Ruthenium Oxide Clusters Immobilized in Cationic Vacancies of 2D Titanium Oxide for Chlorine Evolution Reaction

The development of dimensionally stable anodes (DSAs) has made the chlorine evolution reaction (CER) the most important industrial anode reaction since the 1960s. However, the preparation of DSA depends on the extensive use of precious metals, Ru and Ir, which are expensive and scarce. Herein, a cationic defect adsorption–oxidation anchoring strategy to immobilize oxidized sub‐nano ruthenium clusters on 2D low‐crystallinity titanium oxide (2D TiOx) substrate is reported. Through the metal oxide−support interaction, the 2D TiOx alters the electronic structure of ruthenium oxide (RuOx), improving its activity, selectivity, and stability for CER. Specifically, the mass activity of the RuOx/2D TiOx electrode is 26.5 and 143.5 times higher than that of the state‐of‐the‐art commercial RuO2 and DSA, respectively, at an overpotential of 100 mV. Moreover, the selectivity of the RuOx/2D TiOx electrode to CER is approximately 96.5%, and it exhibits remarkable durability lasting for over 210 h. Therefore, the 2D TiOx substrate holds significant potential for improving the dispersion, active site density, and atomic utilization of oxidized sub‐nano noble metal clusters.


Introduction
[3] These processes include the manufacturing of plastics and pharmaceuticals, paper production, and water treatment. [4,5][8][9] Initially, low-cost graphite rods were used as anodes for CERs.However, their low catalytic selectivity and durability led to their abandonment in the 1970s. [10]Subsequently, researchers proposed dimensionally stable anodes (DSAs), which are Ti metal electrodes coated with RuO 2 , IrO 2 , and TiO 2 mixed oxides. [11,12]DSA has been utilized in commercial production due to its nearly 100% selectivity for CER and exceptional durability under chlorine-saturated acidic conditions. [13]However, commercial DSA requires approximately 30 at% RuO 2 to exhibit good conductivity, and exhibits low CER selectivity at low Cl À concentration. [6,11,14,15]These issues not only hinder the sustainable development of the chloralkali industry in the future but also impede its expansion.
In the last decade, significant efforts have been made to reduce the amount of Ru and eliminate the dependence on Ir in CER electrocatalysts.][17][18][19][20][21][22][23] Despite these efforts, achieving high selectivity in electrolytes with low Cl À concentration while further reducing the amount of noble metal Ru remains challenging.Notably, theoretical study has demonstrated that The development of dimensionally stable anodes (DSAs) has made the chlorine evolution reaction (CER) the most important industrial anode reaction since the 1960s.However, the preparation of DSA depends on the extensive use of precious metals, Ru and Ir, which are expensive and scarce.Herein, a cationic defect adsorption-oxidation anchoring strategy to immobilize oxidized sub-nano ruthenium clusters on 2D low-crystallinity titanium oxide (2D TiO x ) substrate is reported.Through the metal oxideÀsupport interaction, the 2D TiO x alters the electronic structure of ruthenium oxide (RuO x ), improving its activity, selectivity, and stability for CER.Specifically, the mass activity of the RuO x /2D TiO x electrode is 26.5 and 143.5 times higher than that of the state-of-the-art commercial RuO 2 and DSA, respectively, at an overpotential of 100 mV.Moreover, the selectivity of the RuO x /2D TiO x electrode to CER is approximately 96.5%, and it exhibits remarkable durability lasting for over 210 h.Therefore, the 2D TiO x substrate holds significant potential for improving the dispersion, active site density, and atomic utilization of oxidized sub-nano noble metal clusters.
the thin layer of TiO 2 can adjust the surface electron density of the catalyst. [24]In this case, the catalytic activity of oxygen evolution reaction (OER), a side reaction of CER, is suppressed, while the high activity of CER remains almost unchanged, leading to a significant improvement in the CER selectivity.Furthermore, it has been reported that oxidized sub-nano cluster catalysts provide proper platforms for CER. [25]Therefore, the use of ultrathin 2D titanium oxide to immobilize small-sized, highly dispersed sub-nano RuO x clusters as an anode catalyst may be a feasible approach. [26]ere, we report the immobilization of sub-nanometer ruthenium oxide clusters on a 2D titanium oxide (2D TiO x ) substrate with low crystallinity, using a cationic defect adsorptionoxidation anchoring strategy.High-resolution electron microscopy and synchrotron-based X-ray absorption spectroscopy demonstrate that ruthenium oxide is highly dispersed in the form of low nuclear clusters with low oxidation state of ruthenium.Further density functional theory (DFT) calculations elucidate that the 2D TiO x alters the electronic structure of ruthenium oxide (RuO x ) through modulating the metal oxideÀsupport interaction.As a result, the prepared RuO x /2D TiO x electrode exhibits a promising CER performance, with a mass activity that is 26.5 and 143.5 times higher than that of the current state-of-the-art commercial RuO 2 and DSA electrodes, respectively, at an overpotential of 100 mV.Additionally, this electrode shows a high selectivity of approximately 96.5% toward CER, and it has a remarkable durability, exceeding 210 h.Thus, this work provides new insights into the rational design of ruthenium-based CER catalyst with high mass activity and high stability.The sub-nano RuO x clusters immobilized on 2D TiO x nanosheets were synthetized by a cationic defect adsorption-oxidation anchoring method (Figure 1a).First, the cation-deficient 2D TiO x nanosheets were synthesized by the ion intercalation and exchange method of titanate crystals (Figure S1 and S2, Supporting Information, see Experimental Section for details). [27]he resulting TiO x nanosheets have an overall negative charge, and a clear Tyndall effect is observed in a 2D TiO x aqueous suspension (Figure S3a, Supporting Information), which formed a porous low-density structure (Figure S3b, Supporting Information) after lyophilization.Interestingly, during the lyophilization process, the TiO x nanosheets self-assembled and stacked to form a layered titanium oxide, which showed (020) and (040) X-ray diffraction (XRD) diffraction peaks (Figure 1b). [28]ransmission electron microscopy (TEM) images (Figure 1c and S4, Supporting Information) exhibited almost transparent sheet-like morphology, confirming the ultrathin thickness of the obtained 2D TiO x nanosheets.High-resolution transmission electron microscopy (HRTEM) and selective area electron diffraction (SAED) revealed a low crystallinity with the visible lattice spacing of 0.35 nm in the TiO x (101) planes (Figure 1d).A representative atomic force microscopy (AFM) image further demonstrates a thickness of approximately 1.0 nm, indicating the formation of few-layered nanosheets (Figure S5, Supporting Information).

Results and Discussion
Subsequently, the Ru precursor solution (RuCl 3 ) was injected into the TiO x aqueous suspension, and after thorough stirring and adsorption, the lyophilized sample was exposed to air and heated at low temperature for 3 h.The collected product (denoted as RuO x /2D TiO x ) was used as the final catalyst.XRD and HRTEM patterns revealed that the layered order of TiO x nanoflakes was disrupted after ion adsorption and heat treatment and replaced by anatase (101) and (200) crystal faces (Figure 1b,e).No diffraction peaks related to Ru or RuO 2 were observed, ruling out the possibility of corresponding large nanoparticles. [29]RuO x /2D TiO x maintains a 2D morphology, but the heat treatment process increases the layer thickness to 4.14 nm (Figure 1f ).Numerous sub-nanometer RuO x clusters are observed in the high-angle annular dark field-scanning transmission electron microscope (HAADF-STEM) (yellow dotted circles, Figure 1g).The energy dispersive X-ray spectrum (EDS) mappings (Figure 1h) suggest the uniform distributions of Ru, Ti, and O over the entire RuO x / 2D TiO x sample.

Electronic Structure Analysis of RuO x /2D TiO x
The electronic state and surface composition of RuO x /2D TiO x were probed using various spectroscopic techniques.The X-ray photoelectron spectroscopy (XPS) survey spectra (Figure S6 and Table S1, Supporting Information) revealed that the surface ruthenium content of RuO x /2D TiO x was 3.26 at%, which was further supported by inductively coupled plasma optical emission spectroscopy (ICP-OES, Table S2, Supporting Information), demonstrating a mass percentage of 6.56 wt% in RuO x /2D TiO x . [30]In addition, high-resolution XPS spectra of Ti 2p (Figure 2a) and Ru 3d (Figure 2b) were obtained for 2D TiO x , RuO x /2D TiO x , and reference samples of commercial anatase and RuO 2 .The Ti 2p spectra showed that the valence state of Ti in 2D TiO x was 0.3 eV higher than that in TiO 2 due to cation vacancies, [31] while the addition of ruthenium in RuO x /2D TiO x restored Ti to its tetravalent state.In contrast, the Ru 3d spectra showed that the valence state of Ru in RuO x /2D TiO x was 0.4 eV lower than that in RuO 2 .Moreover, Ru 3p 1/2 spectra (Figure S7, Supporting Information) showed that the valence state of Ru in RuO x /2D TiO x was 0.6 eV lower than that in RuO 2 , indicating that Ru atoms in 2D TiO x were not completely oxidized. [32,33]These results, combined with highresolution electron microscopy observations, suggest the presence of sub-nano ruthenium oxide clusters in RuO x /2D TiO x .
X-ray absorption near-edge structure (XANES) and extended X-ray fine structure (EXAFS) spectroscopy techniques were employed to investigate the electronic structure and atomic coordination environment of Ru atoms in RuO x /2D TiO x , and the results were compared with Ru foil and RuO 2 control samples. [34]o evaluate the oxidation state of the Ru cation, we obtained the absorption energy (E 0 ) from the first maximum in the first-order derivative of the K-edge XANES curves of RuO x /2D TiO x (Figure 2c). [35]As shown in the inset, we plotted the absorption energy as a function of oxidation state for the three Ru materials.The average valence of Ru in RuO x /2D TiO x was found to be approximately 3.68, which is consistent with the result of the XPS analysis.38] The coordination environment of Ru atom in RuO x /2D TiO x was further elucidated by Fourier transformations (FT) of k 3weighted for EXAFS (FT-EXAFS) (Figure 2d).The peaks located at 1.5 and 2.5 Å correspond to Ru-O and Ru-Ru, respectively. [32]avelet transform (WT)-EXAFS was used to verify Ru in the RuO x /2D TiO x catalyst (Figure 2e,f ).The WT plot of RuO x / 2D TiO x shows a maximum of 4.4 Å À1 , which is assigned to Ru-O coordination.In addition, the maximum intensity of the Ru-O spectrum shifts toward a slightly lower k than that of RuO 2 , and the detected Ru-Ru coordination signal is weaker than that of RuO 2 , which may be due to the different coordination structure of the Ru atoms or the presence of charge transfer between the RuO x cluster and the 2D TiO x substrate.As shown in Figure S8, Supporting Information, the FT-EXAFS curve of RuO x /2D TiO x was fitted in k and R space to obtain the coordination environment around Ru.As shown in Table S3, Supporting Information, Ru atoms in RuO x /2D TiO x coordinate with oxygen atoms in the form of low-nucleus clusters, which accords well with the experimental and simulation results.

Electrochemical Properties
Electrochemical evaluation was carried out to probe the chlorine evolution performance of RuO x /2D TiO x by coating it on a glassy carbon rotating disk electrode (RDE) under a rotation rate of 1600 rpm in 1 M NaCl (pH = 1) electrolyte at room temperature (%25 °C). [39]Under such conditions, the equilibrium potential for CER (E CER ) is calculated to be 1.36 V versus the reversible hydrogen electrode (RHE). [15]inear sweep voltammetry (LSV) curves (Figure 3a) clearly indicate the superior CER activity of RuO x /2D TiO x compared to 2D TiO x and benchmark CER electrocatalysts, commercial RuO 2 , and DSA catalysts.RuO x /2D TiO x delivers an onset potential of 1.39 V, which is 30 mV higher than the E CER (1.36 V vs RHE for 25 °C).Its overpotential is 98 mV at a current density of 10 mA cm À2 , which is much lower than that of 2D TiO x (339 mV), RuO 2 (125 mV), and DSA (115 mV).Considering the ruthenium content is only 6.56 wt%, the mass activity of RuO x /2D TiO x reaches 0.650 A mg Ru À1 at an overpotential of 100 mV, which is 26.5 and 143.5 times that of RuO 2 (0.0245A mg Ru À1 ) and DSA (0.00453 A mg Ru À1 ) (Figure 3b).The electrochemical surface areas (ECSAs) for the above electrodes were estimated from the electrochemical double-layer capacitance (C dl ) of the catalytic surface (Figure S9-S11 and Table S4, Supporting Information).The ECSAs normalized activity of the RuO x /2D TiO x electrode is 2.47 and 12.36 times higher than those of the RuO 2 and DSA (Figure 3b and S12, Supporting Information).We further confirmed the role of carbon black (VXC-72) in this work (Figure S13, Supporting Information).As the initial potential and limiting current density of VXC-72 are much lower than those of RuO x /2D TiO x , it can be confirmed that it is mainly to increase the conductivity of the catalyst rather than to exist as an active component. [40]afel analysis is a critical means to evaluate the CER kinetics of catalysts (Figure 3c).RuO x /2D TiO x showed a Tafel slope of 43.6 mV dec À1 , lower than RuO 2 (53.2 mV dec À1 ) and DSA (53.2 mV dec À1 ).The Tafel analyses indicate that the RuO x /2D TiO x possesses the fastest CER kinetics, confirming the significant contribution of 2D TiO x substrate in promoting the reaction of Ru-O active sites.The electrochemical impedance spectra (EIS) revealed that RuO x /2D TiO x possessed the smaller charge-transfer resistance than RuO 2 (Figure S14, Supporting Information), further proving the improvement of CER kinetics. [5]he selectivity between CER and OER is the most critical factor to determine the Faradaic efficiency (FE) of CER electrocatalysts. [6,24]We performed a LSV measurement for RuO x /2D TiO x in absent Cl À in 1 M NaClO 4 electrolyte (pH = 1) and observed dramatically decreased LSV current densities (Figure S15, Supporting Information), confirming an extremely poor OER electrocatalytic activity of RuO x /2D TiO x .This means that the acquired anodic currents by RuO x /2D TiO x in 1 M NaCl electrolyte are originated from CER.A commonly accepted rotating ring-disk electrode (RRDE) method determined an average FE of RuO x /2D TiO x toward CER to be 96.5 AE 1.5% (Figure 3d), which is higher than RuO 2 (76.5 AE 1.8%). [39]This result indicates that the thin layer of TiO 2 in RuO x /2D TiO x successfully tunes the surface electron density of the catalyst, thereby significantly enhancing the CER selectivity. [23,24]Additionally, the CER stability of RuO x /2D TiO x was examined by chronopotentiometry at a constant current density of 10 mA cm À2 (Figure 3e).The overpotential of RuO x /2D TiO x increased by 65 mV after 210 h of operation.The degradation of RuO 2 was much more serious.After only 140 h, the overpotential increased by 110 mV, and then the performance rapidly decayed and failed.It is worth mentioning that at high current densities, such as 50 and 100 mA cm À2 , the advantage of RuO x /2D TiO x is more obvious (Figure S16, Supporting Information).This may stem from the overoxidation of Ru to volatile RuO 4 or soluble RuO 5 2À under anodic potentials (>1.4 V) in aqueous electrolytes. [41,42]However, the RuO x sites in RuO x /2D TiO x possess a lower oxidation state of Ru, which can resist the overoxidation of Ru to some extent.These findings not only confirm the synergistic effect of sub-nano RuO x clusters and 2D TiO x substrate but also reveal that RuO x /2D TiO x is one of the most effective Ru-based CER electrocatalysts in low Cl À concentration (C Cl À ≤ 1 M) electrolytes (Figure S17 and Table S5, Supporting Information).
Furthermore, we extended our investigation by assessing the CER activity and selectivity of RuO x /2D TiO x in a 0.1 M NaCl solution (Figure S18, Supporting Information).When subjected to a current density of 10 mA cm À2 , the overpotential of RuO x / 2D TiO x slightly increased to 112 mV (0.1 M NaCl, pH = 1, E CER = 1.42 V vs RHE) from its previous value of 98 mV (1 M NaCl, pH = 1, E CER = 1.36 V vs RHE).The FE of RuO x /2D TiO x for CER also experienced a reduction, from 96.5 AE 1.5% to 88.5 AE 2.1%.Assuming that the selectivity remains constant at high current densities, Figure S19, Supporting Information, shows the current-voltage relationship including the corresponding parts j CER and j OER .
We further compared the CER activity, selectivity, and durability of RuO x /2D TiO x with DSA at a 4 M NaCl concentration, as depicted in Figure S20, Supporting Information.Even at a heightened chloride ion concentration, RuO x /2D TiO x from our study continues to maintain a significant advantage over the DSA catalyst.The overpotential of RuO x /2D TiO x measures 82 mV (4 M NaCl, pH = 1, ECER = 1.32 V vs RHE), which is lower than that of DSA at 97 mV (Figure S20a, Supporting Information).On one hand, we continued to assess the FE of RuO x /2D TiO x toward CER using the RRDE method (Figure S20b, Supporting Information).Our findings revealed an increase in RuO x /2D TiO x FE from 96.5 AE 1.5% to 98.3 AE 1.5%.On the other hand, we opted for a N-diethylp-phenylenediamine (DPD) colorimetric measurement to gauge the Faraday efficiency of both RuO x /2D TiO x and DSA catalysts.Illustrated in Table S6, Supporting Information, the average FE of RuO x /2D TiO x stood at 97.91%, surpassing the average FE of DSA (92.13%).The aforementioned results underscore the impact of reduced chloride ion concentration on the activity and selectivity of RuO x /2D TiO x .
We selected chronopotentiometry at a current density of 100 mA cm À2 to assess its stability.The results are presented in Figure S20c, Supporting Information.RuO x /2D TiO x demonstrates stable operation for 1000 s, after which the voltage gradually increases until it eventually ceases functioning.After running RuO x /2D TiO x at a current density of 100 mA cm À2 for 20 000 s, XPS testing was conducted on the catalyst.The results, presented in Figure S21, Supporting Information, indicate that the valence state of Ru in RuO x /2D TiO x increased by 0.2 eV compared to its initial state, implying that Ru underwent oxidation to some degree during the CER process.Conversely, the Ti 2p spectra showed no change in the valence state of Ti in RuO x /2D TiO x after 20 000 s of CER.Moreover, we detected 0.186 ppm of Ru ions in the solution after the reaction using ICP, normalized to 11.34% Ru loading on the electrode.This indicates that 11.34% of the Ru metal leached out (Table S7, Supporting Information).The results suggest that the decrease in catalyst activity could be a result of ruthenium oxidation and leaching.In contrast, the commercial DSA demonstrated stability throughout a 90 000 s trial period, with virtually no decline in its performance.DSA's stability is a result of its critical factors: 1) prevention of surface soluble product formation, 2) maintenance of mechanical strength, and 3) promotion of electronic conductivity.Operating within its potential limits, DSAs can have a lifespan of 10 years or longer. [10]

Theoretical Calculations
To understand the origin of the high activity and selectivity of RuO x /2D TiO x , the DFT calculations were performed.According to the XRD and HRTEM analyses, the (101) crystal plane of 2D anataseTiO 2 was adopted for calculations.Moreover, based on previous research and XPS results that the Ti vacancies (V Ti ) were formed during the synthetic process, [28,43] a ratio of V Ti with 0.17 was considered in our simulations, as shown in Figure S22 and S23, Supporting Information.We utilized a Ru 4 O 8 cluster with a C 2v geometry for calculations, as shown in Figure S23b, Supporting Information, which has been experimentally and theoretically demonstrated from previous studies. [44,45]The optimized structure of 2D anatase Ti 0.83 O 2 (101) surface with a Ru 4 O 8 cluster is displayed in Figure 4a.It is found that one of the Ru atoms in the cluster is occupied the lattice position of V Ti , with an adsorption energy of -11.34 eV, indicating a strong interaction between the Ru 4 O 8 cluster and the Ti 0.83 O 2 (101) surface.
The CER is a two-electron process through the Volmer-Heyrovsky mechanism, in which the adsorption and discharge of a chloride anion to form the adsorbed chlorine species *Cl (Volmer step: * þ 2Cl À !*Cl þ Cl À þ e À ) is followed by the direct combination with a chloride anion from the electrolyte accompanied by an electron transfer to produce Cl 2 (Heyrovsky step: * [2,5,[46][47][48][49] In this case, we proposed seven possible active sites on Ru 4 O 8 cluster and the corresponding *Cl adsorption models to compare with commercial RuO 2 (Figure 4a and S24, Supporting Information).Figure 4b and Table S8, Supporting Information, depict the Gibbs free energy changes for CER over all possible sites of Ru 4 O 8 cluster according to Volmer and Heyrovsky steps at the equilibrium potential of 1.36 V versus RHE.It is observed that the ΔG *Cl of Ru 4 O 8 -site 2 is the most favorable, leading to a theoretical overpotential as low as 0.17 V.The *OCl pathway over Ru site (sites 1, 2, 3) on Ru 4 O 8 /2D Ti 0.83 O 2 has been further explored.As shown in Figure S25, Supporting Information, the DFT results reveal that *OCl forms endothermically on site 1 and site 3, leading to an energydemanding step for Cl 2 formation with a thermodynamic overpotential of 1.82 and 0.33 V, respectively, while *OCl formation on site 2 is exothermic with a theoretical overpotential of 0.50 V.All overpotentials of the *OCl intermediate pathway are significantly higher than that of the *Cl pathway on site 2, indicating that *Cl species serves as a precursor for Cl 2 evolution over Ru 4 O 8 /2D Ti 0.83 O 2 .
In comparison, the CER overpotential of the RuO 2 (110) surface is 0.35 V via the *OCl pathway, which is twice as high as that of the Ru 4 O 8 cluster.This suggests that the CER process may occur much more easily on the Ru 4 O 8 cluster, consistent with the experimental results.Further, a projected density of states In addition, we have evaluated the CER performance of Ti on Ru 4 O 8 -Ti 0.83 O 2 as CER active site by using DFT calculations (Figure S28, Supporting Information).Our results reveal that the theoretical overpotential on Ti atom is 0.34 V, which is twice that of Ru 4 O 8 -site 2 (0.172 eV).This result combined with the poor CER experimental activity of 2D TiO x (Figure 3a) suggests that the possibility that Ti of RuO x /2D TiO x served as the active center is low, and the real active site should come from the ruthenium oxide cluster.
In order to clarify the role of Ti vacancy of 2D TiO x , we have prepared RuO 2 -TiO 2 (commercial anatase) and compared its electrochemical CER performance with that of RuO x /2D TiO x in 1 M NaCl at pH = 1 (Figure S29, Supporting Information).The CER performance of RuO x /2D TiO x is much higher than that of RuO 2 -TiO 2 (commercial anatase).At a current density of 10 mA cm À2 , overpotential of RuO 2 -TiO 2 is slightly increased to 214 mV higher than 98 mV of RuO x /2D TiO x .Then, the CER performance on the possible active sites of Ru 4 O 8 -TiO 2 (101) (Figure S30, Supporting Information) has been explored by using DFT calculations.There are seven possible active sites on Ru 4 O 8 cluster and corresponding *Cl adsorption models.As shown in Figure S31 and Table S9, Supporting Information, it is found that *Cl on sites 1, 2, 4 migrated to site 3 and the optimal ΔG *Cl on TiO 2 (101)-Ru 4 O 8 is 0.32 eV, leading to a theoretical overpotential of 0.32 V.This overpotential is higher than that of Ti 0.83 O 2 (101)-Ru 4 O 8 (0.17 V), suggesting that the existence of Ti vacancy can directly modulate the Ru 4 O 8 cluster to facilitate the CER performance.

Conclusion
In summary, a cationic defect adsorption-oxidation anchoring strategy is proposed to fabricate oxidized sub-nano ruthenium clusters on a 2D low-crystallinity titanium oxide substrate.Experimental and theoretical results proved that the presence of TiO x thin layers successfully enhances the dispersion of Ru atoms, optimizes the electronic structure of RuO x , as well as enhances the interaction between RuO x and the 2D TiO x substrate.As a result, the mass activity of the RuO x /2D TiO x increases significantly, by 31.6 and 143.5 times that of RuO 2 and DSA, respectively, along with increased selectivity (FE of 96.5%) and CER stability (over 210 h).This work highlights the tremendous potential of 2D TiO x substrates with cationic defects in improving the atom utilization and active site density of oxidized sub-nano noble metal clusters.
Synthesis of RuO x /2D TiO x Catalyst: First, rutile-form TiO 2 , K 2 CO 3 , and Li 2 CO 3 powders in a molar ratio of 1.73:0.4:0.14 were mixed and calcinated at 1000 °C for 20 h to obtain alkali metal titanate product (K 0.8 Ti 1.73 Li 0.27 O 4 powder) with lepidocrocite-like layered structure.Then, the protonic form, H 1.07 Ti 1.73 O 4 •H 2 O, was obtained by stirring the titanate crystals in 0.5 M HCl solution at room temperature for 2 d.The protonic titanate crystals were collected by filtration, washed with copious amounts of pure water, and air-dried.After H þ exchange, the product was added to a TBAOH aqueous solution (%10% w/v) and mechanically mixed for 2 d at room temperature to obtain monolayer-dominated 2D TiO x nanosheets.Then, the excess TBAOH was removed by centrifugal washing, and 2D TiO x nanosheets were redispersed into ultrapure water.
In order to obtain oxidized sub-nano ruthenium clusters on 2D lowcrystal titanium oxide, RuCl 3 aqueous solution and 2D TiO x dispersion were mixed dropwise under continuous stirring at a determined mass (1:10) ratio.After ultrasonication for half an hour, it was stirred and heated at 60 °C overnight.Then the flocculate was recovered and washed after centrifugation at 6000 rpm and then dried in air.Finally, the obtained Ru ad -TiO x was annealed in air at 300 °C for 2 h to obtain the final product RuO x /2D TiO x .
Material Characterization: XRD patterns were recorded using a Philips X-Pert Pro X-ray diffractometer with Cu Kα radiation (λ Kα1 = 1.5418Å) at 40 kV and 40 mA.Scanning electron microscopy images were obtained by a Hitachi SU8020 operated at an accelerating voltage of 10.0 kV.HRTEM, HAADF-STEM, and EDS measurements were acquired on a JEOL-NEOARM instrument (Anhui University) at 200 kV.The thickness and surface morphology of the 2D TiO x and RuO x /2D TiO x nanosheets were determined by AFM (Cypher ES, Asylum Research, Oxford Instruments) using the tapping mode.ICP results were obtained by an ICP6300 (ThermoFisher Scientific, USA).XPS was performed using an ESCALAB 250 X-ray photoelectron spectrometer (Thermo, USA) equipped with Al Kα 1,2 monochromatized radiation at 486.6 eV X-ray source.Ru K-edge X-ray absorption spectra were carried out using the hard X-ray spectroscopy beamline at the Australian Synchrotron in Australia.All EXAFS fitting was performed using an S 0 2 value obtained by modeling the EXAFS of a reference Ru foil (K-edge at 22 117 eV).Athena and Artemis codes were used to extract the data and fit the profiles. [50]lectrochemical Tests: Chlorine evolution activity.A three-electrode system was built using an H-type cell to separate the working electrode from the counter electrode, in which a reference electrode was placed at the compartment of working electrode.Each compartment of the H-type cell was separated by a Nafifion 117 membrane (DuPont).Prior to use, the Nafifion membrane was pretreated with 5% H 2 O 2 and heated at 60 °C for 1 h.A glassy carbon disk (5 mm in diameter) and a glassy carbon disk (5.5 mm in diameter) with a Pt ring (6.5 mm in inner diameter and 8.For the preparation of working electrode, 2 mg of the catalyst and carbon black (VXC-72 CABOT, 2.0 mg) was dispersed in a 1 mL solution (50 μL Nafion solution was dispersed in 0.95 mL ethanol) under sonification.The above dispersion (25 μL) was then loaded onto a glassy carbon electrode and dried naturally at ambient environment; the loading amount of catalyst is %0.255 mg cm À2 .The activity for the CER was subsequently tested by scanning between approximately 1.0 and 2.0 V with a scan rate of 5 mV s À1 and all the polarized profiles were calibrated with 95% iR compensation.The potentials were converted into the RHE reference scale, after IR compensation, with the following equation: where R is the gas constant (R = 8.314 J K À1 mol À1 ), T is the temperature (in Kelvin), and F is the Faraday constant (F = 96485.3C mol À1 ).In comparing overpotentials of catalysts, the equilibrium potential of CER was calculated using the equation reported. [15] where α(Cl 2 ) was assumed by Guerrini et al. to be 0.01, with respect to the partial pressure of Cl 2 evolving under Ar purging, [51] and α(Cl À ) was dependent on molar concentration of the electrolyte (i.e., α(Cl À ) = 1.0 for 1.0 M NaCl). [52]The E 0 CER value was derived through the following equation: where dE 0 dT corresponds to 0.001248, according to a previous report. [53]The overpotential was calculated by the potential difference between the equilibrium potential for CER and the potential experimentally needed to drive a specific current density.
EIS measurements were made by applying an alternating voltage with an amplitude of 5 mV at frequencies ranging from 10 6 to 0.1 Hz.Chronopotentiometric measurements were conducted with the current density of 10, 50, and 100 mA cm À2 .Commercial RuO 2 catalyst was also measured at a loading of 0.255 mg cm À2 for comparison.All electrochemical measurements were conducted at room temperature (%25 °C).
DPD colorimetric measurement.The DPD colorimetric method for FE Cl 2 measurement was according to the references. [5,54]The Cl 2 selectivity was calculated according to the following relation: The experimental condition and setup were the exact same as those used in RRDE studies.Each part of the H-type cell was filled with 100 mL of electrolyte.The electrolyte was sparged with Ar gas for 20 min before Cl 2 evolution.CA was then performed with the applied potential, which was adjusted to generate a current density higher than 10 mA cm À2 for 120 s.The theoretical yield was calculated to be 6.219Â 10 À6 mol Cl 2 .
Immediately after finishing the CA, 5 mL of the anodic electrolyte was moved into 20 mL vial containing 10 mL NaOH solution (0.1 mol L À1 ) and then injected into a 50 mL colorimeter containing 5 mL DPD (1 g L À1 ).Then 15 mL phosphate buffer solution (PBS, pH = 7) was dripped into and chlorine-demand free water added to the scale line and stirred for 5 min.The Cl 2 concentration was determined by measuring the absorbance at 515 nm using UV--vis spectrophotometer.Measurements were also conducted for the electrolyte that had not been used for electrochemical measurements as a blank.At least three measurements were collected per electrode.In these experiments, the water used is all chlorine-demand free water.The Faradaic efficiency was calculated by the following equation: where j is the current (A); t is the reaction time (s); F is Faraday's constant (F, 96485.3389C mol À1 ); the factor ½ is based on that the 2 mol e À that were transferred for the oxidation of two Cl À ions to per mol Cl 2 molecule; V is the total value of electrolyte (100 mL); and C is the Cl 2 concentration of electrolyte.
Theory Calculation: All spin-polarized DFT calculations were performed using the Vienna Ab initio simulation package. [55,56]The ion-electron interactions were described using the projector-augmented wave [57] potential and the exchange-correlation energy was treated by the Perdew-Burke-Ernzerhof functional at the generalized gradient approximation (GGA) level. [58]61][62][63] A cutoff energy was set to 520 eV, and k-points were sampled using the Monkhorst-Pack mesh with a reciprocal space resolution of 2π Â 0.04 Å À1 for geometry optimization.Moreover, the Coulomb interaction effect (Hubbard U ) is also considered to describe the magnetism of the Ti atoms, and the GGA þ U with U À J = 4.2 eV was used. [28]The vacuum region was set at 20 Å to eliminate artifactual interactions between the periodically repeated images.All atoms were allowed to relax until the Hellmann-Feynman forces were smaller than 0.01 eV Å À1 , and the convergence criterion for the electronic self-consistent loop was set to 10 À5 eV.For all considered adsorbates (i.e., *Cl, *OCl, *OH, *O, and *OOH, * refers to the corresponding catalytic site), each reaction can be written as follows: Then, adsorption energies (ΔE) for each species were calculated at zero potential and standard conditions to obtain following equations: At pH = 0, the Gibbs free energies (ΔG) as a function of applied potential U can be written as follows at 298.15 K: To theoretically investigate the CER activity, the overpotential for CER (η CER ) can be defined as follows: η CER ¼ jΔG ÃCl j e or η CER ¼ jΔG ÃOCl À ΔG ÃO j e (21)

Figure 1 .
Figure 1.Synthesis and characterizations of RuO x /2D TiO x .a) Schematic illustration of the synthesis of RuO x /2D TiO x .b) XRD patterns of 2D TiO x and RuO x /2D TiO x .c) TEM image, d) HRTEM image, and SAED pattern of 2D TiO x .e) HRTEM image and SAED pattern, f ) AFM image, g) HAADF-STEM image, and h) EDS mapping of RuO x /2D TiO x .

Figure 3 .
Figure 3. Electrochemical CER performance of RuO x /2D TiO x .a) LSV curves of RuO x /2D TiO x , 2D TiO x , RuO 2 , and DSA in 1 M NaCl at pH = 1.b) The Ru mass-normalized current and the ECSA-normalized current density evaluated with RuO x /2D TiO x , RuO 2 , and DSA at η = 100 mV.c) Tafel slopes.d) CER selectivity of RuO x /2D TiO x and RuO 2 measured by RRDE.e) Chronopotentiometry curves of RuO x /2D TiO x and RuO 2 under current density of 10 mA cm À2 .

(
PDOS) analysis of seven Ru(4d)/O(2p)-involved orbitals is performed to determine the reason why Ru 4 O 8 -Site 2 shows the best configuration (Figure 4c).Ru 4 O 8 -site 2 exhibits the most PDOS distribution near the Fermi level.Additionally, it has a moderate number of antibonding states, which provides a suitable binding energy for the Cl intermediate.Therefore, Ru 4 O 8 -site 2 demonstrates excellent chlorine evolution activity.Due to the possible O 2 evolution during the CER process, the selectivity of Cl 2 over the Ru 4 O 8 cluster was further investigated, as shown in Figure 4d.The formation of the *O intermediate from *OH on Ru 4 O 8 -site 2 is identified as the most endothermic step, resulting in the potential-determining step with a theoretical η OER of 1.50 V at 1.23 V versus RHE.To reference the η OER to the same potential as CER, it is corrected to 1.37 V [η OER = 1.50 À (1.36 À 1.23) = 1.37 V].This indicates that the occurrence of OER is difficult on the Ru 4 O 8 cluster even in the absence of Cl À .As such, the Ru 4 O 8 cluster supported on the anatase Ti 0.83 O 2 (101) surface exhibits high activity and excellent selectivity toward CER in acidic solution, which agrees with our experimental results.Moreover, the OER over RuO 2 /DSA has been evaluated, as depicted as Figure S26, Supporting Information.The OER reaction energy barrier on the RuO 2 (110) crystal surface was 0.74 eV at 1.23 V and 0.61 eV at 1.36 V (η CER = 0), respectively.In comparison, the CER overpotential on the RuO 2 (110) surface via the *OCl pathway was 0.35 V.It is observed that the bonding states of the active Ru and O on the RuO 2 (110) surface are predominantly situated at deeper energy levels (Figure S27, Supporting Information), while these states shift toward the Fermi level of the RuO x cluster.Therefore, our results demonstrate the effective modification of the electronic structure in the RuO x /2D TiO x system, thereby facilitating the subsequent evolution of Cl 2 .

Figure 4 .
Figure 4. DFT calculations of plausible model systems of RuO x /2D TiO x in CER and OER conditions.a) Side view of the optimized configuration of the Ru 4 O 8 cluster supported on 2D Ti 0.83 O 2 .Seven possible active sites on Ru 4 O 8 cluster and corresponding *Cl adsorption models.b) Gibbs free energy diagrams for CER over the Ru 4 O 8 cluster supported on 2D Ti 0.83 O 2 and RuO 2 .c) The PDOS of Ru(4d)/O(2p)-involved orbitals.The Fermi level is set at 0 eV.d) Gibbs free energy diagrams for OER over site 2 of the Ru 4 O 8 cluster supported on 2D Ti 0.83 O 2 .
5 mm in outer diameter) were used in RDE and RRDE (Pine Research Instrumentation, the collection efficiency of 0.37) experiments, respectively.Pt counter electrode and Ag/AgCl electrode containing the filling solution of 3 M NaCl were used as the counter and reference electrodes, respectively.The electrolytes were prepared by diluting 70% HClO 4 and by adding 99.8% NaCl in 18.2 MΩ cm Millipore water.For experiments under different NaCl concentrations, NaClO 4 •H 2 O was added into the electrolyte to compensate the total ionic strength.The pH values of the acidic electrolytes were adjusted to 1.00 by adding a few drops of 70% HClO 4 .The pH values of all electrolytes were measured using a digital pH metre (FE28-Standard, METTLER).