Your new EV could do with a shot of caffeine and AI

By Tim Culpan

Scientists and investors have poured billions of dollars into researching alternative technologies in a bid to eliminate combustion engines and reduce greenhouse gases. Storing this energy remains the single biggest cost of making an electric vehicle, necessitating unusual approaches to materials development.

Standard lithium-ion batteries currently dominate because they use a stable, well-known technology with decades of history. But these power packs still fall short, with range anxiety and price topping the list of reasons consumers are resistant to buying EVs. Hydrogen fuel cells are among the alternatives being investigated, but they too have drawbacks, including a lack of hydrogen refueling infrastructure.

In one of many novel discoveries uncovered by scientists in the alternative-energy field, researchers in Japan have found a way to boost efficiency and trim costs. Their secret is caffeine.

A fuelcell contains three parts: an anode, where the current starts, an electrolyte through which it flows, and a platinum cathode that connects to an external device to create electricity. It’s at this third stage that hydrogen and oxygen combine to create water — in the process of producing electricity — which is a harmless byproduct compared to the poisonous fumes that billow out the back of most cars. The problem is that over time, the cathode gets encrusted in layers of platinum hydroxide, reducing the efficiency of the electricity-generating process.

But Masashi Nakamura, Ryuta Kubo, and Rui Suzuki from Chiba University found that adding a dash of caffeine to the electrolyte limits formation of platinum hydroxide. The result: an increase in the fuel-cell reaction by as much as 11-fold. Such a boost also means that less platinum is needed in the first place, further reducing cost.

A breakthrough like this won’t remove all the roadblocks to expanding EV production and adoption. One of the key concerns, price, has been ameliorated by economies of scale that have made the cost of making batteries much cheaper. Yet a lack of charging infrastructure and insufficient driving range remain hurdles to purchase, even at ever-lower vehicle prices.

That’s why industry and the scientific community are eager to throw new ideas around to see what sticks. Take cobalt, another crucial element in lithium-ion power cells. Environmental degradation in the mining process, and massive price swings, have industry hoping an alternative may come along. Other compounds including manganese, iron and nickel, are being explored and may end up becoming viable. Researchers have also developed sodium as a cheap replacement for lithium, but it can store less energy per kilogram so its uptake won’t add distance to most cars.

And while human scientists toil away in laboratories, artificial intelligence could be the next link in finding new breakthroughs. Combining AI and high-performance computers, a team at Microsoft Corp. identified approximately 500,000 possible materials in just a few days. Crunching reams of data, they narrowed it down to 18 compounds that might be suitable for use in EV batteries.

After evaluating the list and making physical samples of suggested materials, scientists at the Pacific Northwest National Laboratory came up with one candidate that could get the same results as current batteries but with 70 per cent less lithium. If viable, that approach would slash the cost of EV batteries and alleviate a key supply bottleneck.

Most of the research performed by both humans and AI on emissions-free transport will come to naught; that’s just the nature of scientific discovery. But with a caffeine shot and a powerful computer, there’s nothing humanity can’t achieve.