Stereolithgraphy of Metallic Electrode with Janus Porosity toward Controllable Bubble Behavior and Ultra‐Stable Water Electrolysis

The catalytic performance and the stability of the water electrolysis system have perplexed the practical splitting water. Herein, 3D‐printed metallic electrodes with Janus porosity are reported, which can achieve the efficient bubble emission and provide high surface area. After a further one‐step treatment, the catalysts of Ru–Ni(OH)2 and Fe–Ni(OH)2 nanoarrays are in situ grown on the electrodes. Theoretical calculations and systematical experiments confirm the synergistic effect of the gradient morphology and the in situ grown catalysts. The electrolyzer shows unprecedented activity with an overpotential of 98 mV for hydrogen evolution reaction and 343 mV for oxygen evolution reaction at the current density of 500 mA cm−2 and outperformed stability, which can deliver 500 mA cm−2 at the voltage of 1.638 V for 2100 h with no significant decay, far exceedingly most state‐of‐the‐art electrolyzers. This 3D electrode with controllable bubble motions offers a viable solution for future industrial sustainable hydrogen production technology.


Introduction
A surge in population is bound to be accompanied by a sharp depletion of fossil fuels and environmental pollution, which means that the scientific society needs to address the issues of renewable energy development and storage. [1]The electrochemical process of water splitting for the production of hydrogen represents an intriguing prospect for fostering sustainable and eco-friendly renewable energy solutions. [2][5][6][7][8] As one of the most important performance indicators, the overall overpotential can be divided into three parts: the overpotential required to activate the reaction, the overpotential of the intrinsic cell Ohmic drop due to the circuit resistance and resistance between electrode and electrolyte, and the transmission overpotential caused by the ongoing concentration gradient in the bulk electrolyte or on the electrode surface. [9][20][21][22] In contrast, the activation overpotential can also be decreased by increasing the effective active area of the electrodes.[25][26][27] Moreover, the formation The catalytic performance and the stability of the water electrolysis system have perplexed the practical splitting water.Herein, 3D-printed metallic electrodes with Janus porosity are reported, which can achieve the efficient bubble emission and provide high surface area.After a further one-step treatment, the catalysts of Ru-Ni(OH) 2 and Fe-Ni(OH) 2 nanoarrays are in situ grown on the electrodes.Theoretical calculations and systematical experiments confirm the synergistic effect of the gradient morphology and the in situ grown catalysts.The electrolyzer shows unprecedented activity with an overpotential of 98 mV for hydrogen evolution reaction and 343 mV for oxygen evolution reaction at the current density of 500 mA cm À2 and outperformed stability, which can deliver 500 mA cm À2 at the voltage of 1.638 V for 2100 h with no significant decay, far exceedingly most state-of-the-art electrolyzers.This 3D electrode with controllable bubble motions offers a viable solution for future industrial sustainable hydrogen production technology.
of bubbles covering the electrode surface will block the efficient active area and then increase the resistance between the electrode and the electrolyte, leading to an ongoing concentration gradient, especially at high current densities, which increases the Ohmic overpotential and transmission overpotential of the reaction. [28,29]As the surface covered by bubbles reduces the effective active sites, the potential for electrodes during the water splitting can be expressed as follows [30] where the η θ is the overpotential with the effect of the bubbles taken into account, η is the overpotential without the effect of the bubbles, and θ is the ratio of surface coverage by the bubbles, ranging between 0 and 1.Therefore, it is critical to design a structure that facilitates the rapid emission of bubbles to reduce the Ohmic overpotential and transmission overpotential due to bubble blockage and accumulation. [31][34][35] Triply periodic minimal surface (TPMS), as a typical 3D-ordered structure, is a mathematically defined structure that repeats in three dimensions with zero mean curvature and can therefore be expected to be particularly suitable for enhanced gas evacuation. [36]Meanwhile, the pore gradient structure was shown to be capable of achieving efficient bubble transport. [37]This enables bubble-attached active sites to be reused in the reaction, bringing it closer to its intrinsic electrochemical active sites.The gyroid structure, being one of the TPMS structure, exhibits a homogeneous distribution of curvature, good porosity, and the ability to modulate pore dimensions via parameter tuning.Therefore, it is of great interest to design and fabricate electrodes with gyroid structure and periodically Janus porosity (gradient porosity of two porosity variation) to achieve efficient bubble emission, which can well reduce the transmission overpotential of the reaction.
In addition to catalytic activity, long-term stability is another crucial parameter for practical applications. [38][43][44] Although the catalytic reaction electrodes prepared by these methods show excellent catalytic performance, the catalysts may fall off the electrodes surface during continuous catalytic reactions at high current densities, thus reducing the catalytic efficiency and making it difficult to ensure durability. [45]In contrast, the in situ growth of self-supporting electrodes with seamless integration between the catalysts and the electrode substrate enables the catalysts to operate stably at high current densities over long periods. [46]herefore, the construction of self-supporting electrodes is an effective way to improve electrocatalytic stability.
To facilitate the preparation of low-cost and efficient electrodes, such a technique is needed to meet the requirements of both preparable catalysts materials with low cost and 3D hierarchically structures with Janus porosity.The adoption of gradient electrodes ensures both a sufficiently large electrochemical active surface area (ECSA) while accelerating the emission of gas bubbles through the internal pressure differential driving force generated by the gradient structure.With the rapid development of 3D-printing technology, [47] as an emerging 3D-printing technology, digital light processing (DLP) technology is regarded as a revolutionary technology in a wide variety of applications with low cost and high precision. [48]In addition, component regulation and structure design properties conferred by DLP can allow the fabrication of transition metal nickel (Ni)-based electrodes, and also enable the facile preparation of 3D hierarchically gradient structures.
In this work, we have constructed a 3D-printed self-supporting electrode with Janus porosity followed by in situ corrosionassisted in situ growth of Ru-Ni(OH) 2 and Fe-Ni(OH) 2 .The gradient structure achieves efficient 3D bubble transport to promote the contact between electrode and electrolyte with the help of internal differential pressure drive, while the method of in situ growth gives the catalysts excellent stability.In synergy with the in situ growth method and gradient structure design, it is corroborated that the electrodes reveal superior catalytic performance for hydrogen evolution reaction (HER) with the low overpotential (98 mV to achieve 500 mA cm À2 ) and oxygen evolution reaction (OER) (343 mV to achieve 500 mA cm À2 ), respectively.Given the excellent HER and OER performance, the electrolyzer consisting of Ru-Ni(OH) 2 and Fe-Ni(OH) 2 exhibits the outperformed voltage and stability which can deliver 500 mA cm À2 at the voltage of 1.638 V for 2100 h with negligible attenuation.This work addresses the issue of bubble distress at high current densities and lights up the shadow of fabricating the electrodes for efficient and long-time stable hydrogen production in the future industry.

Results and Discussion
For a given porous structure, the number of active sites is closely related to the surface area, while the surface area varies with the porosity.Therefore, for the purpose of designing such a porous structure with both high surface area to ensure a sufficient number of active sites and high porosity to allow efficient emission of bubbles, the relationship between surface area and porosity for the gyroid structure compared to other periodic structures were calculated as shown in Figure 1a.It can be seen that the surface area of the gyroid increases with increasing porosity, allowing the structure to be designed with a large enough pore size to ensure the timely emission of bubbles while maintaining a large surface area.More directly, the ability of the bubble escape from the structure can be directly expressed in terms of the escape velocity of the bubble in the structure.Therefore, to explore the suitable pore size of gyroid structure that can consider both high surface area and bubble escape velocity, the surface area and bubble velocity of the gyroid structure with a small pore size (note as 3DP Ni-S), large pore size (note as 3DP Ni-L), and gradient pore size, referred to as Janus porosity, which is composed of small pores on one side and large pores on the other (note as 3DP Ni-G), were also compared as shown in Figure 1b.The specific pore size information is shown in Figure S1, Supporting Information.It can be concluded that the surface area of the 3DP Ni-G is close to that of the 3DP Ni-S and much higher than that of 3DP Ni-L, which can ensure enough active sites involved in the reaction.In addition, the bubble velocity of 3DP Ni-G is much higher than that of 3DP Ni-S and 3DP Ni-L, indicating that the gradient structure can promote the rapid escape of bubbles and thus achieve efficient bubble emission.Therefore, the gyroid structure with Janus porosity was fabricated by DLP technology due to the desirable surface area and efficient bubble behaviors.The following corrosion engineering for in situ growth of active materials is shown in Figure 1c.Specifically, the Ni-based gradient electrodes with Janus porosity were simply submerged in RuCl 3 and FeCl 3 solutions at room temperature, respectively.Chlorine ions will adsorb on the passivation film of the metal electrode, thus destroying the film and accelerating the corrosion of the Ni electrode.Moreover, hydroxide ions, which are generated through the reaction between water and dissolved oxygen in the solution, will migrate toward the surface of the 3DP Ni electrode.Owing to the inclusion of Ru 3þ and Fe 3þ ions into the chloride ion-containing solution, layered hydroxides containing Ru and Fe will spontaneously precipitate onto the surface of the metal electrode, respectively.The proportion of Ni and Ru content was presented in Table S1, Supporting Information, and the ratio of Ni and Fe content was shown in Table S2, Supporting Information.This approach is simple and effective without any toxic chemicals involved in the reaction, and the in situ growth of the active material allows for a seamless bond between the catalyst and the electrode substrate. [49]he electrocatalytic performances toward HER of Ru-Ni(OH)  2b), the nuanced distinctions among them may originate from the fact that the sparser bubble coverage in the gradient structure results in a greater number of efficient active sites being engaged in the reaction, thereby facilitating faster reaction kinetics.Consequently, this may give rise to a Tafel slope that is comparatively reduced.The double-layer capacitance (Cdl) value of Ru-Ni(OH) 2 /3DP Ni-G is 94.5 mF cm À2 , lying between the values of Ru-Ni(OH) 2 / 3DP Ni-S (128.2 mF cm À2 ) and Ru-Ni(OH) 2 /3DP Ni-L (87.5 mF cm À2 ) in Figure 2c and S4, Supporting Information, which verifies the Ru-Ni(OH) 2 /3DP Ni-G can maintain a large enough ECSA to promote the reaction.Meanwhile, the low resistance of the electrodes for HER makes the electrode favorable for electron transfer, thus accelerating the mass transfer process (Figure S5, Supporting information).The result indicates that Ru-Ni(OH) 2 /3DP Ni-G exhibits remarkable catalytic performance in synergy with catalyst and structure, which is superior to the recently reported catalyst materials (Table S3, Supporting Information).Concerning the OER performance, the Fe-Ni(OH) 2 /3DP Ni-G displays the lowest overpotential of only 343 mV to achieve the current density of 500 mA cm À2 (Figure 2d) compared to the overpotential of 359 mV in Fe-Ni(OH) 2 /3DP Ni-S and overpotential of 387 mV in Fe-Ni(OH) 2 /3DP Ni-L.It also displayed better catalytic activity when compared to the commercial RuO 2 catalyst (Figure S6, Supporting Information).In addition, it trumps most of the recently reported OER catalysts (Table S4, Supporting Information).Furthermore, it was anticipated that the 3DP Ni-S electrode, due to its greater surface area, would demonstrate the most minimal overpotential value.However, the observed overpotential values for the 3DP Ni-G electrode were analogous to 3DP Ni-S, suggesting that a greater number of highly efficient active sites are being involved in the reaction at the 3DP Ni-G electrode, which reduces the overpotential required for the reaction to proceed.Consequently, it manifests overpotential values that are comparable to those exhibited by the  and thus improve the reaction activity (Figure S9, Supporting Information).
Given the outstanding performance of HER and OER, the potential of the electrolyzer for overall water splitting was demonstrated by using the Ru-Ni(OH) 2 /3DP Ni-G electrode as cathode and Fe-Ni(OH) 2 /3DP Ni-G as an anode in a two electrodes electrolyzer with a 1 M KOH electrolyte.Impressively, the Ru-Ni(OH) 2 /3DP Ni-G || Fe-Ni(OH) 2 /3DP Ni-G attains a current density of 500 mA cm À2 at a cell voltage of only 1.638 V, surpassing most of the recent works (Table S5, Supporting Information).In addition to good activity, excellent stability is also of great significance for energy conversion.The Ru-Ni(OH) 2 /3DP Ni-G || Fe-Ni(OH) 2 /3DP Ni-G can operate stably for more than 2100 h at a current density of 500 mA cm À2 with almost negligible fluctuations in Figure 2g.Meanwhile, the increase in overpotential is only 58 mV after the long-term stability test in Figure 2h, illustrating the superior durability of the device.Moreover, the electrolyzer consisting of only bare 3DP Ni-G is also able to operate stably at the current density of 500 mA cm À2 for more than 2300 h with sufficient performance retention in Figure S10, Supporting Information.Such catalytic performance and stability show great potential for practical applications (the current density of 1000 mA cm À2 at an overpotential of 300 mV and long-term durability over 1000 h for industrial applications). [8]o elucidate the possible sources of the excellent catalytic properties, the surface morphology and surface chemical state and composition were further investigated.The microstructures of Ru-Ni(OH) 2 /3DP Ni-G and Fe-Ni(OH) 2 /3DP Ni-G were characterized by Scanning electron microscope (SEM) and high-resolution transmission electron microscopy (HRTEM).
In Figure 3a, it can be seen the layer hydroxides are uniformly distributed on the substrate.Even after long-term stability testing, the microscopic morphology remains layered and was not severely damaged (Figure S11, Supporting Information).The transmission electron microscope (TEM) images illustrate the uniform distribution of metallic Ru nanoparticles on the Ni(OH) 2 surface (Figure S12, Supporting Information).X-ray photoelectron spectroscopy (XPS) spectra were performed to further evaluate the material composition and surface valence of the active material.In Figure 3b, two peaks located at approximately 463.2 and 485.4 eV can correspond to Ru 0 3p 3/2 and Ru 0 3p 1/2 , while the other two peaks are assigned to the Ru xþ 3p 3/2 and Ru xþ 3p 1/2 , suggesting that Ru exists partly as metal nanoparticles in addition to the oxidized valence state.The doped Ru could induce modification of the electronic structure of Ni(OH) 2 , which may optimize the adsorption energy of reaction intermediates.In contrast, the Ru atoms would function as the active centers toward electrocatalytic reactions to accelerate the reactions. [50]For the Ni 2p spectra (Figure 3c) and O 1s spectra (Figure 3d), it can be concluded that the successful preparation of Ni(OH) 2 .Similarly, it can be seen the layered hydroxide is distributed on the Ni particles in Figure 3e.The microstructure remains intact even after a long-term stability test, demonstrating excellent durability, as shown in Figure S13, Supporting Information.In the XPS spectra of Fe 2p (Figure 3f ), the peaks appealing at 712 and 725.2 eV are indexed to the existence of the Fe 3þ . [51]Meanwhile, the XPS spectra of Ni 2p (Figure 3g) and O 1s (Figure 3h) also prove the successful preparation of Ni(OH) 2 .As confirmed by TEM images of Fe-Ni(OH) 2 , it reveals that the Fe 3þ was uniformly embedded in the layered hydroxide surface (Figure S14, Supporting Information), which can be served as an active site for the reaction and improve the electronic structure of Ni(OH) 2 , thus accelerating the reaction kinetics. [52]Notably, by comparing the SEM images of the samples before and after the growth of the active material, it is clear that the deposition of both Ru-Ni(OH) 2 and Fe-Ni(OH) 2 occurs solely on the surface of Ni particles, taking the form of layered hydroxides.This deposition process does not cause any modification to the microscopic porous structure of the underlying substrate.Moreover, the results of BET testing demonstrate that the surface area of 3DP Ni, Ru-Ni(OH) 2 /3DP Ni, and Fe-Ni(OH) 2 /3DP Ni are 31.89,24.796, and 21.115 m 2 g À1 , respectively (Figure S15, Supporting Information).The observed reduction in surface area can be attributed to the decrease in pore size below 10 nm.Analysis of the pore size distribution diagram reveals that the growth of active materials modified the microstructure from one dominated by micropores below 10 nm to one dominated by micropores of 45 nm.This alteration may result from the layered active materials' coverage of most of the micropores below 10 nm, while the structure continues to exhibit a microscopic porous structure that is principally governed by micropores of 45 nm.
To gain a deep insight into the catalytic mechanism, density functional theory (DFT) calculations were used to clarify the excellent performance of Ru-Ni(OH) 2 in Figure 3i and Fe-Ni(OH) 2 in Figure 3k, respectively.In general, the hydrogen adsorption-free energy (ΔG*H) is considered the key index to determine the activity of HER.In Figure 3j, the Ru-Ni(OH) 2 exhibits a much lower |ΔG*H| value of À0.414 eV than the value of 1.189 eV for Ni(OH) 2 , even lower than the value of À0.786 eV for Pt (111), which represents that it has excellent hydrogenadsorption capacity. [53]In addition, the dissociation of *H 2 O into *OH and *H is also a crucial step (Volmer step) in determining HER activity. [54]According to the calculation results, the free energy change of Ru-Ni(OH) 2 is a downhill change in Figure S16, Supporting Information, suggesting that the synergistic effect of Ru and Ni(OH) 2 can accelerate the HER dynamics by optimizing the electronic structure to enhance the electrical conductivity while lowering the free energy barrier of each intermediate state, thus improving the HER performance.Furthermore, the charge density distribution around Ni(OH) 2 and Ru-Ni(OH) 2 was analyzed as shown in Figure S17, Supporting Information, where the yellow and cyan colors represent areas of electron accumulation and depletion.In contrast to the charge density of Ni(OH) 2 , the Ru-Ni(OH) 2 complex exhibits a greater abundance of yellow and cyan regions, which suggests a robust electronic interaction and electron rearrangement between Ru and Ni(OH) 2 .This notable charge exchange density between the Ru and Ni(OH) 2 species plays a key role in promoting an efficient HER process.The theoretical calculation results show that in synergy with Ru and Ni(OH) 2 , the Ru-Ni(OH) 2 exhibits optimized hydrogen adsorption-free energy and the strong charge exchange density between the Ru and Ni(OH) 2 ensures an efficient HER process.Concerning the reaction kinetics of OER, DFT calculations were then carried out to confirm the change in adsorption energy of the Fe-Ni(OH) 2 for the intermediates in the four-step OER mechanism.The corresponding optimized model structures of Fe-Ni(OH) 2 for OER are displayed in Figure S18, Supporting Information.The thermodynamic analysis demonstrates that the first and fourth steps are exothermic processes, while the formations of *O and *OOH show an uphill energy change.Therefore, the formation of *O from *OH is the rate determining step of the reaction. [55]As shown in Figure 3l, the Fe-Ni(OH) 2 processes a lower |ΔG| value of 0.47 eV than that of Ni(OH) 2 (1.09 eV), which is evidenced that the addition of Fe favors to reduce of the reaction barrier of Ni(OH) 2 and further promotes the OER reaction.
In addition to the high activity of the material, the influence of the different bubble behaviors in the electrode on the performance is worth investigating.To further clarify the possible cause of bubble behaviors, in situ microscopic observation of bubble behaviors in the different structures at the same voltage was performed.In Figure 4a-c, the quantity of the Fe-Ni(OH) 2 / 3DP Ni-L is significantly higher than that for Fe-Ni(OH) 2 /3DP Ni-S, while the diameter of the bubbles in Fe-Ni(OH) 2 /3DP Ni-S is much larger than that of Fe-Ni(OH) 2 /3DP Ni-L.It is noteworthy that both the bubbles size and the number of bubbles in Fe-Ni(OH) 2 /3DP Ni-G are close to those in Fe-Ni(OH) 2 /3DP Ni-L, illustrating that Fe-Ni(OH) 2 /3DP Ni-G and Fe-Ni(OH) 2 /3DP Ni-L have similar bubble escape behaviors due to the large pore size.The same situation can be observed in bare 3DP Ni (Figure S19, Supporting Information).The massive increase in the number of bubbles can be from the higher effective active sites of Fe-Ni(OH) 2 /3DP Ni-L and Fe-Ni(OH) 2 / 3DP Ni-G without bubbles attachment.For Fe-Ni(OH) 2 /3DP Ni-S, the smaller quantity of bubbles and larger size bubbles illustrate the bubbles tended to be trapped into large bubbles.And, the active sites will be covered due to the formation of large bubbles adhering to the electrode surface, resulting in a reduced number of effective active sites.The difference in bubble size may correspond to different overpotentials since the large bubbles on the electrode surface will cover active sites and isolate the contact between the electrode and electrolyte.Meanwhile, the bubbles attached to the electrode surface will induce the ongoing concentration gradient on the electrode surface, affecting the mass transfer process and then increasing the transfer overpotential.
Generally speaking, the total overpotential in water splitting consists of three components, and the relationship between them can be illustrated by the following formula [9] where the η act is the overpotential required to activate the catalytic reaction, which can be obtained by Tafel fitting of the LSV curves (Figure S20, Supporting Information). [28]η ohm,cell and η trans are closely connected with the bubble behaviors (growth and detachment), especially at high current densities.To determine whether the gradient structure favors the transmission overpotential decrease, the overpotential fluctuation and resistance fluctuation were carried out.As is shown in Figure 4d, the overpotential at a constant current of 0.  Fe-Ni(OH) 2 /3DP Ni-G.In addition, when trying to reverse the placement of the small aperture side up and injecting bubbles from the large aperture to observe the bubble escape behaviors in Figure 5d-f, the bubble is unable to pass through such a gradient structure and thus escape from the side of the electrode.The escape time and size both have increased to 133.3 ms and 1.07 mm, respectively.Combining the previous results, it suggests that there is a process of bubble aggregation and growth in Fe-Ni(OH) 2 /3DP Ni-S, resulting in the slow bubble emission and the blockage and aggregation of larger bubbles.One proposed cause for the smaller bubble and larger number of bubbles in Fe-Ni(OH) 2 /3DP Ni-G could be the beneficial effects of the gradient structure in bubble escape behaviors.
To intuitively explain the effect of pore size gradient structure on bubble release inside the electrode, the pressure distribution and velocity distribution of the bubbles inside the gradient structure are simulated.In Figure 5g and S27, Supporting Information, it can be concluded that the internal pressure of small pore structure is significantly higher than that of large pore structure in 3DP Ni-G, resulting in a significantly sudden pressure difference within the gradient structure to drive the hydrogen bubble transport from the small pore structure to the large pore structure.In contrast, the pressure is changing gradually in 3DP Ni-L and 3DP Ni-S (Figure S28 and S29, Supporting Information) with no sudden differential pressure.However, it is interesting that the largest pressure difference was established in the 3DP Ni-S, while the averaged velocity (hydrogen flow escape velocity) is the lowest among the three structures.From the simulation plot of the velocity distribution within both structures in Figure 5h and S30-32, Supporting Information, it can be found that the average velocity within the 3DP Ni-G (10.1 μm s À1 ) is 2.1 and 2.4 times higher than that within the 3DP Ni-L (4.8 μm s À1 ) and 3DP Ni-S (4.26 μm s À1 ), respectively.The lowest velocity in 3DP Ni-S can be explained by the longest flow path in the 3DP Ni-S in Figure S32, Supporting Information.The densest gyroid structure of 3DP Ni-S among the three structures induces the largest flow resistance in it, which greatly damps the pressure driving force and results in the lowest hydrogen escape velocity from the electrode.The high-pressure driving force and low flow path ensure the higher average velocity in 3DP Ni-G, which clearly proves the highest mass transfer efficiency in the gradient gyroid structure.Moreover, for 3DP Ni-G, the average velocity of the bubble inside the upper side large pore structure is about 12.3 μm s À1 , which is 62.9% higher than inside the lower side small pore structure in Figure 5i, which guarantees higher hydrogen escape velocity from the electrode.This is caused by the large pressure difference between the two parts, which plays a dictating role in the hydrogen transportation process.In consequence, the gradient structure can effectively accelerate the emission of bubbles to reduce the phenomenon of blocking and aggregation caused by inefficient bubble escape and to avoid the increase of additional overpotential caused by bubble attachment.

Conclusion
In summary, novel metallic electrodes with Janus porosity by DLP technique have been proven to be able to achieve efficient bubble emission, which ensures that the electrode active sites are close to intrinsic value and reengage in the reaction, thus reducing the transport overpotential.After in situ fabrication of catalysts on the surface, the electrodes presented excellent catalytic performance and durability.The Ru-Ni(OH) 2 /3DP Ni only required an overpotential of 98 mV to attain 500 mA cm À2 for HER, and the Fe-Ni(OH) 2 /3DP Ni achieved an overpotential of 343 mV at 500 mA cm À2 for OER.The assembled electrolyzer consisting of Ru-Ni(OH) 2 and Fe-Ni(OH) 2 provided a cell voltage of 1.638 V at a current density of 500 mA cm À2 for 2100 h with no visible performance degradation, which demonstrates extreme competitiveness in most advanced alkaline electrolyzers.The gradient electrodes with self-supporting catalysts showed great promise for industrial hydrogen productions and would offer a viable solution for future carbon neutral energy systems.

Figure 1 .
Figure 1.The comparison of a) surface area for different periodic structures and b) surface area and bubble velocity for gyroid structures with different pore sizes.c) Schematic illustration of self-supporting electrodes.
3DP Ni-S electrode (Figure S7, Supporting Information).Moreover, the Tafel slope of Fe-Ni(OH) 2 /3DP Ni-G was measured to be 40.7 mV dec À1 (Figure 2e), where the value for Fe-Ni(OH) 2 /3DP Ni-S and Fe-Ni(OH) 2 /3DP Ni-L is 53 and 57.6 mV dec À1 , respectively.The smallest Tafel slope indicates that Fe-Ni(OH) 2 /3DP Ni-G has a superior kinetics process for OER.Furthermore, as depicted in Figure 2f and S8, Supporting Information, the Fe-Ni(OH) 2 /3DP Ni-G exhibits the Cdl value of 101.3 mF cm À2 , where the value is intermediate between Fe-Ni(OH) 2 /3DP Ni-S (103.7 mF cm À2 ) and Fe-Ni(OH) 2 /3DP Ni-L (100.8 mF cm À2 ), describing the initial factor of the excellent OER performance.In addition, the low resistance of the electrodes can allow effective electron migration

Figure 2 .
Figure 2. a) Linear sweep voltammetry (LSV) curves of Ru-Ni(OH) 2 /3DP Ni with different structures for hydrogen evolution reaction (HER) in 1 M KOH.b) Tafel slopes.c) The corresponding fitting plots Cdl of Ru-Ni(OH) 2 /3DP Ni with different structures for HER.d) LSV curves Fe-Ni(OH) 2 /3DP Ni with different structures for oxygen evolution reaction (OER) in 1 M KOH.e) Tafel slopes.f ) The corresponding fitting plots Cdl of Fe-Ni(OH) 2 /3DP Ni with different structures for OER.g) Chronopotentiometry curves of Ru-Ni(OH) 2 /3DP Ni-G || Fe-Ni(OH) 2 /3DP Ni-G at a constant current density of 500 mA cm À2 in 1 M KOH.h) Comparison of overall water splitting performance before and after long-term stability test.
25 A for Fe-Ni(OH) 2 /3DP Ni-S produces a jagged graph accompanied by a maximum fluctuation of 179 mV, while the value for Fe-Ni(OH) 2 /3DP Ni-L is 52 mV and Fe-Ni(OH) 2 /3DP Ni-G is 91 mV.Similarly, the resistance fluctuation for Fe-Ni(OH) 2 /3DP Ni-S at a constant current of 0.25 A also shows greater fluctuation (1.08 Ω) than that of Fe-Ni(OH) 2 / 3DP Ni-L (1.03 Ω) and Fe-Ni(OH) 2 /3DP Ni-G (1.04 Ω) in Figure4e.More precisely, the increases in resistance and overpotential are caused by the bubbles covering the electrode surface, thus isolating the mass transfer between the electrode and the electrolyte.Subsequently, the resistance and overpotential decrease rapidly as large bubbles are removed from the electrode surface.It can be seen the same conditions of resistance and overpotential fluctuations in the bare 3DP Ni in FigureS21, Supporting Information.To more directly represent the effect of bubble behavior on the overpotential, the comparison of the transmission overpotential was calculated as shown in Figure4f.From the calculation results, it can be concluded that the η trans of the Fe-Ni(OH) 2 /3DP Ni-G is lower than that of Fe-Ni(OH) 2 /3DP Ni-S and Fe-Ni(OH) 2 /3DP Ni-L.A higher transmission overpotential Fe-Ni(OH) 2 /3DP Ni-S implies that the bubbles form a blockage during the process of emission and aggregate into large bubbles attached to the electrode surface, generating additional resistance and thus causing a greater increase to the overpotential.Unlike the small pore structure, Fe-Ni(OH) 2 /3DP Ni-G has a large pore structure that allows the bubbles to escape quickly without being hindered during the process of floating up.During the process of water splitting, the attached bubbles can obscure the active site, leading to the increase of the overpotential.Therefore, the transmission overpotential for Fe-Ni(OH) 2 /3DP Ni-G is lower than that of Fe-Ni(OH) 2 /3DP Ni-S due to the efficient bubbles emission which is effective in reducing bubble adhesion.A similar conclusion can be drawn in 3DP Ni and Ru-Ni(OH) 2 /3DP Ni (FigureS22-S24, Supporting Information).The previous results show that the performance of Fe-Ni(OH) 2 /3DP Ni-S is more dependent on the bubble transport behaviors due to the relatively small pore sizes and possible high bubble traffic at high current density.The bubble release efficiency gets increased when the pore size increases to a large pore.Therefore, the performance of Fe-Ni(OH) 2 /3DP Ni-L and Fe-Ni(OH) 2 /3DP Ni-G is less dependent on bubbles.To demonstrate whether the gradient structure plays a beneficial role in the bubble escape process, the bubbles were injected into both Fe-Ni(OH) 2 /3DP Ni-S, Fe-Ni(OH) 2 /3DP Ni-L, and Fe-Ni(OH) 2 /3DP Ni-G at the same rate of 20 mL h À1 to simulate the bubble emission behaviors.In Figure5a-c, the escape time of the injected bubble is only 83.3 ms and the bubble size is about 0.93 mm in Fe-Ni(OH) 2 /3DP Ni-G, as a comparison, the value for Fe-Ni(OH) 2 /3DP Ni-S is 683.3 ms and 1.71 mm (Figure S25, Supporting Information), respectively.It is worth noting that the bubble emission time of Fe-Ni(OH) 2 /3DP Ni-L is only 50.1 ms and the bubble size is 0.96 mm (Figure S26, Supporting Information), which were both similar to the value of

Figure 4 .
Figure 4. Microscope images showing the bubble evolution behaviors at the same voltage of a) Fe-Ni(OH) 2 /3DP Ni-S, b) Fe-Ni(OH) 2 /3DP Ni-L, and c) Fe-Ni(OH) 2 /3DP Ni-G.d) The overpotential fluctuation at a constant current of 0.25 A in Fe-Ni(OH) 2 /3DP Ni with different structures.e) The resistance fluctuation at a constant current of 0.25 A in Fe-Ni(OH) 2 /3DP Ni with different structures.f ) Comparison of transmission overpotential in Fe-Ni(OH) 2 /3DP Ni with different structures.

Figure 5 .
Figure 5. Camera images of bubble release from Fe-Ni(OH) 2 /3DP Ni-G.a-c) Large pore size is placed facing upward in the gradient structure, d-f ) small pore size is placed facing upward in the gradient structure.Blue and red dotted lines highlight the first bubble and second bubble, respectively.The scale bar is 2 mm.g) Distribution of pressure inside the gradient structure.h) Distribution of bubble velocity inside the gradient structure.i) The average velocity of the bubble inside 3DP Ni with different structures.