While the amount of solar energy available is far greater than world marketed energy consumption, by about four orders of magnitude, solar energy mostly only arrives during the day, and sometimes even during the day most of it is blocked by clouds, while existing marketed energy is generally available 24 hours per day. The impending transition to a photovoltaic-solar-energy-based economy therefore has many of the humans worried about the “intermittency problem” or the “storage problem”.
Probably the most important result of the intermittency problem is that the humans will use energy intermittently.
In some cases this is relatively straightforward and already practiced — for example, it’s common for large air-conditioning systems on commercial properties to run their chillers during the night, when electrical energy is cheap, to store up ice, and then use the melting ice during the day to chill a coolant which is circulated to chill air, and electric car owners often have a lot of latitude about when they charge up.
In other cases, it’s a matter of moving demand geographically — perhaps a Google Search query needs to be done in a data center close to the requester, but a large finite-element simulation can be done in whichever data center the computation is cheapest. And in some cases long-distance power transmission can bridge a storm cloud.
In other cases, idling machinery at night is expensive because of the mere depreciation of the machinery, and it might be worth paying higher energy costs.
In still other cases, like life-support machinery and alarm systems, it’s worth paying even extremely high energy costs to guarantee uninterrupted power.
Right now, the popular approach to utility-scale energy storage to reduce solar intermittency is lithium-ion batteries, similar to the Tesla PowerWall or Jehu Garcia’s DIY powerwall projects, but larger. (There are a dozen other possibilities — the Sisyphean train, pumped-water storage, flywheels, compressed-air caverns, vanadium flow batteries, hydrogen or methane fuel cells, molten-metal batteries, and so on — but lithium is the currently popular one.) You might reasonably wonder: is there enough lithium?
(Updated from my comments on the orange website.)
In 2018, the USGS estimated 16 million tonnes of “worldwide identified reserves” of lithium (1.6 × 10¹⁰ kg), but about 53 million tonnes of “resources” of lithium (5.3 × 10¹⁰ kg), which are extractable with current techniques but not necessarily economic at current prices. There’s also 2.3 × 10¹⁴ kg of lithium dissolved in seawater, and the estimates of lithium abundance in the earth’s crust have a low end of 20 ppm. The crust averages about 40 km thick and 3 g/cc in the 29% of the Earth that has continents, which works out to 1.5 × 10¹⁴ m², 5.9 × 10¹⁸ m³, 1.8 × 10²² kg of rock, and 3.6 × 10¹⁷ kg of lithium.
World lithium “production” (which is to say, extraction) is about 4.3 × 10⁷ kg per year, which would take 350 years to exhaust the “worldwide identified reserves”. It seems like a safe bet that some new mining technologies will become available before even 2100, let alone 2369.
Lithium is not destroyed when it is used — you can recycle worn-out batteries into new batteries — so, like gold, we should expect that eventually the amount circulating is much greater than the amount mined in any given year or decade. But, how much would need to be circulating for the humans to start using lithium batteries to power entire continents overnight? Maybe extraction would have to speed up?
Wikipedia’s table of energy densities says Li-ion batteries contain 0.36–0.88 MJ/kg, or slightly higher if you only count the mass of the lithium rather than the entire battery. Conservatively taking 0.36 MJ per kg of lithium (which assumes battery technology doesn't improve — almost-nonrechargeable lithium-metal batteries get five times that much energy per kg of lithium), the 1.6 × 10¹⁰ kg of “identified reserves” would hold 5.8 petajoules; the 2.3 × 10¹⁴ kg of lithium in seawater would hold 83 exajoules; and the 3.6 × 10¹⁷ kg of lithium in the continental crust would hold 130 zettajoules.
Current world marketed energy consumption is on the order of 5.7 × 10²⁰ joules per year, which is 18 terawatts. Incident solar power on the Earth (the Kardashev Type 1 benchmark) is 130 petawatts. So the 5.8 PJ of “identified reserves” is only five minutes of world marketed energy consumption, but the 83 EJ of seawater lithium is about 1.7 months of world marketed energy consumption. Even so, that's only 10 minutes of total terrestrial insolation. But the 130 zettajoules of continental crustal lithium is 12 days’ worth of total terrestrial insolation.
So it seems likely that known concentrated lithium deposits will not be sufficient to permit the transition to solar over the next decade or two, but there is plenty of lithium in seawater and other, less-concentrated deposits to permit such a transition. New extraction technologies will be needed if lithium batteries are to bridge the intermittency gap. Alternatively, some of the other utility-scale storage technologies might be developed.
| Lithium | Energy storage | World marketed energy consumption | Terrestrial insolation | |
|---|---|---|---|---|
| Identified reserves | 1.6 × 10¹⁰ kg | 5.8 PJ (1.6 million MWh) | 5 minutes | 45 milliseconds |
| “Resources” | 5.3 × 10¹⁰ kg | 19 PJ (5.3 million MWh) | 18 minutes | 150 milliseconds |
| Seawater | 2.3 × 10¹⁴ kg | 83 EJ (23 billion MWh) | 53 days | 11 minutes |
| Crust | 3.6 × 10¹⁷ kg | 130 ZJ (36 trillion MWh) | 230 years | 11 days |