An Electric Car Battery That Will Get You From Paris to Brussels and Back
The metal-air battery carries more energy per kilogram than today’s lithium-ion batteries
Illustration: Elias Stein
Proposition: Electric cars will remain mostly niche products until they have a range of 800 kilometers, or roughly 500 miles, with an affordable battery.
That’s as far as most people would want to drive in a day, and then they have all night to recharge.
That’s how we came up with a figure of 800 km—or a nice round 500 miles—as the goal for our R&D project, Battery 500. It began in 2009 at the IBM Almaden Research Center, in San Jose, Calif., and has grown since then into a multinational partnership with commercial and academic participants in Europe, Asia, and the United States. It is based on metal-air technology, which packs far more energy into a battery of a given mass than today’s state-of-the-art technology, the lithium-ion battery. We are still years away from commercialization, but we have made enough progress to predict that these batteries could be used in cars in the foreseeable future. Why are we so confident? Read on.
Electric motors are ideally suited for powering cars. They’re lightweight and extremely powerful, they achieve efficiencies in excess of 90 percent, they don’t need complex transmissions, and they churn out torque in just the right way, providing full rotational force starting with zero rpms. Internal-combustion engines, by contrast, don’t produce high torque until they’re spinning at thousands of rpms.
But even though they’re propelled by a near-ideal mechanism, electric cars have a huge drawback, which is the low energy content of the batteries. Gasoline packs about 13,000 watt-hours per kilogram; the best production lithium-ion cells store only about 250 Wh/kg. Add the mass of the ancillary battery equipment—including the bus bars, cooling system, and battery management system—and the energy density of the entire system drops by half, giving the batteries a pitiful 1 percent of the raw energy density of gasoline.
This huge gap between the energy densities of gasoline and batteries seemed to make it impossible to build competitive electric cars, but the success of the Tesla Model S has shown that it can be done. One major factor in favor of the electric car is the high efficiency with which it converts battery power to motive power at the wheels—about six times as efficiently as the average for gasoline-fueled cars in the United States. Also, electric car makers put the biggest, heaviest battery they can reasonably fit into their designs. Even so, the ranges fall far short of the 500-mile target. The upshot is that electric-car batteries need to attain at least twice the energy density of Li-ion cells to achieve a range of 800 km.
In The Electrode—Or On It
A lithium-ion battery puts lithium ions inside the electrode; a lithium-air battery compounds lithium and oxygen on top of the electrode, allowing lighter electrodes
Cost is at least as important as energy density. Today’s battery cells run from US $200 to $300 per kilowatt-hour, which means that—given an average range of 5 or 6 km/kWh—a car with an 800-km range will require a 150-kWh battery costing $30,000 to $45,000. By comparison, the base price of a BMW 2 Series car is $33,000. The price per kilowatt-hour must fall to at most $100 if the technology is to gain a serious foothold. At that price point, the car’s much lower operating costs in energy use and maintenance, together with the sheer pleasure of driving such a responsive machine, will assure success in the marketplace.
But how do we get that 800-km-range battery? Well, start with the current state of the art, the lithium-ion battery.
A conventional, or “intercalation,” Li-ion battery is a sealed system that has one electrode made of graphite (the anode) and an opposing electrode (the cathode) typically made from various oxides of transition metals, such as cobalt, nickel, or manganese. Both electrodes are immersed in a liquid organic electrolyte that contains dissolved lithium salts. In this electrolyte, lithium ions travel from one electrode to the other, the direction of travel depending on whether the battery is charging or discharging. In between the electrodes, immersed in the electrolyte, there’s a porous polymeric separator, which prevents the electrodes from short-circuiting. The ions insert themselves between the atomic layers of the electrode material. This process is known as intercalation, and it is reversible—that is, it allows recharging.
If the electrodes are connected through an external circuit, lithium ions will migrate from the negative electrode to the positive electrode while electrons flow through the attached external circuit, thus discharging the battery. An externally applied voltage will reverse the ion flow, recharging the battery. The battery capacity depends on how much material is available for intercalation. In other words, the battery capacity is related to the volume and hence, the mass of the anode and cathode.
Metal-air batteries, however, employ a true electrochemical reaction rather than intercalation. For clarity, we’ll assume that the metal is lithium. During discharge, the metallic lithium anode releases lithium ions; these travel through the electrolyte and combine with oxygen at the cathode, forming lithium peroxide (Li2O2). As in a conventional Li-ion battery, electrons flow through the external load circuit to compensate for the Li-ion flow inside the battery. The lithium peroxide accumulates on the surface of the porous carbon cathode, where the three participants in the reaction (lithium ions, electrons, and oxygen) meet and react. Since the reaction occurs on a surface, the volume or mass of the cathode material does not matter so long as it has a large surface. This is the main reason why this battery type has such high energy density.
Not So Rough
Recharging reverses this order of events: An externally applied voltage breaks up the lithium peroxide, the oxygen diffuses back out into the environment, and the metal ions migrate back to the anode, where they acquire electrons and thus convert back to bulk metal.
This general principle can be applied to several different metals. Lithium-air, sodium-air—an interesting new contender [see sidebar, "Sodium: Less Energy, More Stability"]—and potassium-air are all possible systems because they lend themselves to recharging. Heavier metals, such as zinc, magnesium, iron, or aluminum have proved very hard to recharge, so we will not consider them here.
Our work has focused on lithium and sodium. Let’s start with lithium, which has the greater energy storage capacity. A lot of extraneous chemical reactions can mess things up in these cells. To understand these side reactions, we precisely measure the gases consumed and produced during the cycling of these cells. For this, we used a sophisticated differential electrochemical mass spectrometer that we built at the IBM Almaden Research Center. It features eight stations to measure gas in parallel experiments.
It was this instrument that really gave us key insights. For example, it showed that the early lithium-air cells released far less oxygen on recharge than they consumed during discharge. (For most experiments we use dry oxygen rather than air.)
In an ideal cell, the amount of oxygen consumed during discharge should exactly equal the amount released on recharge. Our finding, therefore, was bad news because it implied that much of the oxygen released during the (desired) breakup of the Li2O2 on recharge was attacking components in the cell itself, notably the electrolyte. The cells weren’t recharging—they were self-destructing!
With the help of our IBM sister laboratory, in Zurich, we traced the source of this parasitic reaction through experiments and computer simulations. We determined that the main problem was our organic electrolyte—it was breaking down. Since then we’ve gone a long way toward solving the problem. When our latest cells recharge, our new electrolyte (we’ll describe it shortly) allows them to release most of the oxygen they take in during discharge. We also carefully monitored the hydrogen and water produced during the cycling of the cells, because their presence indicates that we’re still seeing at least one other parasitic reaction.
We have been able to achieve 200 discharge-charge cycles, although so far we can do so only by limiting the discharge to less than the theoretical maximum.
Here are some of our key findings:
Anode: Unlike the graphite anodes in a standard Li-ion cell, our metallic lithium anodes change their surfaces dramatically during recharging by growing mossy or treelike structures, called dendrites. These dendrites are dangerous because they provide conducting pathways between anode and cathode that can short-circuit the cell.
We’ve had great success in limiting the formation of dendrites by putting a special separator between the anode and the source of the lithium ions. This separator consists of a layer—we’ve tried both organic and inorganic materials—that contains nanometer-scale pores. That’s small enough to uniformly distribute the flowing ions’ current and thus suppress dendrite formation. With this nanoporous separator, the metal remains smooth for hundreds of cycles, whereas with a standard separator it forms dendrites after only a few cycles. Another membrane that combines ion-
conducting glasses with a polymer matrix works even better.
Fortunately, an electric car’s big battery requires only hundreds, not thousands, of full cycles. For example, a car with the 500-mile range needs to be fully recharged only 400 times for a lifetime range of 200,000 miles (roughly 320,000 km).
Electrolyte: The improved electrolyte solvent molecules we use can still be broken apart by oxygen and other compounds generated during the operation of a cell. We haven’t yet discovered any single solvent that’s stable enough for commercially useful lithium-air cells, but we have found a cocktail of solvents that works pretty well.
Cathode: We add traces of LiNO2 (lithium nitrate) to our carbon cathode to minimize the undesirable catalytic effect that accelerates the breakdown of the electrolyte during charging and releases carbon dioxide. Even so, this reaction still requires that the applied charging voltage must be higher—by up to 700 millivolts—than the battery’s operating voltage. Such a high overvoltage reduces electric efficiency—that is, the fraction of the energy pumped into the battery during charging that’s returned during discharge. Although it’s much better than what you get with plain carbon cathodes (over 1,200 mV), it’s still too high for practical use. We’ve had similar results when we replaced carbon with metal oxides.
Catalysts: The pros and cons of deliberately using catalysts in metal-air batteries are subject to much scientific debate. Catalysts often lead to an apparent reduction of overvoltage, but one has to be extraordinarily cautious about claiming a net benefit, because catalysts generally accelerate the destruction of electrolytes. Also, our theoretical studies indicate that the activation energy of the lithium-oxygen reaction—in both directions—is very low. Thus catalysts should not be needed.
Air preparation: We’ve called these devices lithium-air cells, but in fact we’ve mostly been using dry oxygen gas. The emphasis here is on “dry” because we need only remove the water vapor and the carbon dioxide from the air, not the nitrogen, to make it usable. To do that at scale in a commercial battery, we’ll need to put in a substantial engineering effort to create an air-cleaning system that’s sufficiently light, efficient, and reliable to retain the energy advantage of this technology.
Another outstanding engineering task is how to scale up to much larger cells and integrate them into a multicell module and pack, including a tailor-made battery management system. Our original cells measure roughly 13 millimeters in length and 76 mm in diameter; we are testing versions that measure 100 by 100 mm.
The whole project was motivated by the desire to achieve high specific energy density—that is, the energy per unit mass. Where are we now?
The lithium-oxygen reaction has a theoretical (specific) energy density of 3,460 Wh/kg, which is much higher than the theoretical limit of any lithium-ion intercalation chemistry. The practical energy is much lower than the theoretical values for both intercalation and metal-air chemistries due to the inert mass contributed by those cell components that take no part in the reaction. These include the electrolyte, the cell housing, current collectors, and the separator. Furthermore, a lithium-air battery would also include the inert mass of the machinery needed to prepare ambient air for use in the cell. It is these engineering problems that make the practical development of lithium-air batteries for cars such a challenge.
It is too early to quote a practical energy density for the lithium-oxygen technology, let alone lithium-air technology. These numbers depend on engineering details, and the project is still focused on the basic science of the materials and chemistry. However, early results are encouraging. For example, we have measured a specific energy density of 15 kWh/kg of the raw carbon cathode material (5,700 milliampere hours multiplied by 2.7 volts per gram of carbon black). But as we pointed out earlier, the practical energy density will be greatly reduced by the mass of all the other components in the cell. Our current best estimate of what is practically attainable is around 800 Wh/kg at the cell level.
The first practical metal-air batteries might be used in buses, trucks, and other large vehicles that can more easily accommodate the mass of the air-
cleaning machinery. But the most profound change will come when the technology reaches the family car, freeing it from the “range anxiety” of today’s electric cars—and freeing us all from our dependence on oil and the many problems it causes.
Sodium: Less Energy, More Stability
Illustration: Elias Stein
Sodium-air batteries are another interesting possibility, despite having an energy density lower than that of the lithium-air chemistry.
The lower energy reflects the nature of the reaction, which uses only one electron and thus generates a superoxide (NaO2) instead of sodium peroxide (Na2O2). This reduces energy density immediately by a factor of two. The reaction’s theoretical specific energy is approximately 1,100 watt-hours per kilogram.
On the other hand, sodium-air batteries charge up more efficiently than lithium-air batteries because they have a very low overpotential—less than 20 millivolts as opposed to 700 mV. As a result, it’s possible to keep the operating voltage under 3 volts, which protects components from destructive oxidation, notably the electrolyte destruction observed in the lithium-air system. We have proved this by measuring efficiencies above 98 percent. These results make for good stability during cycling: After 50 cycles, the cell’s capacity is essentially unchanged.
There are a few technical challenges to overcome. For instance, because of the nature of the oxidation, the sodium-air battery sucks in twice as much air as its lithium-air equivalent, requiring an airflow comparable to that of a piston engine of the same power. Then there’s the high chemical reactivity of sodium metal, which you may remember seeing demonstrated in high school, in which a small piece of sodium reacts violently with water.
Lithium is comparatively rare, and it isn’t cheap. But sodium is as common as table salt, and it’s not expensive. The materials of a sodium-air cell would likely cost less than a tenth as much as those in a lithium-ion battery. In the long run, lithium-metal batteries promise the best performance, but given the combination of stability, low cost, and still-impressive specific energy, the sodium-air technology might serve to bridge the gap between today’s batteries and those of the more distant future.—W.W.W. & H.-C. Kim
This article appears in the March 2016 print issue as “The 800-km Battery.”