2.1.1. Mechanical Stress Test (MST)

For the evaluation of the stability and structural integrity of the electrode coating with different BPs, the electrode coins were first mechanically tested by applying a mechanical stress test (MST).

Interestingly, after the MST, the aqueous electrode coatings are washed off and dissolved in DI-water (see Table 1a–d) leading to a turbidity of the solution (see Figure 1a). For the electrolyte solution, the MST shows no dissolution of the coating. Only the PVP binder flakes off the current collector in the electrolyte solution, but still does not dissolve. In contrast, the non-aqueous electrode coatings are stable in both DI-water and electrolyte solutions (see Table 1e–g and Figure 1b). The results of the MST follow the observations seen in Chang et al. (2020) for the non-aqueous binder (here: PVDF+NMP) and the aqueous CMC binder [4].

**Figure 1.** Example side views of the beakers after the MST with an aqueous (here: PAN-based LA133, (**a**) and non-aqueous (here: CA (MEK), (**b**) electrode coating in DI-water, respectively. For the aqueous coatings, the MST leads to a dissolution of the coatings and a turbidity of the solutions (the stirring bar is still inside the beaker but not visible anymore), whereas the non-aqueous coatings show no damages.

(**a**) (**b**)


**Table 1.** Mechanical stress test of the electrode coatings with different binder polymers in DI-water and electrolyte (2 M ZnSO4 + 0.1 M MnSO4), respectively.

For the aqueous binders, the different behaviour in DI-water and electrolyte solution could be referred to the reduced solvation free energy of the water molecules in consequence of the ZnSO4 and MnSO4 addition in the electrolyte ([50] and Supplementary Figure S12), retarding and reducing the BP dissolution: Instead of a dissolution, the BP could only swell while still maintaining the structural integrity of the coating.

For the non-aqueous binders, as was expected, both the DI-water and the aqueous electrolyte solution are not affecting the structural integrity of the coating during the MST. Furthermore, in consequence of the non-aqueous processing of the coating, no swelling of the coating should occur here, in contrast to the previous aqueous binder coatings.

Based on the observations of the MST, a different behaviour during the cycling of the aqueous and non-aqueous electrodes in the aqueous ZnSO4/MnSO4 electrolyte is expected and shown in the following section.

#### 2.1.2. Rate Capability Test (RCT)

For the evaluation of the electrochemical performance of the electrode coatings with different BPs and solvents (see Table 2), a rate capability test (RCT) was performed under various current rates. The current rates are given in mA·g−<sup>1</sup> based on the active material mass loading. The RCT results are shown by plotting both the specific discharge (DCH) capacity as well as the specific discharge energy, together with the efficiency values, over the cycle number, which allows a more in-depth analysis of the effects of the different binder types.

The results of the RCT with specific capacities (see Figure 2a) shows significantly higher capacity values for all the aqueous binders compared to the non-aqueous binders.


**Table 2.** Overview of the slurry preparations with aqueous- and non-aqueous-based binder polymer (BP), active material (AM), and carbon black (CB).

**Figure 2.** Results of the RCT plotting the specific discharge (DCH) capacity (**a**) and energy (**b**) under variation of the current rate over the cycle number for the non-aqueous electrode coatings.

The aqueous PAN-based electrode shows the highest specific capacity for all current rates with ~150 mAh·g−<sup>1</sup> at 40 mA·g−<sup>1</sup> and ~90 mAh·g−<sup>1</sup> at 640 mA·g−1. Both the CMC+SBR and the CMC binders show lower capacity values, with the lowest values for the pure CMC binder. The better performance of the CMC binder with SBR addition could

be related to the better adhesion of the coating on the current collector, as well as a better flexibility of the coating through the addition of the SBR elastomer [51]. The PVP binder, which showed a flaking off the current collector in the MST, was not able to finish the RCT, which can be related to the low stability of the electrode coating (see RCT results in Supplementary Figure S6). Therefore, the PVP binder will not be discussed further in the following sections.

The non-aqueous binders show low-capacity values <20 mAh·g−<sup>1</sup> in the first cycles, but interestingly it is with an increasing trend, even if the current rate increases. This is especially prevalent during the last cycles with 80 mA·g−1, as the increasing discharge capacity trend becomes clearly visible. This behaviour can be related to the surface deposition/dissolution of MnO2 as the predominant reaction mechanism for the non-aqueous electrode, whereby a growing MnO2 surface layer could be formed on the positive electrode [19,33,37,40,41]. The lower overall discharge capacity of the non-aqueous binders compared to the aqueous binders could be related to the observations of the MST results of the previous section, where a swelling and wetting of the electrodes in the electrolyte was suggested for the aqueous but not (or to a lesser extent) for the non-aqueous electrode coatings. In consequence, the MnO2 active material loading of the electrode coating could only become available for the aqueous binders by dissolution/deposition processes during charging/discharging, leading to additional capacity. This suggestion is confirmed by the latest literature for ARZIBs [19,33,37,40,41]. For the non-aqueous binders, the active material loading could mainly be isolated by the binder.

The coulombic efficiencies (CE, see Figure 2a) for all the binder types are located above or at ~100%, which again can be related to the MnO2 deposition/dissolution process. The dissolution of the MnO2 active material loading as well as the additionally deposited MnO2 could lead to an additional discharge capacity, which results in CE values >100%. The CE values for the non-aqueous binders are generally located above the values for the aqueous binders, indicating a more dominant MnO2 deposition/dissolution mechanism. This could be related to the inactivity of the isolated MnO2 loading within the positive electrode coating, as no swelling of the binder is suggested (see MST section), which could enable a material deposition/dissolution only on the surface of the electrode with better efficiency values.

The RCT results showing the specific energy (see Figure 2b) additionally reveal the effects of overpotentials during charge/discharge by also considering the potential value (besides the capacity value, see the equation in Supplementary Figure S5) of the particular cycle. As the energy output of a battery system is of even higher importance for the application than only the capacity output, this way of plotting can be regarded as being a more application-oriented plot. For the specific energy plots, the comparison of the aqueous/non-aqueous binder types show the same characteristics as the specific capacity plots. Nevertheless, distinctions in terms of the energy efficiency (EE) values are noticeable: the EE values decrease for increasing current rates. This indicates an increasing inner resistance and overpotential of the cell [52]. For the non-aqueous binders, the energy efficiency values are generally located above the values of the aqueous binders, as seen before in the capacity plot, again indicating a more dominant MnO2 deposition/dissolution mechanism.

In Figure 3, the potential curves for the 10th and 30th cycle, respectively, each with the current rate of 80 mA·g−1, are shown for the non-aqueous binders. Generally, the shape of the discharge curves of all binder/solvent combinations show two potential plateaus divided by a potential bend (see exemplarily Figure 3, marker 1). The charge curve again shows two potential plateaus (see Supplementary Figure S11 for more details). The potential curves of the non-aqueous binders (see Figure 3b) each have a significant increase in the specific capacities compared to the aqueous binders (see Figure 3a). Nevertheless, the potential curves characteristics stay the same, as the characteristic potential bend for ARZIBs (with MnO2 active material) indicates (see Figure 3, marker 1). The potential curves also show that the capacity increase from the 10th to the 30th cycle is the highest for

both of the PAN-based electrodes (aqueous and non-aqueous), which could explain the good overall performance of the aqueous PAN-based electrode in the RCT (see Figure 2a).

**Figure 3.** Overview over the potential curves of the aqueous (**a**) and non-aqueous (**b**) binder polymers for the 10th and 30th cycle, each with the current rate of 80 mA·g<sup>−</sup>1.

By performing SEM+EDX characterizations (see Figure 4), the cross-sections of electrode coatings of PAN (aq) and PAN (DMSO), prepared by ion-polishing, were compared in a pristine state and in discharged state after the RCT procedure (post-mortem state). This way, the influence of an aqueous and non-aqueous BP could be investigated in more detail.

For both the pristine PAN (aq) and (DMSO) electrodes, the MnO2 particles (coloured green) and the CB content (coloured orange) show comparable morphology characteristics such as the homogeneity, the porosity, and the material distribution (see Supplementary Figure S7 for EDX images of the single elements, respectively).

Interestingly, after the RCT procedure (post-mortem) in the discharged state, both electrodes show a surface layer deposition, which is thicker for the PAN (aq) electrode and could be attributed to a MnO2 deposition (see Figure 4d,h, marker 1), as described in the previous literature [19,33,37,40,41]. Furthermore, both electrodes in a post-mortem state (discharged state) show a new flake structure inside the electrode coating (see Figure 4c,d,g,h, circle), which is not visible in the pristine state (see Figure 4a,b,e,f). This could be attributed to ZHS precipitations filling the pores of the electrode coating, as described in the previous literature [14,19,33,37,40–44]. However, this precipitation seems to be less distinctive for the PAN (DMSO) coating. This could be explained by the previously mentioned assumption of less wetting of the electrode coating by the electrolyte and thus a smaller active inner surface, as the CB-containing, electrically conductive coating could mainly act as an electrochemically active surface. Hence, this could result in a lower capacity by the dissolution of the (mostly unavailable) MnO2 active material loading during discharge, compared to the PAN (aq) coating. The flake structure (see Figure 4c, marker 2) on the surface of the electrode can be referred to the separator with electrolyte salt residues, which was stuck on the electrode coating surface (possibly due to the aforementioned MnO2 deposition).

**Figure 4.** SEM and EDX images of the PAN (aq) electrode in a pristine (**a**,**b**) and post-mortem (**c**,**d**, discharged state) state, and the PAN (DMSO) electrode in a pristine (**e**,**f**) and post-mortem (**g**,**h**, discharged state) state. The MnO2 particles are coloured green, and the carbon black is coloured orange, respectively. For the post-mortem state, the PAN (aq) and PAN (DMSO) both show a MnO2 surface deposition, but with a thinner layer thickness for the PAN (DMSO) electrode.

Altogether, the comparison of the SEM+EDX images in a pristine and post-mortem (discharged) state could confirm the existence of a MnO2 deposition during the charge steps, which takes place both in the pores of the electrode coating (if available, depending on the binder/solvent combination, as previously discussed) as well as on the surface of the electrode. In the following discharge steps, the previous MnO2 deposition dissolves again, releasing capacity, which should be higher for those electrodes whose MnO2 active material loading is available for the deposition/dissolution mechanism (here: PAN (aq) electrode coating, see RCT results). Due to efficiency reasons of the dissolution, MnO2 deposition residues seem to stay, leading to an accumulating MnO2 (surface) deposition layer (here: thicker surface layer for PAN (aq) than for PAN (DMSO) coating).

For a better understanding of the MnO2 deposition/dissolution mechanism, further SEM+EDX characterizations of cross-sections of the positive electrode coatings in different states of charge should be carried out to further prove the latter assumptions of this study.

To underline the findings of the RCT tests for the aqueous and non-aqueous electrodes, the electrical impedance spectroscopy (EIS) measurements of the positive electrode (vs. Zn/Zn2+ reference) were analysed for the PAN (aq) and PAN (DMSO) binder in a charged (CH) and discharged (DCH) state to compare the effects of the different solvent categories as an example (for all the EIS results, see Supplementary Figure S8). The choice of these two binder combinations allowed for a direct comparison of the influence of the non-aqueous solvent (DI-water vs. DMSO) without the influence of the binder polymer (here, the *LA133* binder is considered as a PAN-based binder since the specification of PAN in the safety data sheet is >95 wt% of solid content). In Figure 5, the EIS results of the positive electrode are shown.

**Figure 5.** Results of the EIS measurements of the positive electrode for charged (CH, 1.7 V vs. Zn/Zn2+) and discharged (DCH, 0.8 V vs. Zn/Zn2+) state following both the 80 mA·g<sup>−</sup>1-RCT-steps after 10 and 30 cycles, respectively.

207

For the different PAN binder solvents and state of charges, the following qualitative interpretations of the Nyquist plot can be made:


Altogether, the comparison of the aqueous/non-aqueous electrodes generally shows higher discharge capacities/energies for the aqueous binders, which is related to the availability of the MnO2 active material loading. The MnO2 active material loading is not available in the non-aqueous electrodes, as the electrolyte should not lead to an electrode coating swelling as suggested, in contrast, for the aqueous electrodes. Nevertheless, the non-aqueous electrodes show increasing discharge capacities/energies with proceeding cycles, underlining the major role of MnO2 deposition/dissolution processes (also on the surface of the electrode) for the ZIB with ZnSO4/MnSO4-based electrolytes being introduced in the recent literature.

As the aqueous PAN-based *LA133* binder shows the best cycling results, this binder was chosen for the mixing ratio variation shown in the following section.
