Improvement of safe bromine electrolytes and their cell performance in H 2 / Br 2 flow batteries caused by tuning the bromine complexation equilibrium

(a) interact with perfluorosulfonic acid membranes leading to significant reduction of membrane conductivity and (b) they form a low conductive ionic liquid with polybromides, leading to high overvoltage if the formation happens at the electrode. In this work a solution to this problem is proposed by an excess addition of Br 2 to these electrolytes. The excess bromine leads to a permanent bromine fused salt phase in the tank. Bromine formed in the cell stays in the aqueous phase and bromine transfer between the two phases happens in the tank. Transfer of Br 2 without the transfer of [BCA] + cations exists between the phases, while [C2Py] + cations remain in the fused salt and do not influence cell performance. For the first time a posolyte capacity of 179.6 Ah L (cid:0) 1 based on 7.7 M hydrobromic acid with BCA is achieved compared to previous in- vestigations with e.g. 53.9 Ah L (cid:0) 1 .

Organic quaternary ammonium compounds act as bromine complexing agents (BCAs) to lower the vapour pressure of Br 2 and lead to safer bromine electrolytes [9][10][11][12][13][14][15][16]. However, the application of BCAs in the bromine half cell leads to a reduced conductivity of the deployed perfluorosulfonic acid (PFSA) membrane [17,18] and causes a decrease of electrolyte conductivity [18]. In this work the BCA is transferred into a second electrolyte phase outside of the cell, by influencing the BCA solubility equilibrium. It is intended to prevent contamination of the cell by [BCA] + cations while maintaining their effect in lowering bromine volatility.
Both processes lead to a maximum capacity utilization of 30% of a 7.7 M HBr/1.11 M [C2Py] + electrolyte (53.9 Ah L − 1 ) [18], compared to BCA-free posolytes (HBr/Br 2 /H 2 O) with maximum capacities of 179.6 Ah L − 1 [27]. The [BCA] + cation concentration shall be as low as possible in the bromine half cell during operation in order to prevent its negative effect on the cell performance [18].

Work plan
The focus of this work is to develop safe and powerful bromine electrolytes based on [C2Py]Br. In order to prevent [C2Py] + cations to exist in the aqueous phase of the positive bromine half cell, the ambition is to intervene in the solubility equilibrium of the bromine complexation in Eqn. (4). By introducing an excess amount of Br 2 at SoC 0%, nearly all of the [BCA] + cations shall be transferred into the fused salt phase, following Eqn. (4), resulting in a decrease of the [BCA] + cations concentration in the aqueous phase. This idea is implemented using an electrolyte with a high theoretical energy density of 179.6 Ah L − 1 [27]. The excess amount of Br 2 is available within the entire SoC range (Fig. 1a) and is not used for energy storage, but enables a passive and safe bromine storage during the cell operation. The approach is shown in Fig. 1b: In this work, three electrolyte mixtures with different amounts of bromine at SoC 0% (+0.00 M; +1.68 M; +2.51 M Br 2 ) (Fig. 1), which are hereafter referred as "series no. 1, 2 and 3" are investigated on composition and cell performance. Series no. 1 defines the standard case in literature [18] without Br 2 in the electrolyte at SoC 0%. The addition of Br 2 is intended to reduce the [C2Py] + (aq) concentration in series no. 2   (Fig. 1b) in the fused salt phase. During battery operation, the range of absolute Br 2 concentrations shifts to higher bromine concentrations (Fig. 1a) or higher ratio between absolute bromine amounts and absolute [BCA] + amounts (Fig. 1b) in the electrolytes.
The absolute concentrations of HBr and Br 2 based on the SoC and the excess concentrations of Br 2 c(Br 2 , excess) in the electrolyte are defined in Eqn. (5) and Eqn. (6).
The interference in the equilibrium according to Eqn. (4) is comprehensively investigated for all performance-relevant parameters in the bromine half cell of a H 2 /Br 2 -RFB with a PFSA membrane (Nafion117): (i) concentration of [C2Py] + in the aqueous solution, (ii) bromine concentration and polybromide distribution in the aqueous phase, (iii) conductivity of the aqueous solution, (iv) influence of the parameters in (i, ii and iii) on cell performance and (v) change in the composition of the fused salt. The research article discusses three different bromine transfer mechanisms between the two electrolyte phases in the results.

Experimental and methods
A detailed version of the chapter "Experimental and methods" is available in the Electronic Supplementary Information (ESI).

Chemicals and composition of the electrolytes
Electrolyte solutions consist of HBr, water and Br 2 . The used BCA [C2Py]Br is synthesised from Refs. [13,45,46]. The synthesis of [C2Py] Br is confirmed by 1 H NMR and 13 C NMR analysis by comparison with literature [46,47] (ESI). In all electrolyte samples 1.11 M (absolute) [C2Py]Br is available within the whole SoC range. The absolute concentrations of HBr related to the SoC of the electrolyte are defined as follows: SoC 0% with 7.7 M HBr and SoC 100% with 1 M HBr (Eqn. (5)). The concentrations of Br 2 in the electrolytes depend on the chosen excess amount of Br 2 at SoC 0% in series no. 1, 2 and 3 and are shown in Table 1 and calculated for any SoC and excess amounts of Br 2 following Eqn. (6).

Concentration of [C2Py]Br in aqueous electrolyte solution
Concentrations of [C2Py] + (aq) in equilibrated aqueous phase are investigated by Raman spectroscopy from aqueous electrolyte samples for all three electrolyte series and at all 13 SoCs. Details on the Raman spectrometer equipment are shown in the ESI.
[C2Py] + leads to a strong single Raman peak signal at a Raman shift (ν) ν = 1029 cm − 1 [48][49][50]. Comparison of the peak area with the peak area of a reference electrolyte leads to [C2Py] + (aq) concentrations. Detailed information is shown in the ESI.

Bromine concentrations in aqueous electrolyte solution and fused salt
Bromine concentrations in the aqueous electrolytes are determined for all three electrolyte series and chosen SoCs by means of linear chronoamperometry at a rotating disk electrode on a vitreous carbon electrode. Starting from the open circuit potential the potential is reduced to − 1.0 V vs. Ag/AgCl/KCl(sat.) with a scan rate of − 40 mV s − 1 . Current response shows a constant limiting discharge current for low potentials, which is proportional to the Br 2 concentration by Levich equation [51][52][53][54][55]. Concentration of Br 2 are calculated from limiting currents following Eqn. c(Br 2 (aq)) = m⋅I lim Concentrations of Br 2 in fused salt c(Br 2 (fs)) are calculated from c (Br 2 (aq)) and volumes of aqueous phase V(aq) and of corresponding fused salt V(fs) shown in the ESI.

Fractions of Br 2 in different polybromides
The actual polybromide composition in the electrolyte phases is

Conductivities of aqueous electrolytes
Electrolytic conductivities of aqueous electrolyte solutions (series no. 1 to 3 at all chosen SoCs) are determined in a calibrated conductivity cell at ϑ = 23 ± 1 • C by impedance spectroscopy. Conductivity calculations and the determination of the cell constant are explained in the ESI.

Cycling tests and H 2 /Br 2 -RFB test cell
Cell tests are performed by galvanostatic cycling with a current density of ±50 mA cm − 2 or ±100 mA cm − 2 in a H 2 /Br 2 -RFB single cell to evaluate cycling performance of the three electrolytes. Starting from a SoC 100% the cell test operates with a discharge process followed by a charge process. SoC values for the electrolytes after discharge operation and after charge operation are calculated from the cell current and the discharge/charge time for each cycle by summing up the converted charge Q during the experiment (Eqn. (8)). The SoC after charge or discharge SoC (t 2 ) bases on the SoC (t 1 ) before charge or discharge operation, and the converted amount of charges Q (Eqn. (8)). At the beginning of the experiments SoC (t 1 ) is 100%: A Nafion117 membrane with a geometrical active area of 40 cm 2 and coated single-sided with a Pt/C catalyst for the H 2 reactions is used in the cell. The Nafion117 has a thickness of 175 μm in its dry state [60,61] and is applied due to its good proton transport property [62,63]. Due to its thickness, it reduces the crossover of water, polybromides and bromine, as well as H 2 [4,64], and is also used for that reason in other gas/liquid systems such as vanadium-air fuel cells [65,66] or methanol fuel cells [67,68]. Since it is coated with Pt catalyst on one side, it cannot be pre-treated in H 2 O 2 /H 2 SO 4 solutions under boiling, as is the case of Nafion membranes used in a liquid/liquid-RFB like the all-vanadium RFB [60,61] and is only placed for 24 h in pure water. A graphite felt electrode is embedded in a flow frame of the positive half cell. Further material and cell information are provided in the ESI. The hydrogen half cell is operated in a non-recyclable flow through mode with a flow rate of 100 mL min − 1 (dry) during charge and discharge operation. The aqueous electrolyte is pumped continuously through the positive half cell felt with a flow rate of 30 mL min − 1 , while the fused salt remains in the tank. The two phase electrolyte is stored in a sealed glass tank. The cell voltage E Cell i ∕ = 0 under operation, the redox potential of the electrolyte φ(Br 2 /Br − ) redox versus the normal hydrogen electrode (NHE), half cell potentials of the positive bromine half cell φ(Br 2 /Br − ) i ∕ = 0 vs. NHE and the negative hydrogen half cell potential φ(H + /H 2 ) i ∕ = 0 vs. NHE are determined in parallel during cycling experiments. The setup is shown in the ESI. Between discharge/charge operation during the cycling test the ohmic cell resistances are investigated by means of galvanostatic electrochemical impedance spectroscopy (EIS).

Reduction of [C2Py] + cations in the aqueous electrolyte phase
To achieve a transfer and storage of [C2Py] + into the fused salt phase within the entire SoC range, an excess amount of Br 2 is added and should shift the complexation equilibrium (Eqn. (4)) to the fused salt side, causing a large fraction of Br 2 and also of [C2Py] + to pass from the aqueous phase into the fused salt.
[C2Py] + (aq) concentrations in aqueous electrolytes are investigated for chemical equilibrium between the two phases by means of Raman spectroscopy and presented in Fig. 2.
For all electrolyte series, the concentrations decrease with increasing SoC, which is related to the rising absolute Br 2 concentration based on the electrochemical half cell reaction (Eqn. (1) For series no. 3, c([C2Py] + ) is below 5 mol% referred to the maximum possible concentration of 1.11 M for 95% of the electrolytes' operation range (SoC 5-100%), which is close to the objective of a [C2Py] + -free aqueous phase and expected to enhance cell performance.

Bromine concentrations and safety of electrolytes
For reaching high current densities, normally large Br 2 concentrations are required in the bromine half cell, but in parallel a safe electrolyte bases on low amounts of Br 2 in the aqueous phase. A compromise would be to reach moderate Br 2 concentrations during the operation within the SoC range. The Br 2 concentration in the aqueous phase is influenced due to the solubility equilibrium (Eqn. (4)). Br 2 concentrations for all three electrolyte series within the entire SoC range are investigated from equilibrated aqueous phases and depicted in Fig. 3a and Table 2.
In general, Br 2 concentrations increase in the aqueous solution within the entire SoC range (Fig. 3a) for series no. 2 and no. 3 due to rising Br 2 concentrations and the excess of Br 2 compared to the electrolyte samples without excess of Br 2 in series no. 1. The concentrations in no. 2 and 3 are much higher than in no. 1 as the excess amounts of Br 2 are not bound strong enough due to a lack of [C2Py] + (aq) to achieve Br 2 concentrations of series no. 1 with c(Br 2 ) ≤ 0.28 M. Maximum bromine concentrations c(Br 2 ) max are reached at specific SoC between SoC 70% and SoC 90% and have values between 0.28 and 1.26 M Br 2 ( Table 2, column 2 and 3). However, despite the large concentrations of Br 2 up to 1.26 M in the aqueous phase of series no. 2 and 3, more than 74 mol% of the Br 2 passes into the fused salt related to absolute bromine concentrations as shown in Table 2 in column 4 and 5.
No definition for neither maximum concentration of Br 2 nor their vapour pressures concerning safety parameters of the system are mentioned in the literature so far. This is why measurements in this study are based on concentrations only. However, large fractions of the Br 2 are bound in fused salt and the [C2Py] + cations are mostly transferred into the fused salt. The BCA complies the target to reduce the Br 2 concentration in the aqueous phase. It remains to investigate if these electrolytes can still be classified as "quasi-safe", which depends on accident scenarios, which shall be discussed elsewhere.

Bromine solubility equilibrium at high SoC
Polybromides and [C2Py] + cations form the fused salt phase due to the solubility equilibrium (Eqn. (4)). However, for SoC >70% and depending on the electrolyte series, it possible to investigate that the solubility equilibrium Eqn. (4) does not sufficiently describe the processes of Br 2 -transfer between the aqueous and the fused salt from the It is recognized, that from the SoC with highest concentration Br 2 (aq) max (Table 2) to SoC 100% the solubility of Br 2 is limited in the aqueous phase for all series. Br 2 concentrations decrease in this SoC range (Fig. 3a) while the absolute Br 2 concentrations still rise (Fig. 1a). In parallel all aqueous electrolyte phases in this SoC range are free of [C2Py] + (aq). Br 2 is transferred into the fused salt, leading to falling concentration of Br 2 in the aqueous phase, but a transfer according to Eqn. (4) is not possible to take place.
For high SoC, the Br 2 passes into the fused salt, as its solubility in the BCA-free aqueous phase is limited, which is not the case in the fused salt. From SoC >70% the concentration of HBr(aq) are lower than 3.01 M and not sufficient to form polybromides from all Br 2 molecules in aqueous solution. In parallel the fused salt absorbs Br 2 from the aqueous phase. The shown solubility effect is comparable to BCA-free electrolytes in this SoC range, as it has been discussed in Ref. [27]. Therefore, Br 2 reaches saturation in the aqueous phase and is forced to pass directly into the fused salt following Eqn. (9).

Polybromide distribution in the aqueous phase
The influence of the excess bromine on the distribution of Br 2 on the polybromides in the aqueous phase is further investigated, since individual polybromides are involved in equilibrium Eqn. (4) and shown in  [27]. Due to nearly independent polybromide distribution an influence of the polybromide distribution on further parameters cannot be investigated.

Storage of Br 2 in the fused salt phase
As the concentrations of Br 2 in the aqueous phase are low compared to the absolute Br 2 concentration, Br 2 is transferred in majority to the fused salt. In order to gain deeper knowledge about the mechanisms of bromine transfer, the concentration of Br 2 (fs) and the distribution of Br 2 among the individual polybromides in the fused salt phase are considered. Results are shown in Fig. 4.

Br 2 concentration and storage capacity in the fused salt phase
The fused salt phase acts as the actual bromine reservoir and energy storage media as Br 2 is present in the fused salt phase at a fraction >74 mol% (Table 2). Due to the transfer of Br 2 and [C2Py] + into the fused salt (Eqn. (4)), high concentrations of both components are reached, as shown for Br 2 in Fig. 4a. Concentrations of Br 2 in the fused salt between 9.26 ≤ c(Br 2 (fs)) ≤ 15.11 M are rather high compared to those in aqueous electrolytes (Fig. 3a) and even higher than absolute Br 2 concentrations (Fig. 1a). Concentrations of Br 2 in the fused salt tend to increase for all series no. 1 to 3 with rising SoC as observed in Fig. 4a. High concentrations of Br 2 are reached in the fused salt phase due to the absence of water [11]. The fused salt is a anhydrous ionic liquid consisting of a mixture of [C2Py] + cations and polybromides [11]. Raman spectroscopy showed the existence of these three polybromides.
Due to concentrations of Br 2 in fused salts larger than 10 M Br 2 theoretically, capacities larger than 536 Ah L − 1 are achieved related purely to the fused salt volume.

Distribution of Br 2 among polybromides
The existing polybromides in the fused salt phase are increasingly enriched by Br 2 with rising SoC as can be seen from distribution of Br 2 on the polybromides in the fused salt (Fig. 4b). The increasing Br 2 concentration in the fused salt leads to varying distribution of bromine in polybromides, resulting in an increase of higher polybromide amounts from Br 3 − to Br 5 − to Br 7 − . For all fused salt phases, a strong increase of Br 2 fraction in Br 7 − along the SoC is identified in Fig. 4b The fused salt polybromides take up Br 2 by forming higher polybromides with increasing absolute Br 2 concentrations, leading to a storage media of high capacity.

Transfer mechanisms of bromine between aqueous phase and fused salt phase
The transfer of bromine between the aqueous phase and the fused salt phase has been described so far with the solubility equilibrium Eqn. (4), while polybromides and [C2Py] + cations form micelles, which collapse and form a second phase during charge operation. For SoC >70% a further solubility limitation (Eqn. (9)) is introduced. In parallel to these two mechanisms, a further transfer mechanism for Br 2 into the fused salt and vice versa exists, by considering the concentration of Br 2 in the aqueous phase (Fig. 3a), the concentration of [C2Py] + in the aqueous phase (Fig. 2) and the Br 2 distribution among the polybromides in the fused salt phase (Fig. 4b): For series no. 2 (Fig. 2). Fused salts of series no. 2 and 3 are expected only rarely to absorb Br 2 from SoC ≥40% following Eqn. (4). In parallel, the actual bromine concentration in the aqueous phase increases with rising SoC (Fig. 3a), but is stoichiometrically too low, compared to the expected increase of absolute bromine (Fig. 1a). Therefore, Br 2 needs still to be absorbed by the fused salt either when [C2Py] + (aq) is absent. In the fused salt, with rising SoC and absolute Br 2 concentration more Br 5 − and Br 7 − are formed (Fig. 4b).
Therefore, Br 2 still passes the liquid-liquid interface between the two different electrolyte phases. A further transfer mechanism of Br 2 is proposed, which is independent of [C2Py] + (aq) cations and is compared to the mechanism in Eqn. (4).

Two main bromine transfer mechanisms
Next to the solubility limit shown in Eqn. (9), two main types of bromine transfer mechanism (mechanism I and II) between the two phases appear. Both mechanisms are described in Table 3. Mechanism I characterizes the transfer of the polybromide salt into the fused salt phase and vice versa caused by its low solubility (Eqn. (4)) in the aqueous phase (Table 3/column 1), which has been investigated in the literature for Zn/Br 2 -RFB electrolytes [12,15].
Mechanism II describes a transfer of Br 2 at the phase interface between the aqueous and the fused salt phase (Table 3/ Both transfer mechanisms I and II in Table 3 are reversible and take place in parallel along the SoC range. The presence of one or both

Application oriented preparation of BCA containing electrolytes
The amount of [C2Py] + present in the aqueous solution influences the predominating transfer mechanism of Br 2 and should be as small as possible to gain high cycling capacity ranges. When approx. 95 mol% of [C2Py] + cations are stored in the fused salt phase, we assume that the transfer of Br 2 by mechanism II predominates and expect cycling is possible with only limited influence of [C2Py] + (aq) cations on the cell performance. For simple preparation of electrolytes by electrolyte producers and battery operators, a concentration based parameter R is Table 3 Description of the possibly existing mechanisms of Br 2 transfer between the aqueous electrolyte phase and the fused salt phase, including criteria for the appearance of the individual mechanisms and effects on operation in a bromine half cell. introduced ( Eqn. (11)).
Eqn. (11)) defines the electrolyte mixture for a BCA-retention in the fused salt of at least 95 mol%. From the presented concentration of [C2Py] + (aq) in series 2 and 3 (Fig. 2), it can be defined, that the proportion between the absolute amount of Br 2 n(Br 2,absolute ) and the absolute amount of [BCA] + cations n([BCA] + , absolute ) is R. The BCAretention parameter R is defined in Eqn. (11): In series no. 2 for SoC ≥30% and series no. 3 for SoC ≥5% less than 5 mol% of [C2Py] + are dissolved in the aqueous electrolyte leading to R = 2.41 (series no. 2 and 3). R should be higher than R ≥ 2.4, when working with an excess of bromine and the ambition of achieving a high useable electrolyte capacity.
Battery performance with improved electrolyte mixtures no. 2 and 3 is evaluated in Section 3.6 in order to validate the criterion of R ≥ 2.4.

Electrolytic conductivity of the aqueous phases
Due to fast electrochemical reaction kinetics in both half cells, high current densities in the cell are feasible [3,74], requiring high electrolyte conductivities during cell operation. Due to the low conductivity of the fused salt [15,18], only the aqueous phase is applicable. The electrolyte conductivities are investigated for all aqueous electrolytes at ϑ = 23 ± 1 • C and are compared in Fig. 5. In addition, conductivities of pure HBr/H 2 O electrolytes [75] (orange line) and of BCA-free HBr/Br 2 /H 2 O posolytes [27] (green line) are depicted in Fig. 5.
High conductivities between 344.9 ≤ κ ≤ 781.2 mS cm − 1 are measured across all SoCs and aqueous electrolyte series (Fig. 5). High proton concentrations in aqueous solution with c(H + ) ≥ 1 M lead to high conductivity values by means of the Grotthus mechanism [76,77]. The conductivities of series no. 2 and 3 correspond approximately to the conductivity values of pure HBr/H 2 O solutions at ϑ = 20 • C and reach a maximum conductivity due to the excess amount of bromine within the whole SoC range compared to series no. 1.
For SoC ≤66% there is a strong difference in conductivities between electrolyte no. 1 and series no. 2 and 3. The presence of large organic [C2Py] + cations in series no. 1 (κ = 471.3 mS cm − 1 , SoC 0%) decreases the conductivity strongly, while conductivities of the nearly BCA-free aqueous phases of series no. 2 (κ = 681.7 mS cm − 1 , SoC 0%) and no. 3 (κ = 727.2 mS cm − 1 , SoC 0%) remain at high values. The mobility of the organic cation is limited due to its larger hydrodynamic radius. In addition, charge transport by means of the Grotthus mechanism needs to bypass the larger cations. As the aqueous phases of series no. 2 and 3 are nearly BCA-free within the whole SoC range, maximum conductivities independent of the BCA are reached. The excess of Br 2 and rising Br 2 concentrations over the SoC result in an increased transfer of [C2Py] + in the fused salt phase and lead to higher conductivities of the aqueous phase.
For SoC ≥66% the conductivity of all electrolytes in series no. 1 to 3 decreases, in accordance with the values in literature [11,27], and is independent of the amount of excess Br 2 . The proton concentration decreases to 1 M at SoC 100% due to the cell reaction, leading to a falling conductivity. Since [C2Py] + cations are not present in the aqueous electrolyte in this SoC range, they do not influence the conductivity. But as [C2Py] + cations bind large amounts of Br 2 in the fused salt phase, they reduce the amount of polybromides in the aqueous phase. The conductivity in aqueous solution is increased compared to BCA-free HBr/Br 2 /H 2 O electrolyte (Fig. 5 -green line), as the Grotthus mechanism of protons is rarely influenced by polybromides.
The modification of the electrolytes by adding Br 2 at SoC 0% is of high interest for cell application due to rising conductivities in series no. 2 and 3 within the entire SoC range.

Characteristic cycling voltages and useable electrolyte capacities
Galvanostatic cycling tests for all electrolytes are performed in the H 2 /Br 2 -RFB single cell at ±50 mA cm − 2 . Cell voltages, positive half-cell potentials and redox potentials are presented in Fig. 6.
For series no. 1 without Br 2 excess, there is a strong influence of the [C2Py] + cations on the cell performance (Fig. 6a). During charge, a strong voltage peak is present and during discharge, the voltage decreases strongly, so that the lower voltage limit is reached rapidly and a maximum range between SoC 100% and SoC 63% can be cycled. Only a maximum of 37% of the capacity can be used. This is in agreement with results from Ref. [18]. The [C2Py] + cations lead to the formation of poorly conducting fused salt in the bromine half cell during charge and interact with the Nafion117 during discharge, resulting in high overvoltages.
By adding Br 2 to series no. 2 and 3, the both adverse effects of [C2Py] + cations on cell performance are largely limited for cycling of series no. 2 (Fig. 6b) and nearly eliminated for series no. 3 (Fig. 6c). In both cases [C2Py] + cations are mainly stored in the fused salt (Fig. 2) and therefore their drawbacks on cell performance are diminished (Table 3/mechanism II). From cell tests it is investigated, that for both electrolyte series it is possible to carry out complete discharge and charge cycles within the selected SoC range, starting from electrolytes with a SoC 100%. A useable capacity of 179.6 Ah L − 1 is achieved for the first time for this type of battery using a BCA for Br 2 complexation in the electrolyte and a PFSA based Nafion117 membrane.
For electrolyte series no. 2, complete discharge and charge cycles in the SoC range can be achieved, but still a strong influence of the [C2Py] + (aq) cations is evident (Fig. 6b) during charging process. An increase in cell voltage forms a peak, then flattens out again and demonstrates the presence of small amounts of [C2Py] + cations in the aqueous phase. In combination fused salt is formed in the electrode felt and causes an overvoltage in the half cell. The formation of fused salt is available but does not limit the useable electrolyte capacity. The small but rapid voltage and potential fluctuations in Fig. 6 are due to an air Results of cycling test with electrolyte series no. 3 (Fig. 6b) are only slightly effected by low concentrated [C2Py] + cations as more than 95% of the [C2Py] + cations are bound in the fused salt phase. The shape of cell voltage curve is similar with the voltage trends of BCA-free electrolytes from literature [27,78]. Due to the low amount of [C2Py] + (aq) in the half cell the formation of low conductive fused salt is supressed. As no overvoltage is caused during charge, the cell voltage does not tend to show a peak. During discharge the cell voltage decreases stronger than the bromine half-cell potential, but ends at high values of 0.6 V. Full discharge of 179.6 Ah L − 1 is reached. The redox potential of the Br 2 /Br − electrolyte is in parallel to the bromine half cell potential during charge and discharge.
Since at SoC 0% for series no. 2 and 3 there are concentrations of bromine of 0.24 M and 0.44 M, respectively, no mass transport limitation in the bromine half cell occurs during the discharge process when approaching SoC 0%. This is evident by the bromine half-cell potential, that does not decrease while the cell voltage decreases strongly (red vs. black lines in Fig. 6b and c).
Energy efficiency (EE) for the first three cycles of electrolytes of series no. 2 (EE = 60.0%) are lower than for series no. 3

Membrane performance
In order to determine the influence of [C2Py] + cations in the electrolytes on the membrane resistance, which is part of the ohmic cell resistance, the ohmic cell resistance of the cell before and after the discharge process is investigated using EIS and shown in Fig. 7. Before the first cycle the ohmic cell resistance R OHM is 1.10-1.25 Ω cm 2 in all cells. For all cell tests with different electrolyte series, R OHM is higher   Fig. 6a-c. after discharge. From this results it is expected, that during discharge [C2Py] + is released from fused salt into the aqueous phase and interacts with the PFSA membrane [18]: An average value for R OHM after discharge operation of R OHM = 3.48 Ω cm 2 (series no. 1), R OHM = 2.86 Ω cm 2 (series no. 2) and R OHM = 1.96 Ω cm 2 (series no. 3) is calculated from values in Fig. 7. Although all series show an increased ohmic cell resistance, this effect strongly depends on the selected series. During discharge, higher concentrations of [C2Py] + are available for series no. 1 and no. 2, while for series no. 3 less than 0.07 M [C2Py]Br are dissolved in aqueous solution. As the electrolyte conductivity is high for discharged electrolytes at SoC 0% (Fig. 5) and the ohmic cell resistances of further cell materials are assumed to be constant, the rise in cell resistance is most probably associated with the reduced conductivity of the Nafion117 membrane in contact with [C2Py] + cations. The interaction of the cations with the membrane is not reversible during charge [18].
Even low amounts of large organic [C2Py] + cations in solution interact with the sulphonate groups of the Nafion117 membrane, forming an addition bonding and hindering the transport of protons and water through the membrane structure [43,79]. This phenomenon in the H 2 /Br 2 -RFB is described in detail in Refs. [17,18]. The organic [C2Py] + cation dries the membrane of water by forming hydrophobic spaces in its porous structure [79]. The membrane conductivity decreases, and R OHM rises, as shown in Fig. 7.

Efficiencies for long-term cycling and different current densities
In order to test the stability of the electrolytes, long-term cycling tests are performed for electrolytes no. 2 and 3 at a current density of ±100 mA cm − 2 . Cell voltage and mass change of the bromine electrolyte are recorded (Fig. 8a+b). Long term cycling of series 1 was carried out and discussed in detail in Ref. [18].
The long term cycling is performed for 9 cycles only, as the mass of electrolyte decreases strongly during the tests (Fig. 8/a-b). The mass of the electrolyte decreases independently of the electrolyte series by approximately 40 g (29.8%) throughout 9 cycles. At the outlet of the hydrogen half cell, clear but strongly acidic liquid is collected from the cell in a washing bottle. There is a strong crossover of the individual components. This is mainly caused by a crossover of water and Br 2 into the hydrogen half cell such as has been determined for series no. 1 in Ref. [18]. The crossover leads in parallel to an intense release of Br 2 and [C2Py] + cations from the fused salt into the aqueous phase [18]. Thus, the lack of Br 2 in the aqueous electrolyte caused by the crossover of Br 2 into the hydrogen half cell is compensated. Simultaneously, [C2Py] + cations are released and diffuse into the PFSA membrane and interact with the sulphonate groups. The membrane conductivity is reduced with increasing cycle number by an increasing release of [C2Py] + . Ohmic cell resistances rise due to rising membrane resistances.
For series no. 2, the release of [C2Py] + from the fused salt at ±100 mA cm − 2 and the interaction with the Nafion117 membrane lead to strongly increasing ohmic cell resistances and in the discharge process the lower voltage threshold is reached without reaching the complete electrolyte to be discharged (Fig. 8a). For the selected current density, the minimum SoC continues to increase after discharge (Fig. 8a). A complete discharge and thus a cycling of the capacity of 179.6 Ah L − 1 is not possible. This effect is intensified by the crossover or mass loss of the electrolyte. For series no. 3 a complete discharge operation is possible within the first four cycles. For the further cycles, the same effect occurs as for series no. 2, but with a delay and with higher depths of discharge (Fig. 8b).
For the application of this idea using an access amount of Br 2 , it is therefore necessary to use stronger BCAs which are bind more strongly in the fused salt even when Br 2 is released from the fused salt. For example, [C2Py]Br could be replaced by 1-n-hexylpyridin-1-ium bromide, which is stronger interacting with polybromides [11]. However, the crossover must be reduced in later experiments by higher pressures of the hydrogen gas in the hydrogen anode.
Despite strongly decreasing useable capacities, the energy efficiencies of the battery remain relatively constant (58.5-61.9% for series no. 2 and 64.6-72.5% for series no. 3 in Fig. 8c). However, the efficiencies are not very meaningful, as they remain essentially constant for each series. In contrast, the amounts of energy discharged during the discharge process (Fig. 8c) for the different series are significant. They decrease due to the decreasing useable capacities (Fig. 8c). However, the excess bromine in series no. 3 improves the discharge energy density (93-133 Wh L − 1 ) compared to series no. 2. However, due to the increasing influence of the [C2Py] + cations, the discharge energy density decreases strongly within the 9 cycles. Furthermore, energy efficiencies for all electrolytes are investigated at higher current densities (±50, ±75, ±100 and ± 125 mA cm − 2 ) and are measured in the 2nd cycle in each case. The EE (Fig. 8d) for series no. 1 (40.1 ≥ EE ≥ 33.2%), series no. 2 (62.9 ≥ EE ≥ 55.2%) and series no. 3 (74.7 ≥ EE ≥ 60.1%) increase with increasing excess amounts of Br 2 , suggesting better energy efficiency in aqueous [C2Py] + -free series no. 2 and 3. With increasing current density, the energy efficiencies decrease slightly (Fig. 8d). Based on the results in sections 3.1 to 3.6.3 and in Ref. [18] we suggest the following explanation: The slight decrease in EE is attributed to the higher overvoltages of the ohmic cell resistance of the cell at higher current densities. On the one hand (1) the current in the cell increases, which leads to higher overvoltages and at the same time (2) the ohmic resistance increases due to the release of [C2Py] + cations. For higher discharge currents Br 2 is consumed faster in the positive half cell and a faster release of Br 2 from the fused salt into the aqueous phase is expected, causing also an increased transfer of higher amounts of [C2Py] + . Higher concentrations of [C2Py] + in the aqueous phase can interact with the membrane. Charge and discharge times for higher currents shorten and thus the useable capacities. In series no. 1, mass transport inhibition due to a lack of Br 2 in front of the electrode as a limiting step predominates with increasing current density, while for series no. 2 and 3, the ohmic resistances of the cell and membrane as a limiting step predominate for the current densities investigated.

Conclusions
The selective investigation of the bromine electrolyte with BCAs using the example of [C2Py]Br allowed the development of safe and efficient electrolytes for bromine half cells in H 2 /Br 2 -RFB. By adding an excess amount of Br 2 it is possible to actively interfere into the BCAsolubility equilibrium in order transfer large parts (>95 mol%) of the [C2Py] + from the aqueous phase into the second available fused salt phase and to bind it there.
By focusing on the distribution of Br 2 in the fused salt polybromides and the concentration of [C2Py] + and Br 2 in the aqueous phase, it was found that a transfer of Br 2 happens in an absorption process at the liquid-liquid interface. When [C2Py] + (aq) is absent, in a second transfer mechanism only Br 2 was transferred from a polybromide in the aqueous phase to a polybromide in the fused salt phase. By adding an excess of Br 2 , the transfer by solubility limitation of the fused salt is suppressed, while the transfer mechanism without [C2Py] + allowed the transfer of Br 2 between the phases.
The adverse effects of formation of fused salt in the cell and low conductivities of the PFSA membranes in contact with organic [C2Py] + cations are reduced or eliminated, while Br 2 is still stored safely in the fused salt. While earlier measurements without the modification of the electrolyte showed limited cell performance and a useable capacity of only 53.9 Ah L − 1 , with this approach, a useable capacity of 179.6 Ah L − 1 is achieved for the first time including the application of a BCA and a PFSA membrane. Crossover of the electrolyte leads to reduction of this capacity. Despite a significant reduction of [BCA] + content in the aqueous phase, small amount of it remained dissolved, which lowers the membrane conductivity.
For future investigation, a stronger BCA than [C2Py]Br should be combined with excess amounts of Br 2 in the electrolytes to reduce both, Br 2 to moderate concentrations and transfer [BCA] + cations completely to the fused salt phase. Also a focus should be to develop a cell design to operate the cell with aqueous electrolytes at higher current densities and lower electrolyte crossover during cycling.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.