Battery chemistry recharged

With more renewable energy going into electricity grids, energy storage solutions are increasingly needed to balance supply and demand, Anthony King reports 

In 2018, around 9.5GWh of battery storage capacity was added to the world’s electricity grid system, hitting 17GWh at the start of 2019. China and South Korea together installed around 5GWh.

‘We expect the global market to be around 12GW annually by 2025,’ says Julian Jansen, market analyst at IHS Markit. Currently, around 90% of grid-connected stationary battery storage utilises large lithium-ion batteries, largely because of price. 

‘Since 2012, prices have fallen by around 70%, driven mainly by large-scale expansion of manufacturing,’ says battery chemist Bob Sadoway at Massachusetts Institute of Technology (MIT). Demand from automotive firms has pushed prices down, with lithium-ion costs predicted to fall 5%/year until 2025.

However, battery chemists say lithium-ion is suboptimal for grid storage. ‘In some respects, they are too good for stationary storage,’ says Helmut Ehrenberg, battery chemist at the Karlsruhe Institute of Technology in Germany. Developed to pack as much energy into the smallest volume, they rely on expensive lithium, cobalt and copper. Lithium-ion batteries also degrade after each cycle, and deep cycling – especially discharge below 30% capacity – causes significant capacity fade. 

Safety is also an issue. ‘When it comes to massive grid-scale storage, lithium-ion has a volatile, flammable electrolyte. If the batteries overheat, cells can bloat and ultimately burst into flame,’ says Sadoway. In November 2017, a 20MW facility in Belgium caught fire, with more than a dozen battery fires reported in South Korea in 2018. ‘The main challenge is to improve safety,’ says Dirk Henkensmeier, battery chemist at the Korea Institute of Science and Technology in Seoul, South Korea.  

Presently, lithium-ion is mainly used to provide power for less than four hours, for balancing supply and demand on the electric grid and for storing energy from renewables to make up for intermittency. ‘However, as other applications become feasible, especially longer duration applications that require six to eight hours energy storage, then we do see alternative technologies playing a greater role,’ Jansen predicts. The electricity grids of today can cope with limited energy storage, but as renewable penetration increases, a range of electrochemical storage solutions will be needed.

For six to 12 hours storage, energy experts spotlight a type of battery called a redox flow battery as offering a far better mix of characteristics than lithium-ion. A redox reaction is one in which electrons are transferred between chemical species, with oxidation (loss of electrons) and reduction (gain of electrons). The most common commercial variety is a vanadium flow battery, which comprises of two tanks filled with a water-based electrolyte, consisting of vanadium in two different valence states. Pumps force vanadium solution through the battery during charging and discharging. 

Unlike lithium-ion batteries, where power and capacity depend on the dimensions of the battery cell, in flow batteries power is governed by the active area of the electrochemical cell, while capacity is tied to the tank sizes. This decouples power and energy. ‘You can have a large amount of power and small amount of energy, or you can have a small amount of power and large amount of energy,’ says Sadoway. ‘In a conventional battery, you don’t have that flexibility.’ Another advantage is that redox flow batteries can be run for longer than lithium-ion without degradation – and they start up almost instantaneously. 

Still, there are downsides. The solubility of salt in aqueous electrolyte is low, so decent capacity requires quite large tanks. ‘To compensate for that dilution, you force a flow of water past the electrodes,’ says Sadoway. ‘The pumps are in constant motion and require maintenance.’ He adds that as a transition metal, vanadium is hazardous to groundwater and tanks will require double walls and a degree of caution. Another issue is cost. ‘Vanadium batteries today are more expensive that we would like,’ notes Nigel Brandon, an electrochemical engineer at Imperial College London. The tanks contain lots of vanadium, which is roughly half the cost of the battery. Moreover, vanadium prices are tied to steel production and can be volatile. 

Another problem for flow batteries is that there is currently no global supply chain. ‘Companies are making them,’ says Brandon, ‘but they are often quite small companies, without the volumes and supply chains that would drive down costs.’ Jansen agrees, adding that a big challenge for flow battery developers is achieving scale manufacturing to compete with lithium-ion. ‘These are often small, start-up companies, while the lithium-ion supply chain has got trusted manufacturers that can offer guarantees and warranties needed by project developers and utilities using these assets,’ he adds.

In 2017, BASF invested in Oregon-based flow battery company ESS. ESS says its iron-based electrolyte provides over 20,000 cycles of power with little or no maintenance and is ideal for storing four to eight hours of energy. Iron species change valence states in two separate tanks, which use the same non-toxic electrolyte and no rare earth metals. Lockheed Martin’s Gridstar Flow battery, meanwhile, combines abundant transition metals and commodity chemical ligands. 

In the UK, Brandon’s group at Imperial is developing a redox flow battery based on manganese and hydrogen. ‘Manganese is about a tenth of the price of vanadium. Given that vanadium makes up about half the cost of a vanadium flow battery, you can see that this would make a flow battery a lot cheaper,’ Brandon explains. The device works similarly to vanadium flow, with manganese dissolved in acid in one tank and hydrogen gas in the other. Discharging the device would consume hydrogen gas and changing the oxidation state of manganese on the other side. In 2018, Brandon helped spin out a company from Imperial, RFC Power, which is commercialising the system.  Initially, these will likely be 10 to 100kW systems, with the potential to develop MW scale system for grid applications and solar and wind farm support. 

An alternative flow battery setup is hydrogen bromine, with hydrogen generated and stored in a tank as the battery is charged. Also, hydrogen tribromide (HBr3) is formed and mixed in a tank with hydrogen bromide (HBr) as the electrolyte. During discharge, H2 is consumed and energy generated, with H2 combined again with HBr3 to turn to an initial state with a tank of HBr. Yet another flow battery is zinc-bromine, which consists of zinc bromide salt dissolved in water. As the battery charges, metallic zinc is plated from solution onto a negative electrode, while bromide is converted to bromine at the positive electrode.

In Australia, Thomas Maschmeyer at the University of Sydney has developed a zinc-bromine battery with a safer gel rather than liquid electrolyte. ‘It won’t blow up, it is low cost and has a long life. At four hours, it has a slow-ish charging rate, but that’s compatible with solar panels,’ says Maschmeyer. The group’s spin out company Gelion Technologies deployed its first commercial device in 2019 at the University of Sydney. A 2016 investment by Armstrong Energy saw the creation of Gelion UK, which is now the parent company. It plans to make commercial scale batteries of 5MWh by 2020, and 100+MWh by the end of 2021 for grids and renewable energy storage. 

The technology is designed to be less expensive than lithium-ion batteries and to offer at least 3000 cycles from 100% to 0% and back again. ‘The ability to tolerate zero charge means safer transport and installation as well as longer life,’ according to Maschmeyer. ‘We are building the first 2MWh of batteries by mid-2020, with a partner, and currently building the first few tens of kWh modules.’ Gelion has agreements with companies for £35m of batteries. An ability to tolerate zero charge means safer transport and installation of these batteries, and longer life, since the electrode surface can be completely rejuvenated.

However, Brandon notes that bromine is toxic and says any release would cause concern. His group has patented a flow battery with sulfur dissolved in sodium hydroxide, and air on the other side. He believes this technology could offer 24 hours of energy storage far more cheaply than any of the alternatives. 

There are also ways of improving vanadium batteries. Capacity fade is mainly due to cross-over of redox-active species. Henkensmeier believes this issue and the lower energy efficiency can be addressed by developing better membranes than the current standard, Nafion. This perfluorinated cation exchange membrane transports protons, but also lets metal ions drift across, leading to capacity fade. To reduce permeability, Nafion membranes are often used in very thick grades: 180µm thick. ‘This increases the membrane resistance and thus reduces the voltage efficiency,’ says Henkensmeier, who has made membranes with polybenzimidazole (PBI), a fibre-forming polymer used in fire-protective clothing. PBI membranes show lower permeability for vanadium ions, and thus can operate as a far thinner layer, which boosts energy efficiency.

In the US, Sadoway champions liquid metal batteries, which he has developed over the past 15 years. This setup sees a pair of hot liquid metals, separated by molten salt, with the high-density metal below. ‘The metal on top is converted to its ion, and the electrons go to the external circuit. The ions neutralise when they reach the bottom metal,’ Sadoway explains. As the battery discharges, the top layer thins, with the alloying action driving the external circuit. Current then forces the metals to de-alloy, returning the battery to its pristine state. Top metals in the mix have included lithium, calcium, and magnesium, with lead, antimony and bismuth trialled as bottom metals. 

Image: Researcher studying battery technology for storing renewable energy

Marlborough, Massachusetts-based spin-out company Ambri is developing a device with calcium on top and antimony on the bottom. Calcium chloride, the molten salt used here, is cheaper than alternatives such as lithium chloride. The battery must operate at high temperatures, and would typically be charged over a period of four hours – perhaps in the middle of the day from solar panels – rested, and then after sunset discharged over a period of four to five hours. This continuous duty cycle could be enough to maintain the battery temperature, says Sadoway. ‘We’ve perfected the chemistry at MIT. Now it is about mastering the ability to manufacture reliably and cheaply, at scale,’ he explains. This battery could last decades, with little capacity loss, and its materials then refined and re-used at the end of life. Ambri could offer a commercial liquid metal battery by 2021.

Sodium-ion battery chemistry was discovered as long ago as lithium-ion. Due to its higher energy density, lithium-ion was developed and launched commercially in the early 1990s, while Na-ion lagged far behind. ‘People have been looking seriously into sodium since only around 2010, I would say,’ says Magda Titirici, materials scientist at Imperial College London, who is developing carbonaceous material for Na-ion anodes using recycled organic waste. Li-ion and Na-ion battery chemistries share similarities, and manufacturers could tap similar processes to make Na-ion. The advantages of Na-ion are far cheaper sodium; no expensive cobalt in the cathode; and cheaper aluminium rather than copper for the current collector. 

Titirici estimates that a sodium battery would be 40% cheaper than lithium-ion batteries, with the difference especially significant in cathode costs. ‘In another five years, sodium-ion batteries should be on the market for sure,’ she says. However, sodium-ion has similar cycling characteristics and capacity fade issues as lithium-ion, and arguably safety downsides too.  

UK firm Faradion has a prototype sodium ion battery that it says can deliver lithium-ion performance at lead-acid prices. The cells boast an energy density above 140Wh/kg, while Li-ion range from 100 to 250Wh/kg. For now, Na-ion batteries last for more than 2000 cycles, compared with ca 300 cycles for lead acid, and over 3000 cycles for Li-ion. ‘In future we want a sodium ion battery having 300Wh/kg and 5000 cycles, ideally,’ says Titirici.

Today, lithium-ion is popular for short-term energy storage, rather than for five hours to days timescales, and there is no commercial incentive for storing energy longer than lithium-ion’s zone of comfort. ‘But all the forecasts suggest that a market will emerge for long-term energy storage over time,’ says Brandon. ‘We need alternative chemistries, and ‘the more mixture of technologies the better,’ Titirici says.