight absorption and thermal photon emission. 

Imagine one of those rainwater collection buckets people hook up to their gutters with an outflow that might be used to water a garden or whatever. If the outflow is too small or too big, the bucket is either going to go dry, leaving an unwatered garden, or it's going to overflow, wasting water. It's a bit like that, only the ins and outs are mediated spectrally.

Of course, building a rainwater bucket for solar energy isn't quite a matter of pipes and plastic. Finding a material that can do all of the above and that can be produced cheaply using existing methods has thus far proven elusive. 

"This is the first-ever device of this kind that can be fabricated with a method based on current ... techniques," Chou said in a statement provided by MIT, "which means it's able to be manufactured on silicon wafer scales."

Nearly any metal that can withstand the temperatures required will do, but the key to making it work is in how the material is ultimately fabricated. The secret to Chou and his team's success is in embedding the thermophotovoltaic material with nano-scale cavities, which are then filled with some easily polarized material. These nanocavities make it very easy to fine-tune the emission/absorption properties of the material.

The end result, according to the paper, which has just been published in the journal Advanced Materials, is "large-scale, low cost, and efficient solar-thermal energy conversion"—perhaps the first of its kind. Chou estimates that it could be commercially available within five years.