Direct Ink Writing of 3D Zn Structures as High‐Capacity Anodes for Rechargeable Alkaline Batteries

The relationship between structure and performance in alkaline Zn batteries is undeniable, where anode utilization, dendrite formation, shape change, and passivation issues are all addressable through anode morphology. While tailoring 3D hosts can improve the electrode performance, these practices are inherently limited by scaffolds that increase the mass or volume. Herein, a direct write strategy for producing template‐free metallic 3D Zn electrode architectures is discussed. Concentrated inks are customized to build designs with low electrical resistivity (5 × 10−4 Ω cm), submillimeter sizes (200 μm filaments), and high mechanical stability (Young's modulus of 0.1–0.5 GPa at relative densities of 0.28–0.46). A printed Zn lattice anode versus NiOOH cathode with an alkaline polymer gel electrolyte is then demonstrated. This Zn||NiOOH cell operates for over 650 cycles at high rates of 25 mA cm−2 with an average areal capacity of 11.89 mAh cm−2, a cumulative capacity of 7.8 Ah cm−2, and a volumetric capacity of 23.78 mAh cm−3. A thicker Zn anode achieves an ultrahigh areal capacity of 85.45 mAh cm−2 and a volumetric capacity of 81.45 mAh cm−3 without significant microstructural changes after 50 cycles.

The relationship between structure and performance in alkaline Zn batteries is undeniable, where anode utilization, dendrite formation, shape change, and passivation issues are all addressable through anode morphology. While tailoring 3D hosts can improve the electrode performance, these practices are inherently limited by scaffolds that increase the mass or volume. Herein, a direct write strategy for producing template-free metallic 3D Zn electrode architectures is discussed. Concentrated inks are customized to build designs with low electrical resistivity (5 Â 10 À4 Ω cm), submillimeter sizes (200 μm filaments), and high mechanical stability (Young's modulus of 0.1-0.5 GPa at relative densities of 0. 28-0.46). A printed Zn lattice anode versus NiOOH cathode with an alkaline polymer gel electrolyte is then demonstrated. This Zn||NiOOH cell operates for over 650 cycles at high rates of 25 mA cm À2 with an average areal capacity of 11.89 mAh cm À2 , a cumulative capacity of 7.8 Ah cm À2 , and a volumetric capacity of 23.78 mAh cm À3 . A thicker Zn anode achieves an ultrahigh areal capacity of 85.45 mAh cm À2 and a volumetric capacity of 81.45 mAh cm À3 without significant microstructural changes after 50 cycles.
provides interconnected high surface area, whereby electrode surfaces are more accessible to the electrolyte. [18] Moreover, designed internal voids will restrict dendritic protrusion by offering space for volume expansion. [19] Several 3D Zn structures have been created through templating onto scaffolds, such as graphene foams, [20] carbon nanotube networks, [21] porous copper, [22,23] and steel mesh. [24] Nevertheless, these 3D anodes need extra manipulation to reconcile the interfacial affinity to ensure zincophilicity, while retaining a conductive and stable skeleton. [6] Untemplated 3D Zn structures, such as wires, [25] sponges, [26] foams, [27] nanoporous monoliths, [28] and alloys [29] have also been created to optimize cell performance. Unfortunately, these randomly porous structures with length scales ranging from hundreds of micrometers to tens of nanometers impede the mass transport of the electrolyte and increase the ion transfer distance. [30] Their mechanical strength is also insufficient to maintain stable cycling. [31] An improved strategy to achieve long lifespan with high active materials' utilization at high rates appears to be 3D Zn with low tortuosity, but controlled manufacturing of such a Zn anode is still a challenge. [32,33] Recently, 3D methods have been employed to freeform fabricate designed geometries and structures of electrodes and other components in energy storage devices, [34] including aqueous Zn batteries. [35][36][37][38][39] For instance, a 3D Zn metal anode with a multichannel structure is constructed by uniform electrodeposition of Zn onto a 3D current collector, which is created by electro(less) plating of Ni onto a polymer lattice from UV stereolithography. [40] Another example uses an extrusion method to make a reservoir-integrated carbon host for dendrite-free Zn anodes. [41] These structures have similar directional channels with abundant zincophilic active sites, which benefit the electric field uniformity for homogeneous Zn nucleation. [40,41] In addition, a few composite anodes, mostly composed of Zn powders, conductive carbon, and polymer binders, have also been printed for Zn batteries' application. However, a large quantity of polymer used in these composite anodes significantly affects the connection between Zn particles, increases internal resistance, and decreases the accessible surface area and porosity. [42][43][44] Here, we demonstrate a novel direct ink writing approach to additively manufacture freestanding 3D metallic Zn anodes with controllable architectures. Our method is based on the precise deposition of Zn metal inks with 200 μm-scale features. The key challenge is to simultaneously achieve microstructure adjustability and macroscopic structural integrity. To this end, we adapt an innovative colloidal-based ink and 3D printing strategy that allows the creation of porosity-tunable hierarchical Zn cellular structures with excellent electrical and mechanical strength. We then show that these printed structures work as fast rate anodes with high capacity in a Zn||NiOOH cell using a nonspillable alkaline polymer gel electrolyte.

Results and Discussion
Preparation of the ink for direct ink writing is illustrated in Figure 1a. A colloidal ink was designed by first dissolving solid acrylic block copolymers into an organic solvent system composed of tetrahydrofuran (THF, b.p. 66°C) and 2-butoxyethanol (EGBE, b.p. 171°C). The entangled polymeric chains released to form a solvated polymer solution. Zn powders were then thoroughly dispersed into this solution to generate a highly concentrated suspension. Swollen copolymers served as a viscosifier to stabilize the solid-liquid phase preventing separation, while solvent quantity adjusted the viscosity of the ink. We investigated the inks' printability by changing the composition to find the optimal ratio for each component (Figure 1b). For Zn, the volume fraction of Zn powders in the ink should be between 50 and 70 vol% to reduce shrinkage of the prints. Zn inks with a solid loading of >80 vol% became dilatant and nonextrudable due to high yield stress during flow. For the copolymer component, the concentration should be in the range of 15-25 vol% for the homogeneity of the ink. Excess copolymers will cause structural melting and collapse upon annealing. Finally, the proportion of the solvent should be limited in the range of 10-30 vol%. Excess solvent will dilute the ink, making extruded filaments spread and stacked layers slump.
The direct ink writing process is shown in Figure 1c. The as-prepared ink was loaded into a syringe barrel affixed to a micronozzle to deposit inks following a predefined tool path in a layer-by-layer manner. The ideal ink should smoothly flow through the micronozzle without clogging and "set" immediately to maintain the deposited features. Nonvolatile EGBE acted as a humectant to prevent instant drying-induced jamming and cracking. In contrast, highly volatile THF enabled the ink's fast switch from extrusion flow to partial hardening thereafter due to its rapid evaporation. The resolidified copolymers "glued" Zn powders together, leading to a pseudoplastic transition for extruded filaments for better shape retention. EGBE would also slowly dissipate during the drying process resulting in subsequent dilatant conversion. After drying, only Zn powders embedded in copolymers held the structure amongst numerous cavities on the filament surface from residual solvent removal ( Figure S1, Supporting Information). The dried samples were then heat treated to burn away all remaining organic additives by quick annealing to 600°C in air ( Figure S2, Supporting Information insert). Although thermalgravimetric analysis (TGA) was conducted in N 2 , it still indicated that the copolymers were decomposed at %400°C ( Figure S2, Supporting Information). As the temperature increased from 400 to 600°C, the Zn particles started to fuse together at particle-particle contacts and formed a surface oxidation layer that acts as a rigid shell to maintain the printed shape above the Zn melting point. [26] These annealed samples were then quickly quenched to room temperature to avoid further oxidization. Since the Zn powders mass loading was very high, the structure after annealing shrank by less than 10% linearly. The electrical resistances of the printed Zn film after annealing measured by four-point probe are about 5 Â 10 À4 Ω cm, within 100 Ω cm of the Zn intrinsic electrical resistance (%5.5 Â 10 À6 Ω cm).
Appropriate rheology is crucial for the ink to mediate optimal 3D printability. Shear thinning facilitates reliable extruding flow, and viscoelasticity enables self-support against gravity. Figure 1d shows apparent viscosities (η) of Zn inks with varying solid loading (30, 55 and 70 vol%) as a function of shear rates (γ) from 10 À2 to 10 2 s À1 . Each ink exhibited significant shear thinning behavior with respective η ranging from below 10 2 to over 10 4 Pa s at γ of 10 À1 s À1 . The η values of 55 and 70 vol% of Zn inks are around one and two orders of magnitude higher than that of the 30 vol% ink, respectively. To evaluate their viscoelastic properties, the shear stresses (σ) as a function of storage (G 0 ) and loss (G 00 ) moduli were also measured, as shown in Figure 1e. Specifically, the 30 vol% of Zn inks have a short linear viscoelastic region with a plateau value of G 0 below 10 3 Pa with a yield stress (τ y ) of only 2 Pa. In addition, its G 00 is always lower than G 0 , indicating a more liquid-like behavior after yielding. The 55 vol% Zn inks show that G 0 and τ y increase by approximately an order of magnitude to around 10 4 and 30 Pa, which indicates their ability to sustain higher τ y without losing the homogeneity. Finally, 70 vol% Zn inks show G 0 around 2 Â 10 5 Pa, τ y close to 10 2 Pa, and higher G 00 after critical stress that suggests good phase stability compared to those low-concentration inks. To quantify the inks recovery capability, we measure the G 0 evolution of the ink after preshear flow to mimic their dynamics after deposition. As shown in Figure 1f, both 55 and 70 vol% inks present immediate elastic recovery of G 0 from about 10 Pa to over 10 3 Pa in 10 min. The magnitudes of these key rheological parameters are in good agreement with those reported for other colloidal inks designed for this direct ink writing technique. [34,45] We first printed 3D simple cuboid lattices by depositing multiple orthogonal layers of parallel filaments successively ( Figure 2a). These woodpile structures were designed with a center-to-center filament spacing (L) of 800 μm and a filament Figure 1. a) Schematic illustration of the ink's preparation; b) ternary diagram showing that the composition determined inks' printability. Blue symbols and background: the ink solid-liquid phases separated upon too small amount of polymers at high solid loading; yellow symbols and background: the ink has high polymer-to-Zn powder ratio causing the structure to collapse after annealing; red symbols and background: the ink is too soft to hold the shape after printing because of high solvent concentration; purple symbols and background: the ink is too stiff to print as the solid loading is too high; green symbols and background: the good ink formulation that can be printed and maintain the shape; c) schematic diagram of the direct writing process of Zn microlattice; log-log plots of d) storage (G 0 ) (solid symbols) and loss (G 00 ) (empty symbols) modulus as a function of shear stress (σ) showing ink's viscoelastic properties; e) apparent viscosity (η) as a function of shear rate (γ) showing its shear-thinning flow behavior; f ) storage modulus (G 0 ) as a function of time after preshear showing structural recovery of composite ink with varying solid loadings of 30 vol% (gray diamond symbols), 55 vol% (gold square symbols), and 70 vol% (blue circle symbols).
www.advancedsciencenews.com www.small-structures.com diameter (d) of around 200 μm, resulting in a spacing-todiameter ratio (L/d) of 4, as well as each layer height (Δz) of 120 μm (%0.6 d). By simply changing the filament spacing, we printed dense lattices with L of 400 μm, indicating the ability to print Zn structures over a wide range of geometric densities (Figure 2b, S3, S4 and Video S1, Supporting Information). The annealed samples were covered by dense nanowires on their porous surface (Figure 2c,d). The cross section of the printed filament showed that the internal microstructure remains highly porous, a beneficial property for electrolyte penetration (Figure 2e,f ). Both X-ray photoelectron spectroscopy (XPS) (Figure 2g and S5, Supporting Information) and energydispersive X-ray (EDX) mapping ( Figure S6, Supporting Information) also showed that the 3D-printed Zn is comparable to bulk Zn and is not significantly oxidized during printing process (Table S1, Supporting Information). To quantify the mechanical strength of the Zn lattices, we conducted a series of in-plane compression tests by measuring their compressive stress (σ c ) as a function of strain (ϵ). Typical σ c -ε curves of Zn lattices with different relative density (ρ*/ρ s , ρ* is the geometric density of Zn lattices, ρ s ¼ 7.13 g cm À3 is the density of bulk Zn) are shown in Figure 2h. These curves own very similar deformation behavior, including 1) an initial linear elastic increase at low ε (<2.5%) without the presence of peak stress and 2) a plastic hardening region with ε up to 20%. Zn lattices with higher ρ*/ρ s exhibit larger flow stress. A steady increasing flow stress with no significant oscillations or serrations indicates that Zn lattices undergo plastic deformation instead of fracturing under large compression. [46] The quantitative calculation shows that they have yield strength (σ*) of 2-6 MPa, and Young's modulus of 0.1-0.5 GPa with ρ*/ρ s of 0.28-0.46, which are comparable to those reported for Zn foams produced by melt-casting methods. [31,47] The Gibson-Ashby model has also been used to disclose the yield stress dependence of an open-cell architected material on the relative density as follows (Equation (1)). [48] ρ where σ s ¼ 75 MPa is the yield strength of the bulk Zn. The value of the scaling exponent, n, depends on the nature of the deformation mode in the material: n ¼ 1 when the struts deform under axial stresses (stretch dominated), and n ¼ 2-3 when  the struts deform under bending stresses (bend dominated). [48] By scaling the relative yield strength (σ*/σ s ) with ρ*/ρ s on a logarithmic scale ( Figure S7, Supporting Information), n ¼ 1.85 is fitted, suggesting a bend-dominated deformation mechanism in Zn lattices.
To demonstrate the versatile adaptability of this ink, various 3D Zn structures with controllable shape fidelity and precision were printed. Figure 3a shows a large-area multilayer honeycomb structure with high-aspect-ratio wall features with a height/width (h/w) of 6. The repeated passes of printed layers are well bonded to one another. Their height increases linearly with layer numbers, while their width is nearly constant (Figure 3b). We also demonstrated printing of a nonorthogonal triangular lattice composed of an array of rod-like filaments confined within a triangular shape with curved edges (Figure 3c). Subsequent layers were deposited in a similar pattern except that it is rotated by 120°b etween each layer (Figure 3d). Furthermore, a cylindrical lattice with curvaceous lines was obtained by printing concentric circles in one x-y layer, while radial rods were patterned in the adjacent layers (Figure 3e). Cylindrical samples with outer radii of 10 mm and inner radii of 2 mm have been deposited through nozzles with diameter as small as 150 μm by sequential deposition of layers with alternating patterns of concentric rings and circular arrays of radially oriented rods (Figure 3f ). Large-area Zn lattice structures have also been printed, as shown in Figure S8, Supporting Information. After annealing, we found that the 4 cm Â 4 cm and 6 cm Â 6 cm structures have minimal damage, while the 8 cm Â 8 cm sample suffers structural distortion. This might come from the nonuniform solvent evaporation, and the postannealing process needs to further optimize in the future.
To assess the long-term electrochemical and mechanical stability of the printed Zn anodes, we assembled Zn||NiOOH full cells by coupling the periodic porous Zn lattice with a commercial NiOOH cathode. First, a thin Zn lattice with the dimensions of 1 cm (l) Â 1 cm (w) Â 0.5 mm (h) was used as an anode in an alkaline polymer gel electrolyte with a low volume of %0.355 mL between the anode and cathode. This cell operated stably for over 540 h at a nominal 10% Zn depth of discharge (DoD) cutoff (DoD is based on the theoretical capacity of the anode, limiting cycled electrode capacity) under high charge and discharge rates of 25 mA cm À2 with only %100 mV increase in overpotential over time ( Figure S9a, Supporting Information). Figure 4a displays the discharge capacities against the cycle number, showing the stable performance for over 650 cycles without noticeable irreversible degradation. The cell was initially set to cycle for only 500 cycles which it completed and returned to open circuit. This anode consistently delivered an average specific capacity of 73.46 mAh g Zn À1 , corresponding to a high average areal capacity of 11.89 mAh cm À2 and a volumetric capacity of 23.78 mAh cm À3 including the volume of inactive void space in the lattice. In addition, the capacity retention is also noteworthy given the fast charge rate, which is >5 times faster than other Zn||NiOOH alkaline batteries in literature [14] and does not rely on low Zn mass loadings in the anode. These advantageous cycling parameters are further enhanced using an alkaline polymer gel electrolyte, which makes the cell quasisolid state and nonspillable in compliance with transportation and safety regulations. [49] The use of polymer gel electrolytes has also been shown to suppress dendrite formation, eliminating the need for Ca-or Li-based additives to control anode-shape changes. [49] While hydrogel or semi-solidstate electrolytes are typically considered rate limiting at room temperature, the polymer gel used here does not show many negative effects for the cell cycling at 25 mA cm À2 . We believe that the ability to achieve high discharge and charge rate in this system comes from the unique architecture of the 3D-printed Zn. The polymer gel is designed to  penetrate the void space of the 3D Zn anode before gelation occurs, enabling a large surface area contact region between the electrode and electrolyte. This polymer gel incorporation fully utilizes the inherent high surface area of the 3D lattice, further enabling the high rate displayed in the cell and lowering the overvoltages that are known to lead to passivation. [14,50] As shown in Figure 4b, even at a high rate of 25 mA cm À2 , this cell still reaches an average Coulombic efficiency (CE) of 89.45% even at a high rate of 25 mA cm À2 , which we believe can be increased by lowering the discharge voltage as seen in other alkaline cells. [51] The sub-100% CE of the cell is likely a consequence of applying fast rates in the gel system and relying on gravity and the gel alone (rather than external compression) to maintain compression of the anode and cathode. The capacity delivered by the Zn lattice anode can also be translated roughly to an energy density of %120 Wh kg À1 . While this cell does not always achieve the maximum allowed DoD of 10%, the average DoD of 8.96% is still impressive, especially given that theoretical capacity is calculated assuming the anode mass is 100% metallic Zn (disregarding any initial ZnO).
The structural evolution of Zn anodes is critical for underlying cells' cycling stability. Higher Zn utilization exacerbates any morphological and chemical changes occurring at the anode, requiring a more introspective examination of the 3D print. To this point, a thicker Zn lattice with the dimensions of 1 cm (l) Â 1 cm (w) Â 1 mm (h) was paired with a NiOOH cathode for another cycling test at an even higher nominal DoD of 30% with a charge and discharge current density of 10 mA cm À2 ( Figure S9b, Supporting Information). Over 50 cycles, this cell exhibited an average specific capacity of 214.85 mAh g Zn À1 , corresponding to an areal capacity up to 85.45 mAh cm À2 and a volumetric capacity of 854.5 mAh cm À3 with an average CE of %87% (Figure 4c,d). In comparison, both Zn paste and Zn mesh electrodes were assembled in the same construction and cycled at a Zn DoD of 30% with a charge and discharge current density of 10 mA cm À2 to compare to the 3D-printed Zn ( Figure S10, Supporting Information). However, neither paste nor mesh Zn electrode could maintain even a 15% DOD, highlighting the importance of Zn geometry when using a gel electrolyte and trying to achieve a high Zn DOD. Comparison of e) areal capacity as a function of active materials mass loading and f ) cumulative capacity at varying current densities with previously reported works (Zn sponge-1, [53] Zn sponge-2, [55] Zn sponge-3, [26] NP Zn-1, [57] NP Zn-2, [28] Zn-ZnO-1, [12] Zn-ZnO-2, [51] IHCP ZnO, [54] ZnO nanorod, [56] Nano-ZnO, [52] ZnO [14] ).
www.advancedsciencenews.com www.small-structures.com A rate test was also performed on the 3D Zn anode in a cell configuration used in Figure 4d. As shown in Figure S9c, Supporting Information, the cell was first cycled six times at 5 mA cm À2 before proceeding to cycling at increasing rates of 10, 25, and 50 mA cm À2 and then decreasing to 25 and 10 mA cm À2 with 5 cycles at each current density and a set 30% DoD limit. As shown in the figure, the 3D shows an initial loss-break-in at early cycles, but has minimal capacity loss between rates. This is seen by the less-than-average 20% decrease in capacity from 10 to 50 mA cm À2 . The cell also shows the ability to cycle at the capacities when returning to slower rates, with the cell cycling at a slightly higher average capacity at 25 and 10 mA cm À2 (3.53 and 5.60% increase respectively). After 50 cycles, the macroscopic structure of the 30% (nominal) DoD anode retained its integrity. The battery was disassembled for ex situ analysis of the Zn lattice anode (discussed below). X-ray diffraction (XRD) shows that only Zn and ZnO phases existed in the pristine ( Figure S11a, Supporting Information) and cycled electrode ( Figure S11b, Supporting Information). Interestingly, the relative ratio of ZnO did not increase in the cycled cell. While not a quantitative metric, it indicates an inhibition of ZnO forming after the cell is stopped and disassembled. This further confirms a stable condition between the gel and anode, where the alkaline gel electrolyte does not degrade the printed Zn. More informative are the scanning electron microscope (SEM) images of the preand postcycled material. Compared to a pristine Zn lattice ( Figure S12a-d, Supporting Information), the anode after cycling did not show any significant dendrites on the surface, and the lattice structure survived, as shown in Figure S12e-h, Supporting Information. Overall, the ligament became larger and the pores were visibly smaller. The smaller nodule features (%10 μm) also appeared postcycling, although the wispy fine nanowires were absent. This is unsurprising given the nature of the conversion reaction occurring at the electrode interface and high Zn utilization. The stable morphology of the 3D electrode after cycling at a higher DoD explains why this anode can achieve the long cycle life.
The overall cycling parameters of our work and previous reports are summarized in Table S2, Supporting Information. Here, we analyzed two performance indicators: 1) the discharge capacity per cycle and 2) the cumulative capacity cycled (average discharge capacity Â cycle numbers) to evaluate the comprehensive performance of different cells at varying test conditions. Since the capacity values in most literature were calculated based on the mass of active materials only, without inert mass in the electrode (additives, current collector, etc.), we employed the areal capacity to reflect the specific capacity of the anode as a whole. The "thick" electrodes with higher mass loading of active materials are preferred to provide large power with small form factors. Figure 4e compares the areal capacity of different Zn anodes as a function of mass loadings, and our printed Zn anode achieved a high level of 11.89-85.45 mAh cm À2 as the electrode thickness increased from 0.5 to 1 mm, which is among the highest reported. [12,14,26,28,[51][52][53][54][55][56][57] The use of printed Zn lattices successfully breaks the tradeoff between high areal capacity, high gravimetric capacity, and high volumetric capacity. The directional pores also facilitate the ion transport and use of polymer gel electrolyte. We anticipate the areal capacity can be further increased by increasing the electrode thickness until the performance is eventually limited by ion diffusion. The discharge capacity per cycle only reflects the mass ratio of active Zn in the anode and the DoD. Therefore, the cumulative capacity cycled would be used to show the absolute energy output of this Zn anode by integrating three key parameters of cycle life, current density, and capacity retention. [58] During the whole cycling test before 30% capacity fading, this cell delivered a high cumulative capacity of 7.8 Ah cm À2 (%49 Ah g À1 ) at a high rate of 25 mA cm À2 , which is also record high compared to other Ni-Zn batteries (Figure 4f based on the data in Table S2, Supporting Information). [12,26,28,[51][52][53][54][55][56][57] Furthermore, excess electrolyte is another critical factor that can offer Zn anode with high cumulative capacity under low active material loading. The electrolyte volume normalized by the theoretical capacity of the anode (V/T) was used to show the relative amount of electrolyte used in the battery. Our high cumulative capacity was achieved at V/T of 0.0012-0.0024 mL mAh À1 , which confirms that this Zn anode was cycled in an electrolyte-limited cell. These metrics make this 3D Zn superior to many reports of advanced Zn anodes, with the added benefit of a programmable anode design. Our results validate the concept of printing a practically feasible metallic Zn electrode and pave a new way for fabricating highareal-density-capacity Zn anodes for alkaline cells.

Conclusion
In summary, we present a new direct write protocol for the 3D printing of metallic Zn microstructures with high precision, shape fidelity, and excellent material properties, and examine their use as high rate Zn anodes for alkaline cells. Direct ink writing enables micro-patterning over large length scales to form arbitrary-shaped macroscale anodes that are adaptable to complex architectures. The simultaneous control of microstructure and macrostructure is unobtainable by traditional materials processing. The main challenge for successful 3D printing of Zn lies in optimizing the formulation, adjusting ink's rheology for extrusion flow, and post-processing to avoid undesirable structural damage. By addressing these issues, printed Zn lattices are produced with properties beyond those of bulk materials. The strategy described here is facile and easy to scale up. Our method enables to explore the properties and applications of Zn in a self-supporting, structurally tunable and 3D macroscopic form. Furthermore, our 3D printing approach can introduce multimaterials into the structure, which promises another level of structural control for better integration of full batteries assembly. Future work will focus on the acceleration of the design-fabricatetest innovation cycle, guided by inverse design optimization to extend the cycle life of zinc anodes with a high loading amount of active material, high DoD per cycle, and limited electrolytes.

Experimental Section
Ink Preparation: 2 g poly(methyl methacrylate)-poly(acrylate)-poly(methyl methacrylate) (PMMA-PA-PMMA) pellets were fully dissolved into 2 g THF to generate a polymeric solution. Then, 26.7 g Zn powders were mixed with this solution to form a colloidal suspension for 1 min at 2000 rpm using a planetary mixer (ARE-250, Thinky). High-speed mixing concentrated the suspension through the evaporation of THF to 1 g. Finally, www.advancedsciencenews.com www.small-structures.com 1 g EGBE solvent was added to yield a composite ink composed of a total solid loading of >70 vol% with balanced THE:EGBE of 1:1. Rheology Characterization: Viscoelastic properties of the Zn ink were measured using a stress-controlled rheometer (AR 2000ex, TA Instruments) with a 40 mm flat-plate geometry and a gap of 500 μm in the presence of solvent trap to avoid solvent evaporation. A strain sweep from 10 À1 to 10 2 s À1 was first performed to record η as a function of varying γ. A stress sweep from 10 À2 to 10 3 Pa at a constant frequency of 1 Hz was also conducted to record the G 0 and G 00 as a function of sweep stress. τ y is defined as the stress where G 0 fell to 90% of the plateau value. The creep data of G 0 was recorded after applying preshear at 100 s À1 for 1 min as a function of time from 0.1-10 min to evaluate the structural recovery of ink after deposition during 3D printing.
Fabrication of Zn Anodes: The Zn ink was loaded into a 10 mL syringe barrel (Nordson EFD) attached by a Luer-Lok to a smooth-flow-tapered conical nozzle [inner diameter (ϕ), 200 μm]. An air-powered fluid dispenser (Ultimus V, Nordson EFD) provided the appropriate pressure to extrude the ink through the nozzle. The target patterns were printed using a three-axis positioning stage (ABL9000, Aerotech). The 3D Zn green parts were printed onto an alumina plate with an initial nozzle height of 120 μm to ensure adhesion to the substrate. The required pressure depended on the ink rheology, nozzle size, and printing speed, and typical values ranged from 60 to 90 psi at 5-15 mm s À1 . The printed parts were dried at room temperature overnight in the air or soft baked on a hot plate at 80°C for 2 h. Then, they were annealed in a box furnace (KSL-1100X-S-UL-LD, MTI) by heating from room temperature to 600°C in 60 min and holding for 20 min in air. The typical printed sample xyz dimensions were 10 mm Â 10 mm Â 0.5-1 mm. The linear shrinkage (typically 10%) was calculated from longitudinal dimensions prior and after sintering.
Material Characterization: The dimension and weight of the samples were determined with a caliper with an accuracy of 0.01 mm and an ultramicro balance (XP24, Mettler Toledo) with an accuracy of 0.001 mg. The geometric density was calculated from the measured mass and volume of each specimen. 3D Zn lattices with varying geometric densities were produced with dimensions of %6 mm Â 6 mm Â 3 mm (height) for mechanical tests. The uniaxial compression of these Zn lattices was performed under quasistatic conditions using a universal testing machine (Instron 5943) at a strain rate of 1.1 Â 10 À3 s À1 at room temperature. The yield strength (σ*) of the Zn lattice was calculated by the 0.2% offset method (the proof stress corresponding to a permanent plastic strain of 0.2%). The morphology of the printed Zn was observed by an optical camera and field-emission SEM. SEM and EDS measurements were performed on a JEOL 7401-F at 10 keV (20 mA) in secondary-electron imaging mode with a working distance of 2-8 mm. Secondary-electron X-ray-induced image (SXI) of Zn/ZnO showed area of analysis. XPS was performed on a PHI Quantum 2000 Scanning ESCA Microprobe using Al Ka X-ray (1486.6 eV). The real-time videos were obtained by a charge coupled device camera (Thorlabs). Electrical conductivity was measured using the four-probe method. A TGA Q500 (TA Instruments) was used to analyze the degradation temperature of the composite Zn ink. Samples of %0.1 g were heated to 150°C under vacuum (10 À5 Torr) for at least 24 h to remove all absorbed species, cooled under vacuum, and then heated to 450°C at a rate of 1°C min À1 in nitrogen environment.
Battery Assembly: A Zn lattice was combined with a NiOOH cathode and alkaline polyacrylate gel electrolyte to assemble a full Zn||NiOOH battery. To make electrical connection to the anode, Cu mesh was tightly wrapped around the printed lattice and a Ni tab spotwelded for connection. The anode and a commercial NiOOH cathode (Jiangsu Highstar Battery Manufacturing) were then paired together with a cross-linked polyvinyl alcohol separator (25 μm in thickness, received from Urban Electric Power) and a Celgard 5150 separator placed between the electrodes to prevent cell shorting. The electrode-separator sandwich was then placed horizontally in a polypropylene case (purchased from LA containers). The alkaline acrylate gel electrolyte was made using an adapted procedure from a previous report. [49] Briefly, 10 mL of 42 wt% KOH saturated with zincate (5 wt%) in a plastic beaker was stirred by a magnetic stir bar while being cooled in an ice water bath. Separately, 0.01 g of N,N'-methylenebisacrylamide (MBA) was dissolved in 2 mL of acrylic acid. The MBA acrylic acid solution was then added drop wise into the chilled KOH solutions, ensuring that the exothermic reaction of acid and base did not increase the solution temperature, resulting in a solution of 36 wt% KOH. After %3 min, 0.001 g of potassium persulfate was added to the solution. Upon addition of this oxidant, polymerization was initiated. To ensure usability of the gel, 6 mL of solution was added to the cases with the Zn and NiOOH electrodes %1 min after the initiator was added to the constantly stirred MBA-AA-KOH solution. This amount of time was found sufficient to allow pourability of the solution but still ensure thorough mixing, so a gel formed once added to the case. The case volume was oversized compared to the electrode volumes, requiring excess gel to fill void space.
Electrochemical Characterization: Cells were tested on a MACCOR 4000 series %20 h after gel was added. Cells were galvanostatically cycled with an initial charge-discharge-charge of 5 mA cm À2 before switching to a rate of 25 mA cm À2 for discharge and charge. A cutoff voltage of 1.5 V for discharge and 1.93 V for charge along with a capacity limit of 10% were set as bounds during cycling. During charge if the desired capacity was not reached, a voltage hold of 1.9 V was initiated for 3 h or till the current dropped below 10%. Capacity of the cells was based off the mass of the 3D-printed Zn anode, using the assumption that the anode was 100% Zn. For cells used in ex situ analysis the initial charge-dischargecharge of 5 mA cm À2 was kept but then switched to 10 mA cm À2 for continued cycling. The capacity limit was also set for 30%, but the other conditions were kept the same.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.