Self-Assembled Carbon Nanotube Antennas for Solar Power Revolution

NovaSolix’s carbon nanotube (CNT) antennas are small enough to match the nano-scale wavelengths of sunlight. Antennas can convert electromagnetic spectrum much more efficiently than photovoltaic (PV) cells. When perfected, NovaSolix antennas will capture over four times the energy of current solar panels. They will reach nearly 90% efficiency versus ~20% for todays solar panels.

NovaSolix has invented a self-assembling antenna array solar cell which will be 2-4 times more efficient at a less than one-tenth the cost per watt of existing solar.

NovaSolix claims to have demonstrated a proof of concept to third parties that has touched 43% efficiency. That’d suggest a 72 cell solar module near 860 watts, with a 90% solar cell pushing 1700 watts.

They could buy used manufacturing hardware and retrofit them in the early stages of growth. The first manufacturing lines could cost $4.1 million, and would initially produce ~45% efficient modules, at a clip of 20MW/year with a proposed price of 10¢/W. At full efficiency, costs are cut in half and volumes per year doubled.

Solar Powered Car Gets

Sono Motors is a separate company that makes an electric car with built-in solar power supplemental charging. A sunny day can provide 18 miles of driving range on a 24% efficient solar cell. If NovaSolix increased solar cell efficiency to 90% then one day of sunlight driving would be 67 miles.

51 thoughts on “Self-Assembled Carbon Nanotube Antennas for Solar Power Revolution”

  1. “Yet none of them ever seem to work out.”

    Nonsense – you haven’t been looking at the big picture.

    Both solar and batteries have been on an exponentially declining price curve for about 70 years when looking at the best price per lifetime Kwh in any given year. Lifetime price per Kwh stored & released by batteries has dropped by half about every 6.5 years. Lead acid gave way to NiCad to Nickel Metal Hydride to Lithium Ion. Now we’re looking at solid state and iron and sodium and sulfur in the labs, literally thousands being tested. Only one needs to be better to upend things again.

    Solar – same story, even faster progression (about 3.5 years for the panels to drop by half, although the associated electronics haven’t dropped quite as quickly).

    Experts are now converging on the tipping point being “within 9 years.” That of course means that some niches will already be economical (California is now building three battery “peaker plants” instead of 3 natural gas plants, expected to be cheaper right away).

    9 years from now, solar plus batteries will be cheaper than energy generated by fossil fuel plants. 15 years from now, together they will be half that price.

  2. They build it in 10 layers with nanotubes of different lengths. Shorter ones on top; long waves bypass that layer to the next etc.

    Cost – the primary cost for existing solar cells is manufacturing the silicon substrate. These will be roll-to-roll printed on plastic.

  3. Having made and commercialized optical nanotube-containing technologies over the last two decades, I have to say that while it is risky and hard, what this company is doing has a good scientific basis and what remains includes learning (fast enough) the techniques to control processes and scale manufacturing, and of course finding early and sometimes niche applications that can afford to pay for a new technology that is always more expensive at introduction. Investors ought to know that the proof-of-principle stage has already passed and what remains includes finding the quickest paths to early commercialization.

  4. they have demonstrated the device at least twice in public–over a year ago– to hundreds of people–once at Hacker’s Conf and at Asilomar in regular light–what are you talking about? I saw it–pretty impressive!

    Nice description of making PV devices–spoke to their CTO –not what there are doing.
    They are using same equipment as flat TVs and don’t use silicon–should be dirt cheap!

  5. To capture the entire spectrum you could absorb all the energy, then re-emit that energy at a lower notch frequency gated by a tuned device. The convertor could be optimized for the narrow-band emission.

  6. The efficiency is based on the temperature of solar surface. Also to capture the entire spectrum you will need different length nanotubes.

  7. They might only have to be hundreds of times faster. I did a search on rectifier switching speeds, picked a random device and here’s its switching speed “High switching speed: trr ≤ 4ns”.
    Thanks for the reply!

  8. Bryan, it seems like we’ve heard of hundreds of solar “breakthroughs” over the last decade or so. And it’s the same with battery “breakthroughs”. Yet none of them ever seem to work out. For some reason they’re never commercially viable – in almost all cases I think a product is never even introduced.

    What’s your take on this? Is there a way to identify the viable stuff from all the noise? You’ve been reading and sharing tech breakthroughs for a long time, and I wonder if you’ve been able to figure out some heuristics for what to believe and what to ignore.

  9. Does that apply just to “light” or the whole EM spectrum? It sounds like they’re capturing more than visible light.

  10. The laws of thermodynamics do still apply. And 90% efficiency is impossible for a solar cell. 86% is the max. I think they must have rounded up. If you want more info look up landsberg limit and Thermodynamic efficiency limit.

  11. Sunlight is just very small radio wave. The problem is the minute size of the antennas and the high speed of the circuit. The frequency of light is about 5 x 10e14 Hz so the rectifying diode would have to work ten thousand times faster than the fastest computer which would require a new technology.

  12. Infrared rays would be easier to rectified. I read something about it a few years ago. I haven’t heard anything since.

  13. I thought the laws of thermodynamics still applied so that the conversion of light to electric was still limited by the average temperature of the solar surface. 90% conversion seems too high. I would think conversion in the 60-70% range would be max.

  14. So… keep thinking about the problem: WHITE light — multispectral light — is all jumbled together spatially. If the maximum conversion efficiency occurs when the light’s wavelength is the same as the resonant length of the dipoles, how to arrange for a bunch of them to be all over the place, in the right place, to catch all the incoming mixed waves?

    That’s the first significant gotcha.

    Then — surviving that — the next becomes, how to overcome rectification overhead? See my other comment for that.

    Then — surviving that — how to make this stuff for less than 70¢/watt — the present cost by the container-load of Chinese PV cells (already mounted in nice glass panels)?

    Myself, I think this is a persistent “grad-school research topic” that remains awesome sounding on paper, and rather glaringly full of gotchas in real life. Hard ones, that undermine the glib narrative of cheapness, awesomeness and a near-future of radically improved conversion opportunities.

    You know that by recognizing the “selling snakeoil” narrative that always accompanies magic unicorn horn blabberings: ultimate solutions to electric vehicles, prevention of global warming, remarkably transformative agents to redirect our power and energy efforts in the future. Huge markets. Insatiable demand. Overwhelming dominance propositions.

    Just saying,
    GoatGuy

  15. So… since I’m in the Elephant-in-the-Living-Room mode…

    ⇒ WORKING DEMO PRODUCT ⇐

    Is there any? You know, like a single rectantenna element, just one of the things, on a chip, in a lab, under a microscope, illuminated with perfectly tuned laser light (for maximum conversion efficiency), delivering femtoamps of current at millivolts from the integrated ExaHertz rectifier?

    No? Why not?

    Yes? Linkies!!! I want to see ’em, and read the article, measuring output and conversion efficiency.

    ⇒ RECTIFICATION LOSS ⇐

    Its unlikely that most of my readers will know much about this, but it turns out that turning A/C (such as picked up by the rectantenna dipoles from light hitting ’em) into little pulses of DC (which is useful for getting off-chip and onto wires, leading ultimately to batteries, inverters and grid-matching hoohahs) is key.

    And rectification ALWAYS imposes a “toll to the quantum troll” loss-of-rectificatioin. In ordinary semiconductors, with ordinary diodes, this is called the “bandgap work function”. A silicon diode won’t let the AC go through in the “forward” direction without subtracting 0.7 volts or so from the positive-going AC pulses.

    For things like your high-end PC with high-end processor, needing to turn the power-supply 5 (or 3.3) volts to the core’s 0.9 to 1.3 volts, there are many banks of things called “synchronous FET rectifiers” which run a perfectly timed A/C signal to the ‘gate’ of a MOSFET, to turn it ‘on’ just in time for the A/C positive pulse to pass through with MUCH lower forward voltage drop than a conventional ‘dumb rectifier’.

    The problem really is then, what levels of voltage would the dipoles produce?

    Well, more physics… if you can get oscillations induced in an antenna element, when the wavelength matches that of the antenna, the waves reinforce, increasing net voltage. Rectification losses are proporionatedly less with higher driving voltage.

    Problem then is that removing energy from the resonant antenna element in turn rapidly diminishes the resonance energy itself. Harumph… So yet another countering strategy is to only rectify in distinct windows-of-time: No rectification for 20+ waves, rectify for 5 waves, and repeat.

    In any case, none of this is EASY in any conceivable way.
    Its all theory.
    Hence…

    The questions.
    GoatGuy
    ________________________________________

    PS: The other thing is fabrication complexity. Making a conventional photovoltaic cell requires a small plate of silicon to be put in a series of chambers, alterately implanting different ions, then heating it to diffuse them to different depths. Finally, sputtering metals onto the surface to pick up the photoelectrons, and deliver power. All in parallel, all without special lithography, all in bulk. Presently for less than 70¢/watt from China by the container load. That’s not going to be beat easily. No time soon.

  16. SUNLIGHT IS RADIO WAVES they are all electromagnetic waves just at diffrent frequencies this work is real it was first done at idaho national labs in 2011 see “solar cells that work at night” by inhabitat.com

  17. I USED A SIMILAR TECHNOLOGY FOR MAKING POWER FROM AMBIENT EARTH HEAT USING LONG INFRARED NANTENNAS. IT WORKS 24-7 INDEPENDENT OF SUNLIGHT. I RAN A PENDULUM MOTOR WITH IT

  18. The idea of nantennas has been around for a long time. Light and radio are both EM waves, just very different wavelengths. The problem is making such antennas, since they need to be very small to match the much shorter wavelength. More on that in wikipedia /wiki/Optical_rectenna

    Optical heat pumps have been proposed before as well, and there’s some research on it. There was an article on something similar a while back on NBF. But there are still many challenges to make a practical system and scale it up.

  19. Sounds too good to be true, as if sunlight is radio waves… If the nanoscale antennas work for that part of the EM spectrum, why have I never even heard of this possibility before?
    If so, use even smaller such antennas for the blue end and slightly larger for the IR end. Imagine that, “solar cells” working (slightly) at night. Perhaps they can do it in reverse and emit IR at only a certain frequency that escapes through the “window” of infrared absorbing molecules. Global warming solved?

  20. In the past few years, Elon Musk has developed a megawatt storage battery. Maybe Elon should develop a CNT solar panel. That would really make a breakthrough with solar energy.

  21. Large volume production of CNT-based solar panels requires large volume production of CNTs, which is still difficult and expensive. These panels probably need pretty specific type/size/etc of CNTs, which is even more difficult. I expect that to be their primary scaling and pricing challenge.

  22. Anti=reflective Coatings are here and now and in addition to significantly upping efficiency provide such a smooth service that dust particles literally slide off the panel surface. This will reduce maintenance expenses significantly. ARC

  23. You’re correct but for reasons i dont remember, there are different quantum effects that limit the efficiency of nanosized rectifiers even if they are made of conducting carbon nanoantennas and other things analogous to radio components instead of semiconducting materials.

  24. They’re claiming they are conducting the electrons through an antenna effect not due to electron promotion that are typically scene in pn junction based PV. Therefore, they are not subject to the shockley-queisser limit since they will not suffer from Auger recombination but only from the inefficiencies of the antenna itself which is purely dependent on it’s shape, size and material properties

  25. They might have achieved 46% efficiency and which would be revolutionary if their panels are much less expensive than silicon. But efficiencies as high as they are claiming are impossible as a result of quantum inefficiencies.

  26. They might have reached 46% efficiency and that would be a revolution of they could make those panels ten times cheaper than silicon and other semiconductors. Some experts contend that efficiencies as high as they are claiming are impossible as a result of quantum inefficiencies.

  27. Is the key needed tech to make the rectenna approach viable a terahertz or above rectifier then? That seems to be the real show stopper that hasn’t been adequately demonstrated so far…

  28. They might only have to be hundreds of times faster. I did a search on rectifier switching speeds, picked a random device and here’s its switching speed “High switching speed: trr ≤ 4ns”.
    Thanks for the reply!

  29. The laws of thermodynamics do still apply. And 90% efficiency is impossible for a solar cell. 86% is the max. I think they must have rounded up. If you want more info look up landsberg limit and Thermodynamic efficiency limit.

  30. Bryan, it seems like we’ve heard of hundreds of solar “breakthroughs” over the last decade or so. And it’s the same with battery “breakthroughs”. Yet none of them ever seem to work out. For some reason they’re never commercially viable – in almost all cases I think a product is never even introduced.

    What’s your take on this? Is there a way to identify the viable stuff from all the noise? You’ve been reading and sharing tech breakthroughs for a long time, and I wonder if you’ve been able to figure out some heuristics for what to believe and what to ignore.

  31. Sunlight is just very small radio wave. The problem is the minute size of the antennas and the high speed of the circuit. The frequency of light is about 5 x 10e14 Hz so the rectifying diode would have to work ten thousand times faster than the fastest computer which would require a new technology.

  32. I thought the laws of thermodynamics still applied so that the conversion of light to electric was still limited by the average temperature of the solar surface. 90% conversion seems too high. I would think conversion in the 60-70% range would be max.

  33. So… keep thinking about the problem: WHITE light — multispectral light — is all jumbled together spatially. If the maximum conversion efficiency occurs when the light’s wavelength is the same as the resonant length of the dipoles, how to arrange for a bunch of them to be all over the place, in the right place, to catch all the incoming mixed waves?

    That’s the first significant gotcha.

    Then — surviving that — the next becomes, how to overcome rectification overhead? See my other comment for that.

    Then — surviving that — how to make this stuff for less than 70¢/watt — the present cost by the container-load of Chinese PV cells (already mounted in nice glass panels)?

    Myself, I think this is a persistent “grad-school research topic” that remains awesome sounding on paper, and rather glaringly full of gotchas in real life. Hard ones, that undermine the glib narrative of cheapness, awesomeness and a near-future of radically improved conversion opportunities.

    You know that by recognizing the “selling snakeoil” narrative that always accompanies magic unicorn horn blabberings: ultimate solutions to electric vehicles, prevention of global warming, remarkably transformative agents to redirect our power and energy efforts in the future. Huge markets. Insatiable demand. Overwhelming dominance propositions.

    Just saying,
    GoatGuy

  34. So… since I’m in the Elephant-in-the-Living-Room mode…

    ⇒ WORKING DEMO PRODUCT ⇐

    Is there any? You know, like a single rectantenna element, just one of the things, on a chip, in a lab, under a microscope, illuminated with perfectly tuned laser light (for maximum conversion efficiency), delivering femtoamps of current at millivolts from the integrated ExaHertz rectifier?

    No? Why not?

    Yes? Linkies!!! I want to see ’em, and read the article, measuring output and conversion efficiency.

    ⇒ RECTIFICATION LOSS ⇐

    Its unlikely that most of my readers will know much about this, but it turns out that turning A/C (such as picked up by the rectantenna dipoles from light hitting ’em) into little pulses of DC (which is useful for getting off-chip and onto wires, leading ultimately to batteries, inverters and grid-matching hoohahs) is key.

    And rectification ALWAYS imposes a “toll to the quantum troll” loss-of-rectificatioin. In ordinary semiconductors, with ordinary diodes, this is called the “bandgap work function”. A silicon diode won’t let the AC go through in the “forward” direction without subtracting 0.7 volts or so from the positive-going AC pulses.

    For things like your high-end PC with high-end processor, needing to turn the power-supply 5 (or 3.3) volts to the core’s 0.9 to 1.3 volts, there are many banks of things called “synchronous FET rectifiers” which run a perfectly timed A/C signal to the ‘gate’ of a MOSFET, to turn it ‘on’ just in time for the A/C positive pulse to pass through with MUCH lower forward voltage drop than a conventional ‘dumb rectifier’.

    The problem really is then, what levels of voltage would the dipoles produce?

    Well, more physics… if you can get oscillations induced in an antenna element, when the wavelength matches that of the antenna, the waves reinforce, increasing net voltage. Rectification losses are proporionatedly less with higher driving voltage.

    Problem then is that removing energy from the resonant antenna element in turn rapidly diminishes the resonance energy itself. Harumph… So yet another countering strategy is to only rectify in distinct windows-of-time: No rectification for 20+ waves, rectify for 5 waves, and repeat.

    In any case, none of this is EASY in any conceivable way.
    Its all theory.
    Hence…

    The questions.
    GoatGuy
    ________________________________________

    PS: The other thing is fabrication complexity. Making a conventional photovoltaic cell requires a small plate of silicon to be put in a series of chambers, alterately implanting different ions, then heating it to diffuse them to different depths. Finally, sputtering metals onto the surface to pick up the photoelectrons, and deliver power. All in parallel, all without special lithography, all in bulk. Presently for less than 70¢/watt from China by the container load. That’s not going to be beat easily. No time soon.

  35. SUNLIGHT IS RADIO WAVES they are all electromagnetic waves just at diffrent frequencies this work is real it was first done at idaho national labs in 2011 see “solar cells that work at night” by inhabitat.com

  36. I USED A SIMILAR TECHNOLOGY FOR MAKING POWER FROM AMBIENT EARTH HEAT USING LONG INFRARED NANTENNAS. IT WORKS 24-7 INDEPENDENT OF SUNLIGHT. I RAN A PENDULUM MOTOR WITH IT

  37. The idea of nantennas has been around for a long time. Light and radio are both EM waves, just very different wavelengths. The problem is making such antennas, since they need to be very small to match the much shorter wavelength. More on that in wikipedia /wiki/Optical_rectenna

    Optical heat pumps have been proposed before as well, and there’s some research on it. There was an article on something similar a while back on NBF. But there are still many challenges to make a practical system and scale it up.

  38. Sounds too good to be true, as if sunlight is radio waves… If the nanoscale antennas work for that part of the EM spectrum, why have I never even heard of this possibility before?
    If so, use even smaller such antennas for the blue end and slightly larger for the IR end. Imagine that, “solar cells” working (slightly) at night. Perhaps they can do it in reverse and emit IR at only a certain frequency that escapes through the “window” of infrared absorbing molecules. Global warming solved?

  39. In the past few years, Elon Musk has developed a megawatt storage battery. Maybe Elon should develop a CNT solar panel. That would really make a breakthrough with solar energy.

  40. Large volume production of CNT-based solar panels requires large volume production of CNTs, which is still difficult and expensive. These panels probably need pretty specific type/size/etc of CNTs, which is even more difficult. I expect that to be their primary scaling and pricing challenge.

  41. Anti=reflective Coatings are here and now and in addition to significantly upping efficiency provide such a smooth service that dust particles literally slide off the panel surface. This will reduce maintenance expenses significantly. ARC

  42. You’re correct but for reasons i dont remember, there are different quantum effects that limit the efficiency of nanosized rectifiers even if they are made of conducting carbon nanoantennas and other things analogous to radio components instead of semiconducting materials.

  43. They’re claiming they are conducting the electrons through an antenna effect not due to electron promotion that are typically scene in pn junction based PV. Therefore, they are not subject to the shockley-queisser limit since they will not suffer from Auger recombination but only from the inefficiencies of the antenna itself which is purely dependent on it’s shape, size and material properties

  44. They might have achieved 46% efficiency and which would be revolutionary if their panels are much less expensive than silicon. But efficiencies as high as they are claiming are impossible as a result of quantum inefficiencies.

  45. They might have reached 46% efficiency and that would be a revolution of they could make those panels ten times cheaper than silicon and other semiconductors. Some experts contend that efficiencies as high as they are claiming are impossible as a result of quantum inefficiencies.

Comments are closed.