INTRODUCTION
Grid storage, electric vehicles, and portable electronics all benefit from rechargeable battery systems as power sources. Ideally, industrial-scale development of these power sources requires abundant, accessible, and eco-friendly battery materials [
1,
2]. Metallic Zn fulfills these criteria. It also has a high theoretical capacity (820 mAh g
–1), a low electrochemical potential (0.762 V vs. the standard hydrogen electrode), and is chemically stable in air, making it an ideal metal anode material [
3]. Consequently, in the last century, primary Zn batteries such as Zn–MnO
2, Zn–Ag, and Zn–air batteries have been proposed and successfully commercialized [
4–
6].
The proof-of-concept of aqueous rechargeable zinc-ion batteries (ZIBs) was first demonstrated by Xu et al. [
7] (1980); these are still incapable of sustaining stable electrode activity in neutral or mildly acidic electrolyte environments. These systems persistently suffer from severe irreversibility issues, mainly caused by the inevitable growth of dendrites on the anode surface during cycling, as well as the low Coulombic efficiency of the plating/stripping process [
8], both of which affect the cycle-life of ZIBs. In contrast to the intercalation-deintercalation process in the graphite host anode of lithium-ion batteries, the Zn metal anode dissolves and precipitates. Under these circumstances, dendrite formation is inevitable, as the initial nucleation is normally localized as a result of the non-uniform current density caused by surface defects [
9]. Upon charging, Zn ions preferentially deposit at these existing nucleation sites in order to reduce the surface energy, resulting in a mossy or dendritic morphology. Further growth and fracturing of brittle Zn dendrites results in the formation of “dead Zn” and can cause a possible short circuit [
10].
Despite the literature on dendrite suppression is abundant for alkaline Zn batteries [
11–
16], few studies have investigated the issue of dendrite suppression in secondary rechargeable Zn batteries with a mild acidic or neutral electrolyte [
17–
19]. Considering both electrodes, which are positive and negative electrodes, the mild acidic or neutral electrolyte should be utilized in rechargeable ZIBs. A typical example of electrolyte modification is the discovery by Zhang et al. [
20] that Zn dissolution/deposition had a lower onset potential in an aqueous Zn(CF
3SO
3)
2 electrolyte than in a ZnSO
4 electrolyte, facilitating better reversibility, faster kinetics, and increased Coulombic efficiency with cycling. Another approach involves the use of concentrated aqueous electrolytes in which Zn ions are surrounded by electrolyte anions instead of water, preventing H
2 evolution and facilitating dendrite-free Zn plating/stripping with ~100% Coulombic efficiency [
21]. Alternatively, the use of cross-linked polyacrylonitrile (PAN)-based cation exchange membranes as separators results in a homogeneous ion flux that reduces the concentration gradient and promotes excellent zinc reversibility [
22]. Furthermore, strategies to improve the protection of the Zn surface, primarily through the use of porous carbon materials, have been investigated [
23–
25].
While these techniques can help to reduce dendrite formation to some extent, they rarely address the underlying issue of uneven zinc deposition and subsequent growth on the metallic Zn anode surface [
26]. It was recently discovered that the properties of the zinc plating matrix can influence zinc nucleation behavior [
27]. The different morphologies and crystalline structures on the Zn surface can result in different Zn deposition behaviors. Therefore, the surface texture and surface crystal structure can also be tailored to resist dendrite growth, as some crystallographic orientations of the plating matrix are prone to dendrite formation [
12]. A three-dimensional (3D) zinc sponge anode has been reported to confer enhanced cyclability in nickel-3D zinc batteries owing to the minimization of the effective current density [
5]. In addition, 3D Zn anodes for ZIBs with a preferred crystallographic orientation that promotes planar morphological deposition were created using additive-aided electrodeposition [
27]. The main principle behind the aforementioned methods is the improvement of the accessible active surface area. Despite the inherent merits of 3D structures, such as high surface area and more rapid kinetics, the preparation of the aforementioned anode materials involves complex techniques and expensive electrolytes, which hinder their practical applications [
28,
29]. The use of Zn foil as an anode, on the other hand, is a promising alternative. Due to the existence of undesirable surface defects and limited surface area, it is essential to modify the pristine Zn foil physically and structurally prior to electrochemical applications in order to improve reversibility. In LIBs, the strategy effectively used for pristine Li metal foil. Finding a cost-effective and scalable method for fabricating a highly reversible Zn anode from Zn foil, however remains a challenging task.
In this work, we demonstrated the use of a simple and scalable surface modification procedure for the formation of a well-arranged V-shaped valley-like Zn nanoarray on the surface of Zn metal foil that promotes uniform Zn nucleation. The surface-modified Zn foil inherits the merits of a typical 3D structure with a large surface area, which effectively reduces the local current density. In addition, surface modification of the Zn foil also exposes crystallographic orientations that are favorable for planar morphological growth, thereby restricting dendrite growth. As a result, the surface-modified Zn anode exhibits excellent reversibility as compared to bare Zn, with a 4-fold enhancement in running lifespan (over 400 h), even at high areal capacity. Structural and spectroscopic studies revealed that this excellent Zn reversibility originates from a well-ordered nanoarray structure with a favorable crystallographic orientation. In general, this methodology is of particular interest for increasing the reversibility of com-Zn foil, which is often employed as an inexpensive anode material in ZIBs.
RESULTS AND DISCUSSION
The surface morphology of commercial Zn foil (com-Zn) was first observed using FE-SEM.
Fig. 1a,
b shows a dense smooth com-Zn foil surface containing grooves and some splines, while a multitude of surface defects are noticeable at high magnification (
Fig. 1c). Atomic force microscopy (AFM) measured in AC mode was used to determine the 3D surface topology (
Fig. 1d,
e), further confirming the surface heterogeneity of the com-Zn anode on a scale of tens of nanometers. The mountain-like structures on the com-Zn surface serve as ideal nucleation sites for zinc deposition and dissolution [
30].
A native passivation film is always present on the surface of Zn metal as a result of the reaction between the fresh Zn metal and atmospheric contamination during the fabrication process. As a result, the surface chemistry of the native film is solely determined by the atmospheric conditions under which the zinc foil is manufactured. The XPS measurement results (
Fig. S1, Supporting Information) show the full spectrum and the O 1s and Zn 2p photoelectron spectra of the com-Zn foil [
31]. In the full spectrum of the O 1s region (
Fig. S1b), an asymmetric peak with a shoulder at a higher binding energy was observed, which is deconvoluted in three peaks centered at 531.1, 532.4, and 533.1 eV. The 531.1 peak is ascribed to the
O2- ions in the normal wurrtzite ZnO structure. The peak with the highest intensity (532.4eV) corresponds to the oxygendeficient region in the ZnO matrix (defective ZnO
x) or the O–H group absorbed on ZnO. This peak can be ascribed to the existence of partially reduced ZnO at the Zn/ZnO interface [
32,
33]. The higher energy peak at 533.1 eV can be correlated to chemisorbed oxygen such as H
2O or −CO
3 species [
34]. The Zn 2p
3/2 and 2p
1/2 binding energies (
Fig. S1c) for Zn
2+ correspond to the peaks at 1021.2 and 1044.3 eV, respectively. Meanwhile, the Zn 2p
1/2 and 2p
3/2 binding energy peaks located at 1043.4 and 1020.8 eV, respectively, can be attributed to non-lattice zinc. In addition, the broad nature of the Zn 2p
3/2 peak confirms the presence of more than one Zn species. The correct identification of these Zn species is difficult, as Zn exhibits only a small shift in the corresponding binding energy of the Zn 2p
3/2 region. The relative atomic concentrations of Zn and O elements estimated using the corresponding integrated peak areas and sensitivity factors are 41.52% of O and 55.6% of Zn. It has been proven that the oxide film formed on the surface of zinc contains a stoichiometric excess of zinc [
35]. These results indicate that the surface of the Zn foil mainly consists of a polycrystalline ZnO film along with other zinc species such as Zn(OH)
2, Zn
5(CO)
3, and Zn
5(CO
3)(OH)
6, which contribute additional resistance during electrochemical performance. After the application of the facile, cheap, and fast surface modification technique reported in this work, the native surface of com-Zn is inevitably modified or even partially removed, and a nanostructured pattern is formed, which increases the effective surface area. In general, this methodology is of particular interest as it increases the reversibility of com-Zn foil, which is often employed as an inexpensive anode material in ZIBs.
To mimic the V-valley formation mechanism, the Zn foil was chemically etched in a stirred dilute acid bath under laminar flow conditions. This process modified the surface of the Zn foil into a V-shaped valley-like nanoarray structure, improving the overall effective surface area, as shown in the FE-SEM images (
Fig. 2a,
b). The AFM images representing the topography of the E-Zn(S) foil (
Fig. 2c,
d) and com-Zn foil (
Fig. 1d,
e) show stark differences, further confirming the V-shaped valley-like nanoarray on the surface. Considering that V-shaped valleys in the landscape form and develop by erosion driven by the flow of wind or water over the landscape, the nanoarray observed in
Fig. 2a is formed by a nature-mimicking mechanism.
The formation of the nanoarray structure is attributed to the dissolution behavior of the surface zinc oxide (native film) and intermediate dissolution product (i.e., ZnCl
2) during the etching process. Briefly, ZnO dissolves rapidly in water compared to ethanol, and the initial fast local dissolution of surface ZnO occurs due to the small amount of water in the etching bath derived from the acid. Successive dissolution then follows around the initial local dissolution points with a direction determined by the hydrodynamic convection of the oxonium ion (H
3O
+) induced by mild stirring of the liquid in the etching bath [
36,
37].
Fig. 3 shows FE-SEM images of surface-modified Zn at different acid concentrations. Note that a heterogeneous rough surface was formed without stirring (
Fig. S2). A further increase in the acid concentration increases the amount of water in the bath, resulting in a large initial dissolution area. This leads to a larger gap and depth between adjacent channels, as can be seen from the SEM images (
Fig. 3e,
f). The Zn foil etched with a 10% HCl solution exhibited the largest inter-channel distance among all the samples (5% and 2.5% HCl in ethanol). This behavior in dilute strong acid suggests a diffusion-controlled etching process, which is useful for the fabrication of nanostructures [
38].
The morphology of the Zn foil etched with 5% HCl was more regular with a uniform nanoarray. The formation of nanoarrays can be explained by considering the polycrystalline surface ZnO crystals with alternate Zn and O atomic layers along the c-axis [
39]. Initially, the O atoms in the top layer of polycrystalline ZnO react immediately with the electron-deficient oxonium ions (H
3O
+) from the etching bath due to the presence of dangling bond electrons, causing O–Zn bond breaking and etching. Thereafter, O atoms from underneath the exposed Zn atom layer continue to react with H
3O
+ in the etchant, revealing the dangling bonds on the step edges or on dislocation defects. In this way, the etching follows for the next alternating layers of the ZnO crystal, resulting in rapid vertical etching in conjunction with slow lateral etching due to lateral hydrodynamic convection of the oxonium ion, as reported in previous studies.
The chemical composition of the E-Zn(S) surface was then assessed using XPS measurements, as shown in
Fig. S3. Evidently, the O1s full spectrum of E-Zn (S) exhibits an intensity reduction along with peak shifting to lower binding in comparison with the O1s spectra of com-Zn (
Fig. S3b), indicating the effective partial removal of the native surface film. This observation is further supported by Raman analysis (
Fig. S4), where the intensity reduction of the 570 cm
-1 band corresponds to ZnO with oxygen vacancies [
40]. Such systematic etching further alters the surface crystal orientation of the com-Zn foil, as seen from the XRD analysis. The XRD patterns of com-Zn, E-Zn(WS), and E-Zn(S) are shown in
Fig. 4a. The peak positions were identical in all samples without any extra peaks, indicating that the etching process did not induce the formation of any additional impurities. Similar to the com-Zn, the E-Zn(S) and E-Zn(WS) samples also exhibit a strong (101) plane orientation, which is responsible for the high current efficiency during zinc deposition. Therefore, because of the large intensity of the (101) plane, E-Zn(S) may exhibit less polarization when placed in a battery [
11]. This is because the strong (101) orientation decreases the planar packing density, which decreases the corrosion rate owing to the minimization of the activation energy for dissolution. This behavior was experimentally validated (and theoretically corroborated) in the potentiostatic polarization measurements of exposed (001) , (101) single crystals and polycrystalline zinc, where the corrosion rate for polycrystalline zinc was found to be very high [
41]. Furthermore, the (100) and (101) crystallographic orientations of zinc metal are prone to dendrite formation, while the (002) orientation has the highest resistance to dendrite formation, as it supports the basal morphological growth of crystals due to low surface energy [
42].
The (002) peak intensities of all XRD patterns were normalized by considering the intensity of the (101) peak as a reference. The intensity ratios between the XRD peaks at 36.1o for (002) and 43.2o for (101) were found to be 0.68, 0.59, and 0.045 for the E-Zn(S), E-Zn(WS), and com-Zn samples, respectively. Thus, the relative ratio of (002) was significantly improved in the V-valley-shaped nanoarray. With the highest I(002)/I(101) value, E-Zn(S) has the lowest surface energy and is expected to exhibit better electrochemical performance by reducing unfavorable dendrite formation.
To assess the impact of the V-valley-shaped morphology of the Zn metal electrode on the electrochemical polarization process, cycling behavior and impedance analysis of the commercial and surface-modified electrodes were performed.
Fig. 4b–
d illustrate the galvanostatic potential-time profiles of E-Zn(S) electrodes modified with different acid concentrations used in a symmetric 2032 coin cell. The initial polarization of the com-Zn anode was very high (>100 mV), and the cell exhibited serious fluctuations within 40 h before short circuiting (
Fig. S5a). As mentioned earlier, the surface morphological features of com-Zn result in an inhomogeneous current distribution during the plating/stripping process. On the other hand, surface-modified electrodes possess highly reduced overpotential values (~50 mV) (
Fig. 4b–
d). All E-Zn(S) anodes modified with different acid concentrations exhibited improved performance in comparison with the com-Zn anode in terms of cycle stability as well as polarization, with stable performance up to ~100 h without severe fluctuation of the overpotential value. On the other hand, the E-Zn(WS) anode suffers from severe voltage fluctuations (
Fig. S5b) caused by a highly rough surface, as shown in the SEM analysis (
Fig. S2). Based on these results, V-valley-like nanoarray surface modification offers two main synergistic advantages: 1) An elaborate nanostructured alteration of the surface to homogenize the ion flux. 2) An increase in the surface area, thereby reducing the local current density. The uniform nanoarray brings excellent structural homogeneity to the surface of E-Zn(S) etched in 5% HCl, which guides uniform Zn nucleation. This results in superior performance among all the modified Zn anodes with low overpotential values along with stable cycling performance (
Fig. 4c). All further analyses were carried out using an E-Zn(S) anode modified in a 5% HCl solution, representing an optimized sample. The excellent performance of E-Zn(S) etched in 5% HCl was further validated by long-term cycling at different current densities, as shown in
Fig. 5a–
c. The symmetric cell operated steadily over 800 cycles (>400 h) at 0.25 mAh cm
–2 and operated at high current density 1 mA cm
–2 for 260 hours with an overall overpotential value less than 80 mV.
The impedance analysis of com-Zn foil and E-Zn(S) during the initial cycle and after the 30
th cycle was investigated by EIS, as shown in
Fig. S6. The diameter of the semicircle shown in the Nyquist plot represents the surface resistance, and the related equivalent circuit is shown in the inset. The overall high resistance of the coin cell assembled with com-Zn compared to that with E-Zn(S) clearly demonstrates that the surface modification reduces the resistance of the Zn anode. With the same macroscopic area, the V-valley-shaped Zn anode has a higher surface area than com-Zn, which increases the electrochemical reaction rate at the same current, thereby reducing the polarization of the Zn anode. Furthermore, the cell impedance of both electrodes decreased with cycling as the surface area increased due to Zn deposition. This phenomenon is also observed with of other metal anodes due to dendrite formation [
43]. After the 30
th cycle, E-Zn(S) exhibited a much lower resistance than com-Zn, indicating that the unfavorable resistance behavior was highly reduced with E-Zn(S).
Surface modification of the Zn anode promoted the uniform deposition of metallic Zn during the plating/stripping cycles. As shown in
Fig. 6a,
b, uniform Zn deposition behavior along with a flaky Zn morphology was observed over the entire surface of the E-Zn(S) foil during plating at 0.25 mA h
–1. In contrast, non-uniform Zn deposition behavior with Zn protrusions was observed in the case of the com-Zn foil (
Fig. 6c,
d). Additionally, in the case of the E-Zn(S) foil, the overall structural features were well maintained even after 20 plating/stripping cycles (
Fig. 6e and inset), indicating the key role of the valley-like structure in improving the cycling stability. On the other hand, the surface of the com-Zn anode was covered with an abundance of large Zn protrusions after 20 cycles (
Fig. 6f).
To further understand the growth mechanism of metallic Zn during plating/stripping cycles and its role in cycling stability, additional FE-SEM analysis was employed to check the surface morphology changes after repeated Zn plating/stripping processes.
Fig. S7 shows the FE-SEM images of com-Zn and E-Zn(S) after 60 cycles at an areal capacity of 1 mAh cm
–2 and current density of 1 mA cm
–2. Significant differences in morphologies were observed between com-Zn and E-Zn(S). It can be seen from
Fig. S7a,b that a large number of protrusions with a range of dimensions (from several to ten microns) formed on the surface of the com-Zn metal. Furthermore, in case of com-Zn, the obvious increase in overpotential from 98 to ~150 mV (
Fig. 5a) after cycling indicates the formation of “dead” Zn on the surface, which further blocks the Zn-ion diffusion path during cycling [
44]. FE-SEM images of the filter paper separator (
Fig. S7c and inset) reveal the detachment of such “dead” Zn protrusions from the metal surface. This could lead to separator piercing, resulting in a short circuit that is directly accountable for the short lifespan of ZIBs with a com-Zn anode. Overall, Zn deposition on com-Zn is very uneven due to localized Zn accumulation on the lumpy structure of the comZn foil, as previously observed in the AFM analysis. On the other hand, zinc deposition on the E-Zn(S) anode is non-dendritic, which mainly depicts the presence of fine particles on the surface, as shown in
Fig. S7d,e. In this case, the initial stage of Zn deposition proceeds heteroepitaxially along the preferred Zn orientation (i.e., (002)) resulting in the formation of planar Zn structures [
45]. These specifically oriented thin-layered crystals enclosed by hexagonal edges are shown in
Fig. S7f. As the appearance of the thin-layered zinc crystals depends on the substrate orientation, only parallel crystals edges aligned in a specific direction are often observed, as reported previously [
46,
47]. Subsequently, further random growth begins partly on the (002) planes, and finally, random growth of fine (101) Zn particles occurs partly on the (002) planes (indicated by the white circle in
Fig. S7f). These fine particles eventually agglomerated into larger lumps (indicated by red circles in
Fig. S7f). It has been previously suggested that zinc atoms are incorporated one after another into micro-steps or kinks of vicinal surfaces, without forming nuclei [
48]. Therefore, the presence of micro-steps endorses the initial epitaxial planar growth of Zn along the stiffness of the valley-like structure, which is possible only by 2D nucleation. Therefore, initial 3D deposition does not occur on the surface of the modified Zn anode, minimizing early dendrite protrusion formation via the energetically favorable “bunching mechanism,” instead epitaxial planar growth occurs, as shown schematically in
Fig. 6g. Hence, surface modification of the Zn anode not only exposes preferred orientations (planar morphology) but also tailors the surface morphology in such a way that it further facilitates planar growth in order to reduce dendrite formation during Zn plating/dissolution. These FE-SEM results clearly reveal that surface modification plays a key role in the Zn deposition morphology during the plating/dissolution process.
To further elaborate the dendrite suppression phenomenon of E-Zn(S), an asymmetric E-Zn(S)/com-Zn open electrolyte cell was assembled. The cell was cycled at a constant current density of 1 mA cm
–2 with an areal capacity of 1 mAh cm
–2. The 1
st cycle state of the cell is shown in
Fig. S8a. After 40 cycles, black mossy dendrite deposition occurred on the surface of com-Zn (
Fig. S8b). In striking contrast, the surface of the E-Zn(S) electrode remained smooth without any deposition (
Fig. S8c). Black material was only observed at the edges of the foil. Furthermore, some of these Zn protrusions detached from the Zn surface and sank to the bottom. Such detachment results into “dead” Zn.
In order to probe the potential of surface-modified Zn foil as an anode for practical applications, we assembled a full cell with V
2O
5 as the cathode material. The high purity of the purchased V
2O
5 powder was confirmed by XRD (
Fig. S9), which was well-matched with the standard values of the orthogonal phase (JCPDS no. 41-1426, space group Pmmn (59)) [
49]. We used aqueous 3M ZnSO
4 as an electrolyte, as it supports good reversibility compared to the dilute electrolyte [
21].
Fig. 7a,
b display the CV curves of com-Zn and E-Zn(S) anodes with a V
2O
5 cathode scanned at 2 mV s
–1, respectively. The overall shape of the CV curves and the position of the couple redox peaks were almost identical in both cases, with different current responses. The initial cycle profiles are slightly different from those of the successive cycles, which is attributed to the gradual activation of the material. The same activation phenomenon was also observed for MnO
2 and vanadium-based ZIBs [
44,
50,
51]. The anodic peaks centered at approximately 1.1 and 1.40 V correspond to Zn
2+ extraction while cathodic scan peaks centered at approximately 1.1 and 0.67 V are attributed to the gradual intercalation of Zn
2+ into the V
2O
5 structure [
52]. The high current response of the cell assembled with the E-Zn(S) anode compared to that of com-Zn indicates its high current efficiency. A marked difference in the long-term discharge cycling stability was observed between the two anodes operated at 25oC. The surface-modified Zn metal anode exhibited a high discharge capacity and improved capacity retention in comparison with the unmodified anode. The capacity severely deteriorated in the case of the com-Zn anode, even after 200 cycles. After the 600
th cycle at a current density of 100 mA g
–1, average capacity retentions of 64% and 30% were obtained from the E-Zn(S) and com-Zn cells, respectively (
Fig. S9). In addition, the current density plays an important role in determining cycling stability. At a higher current density (500 mA g
–1), the com-Zn cell performed very poorly starting from the initial cycles. On the other hand, the E-Zn(S) cell not only exhibited good discharge capacity (
Fig. 7c) but also showed excellent cycling performance. The similar nature of the CV curves taken after the initial and 500th cycles further confirm the excellent reversibility obtained with the E-Zn(S) anode (
Fig. S10), which retained >120% (from the initial capacity), or >80% capacity (from the maximum capacity) at 500 mAh g
–1 after 500 cycles (
Fig. 7d). The surface-modified E-Zn(S) anode with a high surface area is subjected to a lower current density compared to that of a com-Zn anode under the same operating conditions and is therefore likely to improve the capacity and capacity retention ability.
A post-cycling FE-SEM analysis of the E-Zn (S) and com-Zn anodes is shown in
Fig. 7e,
f, indicating the dendrite and protrusion -free surface of the E-Zn (S) anode. Moreover, the surface is porous, and the original valley-like structure of the E-Zn(S) anode is well-retained after 100 cycles at a current density of 100 mA g
–1 (inset;
Fig. 7f). On the other hand, dense compact surfaces with the irregular morphological growth of side products were observed on the surface of the com-Zn anode, resulting in high resistance (inset;
Fig. 7e). In addition, Energy dispersive X-ray analysis (
Fig. 7e,
f (insets)) shows a higher percentage of oxygenated side products on the surface of com-Zn, which may originate from electrolyte decomposition or irreversible zinc deposition, further reinforcing the earlier observations.