Molten-salt-induced phosphorus vacancy defect engineering of heterostructured cobalt phosphides for efficient overall water splitting

Abstract

Cobalt phosphides (CoP_x) as a representative of transition metal phosphide (TMP) catalysts show great potential for highly-efficient electrocatalytic water splitting for H₂ production. However, the current synthetic strategies of CoP_x are complex and time-consuming. Herein, a novel, facile and fast molten salt strategy for one-step synthesis of CoP_x (CoP/Co₂P) was first proposed. Moreover, the molten salt's strong polarization force endows CoP_x with abundant phosphorus vacancy (PV) defects. The synthesized CoP_x show impressive catalytic activities towards both the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) under alkaline conditions. This enables the overall water splitting to reach the current densities of 10 and 100 mA cm⁻² driven by only 1.75 and 1.90 V voltage, respectively. Mechanism investigation reveals that PV defects lead to the optimization of the electronic structure, decreased energy barriers of intermediates’ formation, and enhanced electronic conductivity, all of which boost the electrocatalytic activity. This study provides a paradigm of using molten salt for the synthesis of advanced TMPs, which not only simplifies the synthetic procedures, but also provides insight into defect engineering that involves the use of molten salt medium to obtain a better activity.

---

1. Introduction

Renewable-energy-driven electrochemical water splitting provides one promising approach towards green H₂ production.¹ This process requires efficient electrocatalysts to overcome the energy barriers. Materials based on noble metals such as Pt and Ru/Ir are known as the state-of-the-art catalysts for the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER), respectively; nevertheless, the scarcity and high cost hinder their large-scale applications.^2,3 In this context, tremendous efforts have been devoted to developing noble-metal-free electrocatalysts with high performance.

Transition metal phosphides (TMPs, M includes Fe, Co, Ni, Cu, Mo, W, etc.) catalyzing energy conversion and storage reactions (e.g. hydrogenation, rechargeable batteries, electrocatalytic water splitting, etc.) have drawn growing attention in recent years due to their Earth abundance, unique hydrogenase-like structure, and excellent catalytic activity.^4–7 Some tactics including nanostructural engineering, heteroatom doping and heterostructural engineering have been proposed to further enhance TMPs’ performance.⁸ Here, a novel molten salt strategy was proposed to synthesize a TMP catalyst of cobalt phosphides (CoP_x) and simultaneously carry out phosphorus vacancy (PV) defect engineering for efficient water splitting. Defect engineering capable of modulating the electronic structure and thus influencing the chemisorption and catalytic process is one important strategy to improve catalysts’ activity.⁹ Though defect engineering has been extensively studied in transition metal oxides (TMOs),^10,11 transition metal dichalcogenides (TMDs),^12,13 and defective graphene supports,¹⁴ relevant research on TMPs is limited. In transition metal compounds, intrinsic defects were classified into anionic and cationic defects on the basis of the vacancy type, and both of them will cause local electron redistribution.¹⁵ PV defects belong to the class of anionic defects, but it requires more research efforts to reveal their regulation and impact on the electronic structure of TMPs. Moreover, compared to the various mature methods for defect engineering of TMOs and TMDs, the existing methods to regulate PV defects over TMPs are only limited to NaBH₄ reduction,¹⁶ plasma etching,¹⁷ and oxygen doping.¹⁸ More distinct and in-depth research studies on PV defect regulation and how it promotes the catalytic process are highly expected.

For the first time, we proposed a molten-salt-mediated phosphidation (MSMP) strategy to synthesize cobalt phosphide (CoP_x, CoP/Co₂P) in only one step, and meanwhile achieve an enrichment of PV defects in CoP_x, resulting in a substantially modulated electronic structure and an improvement in electrocatalytic activity. Molten salt is a phase changing material and represents a significant class of fluids for high-temperature applications.¹⁹ Recent studies showed that molten salt can serve as a medium for the synthesis of advanced functional materials, such as single atom catalysts (SACs),^20,21 semiconductors,²² and two dimensional (2D) MXenes²³ because of its unique liquid feature, strong polarization force and space confinement effect. Given that liquids have the merits of optimal heat control, fast mass transfer, and robust dynamics,²¹ material synthesis in a liquid environment would be more favorable than that in a gas or solid medium.

Using the MSMP strategy, CoP_x nanorods were obtained by directly heating CoSO₄ and NaH₂PO₂ in molten LiCl–KCl. CoSO₄ was first converted into cobalt (Co) oxides and then reacted with NaH₂PO₂-derived PH₃ to form CoP_x. This one-step synthesis strategy greatly simplifies the fabrication procedure of CoP_x. In previous studies, CoP_x synthesis required at least two steps: (i) pre-treating inorganic Co salts to synthesize Co oxides, hydroxides and metal organic frameworks (MOFs) that are usually used as Co precursors and (ii) the phosphidation of the above Co precursors (Table S8, ESI†). The CoP_x nanorods synthesized in molten salt possess abundant PV defects compared to the control and exhibit a boosted catalytic activity towards overall water splitting. Density functional theory (DFT) calculations were conducted to reveal the underlying promotion mechanism of PV defects towards catalytic activity. Formation of nickel phosphides (NiP_x, including NiP, Ni₂P and NiP₂) and ferric phosphide (Fe₂P) with the molten salt strategy confirms that the reported methodology can be extended to other TMPs’ preparation. This study provides a new opportunity for facile, fast, and one-step synthesis of efficient TMPs. Furthermore, the presented MSMP method extends the strategies for defect engineering, which may inspire the synthesis of more advanced materials in the future.

2. Experimental section

Chemicals

All chemicals used in this work were purchased from commercial sources and used without any further treatment. KCl (99.5%) was provided by Aladdin; LiCl·H₂O (≥97.0%) was provided by XL Scientific; CoSO₄·7H₂O (99.5%) was provided by MACKLIN; NaH₂PO₂ (99.0%) was provided by RHAWN. H₂IrCl₆·xH₂O was provided by Aladdin. Carbon fiber paper (CFP) was purchased from TORAY Company of Japan and employed as the catalytic support. Nafion solution (5 wt%) was provided by Alfa Aesar. All aqueous solutions were prepared using Milli-Q ultrapure water (18 MΩ cm).

Synthetic procedures

In this work, the synthesis of CoP_x was based on a novel and one-step MSMP strategy. Firstly, KCl (1.41 g), LiCl·H₂O (1.12 g), NaH₂PO₂ (1.00 g) and CoSO₄·7H₂O (0.53 g) (the molar ratio of phosphorus (P) to Co is approximately 5 : 1) were mixed and ground into a homogeneous mixture in an agate mortar. Afterwards, the mixture was transferred into a ceramic boat and subjected to annealing treatment in a tube furnace under a N₂ flow (100 mL min⁻¹) with a temperature programming of 10 °C min⁻¹. The mixture was kept at desired temperatures for 2 h. Thereafter, the resulting products were washed with deionized water (DI H₂O) and hydrochloric acid (HCl, 100 mL, 1 M) with the aid of bath sonication to remove the frozen salt (LiCl-KCl) and to possibly form cobalt oxides. The products were separated from the solvent by centrifugation at 8000 rpm. Finally, the products were washed with DI H₂O again to remove the impurities and residual HCl and then dried in a vacuum oven at 60 °C overnight. To study the effect of heating temperature on the catalytic performance, the heating temperature was set as 350, 400 and 500 °C, with the corresponding products denoted as MS-CoP_x-350, MS-CoP_x-400 and MS-CoP_x-500, respectively. For further comparison, the samples were prepared by directly annealing the mixture of NaH₂PO₂ and CoSO₄·7H₂O without LiCl-KCl at 350, 400 and 500 °C and denoted as CoP_x-350, CoP_x-400 and CoP_x-500, respectively. Purification procedures of CoP_x-350, CoP_x-400 and CoP_x-500 were the same as that of MS-CoP_x-350, MS-CoP_x-400 and MS-CoP_x-500. MS-NiP_x-350, NiP_x-350, MS-FeP_x-350 and FeP_x-350 were also synthesized and purified with the same procedures as that used for CoP_x.

3. Results and discussion

Microscopy characterization to reveal the structural features

Using the MSMP strategy (Fig. 1a), NaH₂PO₂ and CoSO₄·7H₂O were mixed with LiCl and KCl directly, and then heated under a N₂ flow to give rise to the growth of CoP_x. The samples prepared at 350, 400, and 500 °C in molten salt were denoted as MS-CoP_x-350, MS-CoP_x-400, and MS-CoP_x-500, respectively; while the control samples prepared without molten salt were denoted as CoP_x-350, CoP_x-400, and CoP_x-500, respectively (Fig. S1†). All samples were first characterized by scanning electron microscopy (SEM) to reveal their surface morphology. MS-CoP_x-350 shows a nanorod-like structure with a diameter of ∼100 nm (Fig. 1b), while CoP_x-350 is shown to be amorphous (Fig. S2†). The nanorod formation of MS-CoP_x-350 can be ascribed to the appropriate synthetic temperature and template effect of molten LiCl-KCl.²⁴ The nanorod structure is supposed to provide more catalytically active sites and facilitate charge and mass transfer during the electrochemical reaction, hence generally showing a higher catalytic activity than amorphous ones.^25,26

Fig. 1 Synthesis route and structural characterization of MS-CoP_x-350. (a) Schematic illustration of the MSMP strategy; (b) SEM graph with a scale bar of 200 nm; (c) TEM image with a scale bar of 200 nm; (d) HAADF-STEM image with a scale bar of 100 nm; (e₁–e₄) HRTEM images with a scale bar of 1 nm; (f) SAED patterns with a scale bar of 5 nm⁻¹; (g and g₁–g₄) EDS mappings with a scale bar of 50 nm.

To further understand the structural features, transmission electron microscopy (TEM) was carried out to characterize MS-CoP_x-350 (Fig. 1c). Small nanoparticles and flakes were observed to adhere on or drift away from the nanorods, which further enhances the exposure of active sites. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of MS-CoP_x-350 (Fig. 1d) further shows a nanorod structure. The high resolution TEM (HRTEM) images (Fig. 1e₁–e₄) show clear lattice fringes. Multiple lattices were exposed: lattice spacings of 1.75 Å, 2.47 Å and 2.83 Å correspond to the CoP's (103), (111) and (011) planes, respectively; a lattice spacing of 2.21 Å is assigned to the Co₂P's (121) plane. Selected area electron diffraction (SAED) by TEM (Fig. 1f) reveals several bright rings that arise from CoP (103), CoP (011) and Co₂P (121), as marked by green, yellow and red circles, respectively. This agrees well with the HRTEM data. TEM data of the comparison samples are shown in Fig. S3;† evidence for the polycrystalline feature of MS-CoP_x-350 is shown in Fig. S4 and S5.† The energy dispersive spectroscopy (EDS) mapping images of MS-CoP_x-350 (Fig. 1g and g₁–g₄) suggest that the dispersions of O, P and Co are even and thereby reveal the actual chemical makeup of CoP_x nanorods. By comparing the SEM-EDS results (Fig. S6, S7 and Table S1†), one can note that the CoP_x nanorods synthesized in molten salt demonstrate phosphidation to a higher extent. For example, MS-CoP_x-350 shows a higher P atomic ratio of 35.76% than CoP_x-350 (27.76%). The molten salt medium provides a space confinement effect, which on the one hand promotes the high dispersion of reactant precursors,²⁷ and on the other hand inhibits the fast blowing away of PH₃ by a N₂ flow. The lower phosphidation of MS-CoP_x-400 and MS-CoP_x-500 than that of MS-CoP_x-350 is because higher temperature led to a faster decomposition of NaH₂PO₂ and subsequently a faster loss of PH₃. Overall, the SEM and TEM data demonstrate the successful preparation of nanorod-like CoP_x hybrids with multiple planes’ exposure by the MSMP strategy. Previously, CoP_x fabrication required two or more steps, because Co precursors (oxides, hydroxides, MOFs, etc.) need to be first synthesized by the pre-treatment of inorganic Co salts.²⁸ The novel CoP_x synthesis strategy presented here demonstrates great progress in simplifying the fabrication process.

Spectroscopy characterization to verify the introduction of PV defects

Fig. 2a compares the X-ray diffraction (XRD) patterns of all CoP_x samples. The diffraction peaks at around 31.6°, 35.2°, 36.3°, 46.1°, 48.1°, 52.3°, 55.9°, and 56.8° were indexed to the crystal planes of (011), (200), (111), (112), (211), (103), (020), and (301) of CoP, respectively (Table S2†). The peaks at around 40.7° and 43.1° correspond to the planes of (121) and (211) of Co₂P, respectively. These XRD patterns indicate that the synthesized CoP_x is polycrystalline, in accord with the HRTEM and SAED data. MS-CoP_x-350 shows a better crystallinity than CoP_x-350, and this can be explicated by the enhanced phosphidation due to the space confinement of the molten salt medium. Besides CoP and Co₂P, the XRD patterns of CoP_x-500 and MS-CoP_x-500 also show the presence of Co₂O₃ (220) and Co₃O₄ (220), suggesting the in situ conversion from CoSO₄ to cobalt oxides. From the XRD results, we further reveal the formation mechanism of CoP_x by the MSMP strategy: CoSO₄ was first converted to cobalt oxides, which then reacted with PH₃ to form CoP_x. The enhanced phosphidation extent explains the absence of the Co₂O₃ (220) and Co₃O₄ (220) planes in MS-CoP_x-350. For the Raman spectra (Fig. 2b), the signals within the rectangular region are assigned to CoP;²⁹ the bands centered at around 470, 510, 605 and 660 cm⁻¹ are attributed to cobalt oxides;^30,31 the modes at around 1040 cm⁻¹ are consistent with P–O stretching vibrations in a disordered system.³²

Fig. 2 Spectroscopic characterization of CoP_x. (a) XRD patterns; (b) Raman spectra; (c) EPR spectra; (d–f) XPS spectra of MS-CoP_x-350 (top) and CoP_x-350 (bottom): (d) Co 2p; (e) P 2p and (f) O 1s. The binding energy of XPS spectra was corrected using C 1s 284.6 eV.

Electron paramagnetic resonance (EPR) spectroscopy was employed to identify the formation of PV defects (Fig. 2c). The strong signals at g = 2.007 for both CoP_x-350 and MS-CoP_x-350 indicate the presence of unpaired electrons, which supports the formation of PV.³³ We further quantified the PV concentrations by comparing the EPR intensity,^34,35 and the results suggest that the PV concentration in MS-CoP_x-350 is 2.41 times higher than that in CoP_x-350. Akin to the synthetic temperature of 350 °C, MS-CoP_x-400 and MS-CoP_x-500 have PV concentrations 1.36 and 1.24 times higher than those of CoP_x-400 and CoP_x-500, respectively (Fig. S8†). These results verified the fact that more PV defects are induced by the molten salt medium. Molten salt generally establishes a robust interaction with the precursor and then provides a strong polarization force.²⁷ In more detail, there is an electronic transmutation between Li (LiCl) and P atoms that obeys the “Zintl-rule”, wherein less electronegative Li donates an electron to more electronegative P, making each P bear a negative charge.^36,37 The bonding structure of Li–P tends to weaken Co–P and gives rise to more PV generation. Structural vacancy usually results in electron redistribution and thereby a significant change in the electronic structure.⁹ Accordingly, X-ray photoelectron spectroscopy (XPS) was applied to reveal the difference in the electronic states between MS-CoP_x-350 and CoP_x-350 (Fig. 2d–f and Tables S3, S4†). For MS-CoP_x-350 (Fig. 2d, top), the Co 2p peaks were deconvoluted into six characteristic peaks within Co 2p_3/2 regions at binding energies of 781.39, 783.44, and 787.19 (satellite) eV and Co 2p_1/2 regions at binding energies of 797.53, 799.93, and 804.19 (satellite) eV. The peaks centered at 781.39 and 797.53 eV are ascribed to the Co–P bonding of CoP_x, while other two peaks at higher binding energies of 783.44 and 799.93 eV are assigned to the oxidized Co species upon contact with air.³⁸ The high-resolution P 2p spectra of MS-CoP_x-350 (Fig. 2e, top) can be deconvoluted into two peaks with a doublet centered at 129.73 and 130.36 eV arising from P 2p_3/2 and P 2p_1/2 of CoP_x, respectively. Moreover, a broad peak emerging at 133.72 eV implies the surface oxidation.³⁹ Within the O 1s region of MS-CoP_x-350 (Fig. 2f, top), two deconvoluted peaks appeared at 532.05 and 533.74 eV are attributed to the Co–O and P–O species, respectively,⁴⁰ consistent with the oxidation species formation indicated by Co 2p and P 2p spectra. By comparison to CoP_x-350 (Fig. 2d, bottom), we conclude that molten salt induces a visible negative shift in the binding energy (∼0.6–0.9 eV, Table S3†) of the Co 2p spectra. Correspondingly, the P 2p region of MS-CoP_x-350 was observed to positively shift when compared to that of CoP_x-350 (Fig. 2e, bottom).

This result implies the electron transfer from PV to the Co center. The observed energy shifts also coincide with the EPR results since the unpaired electrons of PV tend to increase the electron density of the Co center. Such modulation of the electronic structure of CoP_x is expected to enhance the catalytic performance (see below). XPS characterization of CoP_x-400, CoP_x-500, MS-CoP_x-400, and MS-CoP_x-500 (Fig. S9, S10 and Table S3†) further confirmed the increased electron density around the Co center upon introducing the molten salt medium. Altogether, the results strongly supported that the molten salt medium enables defect engineering and modulation of the electronic structure on CoP_x.

Electrocatalytic activity under alkaline conditions

The electrocatalytic activity of CoP_x towards the HER and OER was evaluated in N₂-saturated 1 M KOH solution using a standard three-electrode configuration. Pt/C and IrO₂ served as the control for the HER and OER, respectively. As indicated by the HER polarization curves (Fig. 3a), Pt/C shows the highest HER activity with an overpotential of only 22 mV to afford a benchmark current density of 10 mA cm⁻². MS-CoP_x-350 demonstrates a far lower overpotential (∼190 mV) than CoP_x-350 (∼300 mV) at 10 mA cm⁻², implying the positive impact of molten salt on the HER activity. This promotion is also observed at other temperatures (MS-CoP_x-400, ∼215 mV vs. CoP_x-400, ∼243 mV; MS-CoP_x-500, ∼273 mV vs. CoP_x-500, ∼319 mV) (Fig. S12†). The HER is a multiple-step process and occurs through the Volmer–Heyrovsky or Volmer–Tafel mechanisms.^4,8 In the Volmer reaction (adsorption step), a H₂O molecule reacts with an electron and generates an adsorbed hydrogen atom (H*) on the electrode surface. Then, the Heyrovsky or Tafel reaction (desorption step) proceeds. Lower H* coverage prefers the Heyrovsky reaction, while higher H* coverage benefits the Tafel reaction. At room temperature, the theoretical Tafel slopes of the Volmer, Heyrovsky and Tafel reactions are 118, 39 and 29 mV dec⁻¹, respectively.⁴ To explore the HER kinetics and deeply understand the reaction mechanism, the Tafel slope was calculated from the corresponding polarization curves (Fig. 3b). MS-CoP_x-350 has a smaller Tafel slope (84 mV dec⁻¹) than CoP_x-350 (116 mV dec⁻¹), MS-CoP_x-400 (93 mV dec⁻¹) and MS-CoP_x-500 (103 mV dec⁻¹), thus revealing a faster HER kinetics. This also indicates that MS-CoP_x-350 abides by the Volmer–Heyrovsky HER pathway, wherein the Volmer reaction is the rate-limiting step.⁴¹ The superior HER activity of MS-CoP_x-350 is further supported by the electrochemical impedance spectra (EIS) (Fig. 3c and Table S5†). MS-CoP_x-350 exhibits a reduced charge transfer resistance (R_ct) of 187.5 Ω compared with CoP_x-350 (436.8 Ω). Lower R_ct is beneficial for the combination between H* and electrons, which promotes the HER reaction.^42,43 In catalysis, the electrochemical active surface area (ECSA) is proportional to electrochemical double-layer capacitance (C_dl).⁴⁴ Therefore, C_dl (Fig. S13†) was calculated from the cyclic voltammograms measured with varying scanning rates (Fig. S14†) to evaluate the ECSA. The results show that MS-CoP_x-350 has the highest C_dl value of 21.9 mF cm⁻², which is roughly 5, 3, and 4 times greater than those of CoP_x-350 (4.8 mF cm⁻²), MS-CoP_x-400 (7.5 mF cm⁻²), and MS-CoP_x-500 (5.5 mF cm⁻²), respectively. This suggests that MS-CoP_x-350 can provide more active sites during the reaction process and agrees well with the lowest overpotential it required to reach 10 mA cm⁻². All the above results underscore the key role of molten salt in improving the CoP_x's HER activity. The HER performance of MS-CoP_x-350 is comparable to those of the state-of-the-art CoP- and Co₂P-based electrocatalysts (Table S8†). The HER activity of CoP_x was also evaluated in 0.5 M H₂SO₄ (Fig. S15, S16 and Table S6†), and MS-CoP_x-350 surpassed other samples owing to its highest phosphidation extent. The excellent HER activity of MS-CoP_x-350 under both alkaline and acidic conditions reveals its potential to serve as a pH-universal cathodic electrocatalyst to catalyze water splitting.

Fig. 3 Electrochemical characterization of CoP_x towards the HER and OER in 1 M KOH. (a–c) belong to the HER. (a) Polarization curves; (b) Tafel slopes calculated from the corresponding polarization curves; (c) EIS spectra obtained with an overpotential of 100 mV and fitted with the equivalent circuit models shown in Fig. S17.† (d–f) belong to the OER. (d) Polarization curves of CoP_x; (e) Tafel slopes calculated from the corresponding polarization curves; (f) EIS spectra obtained with an overpotential of 300 mV. The HER and OER polarization curves were acquired with a scan rate of 5 mV s⁻¹ (80% IR).

For OER measurements, MS-CoP_x-350 shows a similar electrocatalytic activity (∼300 mV overpotential) to the control of IrO₂ (∼290 mV overpotential) at 10 mA cm⁻², which is obviously higher than that of CoP_x-350 (∼351 mV overpotential) (Fig. 3d). With the assistance of molten salt, the catalysts synthesized at other temperatures also have lower overpotentials of ∼334 mV (MS-CoP_x-400) and ∼310 mV (MS-CoP_x-500) than CoP_x-350 without molten salt. The corresponding Tafel plots (Fig. 3e) show the smallest slope of 68 mV dec⁻¹ over MS-CoP_x-350 among all samples (CoP_x-350, 91 mV dec⁻¹; MS-CoP_x-400, 84 mV dec⁻¹; MS-CoP_x-500, 75 mV dec⁻¹), suggesting its highest OER kinetics. MS-CoP_x-350 also has the lowest R_ct value (24.19 Ω) compared to CoP_x-350 (81.36 Ω), MS-CoP_x-400 (54.83 Ω) and MS-CoP_x-500 (32.5 Ω) under the same overpotential of 300 mV (Fig. 3f and Table S7†). All results suggest that the synthesis within the molten salt medium enhanced the alkaline OER activity of CoP_x. The increase of the molten salt temperature to 400 and 500 °C did not improve the OER activity very obviously (Fig. S12†), corresponding to the reduction of phosphidation at these higher temperatures (Tables S1 and S4†), which further proves that both the HER and OER are dependent on the phosphidation extent. Notably, the as-prepared MS-CoP_x-350 surpasses most of the reported OER catalysts and locates among the most efficient ones (Table S8†).

In addition to CoP_x, we further extended the reported molten salt strategy to the synthesis of other TMPs. XRD characterization (Fig. S18†) suggests the successful preparation of NiP_x including NiP, Ni₂P, and NiP₂. However, besides Fe₂P, there are a lot of impurities in the synthesized FeP_x compounds such as NaH₂PO₄, Fe₃(PO₄)₂ and some ferric oxides, indicating the insufficient phosphidation and by-product generation.

Electrochemical measurements of NiP_x and FeP_x were carried out under the same conditions as those for CoP_x (Fig. S19 and Table S9†). MS-NiP_x-350 shows overpotentials of 220 and 280 mV towards the HER and OER, respectively, which are similar to those of MS-CoP_x-350. By comparison, due to the impurities and insufficient phosphidation, MS-FeP_x-350 shows much poorer catalytic activities towards the HER (414 mV) and OER (440 mV) when compared to MS-NiP_x-350. Although we did not obtain pure FeP_x here, the presence of Fe₂P also supports the occurrence of the phosphidation reaction with molten salt. From the FeP_x case, we further confirm that more research efforts are needed in improving the purity of specific TMPs or even controlling the crystal exposure.

Overall water splitting under alkaline conditions

Given the prominent catalytic activities of MS-CoP_x-350 towards both the HER and OER under alkaline conditions, a two-electrode electrolyzer (MS-CoP_x-350||MS-CoP_x-350) with MS-CoP_x-350 serving as both cathodic and anodic catalysts was assembled to carry out overall water splitting in 1 M KOH solution (Fig. 4). Fig. 4a demonstrates the dependence of current generation on applied cell voltage within the indicated electrolyzer, and it shows that ∼1.75 V produces ∼10 mA cm⁻². This result coincides well with the HER and OER polarization curves over MS-CoP_x-350 (Fig. 4a, inset). The potentiostatic electrolysis examined with varying cell voltages (Fig. 4b) shows that the current density is positively correlated with the applied voltage; in detail, 1.75 V, 1.85 V, and 1.90 V produce current densities of ∼10, ∼50, and ∼100 mA cm⁻², respectively. The durability was examined by the chronoamperometry test at 100 mA cm⁻² for 24 hours (Fig. 4c). The cell maintained a stable voltage with a negligible increase (<10%) subsequent to 24-hour electrolysis, suggesting that MS-CoP_x-350 has an impressive stability. A digital graph of the electrolyzer (1.75 V) shows the generation of H₂ and O₂ bubbles on the electrodes’ surface (Fig. 4d). CoP_x subsequent to 24-hour water electrolysis was further characterized by XPS to explore the changes in catalyst's surface chemistry (Fig. S21 and Table S10†). Both the post-HER and post-OER samples show similar chemical compositions to pristine CoP_x, which further verifies the excellent stability. The Co 2p spectra of post-HER and post-OER CoP_x experienced a slight negative shift, while the P 2p spectra did not show any visible shift, indicating that Co cite may serve as the active site. Post-HER MS-CoP_x-350 was further characterized by XRD and SEM. The good consistence in the XRD patterns (Fig. S22, inset†) indicates that MS-CoP_x-350 was not decomposed during the long-term reaction. The SEM images of post-HER MS-CoP_x-350 (Fig. S23a and b†) prove that the catalyst maintains its originally rod structure basically. The above results agree well with the excellent catalytic stability. In summary, MS-CoP_x-350 is capable of catalyzing the overall water splitting with the catalytic performance comparable to those of other noble-metal-free catalysts (Table S11†), and moreover, it can be easily manufactured.

Fig. 4 Overall water splitting by the MS-CoP_x-350||MS-CoP_x-350 electrolyzer in 1 M KOH solution. (a) Polarization curve of the electrolyzer for overall water splitting, and inset: polarization curves of the HER and OER over MS-CoP_x-350; (b) potentiostatic electrolysis by the electrolyzer with varying cell voltages; (c) 24-hour chronopotentiometry test of the electrolyzer at 100 mA cm⁻²; (d) digital graph of the electrolyzer at 1.75 V cell voltage.

DFT study to reveal the promotion mechanism of PV towards electrocatalytic activity

Density functional theory (DFT) calculations were conducted to explore the underlying promotion mechanisms of PV defects towards catalytic activity. CoP (103) and Co₂P (121) were selected to model CoP and Co₂P, respectively, because, (i) their presence was confirmed by both XRD and TEM; and (ii) they appeared in all samples. The constructed slab models for CoP and Co₂P are depicted in Fig. S25.† We firstly evaluated the adsorption behavior of H₂O over P and Co sites by calculating the Gibbs free energy diagram (ΔG_H2O) (Fig. 5a). The computational results show that the Co sites of CoP and Co₂P give exothermic ΔG_H2O values of −0.555 and −0.809 eV, respectively, which are more negative than the P sites of CoP (−0.125 eV) and Co₂P (−0.179 eV), indicating a better affinity between Co and H₂O. Therefore, Co was suggested to be the active site for both CoP and Co₂P. This is in line with the post-electrolysis samples’ XPS characterization (Fig. S21 and Table S10†), where the Co 2p region experiences a visibly negative shift for both the post-HER and post-OER samples, but no obvious change is observed in the P 2p spectra.

Fig. 5 DFT calculations. (a) Adsorption energy of H₂O on CoP and Co₂P; (b) free energy diagram of the OER on CoP and Co₂P; (c) free energy diagram of the OER on CoP, CoP-PV1 and CoP-PV2; (d) Bader charge states of CoP, CoP-PV1, and CoP-PV2; DOS plots of (e) CoP, (f) CoP-PV1, and (g) CoP-PV2, wherein the Fermi level was set to zero.

It is known that the OER is a more challenging half reaction than the HER because of its sluggish kinetics and complex proton-coupled multielectron transfer process.⁴⁵ Therefore, the OER was employed as the model reaction for DFT calculations. The OER typically consists of four elementary steps under alkaline conditions,^46,47 and we calculated the Gibbs free energy of all intermediates (OH*, O* and OOH*) (Fig. S26 and S27†), yielding the theoretical overpotential of the OER over CoP and Co₂P (Fig. 5b), respectively. The rate-determining step (RDS) is the formation of OOH* through O* reacting with OH⁻ for both CoP and Co₂P, with a ΔG_RDS value of 3.36 and 4.43 eV, respectively. Hence, the OER will proceed more favourably on CoP due to its far lower energy barrier. Next, we constructed slab models for CoP with 1 (CoP-PV1) and 2 (CoP-PV2) PV to simulate different PV concentrations (Fig. S28†). The Gibbs free energy diagrams (Fig. 5c) show that CoP-PV2 shows optimal OER performances (ΔG_RDS = 2.36 eV) compared to CoP-PV1 (ΔG_RDS = 3.33 eV) and CoP (ΔG_RDS = 3.36 eV). This result suggests that the formation of more PV is beneficial for enhancing the catalytic activity, accounting for the experimental observations that CoP with more PV defects shows boosted catalytic activity.

The Bader charge was examined to understand the above phenomenon. Upon introduction of PV defects, the electron transfer from PV to nearby Co occurs, as evidenced by the reduced Bader charge of the Co sites (Fig. 5d). This result agrees well with the XPS results, where we can see an obvious negative shift in the Co 2p region when comparing MS-CoP_x-350 to CoP_x-350. More negative Co sites will promote the dissociation of the H–O bond in OH*, and then facilitate the following generation of O* and OOH*, thus leading to a visibly reduced energy barrier.⁴⁸ To further reveal the electronic states maintained by CoP, CoP-PV1 and CoP-PV2, partial density of states (PDOS) and total density of states (TDOS) were calculated (Fig. 5e–g). DOS plots show that the conducting charges primarily come from the Co 3d electronic states, as implied by the nearly matched curves of PDOS_Co with those of TDOS. The electronic states close to the Fermi level (set to zero) obey the following order: CoP-PV2 > CoP-PV1 > CoP, indicating that more charge carriers can be introduced via increasing the PV concentration⁴⁶ and coinciding with the smaller charge transfer resistance of MS-CoP_x-350 than that of CoP_x-350 (EIS results). The above discussions strongly support that the PV defects benefit the optimization of the electronic structure, and the enhancement of the electronic conductivity, thereby leading to an improvement in the electrochemical activity.

4. Conclusions

This study provides a one-step molten salt strategy to synthesize and simultaneously carry out PV defect engineering over CoP_x (CoP/Co₂P). In molten salt medium, inorganic Co salt was first converted into cobalt oxides, which then reacted with PH₃ generated by thermal decomposition of NaH₂PO₂ to produce CoP_x. Under alkaline conditions, the assembled two-electrode electrolyzer with the bifunctional electrocatalyst CoP_x only needed 1.75 and 1.90 V to drive current densities of 10 and 100 mA cm⁻² for overall water splitting, respectively. The performance decay was negligible after 24-hour continuous operation, indicating the excellent stability of CoP_x. DFT calculations suggested that increasing the PV concentration, induced by molten salt medium, delivered more negatively charged Co sites and more charge carriers, which were beneficial for reducing the energy barriers of reaction intermediates’ formation and enhancing the electronic conductivity, respectively. The successful preparation of NiP_x and Fe₂P by the molten salt strategy proves that this methodology can be further extended to the synthesis of other mono-, or even bi- and tri-metal TMPs. Moreover, the presented molten salt strategy offers an opportunity to gain more insights into TMPs’ defect engineering and electronic structure manipulation.

Author contributions

Z. L. and L. L. conceived the experiments and wrote the manuscript. Z. L., C. Z., C. W and B. Z. conducted the molten-salt-mediated synthesis and electrochemical characterization of the electrocatalyst. Y. Yang, S. P., Y. Yu and W. C. conducted the TEM, SEM, XRD, XPS, Raman and EPR spectra measurements. Z. L. did the DFT calculations.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (22176046), the State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (2021TS13), the Shenzhen Science and Technology Program (KQTD20190929172630447 and JCYJ20210324124209025), the Natural Science Foundation of Guangdong Province (2022A1515012016), and the China Postdoctoral Science Foundation (2021M700997).

References

1. R. d'Amore-Domenech, Ó. Santiago and T. J. Leo, Multicriteria analysis of seawater electrolysis technologies for green hydrogen production at sea, Renewable Sustainable Energy Rev., 2020, 133, 110166.
2. M. Yu, E. Budiyanto and H. Tüysüz, Principles of Water Electrolysis and Recent Progress in Cobalt-, Nickel-, and Iron-Based Oxides for the Oxygen Evolution Reaction, Angew. Chem., Int. Ed., 2022, 61, e202103824.
3. P. Yu, F. Wang, T. A. Shifa, X. Zhan, X. Lou, F. Xia and J. He, Earth abundant materials beyond transition metal dichalcogenides: A focus on electrocatalyzing hydrogen evolution reaction, Nano Energy, 2019, 58, 244–276.
4. Y. Shi and B. Zhang, Recent advances in transition metal phosphide nanomaterials: synthesis and applications in hydrogen evolution reaction, Chem. Soc. Rev., 2016, 45, 1529–1541.
5. Y. Shi, M. Li, Y. Yu and B. Zhang, Recent advances in nanostructured transition metal phosphides: synthesis and energy-related applications, Energy Environ. Sci., 2020, 13, 4564–4582.
6. J. Xu, J. Li, D. Xiong, B. Zhang, Y. Liu, K. H. Wu, I. Amorim, W. Li and L. Liu, Trends in activity for the oxygen evolution reaction on transition metal (M=Fe, Co, Ni) phosphide pre-catalysts, Chem. Sci., 2018, 9, 3470–3476.
7. J. Xu, Y. Liu, J. Li, I. Amorim, B. Zhang, D. Xiong, N. Zhang, S. M. Thalluri, J. P. S. Sousa and L. Liu, Hollow cobalt phosphide octahedral pre-catalysts with exceptionally high intrinsic catalytic activity for electro-oxidation of water and methanol, J. Mater. Chem. A, 2018, 6, 20646–20652.
8. J. Zhu, L. Hu, P. Zhao, L. Y. S. Lee and K.-Y. Wong, Recent Advances in Electrocatalytic Hydrogen Evolution Using Nanoparticles, Chem. Rev., 2020, 120, 851–918.
9. Y. Wu, W. Xu, L. Jiao, Y. Tang, Y. Chen, W. Gu and C. Zhu, Defect engineering in nanozymes, Mater. Today, 2021, 52, 327–347.
10. F. Cheng, T. Zhang, Y. Zhang, J. Du, X. Han and J. Chen, Enhancing Electrocatalytic Oxygen Reduction on MnO₂ with Vacancies, Angew. Chem., Int. Ed., 2013, 52, 2474–2477.
11. J. Kim, X. Yin, K.-C. Tsao, S. Fang and H. Yang, Ca₂Mn₂O₅ as Oxygen-Deficient Perovskite Electrocatalyst for Oxygen Evolution Reaction, J. Am. Chem. Soc., 2014, 136, 14646–14649.
12. H. Li, C. Tsai, A. L. Koh, L. Cai, A. W. Contryman, A. H. Fragapane, J. Zhao, H. S. Han, H. C. Manoharan, F. Abild-Pedersen, J. K. Nørskov and X. Zheng, Activating and optimizing MoS₂ basal planes for hydrogen evolution through the formation of strained sulphur vacancies, Nat. Mater., 2016, 15, 48–53.
13. Y. Yin, J. Han, Y. Zhang, X. Zhang, P. Xu, Q. Yuan, L. Samad, X. Wang, Y. Wang, Z. Zhang, P. Zhang, X. Cao, B. Song and S. Jin, Contributions of Phase, Sulfur Vacancies, and Edges to the Hydrogen Evolution Reaction Catalytic Activity of Porous Molybdenum Disulfide Nanosheets, J. Am. Chem. Soc., 2016, 138, 7965–7972.
14. Y. Jia, L. Zhang, A. Du, G. Gao, J. Chen, X. Yan, C. L. Brown and X. Yao, Defect Graphene as a Trifunctional Catalyst for Electrochemical Reactions, Adv. Mater., 2016, 28, 9532–9538.
15. C. Xie, D. Yan, W. Chen, Y. Zou, R. Chen, S. Zang, Y. Wang, X. Yao and S. Wang, Insight into the design of defect electrocatalysts: From electronic structure to adsorption energy, Mater. Today, 2019, 31, 47–68.
16. R. Sun, Y. Bai, Z. Bai, L. Peng, M. Luo, M. Qu, Y. Gao, Z. Wang, W. Sun and K. Sun, Phosphorus Vacancies as Effective Polysulfide Promoter for High-Energy-Density Lithium–Sulfur Batteries, Adv. Energy Mater., 2022, 12, 2102739.
17. S. Li, Z. Geng, X. Wang, X. Ren, J. Liu, X. Hou, Y. Sun, W. Zhang, K. Huang and S. Feng, Optimizing the surface state of cobalt-iron bimetallic phosphide via regulating phosphorus vacancies, Chem. Commun., 2020, 56, 2602–2605.
18. M. Jin, X. Zhang, M. Han, H. Wang, G. Wang and H. Zhang, Efficient electrochemical N₂ fixation by doped-oxygen-induced phosphorus vacancy defects on copper phosphide nanosheets, J. Mater. Chem. A, 2020, 8, 5936–5942.
19. V. M. B. Nunes, C. S. Queirós, M. J. V. Lourenço, F. J. V. Santos and C. A. Nieto de Castro, Molten salts as engineering fluids – A review: Part I. Molten alkali nitrates, Appl. Energy, 2016, 183, 603–611.
20. M. Xiao, L. Zhang, B. Luo, M. Lyu, Z. Wang, H. Huang, S. Wang, A. Du and L. Wang, Molten-Salt-Mediated Synthesis of an Atomic Nickel Co-catalyst on TiO₂ for Improved Photocatalytic H2 Evolution, Angew. Chem., Int. Ed., 2020, 59, 7230–7234.
21. M. Xiao, B. Luo, M. Konarova, Z. Wang and L. Wang, Molten Salt Synthesis of Atomic Heterogeneous Catalysts: Old Chemistry for Advanced Materials, Eur. J. Inorg. Chem., 2020, 2020, 2942–2949.
22. X. Liu, N. Fechler and M. Antonietti, Salt melt synthesis of ceramics, semiconductors and carbon nanostructures, Chem. Soc. Rev., 2013, 42, 8237–8265.
23. Y. Bai, C. Liu, T. Chen, W. Li, S. Zheng, Y. Pi, Y. Luo and H. Pang, MXene-Copper/Cobalt Hybrids via Lewis Acidic Molten Salts Etching for High Performance Symmetric Supercapacitors, Angew. Chem., Int. Ed., 2021, 60, 25318–25322.
24. H. Cui, M. Jiao, Y.-N. Chen, Y. Guo, L. Yang, Z. Xie, Z. Zhou and S. Guo, Molten-Salt-Assisted Synthesis of 3D Holey N-Doped Graphene as Bifunctional Electrocatalysts for Rechargeable Zn–Air Batteries, Small Methods, 2018, 2, 1800144.
25. B. Zhou, J. Li, X. Zhang and J. Guo, Engineering P-doped Ni₃S₂-NiS hybrid nanorod arrays for efficient overall water electrolysis, J. Alloys Compd., 2021, 862, 158391.
26. L. Zhang, H. Li, B. Yang, N. Han, Y. Wang, Z. Zhang, Y. Zhou, D. Chen and Y. Gao, Promote the electrocatalysis activity of amorphous FeOOH to oxygen evolution reaction by coupling with ZnO nanorod array, J. Solid State Electrochem., 2020, 24, 905–914.
27. N. Li, W. Liu, C. Zhu, C. Hao, J. Guo, H. Jing, J. Hu, C. Xin, D. Wu and Y. Shi, Molten salt as ultrastrong polar solvent enables the most straightforward pyrolysis towards highly efficient and stable single-atom electrocatalyst, J. Energy Chem., 2021, 54, 519–527.
28. J. Wang, Z. Liu, Y. Zheng, L. Cui, W. Yang and J. Liu, Recent advances in cobalt phosphide based materials for energy-related applications, J. Mater. Chem. A, 2017, 5, 22913–22932.
29. L. Su, X. Cui, T. He, L. Zeng, H. Tian, Y. Song, K. Qi and B. Y. Xia, Surface reconstruction of cobalt phosphide nanosheets by electrochemical activation for enhanced hydrogen evolution in alkaline solution, Chem. Sci., 2019, 10, 2019–2024.
30. J. Tyczkowski, R. Kapica and J. Łojewska, Thin cobalt oxide films for catalysis deposited by plasma-enhanced metal–organic chemical vapor deposition, Thin Solid Films, 2007, 515, 6590–6595.
31. V. G. Hadjiev, M. N. Iliev and I. V. Vergilov, The Raman spectra of Co3O4, J. Phys. C: Solid State Phys., 1988, 21, L199–L201.
32. F. H. Saadi, A. I. Carim, W. S. Drisdell, S. Gul, J. H. Baricuatro, J. Yano, M. P. Soriaga and N. S. Lewis, Operando Spectroscopic Analysis of CoP Films Electrocatalyzing the Hydrogen-Evolution Reaction, J. Am. Chem. Soc., 2017, 139, 12927–12930.
33. G. Yuan, J. Bai, L. Zhang, X. Chen and L. Ren, The effect of P vacancies on the activity of cobalt phosphide nanorods as oxygen evolution electrocatalyst in alkali, Appl. Catal., B, 2021, 284, 119693.
34. S. Nadupalli, S. Repp, S. Weber and E. Erdem, About defect phenomena in ZnO nanocrystals, Nanoscale, 2021, 13, 9160–9171.
35. S. Najib, F. Bakan, N. Abdullayeva, R. Bahariqushchi, S. Kasap, G. Franzo, M. Sankir, N. Demirci Sankir, S. Mirabella and E. Erdem, Tailoring morphology to control defect structures in ZnO electrodes for high-performance supercapacitor devices, Nanoscale, 2020, 12, 16162–16172.
36. J. K. Olson and A. I. Boldyrev, Electronic transmutation: Boron acquiring an extra electron becomes ‘carbon’, Chem. Phys. Lett., 2012, 523, 83–86.
37. A. S. Ivanov, A. J. Morris, K. V. Bozhenko, C. J. Pickard and A. I. Boldyrev, Inorganic Double-Helix Structures of Unusually Simple Lithium–Phosphorus Species, Angew. Chem., Int. Ed., 2012, 51, 8330–8333.
38. L. Yang, R. Liu and L. Jiao, Electronic Redistribution: Construction and Modulation of Interface Engineering on CoP for Enhancing Overall Water Splitting, Adv. Funct. Mater., 2020, 30, 1909618.
39. X. Huang, X. Xu, C. Li, D. Wu, D. Cheng and D. Cao, Vertical CoP Nanoarray Wrapped by N,P-Doped Carbon for Hydrogen Evolution Reaction in Both Acidic and Alkaline Conditions, Adv. Energy Mater., 2019, 9, 1803970.
40. Y. Ma, G. Zhou, Z. Liu, L. Xu, D. Sun and Y. Tang, Electronic structural regulation of CoP nanorods by the tunable incorporation of oxygen for enhanced electrocatalytic activity during the hydrogen evolution reaction, Nanoscale, 2020, 12, 14733–14738.
41. X. Xiao, C. Engelbrekt, Z. Li and P. Si, Hydrogen evolution at nanoporous gold/tungsten sulfide composite film and its optimization, Electrochim. Acta, 2015, 173, 393–398.
42. M. A. Lukowski, A. S. Daniel, F. Meng, A. Forticaux, L. Li and S. Jin, Enhanced Hydrogen Evolution Catalysis from Chemically Exfoliated Metallic MoS₂ Nanosheets, J. Am. Chem. Soc., 2013, 135, 10274–10277.
43. Y. Cao, H. Liu, X. Bo and F. Wang, Highly active porous nickel-film electrode via polystyrene microsphere template-assisted composite electrodeposition for hydrogen-evolution reaction in alkaline medium, Sci. China: Chem., 2015, 58, 501–507.
44. Y. Huang, Y. Sun, X. Zheng, T. Aoki, B. Pattengale, J. Huang, X. He, W. Bian, S. Younan, N. Williams, J. Hu, J. Ge, N. Pu, X. Yan, X. Pan, L. Zhang, Y. Wei and J. Gu, Atomically engineering activation sites onto metallic 1T-MoS₂ catalysts for enhanced electrochemical hydrogen evolution, Nat. Commun., 2019, 10, 982.
45. K. Wang, X. Wang, Z. Li, B. Yang, M. Ling, X. Gao, J. Lu, Q. Shi, L. Lei, G. Wu and Y. Hou, Designing 3d dual transition metal electrocatalysts for oxygen evolution reaction in alkaline electrolyte: Beyond oxides, Nano Energy, 2020, 77, 105162.
46. H. Xue, A. Meng, H. Zhang, Y. Lin, Z. Li and C. Wang, 3D urchin like V-doped CoP in situ grown on nickel foam as bifunctional electrocatalyst for efficient overall water-splitting, Nano Res., 2021, 14, 4173–4181.
47. J. Jiang, F. Sun, S. Zhou, W. Hu, H. Zhang, J. Dong, Z. Jiang, J. Zhao, J. Li, W. Yan and M. Wang, Atomic-level insight into super-efficient electrocatalytic oxygen evolution on iron and vanadium co-doped nickel (oxy)hydroxide, Nat. Commun., 2018, 9, 2885.
48. C. Guan, W. Xiao, H. Wu, X. Liu, W. Zang, H. Zhang, J. Ding, Y. P. Feng, S. J. Pennycook and J. Wang, Hollow Mo-doped CoP nanoarrays for efficient overall water splitting, Nano Energy, 2018, 48, 73–80.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details (spectroscopy and structural characterization, electrode preparation, electrochemical measurements, and DFT calculations), XPS, SEM-EDS, TEM, EPR, electrochemical results, and catalytic activity table (comparison with other available CoP electrocatalysts). See DOI: https://doi.org/10.1039/d2qi01902g

---