Control of Growth Front Evolution by Bi Additives ... - ACS Publications

Letter Cite This: Nano Lett. 2018, 18, 1093−1098

pubs.acs.org/NanoLett

Control of Growth Front Evolution by Bi Additives during ZnAu Electrodeposition Jeung Hun Park,†,‡,§ Nicholas M. Schneider,∥ Daniel A. Steingart,*,‡ Hariklia Deligianni,† Suneel Kodambaka,*,§ and Frances M. Ross*,† †

IBM T. J. Watson Research Center, 1101 Kitchawan Road, Yorktown Heights, New York 10598, United States Department of Mechanical and Aerospace Engineering, and Andlinger Center for Energy and the Environment, Princeton University, 86 Olden Street, Princeton, New Jersey 08544, United States § Department of Materials Science and Engineering, University of California Los Angeles, 410 Westwood Plaza, Los Angeles, California 90095, United States ∥ Department of Mechanical Engineering and Applied Mechanics, University of Pennsylvania, 220 South 33rd Street, Philadelphia, Pennsylvania 19147, United States

Nano Lett. 2018.18:1093-1098. Downloaded from pubs.acs.org by UNIV OF SUNDERLAND on 09/29/18. For personal use only.

‡

S Supporting Information *

ABSTRACT: The performance of many electrochemical energy storage systems can be compromised by the formation of metal dendrites during charging. Additives in the electrolyte represent a useful strategy to mitigate dendrite formation, but understanding the mechanisms involved requires knowledge of the nanoscale effects of additives during electrochemical deposition. Here we quantify the effects of an inorganic additive on the morphology of an evolving electrochemical growth front, using liquid cell electron microscopy to provide the necessary spatial and temporal resolution. We examine deposition of ZnAu on Au in the presence of Bi additive, and show that low concentrations of Bi delay but do not prevent the formation of growth front instabilities. We describe a model in which Bi segregates at the growth front and promotes the surface diffusion and relaxation of Zn, allowing better coverage of the initial Au electrode surface. A more precise knowledge of the mechanism of inorganic additive effects may help in designing electrolyte chemistry for battery and other applications where morphology control is essential. KEYWORDS: Dendrites, additives, energy storage, electrochemistry, liquid cell electron microscopy, in situ electron microscopy n-based flow batteries promise an environmentally friendly and low-cost solution for grid-scale energy storage.1 However, the formation of metal dendrites during charging is one factor reducing the performance of these and other secondary batteries.1,2 Strategies to suppress these growth instabilities include electrolyte flow3 and modulation of the charging current4 or potential,5 but the simplest approach may be the use of additives in the electrolyte.6 Several organic additives are known to inhibit dendrite formation to a greater or lesser extent during Zn deposition.1,7−9 However, extensive potential cycling results in consumption of the organic additives and reduced anode capacity and durability that still limit performance.1,9 Inorganic additives can also influence electrodeposition kinetics so as to suppress dendrite formation,1,10 and can be more stable than organic additives9,11 over time. Inorganic additives are therefore widely employed to suppress H2 generation, reduce dendrite formation, and improve discharge and cycling reversibility.1,10,12 However, the mechanism of their interactions is not well-documented,10 partly because it is difficult to measure additive effects on the spatially and temporally developing growth front with sufficient

Z

© 2018 American Chemical Society

resolution. A better understanding of additive effects on growth front morphology is a key goal that may help in developing new strategies for additive design to optimize cycling performance. In order to quantify the evolution of the electrochemical growth front with good spatial and temporal resolution, we have therefore used liquid cell transmission electron microscopy (TEM) to examine additive effects in a Zn-based system. In this technique, an electrochemical process takes place within a thin electrolyte layer held within a sample designed to allow observation in the TEM.13 The evolving morphology can be tracked with simultaneous recording of electrochemical parameters. Liquid cell TEM studies have explored several aspects of battery operation including the formation of the solid−electrolyte interphase14 and growth front morphology during potential cycling.15−18 Dendrite formation and its suppression with organic additives have been studied ex situ for alkaline Zn-based systems.19 However, the role of inorganic Received: November 1, 2017 Revised: December 20, 2017 Published: January 8, 2018 1093

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098

Letter

Nano Letters

Figure 1. In situ progression of the ZnAu growth front. (A−J) Time series of bright field TEM montages: (A−E) electrolyte 0.1 M ZnSO4; (F−J) electrolyte 0.1 M ZnSO4 + 0.25 mM Bi2O3. Images were obtained after each 5 deposition/relaxation cycles at the times indicated. Each cycle consisted of current −20 nA for 10 s and 0 nA for 10 s. The electron flux was ∼50 e−/(Å2 s). The images are ∼10K × 1.2K pixels and show 25 μm of the electrode edge. Image D shows better contrast due to a hydrogen bubble that formed at 289 s; another formed at 469 s. (Image D is therefore after 289 s of deposition rather than 300 s.) We did not observe bubbles in the Bi-containing Zn solution over the entire 500 s experiment. After 10 successive depositions the total charge passed is ∼4.5 × 10−4 C/cm2. For context this would deposit ∼31 nm Zn if distributed uniformly over the electrode surface (at two e− per deposited Zn). The lateral dendrite lengths were typically 400 nm at this time. (E, J) Superimposed outlines of the growth fronts in the two solutions. Black, red, green, and blue lines show the growth front at t = 0, 100, 200, and 300 s, respectively. (K, L) Average length of deposited material in the two solutions. Three regions of interest, A, B, and C, are marked with black square (■), red circle (●), and blue triangle (▲) symbols. Note that the disconnected islands are only loosely attached to the windows, moving and rotating during growth.

Figure 2. Images extracted from Movies M1 and M2 recorded in (A) 0.1 M ZnSO4 and (B) 0.1 M ZnSO4 + 0.25 mM Bi2O3 during cycling at −20 nA for 10 s followed by 0 nA for 10 s. t is the total time elapsed. (C) Applied current and measured voltage during the first 5 cycles. (D) Root-mean square surface roughness of the two growth Movies M1 and M2 vs time. The images were analyzed with a partial field of view (1.5 μm × 1 μm) after the correction of image drift and rotation. The stepwise increase in roughness for each current pulse is visible. With Bi, roughness does not increase until 200 s and then increases in a stepwise manner.

1094

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098

Letter

Nano Letters

additive effects on growth morphology. The electrode edge initially appears fairly smooth. Ramified structures, which we denote generally as dendrites, form in the plain solution within the first 5 cycles and become longer with additional cycles (Figure 1A−D). Similar to observations of Pb,15 Cu,28 and Ag29 growth, each dendrite grows approximately along the direction of the electric field (see Figure S2 for detailed geometry of the liquid cell). The presence of Bi causes clear changes to the growth front evolution (Figure 1F−I). Growth remains smooth during the first 5 cycles, with an irregular growth front beginning to appear after 10 cycles. Without Bi, the dendrite structures are more fernlike while Bi produces more nodular and rounded dendrite tips. A second effect of Bi is in the suppression of hydrogen. Without Bi, large bubbles typically form within 20 cycles (a bubble is present in Figure 1D, accounting for the improved contrast in that image), but such bubbles were not observed in experiments in which Bi was present. We quantify these observations by marking the growth front and superimposing the line profiles (Figure 1E,J). A distinctive difference between the two experiments is that while many points along the interface stop growing in the plain solution, resulting in deep valleys in the growth front, in the Bicontaining solution the growth front continues to progress at most locations, and is thus smoother overall. Based on the electrochemical data in Figure S1, the overall deposition rate in the two solutions is similar. Bi appears to have its greatest effects at early times, promoting deposition on the electrode surface. To quantify the effects of electrolyte thickness on morphology, in Figure 1E,J we divide the electrode edge into three regions of interest (ROI) labeled A−C, at which the corresponding liquid thicknesses are approximately 250 nm, 500 nm, and 1 μm. Note that these liquid thicknesses are larger than those often used in liquid cell TEM experiments. Thick liquid layers were chosen to provide a better model for bulk processes, even though the quality of the images obtained through such thick layers is reduced.30 In each ROI we measure the root-mean square (RMS) roughness and the distance from the initial to the most advanced position of the interface (Figure 1K,L). In the plain solution, the overall lateral growth rate depends on electrolyte thickness: the dendrite length is approximately proportional to electrolyte thickness at each time. Bi additives reduce both average dendrite length and this variation in growth rate, supporting the idea that Bi ions promote more homogeneous deposition. Bi appears to suppress electrolyte thickness effects for at least the first 20 cycles of deposition. After this time, the morphology at the growth front becomes dendritic, and the dendrite length starts to show a stronger dependence on liquid thickness (see Figure S5, with data after 25 cycles and t = 500 s); in other words, the Bi effects disappear after some time. A dependence of overall growth rate on electrolyte thickness in the plain solution is not surprising because variations in electrolyte thickness will change the supply rate of ions. The lack of such an effect in the early stage growth with Bi may reflect a larger role of surface diffusion causing more uniform growth along the electrode. Finally, the effects of Bi in suppressing bubble formation are also seen in this data (Figure 1D, Figure S4), consistent with previous suggestions that Bi suppresses hydrogen bubble generation.31 The growth data at higher spatial and temporal resolution are shown in Figure 2 and Movies M1 and M2. The initial planar growth is highlighted in Figure S4. The lateral growth regime is

additives in acidic Zn-based systems has not been addressed in liquid cell TEM. We thus use the technique to explore the mechanism by which a commonly used additive, Bi,1,10,12,20,21 modifies growth front evolution in a Zn-based electrolyte. We quantify the role of the additive by tracking the morphological evolution of the material, a ZnAu alloy, that deposits onto an Au electrode from acidified zinc sulfate. We use the results to discuss a model for the effect of Bi, in which Bi both smooths the surface and inhibits the deposition of Zn on Zn, favoring deposition on the electrode between existing growth nuclei. Images of the growth front are shown in Figures 1 and 2 and Movies M1 and M2, with the simultaneously measured electrochemical parameters shown in Figure S1. The darker region at the bottom of each image is the Au electrode while the brighter region is filled with electrolyte. These data were obtained in a liquid cell equipped with parallel electrodes composed of electron-transparent, 30 nm thick polycrystalline Au (Figure S2). One electrode serves as the working electrode (WE) and another as a combined counter/reference electrode (C/RE). The electrolyte was 0.5 M H2SO4 + 0.1 M ZnSO4 (pH = 0.4) with or without 0.25 mM Bi2O3. Deposition was carried out under controlled current without deaerating or flowing the electrolyte. To extend the lifetime of the liquid cell by minimizing dissolution of the CE22 and generation of hydrogen bubbles,23 the deposition current was applied in 10 s pulses at typically −20 nA, followed by a 10 s rest time at zero current.4,12,24 The deposition conditions and reproducibility are discussed further in the Supporting Information. Images were recorded under bright field conditions using a 300 keV TEM (FEI model CM30). Low-dose imaging conditions (electron flux ∼50 e−/(Å2 s)) were used to minimize electron beam effects in the experiments while still providing sufficient resolution to track the growth front. Images were recorded at 30 frames per second. Furthermore, after every 5 cycles (total time = 100 s) the current was paused, and the entire electrode edge was recorded as a montage of TEM images, a procedure taking several minutes. This imaging approach allowed the growth front progression to be recorded with reasonable time resolution in one location, but also over a large area every 100 s to evaluate the effect of electrolyte thickness on growth rate and morphology. In the present experiments, the electrolyte thickness varies from 250 nm to ∼1.5 μm along the length of the electrode. The minimum thickness is determined by the spacer layer thickness used in the liquid cell, while the variation in electrolyte thickness due to deflection of the windows25,26 was estimated by elastic beam theory,27 assuming 1 atm pressure difference between the cell interior and the microscope vacuum and a Young’s modulus of 300 GPa for silicon nitride. Compositional analysis of the deposited layer was carried out postgrowth by Xray energy dispersive spectroscopy (XEDS) in a scanning electron microscope (SEM) after flowing deionized (DI) water through the cell then separating and drying the chips (see the Supporting Information). When current first flows we expect the initial deposition to take place over the entire electrode surface. Once the electrolyte directly above the electrode is depleted of ions, this “surface growth” should transition to “lateral growth”, or a growth front that extends outward from the edge of the electrode. The initial surface growth is visible in the data in Figure 1 if the image contrast is stretched to emphasize details on the electrode, as is shown in Figure S3. However, we analyze the lateral growth regime for a more quantitative measure of 1095

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098

Letter

Nano Letters

Figure 3. Postgrowth SEM characterization of the deposited material after 25 cycles, with boxes showing the area used for chemical analysis. (A) Plain solution. (B) Bi-containing solution. (C, D) Magnified SEM images and elemental XEDS maps from the boxes in A and B, used to measure the deposit composition. It is not possible to determine whether O was incorporated during or postgrowth.

deposit). The XEDS penetration depth is in the micrometers range so this composition reflects the bulk of the deposit and not just surface segregation. The deposited Au originates from the CE, as discussed above in the context of nanoparticle formation. The presence of Bi appears to correlate with higher Zn composition in the deposit. The electrochemical, compositional, and morphological data combined allow us to discuss the sequence of events during deposition. In the plain electrolyte, Zn and Au are codeposited because Au is released from the CE during current flow. Although we do not have accurate reference potential measurements in these two-terminal experiments, Figure 2C shows that the voltage becomes more negative over the 10 s current pulse. Initially the imposed current drives a relatively low voltage. The first reaction at the CE is then the reduction of oxygen at approximately −0.3 V, while at the WE, Zn2+ is converted to zero-valent Zn. However, the voltage does not return to its initial value even after the 10 s rest time, and continues to become more negative until a steady state is reached after several cycles (Figure S1B). As Zn2+ is depleted at the Au WE surface, the potential shifts more negative until it reaches a value that supports other reactions at the CE and WE that can supply the desired current. These reactions are Au dissolution at the CE and Au + Zn deposition at the WE. In strong acids, Au dissolution begins with oxidation to form a soluble Au(III) species or film of Au(OH)3, but Au(I) also forms at more negative potentials.36,37 The dissolution of Au is initiated above −1.35 V (ref 38) and occurs until −1.55 V (vs Au plated Pt reference) in 0.5 M H2SO4 (ref 38). For the electrolyte used here, the theoretical redox potential for Au3+ is E° = −1.51 V (ref 39). These reported values are consistent with our observations. The released Au ions diffuse under the electric field gradient and are codeposited at the WE along with Zn that diffuses from

characterized in terms of growth front roughness in Figure 2D and Figure S4C (obtained using a fully automated imaging processing algorithm32). In the plain solution, the growth front roughens quickly, increasing with each cycle, and develops needle-shaped features. In the presence of Bi, the morphology is more globular with some degree of faceting and does not become very rough over the first 100 s, forming a compact deposit (i.e., no deep valleys). A final feature of these data is the apparently disconnected features that form above the electrochemical growth front. In the plain solution, these are needles; in the presence of Bi ions, they are rounder islands. These appear more frequently in areas that are under continuous electron beam irradiation, as is required during video recording, but can be seen in areas with only occasional irradiation. Metal ions in solution are known to be reduced by beam-induced radiolysis species.13,33 Separate experiments34 using various electrolyte chemistries (such as 0.1 M HCl) and higher electron flux (>250 e−/(Å2 s)) suggest that Au3+ ions released from the CE during current flow (a process discussed below) can reach concentrations high enough to allow beam-induced nucleation of nanoparticles. The electrolytes used here can be irradiated for tens of minutes without nanoparticles appearing,35 but nanoparticle formation during electrochemical deposition is consistent with release of Au3+ ions from the CE followed by chemical reaction with radiolytic or other species present in the electrolyte.33−35 The composition of the deposited material is not discernible in the TEM images but was measured postgrowth by XEDS. Zn (and Bi, in the case of Bi-containing electrolyte) are detectable along the edge and over the electrode surface (Figure 3), presumably from the initial deposition. The subsequent lateral growth is an alloy of Zn and Au that is rich in Au, with 11 atom % Zn in the plain solution and 23 atom % Zn in the Bicontaining solution (and Bi concentration ∼0.03 atom % in the 1096

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098

Letter

Nano Letters

containing electrolyte, we note that surface diffusion is a factor that will suppress the development of surface irregularities.51,52 This smoothing effect appears to persist for long times (∼300 s) even after the initial electrode surface is covered. This is consistent with studies in alkaline43 electrolytes that also suggest that Bi at ppm levels enhances mass transport to form smoother surfaces. It is interesting to consider that the morphology of beam-induced nanocrystals, which is formed in the absence of the electrochemical driving force, may provide information that complements measurements of the morphology of the electrochemical growth front. In summary, spatially and temporally resolved images of growth front propagation allow us to evaluate the effects of an inorganic additive on growth morphology. We find that the presence of Bi has several effects on the deposition process. Bi appears to facilitate more homogeneous deposition over the electrode surface, even in the presence of changes in electrolyte thickness and hence ion supply. It delays but does not prevent the growth of ramified features at the ZnAu growth front. It appears to enhance the surface diffusivity of the growth species, producing a morphology (both at the electrochemical growth front and at beam-induced, disconnected crystals) that is smoother at the local scale compared to Bi-free deposition. Bi also appears to suppress hydrogen bubble formation. We have discussed the effect of Bi in terms of a model in which Bi deposits on the growth front prior to Zn deposition, inhibiting continued addition of Zn onto asperities, enhancing surface diffusion and favoring Zn deposition onto Au. We suggest that, to optimize the formation of a smooth and compact deposition morphology, it may be advantageous to combine Bi ions at ppm levels with pulse plating techniques as were used here. Indeed, pulse plating of Zn in acidic electrolytes is already known to have beneficial effects on mass transport, electrode kinetics, and the nucleation of growth centers.53,54 The use of liquid cell observations to develop a model for this particular additive may be a useful strategy in understanding general inorganic additive effects on the control of unstable, dendritic growth, with applications in improving secondary battery performance.

more distant regions of the electrolyte. The deposit composition depends on the relative supply rates of Au and Zn. Under these conditions, the growth front eventually becomes dendritic because in electrochemical models a dendritic morphology is associated with growth that is limited by diffusion of the depositing ions. Phase-field simulations for plain Zn show that ion concentration and electrical potential at the tips of asperities are higher than those at other positions, leading to a concentration gradient and heterogeneity of local current densities at the tips 40 that promote dendrite growth.40,41 The presence of Bi changes this deposition in several ways. At early times (t ∼ 100 s), potentials are generally more negative in the presence of Bi for the same current (Figure 2C). More negative potentials in the presence of additives are attributed to the absorption of additives on the electrode and partial blockage of active sites which decreases the nucleation rate and increases overpotential, and inhibits electrocrystallization onto the adsorbed additive layer.42 Bi also smooths the surface, both in the electrodeposited material and in the isolated nanocrystals. However, the Bi effects appear to diminish with time. After ∼300 s the potential is similar for the two solutions (−1.56 ± 0.05 and −1.58 ± 0.05 V, respectively), and the growth front has become dendritic in both cases. However, deposition is more uniform along the length of the electrode, and there are fewer valleys with no growth. Bi thus does not suppress roughening but appears to delay its onset. The result is a more compact and void-free deposit. This morphological change also agrees with a change in local current distribution near the tips due to additive adsorption, which increases the surface polarization and creates a more uniform current distribution.6 The details of the activity of Bi as an additive are generally understood in terms of its adsorption onto the electrode surface during electrodeposition.12,20,21,43 One can expect Bi ions to be reduced to solid Bi before Zn ions are reduced, due to its larger ° = +0.317 V; EZn ° = −0.761 standard reduction potential (EBi V).44−46 Similar to Bi deposition in alkaline solution,21,43,47 cyclic voltammetry shows that Bi is expected to be deposited onto Au at less negative potentials than those required for Zn deposition onto Au (Figure S1). In the experiments shown here, as the potential becomes more negative during each deposition pulse, the Bi is therefore deposited onto Au before Zn onto Au. Surface energy considerations48,49 suggest that while the current is off, this Bi should migrate toward the electrode surface. As the asperities develop during diffusionlimited deposition, preferential accumulation of Bi ions at the locations with higher potentials (the tips) would lead to the formation of a screening layer (and hence increase in the voltage required for deposition), similar to the action of Cs or Rb additives on Li.41 The screening layer can favor Zn or Au deposition adjacent to the protrusion, rather than at the tip, and thus slow the development of the asperity. This mechanism is consistent with the growth mode observed, in which growth continues over most or all of the front rather than only at a few locations. This model differs from our understanding of leveling in other systems such as Cu with polyethylene glycol additive, which acts as a leveler because it is an ohmic block. The action of Bi instead appears more like the direct plating additives used to assist Cu deposition on Ru.50 A separate effect of Bi may also be an enhanced surface diffusion at the growth front. In particular, to account for the smoother surfaces of the disconnected nanocrystals in the Bi-

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.7b04640. Materials and methods, discussion of experimental parameters for electrochemical deposition of ZnAu dendrites in the continuous flow cell, and details of TEM movies (PDF) Bright field TEM movies of growth interface evolution in additive-free Zn solution (Movie M1) and Bi-additivecontaining Zn solution (Movie M2) (ZIP)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. ORCID

Jeung Hun Park: 0000-0002-8797-1293 Hariklia Deligianni: 0000-0003-2871-2803 1097

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098

Letter

Nano Letters Author Contributions

(26) Ring, E. A.; Peckys, D. B.; Dukes, M. J.; Baudoin, J. P.; de Jonge, N. J. Microsc. 2011, 243, 273−283. (27) Kubiak, T. Static and dynamic buckling of thin-walled plate structures; Springer International Publishing, 2013; pp 27−47, DOI: 10.1007/978-3-319-00654-3_2. (28) Zhang, X.; Wang, G.; Liu, X.; Wu, H.; Fang, B. Cryst. Growth Des. 2008, 8, 1430−1434. (29) Fang, J.; You, H.; Kong, P.; Yi, Y.; Song, X.; Ding, B. Cryst. Growth Des. 2007, 7, 864−867. (30) de Jonge, N.; Ross, F. M. Nat. Nanotechnol. 2011, 6, 695−704. (31) Yano, M.; Fujitani, S.; Nishio, K.; Akai, Y.; Kurimura, M. J. Power Sources 1998, 74, 129−134. (32) Schneider, N. M.; Park, J. H.; Norton, M. M.; Ross, F. M.; Bau, H. H. Adv. Struct. Chem. Imaging 2017, 2, 2. (33) Park, J. H.; Schneider, N. M.; Grogan, J. M.; Reuter, M. C.; Bau, H. H.; Kodambaka, S.; Ross, F. M. Nano Lett. 2015, 15, 5314−5320. (34) Park, J. H.; Steingart, D. A.; Kodambaka, S.; Ross, F. M. Sci. Adv. 2017, 3, e1700234. (35) Park, J. H.; Reuter, M. C.; Kodambaka, S.; Ross, F. M. Microsc. Microanal. 2014, 20 (S3), 1598−1599. (36) Frankenthal, R. P.; Thompson, D. E. J. Electrochem. Soc. 1976, 123, 799−804. (37) Cadle, S. H.; Bruckenstein, S. Anal. Chem. 1974, 46, 16−20. (38) Shrestha, B. R.; Nishikata, A.; Tsuru, T. J. Electroanal. Chem. 2012, 665, 33−37. (39) In the case that the voltage is sufficiently negative, the following electrochemical reactions are expected: 2Au + 3H2O → Au2O2 + 6H+ + 6e−, E° = −1.457 + 0.0591 log[H+] or Au+ + 3H2O → Au(OH)2 + 3H + 3e−, E° = −1.511 + 0.0591 log[H+]. Therefore, the electrochemical dissolution of Au and formation of Au3+ can be described as Au → Au3+ + 3e−, E° = −1.498 + 0.0197 log[Au3+]. (40) Wang, K.; Pei, P.; Ma, Z.; Chen, H.; Xu, H.; Chen, D.; Wang, X. J. Mater. Chem. A 2015, 3, 22648−22655. (41) Ding, F.; Xu, W.; Graff, G. L.; Zhang, J.; Sushko, M. L.; Chen, X.; Shao, Y.; Engelhard, M. H.; Nie, Z.; Xiao, J.; Liu, X.; Sushko, P. V.; Lu, J.; Zhang, J.-G. J. Am. Chem. Soc. 2013, 135, 4450−4456. (42) Esfahani, M.; Zhang, J.; Durandet, Y.; Wang, J. J. Electrochem. Soc. 2016, 163, D476−D484. (43) Gallaway, J. W.; Gaikwad, A. M.; Hertzberg, B.; Erdonmez, C. K.; Chen-Wiegart, Y.-C. K.; Sviridov, L. A.; Evans-Lutterodt, K.; Wang, J.; Banerjee, S.; Steingart, D. A. J. Electrochem. Soc. 2014, 161, A275− A284. (44) Bard, A. J.; Parsons, B.; Jordon, J. Standard Potentials in Aqueous Solutions; Dekker: New York, 1985. (45) Milazzo, G.; Caroli, S.; Sharma, V. K. Tables of Standard Electrode Potentials; Wiley: London, 1978. (46) Swift, E. H.; Butler, E. A. Quantitative Measurements and Chemical Equilibria; Freeman: New York, 1972. (47) Wei, X.; Desai, D.; Yadav, G. G.; Turney, D. E.; Couzis, A.; Banerjee, S. Electrochim. Acta 2016, 212, 603−613. (48) Puckrin, E.; Slavin, A. J. Phys. Rev. B: Condens. Matter Mater. Phys. 1990, 41, 4970−4975. (49) Martínez, A. J. Phys. Chem. A 2014, 118, 5894−5902. (50) Radisic, A.; Vereecken, P. Direct Copper Plating. In Nanostructure Science and Technology, Copper Electrodeposition for Nanofabrication of Electronics Devices; Kondo, K., Rohan, N. A., Dale, P. B., Yokoi, M., Eds.; Springer Science+Business Media: New York, 2014; Chapter 7, pp 131−173. (51) Aogaki, R.; Makino, T. J. Electrochem. Soc. 1984, 131, 40−46. (52) Aogaki, R.; Makino, T. J. Electrochem. Soc. 1984, 131, 46−51. (53) Li, M.; Luo, S.; Qian, Y.; Zhang, W.; Jiang, L.; Shen, J. J. Electrochem. Soc. 2007, 154, D567−D571. (54) Gomes, A.; da Silva Pereira, M. I. Electrochim. Acta 2006, 51, 1342−1350.

F.M.R. and J.H.P. conceived the project and designed experiments. F.M.R. and S.K. supervised the project. J.H.P. performed experiments and analyzed the data. D.A.S. and H.D. provided expertise in electrochemistry and the growth model. N.M.S. provided the computational algorithm for growth front analysis. All authors discussed the results and contributed to the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge funding supports from the National Science Foundation (NSF-GOALI: DMR-1310639) and the BP Carbon Mitigation Initiative, M. C. Reuter for assistance with experimental details, and C. A. Orme and D. Yu for helpful discussions.

■

REFERENCES

(1) Li, Y.; Dai, H. Chem. Soc. Rev. 2014, 43, 5257−5275. (2) Wang, Z.-L.; Xu, D.; Xu, J.-J.; Zhang, X.-B. Chem. Soc. Rev. 2014, 43, 7746−7786. (3) Naybour, R. D. J. Electrochem. Soc. 1969, 116, 520−524. (4) Arouete, S.; Blurton, K. F.; Oswin, H. G. J. Electrochem. Soc. 1969, 116, 166−169. (5) Chin, D.-T.; Sethi, R.; McBreen, J. J. Electrochem. Soc. 1982, 129, 2677−2685. (6) Banik, S. J.; Akolkar, R. Electrochim. Acta 2015, 179, 475−481. (7) Lan, C.; Lee, C.; Chin, T. Electrochim. Acta 2007, 52, 5407−5416. (8) Wang, J. M.; Zhang, L.; Zhang, C.; Zhang, J. Q. J. Power Sources 2001, 102, 139−143. (9) Bass, K.; Mitchell, P. J.; Wilcox, G. D.; Smith, J. J. Power Sources 1991, 35, 333−351. (10) Shin, J. W.; You, J.-M.; Lee, J. Z.; Kumar, R.; Yin, L.; Wang, J.; Meng, Y. S. Phys. Chem. Chem. Phys. 2016, 18, 26376−26382. (11) Banik, S. J.; Akolkar, R. J. Electrochem. Soc. 2013, 160, D519− D523. (12) McBreen, J.; Gannon, E. J. Electrochem. Soc. 1983, 130, 1980− 1982. (13) Ross, F. M. Science 2015, 350, aaa9886. (14) Zeng, Z.; Liang, W. I.; Liao, H. G.; Xin, H. L.; Chu, Y. H.; Zheng, H. Nano Lett. 2014, 14, 1745−1750. (15) White, E. R.; Singer, S. B.; Augustyn, V.; Hubbard, W. A.; Mecklenburg, M.; Dunn, B.; Regan, B. C. ACS Nano 2012, 6, 6308− 6317. (16) Mehdi, B. L.; Stevens, A.; Qian, J.; Park, C.; Xu, W.; Henderson, W. A.; Zhang, J.-G.; Mueller, K. T.; Browning, N. D. Sci. Rep. 2016, 6, 34267. (17) Rong, G.; Zhang, X.; Zhao, W.; Qiu, Y.; Liu, M.; Ye, F.; Xu, Y.; Chen, J.; Hou, Y.; Li, W.; Duan, W.; Zhang, Y. Adv. Mater. 2017, 29, 1606187. (18) Yang, C.; Han, J.; Liu, P.; Hou, C.; Huang, G.; Fujita, T.; Hirata, A.; Chen, M. Adv. Mater. 2017, 29, 1702752. (19) Sun, K. E. K.; Hoang, T. K. A.; Doan, T. N. L.; Yu, Y.; Zhu, X.; Tian, Y.; Chen, P. ACS Appl. Mater. Interfaces 2017, 9, 9681−9687. (20) McBreen, J.; Gannon, E. J. Electrochem. Soc. 1983, 130, 1641. (21) McBreen, J.; Gannon, E. J. Power Sources 1985, 15, 169−177. (22) Chandrasekar, M. S.; Pushpavanam, M. Electrochim. Acta 2008, 53, 3313−3322. (23) Grogan, J. M.; Schneider, N. M.; Ross, F. M.; Bau, H. H. Nano Lett. 2014, 14, 359−364. (24) This current should give an average deposition rate over the surface of the active electrode of ∼0.31 nm/s (assuming 2 electrons per Zn2+ ion). (25) Holtz, M. E.; Yu, Y.; Gao, J.; Abruña, H. D.; Muller, D. A. Microsc. Microanal. 2013, 19, 1027−1035. 1098

DOI: 10.1021/acs.nanolett.7b04640 Nano Lett. 2018, 18, 1093−1098