US startup begins producing 40%-efficient thermophotovoltaic cells
pv-magazine.com
pv-magazine.com
GE's combined cycle turbines can get system level efficiency of around 63% from these sorts of temperatures.
(For those not familiar with them: They're basically aircraft jet engines followed by steam turbines using the hot exhaust. They are in widespread use to generate electricity from gas, but they can also run off any other liquid fuel, or simply off anything that gets very hot.)
By using the residual heat for hot water and for heating in winter and cooling in summer, the global efficiency typically becomes well above 80%.
Even the best Diesel generators may reach around 55% efficiency, while working at much lower maximum temperatures.
For a few seconds I thought that this thermophotovoltaic technique is great, until I have seen that the emitter must have a temperature above 2000 Celsius degrees. For such a great temperature it is very easy to make heat engines with much better efficiency and which might even be less expensive, because these multijunction III-V photovoltaic cells are many times more expensive than normal solar panels.
If these were used for solar power, the concentrators would also be very expensive. Already the concentrators that produce temperatures around 1000 Celsius degrees, which is more than enough for easily reaching 40% efficiency with closed-cycle heat engines, are much more expensive than the concentrators that reach only lower temperatures, like 650 Celsius degrees, which would still be good enough for a steam turbine or for a closed-cycle supercritical carbon dioxide engine.
The only real advantage is that these should need less maintenance than turbo-generators, which may be essential outside Earth or in remote locations, but less important than cost and efficiency otherwise.
Thermal storage at low temperatures is cheap and easy with molten salts.
Thermal storage at over 2000 Celsius degrees will be extremely difficult, due to the difficulty of preventing heat losses.
The best would be for the hot body to be stored in argon, because in vacuum it would evaporate and heavier inert gases are expensive. The storage vessel would be very expensive in any case, being made from multiple layers with high temperature resistance, an external surface with high reflectance in red and infrared and other layers with low thermal conductivity, so it is hard to imagine that it could have a size large enough to store much energy.
Another obstacle is that the available power is determined by the emitting surface, not by the volume of the hot body, which is another obstacle for scaling to large amounts of stored energy.
Another obstacle to scaling is that when the hot body cools down the conversion efficiency drops extremely quickly (fourth power), which means that it could store energy only e.g. by being heated and cooled between 2400 and 2100 Celsius degrees.
So only a very small fraction of the thermal capacitance of the hot body can be used, many times lower than when the heat stored in that body would be used to power a closed-cycle heat engine.
So no, these devices may have some useful applications, but energy storage is certainly not one of them, because they are much worse than almost any alternative. Even storing compressed air in a pressure vessel is much more practical.
However, there is still sleight of hand. The efficiency quoted is when the equipment is new and clean. Fouling and wear both take single digit percentages off.
Also, they use the lower-heating-value for the gas energy supply. That, in my view, is dishonest - the correct energy measure for gas is the higher-heating-value, which is 10.7% more. The difference comes from how you account for the heat in the steam produced by burning gas. In my view, the energy from that steam should be accounted for when considering efficiency - in GE's view, it shouldn't.
The lower heating value corresponds to the heat actually produced during burning.
The higher value is based on the fact that the exhaust gases contain water as a gas, and if that water were condensed into liquid water, an additional quantity of heat would be produced, the latent heat. Due to the low temperature of condensation, that latent heat cannot be easily recovered, except if it is used for hot water production or for home heating or for home cooling (e.g. with absorption chillers).
If instead of using combustion gases, the same heat engine would use heat from an external source, like also the thermophotovoltaic devices discussed here, the heat engine would function in the same way and with the same efficiency, when receiving the amount of heat corresponding to the lower heating value (assuming no heat losses during heat transfer).
Sorry for being nitpicky, but burning natural gas does not remove carbon, it adds new CO2 to the atmosphere (if you release the burn products which is usually done).
You probably mean that if you released that natural gas directly into the atmosphere instead, it would have a larger greenhouse effect than the CO2 released by the burning process. This is true, and one should absolutely choose "burn it" when given the choice of releasing only vs burning and then releasing.
But expressing it in the way of carbon removal is misleading. The number of carbon atoms in the atmosphere is the same in both scenarios, they are just bound in a less greenhouse-y form (also, natural gas decays into CO2 plus water eventually, but it's a very slow process). The number of CO2 equivalents goes up in both scenarios as well, just way less if you burn it before. Maybe some people refer to those through "carbon", idk.
It’s also possible to capture the CO2 rather than releasing it into the atmosphere.
As always to economics, if you can't measure it well, it doesn't exist!
For base load generation at least, these run continuously under “optimal conditions”. The big ones are designed to be switched on and run for decades.
A turbine is a tuned system (its shape is designed for its operating conditions). For base load generation it’s important to fix those operating conditions thus getting the most output from the least input.
(This is quite different for a peaker plant that needs to spin up and down relatively wuickly in response to demand, much less, say, the turbine on an aircraft or locomotive, which go up and down depending on load. They can never be anywhere close to theoretical efficiency).
Just a long winded way of saying that “optimal load” is not as uncommon as one might think. Its no spherical cow.
Being able to replace the combined turbine/alternator assembly with a 'when it gets hot, voltage comes out' unit would give you significant reliability gains and lower operating costs.
I feel like these sorts of things should be useful in situations where there’s a lot of excess heat in some other process that’s typically wasted. Smelters, incinerators, high temperature chemical reactions, etc. Because they are presumably not large you can clad and enclose the high temperature area in these sorts of panels and capture 40% of the wasted energy and divert it back into the process. That would have compounding effects.
At those temperatures, even the best alloys lose a large percentage of their strength. And also, steam this hot is incredibly corrosive.
There is no need for blades to withstand the combustion temperature, because film cooling can keep blades far cooler in high speed laminar gas flow. The challenge is that the operating gas must be dust-free or a spec of dust on the blade surface disrupts the film and causes failure.
Obviously the amount of energy wasted to pumping the film cooling gasses goes up the higher the combustion temperature is, so I assume there is still some upper limit on combustion temperature.
As far as I know, combustion isn’t required for NOx formation. If you want to run hotter without producing NOx, you need to eliminate either nitrogen or oxygen. IIRC there are a couple of proposed designs for doing this: combustion in nitrogen-depleted air and chemical looping combustion.
Am I misunderstanding this or are both of those concepts basically "just inject pure oxygen"? And I'm no economist, but running a powerplant on rocket fuel doesn't sound that feasible to me.
Urea injection into the exhaust gas sounds more sensible to me, and even that is seems economically questionable. Fuel prices would need to be pretty high that the efficiency gains make buying urea (instead of just more fuel and running the combustion colder) worth the investment...
Now what? It doesn't burn, since there is no oxidizer. The only way is to inject some. And since you don't want nitrogen, you can't inject air. So you inject pure oxygen, or something like peroxide?
Sounds expensive.
That wouldn't impact blade temp.
If the latter, it would be a waste of time, if the former then you just want heat, and the fan is there to provide enough oxygen to burn the fuel so why would it impact efficiency?
Edit: I was thinking literal jet engines that have indeed been used as generators, not combined cycle generators.
I believe there were two sets in that plant, and boy, did it get loud.
http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/carnot.htm...
---
An internal combustion engine (ICE) uses the Otto cycle:
https://web.mit.edu/16.unified/www/FALL/thermodynamics/notes...
https://www.sciencedirect.com/topics/engineering/otto-cycle
If you click the Read more arrow under 3.4.1 The Efficiency of an Otto Engine, it states that The ideal Otto cycle achieves the Carnot efficiency of an engine working between the maximum, pre-combustion, temperature and the intake temperature. This means that the ideal Otto cycle cannot achieve the Carnot efficiency determined by the highest and lowest temperature during the cycle.
Tlow = T ambient
Thigh = T at highest compression of piston
efficiency ~= 1 - Tlow/Thigh < (Thigh - Tlow)/Thigh = Carnot efficiency
We can find the pre-ignition temperature at maximum compression:
https://www.physicsforums.com/threads/compression-psi-and-te...
T2 = T1 * ((V1/V2)^(y-1)) where y ~= 1.4 for air
So a 14:1 compression ratio at an ambient room temperature of 293 K (20 C or 68 F) gives a pre-combustions temperature of:
T2 = 293 * (14^(1.4-1)) = 842 K (569 C or 1056 F)
So the maximum efficiency (Carnot efficiency) of an Otto cycle 14:1 compression ICE would be less than:
efficiency < (842-293)/842 < 65%
In practice, an ICE might scavange 50% of that due to losses to entropy, friction and hot exhaust at 600 K (about 300 C or 600 F) and end up at 33% efficiency, not counting drivetrain losses of about 15% to get to maybe 28% at the wheels.
That's why ICE vehicles waste around 75% of the fuel's energy or more. Whereas an electric vehicle will be around 90% percent efficient from batteries to motor and around 75% efficient at the wheels, or at least 3 times better than an ICE vehicle.
---
A gas turbine uses the Brayton cycle:
https://web.mit.edu/16.unified/www/SPRING/propulsion/notes/n...
The Brayton cycle thermal efficiency contains the ratio of the compressor exit temperature to atmospheric temperature, so that the ratio is not based on the highest temperature in the cycle, as the Carnot efficiency is. For a given maximum cycle temperature, the Brayton cycle is therefore less efficient than a Carnot cycle.
Tlow = T ambient
Thigh = T at highest compression after compressor
effiency ~= 1 - Tlow/Thigh < (Thigh - Tlow)/Thigh = Carnot efficiency
A jet engine might reach 40:1 compression:
T2 = 293 * (40^(1.4-1)) = 1281 K (1008 C or 1846 F)
So the maximum efficiency (Carnot efficiency) of a Brayton cycle 40:1 compression turbine would be less than:
efficiency < (1281-293)/1281 < 77%
In practice, a gas turbine might scavange up to 85% of that and end up at 65% efficiency, not counting generator losses of 5%. But 40-65% overall efficiency is more realistic.
A 45% efficient gas turbine would leave 55% of the energy as waste heat in the exhaust. So a steam turbine scavanging that would only need to be about 35% efficient to reach an overall efficiency of 65%. Depending on exhaust temperature, the article's thermophotovoltaic cells would probably be cheaper and more reliable than additional turbine stages.
---
I've never seen a good way to relate Carnot efficiency to quantum efficiency. Thermodynamics measures average emergent behavior like fluid dynamics. So the Carnot efficiency is kind of like the Bernoulli equation or Reynolds number, and may have no analog at microscopic scales. Maybe something like:
https://en.wikipedia.org/wiki/Quantum_heat_engines_and_refri...
This graphic of photovoltaic cell efficiency (non-thermo) is super interesting: it shows the progression from 1976 and current capabilities (and not necessarily commercially viable or available). The panels you'll get for your home are probably around 20% efficient in ideal conditions.
I wonder if you could use this with parabolic mirrors, though. Build a large mirror array, focus sunlight onto a big carbon sphere (maybe coat it with one of those new materials that are transparent for visible light but pretty reflective for IR), cover the top of the sphere in those new panels. They are more efficient than practically all solar cells and get much more power out the same area than solar cells. This should beat a photovoltaic parabolic mirror setup, right?
If you try that with solar radiation, you'll lose most of it back into the sky.
https://www.nature.com/articles/s41586-022-04473-y.pdf
(Equation 1 and surrounding discussion, and the energy-flow diagrams in Figure 1).
My idea was to now enclose this sphere in TPV modules - except in places where the parabolic mirror puts sunlight on it Let's say that's half of the sphere (but you could cut this number down if you chose a parabolic mirror with a long focal point).
Now you lose IR photons in across the part of the surface that you didn't cover in TPV modules. But you could but a IR mirror there, that is transparent for most of sunlight.
Edit: Apparently, yes, current PV solar may also do this:
> We found temperatures over a PV plant were regularly 3–4 °C warmer than wildlands at night
https://www.nature.com/articles/srep35070
But I suspect/conjecture that turning sunlight directly into heat, then turning some of that heat into electricity, would capture even more heat from sunlight.
Because some of the heat in the sphere will re-radiate, but much of it will inevitably be lost to convection and conduction.
Arguably there should be a "white roof" campaign for urban areas, it would be a cheap way of reducing the urban heat island effect.
[https://www.homedepot.com/p/Henry-887-Tropi-Cool-100-Silicon...] [https://maps.app.goo.gl/TwrV57Jt2fGo2obw9?g_st=ic]
With grid scale, above certain dimensions, you can store energy for months while maintaining economic viability.
Even for single days or weeks it makes sense. You need hot water in your home 24/7, but sun doesn't shine every day in most regions.
The same geometry that allows heat to be effectively trapped also prevents its fast extraction as radiation. A fluid that can relatively easily change its volume to surface ratio could have an advantage here, too.
[1]: https://en.wikipedia.org/wiki/Thermal_energy_storage#Molten_...
Flowing sand has issues with blockages and erosion.
Stationary sand has pretty low conductivity, so getting all the energy out of your 200 yard cube of hot sand might be a challenge.
You can progressively sink them in as the sand cools/ you want more energy out.
There's been talk of doing solar concentration -> hot object -> thermophotovoltaic converter. IIRC concentrated solar already exceeds 40% efficiency so it doesn't make sense to add the extra step, unless you are using the solar concentration to "recharge" a heat storage system.
I buy the argument that thermophotovoltaics can become cheaper on a $/kW basis than comparably-efficient fluid/mechanical heat engines. The power per unit area is intrinsically orders of magnitude higher than for direct solar, so even if these cells are pricier than regular solar cells, they have a fighting chance. Also, both the power density and efficiency increase with temperature, and in principle the operating temperature can be higher than that of a turbine (since the materials don't have to simultaneously withstand crazy mechanical stresses and reactive chemical environment).
[0] https://pubs.rsc.org/en/content/articlelanding/2019/EE/C8EE0...
The fact we don't seem to be able to do this yet suggests I'm missing something, probably several things.
The problem is that a tungsten tank filled with molten salt is heavy. If you want to store a couple of MW/h, we're quickly talking about several thousand tons. You have to suspend it somehow inside that thermos. With materials, that don't conduct heat well, but still have some strength at 2000C.
Batteries are faster to charge/discharge and more efficient per wh.
Both usually have less loss over time - it’s very difficult/expensive to avoid significant heat loss over time due to radiative heat loss, if nothing else.
Most batteries or pumped storage have much lower losses over time.
https://en.m.wikipedia.org/wiki/Radioisotope_thermoelectric_...
There’s a story about some loggers finding an abandoned RTG in Siberia and sleeping next to it for warmth. They woke up with severe radiation burns.
One of the proposed ideas for nuclear rocket propulsion would contain a fissioning gas cloud inside a fused-silica glass blub. At a temperature of several thousand degrees (5,000 - 20,000 K), incandescent light escapes the glass, and gets absorbed by the hydrogen propellant (with bits of solid dust flowing through it as an opacifier). That's one-half of your idea: there's no photovoltaic component, but it does examine the "nuclear fission reactor as an optical light source" half.
https://en.wikipedia.org/wiki/Nuclear_lightbulb
https://sci-hub.se/10.2514/6.1991-3512 (pdf) ("Summary of nuclear light bulb development status" (1991))
excerpt:
- "The gas core Nuclear Light Bulb (NLB) propulsion system could provide both the desired thrust and specific impulse. Initial gas core nuclear rocket (GCNR) investigations started in the 1950s when two somewhat different concepts emerged; an open-cycle GCNR and a closed-cycle GCNR (named the Nuclear Light Bulb). The open cycle configuration (Fig. 1) isolates hot fissioning gas from the chamber walls by flowing cooler propellant around the hot gas. Although some nuclear fuel will diffuse into the propellant and exhaust into space, the theoretical amount was considered too small to be of consequence. NASA-Lewis pursued the open cycle by both analysis and laboratory experiments from the ’50s through the early ‘70s.’"
- "At the same time, United Technologies Research Center, then United Aircraft Research Laboratory, explored the closed cycle NLB concept (Fig. 2).[2,3] Hot fissioning fuel in a gaseous state is confined within a transparent wall of fused silica by a vortex of tangentially injected buffer gas that is also transparent to thermal radiation. Hydrogen propellant, seeded with micron-size tungsten particles to absorb the thermal radiation from the hot fuel, flows axially outside the fused silica wall and is exhausted through thrust nozzles. Complete fuel containment requires continuous withdrawal of a small fraction of the mixture of fuel and buffer gas for reprocessing and reinjection into the vortex flow."
Edit: BTW, there are radioisotope thermoelectric generators used for space applications primarily, but they are not true nuclear reactors - they produce short-range radiation that doesn't require much shielding. Nuclear reactors produce neutrons and gammas that require thick shielding.
[0] https://en.wikipedia.org/wiki/Project_Rover#Kiwi_A
[1] https://en.wikipedia.org/wiki/High-temperature_gas_reactor
And its spaceflight cousin: https://www.projectrho.com/public_html/rocket/enginelist2.ph...
Though they are not peltier devices either.
I understand convenience, laziness and inertia (resisting change) but I also think changing the times when we click "power on" is a simpler solution than mining millions of tons of more lithium, no?
Smart demand-response may yet become a thing, but it's not yet a commodity product. You need a system to send out "turn off" notifications, and a system for measuring that in realtime, and a system for paying people. Some grids _do_ have this, but only for very large consumers.
All power hungry devices should at least have that capability (maybe other than the kettle, lol). This is literally a cost of $20 hardware in many cases.
You would have to completely rebuild most of the manufacturing industry, as many plants have startup times measured in hours or days.
It'd be cheaper to just build nuclear.
So yes, we could survive without power at night. We just have to rebuild every building.
As a long-term ideal I don't disagree with you. We should be building for resilience. But that's not a solution to climate change.
I like to heat my home during winter. We have a (modern, highly efficient) heat pump, so we need most electricity during January, just when the least amount of solar insolation is available [1] and when it sometimes stays cloudy and below 0°C continuously for days. But I guess we'll just have to be more flexible and turn off heating, light, and electricity in general for a week, no big deal.
[1] I wonder if there's a causal relation between cold weather and low solar insolation?
That's already the case – the actual heat pump only runs intermittently, on demand. This happens quasi-randomly, so you automatically get some load balancing across a city.
The problem is that this is intra-day load balancing, which doesn't help one bit if there are several days of low supply (windless winter days).
It also isn't synced to supply, instead most people have it set to different temperatures based on when they are home. It would be better to cool or heat the house based on supply. You want the house between 21 and 24c, you don't care when the system is on.
Including other comments on this exact thread where people did exactly that.
When there's no sun, then other countries probably have sun. When there's no sun, there's hydro. When there's no hydro there's wind. When there's no wind, there's tidal. When there's no tidal there's geothermal. And when there's none of those, there's stored energy in batteries.
And if for some reason all of those combined can't satisfy the demand, then there's nuclear as a last resort.
And if there's none of any of those at all, then we've probably got bigger problems anyway
Wind suffers the same problem as the sun in that sometimes there is no wind anyplace close, and it is even less predictable. While tide power is predictable and consistent, I don't live anywhere close to the ocean so we still need 2000km of wire to get it to me. Geothermo is useful when lava is close to the surface, but I don't live in such an area.
Note that my power is 80% wind, but it was done via several decades of building wind turbines, and I live in Des Moines which nobody would call a big city, if you live in a big city you have even more work to build it (you are both way behind us in building wind, and have a lot more to build)
Sure you don't have to have a single wire from one side of the planet to the other, but that's the point in a power grid. Even connecting two solar farms to the grid a time zone apart provides an extra hour of power to each of those time zones and reduces the risk of a single cloud bringing the country down
And as for the rest, you know what the electricity grid is right? Cables already run around the country, and if anything this reduces the strain on single connection as there are lots of smaller generators dotted around the country rather than a huge point of failure connected to a single power plant
On the contrary, Europe, Africa, Asia, the Americas are all wide enough that they can do plenty of solar trading during the daytime. Of course that does not provide energy in the night and that's OK.
Lack of power either way certainly won’t help!
If by design nuclear is offline, fossil fuels are offline, and battery capacity (or other storage) is insufficient for known (but infrequent) weather events, then we’re designing an inevitable and destructive catastrophe as we have no Plan B for when realtime production is insufficient.
And that will happen some day, regardless of how much capacity we build. It’s currently the norm on many days.
One that will kill a lot of people, especially the physically weaker ones, and be very destructive economically.
One that will also play out randomly based on weather, and for which we’ll have no real Plan B.
Being flippant about it seems rather macabre.
I can imagine some sort of orbital parabolic mirror setup that keeps a few acres at 2000C in the dead of winter and vaporizes any birds that fly over the power plant site.
We're nowhere near the point where that would make sense, but it could let us harness more than the earth's surface area of sunlight for energy production.
It will probably have to wait until after we plunge ourselves into some a perpetual winter trying to mitigate climate change. Of course, a ring world or dyson sphere would make it obsolete.
They're much more efficient, I assume much more expensive?
There's a fair amount of niche applications where peltiers are currently used even though they aren't very good.
But no one's mentioning them as an alternative to peltiers.
They launched a pilot in the summer.
For an energy storage system, though, 40% makes little sense.
Sometimes you can sacrifice some efficiency by warming the coolant to e.g. 75°C and distributing that as district heating.
TLDR: Adding more cells won’t help due to economics, geometry, and fundamental laws of physics.
A source of free energy, a "perpetual engine", is indeed impossible.
Gravitational drag, maybe, but only if there is something to drag nearby, and we postulate a vacuum.
A sphere is symmetric and thus does not emit gravitational waves.
I wonder if an effect akin to black hole evaporation could play a role: if one of the two virtual particles gets accelerated towards the sphere, it may bring some momentum to it. But, assuming that the space is isotropic, statistically such momentums should cancel out.
Also, I’m not so sure that black body radiation wouldn’t have an impact on rotation. At some point, the object will have radiated all its heat and have no energy left. If the particles have no energy, how is the sphere rotating which implies kinetic energy? That implies that even when radiating uniformly, black body radiation must take away angular velocity, no?
In other words - does it remove heat from something at high temperature?
I know there are situations where you can't get rid of heat - would this help by removing it electrically?
My understanding is that to get those kinds of temperatures from geothermal you need to drill to currently infeasible depths.
But you don’t need those kinds of temperatures from a geothermal resource to make it very cost-competitive.
Probably fusion power will not be cheaper than renewables inside of 50 years, because fusion power plants will simply be very expensive.
In the next 20 years we need to decarbonize as much as possible. Fusion sadly won’t have much of an impact for that.
But in 30 years when todays new renewables are at the end of their service life, we have an opportunity to replace them with fusion. That said, renewables will be that much cheaper in 30 years. I think for a while fusion will make the most sense for large industrial manufacturing operations that necessarily require large constant amounts of power.
Gives a whole new meaning to "Sun Microsystems"
Given most demand is during the day and early evening solar is a good complement, but the more mixed renewals you have in your grid the better it will tolerate shocks in supply and demand.
"referring to thermal energy grid storage (TEGS) consisting of a low-cost, grid-scale energy storage technology that uses TPVs to convert heat to electricity above 2,000 C"
You all speak in miracles here, the use case seems to be converting thermal energy and energy storage. Why the moon, and what does that have to do with regular photovoltaic efficiency?
Presumably to produce energy at night and avoid the need for storage. Seems like a moonshot, though.
All around, confusing. I didn't even know we had such a thing.
It's a total waste of time to use moonlight. It is a million times dimmer than sunlight.
This excludes any possibility for generating reasonable amounts of energy from moon light. There is a reason why it is much colder during the night.
At those temperatures, this is not very impressive.
it's pretty much saying "efficient (for a photovoltaic) electricity from powerful infrared"
Sun is 6000 degrees, so most of its emergy is in different frequency
Instead, this panel is planned to be used in thermal batteries, I think the article says.
I guess main competition is steam turbine or smth?
Go figure.
[1]: https://en.wikipedia.org/wiki/Shockley%E2%80%93Queisser_limi...
[2]: https://en.wikipedia.org/wiki/Solar-cell_efficiency#Factors_...
vs
"a higher performance than conventional solar cells, and produce 100 times more power than similarly sized devices”
You have some numbers against that to understand the buzz? Because it sounds pretty impressive? Is it comparing the wrong technology, or are the numbers off for prior performance (in this use case) or how?
The performance is not impressive in comparison with the existing alternative for converting heat into electricity, i.e. with electric generators powered by closed-cycle gas turbines or by steam turbines, which can reach a higher efficiency already at much lower temperatures and without using so expensive materials.
Therefore it is not clear whether this direction of research is worth pursuing.
Moreover, triple-junction (very expensive) solar cells with an efficiency of around 45% for the direct conversion of solar light, without passing through heat, have already been demonstrated some time ago. Due to their high cost, they must also be used with solar concentrators, but the concentrators can be much cheaper than those needed to heat something over 2000 Celsius degrees (which requires very precise focusing).