Net zero, part 1: energy
Natural gas makes switching to renewables a lot easier.
This is part of a series of posts on how to pause climate change using technology and policy:
Net zero, part 1: energy (you are here)
Net zero, part 5: stopping climate change
In the short term, it’s better to pause climate change than revert to a pre-industrial climate. To pause climate change, it’s sufficient to achieve net zero emissions across the economy. Here we’re going to focus on getting to net zero in energy production.
The focus here is electricity and heating. Processes that require chemical energy in the form of fuels will be dealt with in a future post.
Hydro, wind, and solar
Everyone seems to like hydro, it supplies about 6% of US electricity, can be turned on and off at will, and offers some degree of inertia for stabilizing the grid. Most of the good sites for hydroelectric generation are in use, but the DOE thinks it can add 50% more capacity by 2050.
A fully electrified energy system would consume less primary energy but more electricity, diluting the effect of this new capacity. On balance, hydro might make up 5% of our electricity generation in the future.
On to wind, an established technology with a pretty low levelized cost of electricity1. Since wind is intermittent, you need to pair it with other energy technologies. Solar is a natural fit since wind power is relatively uncorrelated with solar output. It tends to rise in the winter, peak in the spring, and fall during summer. It’s also geographically uncorrelated, offering high capacity factors in places with fewer solar resources like the midwest.
Wind currently supplies about 11% of our electricity. I expect that share to grow, particularly as a compliment to growing solar capacity2. Indeed, the 2035 report proposes a 90% clean grid which leans on wind for almost half of electricity generation. I’m going to be more conservative and assume wind can provide a third of our electricity.
Solar is the king of renewable generation. It will soon be the cheapest source of energy anywhere in the world and there are still several avenues to lower costs. Brian Potter uses a simple model and concludes that “[a]t $400 a kw solar and $100 a kwh batteries (costs China is probably achieving right now), we could meet 80% of electricity demand with solar PV for roughly current US average combined cycle gas turbine costs”. Which would be cheaper than our electrical generation today.
But let’s give other energy sources a chance by assuming that solar and batteries stop improving. Getting 50% of our electricity from solar seems feasible in any case.
In this ad hoc scenario, solar, wind and hydro make up almost 90% of electrical generation. Other renewables like offshore wind or tidal energy could potentially close the gap, but I’m less optimistic about their chances of delivering low cost power3.
Batteries
Batteries have seen leaps in deployment in the last few years. Lithium iron phosphate (LFP) batteries use earth abundant materials and come in containerized units. They can provide more stabilization per MW of capacity than a spinning turbine. Their price continues to fall.
We are also quietly adding huge amounts of battery storage to the grid via electric vehicles and plug in hybrids. There is little stopping us from charging our cars at work while solar energy is cheap. With cheap energy during the day, it might even be cost effective to drive grid batteries to nearby places that need energy.
Impulse labs wants to do something similar with home appliances. A battery-powered stove that can boil a pot of water in 40 seconds is a way to sneak storage capacity into your home. Extending this idea to the fridge, dishwasher, laundry, and water heater means a fully electric home that works during a power outage.
Transitioning to a grid dominated by solar and batteries is challenging, but done properly, can be more resilient over all. Households with a lot of storage don’t need the grid to work at all times. The digital inertia produced by batteries can be better than the synchronization we use today.
That said, batteries are still kind of expensive. It’s not clear that LFP batteries can get dramatically cheaper than they are today. There’s two battery technologies that might offer an improvement:
Iron-air batteries release their charge by letting iron turn to rust. Iron and air are some of the cheapest materials you can make a battery out of. Startup Form Energy is targeting batteries with a cost of $20/kWh. Startups never quite seem to reach their ambitions, but even falling short of this target would give us a second revolution in storage.
Lithium-air batteries are a more speculative option. Theoretically they could offer 10x the energy density of traditional lithum-ion batteries, making them competitive with fossil fuels for transport. Whether this will be competitive in grid storage remains to be seen, but Ouros energy is making some big claims.
Thermal storage
The other contender for cheap storage is thermal energy storage. Bricks, hot sand or molten salt can store heat for a week or even months with enough scale. This heat can run a steam turbine, offering energy storage for as low as $30/kWh.
In The Case for Brick Thermal Storage Austin Vernon outlines a plan to heat bricks with cheap renewables when they’re available and re-power a coal turbine when energy is needed. Coal turbines are nice because there are a lot of them, apparently there are ~2.2 TW of coal turbine capacity and the world uses about 2.5 TW electricity total! Coal turbines don’t require the high temperatures of natural gas turbines which improves the economics.
Vernon estimates that if on-site solar can get down to $10/MWh, this scheme could provide flexible energy for $35/MWh. Unfortunately, it’s hard to beat natural gas, with prices lower than $4/MCF (they’re $3.24/MCF right now), gas plants can break even at $50/MWh despite low utilization.
Perhaps more interesting, Vernon explores using thermal storage to meet seasonal energy demand. The economics of seasonal storage are pretty brutal, since your system is only getting used once or twice a year. Underground natural gas storage is cheap enough to serve this niche, is there a clean alternative?
He sketches out a plan using a massive amount of cheap material like crushed granite. With some favorable assumptions (summer excess solar at $5/MWh, 5% interest rate) costs could get down to $53-$63/MWh. For now, it seems like natural gas is a better option.
Thermal seems most compelling as a cheaper battery. Flexible power is a good option to explore, especially as we push to make natural gas clean. In the short term the technical risk probably isn’t worth it for such a small advantage over gas.
Baseload power lowers land use
Why can’t renewables get to 100% of our electricity generation? It’s possible, but expensive. The issue is needing to build enough solar for the worst case scenario. If a cloudy winter day gets 10% of the usual solar energy, then you need 10x as many solar panels to make sure you always have enough power. Most of the year it’s overkill, that’s expensive.
People have suggested that “baseload” power sources like natural gas (distinct from “peaker” plants than only operate occasionally), nuclear, geothermal, and someday fusion4 could provide power on low-renewable days. Baseload power is valuable, but not for this reason.
The issue with these baseload sources is that their upfront (i.e. capital) cost is high. They also work best by running at a consistent rate. That makes them particularly unsuited for dancing with renewables. If your nuclear plant needs the capacity to provide 50% of grid electricity to make up for shortfalls in renewable generation but only supplies 10% of the electricity on average, the power will be ~5x as expensive as it is today. And nuclear is already expensive.
These baseload sources should be seen as a way to scale down the amount of renewables you need. If your town requires 1 GW of electricity, parking a 500 MW nuclear reactor nearby means renewables (and flexible power sources) only need to provide 500 MW. This is a boon for places without much available land or few renewable resources. Nuclear already provides 20% of US electricity, scaling it in places without renewables can make the energy transition a lot easier.
One path to making these sources more flexible is to couple them with thermal energy stores. That way they can produce heat at a consistent rate while dispatching from their thermal stores as needed. Moltex energy for example is coupling nuclear power with molten salt thermal. This is a great idea, I just can’t see this trick offering the kind of flexibility or cost required in high renewables locations5.
Gas + point source capture for flex power
So far I’ve argued:
Renewables + batteries can provide a substantial fraction of our electricity. Though they fall short of supplying 100% of our electricity.
Thermal storage might offer even lower battery costs. Though using it as a basis for truly flexible power or seasonal storage is less compelling.
Baseload power sources like nuclear and geothermal scale down land use. But they are too expensive to use for flexible power.
We’re missing a truly flexible and cheap power source that can counter intermittency and reduce the number of batteries we need.
Natural gas provides 43% of U.S. electricity; a share that’s set to grow as oil and coal are phased out. It’s abundant, cheap, and greener than other fuels. We need a lot of natural gas to transition to a clean grid, but at some point it needs to be net zero too.
The best way to do that is point source capture. This is easier to do than direct air capture because you’re handling flue gas with a higher concentration of CO2. If you site your plant over a suitable rock formation, you can inject CO2 into the ground.
The cost of point source capture is critical. A MWh of natural gas generation produces about 0.4 tons of CO2 and costs $70/MWh6. If carbon capture costs $100/ton, that adds $40 per MWh, a 57% increase7.
Fortunately, if enhanced weathering scales, $100/ton is a worst case scenario. Point source capture should be substantially cheaper, I’ve seen estimates around $50/ton. At that price, clean natural gas is quite reasonable at $90/MWh. These plants would only provide 10-20% of power needs so their higher price relative to renewables isn’t going to add a lot of cost.
NET Power is an opportunity to make gas generation even better. It burns gas in a pure oxygen environment and uses the supercritical CO2 as the working fluid for an efficient power cycle89. The high pressure, pure CO2 stream is perfect for carbon capture systems, potentially lowering the cost of capture.
Scaling point source capture and using existing natural gas infrastructure seems like a safer bet for flexible power than thermal storage approaches. Particularly in high latitude regions that need seasonal storage the most.
Residential heat
Cooling and dehumidifying are already electrified10 and demand for these correlates well with solar energy. But heating demand still relies on fuels and peaks when solar resources are low11.
Fortunately we already know that the future of low-carbon heating is heat pumps. They are so efficient that burning gas in a power plant and sending it to the heat pump uses less gas than burning it in a boiler12. Natural gas plus point source capture can give cold regions the flexible energy the grid needs right when heat pumps have to work the hardest1314.
One drawback with heat pumps is that they are more expensive to install than traditional boilers. While heat pumps will save money over the long term, not everyone can front $10K. I think this is a straightforward case for government subsidy since local grids can save money on gas demand and utilities can depreciate their gas infrastructure.
That being said, gas heating is a great catalyst for electrifying everything else. It makes sense to support traditional systems until everything else is in place. Only then should we switch to heat pumps and point source capture.
Additional innovations like cheaper better insulation, underfloor heating, better architecture, superheating your house, gorilla glass panes, cellulose aerogel, and low E glass can lower peak heating demand15.
Industrial heat
Households only really need electricity, but industrial processes also need thermal and chemical energy. Chemical energy will be addressed in a future post.
For low quality heat, electric boilers and heat pumps are sufficient for temperatures up to about 200°C. For extremely high temperatures (>1500°C), much of the industry already uses electric arc furnaces. As energy costs fall, my hope is that these will replace coal fired boilers in industries like steel production. Perhaps electric arc furnaces could be used for cement production as well.
The real challenge is producing heat of middling temperature. A substantial amount of industrial energy demand is for steam at temperatures up to 400°C. This is not economical for heat pumps or electrical arcs.
Austin Vernon reviews this challenge in The Future of Industrial Process Heat and The Case for Brick Thermal Storage. He rules out things like nuclear16, heat pumps, e-fuels, and hydrogen. That leaves two options for clean industrial heat:
Thermal storage, which we covered before, can be pretty cheap if renewables are cheap. He estimates that with $10/MWh solar it would be price competitive everywhere at about $4/MCF.
Indeed, Austin is jumping on this idea with his company Standard Thermal using solar panels to store heat in the ground directly! This might unlock even lower costs than he mentioned in his post.
Enhanced geothermal power is unproven but could potentially offer 24/7 steam at a competitive price. It has the advantage of more location flexibility, simpler equipment, no intermittency, and less land use.
I want to add one other possibility to this story. You’re probably tired of hearing that natural gas with point source capture is a good option. But at today’s price of $3.24/MCF it would cost $3.62-$4.01/MCF with capture costs of $50-$100/ton17. That’s competitive with thermal brick storage using cheap solar.
(EDIT: companies like Via Separations might lower industrial heat demand by replacing some thermal separation processes using graphene oxide membranes.)
(EDIT see also: Some like it hot: Moving industrial electrification from potential to practice)
Conclusion
With renewables and batteries, our grid can reach net zero while being slightly cheaper. Natural gas will get more expensive with point source capture, but the net effect with renewables is lower energy costs.
The one exception is industrial heat. All the options so far are more expensive than natural gas. Only Standard Thermal looks to beat it. Let’s hope they do.
It’s remarkable that natural gas plus point source capture can be a competitive alternative for flexible power and seasonal storage and industrial heating. I hope fracking technology provides cheap natural gas across the world, it will make the green transition so much easier.
With the exception of (eventually) mandating carbon capture for natural gas plants, green electricity doesn’t require regulations or a carbon tax. Supporting the research and deployment of today’s tech is sufficient.
I think wind could get a little cheaper by using direct drive turbines that generate DC electricity directly. But unlike solar, there’s not much room for further cost declines.
It seems that wind and solar can share the same land if the wind turbines are placed so that they don’t shade to panels.
Though these sources are great for land constrained countries with low solar resources like the UK.
I am bearish on fusion providing cost competitive electricity. But that’s a topic for another time. Hopefully I’m wrong.
Two things that could change this: 1. a dramatic simplification of the nuclear thermal process (perhaps by stirring the fuel rods through a salt bath rather than pumping salt through the rods?) and 2. using lightcell’s technology to efficiently convert the salt thermal energy into monochromatic light to collect with specialized cells. This offers both higher efficiency and lower cost.
I’m using Brian Potter’s number from the last chart of this post. Austin Vernon’s numbers from the thermal storage would suggest it’s closer to $40-$50/MWh.
Peaker plants might be a better comparison. They are more expensive so the relative price increase from point source capture is smaller.
In theory, the efficiency remains pretty high even for intermittent loads. Compare the 55% theoretical efficiency (in a single cycle!) to combined cycle gas plants that get 60% efficiency over two cycles.
I wonder if adapting something like Astro Mechanica’s LNG cooled efficient electric generator could squeeze out a little more energy (95%→99%). Starting from liquid methane should increase the expansion ratio too.
There are some low tech solutions to lowering cooling demand like letting in nighttime air, using evaporative cooling in dry areas, and supercooling your house before demand spikes.
Though wind works fine in the winter.
Note that transmitting natural gas is 10x cheaper via pipeline than via transmission lines, see here. The short distance and savings on direct air capture mean transmission wins out.
Since heat pump efficiency falls slightly in cold weather, I wonder if connecting them to an indoor water line or using some of the hot water tank would prevent the efficiency from dropping so much. The hot water tank can be used as a thermal battery as well. You could essentially get district heating or cooling by shifting the cold water line up or down in temperature. See also: Ice-source heat pumps.
Note that Austin Vernon takes a contrary point, suggesting that we should stick with fuels to address winter heating demand and switch to green fuels instead. I think doing point source capture will necessarily be cheaper than making e-fuels.
See point 2 here on how polar amplification is warming cold places more than warm places. That means that heating demand might fall on its own. EDIT: increasing urbanization and the urban heat island effect (which mostly increases evening and nighttime temperatures) may also lower demand.
Though he notes that small modular reactors could improve the economics.
I can’t get the numbers for pyrolysis transported hydrogen to beat point source captured natural gas. In a future post I’ll explain why can be slightly competitive for delivering hydrogen for chemical use.