#1 cleantech news, reviews, & analysis site in the world. Subscribe today. The future is now.


Published on October 6th, 2014 | by Anand Upadhyay


Solar Thermophotovoltaics — Getting To 80% Efficiency

October 6th, 2014 by  

I think we are actually fortunate to have witnessed solar photovoltaic go commercial within our lifetime. With the support of governments around the world and citizens chipping in resources, there is not much doubt that solar is going to be a huge part of our future. But PV as we know it may not last throughout — it has an inherent problem. You see, PV can only convert a small portion of the energy spectrum of sunlight to electricity.

Image - http://www.onlinefast.org/wwutoday/sites/onlinefast.org.wwutoday/files/imagecache/full/Vehicles-of-the-future-3.jpg

Viking 29 runs using a thermophotovoltaic generator. (Image Source: Onlinefast)

Efficiency of Solar PV

The Shockley-Queisser limit, a fundamental to solar energy production, limits the efficiency of an “ideal” solar cell to around 34%. In the real world, the efficiency we get to see in commercial mono-crystalline solar cells is only about 22% at the moment. You might have heard of solar cells achieving more than 40% efficiencies, but those are multi-junction PV cells. For now, they are still too costly for “normal” applications — they’re mostly just applicable in outer space. In fact, NASA’s Mars missions has employed multijunction solar cells.

Going beyond Solar PV

Coming back to our earth, PV basically has a problem absorbing all the photons from the sun because they are at different energy levels. Being a semi conductor, PV can only use photons at a certain energy level. If the energy is too low, it will be wasted; if it is too high, the ‘higher part’ will get wasted.

Hypothetically speaking, what if one could take all the solar photons in a jar, mix it (for homogeneity), and funnel it to the PV? If the right PV panel was selected as per the average photon energy level, one would expect the efficiency to just shoot up! Theoretically speaking, if one could do this, the conversion efficiency could be a mind boggling 80% or even higher! The technology to achieve this is still under development, and it is called Solar Thermophotovoltaics (STPV).

So, What Exactly is Solar Thermophotovoltaics?

To explain in simple terms, STPV has two main elements — an absorber-emitter and a PV cell. The absorber part of the absorber-emitter absorbs solar energy from a concentrated field of reflectors. In doing so, it reaches a high temperature — in the order of 1000-1200°C. The heat collected at this high temperature is used by the emitter to produce photons which are then used by the PV cell to produce electricity.

Current Efficiency of Solar Thermophotovoltaics

The engineering which goes into STPV has a lot to do with selecting the right materials. The absorber should work in the right window of the solar spectrum and the emitter should be paired to emit photons in the region the PV cell can efficiently use.

You can appreciate how difficult the engineering is from the fact that a technology with a theoretical efficiency of 80% has not been able to cross a threshold of 3.2% in the real world! But a lot of developments in recent times have been keeping researchers on their toes.

Key Challenges

STPV working at a reasonable level is still somewhere in the future, as a number of important challenges need to be addressed. The materials should allow absorbing solar energy within specific wavelengths and also be able to withstand high temperatures. MIT has recently reported on its experience with a new class of materials called metallic dielectric photonic crystals, which exhibit both these properties.

In addition, MIT also reports that these materials can be prepared from any metal capable of withstanding high temperatures, using conventional standard manufacturing process. However, a large part of the problem is still being looked at only at the design table. The team pegs a time period of about five years for commercialization of STPV.

Future of STPV

Thermophotovoltaic systems have few, if any, moving parts, and are therefore very quiet and require low maintenance. These properties make them suitable for remote-site and portable electricity-generating applications. They can, and in fact have been used, even without solar energy (just the thermophotovoltaic part of the devices).

Most interestingly, in the real world, they have been used in a car! Viking 29 is a two-seat sports car designed and built by the Vehicle Research Institute. It is powered by thermophotovoltaics using CNG. This brings us to the possibility of hybridization of thermophotovoltaic systems. The same “generator” can be powered using fossil fuel or solar energy as the condition might be. But why bother, the future is solar, right?

PS: I did not get into economics or cost because the projected numbers available right now would not be so relevant once the system comes up for commercialization. However, we all know that the most important metric for the market is cost per kWh of electricity produced.

Tags: ,

About the Author

is an Associate Fellow with The Energy and Resources Institute (TERI, New Delhi) - an independent, not-for-profit research institute focused on energy, environment, and sustainable development. Anand follows the Indian solar market at @indiasolarpost. He also writes at SolarMarket.IN. Views and opinion if any, are his own.

  • Excellent article. Thanks Anand.

  • Filip_S

    Amsterdam based Amolf Research center director Polman is predicting panels with up to 70% percent efficiency in 2020 due to nano particles and light management

  • JamesWimberley

    The high-efficiency multi-junction cells on the market for Mars rovers, Special Forces etc., are indeed expensive. But a lot of effort is going into tandem cells that would be easy to fabricate, for example perovskite vapour-deposited over silicon. These could also break the Shockley-Queisser limit, but at prices that would keep lowering the unit cost of solar.

    • The SQ limit is only for single junction, I think the lab scale efficiency of MJ is already over 40% (source: wiki).

  • GCO

    Yet another stupidly misleading headline… Yes, and cars are “getting to sub 1 second 0-to-100”, it’s just that we’re still very far from that right now.

    No, really, what was the point? Click-bait?

    • Joseph Dubeau

      The “2-in-1” is a foolish article too.

  • Michael G

    This is fascinating – good reporting – thank you.

    If we got just 5% efficiency in every part of every road, roof, parking lot, wall, curtain, and car exposed to the sun at a low enough cost, then I would say, stick a fork in it, we’re done with the GHG problem.

    For interesting physics solutions to other problems, look at magnetocaloric effect for refrigeration to eliminate refrigerator gases (Nobel prize in 1933). Set to hit the market by 2015.



    Electricity storage by superconducting magnets – currently working in Wisconsin to support saw mills with huge spikes in power needs:


    • Bob_Wallace

      “at a low enough cost”

      It’s those pesky details that get in the way of perfect solutions….

    • Bob_Wallace

      The Tres Amigas project in New Mexico was suppose to include superconducting coil storage but that got dropped (or postponed enough that it dropped out of sight).

      This sounds like another interesting idea not ready for prime time.

      • Matt

        So what is the thing holding up the 3 USA grid connect at Tres Amigas? Approved in 2009 and last I heard I thought permits and construction contracts were signed. Is it political, funding, just slow construction, or all of the above and more? Anyone have insight?

        • Bob_Wallace

          Right now it seems to be financing. Which, to me, suggests that there’s not enough demand for moving power from one grid to another to make it profitable.

          And some transmission has been built between ERCOT and the Eastern grid since the Tres Amigas idea was floated. That may be a cheaper way to connect those two grids.

          Connecting the East and West Coasts may not have a lot of value at this time. Perhaps when renewable get to be a major percentage of our supply….

          • eveee

            Yup. Hi Bob, I just reported on that. They are looking for the remaining 25% financing. So says the local newspaper this year.

  • Gabor Pap

    I am all for technological advances, but efficiency is not the only measure that matters for any technology. Simplicity and the efficiency of the manufacturing process counts a lot too. You can have an 80% efficient converter but if it is so complex to manufacture that it costs 10 times as much as a 16% efficient one than which one do you think will be adopted?
    Plants only convert 1% of solar energy, does it mean we stopped growing them because they are so inefficient…

    • Good point Gabor. And as you read, simple or rather I should say, already existing manufacturing techniques can be used for STPV. To add to this since they don’t have any moving parts, reliability goes up. The description may sound tricky, but it is not any more difficult than PV manufacturing used right now. The big idea is basically to use what we are learning from nano tech. Advances in material science have always driven the world, its almost time for a new round of upgrades!

    • Omega Centauri

      Yes. I read about this a few months back. Sounds like you need to do the CPV stuff to get the emitter to the high temp required. Then the emitter has to be very special, otherwise most of the energy would be radiated in the IR band. I’m not surprised they are only getting 3% in actuality. Let them keep experimenting, but don’t hold your breath waiting for a paractical product to come out.

    • RobS

      Exactly right, the efficiency of current solar is adequate for the vast majoirty stationary applications. you can already generate 100% of a moderately efficient home’s energy needs with its available rooftop area, what we need to focus on is cost reduction and cost reduction of storage technology so that renewable penetrations beyond 30-40% are possible. Thought of another way, when installing a system what determines its size in most cases, available rooftop area or expense? I would argue most people size a system based on cost, generally people do not run out of usable rooftop before they run out of available money.

      • GCO

        Yes. Re PV system sizing, I suspect that most people unfortunately go with a leasing company, and as those tend to focus only on the most expensive kW⋅h of a bill (*), their customers end up with an undersized system, and are then stuck with it for the duration of the lease.

        I personally bought what I could afford first, colonizing the best areas of my roof, then expanded onto less desirable places as prices continue to fall. 🙂

        [My system use power optimizers (SolarEdge), which like micro-inverters, offer complete flexibility on the number, type and placement of modules, don’t suffer from partial shading, etc… This obviously helped a lot.]

        (*) Many US utilities have tiered rates, where e.g. the first 500 kW⋅h in a month, or ‘baseline usage’, are at a relatively low tier 1 rate, then the next 200 (100 to 140% of baseline) are a little more (tier 2), the next 300 (up to 200% baseline) even more (tier 3), etc.
        Unsurprisingly, people are ready to pay more to offset those top tiers than the bottom one or two…

  • Marion Meads

    I really thought they have achieved 80% solar radiation to electric energy efficiency. But no, they haven’t even crossed 3.2% yet, what a bummer, it may take decades before a series of breakthroughs can happen. Meanwhile, my solar water heater achieves over 95% efficiency in converting solar radiation into heat, a feat that was achieved by many others several decades ago.

    • Haha.. full points for sarcasm there. Now I hope you meant that your SWH converts 95% of “captured” solar radiation to heat. Otherwise solar to heat efficiency would not be more than 80% under operating conditions. And that in fact is the challenge here. To get 80% solar to electricity (not just heat). Personally I (would like to) believe that the team at MIT would be able to develop a commercial product within a decade.

      • Marion Meads

        You can theoretically convert 99% of intercepted solar radiation into heat energy. Use water that is a lot colder than air to start with, and you will have net black body radiation and sensible heat transfer even without the sun.

    • GCO

      my solar water heater achieves over 95% efficiency

      Oh really? That’s miraculous. And decades ago? Please, tell us what systems you’re talking about.

      Mine has a 2.32 m^2 collector and receives 4.7 kW⋅h/m^2/day of sunlight on average, or ~4 MW⋅h/year.
      The system’s output (OG-300 rating) is 1.85 MW⋅h/year — 46% efficiency.

      While I trust that a more elaborate installation would reach 60 or 70%, we’re still very far from your crazy-looking claims.

Back to Top ↑