How many EV batteries can we make?

 kw: analysis, lithium ion batteries, cobalt, resource limits

When I lived in South Dakota I sometimes visited the sites of abandoned mines. A number of these had mined spodumene, a lithium-bearing mineral that is famous in the Black Hills for forming gigantic crystals such as this one in the Plumbago pegmatite. It is 12 feet long and 3-4 feet wide (~4m x 3m). Crystals approaching 50 feet long (~15m) have been reported.

A pegmatite is a special igneous (crystallized from molten rock) deposit that formed in the presence of high-pressure water and cooled slowly, so large crystals were produced. Many gemstone mines are in pegmatites. For example, while there are often ruby or sapphire crystals in granitic rocks, they are usually microscopic. It takes "pegmatite conditions" for large (cm size) gem crystals to form.

Spodumene is the primary source of lithium around the world. Pure spodumene has the chemical formula LiAl(SiO₃)₂. To understand the physical economics of lithium mining, it helps to know just how much lithium the mineral contains. This table will help:

This presumes the pure mineral. Elemental lithium content is less than 4%. Ore grades are reported as the oxide, Li₂O, which comprises 16% of the pure mineral.

Spodumene in bulk is seldom pure. Its gemstone varieties, hiddenite and kunzite, are nearly pure and transparent, with coloration due to trace impurities. The "log" of spodumene shown in the image above is probably no more than 60% pure, containing iron oxide and iron silicate and other impurities. Spodumene ore is considered "good" when its composition of Li₂O is above 8%.

Also, a spodumene pegmatite seldom contains more than 15% of such "good" spodumene. The best spodumene ore in the world, found in Australia, has a Li₂O content of 1.6%. About 1% is more common. See this Wikipedia article for more detail.

What is needed to produce an EV battery? From a number of sources I find that a "typical" EV battery contains 8 kg of lithium, 14 kg of cobalt, and 20 kg of manganese in its cathode, the + electrode. Such a battery as a whole weighs about 250 kg, much of which is the casings (metal shells of the cells). By contrast, the battery in a Tesla Model S contains 62.4 kg of lithium. I haven't discerned whether the proportions of cobalt and manganese are the same, but that is probably so. The packaging is more efficient, so the Model S battery weighs 900 kg (about a ton).

What do the cobalt (Co) and manganese (Mn) do? They are part of the cathode, and increase its energy density. Much experimentation and optimization have resulted in this composition. One restriction is the scarcity of cobalt, seen in figures for known world reserves for these metals:

• Li: 89 billion kg
• Mn: 1,700 billion kg
• Co: 8.3 billion kg (9% compared to Li)

A battery made with Li and Mn but without Co is less efficient, but it appears we'll need to go that way anyway; batteries aren't the only use we have for cobalt.

So, what are the implications of replacing every vehicle with an electric vehicle? In the US there are nearly 280 million registered vehicles; worldwide, about 1.4 billion. Let's assume as an initial estimate that each one needs 8 kg Li, 14 kg Co, and 20 kg Mn. That works out to

• Li: 2.24 billion kg for the US and 11.2 billion kg for the world. So far, that's doable (NOT easy!)
• Mn: 5.6 billion kg for the US and 28 billion kg for the world. Ditto.
• Co: 3.92 billion kg for the US and 19.6 billion kg for the world. OOPS! Almost 2.4 times the known supply.

Using all known cobalt to make EV batteries using the currently-accepted technology, the number would be 590 million. Almost half of that would be snarfed up by the US, if that were allowed.

Extracting the elements produces lots of waste. In the case of Lithium, each kg produced is accompanied by 27.3 kg of waste if the spodumene is pure, and usually by 98-99 kg of waste. Cobalt and Manganese are heavier and their ores are a bit richer, but their production also yields enormous piles of waste material. By the way, seabed mining can't help much. The "manganese nodules" do contain a lot of manganese, but usually a nearly equal amount of iron; they contain less than a percent of cobalt.

All this is big, but it is actually the tail wagging the elephant in the room. Transportation uses about 30% of US energy consumption, and half of that is for private autos. In the world, the figures are 25% and 12%, which means for the non-US countries, private auto transportation is even less of the total, around 10%. Electric generation would have to increase mightily to support another half billion or billion EV's. What could be done to replace petroleum and coal used for electric generation, the other 80-90%?

Burning coal, oil, and natural gas are three of the four stable forms of electric generation. The fourth is nuclear fission. None of these depend on daylight, or wind, or tides. I am disregarding geothermal because its practical, widespread use is two or more generations in the future, at the very best. Hydropower is comparatively stable, but nearly all the hydropower available in the US is already being used, and worldwide, it cannot provide more than a few percent of the need.

The two big technologies that seem able to produce large amounts of electricity are solar and wind. Solar reached energy breakeven 10-20 years ago, and improvements continue. Wind power does not yet generate enough new electricity to replace what is used to manufacture, transport, erect, and maintain the equipment. Wind turbines also kill birds, lots of them, including eagles and numerous endangered species, but the industry gets a pass because of climate-change activism. Both technologies have variable output that depends on the time of day and the local weather. Thus, utility-scale battery storage is required. Tons and tons and millions of tons of batteries.

I could get into a rant here, but I'll stop with this admonition: we need new battery technologies that depend on neither lithium nor cobalt nor any "rare earth" elements (which aren't rare, just hard to separate from their ores). A 10x-100x increase in funding for battery research is definitely worth pursuing.