High-entropy nitrides from dual entropic and enthalpic forces for high-efficiency oxygen evolution reaction

Materials

Polyethylene glycol (PEG) 200, Co(NO₃)₂·6H₂O (≥ 99%), barbituric acid (BA), potassium hydroxide (KOH, ≥ 90%), and ethanol were sourced from Sinopharm Chemical Reagent Co., Ltd. Fe(NO₃)₂·9H₂O (99.99%), Zn(NO₃)₂·6H₂O (99.99%), and Ni(NO₃)₂·6H₂O (99.99%) were acquired from Shanghai Aladdin Biochemical Technology Co., Ltd., while Cu(NO₃)₂·3H₂O (99%) and ruthenium oxide (RuO₂, 99.9%) were obtained from Shanghai Macklin Biochemical Co., Ltd. All chemicals were used without further purification, and ultrapure water with a resistivity of 18.2 MΩ·cm was used in all experiments.

Preparation of (FeCoNiCuZn)N and the related single metal nitrides

(FeCoNiCuZn)N was obtained via a one-step eutectic-solvothermal process. Typically, 2 mmol each of Fe(NO₃)₃·6H₂O, Co(NO₃)₂·6H₂O, Ni(NO₃)₂·6H₂O, Cu(NO₃)₂·6H₂O, and Zn(NO₃)₂·6H₂O were dissolved in 30 mL PBDESs. The formed dark-brown solution was then transferred into a 100 mL Teflon-lined autoclave and heated at 200 °C for 16 h. After cooling, the samples were collected, washed and dried at 60 °C for 2 h. In addition, unary metal nitrides (UMNs) were synthesized using the same method as for (FeCoNiCuZn)N, dissolving 10 mmol of each corresponding nitrate in PBDESs.

Materials characterization

Thermal analysis was conducted using a METTLER TOLEDO TGADSC3+ (Germany) system, with a heating rate of 3 °C min^-1 from 30 to 450 °C under an Argon atmosphere. Fourier transform infrared spectroscopy (FT-IR) was performed on a Shimadzu Prestige 21 (Japan, DTGS detector) using the KBr pellet technique, with a measurement range of 750 to 4,000 cm^-1. Raman spectra were obtained using a LabRAM HR Evolution with a 638 nm laser.

Scanning electron microscopy (SEM) was carried out using a Hitachi SU8010 field emission SEM (Japan), while transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) were performed using an FEI Tecnai G2 F20 field emission microscope. Dark-field scanning TEM (STEM) was performed on FEI Themis Z, Titan Cubed Themis G2300, JEM-ARM200F.

X-ray powder diffraction (XRD) was done with a Y-2000 diffractometer (λ = 1.5406 Å) using Cu Kα radiation at a scanning rate of 5° min^-1. X-ray photoelectron spectroscopy (XPS) was conducted using a Thermo Fisher Scientific ESCA Lab250 (USA) with an Al Kα source. All XPS data were calibrated using the C 1s line at 284.6 eV and fitted using the Gaussian fitting method.

X-ray absorption near-edge structure (XANES) and X-ray absorption fine structure (XAFS) spectra were collected at the TPS44A1 beamline of the National Synchrotron Radiation Research Center (NSRRC), operating at 3.5 GeV with a constant current of 260 mA. The energy calibration was performed using Co foil. Data extraction and line fitting were conducted using the Athena and Artemis software. For XANES, the absorption coefficients μ(E) were processed through background subtraction and normalization, reported as “normalized absorption” with E₀ for all samples. XAFS Fourier-transformed (FT) data in R space were analyzed using the Co foil model for Co-Co contributions. The passive electron factors (S₀) were determined by fitting the Co foil data with a fixed Co-Co coordination number (CN) and then applied to the analysis of the measured samples.

Electrochemical performances

All electrochemical experiments were performed using a CHI 760E electrochemical station (ChenHua Instrument, Shanghai, China) in a three-electrode configuration. The OER performance was evaluated in 1 M KOH, with (FeCoNiCuZn)N-modified glassy carbon and Ni foam as the working electrodes, a graphite rod as the counter electrode, and an Ag/AgCl electrode (in saturated KCl) as the reference. For comparison, UMN-modified glassy carbon electrodes were also tested. All potentials were calibrated against the reversible hydrogen electrode (RHE) using E_(RHE) = E_(Ag/AgCl) + 0.197 V + 0.059 × pH, and polarization curves were corrected for 80% iR drop.

Linear sweep voltammetry (LSV) was performed from 1.0 to 1.7 V vs. RHE at a scan rate of 5 mV s^-1. Tafel slopes were derived from the LSV curves. Cyclic voltammetry (CV) was conducted at scan rates ranging from 1 to 200 mV s^-1 within a potential window of -0.1 to 0.1 V (vs. Ag/AgCl), where only non-faradaic currents were observed. Electrochemical impedance spectroscopy (EIS) was measured in potentiostatic mode over a frequency range of 0.1 to 100 kHz. Electrochemical stability tests of the samples were performed by utilizing multi-potential steps, multi-current steps and amperometry i-t experiments.

Fabrication of HENs (FeCoNiCuZn)N

DES-aided strategy for HEN synthesis at low-temperature can be schematically generalized with an associated Gibbs free energy (ΔG) [Figure 1A]. The decomposing ΔG^o involves an enthalpy contribution (ΔH) and an entropy contribution (TΔS), and it must be negative under the experimental conditions. In detail, the differential scanning calorimetry (DSC) curves [Figure 1B] of PBDESs with or without metal salts reveal their related endothermic peaks. For PBDESs, an endothermic peak at 473 K is observed, indicating PBDESs intensively decomposed at 473 K, which is in accordance with the thermogravimetric analysis (TGA) results [Supplementary Figure 1]. For comparison, the thermal destruction of salt-in-PBDESs becomes less intensive, as proved by the decreased decomposition enthalpy and reduced decomposed temperature. The normalized decomposition enthalpies of PBDESs with or without five metal salts are -332.99 and -201.06 J g^-1, respectively. In other words, in the presence of metal salts, the endo effect of thermo-destruction occurs at lower temperatures and the enthalpy value decreases by 1.5-fold. In addition, the minimum endothermic peak for the decomposition shifts from 473 to 458 K [Figure 1B, Supplementary Figure 1]. The lower temperature and enthalpy peak of the pyrolysis process indicate that the salts-in-PBDES system has a less perfectly ordered structure, which is caused by the interaction between metal ions and PBDESs. Apart from the enthalpy contribution, a substantial positive entropy change arises from the increased degree of disorder. Under ideal mixing conditions, the entropy could be calculated by $$\Delta S_{m i x}=-R \displaystyle{\sum_{i=1}^{n}} X_{i} \ln X_{i}$$, where X_i represents the molar fraction of each element in the system and R is the gas constant. The configurational entropy of five salts-in-PBDES and PBDES is -13.381 and 0 J K^-1 mol^-1, respectively [Figure 1C]. During the pyrolysis of PBDES in the presence of metal salts, the reaction proceeds through multiple thermal stages, including distinct endothermic and exothermic events. These thermal transitions play a crucial role in forming HENs at a record low temperature of 473 K. The decomposition of PBDES, coupled with the controlled release of gases, enhances reaction kinetics, facilitating the formation of the nitrides at this low temperature. The efficient assembly interaction between metal salts and PBDESs was realized in the homogeneous salts-in-PBDES system, which can be proved by the Fourier transform-infrared spectroscopy (FT-IR) [Supplementary Figure 2, Figure 1D] and Raman results [Figure 1E]. As depicted in Supplementary Figure 2, the -OH peak derived from PEG 200 shifts from 3,406 to 3,386 cm^-1, and the related half peak width is increased, suggesting the formation of a supramolecular network via H bonding between the -OH group and the C=O group. After adding metal salts, the asymmetric stretching peak of -C=O in Figure 1D shifts from 1,720 cm^-1 to approximately 1,738 cm^-1, which is caused by the complexation between metal ions and BA^[27]. Additionally, the complexation is further proved by Raman result. As Figure 1E depicts, a new band at 1,038 cm^-1 is attributed to the M-N complex^[28,29]. The complex interaction between metal ions and BA as well as the BA-PEG hydrogen bonding interaction are thermally in favor of the pyrolysis process. Furthermore, the generation of high-temperature conditions is caused by the combustion of volatile and light species accompanied by the cracking, dehydrogenation, and dealkylation of PBDES, providing sufficient thermal energy for multielement metal nitride formation^[30,31]. Therefore, the synthesis of PBDES-assisted HMNs begins with the pyrolysis of PBDES in the presence of metal salts. This process involves a series of endothermic and exothermic steps, ultimately resulting in the formation of the target (FeCoNiCuZn)N at a relatively low temperature of 473 K.

Figure 1. Schematic representation of the formation of (FeCoNiCuZn)N via a low-temperature solvothermal process. (A) Schematic mechanism of (FeCoNiCuZn)N; (B) DSC results and (C) the calculated entropy values, (D) Magnified FT-IR spectrum, and (E) Raman spectrum of salts-in-PBDES and PBDES systems. PBDES: Polyethylene glycol-barbituric acid deep eutectic solvent.

Structural and compositional characterization

To elucidate the crystallographic properties of the prepared HENs, powder X-ray diffraction (PXRD) analysis was conducted [Figure 2A]. The obtained data closely match the hexagonal FeN_0.0589 phase (JCPDS 01-075-2130), with a space group of P63/mmc (No. 194), confirming the successful formation of single-phase nitrides. Despite the low concentration of nitrogen atoms integrated into the metal lattice in this structure, the material is designated as (FeCoNiCuZn)N to emphasize its composition and highlight its high-entropy characteristics. To further investigate the structure of (FeCoNiCuZn)N, SEM and TEM technologies were utilized. Figure 2B and C shows that the average size of the HEN nanoparticle is approximately 100 nm. This uniform size is primarily due to the templating effect of PEG 200, a common surfactant, which not only regulates nanoparticle size but also reduces agglomeration and enhances the stability of the synthesized nanoparticles. HRTEM images [Figure 2D] depict clear lattice fringes, indicative of the high crystallinity of (FeCoNiCuZn)N. The observed interplanar spacing of 0.21 nm [Inset of Figure 2D] corresponds well with the d(111) plane of FeN_0.0589, which exhibits a hexagonal crystal structure and is well consistent with the XRD results. This is further corroborated by the fast Fourier transformation (FFT) image [Figure 2E] and inverse FFT (IFFT) pattern [Figure 2F]. Additionally, the uniform distribution of the five metal species (Fe, Co, Ni, Cu, and Zn) and the non-metal element (N) was visualized by energy-dispersive X-ray (EDX) mapping [Figure 2G] and inductively coupled plasma atomic emission spectrometry (ICP-AES) results [Figure 2H, Supplementary Table 1].

Figure 2. Structural characterization of (FeCoNiCuZn)N. (A) PXRD pattern, (B) SEM, (C) TEM and (D) HRTEM images of (FeCoNiCuZn)N; Inset (D) was the magnified area in blue region. (E) FFT and (F) IFFT pattern of the red region derived from (D); (G) EDS elemental mapping of (FeCoNiCuZn)N: Fe (yellow), Co (green), Ni (cyan), Cu (blue), Zn (purple) and N (red); (H) Mole fraction of (FeCoNiCuZn)N obtained from the ICP-AES result.

The elemental composition and chemical states of (FeCoNiCuZn)N were analyzed using XPS. The full survey spectrum [Supplementary Figure 3] proves the presence of Fe, Co, Ni, Cu, Zn, N, C, and O in the synthesized material. The high-resolution Fe 2p spectrum [Figure 3A] reveals two spin-orbit doublets and four shake-up satellites^[32]. The first doublet at 712.4 and 725.5 eV corresponds to Fe³⁺, while the second at 710.3 and 723.4 eV is attributed to Fe²⁺. In the high-resolution Co 2p spectrum [Figure 3B], four subpeaks corresponding to Co²⁺ and Co³⁺ are deconvoluted, with peaks at 780.2 and 795.6 eV for Co³⁺, and 781.3 and 798.0 eV for Co^2+[27]. Similarly, the co-existence of Ni²⁺ and Ni³⁺ in (FeCoNiCuZn)N is also confirmed by the high-resolution Ni 2p spectrum [Figure 3C], where the peaks located at 854.6 and 872.0 eV are related to Ni²⁺, and those at 856.4 and 873.9 eV corresponded to Ni^3+[33]. The Cu 2p spectrum [Figure 3D] shows Cu 2p₃/₂ (934.2 eV), Cu 2p_1/2 (953.9 eV), and a characteristic shake-up peak (942-944 eV) indicative of Cu^2+[34]. The Zn 2p spectrum [Figure 3E] displays two spin-orbit peaks at 1,045.7 and 1,022.5 eV, corresponding to Zn²⁺ 2p_3/2 and 2p_1/2, respectively^[35]. In the N 1s spectrum [Figure 3F], four distinct peaks appear at binding energies of 398.5, 399.3, 400.2 and 401.4 eV. These peaks correspond to pyridinic nitrogen, nitrogen-metal bonds, pyrrolic nitrogen, and graphitic nitrogen, respectively^[36]. The valence states of nitrogen in pyridinic, nitrogen-metal, and pyrrolic forms are -3, while graphitic nitrogen exhibits a valence state of 0 due to differences in bonding environments, ensuring the overall charge balance within the systems.

Figure 3. XPS spectra of (FeCoNiCuZn)N: (A) Fe 2p, (B) Co 2p, (C) Ni 2p, (D) Cu 2p, (E) Zn 2p, and (F) N 1s.

Evaluation of electrocatalytic performance

To identify the structural advantages of the obtained samples, the electrochemical behaviors of (FeCoNiCuZn)N and related UMNs including (Fe)N, (Co)N, (Ni)N, (Cu)N, and (Zn)N were compared and evaluated utilizing a classical three-electrode system in 1 M KOH. LSV curves of all samples were listed in Figure 4A and Supplementary Figure 4. The OER activity of (FeCoNiCuZn)N outperforms all the related UMNs, performing an excellent overpotential of 223 mV (at 10 mA cm^-2) and 328 mV (at 100 mA cm^-2). This performance is also superior to those of previously reported transition metal-based materials [Supplementary Table 2]. Furthermore, the kinetic processes of those catalysts were also evaluated by the Tafel slopes derived from the LSV curves [Figure 4B]. The Tafel slopes of those samples follow this order: (FeCoNiCuZn)N (57 mV dec^-1) < (Co)N (68 mV dec^-1) < (Ni)N (71 mV dec^-1) < (Fe)N (90 mV dec^-1) < (Cu)N (91 mV dec^-1) < (Zn)N (92 mV dec^-1). To investigate the OER kinetics, EIS tests of (FeCoNiCuZn)N and (Co)N were performed and the Nyquist plots were shown in Figure 4C. Apparently, (FeCoNiCuZn)N exhibits a relatively low charge transfer resistance (R_ct) value, which is much lower than that of (Co)N. Moreover, the electrochemically active surface area (ECSA) was calculated [Figure 4D and Supplementary Figure 5]. The double-layer capacitances (C_dl) of (FeCoNiCuZn)N and (Co)N are determined to be 0.0022 and 0.0013 mF cm^-2, respectively, indicating that (FeCoNiCuZn)N has relatively high ECSA, almost 2-fold than that of (Co)N. ECSA-normalized LSV curves show that (FeCoNiCuZn)N also exhibits the best OER activity [Supplementary Figure 6], indicating that both the enhanced intrinsic activity and the increased number of active sites enhance the OER performance.

Figure 4. Electrochemical performance of the obtained samples. (A) Geometrical area normalized LSV curves, (B) Tafel plots, (C) EIS, and (D) EASA results of (FeCoNiCuZn)N (red) and (Co)N (blue) modified glass carbon electrodes; (E) Multi-potential steps (left, orange) and multi-current steps (right, blue) of (FeCoNiCuZn)N; (F) Chronopotentiometry curve of (FeCoNiCuZn)N modified Ni foam at current density located at 400 mA cm^-2 for 30 h.

Apart from the excellent activity, the electrochemical stability of (FeCoNiCuZn)N was evaluated by multi-potential and multi-current step tests (rapid response and short-term stability), as well as the chronopotentiometry curve (long-term stability). It is found that the cycling stability of (FeCoNiCuZn)N was studied at various potentials [Figure 4E, left] and current densities [Figure 4E, right], in which the selected current densities and potentials were well retained, respectively. In addition, the potential change was almost negligible when the electrode was measured at a constant current density of 400 mA cm^-2 for 30 h [Figure 4F], verifying their preliminary yet substantial assessment of potential practical applications.

Mechanistic investigations

To investigate the mechanism behind the enhanced electrochemical performance, XPS, XANES, and extended EXAFS measurements were conducted. As illustrated in Supplementary Figure 7 and Figure 2B, unary (Co)N only shows Co²⁺, while Co²⁺ and Co³⁺ co-exist in the (FeCoNiCuZn)N samples. Furthermore, the Co 2p spectrum exhibits positive shifts with an increasing number of metal elements. Meanwhile, the Fe, Ni, Cu, and Zn peaks in (FeCoNiCuZn)N shift to the higher binding energy compared with those of the related UMNs [Supplementary Figures 8-11]. In other words, accompanied by metal addition, the metal ions in (FeCoNiCuZn)N exhibit higher oxidation states, indicating the change in the local coordination situation^[11].

The electronic structure and coordination environment of (FeCoNiCuZn)N were further confirmed through XANES and EXAFS experiments. In the Co K-edge XANES [Figure 5A], the Co K-edge position of (Co)N closely matches that of CoO, indicating a +2 oxidation state. However, for (FeCoNiCuZn)N, the Co K-edge lies between CoO and Co₃O₄. A slightly enhanced and right-shifted white line peak in (FeCoNiCuZn)N compared to (Co)N suggests an increased oxidation state of Co^[37,38]. Figure 5B presents a typical EXAFS fitting of (FeCoNiCuZn)N, which renders a bond distance of 2.10 Å and a CN of 7.55. The local coordination geometry of the two samples is further confirmed by their FT R-space spectra [Figure 5C], where both (Co)N and (FeCoNiCuZn)N display a prominent Co-N bond scattering feature that is inconspicuous as that of the Co foil. FT curve fitting [Supplementary Table 3] reveals that the Co-N coordination number decreases from 6.4 in (Co)N to 5.6 in (FeCoNiCuZn)N, while the bond length increases from 2.09 to 2.13 Å, indicating the existence of dangling bonds and structural distortion. It was also observed that the first shell peaks in the EXAFS spectrum of R space of all Fe, Co, Ni, Cu, and Zn elements exhibit only a slight mismatch [Figure 5D], indicating that the (FeCoNiCuZn)N is not completely in random and some short-range chemical ordering exists^[39,40]. This structural distortion is further supported by wavelet transform (WT) analysis [Figure 5E], which highlights differences in the local atomic arrangements of cobalt. Nitrogen functions as a structural and electronic modifier within the HEN framework, improving both chemical stability and electrical conductivity, which are crucial for enhancing catalytic efficiency in the OER^[18,27]. Obviously, the severe structural distortion is positive for the high oxidation state of Co, which is in favor of the further electrocatalysis process^[41-44].

Figure 5. X-ray absorption spectroscopy of samples. (A) Co K-edge XANES spectra of Co foil, CoO, Co₃O₄, (Co)N, and (FeCoNiCuZn)N; (B) EXAFS fitting of (FeCoNiCuZn)N; (C) Fourier transform (FT) of the Co K-edge EXAFS spectra of Co foil, (Co)N, and (FeCoNiCuZn)N; (D) The corresponding EXAFS fitting curve of Co-(FeCoNiCuZn)N: Fe (yellow), Co (grey), Ni (green), Cu (pink), and Zn (blue); (E) the WT contour plot of Co foil, (Co)N, and (FeCoNiCuZn)N.

The structure and composition of (FeCoNiCuZn)N after long-term stability were also evaluated by SEM, TEM, and XPS tests. As Supplementary Figure 12 shows, the spherical morphology of (FeCoNiCuZn)N is retained after the OER test. Additionally, the energy-dispersive X-ray spectroscopy (EDS) analysis further confirms the uniform distributions of all the metals [Supplementary Figure 13]. The survey and related high-resolution XPS spectrum of the (FeCoNiCuZn)N after OER were shown in Supplementary Figure 14. It is apparent that the Co³⁺ species increase after OER reaction [Supplementary Figure 14A], indicating that the oxidation of cobalt takes place and Co serves as an electron donor. In addition, the oxidation state of iron and nickel elements shows negligible variation in composition [Supplementary Figure 14B and C], implying Fe and Ni may play a role in structure stability. Furthermore, both Cu 2p and Zn 2p peaks of the (FeCoNiCuZn)N after OER shift to lower binding energy [Supplementary Figure 14D and E], which implies that Cu and Zn would act as electron acceptors to Co in the electrocatalysis process. Given the above analysis, we can clearly see that the synergy effect among the metals in (FeCoNiCuZn)N is conducive to boosting OER performance. Moreover, the metal-N content shows no significant variation [Supplementary Figure 14F], indicating the preservation of the original high-entropy metal nitride structure, suggesting that the catalytic performance is primarily affected by the engineered composition and nanostructure of the HEN. The following key aspects of high-entropy engineering contribute to the superior OER performance of (FeCoNiCuZn)N. First, (FeCoNiCuZn)N demonstrates enhanced structural stability due to lattice distortion caused by the incorporation of multiple elements. This structural reinforcement improves stability during electrochemical reactions. Second, synergistic effects arise from the diverse metal atoms in HENs, introducing a variety of catalytic sites and coordination environments, which enhance OER catalytic performance. Finally, high-entropy engineering modifies the electronic structure of the material, optimizing charge transfer, increasing the density of active sites, and lowering reaction overpotentials, all of which are crucial for efficient OER performance. Therefore, these results above demonstrate the synergistic effect of electron modulation and entropy stabilization on the (FeCoNiCuZn)N, which are closely correlated with the superior OER of (FeCoNiCuZn)N.