Dear Aventine Readers, 

In this issue we’re looking at batteries — specifically, what it will take for sodium-ion batteries to displace their lithium-ion counterparts for certain purposes as the demand for batteries skyrockets. Lithium-ion batteries have many advantages, like packing a lot of energy into relatively small and light spaces. But lithium can be expensive and difficult to get. Sodium-ion batteries, meanwhile, are bigger and heavier, but sodium is one of the most plentiful resources on earth. Given the competing advantages, will sodium-ion batteries become the default option in situations where size and weight are not key factors? Read on to find out. 

Also in this issue: Researchers have figured out a method of removing carbon from the air that creates green hydrogen in the process; excess power from home solar panels and EVs could be harnessed to power the grid; and artificial intelligence — which has proved itself to be excellent at manufacturing false information — is now debunking conspiracy theories.

Thanks for reading! 

Danielle Mattoon 
Executive Director, Aventine

Subscribe

Subscribe to our newsletter and be kept up to date on upcoming Aventine projects

A potential solution to humanity’s reliance on lithium-ion batteries may lie in one of the planet’s most abundant elements.

A new breed of sodium-ion batteries, previously sidelined because of the superior performance of their lithium-ion counterparts, is gaining traction. Over the past ten years the global demand for batteries has soared and is expected to quadruple in this decade. As governments have grappled with the geopolitical and ethical complexities of procuring lithium, the performance of sodium-ion batteries has improved enough that major players in the battery industry are betting that the once neglected technology will play a meaningful role in our energy future. 

Over the last two years, battery makers such as China’s BYD and CATL, America’s Natron Energy and Peak Energy, Sweden’s Altris and the U.K.'s Faradion have all made announcements about new sodium-ion battery technologies and investments. This summer, the news has gotten more significant. Natron announced that it would invest $1.4 billion to build a 1.2 million square foot sodium-ion battery factory in North Carolina that will enable it to produce 40 times more batteries than its current facilities allow. Meanwhile, the first phase of the world’s largest sodium-ion energy storage facility — potentially capable of storing enough electricity to power 12,000 homes on a single charge — went online in the Hubei province of China.

Battery experts who spoke with Aventine believe that sodium-ion battery technology is now mature enough for commercial use in certain applications. They also said that while it is unlikely that a sodium battery will ever outperform its lithium counterpart, there is still ample room for it to improve dramatically enough to become popular. “I think commercially sodium-ion will be successful,” said Robert Armstrong, a battery researcher at the University of St Andrews in Scotland. “I think what's in doubt at the moment is the time scale over which that happens.”

Ultimately, the speed at which sodium-ion batteries enter the mainstream may rest as much on global geopolitical currents as on technological advancement. At the moment, the overwhelming majority of lithium is mined in three countries: Australia, Chile and China. And according to the International Energy Agency, one country dominates the production process: China. The country performs “well over half” of all global raw material processing for lithium and is home to 85 percent of global battery cell production capacity. None of this sits well with governments around the world which are nervous about relying on other nations — particularly China — to secure their energy futures. See, for instance, the $3.5 billion that the Biden-Harris administration has committed to strengthening battery manufacturing in the U.S., which priortizes “next-generation technologies and battery chemistries, in addition to lithium-based technologies.” 

Sodium vs. lithium 

Compared to lithium, sodium is very easy to come by: It’s the sixth most common element in the Earth’s crust, making up about 2.8 percent of it compared to lithium’s 0.002 percent. It is also cheap and easy to obtain. And sodium-ion battery technology isn’t new: it was being actively researched at the same time as the development of lithium-ion decades ago. But lithium batteries were commercialized first — by Sony, for use in camcorders — and that commercial edge, combined with lithium-ion’s performance advantage, dampened interest in sodium-ion alternatives. 

The electrochemical processes that take place in both lithium-ion and sodium-ion batteries are broadly the same. But the exact chemistry inside them differs because of the different materials; as a result, sodium-based batteries end up having a lower energy density. This means that to deliver the same amount of power that a lithium-ion battery does, a sodium-ion battery may need to weigh up to twice as much. Though basic physics suggests that sodium will never beat lithium in absolute performance, experts believe that research and development could allow it to narrow the gap. (To get a sense of what that gap currently looks like, the current best energy density of sodium-based batteries in 2022 was about equivalent to the performance of low-end lithium-ion batteries in 2012, according to analysis of data from Bloomberg New Energy Finance undertaken by MIT Technology Review.)

So: what’s happening to improve sodium-ion technology? “There's questions around every part” of the battery, said Armstrong. While there could be an advance that dramatically changes the energy density of sodium-ion batteries overnight, experts Aventine spoke with agreed that it is more likely that improvements will accrue over time. “[We see] low single digit increase in energy density and low single digit decreases in cost per year for lithium-ion, but it's been doing that year-over-year for like, 30 years,” said Jay Whitacre, the Trustee Professor of Energy at Carnegie Mellon University and previously the founder of Aquion Energy, a sodium-ion battery company. “It's amazing, and that continues,” he said. A similar trajectory, he added, is expected to play out for sodium-ion batteries, with multiple R&D cycles required to improve their performance.

The case for an inferior battery

A key driver of the current enthusiasm for sodium-ion batteries is that they may not need to directly outperform lithium-ion in order to be useful. Crucially, there are plenty of applications in which energy density is not the most important factor in a battery. Take grid storage: Batteries deployed to soak up excess energy from solar and wind power do not need to be small and light in the same way a battery used in a high-performance electric vehicle does. “It doesn't matter if you're going to produce a big battery for the [storage of] renewables,” said Armstrong. Altris and Natron, for instance, are primarily focused on building sodium-ion batteries for grid storage and industrial applications. (Neither company responded to a request for an interview.) 

Proponents of sodium-ion technology argue that sodium-ion batteries have plenty of other advantages over lithium-ion alternatives: They can be recharged more times over their lifetimes and can be fully drained before degrading, they are less prone to overheating and catching fire and they perform more reliably at low temperatures. Armstrong cautioned that the jury is still out on some of these claims, with further research needed to establish whether they stand up. Whitacre, meanwhile, pointed out that it’s now possible to design lithium-ion batteries with those attributes as well. One benefit that is not questioned is that sodium-based batteries are well understood in terms of manufacturing: “It's a nondisruptive and drop-in technology,” said Armstrong. “You can make sodium-ion batteries on a lithium-ion production line.”

Cheap, short-range transport — such as e-bike or e-scooters — is commonly cited as another obvious use for the technology because power and range are less important than for EVs, though this assumes that the cost of sodium-ion batteries can be driven down.

But will it make economic sense? 

The final — and likely determinative — variable in the question of whether sodium-ion batteries will become a permanent part of the energy landscape boils down to whether they can compete on price. 

While there is far less lithium on the planet than there is sodium, it’s not the total amount of lithium that is the problem. The U.S. Geological Survey notes that the U.S. has a lithium resource of 14 million tons and that other countries have an additional 91 million tons — plenty to facilitate the global electrification now occurring. Putting aside the ethical dimensions of lithium mining, which are significant, the challenge is extracting it quickly and affordably.

As demand for lithium has increased over recent years, its price has become increasingly volatile. From 2021 to 2023, for example, the price of lithium bounced from $12,600 up to $68,000 and down to $46,000 per metric ton and remains erratic, reacting to news such as the discovery of new resources. This kind of raw material price volatility “can make or break battery economics,” a researcher at the Oxford Institute for Energy Studies wrote earlier this year.

The experts Aventine spoke with cited this behavior as a key driver in the resurgence of interest in sodium-ion battery technology, but also as a potential barrier to its widespread adoption. If the price of lithium stabilizes at a relatively high price, sodium-ion technology has room to develop economies of scale and become the cheaper option. But if the price of lithium drops and stays low, it will be much harder for competing technologies to gain a foothold. While some nations, such as France and Japan, may look to subsidize sodium-ion battery technology to help secure their energy futures, the global outlook is likely to be rather more pragmatic. “Customers would always go for low cost,” said Magda Titirici, Professor and Chair in Sustainable Energy Materials at Imperial College London. The trouble for sodium-ion batteries is that despite the fluctuating price of lithium, lithium-ion batteries — thanks to accrued improvements — have fallen in price by about 97 percent since their commercial introduction, and continue to do so according to our sources. Meanwhile, sodium-ion battery production is not yet developed enough to be economically competitive. 

“Look how expensive lithium-ion was 20 years ago, and now look, look where it is,” said Whitacre. “Can sodium-ion follow the same path? Maybe. But I'll tell you, it's gonna take 20 years; it's not like it's gonna happen in a year.”

 Blake Wish/Unsplash

A method of removing CO2 from the air that creates green hydrogen in the process. The world’s oceans are an unlimited source of hydrogen, and chemists have long wondered how to efficiently and affordably tap that resource for fuel. A new approach goes some way to making that a possibility while also removing CO2 from the air, reports Heatmap. With significant backing from the Department of Energy’s Advanced Research Projects Agency-Energy, or ARPA-E, a startup called Equatic has developed a process that uses electrolysis to split seawater into hydrogen and oxygen while producing a stream of alkali water that, when exposed to the air, soaks up CO2. Some of the carbon is dissolved safely back into the water as carbonates and bicarbonates; some forms solid calcium carbonate, which is removed. A turning point for the company was developing a new piece of hardware that minimizes the effects of chlorine in the process, which previously has led to the creation of large quantities of corrosive chlorine gas in electrolysis reactions with seawater. The new process uses electricity, and a lot of it: to remove 1 metric ton of carbon from the atmosphere, the process requires 2.5 megawatt-hours of electricity. But it also produces the equivalent of 1 megawatt-hour of energy in hydrogen, so the net energy use is approximately 1.5 megawatts — far lower than existing direct air capture carbon removal systems, which used about 2.6 megawatts. The big question for Equatic is whether it can scale up its process, which two new plants — one in Singapore and one in Quebec — will seek to demonstrate. (To learn more about green hydrogen and CO2 removal, listen to our podcast series, which has episodes devoted to each. It’s also available on Apple, Spotify and other podcast sources.) 

OpenAI built an LLM that it claims can reason. The company behind ChatGPT has unveiled a new kind of large language model, called o1, that has been designed to solve difficult problems step-by-step. Though the company has released few technical details about the architecture of the model — and has even warned users not to probe its inner workings — it has described how the system is designed to reason through a problem as a human might, a potentially monumental shift in the capabilities of AI. The software is trained using a well-established technique called reinforcement learning, an artificial intelligence approach in which models are rewarded for correct answers and penalized for wrong ones. OpenAI claims that this makes o1 far better at tackling complex math and science problems than its predecessors, noting that the model solved 83 percent of problems on a qualifying exam for the International Mathematics Olympiad, compared to GPT-4o, the most advanced of OpenAI’s other models, which solved just 13 percent. The new model is slower than its GPT models and lacks many of their features, such as web searching and image handling. Perhaps most interesting is that the company sees o1 as a complementary technology to its GPT models, telling Wired that its forthcoming and most advanced model yet, GPT-5, will make use of the technology that underpins ChatGPT as well as that used in o1. The new model is available for paying ChatGPT users, and only over the coming weeks and months will people get a better understanding of the tool’s abilities — and whether the claim that it can reason holds water.

AI can talk people out of conspiracy theories. We know large language models can create misinformation. Researchers have now shown that they can battle it, too. Researchers from MIT, Cornell and American University have built a tool called DebunkBot that is designed to provide conspiracy theory devotees with counterarguments to convince them that their beliefs are misplaced. The team asked 2,190 participants to tell DebunkBot about a conspiracy theory they found credible and provide evidence that supported their view. They were also asked to rate their confidence in the theory on a 100-point scale. The participants then chatted with the bot — which had been instructed to persuade users against the theory they described — three more times and were asked again to rate their confidence in whatever conspiracy theory they originally espoused. The results, published in Science, show that participants on average experienced a 20 percent decrease in their convictions, an effect that persisted when the researchers followed up with them two months later. Speaking to Ars Technica, the team explained that they thought the success lay in the tool’s bespoke approach, as it was able to target the specific origins of a person’s belief. The team tested the tool in the wake of the July 2024 assassination attempt on Donald Trump and it was found to be less effective, possibly because there is typically less evidence available to refute false claims soon after an event has happened. The bigger challenge may be how to best make use of such a tool. One suggestion is that it could be embedded in search engines or social media bots to provide targeted information to users in the future.

The End of the Lab Rat? from Scientific American
3,700 words, or about 15 minutes

It’s hard to get an accurate read on how many animals are used in labs, but some estimates suggest that it’s as many as 50 million per year in the U.S. alone. In addition to ethical concerns over animal cruelty, there are other problems with science’s reliance on experimenting on living creatures: It's time intensive, expensive and the results are often not translatable to humans. This story takes a close look at how academics, private companies and lawmakers are increasingly getting behind new technologies, including those known as organs-on-a-chip, that can allow researchers to perform experiments without the need for animal models. These technologies have matured to the point that they sometimes offer more accurate results than those performed on lab animals. The challenge facing wider adoption is that there is significant inertia around the use of animals in labs, which means the transition away from them will be long and difficult.

Here’s a blueprint for building virtual power plants in every state, from Canary Media
2,500 words, or about 10 minutes

Domestic battery packs, electric vehicles and smart home devices are all separately doing their bit for humanity’s transition toward an all-electric economy, but they can also all be bought together to achieve something even more impressive. With an additional layer of hardware and software in place, these systems can be used to store electricity during times of excess supply and share it during times of demand — not just for the house or vehicle they’re connected to, but across whatever larger electrical grid they are attached to, working as a so-called virtual power plant, or VPP. If rolled out correctly, the technology could reduce the amount of storage capacity that will need to be built, potentially saving utilities and consumers billions of dollars. The trouble, as this story explains, is agreeing on how exactly that should be done given state-by-state regulations and patchwork grid infrastructure in the U.S.

The biology of smell is a mystery — AI is helping to solve it, from Nature
2,600 words, or about 11 minutes

Our noses are mind-bogglingly complex: The eye has two types of receptors to detect light from the outside world; the nose has 400. And the science of scents is also bewildering: Molecules that have similar structures can smell entirely different, while molecules that have radically different structures can smell almost identical. This complexity has made it fiendishly difficult for scientists to understand the science of smell, but artificial intelligence is increasingly helping researchers understand both the odor receptors in our noses and the aromatic compounds that interact with them in order to develop a better understanding of how our sense of smell really works.