Another High Efficiency Solar Cell Material for 29+% Efficiency

Researchers combined an aluminum source into their hydride vapor phase epitaxy (HVPE) reactor, then demonstrated the growth of the semiconductors aluminum indium phosphide (AlInP) and aluminum gallium indium phosphide (AlGaInP) for the first time by this technique.

The current world efficiency record for MOVPE-grown GaAs solar cells that incorporate AlInP window layers is 29.1%. With only GaInP, the maximum efficiency for HVPE-grown solar cells is estimated to be only 27%.

Now that aluminum has been added to the mix of D-HVPE, the scientists said they should be able to reach parity with solar cells made via MOVPE.

ACS Applied Energy Materials – Growth of AlGaAs, AlInP, and AlGaInP by Hydride Vapor Phase Epitaxy

Researchers demonstrate hydride vapor phase epitaxy (HVPE) of AlxGa1–xAs, AlxIn1–xP, and AlxGayIn1–x–yP using an AlCl3 precursor. We study the growth of the AlxGa1–xAs alloy system to elucidate the effects of deposition temperature, V/III ratio, and group V precursor species on Al solid incorporation via AlCl3. Crucially, the presence of group V hydride at the growth front kinetically promotes the solid incorporation of Al. We use these insights to demonstrate controlled deposition of AlxGa1–xAs, and for the first time by HVPE, AlxIn1–xP and AlxGayIn1–x–yP. These results create exciting implications for HVPE-grown high-efficiency III–V solar cells and devices with reduced cost.

31 thoughts on “Another High Efficiency Solar Cell Material for 29+% Efficiency”

  1. ‘Nuclear isn’t remotely cost effective.’ Except that somehow Ontario sells power at a cost on par with Texas, half that of California, and with emissions between five and ten times lower than either. (Texas hasn’t got enough wind to do much for their emissions, but it has enough they’re having to start installing big Wartsila piston engines to keep the beat steady.)

  2. It’s been working on my firefox most of the time, with just occasional outages for a few hours at a time. Maybe a cache or cookies issue?
    edit: Also need to make sure the comments domain ( isn’t being adblocked.

  3. One problem we don’t consider …

    Except that we consider this all the time. It comes up with every solar power discussion.

  4. I’ve had to swap from Chrome to Safari to Firefox to Chrome over the past year or so to keep the comments working.

  5. There are what are called “capacity markets” in the PJM grid and generators bid into them and everybody else pays for it. Texas has another mechanism they use that is different/opposite but achieves the same result (they allow massive micro spikes in kwh costs but spread that cost/pain around).
    So already considered, already levied.
    Nuclear isn’t remotely cost effective.

  6. 32% efficient GaInP/GaAs/Ge triple layer cells are already standard for space applications. So this would be at most a mod to them:

    Note: I’ve been absent from the forums because for some reason the comments section wouldn’t display, no matter what I tried. I just installed a new PC yesterday, and now it works.

  7. One problem we don’t consider is the backup plants are spinning all the time. That cost should be levied against panels from a policy view, but they don’t. They just say, “green”, and everybody thinks it’s true. Panels are ugly, take up a lot of space. Nuke is the way to go. Batteries may mitigate the backup issue, but at what cost? Existing plants are a sunk cost, so let’s spend more?

  8. Well … doing the math on that, 1200 mT, at 114 g/mol, is about 10.5 megamoles of In. 

    If we take the active layer of InP (or a tertiary III-IV composition) to have a unit-cell of about 0.10 nm³, which is 4.65×10⁻¹⁰ m on a side, and we need 100 unit-cells thickness (46 nm) for it to do its PV magic, then 1 ton of Indium without loss goes onto 8,940,000 m² of cells, which generate 14.5 GWh/day at nominal diurnal insolation patterns, packaging losses, yada, yada, yada.

    And that’s for ONE ton of Indium. 14.5 GWh/day is about 65% of what a 1 GWe nuclear power plant would produce, per day. So… it’d take about 8 metric tons of ‘In’ to make enough PV cells to compete with 1 large installation (5 GwE) nuke plant. But that’s “only” 8/1200 or 0.7% of the world’s annual supply of the stuff, according to your stats.

    Seems to me that even a rather rare resource could go into a LOT of PV solar cells, if we decided it was critical. A lot. Of course pulling “100 unit cells thick” is directly from my pipe. Who knows what the actual minimal “active layer” thickness is. 

    Doubtless, a well guarded trade secret. 

    ⋅-⋅-⋅ Just saying, ⋅-⋅-⋅
    ⋅-=≡ GoatGuy ✓ ≡=-⋅

  9. No, Indium is so rare that the current supply won’t certainly suffice (1200 T/year worldwide). If demand increases, prices will skyrocket. You can’t open new mines of indium since there is not such a thing. Indium is usually a byproduct of base metal extraction and is always in very low quantity, regardless of the type of deposit (SEDEX or VMS).

  10. Utility scale solar projects are already being built in the USA with the peak power being 50% higher DC power than the DC->AC inverter can send to the grid (so it is thrown away). That’s because more money is being made by selling power into off-summer and off-noon times when there is excess solar and thus depressed prices and panels are really that cheap now. The idea is they’ll add storage later if/when it makes financial sense.

  11. Space. 

    There’s an application — providing the long term open-space durability of the new stuff is at least “on par” with present space-tech.  

    Multi-story apartment complexes

    … because the roof is a substantially smaller proportion to the number-of-occupants (and by proxy, their total energy use), so as a domestic energy-cost mitigator, the roof area becomes quite limiting. 

    Similar high-intensity-per-available-PV-area-limitations’ applications.  

    Such as ships, maybe. Or upper faciae of skyscrapers. Or wings of perpetual-loiter drones. Or buoys. Or the endless outcropping of roadside powered-signage. Some to tell you how fast you’re going. Others craftily acting as small-cell cell-phone or WiFi towers. 

    Or cut into strips for hyperbolic reflector concentrator-based PV.  

    IF 29% at 1-Sol, very likely rises to 35% at 20-Sols. Excellent for municipal PV production, where economies of scale make having sun-tracking reflective-concentration technology a good economic choice.

    ⋅-⋅-⋅ Just saying, ⋅-⋅-⋅
    ⋅-=≡ GoatGuy ✓ ≡=-⋅

  12. Mmmm… not quite.  

    {engineering} is identical.  
    {permits} is mostly the same.  
    {project management} is almost same.
    {shipping}, if less than 1 truckload, the same.
    {test, certification} the same.
    {power inversion}, the same.
    {power transmission}, the same.


    {maintenance} between ½ and unity
    {fail rate} slightly more than ½
    {dust losses}, ½
    {heat buildup}, ⇐ (1 × (1 – 29%)) / (2 × (1 – 15%))  (wicked!)
    {panel-frames} ½
    {panel-electrical} ½

    So, then it depends a lot on estimating the monthly $/kWh all-in cost, including the build-out, ongoing maintenance, product acquisition and insurance.  

    My spidey-sense-meter is wiggling… I don’t think that kilowatt-hour costs drop by even a small fraction.  Especially if the manufacture’s production-and-investor-return realities cause the $/kW of panels to substantially exceed the future market-rate single crystal Si commodity-market PV from government subsidized foreign PF makers. 

    ⋅-⋅-⋅ Just saying, ⋅-⋅-⋅
    ⋅-=≡ GoatGuy ✓ ≡=-⋅

  13. Yah, I was thinking the inverse … “what’s its $/m² plus-up cost?” … ⋅-=≡ GoatGuy ✓ ≡=-⋅

  14. Another case of ‘that’s the other side’s thing’ therefore we will oppose it at all levels, ignoring the common good that ‘it’ is capable of.

    i have no doubt commercially available panels in the next few years will have extraordinary efficiencies. But instead of halving the number of panels to get the same power, keep the same number you have today and double your power. Why not? Prices will continue to plunge, take advantage of cheap energy.

  15. This technology is deemed to stay in a niche. The indium worldwide production and reserves are so tiny that it makes that whole technology irrelevant.

  16. Will probably take quite a few years at least to hit the market. So don’t expect miracles yet.
    Also, the big cost of solar is in the storage and the installation these days. Everything else is secondary to that. Granted, smaller panels could (somewhat) decrease the installation cost, but not by much. Renewables have an energy storage problem, most of all.

  17. I have no clue what could happen, but I could also say take two known household products like toilet cleaner and clorox and mix them together for a massive boost in cleaning power. Oh yeah, toxic gas could result from these two products. To quote the Goat, “Just sayin”.

  18. Because the technology in question (energy related) clashes with their ideology and politics. They would find blame with it if god himself brought it down.

  19. It’s a previous, working recipe with some aluminum thrown in. Why would you assume it’s suddenly some nefarious concoction of toxicity.

  20. The panels they were pushing just over seven years ago were at 15% or 16%. This effectively would double my output without an increase in the number of panels. Being an early adopter makes you envious of what becomes available in a short time.

Comments are closed.