Operando pH Measurements Decipher H+/Zn2+ Intercalation Chemistry in High-Performance Aqueous Zn/δ-V2O5 Batteries.

Vanadium oxides have been recognized to be among the most promising positive electrode materials for aqueous zinc metal batteries (AZMBs). However, their underlying intercalation mechanisms are still vigorously debated. To shed light on the intercalation mechanisms, high-performance δ-V2O5 is investigated as a model compound. Its structural and electrochemical behaviors in the designed cells with three different electrolytes, i.e., 3 m Zn(CF3SO3)2/water, 0.01 M H2SO4/water, and 1 M Zn(CF3SO3)2/acetonitrile, demonstrate that the conventional structural and elemental characterization methods cannot adequately clarify the separate roles of H+ and Zn2+ intercalations in the Zn(CF3SO3)2/water electrolyte. Thus, an operando pH determination method is developed and used toward Zn/δ-V2O5 AZMBs. This method indicates the intercalation of both H+ and Zn2+ into δ-V2O5 and uncovers an unusual H+/Zn2+-exchange intercalation-deintercalation mechanism. Density functional theory calculations further reveal that the H+/Zn2+ intercalation chemistry is a consequence of the variation of the electrochemical potential of Zn2+ and H+ during the electrochemical intercalation/release.

Aqueous zinc metal batteries (AZMBs) are a promising complementary technology to the state-of-the-art lithium-ion batteries for large-scale stationary energy storage because of their sustainability,[1] low cost,[2] safety,[3,4] and high energy density among aqueous batteries.[5−7] Since σ-Zn0.25V2O5 was employed as an intercalation-type cathode material for AZMBs in 2016,[8] vanadium oxides have been extensively investigated.[3,9,10] The R&D efforts resulted in the rational design of their structure,[11,12] morphology,[13,14] and addition of secondary components,[15,16] enabling high specific capacities (more than 300 mAh g–1 at 50 mA g–1) and prolonged cycle life (hundreds of cycles at 500 mA g–1).[4,17] Nevertheless, the fundamental questions, i.e., which/how species are intercalated into these vanadium oxides, are still under debate.

Although many different characterizations are used in the studies mentioned above, typically, X-ray diffraction (XRD) is the main tool providing the evidence supporting the respective mechanisms. During the operation of many reported vanadium oxides,[7,8,26−29,18−25] two main phenomena, reversible peak shifts of the initial phase and reversible appearances of basic zinc salts (BZSs) at low potentials, are often observed in the XRD patterns. Different explanations are given in the literature for these phenomena, involving the intercalation of H+,[29] Zn2+,[7] and their combination.[21] The peak shifts indicate solid solution processes, going along with a changing interlayer distance. Zn2+, the main cation in the commonly used aqueous electrolytes, is widely considered responsible for the shrinking interlayer distance during the discharge process.[8,18,19,25] However, the possibility of H+ intercalation, which could also contribute to the interlayer shrinking, is neglected. The appearance of BZSs, e.g., Zn12(CF3SO3)9(OH)15,[30] upon discharge depends on the employed aqueous electrolytes.[24,27,29,31,32] Because the formation of BZC compounds is favored by the increasing electrolyte pH values upon discharge, they have been proven to be an indicator for the reversible H+ intercalation.[27,31] Thereafter, the Zn2+ intercalation process was widely doubted by the scientific community,[33] as the intercalation of multivalent cations is believed to be difficult and sluggish,[34] whereas vanadium oxides exhibit a decent specific capacity and rate capability in AZMBs. Apparently, the role of H+ and Zn2+ intercalations in the vanadium oxides-based AZMBs still needs to be clarified, requiring further experimental evidence. Deciphering the intercalated carriers is not trivial, as it may guide scientists to further optimize cathode materials and electrolytes for better-performing AZMBs.

In this work, high-performance δ-calcium vanadium oxide bronze/reduced graphene oxide (CVO/rGO) was used as a model material to investigate the widely explored but still debated intercalation mechanism.[24,35,36] CVO/rGO was prepared via an efficient microwave reaction (Figure S1) and has been confirmed to be composed of submicrometric δ-Ca0.34V2O5·1.16H2O rods on micrometric rGO sheets (1.9 wt %) via XRD, SEM, EDS, Raman spectra, ICP-OES, and TGA (as shown in Figure a,b and S2–S4).

Figure 1

(a) XRD refinement of as-prepared CVO/rGO and the crystal structure model of the CVO. (b) SEM image of as-prepared CVO/rGO. (c–f) Electrochemical performance of the as-fabricated CVO/rGO electrode. (c) CV curves measured at 0.05 mV s–1. (d) Charge–discharge profiles at 50 mA g–1. (e) Specific capacity and Coulombic efficiency obtained at different specific currents. (f) Specific capacity and Coulombic efficiency of cycling at 2000 mA g–1. Coin cell, 3 m Zn(OTF)2 aqueous electrolyte.

The electrochemical performance of as-prepared CVO/rGO as cathode material for AZMBs was evaluated in coin cells employing a zinc foil as the negative electrode and a 3 m Zn(OTF)2 aqueous solution as the electrolyte. Figure c displays the cyclic voltammetry (CV) curves measured at a scan rate of 0.05 mV s–1. The first cycle is different from those following, indicating an activation process occurs. Also, the presence of multiple peaks indicates a multistep electrochemical process (Supplementary Note 1). The charge–discharge profiles recorded during the initial three cycles (at 50 mA g–1) are displayed in Figure d. In agreement with the CV results, there is an evident increase in the delivered capacity after the first cycle showing a charge capacity higher than the discharge capacity. However, a highly reversible behavior is observed in the following two cycles in which a specific capacity of 357 mAh g–1 is delivered. The different electrochemical response in the first cycle is most likely the result of the deintercalation of calcium ions from CVO, occurring upon the first charge above 1.2 V. This is strongly corroborated via XPS measurements, which demonstrate a significant decrease of the relative signal intensity of Ca 2p against V 2p after the first cycle (Figure S5).[24] The specific capacity and charge–discharge profiles at various current densities, measured via GCPL, are displayed in Figures e and S6a, respectively. At specific currents of 500 and 1000 mA g–1, specific capacities of 267 and 215 mAh g–1 are delivered, respectively. After 200 cycles at 500 mA g–1 (Figure S7) and 2000 cycles at 2000 mA g–1 (Figures f and S6b), the cell retains 92% and 87% of the initial capacities, respectively, thus showing a good cyclability. The CVO/rGO composite, exhibiting satisfactory properties as cathode material, is suited to serve as a model material for the investigation of the intercalation mechanism.

Following the exploration route in earlier works, we start with the ex situ XRD characterization of the CVO/rGO electrode in Zn(OTF)2 aqueous electrolyte, as shown in Figure a. Upon charge and discharge, the XRD patterns exhibit a reversible shift of the CVO’s (001) reflection. Moreover, additional peaks appear/disappear during cycling, suggesting the formation of additional phases. The reversible shift of the CVO’s (001) plane indicates the typical solid solution mechanism originates from cation intercalations. Because of the opposite charge, the intercalated cations reduce the repulsion of the V–O layers and lead to a decrease of the interlayer spacing from 10.6 to 9.8 Å during the discharge process. An expansion back to 10.6 Å is, in fact, observed during the subsequent charge process. At high voltage, the additional reflection appearing at 12.3° can be indexed to the (001) plane of Zn3(OH)2V2O7·2H2O (12.29°, PDF 50-0570), which has also been investigated as an intercalation cathode material.[37,38] Upon discharge, the Zn3(OH)2V2O7·2H2O reflection disappears because of the dissolution of the low oxidation state vanadium species in the aqueous electrolyte.[24] At low voltage (0.2–0.5 V in discharge, and 0.2–0.7 V in charge), a series of reflections at 6.4°, 12.9°, and 19.5° (highlighted in blue) can be indexed to Zn12(CF3SO3)9(OH)15·xH2O, according to previous works.[30,39] The ex situ SEM (Figure S8) and ex situ XPS (Figure S5) further prove the reversible formation of this BZS which can be obtained via the titration of 3 m Zn(OTF)2 aqueous solution with NaOH aqueous solution (Figure S9). Therefore, the formed BZS is indeed a good indicator for H+ intercalation occurring in the low-voltage region (0.2–0.5 V) upon discharge. However, the involvement of H+ in the higher-voltage region as well as of Zn2+ ions along the entire discharge and charge processes are still unclear. Despite the presence of Zn3(OH)2V2O7·2H2O taking up a minor portion of the capacity, the energy storage mechanism of CVO is mainly based on solid solution processes. Although a variation of the diffusion coefficient upon charge/discharge is observed (Figure S10), this cannot be used to distinguish the shuttling cation (Zn2+ and H+), as discussed in the Supporting Information.

Figure 2

(a) Ex situ XRD patterns of cycled CVO/rGO electrodes at different states of charge during the first and second cycles (coin cell, 3 m Zn(OTF)2 aqueous electrolyte). Charge–discharge profiles and ex situ XRD patterns at different states of charge of CVO/rGO electrodes in (b and c) 0.01 M H2SO4/H2O and (d and e) 1 M Zn(OTF)2/ACN electrolytes, respectively. CE is the counter electrode. RE is the reference electrode.

To further clarify the role of H+ and Zn2+ intercalation in CVO, three-electrode Swagelok-type T cells employing either 0.01 M H2SO4 aqueous solution or 1 M Zn(OTF)2 in ACN as the electrolyte were assembled. For the H+ intercalation investigation, free-standing activated carbon (AC) electrodes and leakless Ag/AgCl electrodes were used as the counter and reference electrodes, respectively, to exclude the influence of Zn2+. Because of the limited electrochemical stability window of the 0.01 M H2SO4 aqueous solution, severe irreversible hydrogen evolution is observed in the discharge curve shown in Figure b. Nonetheless, even within the narrow potential window (−0.3–0.73 V vs Ag/AgCl, corresponding to 0.57–1.6 V vs Zn2+/Zn), the CVO/rGO electrode uptakes H+ up to 45 mAh g–1 as observed in the following charge. In the XRD patterns of the CVO/rGO electrode discharged to −0.3 V vs Ag/AgCl and charged to 0.73 V vs Ag/AgCl (Figure c), only peaks belonging to CVO can be observed. The (001) reflection exhibits a typical lattice breath, which proves the solid solution process occurs for CVO in 0.01 M H2SO4. For the investigation of Zn2+ intercalation in CVO, cells employing Zn foils as both the counter and the reference electrodes and 1 M Zn(OTF)2 in ACN as the electrolyte were used. Although with a lower specific capacity (22 mAh g–1) and a stronger shift of the (001) peak (Figure d,e and Supplementary Note 2), the CVO/rGO electrode clearly shows Zn2+ storage ability. From the results mentioned above, H+ and Zn2+ can be independently intercalated into CVO through solid solution processes. However, the structural and electrochemical behaviors of each ion cannot be directly transferred to the Zn/CVO cell using 3 m Zn(OTF)2 aqueous solution because of the inherent differences between the employed electrolytes (Supplementary Note 3).

Overall, a lack of effective characterization tools to discriminate the H+ and Zn2+ intercalation processes appears in the Zn/CVO cell using 3 m Zn(OTF)2 aqueous solution. In fact, the intercalation of both H+ and Zn2+ occurs via solid solution processes, which eliminate the support of XRD, as well as high-resolution transmission electron microscopy (both supplying interlayer distances according to their operation mechanisms). On the other hand, the formation of the BZSs containing Zn species but not related to Zn2+ intercalation in CVO eliminates all elemental analysis techniques, e.g., XPS, EDS, and ICP-OES.

However, it is important to state that the intercalated H+ must originate from the 3 m Zn(OTF)2 aqueous electrolyte. This means that an increase of the electrolyte’s pH occurs when H+ is intercalated in CVO, but not upon Zn2+ intercalation. Therefore, monitoring the electrolyte’s pH appears to be an appropriate, indirect tool to discriminate the H+ and Zn2+ intercalation processes.

Therefore, a cell for operando pH determination has been designed, and its scheme is shown in Figure a. A stainless steel strip and a porous carbon paper (the left one) are used as a current collector for the working electrode. This additional carbon paper disc can increase the contact of the CVO electrode with the stainless steel strip because the size of the strip and the electronic conductivity of CVO are limited. CVO/rGO slurry cast on another porous carbon paper disc (the right one) is applied as a working electrode with the active material side facing the current collector. The back of the electrode, the porous carbon paper substrate, is in direct contact with a pH electrode that has a flat membrane and therefore can measure the pH value of the electrolyte on the carbon paper surface (Figure S11). The use of this configuration guarantees a high sensitivity of the pH measurement, because it detects in close contact of the active material.[40] Moreover, the limited amount of electrolyte (only 100 μL) used in this cell guarantees a significant variation of pH, which is not achievable in the beaker cell’s configuration.[41]

Figure 3

(a) Scheme of the cell configuration for operando pH measurements. (b) Evolution of the cell voltage and the electrolyte pH value upon the initial three charge–discharge cycles at 50 mA g–1. (c) Evolution of pH versus voltage during the initial three charge–discharge cycles. 3 m Zn(OTF)2 aqueous electrolyte.

The cell was subjected to a GCPL test at 50 mA g–1 while the pH value was recorded with a pH meter every 5 min. Figure b displays the typical evolution of the cell voltage and electrolyte pH during the initial three cycles. The cycling performance of the CVO electrode in this operando cell is inferior compared to those obtained in the coin cell tests. Lower efficiency and reversibility are detected, but they are associated with the imperfect configuration of the operando cell (Supplementary Note 4). The limited reversibility of CVO in the operando cell and the hydrogen evolution on Zn metal negative electrodes[6] can explain the drift of the pH value, which does not recover to its initial value after each cycle. However, the correlation of the pH variation with the cell voltage upon cycling is very evident, demonstrating that the pH variation is tightly associated with the electrochemical reaction of the CVO.

At the beginning of the discharge process, a sharp increase of the pH value is observed (Figure b). Despite the pH drift affecting the absolute value, the same occurs in the following discharges and always ends at around 0.87 V, once more implying a strong correlation between the pH and voltage. To have a better overview of the relation between voltage and pH, Figure c illustrates the pH versus voltage trend upon each cycle. In the first discharge, two regions (A–B and C–D) can be identified by increasing pH values, while one region (B–C) shows a rather stable pH. Increasing pH values can be the result of the intercalation of H+ into CVO or/and hydrogen evolution, oxygen reduction side reactions. In the plateau region (B–C), accounting for a significant fraction of the CVO’s capacity (about 35%), the stability of the pH value excludes H+ intercalation. Hence, Zn2+ intercalation is most likely taking place in this region. The noninfluence of Zn2+ intercalation in pH evolution can also be extended to the deintercalation of Zn2+ upon the charge process.

Similarly, in the first charge process, two regions (D–E and F–G) are characterized by decreasing pH value, while two other regions (E–F and G–H) evidence steady pH values. The E–F region can be associated with the deintercalation of Zn2+ ions (previously intercalated in the B–C region). The G–H region, i.e., above 1.2 V, is related to the deintercalation of Ca2+ ions from the CVO (see Figure S5 and related text in the Supporting Information). One can conclude that deintercalation of H+ ions from CVO is the only event leading to the pH decrease in these two regions (D–E and F–G), for the following two reasons:

The voltage in these two regions is much lower than the onset potential of oxygen evolution reaction in 3 m Zn(OTF)2 aqueous electrolyte (higher than 2.3 V vs Zn2+/Zn);[42]

It was evidenced above that the intercalation/deintercalation of Zn2+ does not significantly affect the measured pH.

The pH decrease in the D–E and F–G regions (charge) well relates with the pH increase in regions A–B and C–D (discharge), taking place at the same voltage, respectively. Therefore, the occurrence of deintercalation of H+ in D–E and F–G regions also indicates the presence of H+ intercalation into CVO in regions A–B and C–D.

These results demonstrate that both the H+ and Zn2+ intercalations take place and contribute to the specific capacity of CVO in AZMBs. However, the contribution of Zn2+ intercalation/release in the regions where the pH varies, e.g., A–B, C–D, D–E, and F–G, cannot be excluded.

Upon the following (seconnd and third) cycles, the pH evolution is generally similar to that observed in the first cycle, apart from region I (blue circle in Figure b), in which the pH is seen to increase upon charging. In fact, the intercalated H+ and Zn2+ cations should be released back into the electrolyte during charging. However, the increasing pH value supports for H+ intercalation. This phenomenon is also observed in the third cycle. Moreover, in regions II and III (also marked with blue circles in the discharge processes), the pH decreases rather than stabilizes or increases. These imply that the Zn2+ deintercalation in region I is accompanied by H+ intercalation, while the Zn2+ intercalation in regions II and III is accompanied by H+ deintercalation. Such an H+/Zn2+-exchange has seldom been reported in battery systems. DFT calculations were performed to gain insights into the mechanism responsible for the H+/Zn2+ intercalation chemistry observed above. For this purpose, a large number of model structures with different intercalation chemistry were optimized by applying the VASP code (see the Supporting Information for details).[43,44]

As the Ca2+ is released into the electrolyte during the first charge (Figure S5), δ-V2O5, i.e., the V–O skeleton of CVO, was used. In a first approach, structural water was neglected, meaning that only the intercalation of H+ and Zn2+ ions was considered. Interestingly, for this scenario, Zn prefers to be tetrahedrally coordinated by oxygen atoms from the V2O5 layers, while hydrogen coordinates on the V–O layers trying to maximize its distance to the next Zn atom (Figure S12). This is exemplarily shown for three different compounds (see Figure a). Moreover, it is observed that the structural integrity of the layers is lost when more than 0.5 H+/fu (formula unit of V2O5) are inserted, while the intercalation of one Zn2+/fu is still possible. The concept of the computational hydrogen electrode (CHE) was adapted to access the respective stability of the intercalation compounds in an electrochemical environment.[45,46] With this approach, it is possible to access the phase stability with respect to the electrochemical potential, however, without having to explicitly calculate solvated ions (see the Supporting Information for a detailed discussion). Hence, this allows constructing a phase diagram as a function of the zinc and hydrogen electrochemical potentials (Δμ̃Zn and Δμ̃H, respectively), which depend on the temperature, pressure, and corresponding ion concentrations (e.g., Zn2+ and H+). The phase diagram obtained with this approach is depicted in Figure a. A closer look at the phase boundaries of the respective phases reveals that, depending on the electrochemical potential, different intercalation compounds are stable. In this graph, an increasing H+ concentration (decreasing pH) and otherwise unchanged parameters correspond to displacement along the y-axis. Hence, many regions are seen in the phase diagram for which an increasing pH value will indeed result in H+ moving out from the structure, while Zn2+ will move in. This obviously corresponds to the experimentally observed exchange mechanism and dual-carrier intercalation chemistry.

Figure 4

Calculated phase diagrams for hydrogen and Zn intercalation in (a) water-free δ-V2O5 and (b) δ-V2O5·H2O as a function of the respective electrochemical potential. The stoichiometry of the stable phases is denoted in the corresponding area of the phase diagram. Structure images depicting hydrogen intercalation, Zn intercalation, and hydrogen–Zn cointercalation are shown as examples next to the respective regions of the phase diagram. The dashed red box is a rough estimate of the chemical potential range relevant to the experiment. The yellow arrows point in the direction of increasing Zn2+ and H+ concentration (i.e., decreasing pH).

In a subsequent step, the structural water has been introduced in the DFT model, thus allowing for a more realistic modeling of the intercalation process. The investigated host structure is δ-V2O5·H2O. Indeed, when water is present in the structure, H+ ions either form H3O+ ions with the structural water or coordinate on the V–O layers as in the water-free scenario (see Figures b and S13). However, now the structure can accept more than 1 H+/fu, which is clearly a consequence of the structural water. Zn2+, on the other hand, mostly occurs in 4-fold or 5-fold environments. The oxygen from the V–O layers and structural water are involved in the coordination, as depicted in Figure b. Interestingly, at high Zn content, it is observed that even dissociation of structural water is possible, resulting in the formation of O–H groups in the coordination polyhedra of the Zn2+ (see Figure b). Revisiting the phase diagram, one can observe that the phase fractions and the potential ranges in which they occur have changed; however, the overall trend remains the same. Still, several phase boundaries are observed where the exchange of H+ and Zn2+ may occur, which can also explain the anomalous occurrence of a Zn2+ intercalation step between two H+ intercalation steps. As a side note, it must be mentioned that different orders and further Zn2+ to H+ ratios may produce additional phases, which are stable in certain regions of the phase diagram. Similarly, changing the water content will also affect the details of the phase diagram. However, phase boundaries that result in H+/Zn2+ exchange will always be present, and therefore our findings strongly corroborate the proposed exchange mechanism.

In summary, with selected electrolytes and designed cell configurations, it was proved that both Zn2+ and H+ can independently intercalate into CVO. Operando pH measurements were employed witnessing the intercalation of both H+ and Zn2+ into CVO from Zn2+-based aqueous electrolyte and further revealing an unusual H+/Zn2+-exchange mechanism upon charge and discharge. DFT calculations revealed that the observed H+/Zn2+ intercalation chemistry is a consequence of the variation of the electrochemical potential of Zn2+ and H+ during the electrochemical intercalation/release.