Self-Recovery Chemistry and Cobalt-Catalyzed Electrochemical Deposition of Cathode for Boosting Performance of Aqueous Zinc-Ion Batteries.

Rechargeable Zn-ion batteries working with manganese oxide cathodes and mild aqueous electrolytes suffer from notorious cathode dissolution during galvanostatic cycling. Herein, for the first time we demonstrate the dynamic self-recovery chemistry of manganese compound during charge/discharge processes, which strongly determines the battery performance. A cobalt-modified δ-MnO2 with a redox-active surface shows superior self-recovery capability as a cathode. The cobalt-containing species in the cathode enable efficient self-recovery by continuously catalyzing the electrochemical deposition of active Mn compound, which is confirmed by characterizations of both practical coin-type batteries and a new-design electrolyzer system. Under optimized condition, a high specific capacity over 500 mAh g-1 is achieved, together with a decent cycling performance with a retention rate of 63% over 5,000 cycles. With this cobalt-facilitated deposition effect, the battery with low concentration (0.02 M) of additive Mn2+ in the electrolyte (only 12 atom % to the overall Mn) maintains decent capacity retention.

Introduction

Since their first commercialization in the 1990s, rechargeable lithium-ion batteries (LIBs), which are characterized by high voltage, high energy density, and long cycling stability, have been widely used as portable power source in many aspects of our daily life during the past three decades, in particular in the field of personal electronics (Lin et al., 2017, Du et al., 2019, Zhao et al., 2015). The state-of-the-art LIBs use organic-based electrolytes, which, however, can introduce serious safety concerns for large-scale applications, such as electrochemical energy storage for electric grid and power source for electric vehicles (Hu and Xu, 2014). In addition, the limited resource of lithium also brings a big concern about the cost of LIBs when their widespread application is realized. Therefore, battery systems with higher safety and lower cost are highly demanded (Lukatskaya et al., 2016, Ming et al., 2019, Zhong et al., 2019, Lu et al., 2019). Compared with conventional organic electrolytes, inorganic, aqueous electrolytes usually have improved ionic conductivity, higher safety, and lower fabrication cost (Lu et al., 2011, Fang et al., 2018). Aqueous batteries based on different metal ions, e.g., Li+, Na+, Zn2+, Mg2+, and Al3+, have been developed and intensively studied over the past decades (Luo et al., 2010, Li et al., 2015, Verma et al., 2019). In particular, zinc-ion (Zn-ion) batteries have attracted increasing attention owing to the abundance of Zn resource and the relatively high electrochemical potential of Zn metal in neutral or mildly acidic aqueous electrolytes, which is beneficial to the stable performance (Konarov et al., 2018, Song et al., 2018, Zeng et al., 2019). The concept and the fundamental chemistry of the rechargeable Zn-ion batteries based on the migration of Zn2+ between cathode and anode was first proposed by Kang and co-workers in 2011 (Xu et al., 2012, Li et al., 2019a). In recent years, intensive efforts have been made to the investigation of the composition of aqueous electrolytes, the preparation and protection of Zn anodes, and the development of high-performance and stable cathode materials (Zeng et al., 2019, Yang et al., 2019, Zhang et al., 2017).

Compared with metallic Zn anode, the capacity of the cathode is usually much lower, which becomes one of the bottlenecks for the commercialization of aqueous Zn-ion batteries. Till now, most of the research activities on Zn-ion batteries have been focused on the development and modification of the cathode. Transition metals with high valent states, e.g., Mn (Cheng et al., 2005, Xu et al., 2012, Alfaruqi et al., 2015, Pan et al., 2016, Pan et al., 2018, Zhang et al., 2016, Zhang et al., 2017, Huang et al., 2018, Wu et al., 2018), V (Kundu et al., 2016, Kundu et al., 2018, Wan et al., 2018, He et al., 2017, Yan et al., 2018, Wan and Niu, 2019), and Mo (Chae et al., 2016, Li et al., 2019b, Kaveevivitchai and Manthiram, 2016), are more suitable cathode materials for Zn-ion batteries as a bivalent ion battery system, and therefore they have been widely studied (Verma et al., 2019). Actually, Zn/MnO2 battery is one of the earliest systems for primary Zn-based batteries, which has achieved commercialization many years ago (Li et al., 2019a). Owing to the fast kinetics of Mn redox reactions, Mn compounds (e.g., MnO2) show promising capability as high-performance cathodes for rechargeable Zn-ion batteries (Cheng et al., 2005, Pan et al., 2016, Ha and Lee, 2016). In recent years, a variety of modification strategies (e.g., pre-intercalation, heteroatom doping, and defect engineering) have been proposed for improving the performance of the Mn-based cathodes (Wang et al., 2019a, Long et al., 2019, Xiong et al., 2019, Fang et al., 2019). For example, pre-intercalation of ions or molecules (e.g., Na+, Zn2+, polyaniline or water) in the layered MnO2 can boost the redox kinetics by facilitating the charge storage (Wang et al., 2019a, Wang et al., 2019b, Huang et al., 2018). Doping cathode materials with cobalt or nickel is beneficial to improve the energy density of the cathode by increasing the discharge voltage or improving the electronic conductivity (Long et al., 2019, Pan et al., 2018, Ma et al., 2019).

Despite significant advantages (e.g., high theoretical capacity, low cost, low toxicity), the notorious instability of the Mn-based cathodes during recharging cycles severely impedes their use in practical Zn-ion batteries (Nam et al., 2019, Fang et al., 2019). The instability could be mainly associated with the disproportionation of Mn(III), generating Mn(IV) and highly soluble Mn(II) (Song et al., 2018, Yu et al., 2019). It is important to note that for MnO2, regardless of its initial phase structure (e.g., α-MnO2, β-MnO2, γ-MnO2 or δ-MnO2), this charge disproportionation is very likely to happen, leading to the same result of the dissolution and loss of active Mn compound and, consequently, deterioration of the capacity. To address this issue, additional Mn2+ at high concentrations (usually 0.1–0.5 M) was proposed to be introduced into the electrolyte to suppress the disproportionation reaction (Pan et al., 2016, Wu et al., 2018, Huang et al., 2018, Huang et al., 2019a, Huang et al., 2019b, Zhang et al., 2017, Fang et al., 2019). The cycling stability of Mn-based cathodes has been greatly enhanced by adopting this strategy. However, this enhancement is at the expense of reduced mass energy density of the battery, given that the large proportion of the extra Mn2+ occupies additional weight but might not provide decent capacity.

To date, mechanisms behind the enhanced stability of Zn-ion batteries with Mn2+-containing electrolytes are still unclear. Recent reports have provided some clues that Mn2+ in the electrolyte can be oxidized, inducing the deposition of active Mn compound on the cathode surface (Gibson et al., 2016, Li et al., 2019c). This suggests that the deposition of Mn compound from the Mn2+-containing electrolytes is also an important phenomenon that deserves more research attention. Such a self-recovery phenomenon resulting from the deposition of Mn compound on the cathode may have a significant effect on the capacity and stability of Zn-ion batteries. However, most of the previous reports evaluated the capacity only based on the weight of the Mn-based cathode. In addition, only the Mn2+ concentration in the electrolyte was given, whereas information about the volume of the electrolyte and the Mn ratio (atomic/weight) of the solid Mn-based cathode to the aqueous Mn2+ in the electrolyte is lacking. This could lead to an unreliable comparison of the results across different publications. Owing to the potential contribution of Mn2+ in the electrolyte to the capacity, retention rates of the cathode much higher than 100% were observed in some literature reports where the capacity was calculated just based on the active material at the cathode (Wu et al., 2018, Fu et al., 2018, Zhu et al., 2018, Wang et al., 2019b).

All these facts emphasize the motivation of this work, i.e., to study self-recovery chemistry in the cathode of Zn-ion batteries, including the influencing parameters for the process and their interactions and the mechanism behind the electrochemical deposition of Mn compound during the battery cycling. Based on the understanding of the self-recovery chemistry in the cathode, we further propose a cobalt-modified δ-MnO2 as a cathode material with a redox-active surface for fast deposition of Mn compound. To emphasize the importance of the deposition rate of the Mn-active compounds for the self-recovery cycling capacity, an equation that describes the dynamic capacity of the Mn-based cathode is also demonstrated. The cobalt-induced facilitating of deposition is confirmed by the electrochemical and chemical characterizations of both practical coin-type batteries and a newly designed electrolyzer system. Insights into the catalysis routes of the Mn compound deposition are also revealed. This work provides guidance of the utility of the Mn2+-containing electrolyte for the development of Zn-ion batteries and may provide inspiration for materials engineering for future functionalized cathodes, especially for other conversion-type battery cathodes, which have similar dissolution issues.

Results and Discussion

Self-Recovery Chemistry in MnO2 Electrodes of Zn-Ion Batteries

Figure 1A presents the galvanostatic cycling stability of aqueous Zn-ion batteries working with several different MnO2-based cathodes. Here, the capacity was calculated based on the mass of the initial weight of solid MnO2 on the cathode. For fair comparison, parameters including (1) the mass loading of cathode (area: 0.5 cm2, areal loading of MnO2: ∼2.0 mg cm−2), (2) the concentration of MnSO4 additive (0.2 M), and (3) the volume of electrolyte (200 μL for each battery) were well-controlled for all batteries. The δ-MnO2 cathode synthesized via a molten-salt-assisted method (see Transparent Methods) showed an initial discharge capacity of 252 mAh g−1, and reached 317 mAh g−1 after a five-cycle activation process (Figure 1A). Unlike most of the previously reported works, it is interesting that the discharge capacity of the δ-MnO2 cathode kept increasing during subsequent cycling at a current density of 300 mA g−1, and reached a discharge capacity of 403 mAh g−1 after 100 cycles. α-MnO2, another cathode material widely utilized in Zn-ion batteries, was also evaluated for its cycling stability. Although showing better discharge capacities in the first five cycles (first discharge: 325 mAh g−1 and fifth discharge: 339 mAh g−1), the α-MnO2 cathode exhibited inferior capacity retention to that of the δ-MnO2 cathode. It suggests that different cathodes delivered different electrochemical behaviors in Zn-ion batteries.

Figure 1

Comparison of Performance and the Self-Recovery Chemistry in a MnO2 Cathode

(A) Comparison of cycling performance of α-MnO2, δ-MnO2, and cobalt-modified δ-MnO2 cathodes with a 2 M ZnSO4 electrolyte containing 0.2 M MnSO4.

(B) Comparison of retention rates and Coulombic efficiencies of the MnO2 cathodes with or without 0.2 M MnSO4 in the electrolyte (the calculation of retention rate is based on the 10th discharging of the cathodes).

(C and D) (C) First five galvanostatic discharging and charging profiles of a cobalt-modified δ-MnO2 cathode and (D) the corresponding dQ/dV profiles from 1.55 to 1.80 V.

(E) A schematic and an equation for describing the dynamic capacity retention of a Mn-based cathode.

(F) Comparison of the fifth charging of the MnO2 cathodes; inset shows their corresponding dQ/dV profiles from 1.62 to 1.80 V.

(G) Comparison of the rate performance of α-MnO2, δ-MnO2, and cobalt-modified δ-MnO2 cathodes with a 2 M ZnSO4 electrolyte containing 0.2 M MnSO4.

The different capacity variation tendencies of these two cathodes point to a fact that is seldom noticed: The capacity of Zn-ion battery cathodes may not be simply determined, as previously reported, by the capability of suppressing Mn-based cathode dissolution. The intrinsic property of the cathode (e.g., the specific surface area, the valent state of Mn cations, and Mn-site defects) may result in the change of parameters that were not given sufficient attention, i.e., the oxidation of the Mn2+ ion in the electrolyte that can lead to the deposition of active Mn compound on the cathode. Considering the abundant defects in the δ-MnO2 cathode material as synthesized under an unstable, highly oxidizing atmosphere generated by the molten sodium nitrate salt (Hu et al., 2017), modification of δ-MnO2 was conducted using a facile method where the oxide was stirred in 1 M CoCl2 at room temperature (see Transparent Methods). Compared with the original δ-MnO2, the resulting cobalt-modified δ-MnO2 showed an even more pronounced capacity enhancement (Figure 1A), increasing from the initial discharge capacity of 196 mAh g−1 to 258 mAh g−1 after five cycles and further to 435 mAh g−1 at the 100th discharge. This capacity is surprisingly high when compared with the previously reported values (∼300 mAh g−1 in the 100th discharge) (Pan et al., 2016, Huang et al., 2018, Zhang et al., 2017).

The increment of the capacity could be highly associated with the oxidation of Mn2+ in the electrolyte, which was evidenced by the analysis of Coulumbic efficiency and the differential capacity (dQ/dV). All the cathodes showed high Coulumbic efficiency close to 100%, whereas the cobalt-modified δ-MnO2 showed slightly lower Coulumbic efficiencies during the recharging cycles (Figure 1B). Without the MnSO4 additive, the Coulumbic efficiency of all samples is similar and near 100% (Figures 1B, capacity data shown in S1). This indicates that the additional Mn2+ in the electrolyte was actually involved in the redox process on the cathode. To further investigate this issue, the first five discharge-charge profiles of the cobalt-modified δ-MnO2 cathode are presented in Figure 1C. Two conventional plateaus at potentials of 1.4 and 1.3 V versus Zn/Zn2+ reflect the H+ intercalation process and the Zn2+ intercalation process, respectively (Huang et al., 2018). In each subsequent charging process, an additional plateau/slope was observed at potential around 1.6–1.8 V, besides the two conventional plateaus that represent the reverse process of Zn2+ and H+ intercalation. In Figure 1D, differential capacity (dQ/dV) analysis more clearly shows this additional oxidation process happening at a voltage range of 1.62–1.80 V. This could be ascribed to the oxidation of Mn2+ to Mn3+, as will be discussed in the following sections. In a mild aqueous electrolyte environment, subsequent hydrolysis of Mn3+ could result in the deposition of Mn compound on the cathode (Gibson et al., 2016). These compounds were involved in the subsequent discharge, thus providing additional capacity.

Figure 1E gives a schematic and an equation that describe the dynamic capacity of a Mn-based cathode. Based on the fact that no matter what be the phase structures of the Mn-based cathode (e.g., α-MnO2, δ-MnO2), the discharge capacity of the cathode mainly originates from the redox of Mn species between different valent states (from +4 to +3 or even to +2). Therefore, the discharge capacity of a Mn-based cathode on a specific cycle is determined by three essential parameters and their interaction (Figure 1E): (1) the initial amount and valence of Mn in the solid Mn-based cathode before cycling, (2) the dissolution amount of the active Mn compound during each discharge/charge cycle, and (3) the deposition amount of active Mn compound during each charge. In these parameters, the initial valent state of the Mn element is determined by the cathode material itself, which has a significant influence on the first discharge capacity. On the subsequent cycling, the dissolution rate and deposition rate will negatively and positively, respectively, influence the amount of Mn on the cathode, thereby playing a more important role in the capacity retention. Based on this equation, developing a cathode with a low dissolution rate and a high deposition rate is crucial to the stable operation of Zn-ion batteries.

The synthesis of δ-MnO2 and its further modification by cobalt were illustrated in Figure 2A (details are provided in the Supplemental Information) (Hu et al., 2017). As indicated by the qualitative analysis of the Mn L-edge near-edge X-ray absorption fine-structure (NEXAFS) spectra (Cowie et al., 2010, Gann et al., 2016) (Figure 2D), cobalt-modified δ-MnO2 showed the lowest Mn valency among the three different MnO2-based cathodes. Therefore an expected lower initial capacity was observed for cobalt-modified δ-MnO2. For the batteries without any additional MnSO4 in the electrolyte, the dissolution rate is reflective of the capacity retention. As shown in Figure 1B, the dissolution rate of cobalt-modified δ-MnO2 is only slightly lower than those of the other two cathodes. Based on these results, the significant capacity increment of cobalt-modified δ-MnO2 during cycling could be attributed to its superior deposition rate of Mn compound. A comparison of the dQ/dV profile of the fifth charging (Figure 1F) also indicates that the additional oxidation process on a δ-MnO2 cathode with cobalt modification is more significant than that without cobalt modification. Furthermore, the cobalt-modified δ-MnO2 cathode shows excellent rate performance and rate retention (over 500 mAh g−1 at 0.1 mA g−1 and ∼125 mAh g−1 at 5 mA g−1) after 100 cycles of cathode recovery (Figure 1G). In the following sections, we analyze the property of the material and design a first-reported home-made electrolyzer system to offer more insights into the role of the cobalt-containing species on the mechanism behind the unique performance.

Figure 2

Synthesis and Characterization of Cobalt-Modified δ-MnO2

(A) Illustration of the synthesis procedure of cobalt-modified δ-MnO2, and schematic demonstration of incorporating positions of the Co species on the layered δ-MnO2.

(B) XPS survey profiles of δ-MnO2 and cobalt-modified δ-MnO2.

(C) Co L-edge NEXAFS spectra (TEY mode) of cobalt-modified δ-MnO2.

(D) Mn L-edge NEXAFS spectra (TEY mode) of the three MnO2 materials.

The existence of the cobalt element and its chemical state is crucial to the deposition rate of the cobalt-modified δ-MnO2 cathode material. The presence of the Co was revealed by Co 2p peaks (2p3/2 at 795.5 eV and 2p1/2 at 780.7 eV) in the spectra of X-ray photoelectron spectroscopy (XPS) (Figures 2B and S2). The quantity of cobalt element was further evaluated with inductively coupled plasma-optical emission spectrometry (ICP-OES), which gave a Co/Mn atomic ratio of 0.063:1 after cobalt modification (Table S1). A similar result of Co/Mn ratio (0.052:1) was also obtained by the compositional analysis of the XPS survey spectra (Table S2). These results from both bulk-sensitive ICP and surface-sensitive XPS indicate that Co element is evenly decorated on the δ-MnO2. Evidence from the Co L-edge NEXAFS spectra (Figure 2C) suggests that Co species on the cobalt-modified δ-MnO2 has a valent state between +2 and +3. This indicates that during the modification process, Co2+ in the CoCl2 solution was partially oxidized by the high-valent-state (+4) Mn cations in the δ-MnO2. Meanwhile, the valent state of Mn cation was decreased, as evidenced by the Mn L-edge NEXAFS spectra (Figure 2D). Owing to the presence of Mn-site defects, cobalt-containing species (e.g., Co(II) and Co(III)) can be incorporated in δ-MnO2 nanosheets (as illustrated in Figure 2A). (Wang et al., 2018) In addition, X-ray diffraction (XRD, Figure S3) and selected area electron diffraction (SAED, Figure 3C) patterns of the material show no other well-defined crystalline phase. Detailed observation of the transmission electron microscopic (TEM) and high-resolution TEM (HR-TEM) images did not reveal the existence of any amorphous phase other than the layer-structured manganese oxide. Energy-dispersive X-ray (EDX) mapping in Figure 3F indicates that cobalt element is well dispersed on the cobalt-modified δ-MnO2 with no significant agglomeration. In short, after cobalt modification, cobalt cations with valent states +2 and +3 in the form of cobalt-containing species is incorporated in the highly defective δ-MnO2 nanosheets. A similar conclusion was also drawn by Wang et al. based on analysis of the quick extended X-ray absorption fine structure data (Wang et al., 2018).

Figure 3

Microscopic Characterization of Cobalt-Modified δ-MnO2

(A–F) (A) Scanning electron microscopic, (B) TEM, (C) SAED pattern, (D and E) HR-TEM, and (F) EDX mapping of cobalt-modified δ-MnO2.

Resulting from the removal of the salts after the molten-salt synthesis, the δ-MnO2 nanosheets formed loose-packed particles with a size of several micrometers (Figure S4). The loose packing morphology results in abundant secondary pores, which demonstrates a high specific surface area (139 m2 g−1) and a large total pore volume (0.50 cm3 g−1), as presented in Figure S5. Owing to the mild condition utilized for the cobalt modification, the resulting material shows a flower-like morphology (Figure 3A), exhibiting similar surface area (161 m2 g−1) and porosity (0.48 cm3 g−1) (Figure S5). Based on the observation of TEM images, the thickness of the cobalt-modified δ-MnO2 nanosheets is approximately 5 nm (Figures 3B and S6). This value was in accordance with the linear analysis of the 2D height sensor signals using atomic force microscopy (Figure S7), which indicates that the thickness is in the range 3–5 nm. HR-TEM in Figures 3D and 3E demonstrates the [001] view direction of the cobalt-modified δ-MnO2 nanosheet, which clearly shows the (010), (100), and (110) planes with a lattice distance of 2.4, 2.4, and 1.4 Å, respectively. This result is also confirmed by the SAED (Figure 3C) and XRD (Figure S3) patterns.

The Catalytic Effect of Cobalt Components in the Deposition Process of Active Mn Compound

In this section, the role of cobalt components in the deposition of active Mn compound is further discussed using cyclic voltammetry (CV) and linear sweep voltammetry (LSV) for a practical battery configuration and a home-made electrolyzer cell. Compared with the galvanostatic charging process where the time duration in the range of 1.6–1.8 V for each charging is very short (causing a fast voltage buildup due to the full charge of the cathode), the CV and LSV methods allow equal time at every voltage step. Therefore, well-defined peaks could be identified from a CV or an LSV curve, when compared with the dQ/dV profile extracted from a galvanostatic charging profile.

Using a conventional Zn|electrolyte|MnO2 coin-type battery configuration, an oxidation peak at around 1.65 V was observed, indicative of the oxidation of Mn2+. It is no surprise that without introducing MnSO4 into the electrolyte, the oxidation of Mn2+ is much less significant (Figure S8). Also, in the same electrolyte composition, the oxidation peaks at 1.65 V of the cobalt-modified δ-MnO2 cathode were more noticeable than those of the δ-MnO2 cathode, as presented in Figure 4A.

Figure 4

Catalytic Effect of Cobalt Components in the Deposition Process of Active Mn Compound

(A) Comparison of CVs (0.1 mV s−1) of coin-type batteries using δ-MnO2 cathodes with or without cobalt modification.

(B) Comparison of LSV profiles (positive sweep, 1 mV s−1) of stainless steel (SS) and SS covered with pre-electrodeposited Co(OH)2.

(C and D) (C) Mn 2p3/2 XPS profiles and (D) digital photographs of the samples after the LSV positive sweep to specific potentials.

(E) First two CV cycles (1 mV s−1) of SS with pre-electrodeposited Co(OH)2.

(F) Illustration of mechanism of the catalysis effect via Mn self-catalysis (Route 1) and via cobalt-induced catalysis (Route 2).

Note that results in (B–E) are obtained based on a home-made electrochemical cell (See Figure S9A).

To further reveal the role of cobalt components, a home-made electrolyzer cell with a two-electrode configuration was designed (Figure S9A). A stainless-steel (SS) electrode or a Co(OH)2-coated SS electrode (denoted as SS-Co(OH)2, preparation setup shown in Figure S10) was used as the working electrode. A Zn plate was used as both the reference and the counter electrode. A larger quantity (10 mL) of the electrolyte with the same composition (2 M ZnSO4 + 0.2 M MnSO4) was used. An LSV scan from 1.0 to 2.3 V (oxidizing process) was applied to the SS electrode and SS-Co(OH)2 electrode.

For SS electrode, with the onset voltage at ∼1.85 V, an oxidation peak centered at ∼2.05 V was observed (Figure 4B). Without any Mn compound on the original electrode, this process can be attributed to the oxidation of the Mn2+ in the electrolyte. After scanning to 2.3 V, the electrode (denoted as ▪ in Figures 4B–4D) was collected and characterized with XPS. The Mn 2p3/2 profile of the sample indicates that the deposited Mn compound was mainly in high valent states of +3 and +4 (Figure 4C). For the SS-Co(OH)2 electrode, two additional peaks centered at ∼1.65 V and ∼1.80 V were identified, which could be ascribed to the oxidation of Mn2+ in the electrolyte and the further oxidation of Mn compound (e.g., from +3 to +4) on the solid electrode (Figure 4B). The Mn 2p3/2 profile (Figure 4C) of the SS-Co(OH)2 electrode after scanning to 1.85 V (denoted as ★) shows that the chemical state of the deposited Mn compound was very similar to those of SS and SS-Co(OH)2 after scanning to 2.3 V (denoted as ▪ and ▲). XRD profiles of the SS-Co(OH)2 after scanning to 1.85 and 2.3 V (denoted as ★ and ▲) indicate that their phase strutures are similar to those of the layered δ-MnO2 and cobalt-modified δ-MnO2 (Figure S11). As presented in the digital photographs (Figure 4D), the major color and texture changes of the SS electrode and SS-Co(OH)2 electrode before and after the LSV scan to a certain voltage also provide visual support for the deposition of the Mn compound.

The high valent state of the Mn compound of the SS-Co(OH)2 sample after scanning to 1.85 V (★) indicates the readiness to be reduced and to provide capacity in a subsequent discharge process. In Figure 4E, the SS-Co(OH)2 electrode was CV scanned (1 mV s−1). No reduction peak was found at the first negative scan from open circuit voltage (∼1.28 V) to 1.0 V, indicating that no significant discharging capacity was provided from SS-Co(OH)2. The following positive (oxidation) scan from 1.0 to 1.85 V shows the oxidation of Mn2+ and further oxidation as mentioned earlier, after which the subsequent negative scan shows two well-defined peaks at ∼1.4 and ∼1.2 V (Figure 4E), matching well with the reduction peaks observed from the cobalt-modified δ-MnO2 cathode in a typical Zn-ion battery (Figure 4A).

It is important to note that cobalt element in the state of unbound Co2+ in the electrolyte did not facilitate the deposition of active Mn compound. As presented in Figure S12, LSV of SS in the electrolyte with and without additional 0.01 M CoSO4 demonstrated similar onset potentials of ∼1.9 V, indicating no beneficial effect on facilitating the Mn compound deposition. Therefore, the promoted deposition of active Mn compound would be ascribed to the cobalt hydroxides and cobalt-containing species that bound on the solid cathode. This result is vital for the analysis in the following section.

The mechanism behind the facilitated deposition of active Mn compound could be ascribed to dual-catalysis effects from Mn(IV) species and Co(III) species on the cathode. Usually, Mn2+ in the electrolyte is more stable than most of the Mn cations due to the high tendency of Mn3+ disproportionation (Gibson et al., 2016). A recent research reported by Qiao's group indicated that the deposition of MnO2 on a carbon fiber using a Mn2+-containing electrolyte took place at a high potential of 2.2 V versus Zn2+/Zn in a chronoamperometry charge process (Chao et al., 2019). The sluggish kinetics of the direct oxidation of Mn2+ could result in a large overpotential. Our LSV scan for the SS electrode in the Mn2+-containing electrolyte (Figure 4B) also indicates a similar result where an onset potential of deposition of the active Mn compound was higher than 1.85 V versus Zn2+/Zn, which is beyond the voltage window for a typical Zn-ion battery. The unique facilitated deposition observed in this work within the voltage range below 1.85 V could be ascribed to the following two different catalysis routes (Figure 4F):

Mn self-catalysis (Route 1): where oxidation of Mn2+ was achieved by the comproportionation of Mn(IV) and Mn2+Mn(III) → Mn(IV) + eMn(IV) + Mn

Co-induced catalysis (Route 2): where oxidation of Mn2+ was enhanced by cobalt-facilitated charge transferCo(II) → Co(III) + eCo(III) + Mn

After the oxidation process, the generated Mn3+ is then hydrolyzed and deposited as solid Mn(III) compounds on the cathode (Gibson et al., 2016).

For cobalt-modified δ-MnO2 cathode, both catalysis routes contribute to the deposition process. Of importance, the well-dispersed cobalt-containing species could serve as preferred active sites by providing enhanced charge-transfer capability, thereby demonstrating a much superior deposition rate. On the contrary, the deposition of active Mn compound was still observed but less significant (Figure 4A) for the cathode without cobalt modification, as the catalysis could only rely on the Mn self-catalysis route.

It is worth noting that the cobalt content can influence the deposition of Mn compound. Manganese oxide cathodes with different cobalt contents were compared in Figure S13. The increase in cobalt content provides a better capability for the deposition of Mn compounds on the cathode, thereby leading to faster increases of capacities (Figure S13B). The improvement of the deposition of Mn compounds is not very significant when the Co/Mn ratio is higher than 0.10:1 (atom %), indicating that the cobalt-induced facilitation of the deposition of Mn compounds may have a limitation. Also, the deposition of Mn compounds may be influenced by the concentration of the Zn2+ ion. As presented in Figure S14, batteries with lower Zn2+ concentrations demonstrated higher retention rates. This could be associated with the higher ionic conductivities and higher pH values of the electrolytes, which are both beneficial to the deposition of Mn compounds (Zhang et al., 2016, Poyraz et al., 2019).

The Dynamic Movement of Cobalt Components for Self-Recovery of the Mn-Based Cathode

As shown previously in Figure 1A, the cobalt-modified δ-MnO2 cathode exhibits a superior rate toward the deposition of active Mn compound during the entire cycling test. This result indicates the continuous influence of cobalt components on facilitating the recovery of the Mn-based cathode. The authors once considered that the cobalt components could be only confined on the initial δ-MnO2 cathode. In other words, the newly deposited Mn compound (Figures 5A and 5B) on the surface might not contain cobalt components. Interestingly, however, after 10 and 100 cycles, identical cobalt signals were still observed both in bulk-sensitive EDX spectra (Figures 5C and S15) and in surface-sensitive Co 2p XPS profiles (Figure 5D), although both signals were decreased as expected. This could be attributed to a dynamic movement of cobalt components, similar to the dissolution-deposition process of the MnO2 cathode.

Figure 5

Investigation of the Cobalt Component Before and After Cycling

(A–C) (A) Scanning electron microscopic, (B) TEM, and (C) EDX mapping images of cobalt-modified δ-MnO2 cathode after 100 cycles.

(D) Co 2p XPS profiles of cobalt-modified δ-MnO2 cathode at open circuit voltage before cycling and charged state after 10 and 100 cycles.

It is important to note that the pH value of the electrolyte could be one of the important parameters that may cause the dissolution and redeposition of cobalt components. The pH value of the electrolyte (2 M ZnSO4 + 0.2 M MnSO4) is ∼5 (Sun et al., 2017), a mild acidic environment under which the cobalt components (e.g., Co(OH)2) could be unstable (Gayer and Garrett, 1950). As indicated in previous reports, the electrolyte pH of Zn-ion batteries may vary due to the intercalation/de-intercalation of H+. In a system with a large volume of electrolyte (e.g., 10 mL, experiment shown in Figure 4E), the intercalation/de-intercalation of H+ is not significant enough to largely influence the pH value. Decomposition of Co(OH)2 was observed only after five CV cycles (1.0–1.85 V, Figure S16). However, in a coin-type battery with a small amount of electrolyte (200 μL in each of our battery), the pH value of the electrolyte could be significantly increased after being fully discharged (Lee et al., 2016). The pH increase after the first discharge to 1.0 V was evidenced by the formation of zinc hydroxide sulfate hydrate (which only forms in high pH environment) on the cathode surface, as confirmed by XRD analysis (Figure S17). (Huang et al., 2018, Pan et al., 2016)

In Figure 6, a bigger picture of the role of cobalt components in facilitating the recovery of the cathode and their dynamic movement is demonstrated. During the cycling, cathode reconstruction occurs in the battery, involving the dissolution and deposition of both Mn and Co components. Deposition of the Mn-based cathode is driven by an electrochemical charge process, which is further facilitated by cobalt components bound on the Mn compound. At the same time, the deposition/dissolution of the Co component is a pH-sensitive process. Owing to the weak acidic environment, a small fraction of cobalt component was dissolved in the electrolyte. During battery discharging, H+ was inserted in the Mn-based cathode (Sun et al., 2017), resulting in the increase of pH value and consequently the re-deposition of cobalt component on the reconstructed cathode surface. During each charge process, H+ was reversibly released, causing the re-dissolution of cobalt components to some extent. This dynamic movement of the cobalt component in the solid electrode and liquid electrolyte is important for the persistent facilitation of cathode recovery.

Figure 6

Schematic Illustration of the Whole Picture of This Work: the Dynamic Movement of Cobalt Components during Discharge and Charge of the Batteries and Its Influence on Facilitating Deposition of Active Mn Compound

It is worth mentioning that apart from the facilitating effect from cobalt components, the overall capacity recovery rate of the Mn-based cathode is also limited by many other factors, e.g., (1) limited conductive cathode surface, (2) decreased Mn2+ concentration in the electrolyte after many discharging-charging cycles, (3) the ionic conductivity and pH of the electrolyte, and (4) deterioration of the Zn anode (Ha and Lee, 2016, Selvakumaran et al., 2019). From the consideration of the first factor, the increment of the capacity for a battery with the cobalt-modified δ-MnO2 cathode (as shown in Figure 1G) gradually slowed down during the 50–100 cycles. In addition, a comparison of the morphologies of cathode surface before cycling and after 100th charge provides clear evidence that most of the conductive surface (e.g., current collector and conductive carbon) was heavily covered with dense, newly deposited active Mn compound (Figures S18C and S18E). The deterioration of the charge-transfer situation of cathode surface could further limit the deposition of active Mn compound.

Cycling Performance and More Practical Battery Configurations

The practical viability of the cobalt-modified δ-MnO2 cathode was further demonstrated with a prolonged cycling test based on a higher current density of 2 A g−1 in a 200 μL electrolyte of 2 M ZnSO4 and 0.2 M MnSO4 (Figure 7A). A commercially applicable battery cycle life (80% of capacity retention) of 2,000 discharge-charge cycles is presented. Even after 5,000 cycles, the battery still shows decent capacity retention of approximately 63% (compared with the 10th discharge capacity). As mentioned in the Introduction section, as the Mn2+ additive is also active during cycling, it should also be considered for the evaluation of the overall capacity of the cathode. From the practical utility aspect, the final objective is to stabilize the battery performance with minimum additional Mn2+. Using a smaller electrolyte volume of 70 μL, we then investigate the capacity retention of batteries in 2 M ZnSO4 with five different concentrations of MnSO4 additives (from 0 to 0.2 M as demonstrated in Figure 7B) based on the total Mn compound mass (i.e., MnO2 in the solid cathode and MnSO4·H2O added in the electrolyte). As indicated in Figure 7C, due to its good cathode recovery capability, the cobalt-modified δ-MnO2 cathode enabled a more stable operation of rechargeable aqueous Zn-ion batteries. With a low MnSO4 additive concentration of only 0.02 M (corresponding to an atomic percentage of 12% of the Mn additive in the electrolyte), an average capacity fading rate of 0.23% per cycle was achieved between the 7th and 100th cycles, much lower than that without MnSO4 additive (0.35% per cycle, 1.5 times faster fading). We further demonstrate that batteries with a gel-polymer electrolyte and a cobalt-modified δ-MnO2 cathode also presented decent performance and a similar trend of capacity increase during a 100-cycle stability test (Figure S19). This indicates that the cobalt-catalyzed deposition strategy can also be applied to a hydrogel electrolyte battery system.

Figure 7

Investigation of Long Cycling Performance and the Electrolyte Compositions

(A) Cycle life and corresponding Coulombic efficiency over 5,000 cycles at 2 A g−1 of the batteries with cobalt-modified δ-MnO2 cathode and 200 μL electrolyte.

(B) Comparison of atomic contents of Mn element in electrolyte and cathode.

(C) Comparison of cycling performance for the batteries with 70 μL of aqueous electrolyte containing 2 M ZnSO4 and MnSO4 additives at different concentrations.

Conclusion

Stable performance of Zn-ion batteries was realized through a first reported cobalt-catalyzed recovery strategy, which was achieved by facilitating the deposition of active Mn compound in the cathode. The discharge capacity of a Mn-based cathode was determined by (1) the initial amount and valence of Mn in the solid Mn-based cathode before cycling, (2) the dissolution amount of the Mn cathode, and (3) most importantly, the deposition amount of active Mn compound. The facilitation of deposition is achieved by utilizing a facile cobalt modification strategy that integrated the well-dispersed, cobalt-containing species into the defective δ-MnO2 nanosheets. Verified by both practical coin-type batteries and a newly designed home-made electrolyzer system, the deposition of active Mn compound was confirmed to be significantly facilitated with cobalt-based compounds (e.g., Co(OH)2) and the bound cobalt-containing species (e.g., Co(II) and Co(III)) in the cobalt-modified δ-MnO2 cathode, whereas the unbound Co2+ in the electrolyte showed no significant impact on the deposition process. In addition, the concentration of Mn2+ additives and the volume of the electrolyte are also vital parameters for capacity recovery of the cathode. For a fair performance comparison, these two parameters should be well controlled and clearly indicated in future publications of Mn-based cathode for Zn-ion batteries. Overall, this work provides a deeper insight into the deposition behaviors of the active cathode material and its significant influence on the stable performance of aqueous Zn-ion batteries. This work could also provide inspiration for the materials engineering for future functionalized cathodes, especially for other conversion-type battery cathodes that suffer from the similar dissolution issue, for example, manganese-based cathodes for Li-ion and Na-ion batteries and sulfur cathodes for Li-S and Na-S batteries.

Limitations of the Study

Our work has demonstrated how cobalt modification can facilitate the deposition of Mn compounds on the δ-MnO2 cathode and lead to stable performance of Zn-ion batteries. Although the catalytic effect of cobalt species was evidenced by electrochemical analysis and supported by multiple ex situ characterizations, in-depth understanding of the catalytic mechanism is still needed. Further investigation with advanced in situ techniques and theoretical calculations may provide a deeper insight.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.