Atomically Adjustable Rhodium Catalyst Synthesis with Outstanding Mass Activity via Surface‐Limited Cation Exchange

Rh has been widely studied as a catalyst for the promising hydrazine oxidation reaction that can replace oxygen evolution reactions for boosting hydrogen production from hydrazine‐containing wastewater. Despite Rh being expensive, only a few studies have examined its electrocatalytic mass activity. Herein, surface‐limited cation exchange and electrochemical activation processes are designed to remarkably enhance the mass activity of Rh. Rh atoms were readily replaced at the Ni sites on the surface of NiOOH electrodes by cation exchange, and the resulting RhOOH compounds were activated by the electrochemical reduction process. The cation exchange‐derived Rh catalysts exhibited particle sizes not exceeding 2 nm without agglomeration, indicating a decrease in the number of inactive inner Rh atoms. Consequently, an improved mass activity of 30 A mgRh−1 was achieved at 0.4 V versus reversible hydrogen electrode. Furthermore, the two‐electrode system employing the same CE‐derived Rh electrodes achieved overall hydrazine splitting over 36 h at a stable low voltage. The proposed surface‐limited CE process is an effective method for reducing inactive atoms of expensive noble metal catalysts.


Introduction
[7] Conventional water splitting, including hydrogen evolution reaction (HER) and oxygen evolution reaction (OER), theoretically requires 1.23 V to decompose water into hydrogen and oxygen.10] This can be achieved by introducing nanostructures, [11,12] increasing the overall cell size and accelerating the OER kinetics, which is still challenging.[15][16] Among them, the hydrazine oxidation reaction (HzOR, Equation 1) has a noticeable advantage of exhibiting a remarkably lower theoretical thermodynamic potential (−0.33 V vs RHE) than that of OER (1.23 V vs RHE) and is also a carbon-free reaction. [17]ydrazine has been widely utilized in the pharmaceutical and chemical industries and has been mainly treated as a by-product of ammonia. [18]However, the wastewater including hydrazine is a highly toxic pollutant. [19]Therefore, overall hydrazine splitting (OHzS, Equation 2), replacing OER in overall water splitting (OWS, Equation 3) with HzOR, can potentially be a promising novel approach for energy-saving hydrogen production and green environment.
][22] Among them, Rh is considered to promote a 4-electrons reaction to decompose hydrazine and has the lowest overpotential. [23]In addition, owing to previous reports suggesting that Rh exhibits catalytic properties for HER, we infer that Rh-based catalysts can be considered for simultaneous reactions of both HzOR and OHzS. [22,24,25]However, as Rh is one of the most expensive noble metals (more expensive than Pt, Pd, and Ir), the catalytic properties considering the used amount (mass activity) should be preferentially compared, more stringently than in the case of other precious metals.Nevertheless, most of the previous studies focused on geometric area-based current density, whereas very few studies reported the mass activity Rh has been widely studied as a catalyst for the promising hydrazine oxidation reaction that can replace oxygen evolution reactions for boosting hydrogen production from hydrazine-containing wastewater.Despite Rh being expensive, only a few studies have examined its electrocatalytic mass activity.Herein, surface-limited cation exchange and electrochemical activation processes are designed to remarkably enhance the mass activity of Rh.Rh atoms were readily replaced at the Ni sites on the surface of NiOOH electrodes by cation exchange, and the resulting RhOOH compounds were activated by the electrochemical reduction process.The cation exchangederived Rh catalysts exhibited particle sizes not exceeding 2 nm without agglomeration, indicating a decrease in the number of inactive inner Rh atoms.Consequently, an improved mass activity of 30 A mg Rh −1 was achieved at 0.4 V versus reversible hydrogen electrode.Furthermore, the two-electrode system employing the same CE-derived Rh electrodes achieved overall hydrazine splitting over 36 h at a stable low voltage.The proposed surface-limited CE process is an effective method for reducing inactive atoms of expensive noble metal catalysts.
characteristics of Rh. [16,26] Furthermore, Rh-based catalysts have been produced either by hydrothermal synthesis or with particle shapes through heat treatment-based methods. [22,24,25,27]Because Rh particles exist in the form of differently sized clusters, the inner Rh atoms cannot directly establish an electrochemical connection with the electrolyte.Therefore, a low mass activity is expected owing to the considerable agglomeration of the inactive Rh atoms during thermal annealing (Scheme 1a).Moreover, because catalytic particles are usually loaded onto the conductive substrates through the Nafion binder, a fraction of the active sites decreases owing to Nafion coverage, and stability problems from the detachment with the substrate may occur during long-term operation. [28]Therefore, a novel synthesis strategy that aims for a superior increase in mass activity and self-supported electrocatalysts is necessary to reduce Rh consumption.Cation exchange (CE), which has been recently developed as an alternative method to the conventional colloidal synthesis, is a synthesis method that converts the parent element into another by replacing the parent cation in the lattice. [29]The CE process is initiated by immersing template compounds in electrolyte solutions including other cations, and this process can overcome the limitations of traditional methods to manipulate structures and compositions.Especially, the CE synthesis by the rapid solid-state diffusion is beneficial for non-equilibrium nanoparticle synthesis.Thus, new compounds such as chalcogenide-based nanocrystals and heterostructures with complex structures and metastable phases, which are conventionally not accessible, have been preferentially considered in the CE process. [30]During the CE process, the anion basis with a relatively large ionic radius tends to stably maintain the lattice structure, morphology, and dimension of the parent material. [31]Notably, this process requires extremely low input energy and is spontaneous even at room temperature.In typical nanomaterial synthesis using CE, excessive cations can be substitutionally introduced via cation diffusion up to the deep bulk of the parent material.However, when only a small amount of cation is substituted, the surface of the parent material becomes the dominant reactive site owing to diffusion limited reaction. [32]Consequently, if the cation amount substituted for CE can be artificially controlled, the location of the substituted cations can be limited to only a few layers close to the surface.Especially, oxide-or hydroxide-based materials with a smaller ionic radius than sulfur or selenium are advantageous to obtain a shallow CE region.Ultimately, the formation of a shallow CE region is expected to result in excellent mass activity because most of the injected cations can be used as active catalytic sites after proper activation.
In this study, we propose two novel ideas for attaining the atomicscale Rh catalysts with outstanding mass activity and long-term stability: i) the surface-limited CE process for intermediate hydroxide compounds capable of atomically injecting and distributing Rh and ii) the electrochemical surface reduction process for the vigorous activation of Rh (Scheme 1b).The NiOOH compounds were electrochemically prepared on carbon cloth substrates as a template for Rh CE.Here, the Ni 3+ ion in NiOOH has an ionic radius of 60 pm and exhibits a 10% difference against the Rh 3+ ionic radius (66.5 pm), indicating the feasibility to supply facile CE sites with Rh in the NiOOH compound. [33,34]hus, the prepared NiOOH electrodes were immersed in Rh 3+ aqueous solution to induce cation exchange between Ni 3+ and Rh 3+ , and the preparation of the designed RhOOH compounds was successfully achieved.Next, for catalytic activation of the inactive RhOOH, phase conversion into metallic Rh was atomically achieved with excellent catalytic properties through the electrochemical reduction process that selectively reduces the Rh atoms existing on the surface.Compared with the inactive cation-exchanged Rh (ICE-Rh) sample, the activated cation-exchanged Rh (ACE-Rh) electrode exhibited a 100-fold increase in the HzOR catalytic properties.Therefore, ACE-Rh demonstrated an excellent mass activity of 30 A mg Rh −1 for HzOR in 1.0 M KOH + 0.5 M N 2 H 4 solution under 0.4 V versus RHE.In addition, a robust stability was maintained for 36 h under the condition of 1 A mg Rh −1 for the OHzS test, where both the anode and the cathode were composed of the same ACE-Rh.

Characterization of Cation Exchanged RhOOH
The morphology and structure of the electrodeposited NiOOH-based samples before and after the CE process were compared using scanning electron microscopy (SEM) (Figure 1a,b and Figure S1a-c).The CE process of NiOOH was conducted in aqueous solutions with various Rh 3+ concentrations from 0.1 to 1.0 mM.The NiOOH on carbon cloth is in the form of a nanosheet and retains its shape even after the CE process regardless of the Rh 3+ concentration.It was confirmed that the CE of Ni 3+ and Rh 3+ ions was a stable reaction that did not disrupt the structure of the parent material.In contrast, when Rh metal was directly electrodeposited on the conductive substrates, it was observed in the form of spherical clusters inducing limited active sites (Figure S1d).For the electrodeposition of noble metals, the electric field was selectively concentrated on the initially formed metallic seeds, so that the Rh catalysts preferentially grow around the cluster, which significantly decreases the ratio of the active catalytic sites to the amount of deposited Rh metals. [35]The crystallographic structure of the NiOOH electrodes before and after cation exchange in 0.6 mM Rh 3+ solution was confirmed by high-resolution transmission electron microscopy (HRTEM).
The HRTEM image and fast Fourier transform (FFT) diffraction pattern for the electrochemically synthesized NiOOH show a circle of 0.208 nm corresponding to the (210) plane of α-NiOOH (JCPDS No. 27-956) (Figure 1c). [36,37]For the CE-processed NiOOH, circles with distances of 0.208 and 0.223 nm were involved with the NiOOH (210) plane and the RhOOH (111) plane (JCPDS No. 27-1383), respectively (Figure 1d). [38,39]The NiOOH and RhOOH circles are distinguished, and weak hazy circles appear owing to the gradual increase of interplanar distance caused by Rh substitutions in the NiOOH lattice.Furthermore, atomic structures with interplanar distances of 0.223 and 0.208 nm are observed in the HRTEM image (Figure 1e).Therefore, we infer that the orthorhombic RhOOH crystal was synthesized through CE between Rh and Ni from the α-NiOOH with the same orthorhombic structure.In addition, the elemental mapping of high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) confirms that Ni and Rh are homogeneously distributed in the compounds (Figure 1f-h).Owing to homogeneous Rh exchange with Ni, agglomerated particles with single Ni or Rh elements are absent.) were vigorously activated in the metal phase.Although the CEprepared RhOOH was in direct contact with the electrolyte, it developed catalytic properties with difficulty owing to its -OOH bonding.Therefore, to activate the catalytic properties of ICE-Rh, it is necessary to partially metalize the RhOOH on the surface that is in direct contact with the electrolyte.The method for selectively metalizing RhOOH without changing the bulk region is distinctly involved with diffusion reaction and the electrochemical design of the experimental condition is theoretically inspired by the Pourbaix diagram (Figure S2). [40]ccording to the Pourbaix diagram regarding Rh and RhOOH, Rh 3+ ions can be electrochemically reduced to the Rh 0 metal phase, and the reaction can be described through the following equation:

Electrochemical Activation of RhOOH
Thus, to find the optimal electrochemical condition for reducing RhOOH into metallic Rh, the electrochemical properties of NiOOH and ICE-Rh electrodes were characterized using cyclic voltammetry (CV) in a 1.0 M KOH solution (Figure 2b and Figure S3).The pristine NiOOH without Rh (black line) shows no reduction properties with almost zero HER current in the range of −0.3 to 0.9 V (vs RHE).In contrast, for all ICE-Rh electrodes, HER current and H desorption peaks appear in the first cycle, and slightly increase with continuing cycle numbers.The H desorption peak near 0.2 V is attributed to the current generation owing to the detachment of H + ions adsorbed to the noble metal during the HER process.Thus, it is observed that the RhOOH phase is reduced to metallic Rh in the CV process through the increase of the H desorption peak.In addition, the increase in the HER current is closely associated with the formation of the metallic Rh phase instead of the inactive RhOOH, thereby contributing to catalytic behavior. [38]Therefore, it is expected that the RhOOH on the electrode's entire surface can be uniformly reduced to metallic Rh using the electrochemical galvanostatic process.The Rh exchanged with NiOOH and Rh distribution require adjustments to attain the maximum mass activity.Thus, various ICE-Rh electrodes prepared at different Rh 3+ concentrations were electrochemically reduced in 1.0 M KOH solution (Figure 2c).The current density for the reduction was set to −50 mA cm −2 , corresponding to the maximum value in the CV measurement.The applied bias at the initial stage of the reduction reaction exhibited large negative values which gradually reduced in magnitude because the electrocatalytic HER simultaneously improved with the galvanostatic reduction of RhOOH into metallic Rh, and accordingly, the HER overpotential decreased.Even after the applied bias was stabilized at approximately 200 s, the reaction persisted longer for a sufficient reduction of RhOOH.
To verify the phase change of the surface Rh metallization for electrocatalytic activation of ICE-Rh, X-ray photoelectron spectroscopy (XPS) and TEM analyses were conducted with pristine NiOOH, ICE-Rh, ACE-Rh, and ED-Rh electrodes, as shown in Figure 3 and Figure S4.Typical Ni2p 3/2 and Ni2p 1/2 peaks appeared in all samples owing to the existence of NiOOH bonding near the surface (Figure 3a). [41]After CE, the intensity of the Ni2p 3/2 peak decreased owing to the replacement of Ni atoms with Rh 3+ ions on the surface, and the Ni 0 peak was observed at 852.8 eV in only the ED-Rh sample owing to the reduction of NiOOH during the ED process of Rh. [42] For the Rh3d spectra, unlike NiOOH in the absence of a Rh3d peak, it is confirmed that the ICE-Rh electrode revealed an Rh3d 5/2 peak at 309.9 eV, indicating the generation of RhOOH bonding (Figure 3b).After the electrochemical Rh metallization process, it was observed that the RhOOH peak intensity decreased and additional peaks at 308.7 and 307.1 eV corresponding to Rh 2 O 3 and Rh metal, respectively, were produced. [43]According to the XPS spectra of Rh3d, it was confirmed that the phase change from RhOOH to metallic Rh was successfully achieved at the surface region using the electrochemical reduction process.Moreover, as shown in the inset of Figure 3b, the intensity of the metallic Rh peak (13.7%) is significantly lower than that of Rh 2 O 3 (37.2%).Here, the Rh 2 O 3 peak originates from native oxidation on the surface of the Rh atoms, which is a common result for atomically dispersed catalysts. [44,45]This result indicates that the Rh particles produced from RhOOH are very thin and small, which is consistent with our expectation regarding surface-limited CE and electrochemical activation reaction.Conversely, the ED-Rh electrode demonstrated the presence of a very strong metallic Rh peak and a weak Rh 2 O 3 peak because excessive Rh was heavily loaded in the form of agglomerated particles on the NiOOH surface, as previously confirmed by SEM.For the O1s spectra, the peak at 531 eV regarding the M-OH bond was positively shifted by 0.2 eV after the CE process of the NiOOH electrode and finally returned to the original peak position after the Rh metallization process (Figure 3c).This shift occurred because when the surface NiOOH transformed to RhOOH, the electron-binding energy of O1s increased owing to the high electronegativity of Rh ion, which was stronger than that of Ni ion. [46]In the XPS spectra of the ED-Rh electrode, a weak M-O bonding peak appeared owing to the formation of a natural native oxide layer on the Rh surface.
Subsequently, HRTEM analysis was performed for further analysis on the existence of metallic Rh in ACE-Rh (Figure S4a-d).Similar to the previous HRTEM result of cation-exchanged Rh, a band-shaped circle corresponding to NiOOH and RhOOH was observed in the FFT diffraction pattern of the HRTEM image of ACE-Rh sample.In addition, the (111) lattice lines of the metallic Rh phase were observed with nanoparticle shapes in the HRTEM image, [38] and the Rh particle sizes did not exceed 2 nm without significant agglomeration.Owing to the suppressed aggregation of Rh, no obvious peaks were observed in the X-ray diffraction measurements (Figure S4e), and no clusters were observed in the SEM image (Figure S4f).It was found that the reduced Rh from RhOOH was well distributed with ultrafine nanoparticles of a few-nanometer atomic size on the surface, instead of the typical cluster form.Furthermore, the electrochemical reduction process of the RhOOH maintained the nanosheet structure of the pristine NiOOH without significant structural collapse.

Electrocatalytic Mass Activity for HzOR and HER
The electrocatalytic HzOR properties of the ACE-Rh electrodes prepared via galvanostatic activation of ICE-Rh were investigated using LSV in a 1.0 M KOH + 0.5 M N 2 H 4 electrolyte (Figure 4a).Therefore, the electrochemical current density increased by more than 100 times than that of ICE-Rh, and the excellent HzOR catalytic properties can be ascribed to the successful formation of metallic Rh catalysts.In addition, the higher the Rh 3+ concentration in the CE solution, the greater the amount of Rh in ACE-Rh during the activated reduction process, and the better the electrocatalytic HzOR current density.The corresponding Tafel slopes of ACE-Rh were calculated as 56.0, 46.6, 45.7, and 45.2 mV dec −1 for ACE-Rh0.1,ACE-Rh0.3,ACE-Rh0.6, and ACE-Rh1.0,respectively (Figure 4b).Meanwhile, the HzOR LSV tests of ACE-Rh electrodes indicate that the current density of ACE-Rh linearly increased in proportion to the concentration of CE Rh solution at 0.2 V versus RHE, but the ACE-Rh1.0 sample exhibited a deviation from the linear increment despite a slight increase in current density (Figure 4c).This implies that the amount of exchanged Rh increased with the concentration of the CE Rh solution, and a considerable amount of the loaded Rh at 1.0 mM condition did not contribute to the catalytic reaction even when cation exchange occurred continuously.To quantitatively confirm this, the electrochemical surface area (ECSA) of ACE-Rh electrodes was characterized through double-layer capacitance (C dl ) measurements (Figure 4d).As ECSA is directly proportional to C dl , electrochemically active sites can be compared by calculating C dl . [47]The C dl values of ACE-Rh were obtained by measuring CV in a 1.0 M KOH solution from 0.75 V to 0.85 V while increasing the scan rate from 20 to 100 mV s −1 (Figure S5).The ACE-Rh0.1, ACE-Rh0.3,ACE-Rh0.6, and ACE-Rh1.0 samples exhibited C dl values of 2.1, 2.5, 3.0, and 3.1 mF cm −2 , respectively.It was confirmed that the ECSA of ACE-Rh increased with increasing concentration of CE Rh solution but nearly saturated when the concentration exceeded 0.6 mM.Likewise, in the Tafel plots of the HzOR current density of the ACE-Rh electrodes, the Tafel slopes significantly increased from ACE-Rh0.1 to ACE-Rh0.6 but remained nearly constant for ACE-Rh1.0.This indicates that the ACE-Rh1.0 contains over-saturated Rh and the inactive Rh atoms may cause a decrease the mass activity of Rh.
To calculate the mass activity of Rh, the real mass of Rh contained in the ACE-Rh electrodes was measured using inductively coupled plasma-optical emission spectroscopy (ICP-OES) (Figure 5a).As a result, the loaded masses of Rh in the ACE-Rh electrodes were calibrated to 5.2, 16.6, 26.3, and 47.3 μg cm −2 for ACE-Rh0.1,ACE-Rh0.3,ACE-Rh0.6, and ACE-Rh1.0,respectively.These masses were converted according to the relative atomic ratio of Ni and Rh, resulting in Rh atomic fractions of 7.3%, 18.4%, 31.9%, and 41.0%, respectively.Considering the measured mass of Rh, the HzOR current densities in the LSV curves of ACE-Rh and ED-Rh electrodes can be represented as mass activity instead of the typical current density without considering the amount of loaded Rh atoms (Figure 5b).Consequently, the ACE-Rh0.1,ACE-Rh0.3, and ACE-Rh0.6 electrodes demonstrated a similar mass activity, but the ACE-Rh1.0 and ED-Rh electrodes exhibited relatively low mass activity by individually normalizing the consumed Rh amounts.This indicates that all exchanged Rh atoms did not act as active sites in the ACE-Rh1.0 sample, and similarly, the mass activity of the ED-Rh electrode was considerably low owing to the contribution of inactive Rh atoms inside the agglomerated particles with a large dimension.Specifically, the ACE-Rh0.6 electrode displayed an outstanding mass activity of over 30 A mg Rh −1 at 0.4 V versus RHE, which is the best HzOR mass activity of Rh among the reported results (Table S1).The potentials required for mass activities of 5 and 10 A mg Rh −1 in all ACE-Rh electrodes were superior to 0.174 and 0.283 V of the ED-Rh electrode, and the ACE-Rh0.6 electrode exhibited the lowest potentials of 0.141 and 0.201 V (Figure 5c).This is because all exchanged Rh atoms were dominantly distributed in the surface region and acted as catalytic active sites when the concentration of CE Rh solution dropped below 0.6 mM and the electrodes became electrochemically activated.However, if the concentration increased beyond 0.6 mM, the presence of inactive Rh atoms decreased the mass activity, as already verified by ECSA.Therefore, it can be concluded that the optimal Rh concentration of CE solution is 0.6 mM for the effectiveness of mass activity at the appropriate Rh amount.A galvanostatic HzOR test was performed under the condition of 10 mA cm −2 for 12 ks to evaluate the stability of the ACE-Rh electrodes (Figure 5d).The ACE-Rh0.6 and ACE-Rh1.0 electrodes displayed a stable operation, while the ACE-Rh0.1 and ACE-Rh0.3 showed a marked increase in the applied voltage required for maintaining the same current density.This is because the Ni sites existing on the ACE-Rh0.1 and ACE-Rh0.3 surfaces, not saturated with Rh ions, were readily oxidized under the HzOR condition so that the active catalytic sites were significantly reduced (Figure S6).Therefore, considering the mass activity and stability performances, the Rh-saturated ACE-Rh0.6 electrode is the best HzOR electrocatalyst among ACE-Rh electrodes, which effectively performs the cation exchange and electrochemical activation processes.
As Rh is commonly accepted as an excellent electrocatalyst for HER, the HER current density was measured by LSV curves in KOH 1.0 M solution to confirm the HER properties in view of the ACE-Rh0.6 mass activity (Figure 6a).The pristine NiOOH electrode with low HzOR performance was considerably reactive for HER below −0.3 V versus RHE.Moreover, the ACE-Rh0.6 electrode demonstrated significantly increased HER current density owing to the intrinsic HER properties of Rh, whereas the ED-Rh exhibits a higher current density.The NiOOH, ACE-Rh0.6, and ED-Rh electrodes exhibited Tafel slopes of 225.1, 106.3, and 91.2 mV dec −1 , respectively, implying gradual improvement of HER via Rh addition (Figure 6b).Specifically, ACE-Rh0.6 showed better properties than ED-Rh when represented with mass activity (Figure 6c).To achieve mass activities of 1 and 5 A mg Rh −1 , ACE-Rh0.6 required overpotentials of 152 and 289 mV, whereas the ED-Rh required overpotentials of 191 and 377 mV, respectively.This result demonstrated the same trend as that of HzOR, indicating that Rh in the form of agglomerated particles is unfavorable for high mass activity.Galvanostatic HER tests were performed at −10 mA cm −2 for 30 ks to confirm the HER stability of ACE-Rh0.6.As shown in Fig- ure 6d, the electrocatalytic HER operation reveals the stable maintenance of the applied bias.Further, to check whether the addition of hydrazine in the electrolyte affects HER properties, the HER current density was characterized in both 1.0 M KOH solution and 1.0 M KOH + 0.5 M N 2 H 4 solution (Figure S7).Interestingly, no significant difference was observed in HER performances regarding hydrazine addition by demonstrating almost identical electrochemical curves.

OHzS Tests
The ACE-Rh0.6 electrode prepared through the CE and the electrochemical reduction processes was distinctly confirmed to possess excellent electrocatalytic mass activity for both HzOR and HER.Therefore, overall hydrazine splitting tests were performed in a twoelectrode system wherein both the anode and cathode consisted of the same ACE-Rh0.6electrodes.First, the OHzS current density and mass activity were characterized through LSV curves in 1.0 M KOH + 0.5 M N 2 H 4 solution (Figure 7a).Potentials of 0.266 and 0.471 V were required to achieve mass activities of 1 and 5 A mg Rh −1 , respectively, and a high mass activity of 9.9 A mg Rh −1 could be achieved even at a voltage as low as 0.6 V.When comparing the current densities of OWS and OHzS, the difference between their required voltages was more pronounced (Figure 7b).The OHzS operation could remarkably reduce the driving voltage of 1.44 V compared to that of OWS at 10 mA cm −2 , which is highly advantageous for energy-saving hydrogen production.To check the electrochemical stability of entire cells, the chronoamperometry of individual OWS and OHzS electrodes was characterized for long-term stability under the same condition of 1 A mg Rh −1 (Figure 7c).
The cell operation in the OHzS exhibited superior stability for 36 h owing to the excellent adhesion and chemical stability of cation exchanged Rh, similar to the stability test results of HzOR and HER.In addition, the applied voltage reduction of more than 1.4 V was achievable even during long-time operations, compared to the result of traditional OWS operation.For further analysis, the XPS analysis of the ACE-Rh0.6 was performed after the long-term stability test (Figure S8).Consequently, no significant phase change was observed after the OHzS reaction for both the ACE-Rh electrodes, which were used as anode and cathode.The amounts of hydrogen and nitrogen generated through the OHzS were measured using gas chromatography (GC) to confirm the complete decomposition of hydrazine by the ACE-Rh0.6 electrode (Figure 7d).The Faradaic efficiency of hydrogen and nitrogen evolution was nearly 100%, approaching the theoretical amount of hydrogen and nitrogen calculated from the electric charge and the detected gas amount.This Faradaic efficiency confirmed that the complete decomposition of hydrazine occurred with ultimate catalytic activity in the ACE-Rh. [23]Because the two-electrode OHzS system composed of ACE-Rh0.6 can sufficiently operate even at voltages less than 0.6 V, a solar cell-driven OHzS test was conducted using a copper-indium-gallium-selenide (CIGS) solar cell (Figure 7e).The CIGS solar cell had an open-circuit voltage of 584 mV and a short circuit current of 19 mA cm −2 under the 1-sun illumination condition (AM 1.5 G, 100 mW cm −2 ) (Figure S9).It is observed that active hydrogen and nitrogen production occurred only through a single solar cell (Video S1).Therefore, unlike OWS, which requires a large voltage by connecting several solar cells in series for the mass production of hydrogen, the OHzS operation that is possible by CE and electrochemical activation processes of Rh catalysts is very advantageous for energy-saving hydrogen production using solar energy without carbon emission.

Conclusion
In this study, a surface-limited CE method was used to remarkably increase the electrocatalytic mass activity of Rh in the HzOR and OHzS reactions.The NiOOH template was electrochemically deposited on a carbon cloth substrate, and substitutional CE between Ni 3+ and Rh 3+ was induced in an aqueous Rh 3+ solution to synthesize RhOOH.The same orthorhombic structure of NiOOH and RhOOH was confirmed through HRTEM and diffraction pattern, suggesting the accomplishment of the designed RhOOH compounds using CE.However, the CE-prepared RhOOH was inactive for hydrazine oxidation.Thus, an electrochemical reduction process was suggested for the galvanostatic activation of ICE-Rh, and the reduction of RhOOH to thin Rh nanoparticles was confirmed by analyzing the chemical phase of the surface using XPS.These Rh atomic particles exhibited excellent electrocatalytic properties for HzOR and HER.Especially, when Rh was saturated on the NiOOH surface, a maximum mass activity of 30 A mg Rh −1 for HzOR was obtained with the ACE-Rh0.6 electrode at 0.4 V vs RHE, which was double compared to the Rh catalysts prepared by the general  , respectively.Additionally, the two-electrode system using the same CEderived Rh anode and cathode exhibited an excellent mass activity of 9.9 A mg Rh −1 at 0.6 V for the OHzS reaction.The OHzS system induced the reduction of the operating voltage by more than 1.4 V compared to that of the general water splitting.

Experimental Section
Materials: Ni(NO Synthesis of electrodes: NiOOH templates for Rh cation exchange were electrodeposited on a carbon cloth substrate.The potentiostatic electrodeposition was conducted at −1.0 V versus Ag/AgCl (saturated KCl) for 200 s in a 0.04 M Ni (NO 3 ) 2 aqueous solution.Then, cyclic voltammetry was performed for voltages from 0.2 to 0.6 V versus Hg/HgO (1.0 M NaOH) for 50 cycles in a 1.0 M KOH solution.The NiOOH electrode was washed with water and ethanol, then dried in air.
Cation exchange between Ni and Rh was conducted in RhCl 3 aqueous solutions of various Rh 3+ concentrations.The prepared 1 cm 2 NiOOH electrode was soaked in a 5 ml RhCl 3 solution for 12 h at 25°C to achieve enough cation exchange.Subsequently, the ICE-Rh electrodes were washed with water and ethanol, then dried in air.
The electrochemical metallization process of cation exchanged RhOOH was conducted using the galvanostatic method (−50 mA cm −2 ) for 400 s in 1.0 M KOH to reduce RhOOH into metallic Rh.The ACE-Rh electrodes were subsequently washed with water and used for electrochemical measurement immediately.
The ED-Rh electrode was prepared by the electrodeposition of Rh on the prepared NiOOH substrate.The potentiostatic electrodeposition was conducted at −1.0 V versus Ag/AgCl (saturated KCl) for 3 h in a 1.0 mM RhCl 3 aqueous solution.The ED-Rh electrode was washed with water and used for electrochemical measurements immediately.
Material characterization: Morphological information of the prepared samples was obtained using scanning electron microscopy (JSM-7600F, JEOL, Japan).The crystallographic information and elemental mapping were acquired using a high-resolution transmission electron microscope (Titan G2 ChemiSTEM Cs Probe, ThermoFisher Scientific, USA) with an accelerating voltage of 200 kV.The elemental and chemical states were analyzed using X-ray photoelectron spectroscopy (NEXSA, ThermoFisher Scientific, USA).The mass of Rh was determined by inductively coupled plasma-optical emission spectroscopy (OPTIMA 8300, Perkin-Elmer, USA) for calculating mass activity.
Electrochemical methods: All electrochemical methods and measurements were performed using a 3-electrode system potentiostat (VersaSTAT4, Princeton Applied Research, USA) in Ar-purged electrolytes.The prepared NiOOH or RhOOH electrodes were used as working electrodes, a platinum mesh was used as a counter electrode, and Hg/HgO (1.0 M NaOH) served as a reference electrode.All measurements involving hydrazine oxidation reaction were conducted in a 1.0 M KOH + 0.5 M N 2 H 4 solution, while the others were conducted in 1.0 M KOH.All potentials mentioned in the manuscript were converted to reversible hydrogen electrode, followed by the Nernst equation (E RHE = E Hg/ HgO + 0.059 × pH + E°H g/HgO ).All linear sweep voltammetry curves were corrected by 95% iR-compensation.The gas chromatograph (7890B GC, Agilent, USA) was used to determine the evolution rate of hydrogen and nitrogen.
Solar cell-driven overall hydrazine splitting: A CIGS solar cell (SF-CIGS-10G3, SolarFLEX, Republic of Korea) was used for the solar cell-driven overall hydrazine splitting.The solar cell and the ACE-Rh0.6 electrodes were connected by copper wire, and the solar cell was subjected to a 1-sun illumination (AM 1.5 G, 100 mW cm −2 ) using a solar simulator (HAL-320, ASAHI SPECTRA, USA).The light-exposed area of the solar cell was 25 cm 2 (5 cm × 5 cm).

Figure 1 .
Figure 1.Top-view SEM images of a) pristine NiOOH and b) NiOOH after CE process in 0.6 mM Rh 3+ solution.HRTEM images and FFT patterns of c) NiOOH and d) cation exchanged NiOOH.e) Enlarged HRTEM image of the selected area in d).f-h) HAADF-STEM images of cation exchanged NiOOH and corresponding elemental mapping for Ni and Rh, respectively.
HzOR characteristics were evaluated using linear sweep voltammetry (LSV) in a 1.0 M KOH + 0.5 M N 2 H 4 solution.Pristine NiOOH, four types of ICE-Rh, and electrodeposited Rh (ED-Rh) electrodes were used for the comparative tests where all substrates were same NiOOH.As shown in the enlarged graph of Figure 2a, pristine NiOOH (black line) did not exhibit any meaningful current for HzOR, and the reactivity of each ICE-Rh slightly increased compared to that of pristine NiOOH.ICE-Rh0.1,ICE-Rh0.3,ICE-Rh0.6, and ICE-Rh1.0 prepared in CE solutions with Rh concentrations of 0.1, 0.3, 0.6, and 1.0 mM, respectively, exhibited current densities of 4.2, 6.1, 7.4, and 8.4 mA cm −2 at 0.4 V versus RHE, respectively.Nevertheless, all ICE-Rh electrodes still demonstrated considerably lower HzOR current than the performance of ED-Rh (blue line) coated directly with metallic Rh, implying insufficient catalytic behaviors of ICE-Rh samples.In general, the electrochemical catalytic properties of various metals (Pt, Rh, Pd, etc.

Figure 4 .
Figure 4. a) LSV curves for electrocatalytic HzOR and b) corresponding Tafel plots of ACE-Rh electrodes.c) HzOR current densities of ACE-Rh electrodes at 0.2 V versus RHE.d) C dl values of ACE-Rh obtained from capacitive Δj as a function of scan rates in CV.

Figure 5 .
Figure 5. a) Measured mass and atomic ratio of the Rh loaded on ACE-Rh by ICP-OES.b) Mass activities and c) applied potentials maintaining 5 and 10 A mg Rh −1 of ACE-Rh.d) The long-term stability test of ACE-Rh for the HzOR at 10 mA cm −2 in 1.0 M KOH + 0.5 M N 2 H 4 .

Figure 6 .
Figure 6.Electrocatalytic HER characterizations.a) LSV curves for HER in 1.0 M KOH at scan rate of 10 mV s −1 , b) corresponding Tafel plots, and c) mass activity.d) The long-term stability test for HER at −10 mA cm −2 in 1.0 M KOH.

Figure 7 .
Figure 7. Two-electrode OHzS tests using the same ACE-Rh0.6 as both anode and cathode.a) LSV curve and corresponding mass activity for OHzS.b) Comparison of LSV curves for OHzS and OWS.c) Comparison of the long-term stabilities and operating potentials at 1 A mg Rh −1 for OHzS and OWS.d) The amounts of hydrogen and nitrogen gas obtained from theoretical calculation and experimental measurements under galvanostatic operation of 1 A mg Rh −1 .e) Photograph of solar cell-driven OHzS test using CIGS solar cell under 1 sun illumination.