Expand the effective carriers in solid polymer electrolytes via anion-hosting cathode

The non-reactive anion migration deteriorates the limited ionic conductivity of the solid polymer electrolytes (SPEs) and accelerates solid-state batteries failure. Here, we introduce an integrated approach in which polyvinyl ferrocene (PVF) cathode encourage anions and Li + to act as effective carriers simultaneously. The concentration polarization and poor rate performance, caused by insucient effective carriers, were addressed by the participation of anions in electrode reaction. Specically, the PVF|Li battery matched with unmodied SPE (PEO-LiTFSI) showed 107 mAh g − 1 initial capacity at 100 µA cm − 2 and maintained 70% retention for more than 2800 cycles at 300 µA cm − 2 and 60°C. Moreover, the slight capacity decrease at 1000 µA cm − 2 and the successful batteries operation at minimal ionic conductivity (8.13×10 − 6 S cm − 1 ) show that the current carrying capacity of SPEs was greatly improved without complex design. This strategy weakens the strict requirements for ion conductance and interface engineering of SPEs, and provides an ecient scenario for constructing advanced polymer-based all-solid-state batteries.

anion migration 22,23 . By constructing polymer segments with weak interaction with Li +24 or grafting anions to the polymer backbone 25 , SISPEs can achieve high t Li+ (>0.9). However, this comes at the cost of ionizing less free Li + and also the low ionic conductivity due to the lack of solvation ability.
The participation of anions in electrode reaction has promoted the development of attractive energy conversion systems in the liquid or extended gel phase [26][27][28] . These systems were intended as low-cost alternatives to lithium-ion batteries, focus on some key parameters such as sustainability and material availability 29,30 . Apart from the requirement for high electrochemical stability, the electrolyte of the anion reaction system is similar to that of lithium-ion batteries. The bene t of the effective carrier expansion is hidden by the abundant ionic conductivity of the liquid/gel electrolyte. However, for solvent-free SPEs with limited ion movement, the strategy of enhancing the correlation between ion migration and electrode reaction is expected to play a more crucial role. In this work, we introduced anion-hosting cathode to overcome the negative impact owing to non-reactive anion migration in SPEs. The ferrocene unit, anchored to the long-chain polymer, encourage anions as the effective charge carrier similar to Li + . The expansion of carriers signi cantly improves current carrying capacity of unmodi ed SPEs (Fig. 1), and avoids the short-circuit failure lead by concentration polarization. Besides, the impact of anion species on ion mobility and interacting with the cathode is also investigated in depth.

Materials design and characterization
The electronic structure of the cyclopentadiene and iron atom hybrid orbital provides ferrocene with stable and reversible redox properties (Fig. 2b) 31,32 . To avoid the diffusion of active units in a non-ow system, we anchored the anion-hosting unit to the polymer chain by free-radical polymerization of vinyl ferrocene, as illustrated in Fig. 2a. The EDX mapping proves the homogeneous distribution of Fe element across the PVF. The polymerization process was con rmed by the Fourier transform infrared spectroscopy (FT-IR) in Fig. 2c, which shows a signi cant weakening of double bond vibration peaks at 1625 cm -1 after polymerization. The gel permeation chromatography (GPC) results (Table S1) prove that PVF has a high molecular weight (~4800 g mol -1 ) with wide distribution (M w /M n = 1.71). The redox of PVF can provide a theoretical capacity of 124 mAh g -1 (Fig. 2d), locating at the top level of the anionhosting organic cathode 33,34 . Moreover, the theoretical redox potential of ~3.45 V vs. Li + /Li can be tolerated by most SPEs. The thermogravimetric analysis (TGA) in Fig. S1 shows that PVF did not undergo signi cant thermal weight loss or phase change below 300 °C, ensuring the electrode stability in the case of high temperature operation.
Classic SPEs are composed of lithium salt and polymer with solvation ability. Bene ting from the high dielectric constant and chain exibility, PEO is one of the most widely studied polymers matrix 10,11,35 .
Based on the well-designed anion hosting material, the PVF|Li battery matched with PEO-LiTFSI electrolyte exhibits excellent cycle stability. It maintained 70 % capacity retention after 2800 cycles ( Figure 2e) at 60 °C, while avoiding battery failure for more than 4000 cycles (Figure 2e, 2f). Considering the current case, where the migration of anions and cations is related to the electrodes' reaction, we set several types of lithium salt in SPEs to obtain a deep insight into the anion electrode reaction. The ionic conductivity of the SPEs was conducted through AC impedance (seen in Fig. S2). Fig. S3 plots the ionic conductivity as a function of temperature. Owing to the high delocalized negative charge of anions, PEO-LiTFSI displays the highest ionic conductivity (3.53×10 -4 S cm -1 at 333 K). Except for LiClO 4 , other electrolytes have comparable conductivity at high temperature. The differential scanning calorimetry (DSC) results of SPEs correspond to the ionic conductivity, where PEO-LiTFSI has the lowest melting point (Fig. S4).
Ine cient utilization of ions in SPEs severely restrict the batteries performance. To explore the ion mobility in uence in the designed systems, we measured the lithium-ion transference number (t Li+ ) of SPEs through the steady-state current method (results are shown in Fig. S5, Table S2). As seen in Fig. 3a, the electrolytes with LiTFSI, LiFSI and LiClO 4 as salts pose low t Li+ which are all-around 0.1, proving that anions contribute the most ion movement in these SPEs. Note that the calculated t Li+ of PEO-LiBOB electrolyte is negative (-0.38) differ from other threes. In fact, negative cation transfer numbers are not rare, and usually attributed to the presence of ionic aggregates 36,37 . As shown in Fig. 3b, while the negatively charged ion clusters dominate the charge transfer in electrolyte, the t Li+ of SPEs can be low to negative due to the short-range interactions 38,39 . The domination of ion clusters could reduce the carrier loading capacity at high current density, which affects the battery performance, as will be discussed later. The AC impendence results of PVF|Li batteries shown in Fig. 3c proves that negatively charged clusters signi cantly increase the batteries' charge transfer resistance with PEO-LiBOB electrolytes, despite the high ionic conductivity compared to other SPEs. Generally, for a typical Li + -hosting cathode, SPE with low t Li+ operate poorly. However, the anions participation in the electrode reaction breaks through the strict requirement on ion aggregation in the SPEs 20 . The PVF|Li batteries assembled with SPEs all exhibit reversible charge-discharge process shown in Fig. 3d, and the overpotential of batteries is controlled by the ion conductivity of SPEs. Moreover, the anion species lead a difference on operating voltage of batteries. This nding prompted us to further study of anion impact.

Anion impact on electrode reaction
The electrochemical stability of the SPEs was evaluated using Li|SPE|stainless steel cells. The linear scanning voltammetry (LSV) curves in Fig. S6 shows a low redox current within a voltage window up to 4 V vs. Li + /Li, ensuring that the electrode reaction is not disturbed by electrolyte oxidation. The electrochemical behavior of PVF with anions was evaluated through cyclic voltammetry (CV) tests. As shown in Fig. 4a, the PVF cathode shows good redox reversibility provided by the active unit. Notably, the stabilization of the unit into polymer engages a crucial role in the electrode reaction reversibility. With the free ferrocene electrode (Fig. S7), even with a high content of the conductive agent, the assembled solidstate battery exhibits oxidation peak only in the rst cycle. This indicates that the ion pair formed by ferrocenium and anion cannot undergo further reduction. In contrast, PVF with its long-chain structure prevents the diffusion of active materials by anchored ion pairs into the polymer. The peak potential separation (ΔE) shown in CV results is informative of the electrochemical reaction kinetics. Judging from Fig. 4c, the highest ionic conductivity SPE, PEO-LiTFSI, has the lowest potential gap (∆E = 0.103 V), while PEO-LiClO 4 has the highest ∆E (Fig. 3d), corresponding to the ionic conductance trend.
The in uence of anion species on the electrode reaction in liquid electrolyte, related to anion couples, was previously investigated by Redepenning et al 40 . The results show that the ion pairs' formation could negative shift the electrode potential from theoretical 41,42 . To determine the ion pair effect on the electrode in SPEs, we conducted density functional theory (DFT) simulations to calculate the binding energy (BE) of different anions to the cathode. The cathode was simpli ed by substituting ethyl ferrocenium for the polyvinyl ferrocenium (Table S4). The calculations indicated that ethyl ferrocenium has the highest BE to ClO 4 -, and decrease with the order FSI -, BOB -, and TFSI - (Fig. 4b). property of PVF. Besides, the combination of PF 6 and cathode also shifts voltage plateau to a lower value (~3.27 V) than the theoretical, which can be explained by the combined effects discussed above.
Enhanced batteries rate performance and anode stability Motivated by the results that the batteries with PEO-LiTFSI and PEO-LiBOB electrolytes have small polarization and high capacity (Fig. 3d), we focused on these two SPEs for further studies of long cycles. After a few cycles of cathode activation in PEO-LiBOB electrolyte, the PVF capacities were measured to be 112, 104, and 107 mAh g -1 at currents of 20, 50, and 100 μA cm -2 , respectively (Fig. S10). The capacity has not been signi cantly attenuated after a long period cycling. However, as shown in Fig. S11, the battery with PEO-LiBOB electrolyte shows substantial polarization and signi cant capacity fade at higher current density (~67 mAh g -1 at 300 μA cm -2 ), which is as weak as the SPEs containing ClO 4 and FSI - (Fig. S12), despite the ionic conductance disparity among threes. The poor rate performance of batteries with LiBOB salt can be ascribed to the formation of ions aggregations demonstrated in the previous section. The sluggish electrode reaction deteriorates the battery performance at high current densities.
Bene t from high ionic conductivity, large anion structure and the ability to prevention aggregation, the battery with PEO-LiTFSI showed excellent cycle stability and rate performance (Fig. S13, S14).
Speci cally, the initial capacity reaches 97 mAh g -1 at 300 μA cm -2 and maintained 90 % after 1000 cycles (Fig. 2e). Bene ting from the extended carriers, the battery exhibits good rate performance, with capacities of 108, 107, 97, and 94 mAh g -1 at current densities of 100, 200, 300, and 500 μA cm -2 , respectively (Fig. 5f). Even when the current increases to 1 mA cm -2 , which is intolerable by many reported advanced SPEs, the battery still maintains 78 mAh g -1 capacity with a stable plateau. The PVF electrode exhibits pseudo-capacitance, which also contributes to the excellent rate performance to a certain extent (Fig. S15). It is worth noting that the electrode reaction rate is not the rate-determining step of most solidstate device owing to the limited ion transport capability of solid electrolytes. Despite using a large amount of the conductive agent to overcome PVF low conductivity, the result of the control experiment proves that the capacity contribution of the conductive agent at the tested voltage range is negligible (Fig.  S16). The decrease of conductive agent proportion reduces the battery's rate performance signi cantly (Fig. S17). However, this does not affect the effectiveness of the strategy proposed in this report. The extended carrier makes the ionic conductivity of SPEs no longer a bottleneck of the battery's performance at high current loading. The addition of single-walled carbon nanotubes in the electrode can signi cantly increase the proportion of active materials with the reduced polarization (Fig. S18), which provides massive space for the high conductive anion-hosting electrode.
Concentration polarization could induce serious consequences in SPEs when only Li + act as effective carriers, signi cantly accelerate anode degradation and battery failure especially at high current density (Fig. 5b). Taking the charging process as an example, the Li + -hosting cathode (e.g., LiFePO 4 ) undergoes an anodic reaction. In an ideal situation, the ion number in SPEs keep constant throughout the process.
The concentration of Li + increase on the cathode side and decrease on the counter. This leads to a salt concentration gradient and ion diffusion barriers (Fig. S19a). The SPEs can be regarded as liquid electrolyte with extremely high viscosity. The minimal ion diffusion results in a stable massive concentration gradient, which is more severe than that in liquid. In the present work where anions were involved in energy storage, the ions and ions clusters migration under electric eld is similar to that of LiFePO 4 . The difference between two cases is that the PVF electrode reaction consumes anions similar to Li + on the anode in SPEs (Fig. 5a). Therefore, the distribution of SPEs salt concentration is homogeneous, facilitating the ions' diffusion and avoiding the detrimental consequences of the concentration polarization (Fig. S19b).
To verify that minimizing the concentration polarization could effectively enhance the lithium anode stability, we tested LiFePO 4 |PEO-LiTFSI|Li batteries under the same condition (Fig. S20). As shown in Fig.   5c, 5e, a micro short circuit occurred in LiFePO 4 |Li at the 24th hour at 300 μA cm -2 . The short circuit has not been repaired in subsequent cycles, leading to a continuous decayed in coulombic e ciency (Fig.  S20b). Similar failures are common in other polymer-based solid-state batteries, even for SPEs with improved ionic conductivity and interface properties. In contrast, the coulombic e ciency of the PVF|PEO-LiTFSI|Li battery maintained about 99.7 % with no short circuit observed over 4000 cycles at 300 μA cm -2 ( Fig. 5c, d), which is the maximum level for most polymer-based solid-state-batteries. The excellent cycle stability proves the success of our concept in controlling the polarization and suppressing lithium anode deterioration.

Performance with insu cient ionic conductivity
As an unmodi ed SPE, the combination of PEO-LiTFSI does not have advantages in terms of ionic conductivity and interfacial properties, usually performed as negative comparison. However, the expansion of the effective carrier through the anion-hosting cathode eliminates the demand of complex design for advanced SPEs. We compared the rate performance of this work with the reported advanced SPEs in Fig. 5g 35,[44][45][46][47][48][49][50][51] . The strategy we proposed strongly enhances capacity retention at high current density with simple PEO-LiTFSI without any modi cations, better than other SPEs of complex design (details in Table S5).
The improvement of the current-carrying capacity of SPEs inspired us to examine the battery performance with very low ionic conductance. PEO-LiTFSI exhibits extremely limited ionic conductivity of 2.65×10 -7 S cm -1 at 30 °C, far from the usual battery test requirement. In this condition, PVF|Li batteries exhibited capacities of 83, 60, 48 mAh g -1 at 10, 30, and 50 μA cm -2 with increased polarization, respectively (Fig 6c). The inferior ionic conductivity leads the unsatisfactory performance. As an ideal additive, succinonitrile (SN) could strengthen the segment movement ability of polymer thereby enhancing SPEs' ionic conductivity 52,53 . After doping PEO-LiTFSI with 5% SN, the ionic conductivity increased to 8.13×10 -6 S cm -1 at 30 °C (Fig. S21a). Therefore, the battery assembled by PEO-LiTFSI-SN electrolyte show a lower impedance compared with the electrolyte without plasticizer (Fig. S21b). The CV curves exhibit stable and reversible redox performance (Fig. 6a). As shown in Fig. 6b, 6c, the recorded capacities at 10, 30, and 50 μA cm -2 rise to 94, 86, and 70 mAh g -1 , respectively. The PVF|Li batteries with these two SPEs both maintain more than 100 cycles at 30 μA cm -2 without signi cant capacity decayed. The anion-hosting cathode makes full use of each dissociated ion in electrolytes, resulting in a battery system with high tolerance to SPEs with very low ionic conductivity. In short, this strategy avoids plenty problems devoted to the low ionic conductivity and utilization faced by previous reported SPEs 5,6,23 (Fig.   6d), reinforcing the correlation between ion movement and electrode reaction.

Discussion
In summary, we developed advanced polymer-based solid-state batteries by inducing anions as effective carriers simultaneously with Li + . The anion-hosting cathode PVF put the entire ion movement of SPEs into energy storage, which produces an updated rate performance and promotes batteries operation at very limited ionic conductance. In addition, the ultra-stable cycles of PVF|Li batteries prove that the anode deterioration, mainly contributed by concentration gradients, were avoided effectively by reactive anion migration, which is essential in building safer metal anode batteries. Besides, experiments and theoretical calculations clari ed the effects of anion structure, binding energy and ion aggregation on battery performance. This work provides a pioneering strategy for the design of advanced solid-state energy storage systems. However, future investigations into high capacity and conductivity anion-hosting cathode are certainly warranted. Since the migration and aggregation of anions differ from Li + in most SPEs, the disparity of reaction status between cathode and anode could present some minor complications and addressing this issue could be the subject of the following study.

Synthesis of polyvinyl ferrocene and solid polymer electrolytes
Polyvinyl ferrocene (PVF) was synthesized by free-radical polymerization. In a typical process, vinyl ferrocene was dissolved in dry toluene, and AIBN was used as the initiator. The ratio of Polyethylene oxide and lithium salts were dissolved in acetonitrile. The obtained solution was coated on a polytetra uoroethylene plate and dried under reduced pressure to obtain a self-supported lm. The electrolyte membrane was kept in Ar atmosphere glove box to prevent moisture contamination. The solid electrolyte containing succinonitrile adopts the same preparation method.

Materials characterization
The thermal gravimetric analysis was carried out with METTLER TOLEDO TGA/DSC 3+ at a temperature range of 30-800 °C under nitrogen atmosphere, with a heating rate of 10 K min -1 . Fourier transform infrared (FT-IR, Bruker Tensor 27) were recorded between 400 and 4000 cm -1 . A eld emission scanning electron microscope (FE-SEM, SU-6600) equipped with an energy dispersive spectrometer was used to characterize the samples' morphology. All the electrochemical characterization was performed using the electrochemical workstation (PARSTAT 1000 and CHI 660E).

Electrochemical measurements
The active material (PVF), conductive agent (Super P) and binder (PVDF) were dispersed in N-methyl pyrrolidone at a ratio of 4: 5: 1 or 6: 3: 1. The dispersed slurry was coated on aluminum foil and dried under reduced pressure at 60 °C. The prepared electrodes were stored in a glovebox until use. The same method was used to prepare LiFePO 4 electrodes with a ratio of 7:2:1 and the Super P electrode in a ratio of 85:15 (conductive agent: binder). The mass loading of the active material was controlled at 1.0 mg cm -2 . The electrodes were cut into discs (d = 12 mm) for subsequent testing. The cells assembly was carried out in a glove box lled with argon gas (H 2 O, O 2 < 0.1 ppm), using metallic lithium as the counter electrode.
The ionic conductivity of the electrolyte was measured by sandwiching the polymer electrolyte lm between two stainless steel electrodes and then record the electrochemical impedance. The ionic conductivity was calculated by the following equation l is the thickness of the polymer electrolyte, R is the bulk resistance of the polymer electrode, and S is the electrolyte area.     between this work and the other reported advanced SPEs, classi ed in the graph by design strategy (detail seen in Table S5).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.