As energy storage becomes an increasingly integral part of a renewables-based system, interest in and discussion around non-lithium (and non-pumped hydro) technologies increases. A team of experts from CENELEST, a joint research venture between the Fraunhofer Institute for Chemical Technologies and the University of New South Wales takes a deep dive into redox flow batteries.
In the last 15 years, the increase in renewable energy sources such as photovoltaic and wind energy has accelerated significantly. At the same time, manufacturing and installation costs, especially for PV systems, have fallen significantly, making this energy source one of the cheapest and cleanest forms of energy, depending on the location. With the increase of fluctuating renewable energies in an electrical grid, the need for compensation possibilities at times when renewable energies are not available increases . One possibility is the use of electrochemical energy storage such as lithium-ion, lead-acid, sodium-sulfur or redox-flow batteries. Additionally, combinations of hydrogen electrolysis and fuel cells can be used . Batteries can be adapted in a flexible and decentralized manner depending on the respective requirements and are scalable from a few kW/kWh for e.g. domestic storage up to systems of several MW/MWh for grid storage. The different types of electrochemical energy storage systems have different physical/chemical properties, which affect the cost of the system. It is important to note that the cost of the storage system over its lifetime (levelised cost of storage – LCOS) is a critical factor used in selecting the most suitable system for a particular application . For example, the investment costs for lead-acid batteries are significantly lower than for all other technologies, but the service life is very short.
Technologies with similar investment costs at higher lifetimes result in a lower levelised cost of storage, but to be precise additional factors such as recycling, energy efficiency, and maintenance costs have to be considered. A battery with high efficiency, low recycling effort, low investment and maintenance costs and great freedom of scalability to meet the requirements of the application would be an ideal system. In electrical networks, there are different storage time requirements: short-term storage, medium-term storage, and long-term storage. The shorter the storage time, the more suitable are physical storage devices such as capacitors. Batteries are suitable for applications ranging from a few minutes to several hours. In addition, mass storage systems such as electrochemical hydrogen generation (power-to-gas) are particularly suitable for long-term storage of several weeks.
Redox flow principles
All electrochemical energy storage systems convert electrical energy into chemical energy when charging, and the process is reversed when discharging. With conventional batteries, the conversion and storage take place in closed cells. With redox flow batteries, however, the conversion and storage of energy are separated . Redox flow batteries differ from conventional batteries in that the energy storage material is conveyed by an energy converter. This requires the energy storage material to be in a flowable form. This structure is similar to that of fuel cells, whereby in redox flow batteries, charging and discharging processes can take place in the same cell. Redox flow batteries thus have the distinguishing feature that energy and power can be scaled separately. The power determines the cell size or the number of cells and the energy is determined by the amount of the energy storage medium. This allows redox flow batteries to be better adapted to certain requirements than other technologies. In theory, there is no limit to the amount of energy and often the specific investment costs decrease with an increase in the energy/power ratio, as the energy storage medium usually has comparatively low costs. Figure 1 shows the general operating principle of redox flow batteries. The energy conversion takes place in an electrochemical cell which is divided into two half cells. The half cells are separated from each other by an ion-permeable membrane or separator so that the liquids of the half cells mix as little as possible. The separator ensures a charge balance between positive and negative half cells, ideally without the negative and positive active materials coming into direct contact with each other. In fact, however, separators are not perfect so some cross-over of the active materials always occurs and this leads to the self-discharge effect.
In a single cell, there is always one positive and one negative half-cell. The electrochemical reactions for charging and discharging take place at the electrodes of the half-cells. The electrodes are the phase transitions of ionic and electronic conductors. In redox flow batteries, the electrodes should not participate in the reactions for energy conversion and should not cause any further side reactions (e.g. undesirable gas formation). Most redox flow batteries are therefore based on carbon electrodes.
The difference between the voltages of the positive electrode and the negative electrode is the cell voltage and is between 0.5 and 1.6V in aqueous systems. During the charging process, ions are oxidized at the positive electrode (electron release) and reduced at the negative electrode (electron uptake). This means that the electrons move from the active material of the positive electrode to the active material of the negative electrode. When discharging, the process reverses and energy is released. The active materials are redox pairs, i.e. chemical compounds that can absorb and release electrons.
In redox flow batteries, the energy storage medium is often referred to as an electrolyte. However, there are redox flow batteries that use a gas that is not an electrolyte (e.g. H/Br-RFB) as with hydrogen. As with all other aqueous batteries, aqueous energy storage media from redox flow batteries are also subject to water limitations. In case of too high voltages or more precisely too high or too low half-cell potentials, the water is decomposed into its components, hydrogen, and oxygen. The generation of hydrogen, in particular, is often present as a very small but undesirable side reaction and causes a charge carrier imbalance between positive and negative half-cells, which leads to a slow loss of capacity. Due to the flowability of the energy storage medium, the reaction products that would normally remain in the half-cell can be transported out of the cell and stored in separate tanks thus allowing the capability for a higher capacity than that attainable with conventional batteries.
Hybrid redox flow batteries
In redox flow batteries there are normally no phase transitions with solid active materials as with other batteries. This can significantly increase battery life because no lattice structures have to be rebuilt each time the battery is charged or discharged, and all materials are in a solution. The best-known representative of redox flow batteries today is the vanadium redox flow battery. However, there are also flow batteries in which solids are deposited and dissolved at one or both electrodes. A typical representative is the zinc/bromine redox flow battery and patented in 1885 by Charles Bradley . Such batteries are called hybrid redox flow batteries. In contrast to redox flow batteries, power and energy are not separately scalable, as the amount of possible solids deposition is limited by the cell geometry.
Hybrid redox flow batteries also usually have two electrolyte circuits like conventional redox flow batteries, but too much active material would lead to a too-thick layer of solids in the half-cell and thus to clogging of the fluidic system or to so-called dendrites, which are uneven deposits and can lead to short circuits through the separator. As with all other batteries, however, the power density decreases with the layer thickness of the half cells. For this reason, the space for the deposition and thus the layer thickness is optimized with regard to the power density and usually leads to storage times of approximately 4-8 hours. The advantages of the Zn/Br redox flow battery are the low costs of the active materials, zinc and bromine, and the high energy density of approximately 70-80Wh/liter. The disadvantages are above all the use of bromine and the relatively short cycle life with several thousand charging and discharging cycles. Elementary bromine is produced when charging the battery at the positive electrode. In order to reduce the risk potential and self-discharge, organic complexing agents are added to the energy storage medium to bind the bromine and prevent it from escaping . The complexing agent is relatively expensive and the subject of research to reduce battery costs. The used redox pair Br/Br- has a very high reaction speed and is ideal for batteries from an electrochemical point of view. In the 1970s, Exxon and General Electric launched relevant commercialization efforts in the USA and led to stack concepts, materials and production technologies that are still relevant today.
Later, starting in the 1980s, commercialization efforts were made mainly by ZBB Corp. Australia, which developed modular multi-megawatt battery systems but stopped working a few years ago. In the 1980s, developments were also made for use in electric vehicles, primarily by the Austrian companies SEA and Powercell. The batteries were used experimentally in various commercial vehicles and buses. A vehicle with a Powercell battery finished first in the EV Division of the 1994 and 1995 World Clean Air Vehicle Rallies in California  Today only one company is selling Zn/Br-RFBs.
The Ti/Fe-RFB uses Fe2+/Fe3+ as negative and Ti3+/TiO2+ as positive redox pair. The reactions of Fe2+/Fe3+ are very fast, but the reactions of Ti3+/TiO2+ are much slower, which makes the energy efficiency and power density relatively low. The Ti/ Fe-RFB has the disadvantage of a low cell voltage of about 0.8V and a low concentration of active materials. The maximum achievable energy density is thus approximately 10Wh/L and practically much lower. For these reasons, the Ti/Fe-RFB has never been successful.
With the beginning of the first oil price crisis in 1973, a rethinking of the energy supply began. Investments in regenerative energy sources and the necessary research and development of storage systems for fluctuating energy producers led to the development of the iron/chromium redox flow battery at NASA by Thaller . Thaller was also the first to use the term “Redox Flow Cell”. In the Fe/Cr-RFB, as in the Ti/Fe-RFB, the redox pair Fe2+/Fe3+ is used, but on the positive electrode. As already mentioned above, the reaction rate of the redox pair is high and thus are the achievable power density and energy efficiency. Iron is also an extremely inexpensive material for energy storage. By far the greatest challenges occur with the reactions of chromium ions at the negative electrode. The redox reactions of Cr2+/Cr3+ are very slow and are close to hydrogen generation, so the efficiency of the reactions is very low. NASA’s work was therefore primarily concerned with these reactions and their acceleration by catalysts and the suppression of hydrogen formation by inhibitors. Prior to this, however, there was again a screening of possible candidates as active materials for redox flow batteries. Again, vanadium was considered on a theoretical basis, but ultimately due to the cost was not further studied. Iron and chromium were selected because of potentially low costs . NASA’s work led to a demonstration system with an output of 1kW/13kWh as a domestic storage system coupled with a PV system  and lasted until around the mid-1980s. With the lowering of crude oil prices, the general interest in renewable energies and storage facilities decreased so that no commercialization took place. It was not until the mid-2000s that various companies attempted to commercialize Fe/Cr-RFBs again, but these were discontinued.
In 1981, Hruska and Savinell published an article about a hybrid redox flow battery that only uses iron as an active material . The motivation was the use of an energy storage material that was as inexpensive as possible, which is almost unsurpassable with iron. One kilogram of iron corresponds to approximately 500Wh, or 1kWh would cause approximately US$5 in active material costs. The Fe/Fe-RFB uses the soluble redox pair Fe2+/Fe3+ at the positive electrode but the redox pair Fe/Fe2+ at the negative electrode just like the two iron-based RFBs discussed above.
The initial solution is a relatively simple and widely available Fe(II) salt solution, similar to that used on a large scale in wastewater treatment. Solid and elemental iron is deposited on the negative electrode during charging and dissolved again during discharge. The challenges with this battery are the hydrogen generation at the negative electrode and the low reaction rate of Fe/Fe2+ at the negative electrode.
The generation of hydrogen leads to a significant loss of capacity, which can, however, be prevented or reversed by appropriate measures. The low reaction rates result in low efficiency, which can be increased by operating at elevated temperatures between 50-80°C. The low reaction rates result in low efficiency. The energy efficiency is approx. 60-70%, but it should be noted that energy efficiencies must be considered in connection with the application. The potentially low investment costs of the battery, together with compensation for losses from low-cost renewable energy, can result in a lower levelised cost of energy than other storage technologies. In the last 10 years, research and commercialization activities have increased, albeit at a very low level.
Vanadium redox flow batteries
At about the same time as NASA’s developments came to an end, the University of New South Wales conducted investigations into vanadium ions as an active material for redox flow batteries. Maria Skyllas-Kazacos et al finally found the possibility of using four different oxidation states on carbon electrodes in fundamental electrochemical investigations [17, 18] and later as battery experiments .
In the following years, almost all aspects of the vanadium redox flow battery (VRFB) were investigated at the UNSW. Efficient manufacturing processes were developed from inexpensive and widely available raw materials for the active material, the electrochemical and chemical basics were investigated, the energy density was increased through the use of concentrated solutions and stabilizing agents, inexpensive separators and membrane modification methods were investigated and the power density was increased through electrode treatment processes. Stack cost reduction was also achieved by the development of an inexpensive conducting plastic bipolar electrode substrate that can be welded to the cell flow frame. In 1989, the work led to a battery with 1kW power and its performance was reported in 1991 . This was followed by the installation of a 15kWh VRFB by the UNSW team in a Solar Demonstration House in Thailand and the licensing of the UNSW patents and technology to Mitsubishi Chemicals/ Kashima Kita Power Corporation (MC/ KKEPC), Japan in 1994. This was followed by a five-year R&D collaboration program between UNSW and MC/ KKEPC that led to the installation and testing of a 200kW/800kWh VRFB at Kashima Kita in 1995. In the following years, more and more work was carried out by other research groups and further commercialization trials were carried out by UNSW’s licensees. Particularly from the 2000s onwards, interest in research and industry increased sharply.
A further advantage is the simple recyclability of the batteries. Due to the high content of vanadium in the liquid electrolyte, the vanadium can easily be reintegrated into process chains and the existing value reused. The energy storage solution consists primarily of vanadium sulphate in a diluted (2mol/L) sulphuric acid containing a low concentration of phosphoric acid and is therefore roughly comparable to the acid of lead/acid batteries. The energy density is limited by the concentration of the pentavalent + VO2.
Unfortunately, pentavalent vanadium ions have a tendency to react with each other, which leads to the formation of larger molecules that precipitate as solids and can thus damage the system. The reaction depends on the temperature and the concentration of VO2+ (state of charge), but is also a function of the proton concentration. Increasing the acid concentration increases the stability of VO2+, but this reduces the solubility of the V(II), V(III) and V(IV) sulphates. The vanadium and total sulphate concentrations are thus set at around 1.8 and 4.2mol/L respectively in order to achieve an acceptable operating temperature range. With a high state of charge and elevated temperature, the tendency to form solids in the positive half-cell increases, which is why the electrolyte temperature is usually limited to a maximum of 40°C. To minimize the risk of precipitation of the other species at low temperatures, a lower limit of 10°C is usually recommended. Alternatively, the SOC limits can be adjusted to handle temperatures outside this range.
Like all other RFBs, VRFBs also has battery management. A battery management ensures optimum and safe conditions for battery operation. Often a heat management system is integrated to avoid too high or too low temperatures. The classic VRFB has undergone several further developments. First, the V/Br-RFB called Gen 2 was developed by Skyllas- Kazacos et al. at the UNSW .
By using the fast redox pair Br2/Br-, the objectives were a higher energy density and better temperature stability, as well as possibly higher energy efficiency. Gen 3 was developed at the Pacific North West National Laboratory (PNNL) and uses a mixture of hydrochloric acid and sulfuric acid as a solvent for the vanadium ions . This increases temperature stability and energy density up to 50Wh/L. A disadvantage is the internal formation of chlorine gas in the system, which has a higher demand on the stability of materials. Vanadium/oxygen cells are referred to as Gen 4, in which vanadium ions are oxidized by oxygen (e.g. from the air) during discharge and energy can thus be generated. The process is reversed in the charging process. Theoretically up to 150Wh/L energy density can be achieved . Today’s VRFBs range from a few KW/KWh to several hundred MW/MWh. The applications range from home storage to industrial plants as large storage units in the grid. Several companies sell VRFBs today in different size classes.
Commercialization and ongoing research
From the 2000s onwards, the number of scientific publications and commercialization efforts for RFB types other than those mentioned here increased significantly. Basically, there are many possible combinations for inorganic RFBs. By 2015, there were about 78 different types of RFB, but only a few of them have or will ever have commercial relevance . Of the many different combinations, Pb/Pb- and Zn/ Ce- and H/Br-RFBs were the most studied in addition to the ones mentioned above [24, 25, 26]. The major challenges for new aqueous inorganic RFBs are above all the electrochemical window limited to a maximum of 2.1V, in which the redox pairs can function largely without hydrogen and oxygen formation, side reactions of the redox pairs with the solvent water and costs of the active materials. Especially the limitation of the voltage by the use of water led to investigations on the use of non-aqueous alternative solutions and redox pairs. The maximum possible voltage correlates with the maximum possible energy of an RFB. Doubling the voltage to 4V, as with lithium-ion batteries, doubles the maximum possible energy density. However, it should be noted that the cell resistance significantly determines the real energy density and often only very low energy densities can be achieved with very high cell resistances. Liu et al at the University of Michigan trialed the approach of using organic solvents to increase the energy density for the first time in 2009 with a vanadium-based organic acetylacetonate complex in acetonitrile . Cell voltage versus aqueous inorganic VRFB increased from 1.6V to 2.2V.