A joint research team from Shanghai University and the Qinghai Institute of Salt Lakes of the Chinese Academy of Sciences has recorded a high discharge capacity of 156.2 Ah per liter of positive electrolyte in a redox flow battery with solid manganese dioxide (MnO₂) added to the electrolyte. A redox flow battery charges and discharges by circulating electrolytes stored in tanks through a cell. In this study, the team increased the amount of electricity that can be extracted by using not only materials dissolved in the electrolyte but also solid MnO₂. However, under the conditions that produced the maximum capacity, capacity dropped to 38% of the initial baseline after 50 charge-discharge cycles, so combining high capacity with long life remains a challenge. The results are presented on Shanghai University's official website.
The research, by Xin Liu, Zhang Chen, Daiqi Zhou and colleagues, was published online in the peer-reviewed journal Nano-Micro Letters on August 29, 2026. It appears in Volume 19 for 2027, with the DOI 10.1007/s40820-026-02349-8. However, what the original paper reports are measurements from a small laboratory cell, and the 156.2 Ah/L figure is based only on the volume of the positive electrolyte. It is not a figure for the energy density of the battery as a whole.
A small flow battery records 156.2 Ah per liter of positive electrolyte
A defining feature of redox flow batteries is that the cell where the electrochemical reactions take place and the tanks that store the electrolyte can be designed separately. Typically, the positive and negative electrolytes, separated by a membrane, are each circulated by pumps. In this design, increasing the amount of electricity stored per unit volume of electrolyte could help make stationary storage systems more compact. The National Institute of Advanced Industrial Science and Technology (AIST) also explains the characteristics of flow batteries, but this study has not demonstrated by what percentage tanks could actually be made smaller.
The team prepared a semi-solid slurry by adding commercially available MnO₂ particles to a positive electrolyte containing manganese sulfate. A slurry is a flowable mixture of solid particles dispersed in a liquid, and this approach does not require the addition of conductive carbon. The base electrolyte used 1 M MnSO₄, 0.5 M Fe₂(SO₄)₃, 0.1 M sodium pyrophosphate and 3 M sulfuric acid. M denotes molar concentration (mol/L), and the amount of solid MnO₂ added is also expressed in moles per liter of liquid.
In the experiments, the graphite felt electrode area was 4 cm² or 25 cm², and 15 mL or 20 mL of positive electrolyte was used. The membrane separating the electrolytes was Nafion 212, the circulation flow rate was 80 mL per minute, the upper charge voltage limit was 1.1 V and the lower discharge voltage limit was 0.1 V. On the negative side, the redox reaction between Fe³⁺ and Fe²⁺ was mainly used, with a larger volume of negative electrolyte than positive electrolyte.
Without MnO₂, a discharge capacity of 38 Ah/L was obtained at an electrode area of 4 cm² and a current density of 30 mA/cm². By contrast, with 4 M MnO₂ added and the electrode area enlarged to 25 cm², capacity reached 156.2 Ah/L. However, because the electrode area was changed along with the amount of MnO₂ added, the increase in capacity cannot be attributed entirely to MnO₂. Also, 30 mA/cm² is a current per unit of electrode area, not a statement that the absolute current was the same in both conditions. The paper reports a capacity above 151 Ah/L and an energy efficiency above 71% at 30 mA/cm², and a capacity of 125 Ah/L even when the current density was raised to 40 mA/cm².
How poorly soluble MnO₂ is put to work in the battery reaction
In conventional systems that store electricity using only materials dissolved in the electrolyte, the solubility of those materials limits capacity. In this study, the team aimed to use the manganese contained in the solid as well, by converting the solid MnO₂ added to the electrolyte into Mn²⁺ in solution during discharge. In fact, after two charge-discharge cycles, unreacted solids were separated by centrifugation and the concentration of soluble manganese in the liquid was measured by inductively coupled plasma (ICP) analysis; with 4 M MnO₂ added, it reached 3.76 M. This result supports the idea that adding solid MnO₂ increased the manganese available in solution. However, measuring soluble manganese concentration alone does not clarify the entire reaction pathway.
The team also analyzed the MnO₂ generated after charging using X-ray diffraction and transmission electron microscopy to examine structural changes. The specific surface area of commercial MnO₂ was 12.1 m²/g, whereas the MnO₂ produced by the electrochemical reaction showed 46.3 m²/g, about 3.8 times higher. When each was stirred for two hours, followed by solid-liquid separation, the capacity obtainable from the supernatant was 0.72 Ah/L for commercial MnO₂ and 2.58 Ah/L for the generated MnO₂. This indicates that MnO₂ produced electrochemically more readily yields components in the liquid that can take part in the reaction.
The team further performed electrochemical measurements of the relationship between current and potential, impedance measurements, and comparisons of ultraviolet-visible absorption at about 500 nm. Based on these results, they suggest that a "reverse disproportionation reaction," in which Mn³⁺ forms from Mn²⁺ and Mn⁴⁺, is involved. Specifically, Mn³⁺ arises in the liquid from the MnO₂ generated by charging, and the reaction then proceeds between Mn³⁺ and Mn²⁺. However, this is a reaction pathway the authors proposed to explain multiple experimental results, and it must be evaluated separately from the observed capacities and structural changes themselves.
Combining high capacity with long life remains a challenge
Even if adding MnO₂ increases the electricity that can be extracted in a single discharge, there is no guarantee that performance can be maintained over the long term. The team repeatedly charged and discharged the cells while varying the MnO₂ concentration under common conditions of a 25 cm² electrode area and a current density of 30 mA/cm². The results shown in Figure 5 of the original paper are as follows.
| MnO₂ concentration | Coulombic efficiency (CE) | Energy efficiency (EE) | Capacity retention | Cycles measured |
|---|---|---|---|---|
| 1M | 99.0% | 76.0% | 100% → 59% | 100 |
| 2M | 99.3% | 73.9% | 100% → 50% | 150 |
| 3M | 99.0% | 71.9% | 100% → 51% | 50 |
| 4M | 98.2% | 71.1% | 100% → 38% | 50 |
Capacity retention is the percentage relative to the baseline capacity of each condition, set at 100%. The absolute baseline capacity (in Ah) for each condition is not listed individually in the main text of the paper.
These results show that the ability to maintain capacity differs greatly depending on the MnO₂ concentration. Under the 4 M condition, which gave the maximum capacity, capacity fell to 38% of baseline after 50 cycles. By contrast, the 1 M condition retained 59% after 100 cycles and the 2 M condition retained 50% after 150 cycles. The results suggest that adding more MnO₂ to obtain a high discharge capacity makes capacity loss during repeated use a problem.
It is worth distinguishing among Coulombic efficiency (CE), energy efficiency (EE) and capacity retention. CE is the proportion of the electric charge put in during charging that can be recovered in discharge, and EE is the proportion of the energy put in that can be recovered. Capacity retention, on the other hand, shows how much of the baseline discharge capacity is maintained after repeated cycling. Under the 4 M condition, CE was high at 98.2%, yet capacity retention after 50 cycles was only 38%. In other words, high charge-discharge efficiency in each individual cycle does not guarantee that capacity will be maintained over a long period.
However, the number of cycles measured differs between conditions, and no variation among measurements from independent cells is shown. For that reason, these results alone cannot be used to rank lifespan strictly by additive concentration or to judge statistical significance.
Some clues to the cause of capacity loss have also emerged. Under the 1 M condition, the manganese concentration in the positive electrolyte fell from 1.83 M to 1.15 M after 100 cycles. In the initial state, there was a concentration difference across the membrane of 1.83 M versus 0.67 M, and the team points to the possibility that manganese crossing the membrane and particle aggregation contribute to the capacity loss. However, a single cause has not been identified. Improving the membrane and using dispersants have been proposed as countermeasures, but for now they remain subjects for future research.
Whole-battery energy density and practicality need to be verified
The reported figure of 156.2 Ah/L represents the electric charge that can be extracted per unit volume of positive electrolyte. But to evaluate the energy a battery can store, discharge voltage must be considered as well as charge. Ah/L, which expresses charge, and Wh/L, which expresses energy, are different metrics. Moreover, an actual storage system needs negative electrolyte as well as positive electrolyte, so a figure based only on the positive electrolyte cannot be treated as the energy density of the entire battery system.
The team also ran a test with 1 M MnO₂ added in which the negative electrode reaction was changed from the iron-based Fe³⁺/Fe²⁺ to the vanadium-based V³⁺/V²⁺. As a result, discharge voltage at a current density of 30 mA/cm² rose from 0.71 V to 1.45 V, and the discharge energy per liter of positive electrolyte increased from 48.2 Wh/L to 97.1 Wh/L. This suggests that changing the negative-side reaction could increase the energy that can be extracted. However, this result was obtained under the 1 M MnO₂ condition, and it would not be appropriate to combine it with the 156.2 Ah/L of the 4 M condition to calculate an energy density that was not measured.
On cost, the paper gives a figure of US$33/kWh for a 4 M electrolyte, estimated using a model by Darling and colleagues. However, this is not a measured manufacturing cost, nor does it represent a selling price for a storage system. Because equipment costs for cells, membranes, pumps and so on are not included, the economics of a commercial storage system cannot be judged from this figure alone.
The results were also obtained in a small laboratory cell. The main text does not give the number of independent cell tests, the number of repeated measurements or statistical errors, nor does it state the specific temperature during cell testing. No third-party replication has been confirmed so far, and it is unknown whether the same performance can be maintained in larger devices or how much variation the measurements would show.
In the paper's conclusion, the research team itself states that extending cycle life should be a priority for future work. The significance of this result lies in showing, through the use of solid MnO₂, a possibility of high capacity that differs from conventional approaches limited by solubility. At the same time, if that capacity cannot be maintained after many charge-discharge cycles, the advantages as a practical storage system will be limited. The next focus for commercialization will be whether long-term capacity retention and reproducibility can be verified, and whether performance of the whole system, including the volumes and voltages of both positive and negative electrolytes, can be clarified.