Enhanced redox kinetics of Prussian blue analogues for superior electrochemical deionization performance

Prussian blue analogues (PBAs), representing the typical faradaic electrode materials for efficient capacitive deionization (CDI) due to their open architecture and high capacity, have been plagued by kinetics issues, leading to insufficient utilization of active sites and poor structure stability. Herein, to address the conflict issue between desalination capacity and stability due to mismatched ionic and electronic kinetics for the PBA-based electrodes, a rational design, including Mn substitution and polypyrrole (ppy) connection, has been proposed for the nickel hexacyanoferrate (Mn–NiHCF/ppy), serving as a model case. Particularly, the theoretical calculation manifests the reduced bandgap and energy barrier for ionic diffusion after Mn substitution, combined with the increased electronic conductivity and integrity through ppy connecting, resulting in enhanced redox kinetics and boosted desalination performance. Specifically, the optimized Mn–NiHCF/ppy demonstrates a remarkable desalination capacity of 51.8 mg g−1 at 1.2 V, accompanied by a high charge efficiency of 81%, and excellent cycling stability without obvious degradation up to 50 cycles, outperforming other related materials. Overall, our concept shown herein provides insights into the design of advanced faradaic electrode materials for high-performance CDI.


Introduction
The global freshwater shortage has become one of the most urgent challenges for human society with the rapid development of industry and serious environmental pollution. 1,2To alleviate the water crisis, it is necessary to explore and develop novel and low-cost technologies for the desalination of brackish and saline water. 3[9][10] Given this, exploring high-performance electrodes, featuring high desalination capacity and good cycling stability, is promising yet challenging for the practical application of CDI.
Recently, replacing the Na + capture side from the carbonaceous electrode with a faradaic electrode has drawn increasing attention, which involves the redox reaction for CDI process, and dramatically improves the desalination capacity. 11,129][20] Particularly, such structure advantages render, for example, the nickel hexacyanoferrate (NiHCF) to be a promising electrode candidate with remarkable desalination performance. 11,21Promoted by the structure advantages of NiHCF for CDI application, smartly designing the framework of NiHCF and assembling it on a conductive substrate to fabricate a hybrid electrode, aiming to facilitate the desalination performance, has become the mainstream. 11,16,18,20However, lacking rational control will cause serious aggregation and stacking of NiHCF upon selfassembly on a conductive substrate, restricting the migration of electrons/ions and inducing high concentration polarization, which in turn degrades the desalination performance. 20Based on literature review, the application of NiHCF for CDI, even with modications of nano-structure design and conductive matrix hybridization, suffers from unsatised desalination performance (Fig. 1): [21][22][23][24][25][26][27] (1) limited desalination capacity (usually less than 50 mg g −1 ) due to low electrochemical activity; (2) sluggish desalination rate owing to poor electronic conductivity and intrinsic structural defects; and (3) degraded capacity retention upon cycling (usually less than 90% # 50 cycles) resulting from lattice collapse.Of note, the mismatched kinetics between ions and electrons should be responsible for such degradation on desalination, and more sophisticated structural control and hybridization are required for NiHCF.
To create novel nanomaterials through the self-assembly of building blocks on a conductive substrate is very promising for synthesizing hybrid structures with controlled composition, structure, and morphology. 28Specically, for the atomic-level compositional control, rational substitution at the atomic level for the active material at target sites can dramatically improve electrochemical performance. 29For the optimization of structure and morphology, it is crucial not only to achieve the uniform distribution of building blocks on the substrate but also to ensure rational interactions between these building blocks.This is essential for avoiding aggregation and stacking. 30,31The comprehensive design of CDI electrode materials, spanning from atomic scale to nanoscale and meso/ macroscopic scale, is anticipated to enhance electronic conductivity between the building block and substrate.Simultaneously, improved ionic conductivity among building blocks is expected, reducing the disparity between ionic and electronic kinetics. 30This modication in turn enhances redox kinetics and facilitates faradaic reactions. 31his work proposes the tailoring of Mn substitution and ppy connection, aiming to achieving rapid redox kinetics and high desalination performance and using NiHCF as the model case (Mn-NiHCF/ppy).The superior electron/ion migration capability of the target Mn-NiHCF/ppy is conrmed through density functional theory (DFT) calculation, combined with systematic characterizations.Additionally, the enhanced desalination performances under different working voltages and feed concentrations have also been demonstrated.As a result, the optimized Mn-NiHCF/ppy achieves a high desalination capacity of 51.8 mg g −1 at 1.2 V in a feed solution of 500 mg L −1 , along with a stable cycling up to 50 cycles and only 0.11% decay per cycle, outperforming other MHCF counterparts.

Experimental
Preparation of the Mn-NiHCF/ppy: according to previous literature, 11,32 Ni(NO 3 ) 2 $6H 2 O (1.0 mmol), Mn(NO 3 ) 2 $4H 2 O (0.1 mmol), and sodium citrate (2.25 mmol) were dispersed in deionized water (50 mL) under vigorous stirring to form solution A. Meanwhile, a certain amount of as-prepared ppy (50, 100, and 150 mg) was added into the above solution, respectively.In a separate step, K 3 Fe(CN) 6 (2.25 mmol) was dispersed in 100 mL of deionized water to form solution B. Aerward, solutions A and B were simultaneously poured into 50 mL of deionized water, which was continuously stirred for 5 hours and then set aside for 24 hours.The resulting mixture was then thoroughly washed with deionized water and dried overnight at 60 °C to obtain the target Mn-NiHCF/ppy.Noting that the samples with the addition of 150, 100, and 50 mg of ppy were denoted as Mn-NiHCF/ppy-1, Mn-NiHCF/ppy-2, and Mn-NiHCF/ppy-3, respectively.For comparison, Mn-NiHCF was synthesized using the same precursor as mentioned above, but without the addition of ppy.Similarly, the pure NiHCH was prepared without the manganese source and ppy in the precursor.Please refer to the ESI † for more details of the preparation of ppy tubes, material characterization, electrochemical testing, and desalination experiments.

Results and discussion
Before comparing the discrepancies on ionic kinetics and electronic properties aer Mn substitution based on DFT calculation, the optimized substitution sites of Mn in the NiHCF framework should be determined.As shown in Fig. S1 (ESI †), the results reveal that the total energies for Mn substitution at the high-spin Ni sites and low-spin Fe sites are calculated to be −0.32 and −0.23 eV, respectively.This analysis establishes that Mn substitution at the high-spin Ni sites is more stable compared to the low-spin Fe sites.3][34][35][36] Based on the results of substitution site calculation, the positive effect of Mn substitution for the ionic diffusion and energy barrier in NiHCF and Mn-NiHCF has been compared.The possible diffusion pathway and the corresponding energy barrier for Mn-NiHCF are illustrated in Fig. 2a and b, with the calculated results for NiHCF shown in Fig. S2 (ESI †).It is evident that the energy barrier of the substituted sample is signicantly lower than that of pristine NiHCF.The respective values are 0.25 eV for NiHCF and 0.21 eV for Mn-NiHCF.This discrepancy highlights the altered ionic diffusion kinetics attributed to the Mn substitution. 18,37,38Apart from the ionic kinetics, the electronic properties of NiHCF and Mn-NiHCF have also been compared through the total density of state (TDOS) and partial density of state (PDOS).From the TDOSs of NiHCF and Mn-NiHCF (Fig. 2c and d), the bandgap of NiHCF (2.45 eV) is reduced aer Mn substitution, indicating the modied electronic conductivity for the Mn-NiHCF. 32Moreover, a detailed investigation of the PDOSs for the two samples (Fig. 2e and f) reveals that it is the impurity energy levels induced by Mn 3d states, as well as its interaction with the Ni 3d states, lead to the reduction of the bandgap. 39,40Of note, the Fermi level almost shis to the newly formed conduction band for the Mn-NiHCF, further combining with the enhanced ionic diffusion kinetics, which highlights the modied ionic/ electronic dynamics aer Mn substitution. 32Particularly, the Mn substitution shows advantages for increasing the ionic/ electronic conductivity of the NiHCF, hence facilitating the redox reaction upon desalination. 32,41Given this perspective, the strategic doping of metals, aimed at achieving modied electronic and ionic transfer concurrently, emerges as a promising strategy for altering the internal structure and properties of PBAs.
The DFT calculation results above have proved the superiority of Mn-NiHCF on ionic/electronic kinetics, thus verifying the signicance of rational Mn substitution for NiHCF.To get a clear understanding for the synthesis of the target Mn-NiHCF/ ppy, the schematic illustration has been shown in Fig. 3a.First of all, the tubular ppy was fabricated through a reported method, employing the ferric trichloride as the initiator. 424][45] Fig. 3b depicts the X-ray diffraction (XRD) patterns of the as-prepared samples.As shown, the diffraction peaks for the ve samples can be well indexed to the JCPDS no.46-0908, thus verifying the Mn substitution and introduction of ppy show little effect for the Prussian blue crystal structure. 7,16Notably, the closed investigation for the (200) peak at around 17.6°reveals that the Mn substitution leads to its le shi (Fig. S3, ESI †), which means expanded interplanar spacing, hence benetting the Na + migration. 46,47n regard to the X-ray photoelectron spectroscopy (XPS) measurements, the obvious Mn signal at around 643.2 eV present in the survey spectra (Fig. 3c) of Mn-doped samples evidently veries the Mn substitution in the framework.(708.5 eV) and Fe(III) 2p 3/2 (710.1 eV), indicating the mixed states of Fe 2+ and Fe 3+ . 24For the Ni 2p spectra, four main peaks can be obtained aer deconvolution, including Ni(II) 2p 1/2 (874.1 eV), Ni(III) 2p 1/2 (876.2eV), Ni(II) 2p 3/2 (856.4 eV), and Ni(III) 2p 3/2 (858.4 eV). 22In addition, the deconvolutions of Mn 2p 1/2 at around 651.9 and 653.4 eV correspond to the Mn(II) 2p 1/2 and Mn(III) 2p 1/2 , respectively, while those located at around 640.9 and 645.3 eV represent for Mn(II) 2p 3/2 and Mn(III) 2p 3/2 . 48In principle, the Mn substitution in the framework, activating the electrochemical activity of NiHCF, provides abundant active sites and facilitates the faradaic redox reactions upon desalination. 49ig. 3d depicts the Fourier transform infrared (FTIR) spectra of as-prepared samples.For the NiHCF, the characteristic infrared bands at 3640 cm −1 can be attributed to the stretching of -OH, while the bands at around 2163 and 2096 cm −1 correspond to the ngerprints of divalent iron (Fe(II)) and trivalent iron (Fe(III)), respectively, assigning to Fe 3+ -CN and Fe 2+ -CN. 7,11he band at 590 cm −1 represents the cyano group (-C^N) stretching, whereas the metal-carbon bonding can be detected at around 455.0 cm −1 . 33Aer Mn substitution and ppy connection, the stretching of -OH shows a slight right shi due to the enhanced interaction between -OH and -C^N (-OH/ N). 11 Additionally, Mn substitution leads to the amalgamation of Fe 2+ -CN and Fe 3+ -CN into a single broad peak with a le shi, centered around 2066 cm −1 . 50,51][52] Moreover, the alterations in bonds congurations observed in the FTIR spectra provide additional conrmation of the substitution of Mn for Ni within the framework.Based on the above characterizations, the compositional and spatial control not only optimizes the inner structure of NiHCF but also induces improved interaction between the active framework and ppy substrate, thereby contributing to enhanced redox kinetics and high integrity of the hybrid electrode and enabling high desalination performance. 53,54pecically, for the NiHCF and Mn-NiHCF, the ion/electron transfer in electrode largely depends on the contact of particles, that is, a particle-to-particle conductive pathway (Fig. S6a, ESI †). 55Nevertheless, the unsatised electronic conductivity of NiHCF and Mn-NiHCF not only restricts the charge transfer but also limits the utilization of redox sites, leading to poor desalination performance. 42Field emission scanning electron microscope (FESEM) observation for the NiHCF and Mn-NiHCF reveals that both of them show well-dispersed cube-shaped particulates with a size of about 150 nm (Fig. S6b-e †), which is in good agreement with earlier investigation, 24 thus showing the successful synthesis of NiHCF and little inuence of Mn substitution for the NiHCF framework.
Aer the introduction of ppy, the electronegativity of ppy induces the in situ nucleation of Mn-NiHCF on the surface of ppy, 56 which can be clearly observed in Fig. 4. Compared with the pure NiHCF and Mn-NiHCF, the cubic particles on the ppy display similar size, and the number of detected particles  gradually increases with the reduced addition of ppy in the precursor, showing high consistency with our design.Indeed, the tight hybridization between Mn-NiHCF and ppy can ensure high electronic conductivity. 23However, only short-range electronic/ionic conductivity at the contact area can be obtained for Fig. 4a-c, and the separated dispersion of Mn-NiHCF particles induces the lack of long-range conductivity, especially for ions. 42Furthermore, the substantial loading of ppy may potentially compromise the desalination capacity of the hybrid electrode, given that the primary contributor to desalination capacity is the NiHCF. 44Increasing the loading of Mn-NiHCF in the composite induces the homogeneous dispersion of Mn-NiHCF on the tubular ppy (Fig. 4d-f), simultaneously obtaining good contact between ppy and Mn-NiHCF particles, which leads to enhanced ionic/electronic conductivity. 32,46However, the excess of Mn-NiHCF causes aggregation of Mn-NiHCF particles (Fig. 4g-i). 21Noting that controlling the ppy loading changes the morphology of Mn-NiHCF hybrids, ensuring modied electronic/ionic conductivity and improved electrochemical performance. 22Concerning the elemental mapping, the uniform and co-existence of Ni, Mn, Fe, C, and N for the Mn-NiHCF/ppy-2 are obviously shown in Fig. 4j, which conrms the tight combination between Mn-NiHCF and ppy.As observed from Fig. S7 (ESI †), the electronic conductivities measured from the four-probe tests further reveal the increased electronic conductivity aer Mn substitution and ppy connection, and the higher loading of ppy results in higher electronic conductivity.Given this, integrating conductive agents into PBAs serves the dual purpose of bolstering electronic conductivity while maintaining the structural integrity of the electrode, and this strategy is expected to notably elevate the electrochemical performance of the target electrode.
To investigate the electrochemical properties of as-prepared samples, cyclic voltammetry (CV) and galvanostatic discharge/ charge (GCD) measurements based on the three-electrode system in 1.0 M Na 2 SO 4 solution have been performed.As shown in Fig. 5a, the redox reaction of high-spin Fe(III)/Fe(II) coordinated with N of cyano group (Fe-N^C) upon Na + insertion/desertion results in the redox pair at around 0.6 V, which can be clearly seen for all samples, thus revealing the reversible faradaic reactions of those NiHCF-based electrodes. 57he modications in both composition and structure show different CVs compared with the pristine NiHCF electrode. 58esides, the Mn-NiHCF/ppy-2 shows the maximum integrated area under the relevant CV curve, thus indicating the highest ion storage ability and thereby excellent desalination performance. 44Additionally, the higher ion storage ability of Mn-NiHCF/ppy-2 can be also demonstrated from the GCD proles at 1.0 A g −1 (Fig. 5b) because the optimized Mn-NiHCF/ppy-2 displays the longest voltage plateau and charge/discharge time, further conrming its superiority in ion storage among all samples. 7Particularly, the specic capacitance decreases in the order of Mn-NiHCF/ppy-2 (338 F g −1 ) > Mn-NiHCF/ppy-3 (300 F g −1 ) > Mn-NiHCF/ppy-1 (265 F g −1 ) > Mn-NiHCF (221 F g −1 ) > NiHCF (190 F g −1 ) at 1.0 A g −1 .Noting that the GCD proles vary with different current densities, and the duration of charge/discharge plateau decreases at high current densities due to the fact that partial active sites cannot be utilized effectively (Fig. S8, ESI †). 6,11Moreover, the evident charge/discharge plateaus exhibiting high symmetry in the GCD proles at various current densities for all samples, despite differences in duration, underscore the high electrochemical reversibility of the faradaic reaction within these electrodes. 18,20Fig. 5c compares the specic capacitances of the as-fabricated ve samples at various current densities, where the target Mn-NiHCF/ppy-2 exhibits superiority over those other samples on capacitances at all current densities (346, 338, 317, 312, and 308 F g −1 at 0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 A g −1 , respectively), showing excellent ion adsorption ability and potentially high desalination capacity. 56enerally, it is considered a typical capacitive-dominated electrochemical process (eqn (1)) when the specic current (i) in the CV curves shows a linear relationship with the scan rate (v). 6While the specic current varying with the square root of the scan rate (v 1/2 ) represents the diffusion-controlled process. 22pacitive-controlled electrochemical process: Diffusion-dominated electrochemical process: where k 1 and k 2 are parameters varied by voltage, i corresponds to the specic current at a xed voltage, and v represents the scan rate, separately. 59Specically, both the diffusion and capacitive behaviors contribute to the current in CV except for the ideal battery and electrical double-layer capacitance materials, and the correlation between them can be explained as follows: which can be further modied into eqn (4) to facilitate the subsequent calculation.
Based on the above discussion, aer solving for k 1 and k 2 from the slope and the Y-axis intercept by tting a straight line across multiple scan rates at each specic potential, the contributions for the current induced by capacitive or diffusion behaviors can be determined, hence allowing to separate the CV into the capacitive-controlled and diffusion-dominated regions. 60In principle, the surface-related kinetics behaviors correspond to the capacitive process, while the solid-state diffusion process oen accompanies with slow mass transfer. 61Based on the CV curves of Mn-NiHCF/ppy-2 electrode at various scan rates (Fig. 5d), the capacitive contributions at 5.0 and 10.0 mV s −1 are 85.1% and 90.7%, respectively, aer data collection and tting (Fig. 5e).Besides, the comparison of capacitive and diffusion contribution of all prepared NiHCF electrodes at 5 and 10 mV s −1 according to the CV curves in Fig. S9 (ESI †) has been illustrated in Fig. 5f and S10 † (ESI †), where the capacitive contribution follows the order: Mn-NiHCF/ppy-2 > Mn-NiHCF/ppy-3 > Mn-NiHCF/ppy-1 > Mn-NiHCF > NiHCF.It's important to highlight that the clustered Mn-NiHCF on the ppy surface in Mn-NiHCF/ppy-3 diminishes the exposed surface area, limiting ionic/electronic migration.Consequently, this results in a lower capacitive contribution compared to Mn-NiHCF/ppy-2. 62The substantial capacitive contribution observed in the charge storage of Mn-NiHCF/ppy-2 underscores the efficacy of the rational design involving Mn substitution and ppy connection.This design proves advantageous in achieving rapid redox kinetics and a high ion storage capacity, thereby facilitating the subsequent desalination process. 11y means of CV curves at various scan rates, the investigation of Na + transportation process on the surface of various electrodes was also performed.The higher peak currents aer Mn substitution and ppy connection at the same scan rate indicate the improved charge transfer and increased electrochemical activity. 63,64Regarding the linear tting results of peak current versus square root of scan rate (Fig. 5g), the high index (R 2 > 0.99) for the tting reveals that all the ve samples undergo the same ionic diffusion mechanism, all of which is controlled by surface diffusion. 65Specically, the exact ionic diffusion coef-cient can be obtained based on the following equation: where, I p (A), n, S (cm 2 ), v (mV s −1 ), and C 0 (mol L −1 ) represent the peak current, the number of electrons involved in the redox reaction, the load area of active material, the scan rate, and the concentration of Na + . 66Table S1 (ESI †) compares the Na + diffusion coefficient of as-prepared ve samples, where the Mn substituted sample (Mn-NiHCF) exhibits a larger average coef-cient (1.73 × 10 −10 cm 2 s −1 ) compared with pristine NiHCF (7.93 × 10 −11 cm 2 s −1 ).Additionally, the introduction of ppy induces the further increase of diffusion coefficient, and the desired Mn-NiHCF/ppy-2 exhibits the highest diffusion coefficient of 5.76 × 10 −10 cm 2 s −1 . 66Additionally, the Nyquist plots of all as-prepared samples (Fig. 5h) reveal that the target Mn-NiHCF/ppy-2 possesses the lowest charge transfer resistance (R ct ) based on the smallest semi-circle in the high-frequency region, demonstrating superior charge transfer of Mn-NiHCF/ ppy-2, which shows high agreement with the calculation and characterization results. 67n situ EIS measurements were further employed to estimate the detailed evolution of charge transfer kinetics, taking the optimized Mn-NiHCF/ppy-2 as an example.For the charging process, as shown in Fig. S11a (ESI †), the quasi-semicircle in the high-frequency region, corresponding to the R ct , is restricted, while the obvious slope line in the low-frequency region indicates favorable ionic diffusion on the surface of the electrode at the initial charge, demonstrating ion storage by EDL in this stage. 3,60Aer charging to 0.6 V, the obvious difference can be detected from the appearance of the quasi-semicircle and the restriction of the slope line, which manifests that the charge transfer resistance dominates the faradaic reactions. 65Further charging to 1.2 V, away from the redox region, results in the vanish of quasi-semicircle and convenient ionic diffusion.Similar to the charging process, the in situ EIS measurement for the discharge process (Fig. S11b, ESI †) also displays the capacitive feature at the region away from the redox reaction, and the charge transfer controls the faradaic reaction. 68In summary, the ion storage by synergistic effect of faradaic reaction and EDL can not only guarantee high capacitance but also induce fast redox kinetics, hence contributing to modifying the desalination performance. 6n general, the batch mode was employed for all desalination experiments, utilizing the active carbon and NiHCF-based composites as the positive and negative electrodes, respectively.A feedwater NaCl concentration of 500 mg L −1 and a voltage of 1.2 V were applied for the CDI cell to initially evaluate the desalination performance of the fabricated samples.As depicted, two typical stages, which involve electrochemical adsorption followed by desorption over one charge/discharge cycle (i.e., one desalination-regeneration cycle), could be observed from the conductivity-time proles for all NiHCF electrodes (Fig. 6a).The Mn-NiHCF/ppy-2 electrode exhibites a much faster and higher Na + ion capture ability compared to other electrodes during the desalination process.Besides, further comparison on the salt adsorption capacity (SAC) aer quantitative calculation (Fig. 6b) demonstrates that the NiHCF/ ppy-2 delivers the highest SAC of 51.8 mg g −1 , showing superior to other electrodes (37.0 mg g −1 for NiHCF, 40.7 mg g −1 for Mn-NiHCF, 44.4 mg g −1 for Mn-NiHCF/ppy-1, and 48.1 mg g −1 for Mn-NiHCF/ppy-3).In principle, the Mn substitution promotes both the ionic diffusion and electronic conductivity in the framework, combining with the optimized loading of ppy to guarantee the stability and integrity of the electrode, contributing to the high Na + capture. 44Moreover, the in situ growth of Mn-NiHCF on tubular ppy not only enhances the interfacial charge transfer but also improves the interaction between them, thus simultaneously obtaining modied redox kinetics and tight combination. 7,59Noting that the SACs of Mn-NiHCF/ppy electrodes are found to have a signicant dependence on ppy loading since insufficient ppy connection could not ensure good electronic conductivity or provide enough buffering matrices, while excessive ppy loading would sacrice the high electrochemical activity of NiHCF. 7ig.6c compares the salt adsorption rate (SAR) and SAC based on CDI Ragone plots (SAR versus SAC) among the prepared samples, from which we can get the conclusion that superior desalination performance can be obtained when the SAR versus SAC prole gets closer to the upper right corner in the Ragone plots. 68Noting that the distance of these SAR versus SAC proles follow the order of NiHCF < Mn-NiHCF < Mn-NiHCF/ppy-1 < Mn-NiHCF/ppy-3 < Mn-NiHCF/ppy-2, thus manifesting that the desired Mn-NiHCF/ppy-2 possesses both the fastest and highest Na + capture ability, further highlighting the enhanced redox kinetics and modied electrochemical activity for NiHCF.Fig. S12 (ESI †) depicts the current response upon desalination of various NiHCF-based electrodes at 1.2 V in 500 mg L −1 feedwater, based on which the charge efficiencies are calculated to be 0.50, 0.60, 0.68, 0.81, and 0.75, respectively, for NiHCF, Mn-NiHCF, Mn-NiHCF/ppy-1, Mn-NiHCF/ppy-2, and Mn-NiHCF/ppy-3.Overall, the modied redox kinetics and improved integrity of the electrode endow the target Mn-NiHCF/ppy-2 with improved charge transfer and enhanced desalination performance. 69o reveal the potential for practical application of the asfabricated CDI cell, the desalination performance under various voltages and feed concentrations was evaluated.Fig. 6d and S13 (ESI †) show that the conductivities and corresponding SACs of all NiHCF-based electrodes are strongly depended on the voltage employed from 0.6 to 1.2 V.For all samples, the decline in conductivity indicates the capture of Na + , while the various modications applied to the NiHCF show differing in SACs.Specically, for the optimized Mn-NiHCF/ppy-2, the SAC decreases from 52.2 to 25.2 mg g −1 when the voltage drops from 1.2 to 0.6 V (Table S2, ESI †).It is worth noting that when the voltage is returned to 1.2 V, the SAC of Mn-NiHCF/ppy-2 can be recovered to 51.8 mg g −1 .In principle, the higher voltage employed for the CDI cell can improve performance due to enhanced sorption ability. 70However, side reactions also occur, degrading the cycling performance and reducing the charge efficiency. 7Based on the literature review of Na + capturing electrodes for CDI, the cell voltage of 1.2 V was oen selected to estimate the desalination performance, and therefore was adopted as the optimized voltage to investigate the effect of feedwater salt concentration for SAC (Fig. 6e and S14 in ESI †). 60ith the increase of salt concentration, the electrodes manifest rapid adsorption/desorption equilibrium as well as improved SAC.[23] Fig. 6  Additionally, the target Mn-NiHCF/ppy-2, featuring modied redox kinetics and good electrode integrity, exhibits the highest SAC among all NiHCF-based electrodes, showing desalination capacity of 38.8, 51.8, and 55.5 mg g −1 with the feedwater salt concentration of 250, 500, and 1000 mg g −1 , respectively (Table S3, ESI †).Compared with other MHCF electrodes reported previously, the optimized Mn-NiHCF/ppy-2 demonstrates superiority on SAC at 1.2 V with the feedwater salt concentration of 500 mg L −1 (Fig. 6f), thus indicating the signicance of compositional and structural control to modify the desalination performance of NiHCF. 57xcept for the SAC, another index to evaluate the CDI performance is cycling stability, which plays an important role in practical application of CDI.In this regard, the cycling performances of CDI cells with various NiHCF-based electrodes have been observed through repeated charge and discharge in 500 mL g −1 salt concentration at 1.2 V. Fig. 7a shows the feedwater conductivity and corresponding SAC plotted as a function of cycling number for all NiHCF-based electrodes.For the pristine NiHCF electrode, the SAC reduces from 37.0 to 22.9 mg g −1 from the initial cycle to the 50th cycle, with a capacity retention of 61.8% (Fig. 7b), showing degraded cycling stability.Interestingly, the Mn substitution (Mn-NiHCF) modies the cycling stability, with a desalination capacity of 32.9 mg g −1 aer 50 cycles and 80.8% capacity retention.Further combined with ppy, the high conductivity and efficient buffering induce enhanced desalination capacity and capacity retention, and the target Mn-NiHCF/ppy-2 shows the highest SAC of 49.2 mg g −1 aer 50 cycles and 94.9% capacity retention, further highlighting the signicance of composition/structure modication for desalination performance.Noting that the excess loading of NiHCF (Mn-NiHCF/ppy-3) displays inferiority on both SAC and retention compared with Mn-NiHCF/ppy-2, which can be ascribed to the fact that a large amount of NiHCF on ppy may restrict the migration of Na + , resulting in increased electrochemical polarization. 69o highlight the advantages of Mn-NiHCF/ppy-2 for other NiHCF-based electrodes in this work, a detailed comparison of electrochemical and desalination performance (Fig. 7c) demonstrates that the Mn-NiHCF/ppy-2 shows the superiorities on specic capacitance, capacitive contribution, SAC, SAC retention aer 50 cycles, and charge efficiency over the other NiHCF electrodes.To conrm the enhanced structure stability of the NiHCF electrode aer Mn substitution and ppy connection, the characterizations on XRD and FTIR before and aer cycling for the Mn-NiHCF/ppy-2 have been conducted.As shown in Fig. 7d, the crystal structure of NiHCF is well maintained aer 50 cycles since no new phases can be detected, while the results of FTIR reveal that the main functional groups of the pristine electrode can be well-maintained aer 50 cycles (Fig. S15, ESI †), thus showing the excellent structure stability of Mn-NiHCF/ppy-2 and its high promise for efficient desalination.Overall, the outstanding desalination performance exhibited by Mn-NiHCF/ppy-2 underscores the synergistic effect of Mn substitution and ppy connection, and such modi-cations represent typical compositional and morphological tailoring approaches, showcasing their feasibility and promise in enhancing the desalination performance of the PBAs electrode.

Conclusion
Herein, we strategically prepare a new hybrid structure of Mn-NiHCF/ppy.From both theoretical calculation and experimental measurement, we have demonstrated modied dynamics for both ions and electrons, improved electrochemical activity, and enhanced desalination performance.Particularly, the introduction of partial Mn into the NiHCF reduces both the diffusion energy of Na + in the cubic framework and the bandgap between conduction and valence bands, while the ppy substrate connects the separated NiHCF, further guaranteeing the conductivity and integrity of the hybrid electrode.Our concept leads to enhanced electrochemical activity and desalination performance.Importantly, the tailoring strategies reported in this work are also feasible for other faradaic electrodes for efficient desalination, hence promoting the development and practical application of CDI.

2 Fig. 2
Fig. 2 (a) Possible Na + diffusion pathway inside ideal Mn-NiHCF structure; (b) energy barrier profiles of ionic diffusion in NiHCF and Mn-NiHCF; total density of states and partial density of states of (c and e) NiHCF and (d and f) Mn-NiHCF.

Fig. 5
Fig. 5 Electrochemical measurements of the NiHCF-based electrodes in 1.0 M Na 2 SO 4 solution: (a) CV curves at 10 mV s −1 ; (b) galvanostatic discharge/charge profiles at the current density of 1.0 A g −1 ; (c) comparison of specific capacities at various current densities; CV curves of Mn-NiHCF/ppy-2 at (d) various scan rates and (e) capacitive contributions at 5.0 and 10.0 mV s −1 ; (f) comparison of capacitive contributions of the as-prepared five samples at 5.0 and 10.0 mV s −1 ; (g) fitting curves of peak current and square root of scan rate for all samples; (h) Nyquist plots for all samples with the inset of highfrequency zoom-in region.
Fig. 6 (a) Conductivity-time plots of the first desalination/regeneration with various NiHCF-based electrodes in 500 mg per L NaCl solution at 1.2 V and (b) corresponding salt adsorption capacities; (c) desalination Ragone plots of as-prepared five NiHCF-based electrodes; (d) conductivity versus time profiles and SACs of the Mn-NiHCF/ppy-2 electrode in 500 mg per L NaCl solution with various voltages in CDI cell; (e) SACs of NiHCF-based electrodes versus salt concentration; (f) comparison of desalination performance between the target Mn-NiHCF/ppy-2 and other MHCF-based electrodes in CDI.