Introduction

High-entropy oxides (HEOs) are regarded as a solid-solution phase, which can be classified into spinel, rock salt, fluorite, and perovskite structures1,2,3,4,5. They have emerged as a significant focus of research due to their disordered arrangement, interactions among multiple principal elements, and the promotion of entropy6,7,8,9. Moreover, the configurational entropy of high-entropy oxides significantly exceeds that of individual monometallic oxides, suggesting that high-entropy oxides exhibit greater thermodynamic stability than their monometallic counterparts at finite temperatures10,11,12,13. The previously reported synthesis methods for HEOs are characterized by prolonged heating durations, elevated temperatures, and substantial energy consumption. For instance, they involve the mixing of metal salts followed by heating to temperatures of 900 °C or higher to achieve sintering into a single phase14,15. However, the high temperatures associated with this process can lead to irregular molten bulk formation, limited adjustability of components, and a propensity for phase separation, which may hinder the accessibility of active sites and impair the catalytic reaction process16,17. Consequently, there is an urgent need for a synthesis strategy that is both rapid and controllable, aimed at minimizing heating duration and temperature while effectively managing morphological characteristics18,19.

The utilization of metal hydroxides as precursors constitutes a common approach for the synthesis of metallic oxides. It has been reported that monocomponent, binary, or even ternary metal hydroxides with various components and morphologies can be fabricated through the solution method. Subsequently, the corresponding metal oxides are obtained via low-temperature heat treatment while retaining the original morphologies of the metal hydroxides20,21,22,23,24. Nevertheless, the incorporation of five or more metallic elements into metal hydroxide precursors poses a significant challenge due to the complex co-precipitation kinetics arising from the disparate physicochemical properties of different metal cations and standard solubility product constants (Ksp) of metal hydroxides25. Furthermore, in a multiple metal hydroxide precursor system, the precipitation sequence of metal species cannot be solely evaluated by Ksp. It may also be influenced by factors such as coordination environments, surface energies, organic additives, and the selection of precipitating agents. These variables complicate the control of nucleation and crystal growth processes, which can further affect the morphology of the resulting products20,26. Consequently, it is crucial to regulate the co-precipitation process and the morphology of metal hydroxide precursors to achieve the formation of single-phase high-entropy oxides.

Morphology is a critical property of catalysts, influencing the mass transfer rate as well as the adsorption and desorption of intermediates and products, thereby exerting a significant impact on catalytic performance27. Morphological engineering, particularly with respect to regulating dimensionality and size, has been demonstrated as an effective strategy for enhancing physicochemical properties and introducing novel functionalities in catalysts28. In particular, three-dimensional (3D) hollow architectures that expose high-density active sites and provide additional channels for ionic transport and solution diffusion can shorten the charge transfer pathway and enhance the interaction between reactants and catalysts, thereby significantly improving catalytic efficiency29,30. However, the construction of three-dimensional hollow architectures in HEOs is seldom reported due to the unsuitability of synthesis methods. In general, the template removal methods or kinetically controlled self-assembly strategies were considered typical methods for synthesizing the three-dimensional hollow nanomaterials31,32,33,34. However, these approaches demand rigorous precision in precursor stoichiometry, reaction kinetics, and thermal processing parameters. Conventional multi-step synthetic routes are prone to impurity incorporation or structural integrity compromise during template removal and crystallization stages35,36,37. Furthermore, ‌another critical challenge lies in maintaining high-entropy characteristics throughout the outer shell of 3D hollow structures. Achieving simultaneous control over homogeneous elemental distribution across multiple metallic components in HEOs architectures typically necessitates. Multi-metallic systems exhibit thermodynamic tendencies toward phase segregation during high-temperature processing. For instance, single-phase spinel HEOs require thermal treatment at ~900 °C for phase stabilization, yet such extreme conditions inevitably induce hollow architecture collapse or pore channel closure through sintering effects. This inherent contradiction between phase stability and structural preservation fundamentally limits the achievable compositional complexity while maintaining desired morphological features. Therefore, low-temperature and facile methods to synthesize high-entropy spinel oxides with uniform compositions featuring hollow structures are vital items for catalytic reactions.

In this study, we present a versatile synthetic method for fabricating hollow nanocube high-entropy spinel oxides composed of up to eight metallic elements, by a template-assisted route inspired by coordinating etching. The essence of this distinctive strategy hinges on (1) the preparation of the Cu2O nanocube template, (2) the release of hydroxide ions (OH−) through coordinating etching reactions between a soft base and a soft acid, (3) the co-precipitation of high-entropy metal hydroxides (HE-OH) onto the nanocube shell, and (4) the thermal treatment of HE-OH precursors. This method is achieved by meticulously controlling the balance between the precipitation rate of metal hydroxides and the synchronous coordinating etching rate towards the soft acid. As a demonstration of the catalytic concept, the quinary NiCoFeCdCr-O exhibits a high rate constant (k) of 1.79 min−1 for the hydrogenation of p-nitrophenol. Furthermore, NiCoFeCdCr-O demonstrates stability over 10 cycles, with conversion rates maintained above 95%. Density functional theory (DFT) simulations reveal that the good performance can be ascribed to a more continuous density of states (DOS) near the Fermi level and a more favorable d-band center in HEOs. Our systematic investigation provides valuable insights for the future design of multi-component catalysts with tailored morphologies.

Results

Preliminary investigations

The template-assisted route inspired by the coordinating etching strategy can be utilized to synthesize metal high-entropy hydroxide precursors, achieved by carefully balancing the precipitation rate of metal hydroxides with the synchronous coordinating etching rate towards the soft acid20,38,39,40. The synthetic route is illustrated in Fig. 1a, and the general chemical process can be described as follows:

$${{Cu}}_{2}{{\mbox{O}}}+x{{S}_{2}{O}_{3}}^{2-}+{H}_{2}{{\mbox{O}}}\to {\left[{{Cu}}_{2}{\left({{S}_{2}{O}_{3}}^{2-}\right)}_{x}\right]}^{2-2x}+2{{OH}}^{-}$$
(1)
$${{S}_{2}{O}_{3}}^{2-}+{H}_{2}O\rightleftharpoons {{H}}{{{{\rm{S}}}}_{2}{O}_{3}}^{2-}+O{H}^{-}$$
(2)
$${M}^{x+}+x{{OH}}^{-}\to {M\left({OH}\right)}_{x}$$
(3)
Fig. 1: Synthesis route and theory prediction.
figure 1

a Schematic illustration of the fabrication of HEOs hollow nanocube. b The pKsp and ionic radius of different metal ions. c A periodic table of elements highlighting those that can form hollow nanocube HEOs in this study. d, e Total and partial density of states for NiCoFeCdCr-O and NiCoFe-O. f The d-band centers of NiCoFeCdCr-O (blue), NiCoFeCr-O (pink), NiCoFeCd-O (yellow), and NiCoFe-O (purple).

Full size image

Initially, the nanocubes Cu2O were synthesized to function as a soft acid, and the soft base Na2S2O3 was adopted as the coordinating etchant. In accordance with Pearson’s hard and soft acid-base principle, soft Lewis acids can form stable complexes with soft bases. Consequently, the soft acid characteristic of Cu+ within the Cu2O nanocubes reacts with the soft base ligand (S2O32-) to form the soluble complex [Cu2(S2O3)x]2-2x (Eq. 1), since the soft-soft interaction of Cu+−S2O32- is significantly stronger than the soft-hard interaction of Cu+−O2- within Cu2O nanocubes. Apart from the OH− released during the etching of Cu2O, there are certain OH− originating from the hydrolysis of S2O32− (Eq. 2). Consequently, the other added metal ions (Mx+) reacted with OH− and concurrently began to precipitate (Eq. 3). The shell structure M(OH)x (denoted as HE-OH) prefers to form around the etching interface where the local concentration of OH− is maximized. Upon the dripping of the Na2S2O3 solution, the pH value of the system exhibited a sharp increase, indicating that S2O32- ions rapidly reacted with Cu+, resulting in the release of a significant amount of OH− into the solution, which in turn caused the pH to rise until Cu+ was entirely consumed (Supplementary Figs. 1–3). The synchronous chemical reactions delineated herein guarantee that the exterior of the HE-OH shell precisely replicates the geometry of the Cu2O nanocubes.

An alkaline environment is crucial for the co-deposition of multi-metal ions to form metal hydroxides. In this system, the control of the quantities and ratios of Cu2O and Na2S2O3 exerts a substantial impact on the concentration of OH−, further affecting the formation of HE-OH. It’s worth noting that the precipitation sequence of the added metal ions has a significant impact on the formation of HE-OH during the coordinating etchant process. The Ksp is typically utilized to characterize the precipitation sequence of metal ions to metal hydroxides; the smaller the value of Ksp, the more metal ions can be preferentially precipitated readily in the form of metal hydroxides41. As depicted in Fig. 1b, the considerable variation in the pKsp (pKsp = -lgKsp) values indicates that the precipitation sequence requisite for these ions to hydroxides varies markedly, with Cr3+ (30.2) > Sn2+ (27.26) > La3+ (18.7) > Zn2+ (16.5) > Fe2+ (16.31) > Ni2+ (15.26) > Co2+ (14.23) > Cd2+ (14.14). This suggests that Cr3+ has the highest precipitation priority. Indeed, the conventional co-precipitation approach faces significant difficulties in the formation of high-entropy hydroxides composed of eight metal elements. This is mainly due to the considerable difference in the precipitation sequence of metal ions. To verify the controllability of the coordinating etchant method, the traditional precipitation method of a pure metal ion reaction with NaOH was carried out as a comparison test, and the products were denoted as NiCoFeCdCrLaSnZnOHNaOH. X-ray diffraction (XRD) results indicate that the eight-element hydroxide fabricated by the traditional method is not a single phase, instead, a phase separation occurs (Supplementary Fig. 4). In this system, as a consequence of the precise regulation of the coordinating etchant strategy, which encompasses the concentration of metal ions, soft acid, and soft base, the co-precipitation of eight metal ions into HE-OH is achieved. Furthermore, an excessive difference in ion radius is prone to cause phase segregation. The coordinating etchant strategy presented herein also exhibits extensive applicability in the formation of HE-OH across a broad range of metal ionic radii, even for the minimum Cr3+ (0.69 Å) and the maximum La3+ (1.06 Å). Based on our findings, divalent and trivalent metal ions with ionic radii ranging from 0.69 Å to 1.06 Å and pKsp values between 14.14 and 32.2 are suitable for synthesizing high-entropy spinel oxides.

Ultimately, the HEOs are synthesized through a thermal treatment process at 873 K in air for HE-OH. Consequently, the template-assisted route inspired by the coordinating etching strategy enables the synthesis of a diverse library of HEOs. As illustrated in Fig. 1c, a series of spinel oxides, including ternary NiCoFe-O, quaternary NiCoFeCd-O, quaternary NiCoFeCr-O, as well as HEOs systems such as quinary NiCoFeCdCr-O, senary NiCoFeCdCrLa-O, septenary NiCoFeCdCrLaSn-O, and octonary NiCoFeCdCrLaSnZn-O, have been successfully synthesized.

In addition, the potential benefits of the synthesized HEOs for catalytic applications can be elucidated through density functional theory (DFT) calculations. As illustrated by the density of states (DOS) in Fig. 1d, the quinary high-entropy oxide exhibits a more continuous electronic structure compared to ternary (Fig. 1e) and quaternary materials (Supplementary Figs. 5–10), which facilitates the formation of highly active sites and broadens the range of adsorption energies, thereby enhancing its catalytic properties42,43,44. In addition, the significant overlap observed among the orbitals of all elements indicates strong bonding, which not only facilitates electron transfer between various metal sites but also provides multiple active sites for reactions45. In accordance with the d-band theory, the adsorption energy of reaction intermediates can be optimized by modifying the d-band center of metal sites, which can be influenced by the interaction with other metal dopants45,46. Simultaneously, the centers of the Ni-3d and Co-3d orbitals exhibited a negative shift, while the Fe-3d orbitals displayed a positive change upon the introduction of Cd and Cr (Supplementary Fig. 11). Therefore, the Cd and Cr dopant plays a crucial role in modulating the electronic structure of Ni, Co, and Fe, thus leading to an optimal d-band center of Ni, Co, and Fe in NiCoFeCdCr-O for the favorable adsorption of intermediates. As illustrated in Fig. 1f, the formation of the quinary HEOs results in a more optimal d-band center for NiCoFeCdCr-O at −2.372 eV. In comparison to NiCoFe-O (−1.325 eV), NiCoFeCd-O (−2.791 eV), and NiCoFeCr-O (−1.333 eV), this configuration of NiCoFeCdCr-O exhibits moderate binding strength with reaction intermediates, thereby facilitating product release from the surface and enhancing catalytic activity.

Composition and structure of NiCoFeCdCr-O

The scanning electron microscopy (SEM) and XRD results presented in Supplementary Fig. 12a and 12b indicate that Cu2O exhibits a nanocube morphology, with the XRD diffraction peaks accurately corresponding to the cuprite phase of Cu2O (JCPDS Card Number 05-0667). The transmission electron microscopy (TEM) images further reveal a solid nanocube structure with an average size of approximately 300 nm (Supplementary Fig. 13). As shown in Supplementary Fig. 14a, NiCoFeCdCr-OH exhibits a morphology of hollow nanocube, and the XRD results indicate that the diffraction peaks can be accurately assigned to classical hydroxides (Supplementary Fig. 14b)41. Additionally, the TEM images and elemental distribution of NiCoFeCdCr-OH are presented in Supplementary Fig. 15, demonstrating that the five metal elements and oxygen are uniformly and randomly distributed throughout the framework, while the interior is observed to be hollow. The XRD and TEM patterns primarily indicate that the single-phase NiCoFeCdCr-OH precursor was successfully synthesized using the coordinating etchant method. To further elucidate the coordinating etchant process in detail, a 2-minute reaction was conducted, and the product was annealed at 600 °C to obtain NiCoFeCdCr-O2min. The corresponding TEM images and element distribution, as shown in Supplementary Fig. 16, the Cu element occupies the core of the nanocube, revealing that the Cu2O is not completely coordinated etching and predominantly remains after a 2-minute reaction time. The Ni/Co/Fe/Cd/Cr elements are mainly distributed edges around the cube in the form of a shell composed of small nanosheets, indicating that few NiCoFeCdCr-OH derivatives are still formed even after 2 min reaction between Cu2O and Na2S2O3. This observation further corroborates the efficacy of the coordinating etchant process. To further investigate the effects of the thermal treatment process on the products, various temperatures were analyzed. As illustrated in Supplementary Fig. 17, with a gradual increase in synthesis temperature from 300 to 700 °C, the hydroxide precursor undergoes thermal decomposition and subsequently transforms into a single-phase spinel structure at elevated temperatures, driven by entropy47. The formation of a single-phase hydroxide precursor enables subsequent annealing to facilitate the formation of high-entropy oxides via dehydration and oxidation, ultimately yielding a high-entropy oxide solid solution. The driving force for the annealing of high-entropy hydroxides in air to form high-entropy oxides primarily originates from their enhanced thermodynamic stability.

Figure 2a and Supplementary Fig. 18a present the TEM images of NiCoFeCdCr-O, illustrating the morphology of the hollow nanocube. The nanosheets surrounding the surface of the hollow nanocube exhibit a degree of shrinkage after the annealing treatment compared to NiCoFeCdCr-OH. Furthermore, corresponding to Fig. 2b, the atomic strain distribution pattern derived from geometric phase analysis (GPA) indicates an uneven strain distribution (Fig. 2c), featuring numerous discontinuous yellow compressive strain regions and blue tensile strain regions. This phenomenon is linked to the incorporation of metal elements with differing ionic radii, suggesting that lattice distortion occurred in NiCoFeCdCr-O48,49,50,51. The element mapping presented in Fig. 2d illustrates the homogeneous distribution of Ni, Co, Fe, Cd, Cr, and O elements within the quinary HEOs. Furthermore, the results obtained from inductively coupled plasma-optical emission spectrometry (ICP-OES) in Supplementary Table 1 indicate that the metal content of the five elements is approximately uniform, thereby reinforcing the notion of a consistent elemental distribution within the high-entropy phase with the configurational entropy of 1.573 R. Additionally, the magnified TEM and HAADF-STEM images depicted in Supplementary Fig. 18b and Fig. 2e reveal that NiCoFeCdCr-O is composed of various nanosheets oriented differently and interconnected by grain boundaries. The illustrated model of grain boundaries is presented in Fig. 2f10,52. As illustrated in Supplementary Fig. 19, the XRD pattern of NiCoFeCdCr-O demonstrates the establishment of a spinel structure in HEOs without any evidence of phase segregation53,54. Furthermore, the Fig. 2g and enlarged TEM images in Supplementary Fig. 20 confirm the clear spacing lattice fringes of NiCoFeCdCr-O. Notably, the spacing lattice fringes of 0.245, 0.238, and 0.296 nm correspond to the typical (222) and (220) planes of spinel structure (Fd-3m) CoFe2O4 (PDF#22-1086).

Fig. 2: Morphology and structure of hollow nanocube NiCoFeCdCr-O.
figure 2

a TEM image. b Aberration-corrected high-angle-annular-dark-field scanning transmission electron microscopy (AC HAADF-STEM) image. c The related atomic strain distribution originated from (b). d HAADF-STEM and corresponding element mapping images. e HAADF-STEM image with abundant grain boundaries (GB) marked with white dashed lines. f The model of NiCoFeCdCr-O hollow nanocube with GB (Multiple colors only represent high-entropy characteristics. GB is arbitrary and arbitrarily defined, with no specific color assigned to it.) g Integrated pixel intensities of the HEO phase taken from d1 and d2 in (e).

Full size image

Apart from this, elemental mapping images of ternary NiCoFe-O, quaternary NiCoFeCd-O, and NiCoFeCr-O displayed a uniform distribution of elements in Supplementary Figs. 21–23. Besides, the XRD patterns exhibit similar information to quinary NiCoFeCdCr-O (Supplementary Fig. 24). The TEM images were provided to demonstrate that the spacing lattice fringes align with the typical planes of the spinel structure CoFe2O4 (PDF#22-1086) in Supplementary Figs. 25–27, suggesting the synthesis strategy is also suitable for low-component spinel oxides. From the above results, it can be concluded that different components of oxides were synthesized by this method, and the spinel structure was retained even with the different incorporation elements.

Library synthesis for HEOs

Our method for synthesizing HEOs exhibits universal applicability and can be employed to produce senary NiCoFeCdCrLa-O (Fig. 3a), septenary NiCoFeCdCrLaSn-O (Fig. 3b), and octonary NiCoFeCdCrLaSnZn-O (Fig. 3c) with a hollow nanocube structure. Furthermore, the magnified TEM results presented in Fig. 3a–c and Supplementary Figs. 28–30 indicate that the lattice fringe spacings correspond to the characteristic planes of the spinel structure (Fd-3m) CoFe2O4 (PDF#22-1086). In Fig. 3d, Synchrotron radiation X-ray diffraction (SXRD) patterns of all samples also reveal similar major diffraction peaks associated with the spinel phase structure (Fd-3m) without any phase separation, which aligns with the TEM findings. As illustrated in Supplementary Fig. 31, the XRD pattern of octonary NiCoFeCdCrLaSnZn-OH undergoes a transition to that of octonary NiCoFeCdCrLaSnZn-O at varying annealing temperatures, thereby further corroborating that this strategy offers a universal and rapid synthesis method for a wide range of HEOs.

Fig. 3: Morphology and composition of HEOs.
figure 3

TEM, HAADF-STEM, and element mapping images of a NiCoFeCdCrLa-O, b NiCoFeCdCrLaSn-O, c NiCoFeCdCrLaSnZn-O (scale bar: 200 nm). d SXRD patterns of the obtained senary, septenary, and octonary HEOs.

Full size image

Furthermore, Synchrotron radiation X-ray diffraction pair distribution function (PDF) analysis was added to investigate the phase structure of the spinel HEOs system. As illustrated in Supplementary Fig. 32a, an increase in the metal component content has been observed to lead to the broadening of characteristic peaks, particularly in the octonary NiCoFeCdCrLaSnZn-O system. This phenomenon suggests a higher degree of structural disorder, which is likely attributable to inhomogeneous lattice strain resulting from differences in atomic sizes within the HEOs system55,56. Moreover, an analysis of the local atomic structure within 6 Å was conducted (Supplementary Fig. 32b). The principal features observed in the PDF correspond to the nearest neighbor M–O bonds (~2 Å), MO − MO pair (~3 Å), and possibly a peak attributed to MT − MO pair (~3.5 Å), all of which align with the respective interatomic distances reported for the spinel CoFe2O4 structure57,58,59. This consistency supports the uniform formation of the spinel structure and confirms the successful synthesis of high-entropy spinel oxides.

The electronic structure of HEO

X-ray photoelectron spectroscopy (XPS) was utilized to analyze the surface chemical compositions and valence states of the elements. As shown in Supplementary Fig. 33, the XPS survey spectrum confirms the successful synthesis of ternary NiCoFe-O, quaternary NiCoFeCd-O, quaternary NiCoFeCr-O, and quinary NiCoFeCdCr-O. In Supplementary Fig. 34, it is evident that the electronic structures of Ni and O are influenced by the incorporation of Cr in both quaternary NiCoFeCr-O and NiCoFeCdCr-O when compared to ternary NiCoFe-O and quaternary NiCoFeCd-O. The emergence of a peak at O1 position from the O 1s XPS spectrum is attributed to strong interactions between Cr and O. This Cr–O interaction may also affect the electronic structure of Ni60,61. Analysis of the Cr 2p, Co 2p, Fe 2p, and Cd 3d XPS spectra presented in Supplementary Figs. 35–36 indicated that Cr exists as Cr3+, while Co, Fe, and Cd are present as Co2+, Fe3+, and Cd2+ in NiCoFeCdCr-O, respectively. Moreover, the high-resolution XPS spectra for ternary NiCoFe-O, quaternary NiCoFeCd-O, and NiCoFeCr-O demonstrate consistent valence states among these metal elements (Supplementary Fig. 37), comparable to those observed in quinary NiCoFeCdCr-O, thereby indicating a similar structural configuration across samples synthesized via this strategy.

The electronic structure of the catalysts was further elucidated by X-ray Absorption Fine Structure (XAFS) analysis. Analysis of the X-ray absorption near-edge structure (XANES) presented in Fig. 4a, c, e, and i reveals that the adsorption edges of Ni, Co, Fe, and Cr K-edges in ternary NiCoFe-O, quaternary NiCoFeCd-O, quaternary NiCoFeCr-O, and quinary NiCoFeCdCr-O closely resemble those of NiO, Co3O4, α-Fe2O3 and Cr2O3. This observation indicates that their oxidation states are analogous to those found in the respective reference oxides47. Furthermore, the Cd K-edge XANES spectra elucidated the similar electron structure of Cd in quaternary NiCoFeCd-O and quinary NiCoFeCdCr-O. To further investigate the alterations in coordination configuration, an extended X-ray absorption fine structure (EXAFS) analysis was conducted. As illustrated in the Ni EXAFS spectra (Fig. 4b), the first shell EXAFS peak at 1.63 Å corresponds to Ni-O scattering, while peaks at 2.55 Å are associated with Ni-M pathways61. Notably, shorter Ni-O and Ni-M bond distances were recorded for samples incorporating other metals, indicating that metal incorporation induces changes in both Ni-O and Ni-M lengths62. The Co K-edge EXAFS spectra of NiCoFeCdCr-O and Co3O4 show similar oscillation patterns, reflecting structural similarity63. The first shell EXAFS peak observed at approximately 1.44 Å is attributed to Co-O scattering (Fig. 4d). Peaks corresponding to Co-M1 and Co-M2 pathways were detected at 2.42 and 2.98 Å, respectively. It is noteworthy that longer Co-O and Co-M bond distances were recorded for samples incorporating other metals, indicating that variations in the local coordination environments surrounding Co atoms due to metal incorporation, thereby leading to an increase in both Co-O and Co-M lengths64. As illustrated in Fig. 4f, the first coordination shell EXAFS peak at 1.47 Å corresponds to Fe-O scattering, while peaks at 2.52 Å associated with Fe-M pathways were also identified65. The shorter Fe-O bond and longer Fe-M bond are observed for samples incorporating other metals66,67,68. In Fig. 4h, NiCoFeCd-O exhibits Cd-O and Cd-M peak positions at 1.72 and 3.01 Å, respectively. However, the Cr incorporation causes a dramatic decrease in Cd-O length and an increase in Cd-M. The Cr-O coordination is observed in Fig. 4j, suggesting the strong effect of Cr-O69. Besides, contrasting with the M-O length, major changes in Ni-M, Co-M, Fe-M, Cd-M, and Cr-M bond lengths are observed in ternary NiCoFe-O, quaternary NiCoFeCd-O, quaternary NiCoFeCr-O, and quinary NiCoFeCdCr-O. Importantly, similar results are further corroborated by the EXAFS analysis of NiCoFe-O, NiCoFeCd-O, NiCoFeCr-O, NiCoFeCdCrLa-O, NiCoFeCdCrLaSn-O, and NiCoFeCdCrLaSnZn-O (Supplementary Figs. 38–45). Overall, the aforementioned characterization results suggest that the high-entropy system exhibits lattice distortion upon the incorporation of various elemental atoms, which modifies interatomic coordination and electronic structure, potentially resulting in good catalytic activity. Bader charge analysis further offers insight into the varying numbers of transferred electrons at the active site for ternary NiCoFe-O, quaternary NiCoFeCd-O, quaternary NiCoFeCr-O, and quinary NiCoFeCdCr-O. It indicates that oxygen tends to accept electrons, whereas Fe and Cr are more prone to electron loss, with Cd exhibiting the least tendency to lose electrons. In comparison to NiCoFe-O, the number of electrons received by oxygen in NiCoFeCdCr-O appears to have decreased, suggesting that electron transfer may have occurred among different elements (Fig. 4k).

Fig. 4: Electron structure of HEO.
figure 4

a Ni K-edge XANES spectra and corresponding b EXAFS spectra for NiCoFe-O, NiCoFeCd-O, NiCoFeCr-O, and NiCoFeCdCr-O with NiO as references. c Co K-edge XANES spectra and corresponding d EXAFS spectra for NiCoFe-O, NiCoFeCd-O, NiCoFeCdCr-O, and NiCoFeCr-O with Co3O4 as references. e Fe K-edge XANES spectra and corresponding (f) EXAFS spectra for NiCoFe-O, NiCoFeCd-O, NiCoFeCr-O, and NiCoFeCdCr-O with α-Fe2O3 as references. g Cd K-edge XANES spectra, and corresponding h EXAFS spectra for NiCoFeCd-O and NiCoFeCdCr-O. i Cr K-edge XANES spectra and corresponding (j) EXAFS spectra for NiCoFeCr-O and NiCoFeCdCr-O with Cr2O3 as references. k Bader charge of different metal atoms, including O, Cd, Cr, Fe, Co, and Ni among NiCoFe-O, NiCoFeCd-O, NiCoFeCr-O, and NiCoFeCdCr-O.

Full size image

Performance of catalytic hydrogenation for p-nitrophenol

As proof of demonstration, the catalytic performance of quinary NiCoFeCdCr-O was explored as a catalyst for the catalytic hydrogenation of p-nitrophenol (4-NP). As illustrated in Supplementary Fig. 46a, the aqueous solution of 4-NP exhibits an absorption peak at approximately 317 nm70. The introduction of sodium borohydride leads to the appearance of a new absorption peak around 400 nm. Following the addition of the catalyst to the solution, a decrease in the absorbance of 4-NP is observed, accompanied by the emergence of a distinct peak near 310 nm, which corresponds to 4-aminophenol (4-AP). Figure 5a illustrates the typical absorbance changes of 4-NP during the reactions conducted at 25 °C. The NiCoFeCdCr-O catalyst achieves a remarkable 100% conversion efficiency of 4-NP within 150 seconds, which is superior to ternary NiCoFe-O, quaternary NiCoFeCd-O, and quaternary NiCoFeCr-O in Supplementary Fig. 46b–d. Furthermore, a linear relationship is established in Fig. 5b by fitting the values of ln(Ct/C0) against the corresponding reaction time, which aligns with first-order kinetic behavior. Consequently, the calculated apparent rate constant (k) for NiCoFeCdCr-O was 1.79 min−1 (Fig. 5c), which is 7.40, 2.82, and 1.79 times that of NiCoFe-O (0.242 min−1), NiCoFeCd-O (0.635 min-1), and NiCoFeCr-O (1.00 min-1), respectively. Moreover, NiCoFeCdCr-O exhibited a high specific constant (K) of 1.53 × 106 min-1 g-1. As illustrated in Fig. 5d and Supplementary Tables 2–3, NiCoFeCdCr-O outperforms most reported noble metal-based catalysts and high-entropy materials for the direct hydrogenation of 4-NP. For further exploration, a comparative analysis was conducted on NiCoFeCdCr-OH, NiCoFeCdCrLaSnZn-OH, and NiCoFeCdCrLaSnZn-O, as well as NiCoFeCdCr-Ox (x = 300/400/500/700 °C) produced at different annealing temperatures (Supplementary Figs. 47–49). The conclusion drawn from the research is that NiCoFeCdCr-O exhibits optimal performance when subjected to annealing at a temperature of 600 °C. Furthermore, a linear relationship can be derived by fitting the value of ln(Ct/C0) under the various reaction temperatures in Fig. 5e. The values of k demonstrated an increase with elevated reaction temperatures, and this observation can be explained by collision theory. As the system temperature rises, the intense motion of the catalyst and 4-NP increases the likelihood of collisions, resulting in faster reaction rates. The final activation energy (Ea) for NiCoFeCdCr-O was calculated to be 49.7 kJ/mol from the calculated k at different reaction temperatures using the Arrhenius equation. The corresponding activation energy values for several additional high-entropy and noble metal-based catalysts as presented in Supplementary Table 4. These results clearly indicate that NiCoFeCdCr-O exhibits relatively low activation energies within the category of high-entropy alloy catalysts and is also comparable to numerous noble metal-based catalysts. This observation is consistent with the superior reactivity of the catalysts. The lower the activation energy indicates the less energy is required for the reaction, which is more conducive to the reaction71,72. Additionally, the performance of various 4-NP concentrations was evaluated to assess the catalytic suitability of the catalyst (Supplementary Fig. 50). Reusability is widely recognized as a critical parameter for catalysts. As illustrated in Fig. 5f, the NiCoFeCdCr-O catalyst exhibited sustained high activity over 10 cycles, with the conversion rate consistently maintained above 95%. There is no change in the valence states of these metal elements in the NiCoFeCdCr-O system after the catalytic reaction, confirming the stable phase structure of high-entropy oxides NiCoFeCdCr-O (Supplementary Fig. 51). These findings indicate that the NiCoFeCdCr-O catalyst not only demonstrates high catalytic activity but also exhibits good recyclability. To further explore the general applicability of the NiCoFeCdCr-O catalyst in the reduction of nitrophenolic pollutants to aminophenol, a range of nitrophenol with diverse structures was selected as probe molecules, including 2-nitrophenol (2-NP) and 3-nitrophenol (3-NP). As illustrated in Supplementary Fig. 52, a series of successive UV–vis spectra clearly demonstrates that the NiCoFeCdCr-O catalyst facilitates the reduction of nitrophenol to its corresponding aminophenol products in the presence of NaBH4.

Fig. 5: The catalytic reduction of 4-NP to 4-AP over as-prepared catalysts.
figure 5

a The UV–vis absorbance spectra of 4-NP solutions with a change of reaction time at 25 °C for NiCoFeCdCr-O. b Plots of ln(Ct/C0) against the reaction time of as-measured catalysts. c The corresponding histogram of rate constants. d Comparison of rate constants (k) and ratio constant (K) for catalytic reduction of 4-NP over NiCoFeCdCr-O with previously reported noble metal-based nanocatalysts81,82,83,84,85,86,87,88,89,90. e Plots of ln(Ct/C0) against the reaction time of NiCoFeCdCr-O at different temperatures (Inset: the Arrhenius plots). f The recycling tests of NiCoFeCdCr-O for the hydrogenation of 4-NP over ten cycles. g The calculated adsorption energies of p-nitrophenol on various metal sites of NiCoFeCdCr-O. h Stepwise hydrogenation and potential energy profiles of 4-NP hydrogenation on NiCoFeCdCr-O.

Full size image

To demonstrate the superiority of the coordinated etching and precipitation method, as well as the limited reactivity of CuO, NiCoFeCdCrONaOH was synthesized using NaOH solution in place of Na2S2O3 solution, while all other synthesis and testing conditions remained constant. As shown in Supplementary Fig. 53a, the catalytic activity of NiCoFeCdCrONaOH exhibits a precipitous decline compared to NiCoFeCdCr-O, with 4-NP being entirely degraded for more than 40 minutes. The XRD results of NiCoFeCdCrONaOH sample confirmed the presence of CuO as the dominant crystalline phase, further verifying that NaOH does not chemically react with Cu2O. The unreacted Cu2O subsequently transforms into CuO during thermal treatment. No diffraction peaks corresponding to the NiCoFeCdCr oxides were observed, attributable to their low concentration compared to CuO within the system (Supplementary Fig. 53b). These findings indicate that the high entropy oxides component NiCoFeCdCr-O, rather than CuO, serves as the primary active species and demonstrates superior catalytic activity for hydrogenation of 4-NP. Furthermore, to confirm the inherent inertness of CuO, the pure CuO was obtained by direct thermal treatment of Cu2O and tested for performance. The pure CuO exhibited significantly lower catalytic activity compared to NiCoFeCdCrONaOH (Supplementary Fig. 53c), as evidenced by the incomplete degradation of 4-NP even after 150 min of reaction, thereby demonstrating the inert activity of CuO in the catalytic hydrogenation of 4-NP. Moreover, the XRD patterns indicated the formation of a high-crystallinity CuO phase without impurities (Supplementary Fig. 53d). Consequently, these results further prove the aforementioned statement that CuO demonstrated inert activity for the hydrogenation of 4-NP. In essence, the coordinating etchant method ensures the overwhelming consumption of Cu2O and circumvents more residues of CuO, thereby conferring upon the hollow nanocube NiCoFeCdCr-O a dominant role in the catalytic process, while Cu species do not affect the catalytic activity of high-entropy materials.

To further elucidate the catalytic mechanism of NiCoFeCdCr-O towards the efficient reduction of 4-NP, DFT analysis was conducted. First, the most stable molecular configurations of 4-NP on various metal sites in NiCoFeCdCr-O, along with the binding energies between 4-NP and these metal sites (Supplementary Fig. 54), were determined through structural optimization. As illustrated in Fig. 5g, the adsorption energy of 4-NP on the Fe site (−2.62 eV) is significantly lower than that on the Cd site (-1.98 eV), Co site (-2.07 eV), Ni site (−2.23 eV), and Cr site (−2.37 eV), indicating a stronger affinity of 4-NP for the Fe site. DFT calculations were further carried out to investigate the hydrogenation reaction pathway of 4-NP over the NiCoFeCdCr-O catalyst. The free energy profile of the most favorable reaction path, as well as the adsorption energies of potential reaction intermediates at the Fe sites of NiCoFeCdCr-O, are presented in Fig. 5h. The adsorption energy of 4-NP on the NiCoFeCdCr-O catalyst (denoted as GR-NO2*) is defined as 0 eV73. GR-NO2* undergoes the first hydrogenation step to form GR-NOOH*, with an adsorption energy of -0.339 eV, indicating that this transformation occurs readily. In the second hydrogenation stage, two possible intermediates GR-NO* and GR-NOHOH* intermediates are considered. The significantly lower adsorption energy of GR-NOHOH* (−1.55 eV) compared to GR-NO* (2.42 eV) suggests that GR-NOHOH* is the dominant intermediate in this step. For the third hydrogenation step, GR-NOH* (-0.503 eV) exhibits greater stability than GR-NHO* (−0.274 eV), making it the favored species. During the subsequent three hydrogenation steps, GR-NOH* is sequentially reduced to GR-NHOH* (−0.261 eV) and GR-NH* (−3.53 eV), ultimately forming GR-NH2* (adsorbed 4-AP)73. Overall, the reduction pathway proposed by our DFT calculations proceeds as follows: GR-NO2* (adsorbed 4-NP) → GR-NOOH* → GR-NOHOH* → GR-NOH* → GR-NHOH* → GR-NH* → GR-NH2* (Supplementary Figs. 55–57). The good catalytic performance of the NiCoFeCdCr-O catalyst can be attributed to the synergistic effects among multiple metal atoms, which modulate the electronic structure and enhance the hydrogenation efficiency of reaction intermediates. Additionally, the three-dimensional hollow nanocube architecture provides a large accessible surface area, thereby facilitating efficient mass transfer between the catalyst and reactants.

Discussion

In summary, we have established a library of multicomponent hollow nanocubes high-entropy spinel oxides, spanning ternary to octonary compositions, synthesized through the template-assisted route inspired by coordinating etching. Our detailed investigations into the underlying synthetic mechanisms elucidate that the formation of high-entropy spinel oxides transpires via four pivotal steps: (1) The preparation of the nanocube Cu2O temple, (2) the liberation of hydroxide ions (OH−) through coordinating etching reactions between soft bases and soft acids; (3) The co-precipitation of various metal cations onto the nanocube shell, leading to the formation of high-entropy metal hydroxide (HE-OH); and (4) subsequent thermal treatment of HE-OH. Experimental results indicate that the formation and stabilization of HEO nanocubes are critically dependent on the generation of HE-OH, which is achieved by meticulously controlling the balance between the precipitation rate of metal hydroxides and the synchronous coordinating etching rate towards the soft acid. Furthermore, the DFT simulation reveals a more continuous DOS near the Fermi level and a more moderate d-band center for the representative quinary NiCoFeCdCr-O hollow nanocube. Consequently, in the catalytic hydrogenation of 4-NP, this quinary NiCoFeCdCr-O demonstrates good catalytic activity and cyclic stability. The proposed synthesis methodology establishes a comprehensive platform for the advancement of HEOs.

Methods

Reagents

Cupric sulfate pentahydrate (CuSO4, 99%), sodium hydroxide (NaOH, 98%), L(+)-Ascorbic acid (99.7%), ethanol absolute (CH3CH2OH), nickel chloride hexahydrate (NiCl2·6H2O, 98%), cobalt chloride hexahydrate (CoCl2·6H2O, 99%), ferrous sulfate (FeSO4·7H2O, 99%), cadmium chloride hemi(pentahydrate) (CdCl2·2.5H2O, 99%), chromic nitrate nonahydrate (Cr(NO3)3·9H2O, 99%), lanthanum nitrate hydrate (La(NO3)3·nH2O), stannous chloride dihydrate (SnCl2·2H2O, 98%), zinc chloride (ZnSO4·7H2O, 98%), polyvinylpyrrolidone K-30 (PVP K-30, 99%), sodium thiosulfate (NaS2O3·5H2O, 99%), p-nitrophenol (4-NP, analytical grade,), 3-nitrophenol (3-NP, 99%), 2-nitrophenol (2-NP, 99%), sodium borohydride (NaBH4, 98%) and sodium citrate were used without any further purification.

Synthesis of Cu2O nanocube

0.75 g CuSO4·5H2O and 0.3 g sodium citrate were dissolved in 200 mL ultra-pure water. Then, 2 g NaOH was added to the above solution under stirring for 0.5 h, and 100 mL of 0.03 M ascorbic acid was dropped into the above solution. After being stirred for 0.5 h, the Cu2O nanocubes were centrifugally separated and dried in a vacuum oven at 60 °C.

Synthesis of precursor NiCoFeCdCr-OH

40 mg of the as-prepared Cu2O nanocubes was dispersed into the mixture solution of 40 mL ultra-pure water and 40 mL ethanol. Afterward, 666.6 mg PVP-K30, 4.8 mg NiCl2·6H2O, 4.8 mg CoCl2·6H2O, 5.6 mg FeSO4·7H2O, 4.6 mg CdCl2·2.5H2O, and 8 mg Cr(NO3)3·9H2O were dissolved in the above solution under stirring for 15 min. Finally, 32 mL of 1.0 M Na2S2O3 was added dropwise slowly into the solution system. After being stirred for 0.5 h, the NiCoFeCdCr-OH were centrifugally separated and dried by vacuum freezing. The NiCoFe-OH, NiCoFeCd-OH, and NiCoFeCr-OH were prepared with the same method, just without CdCl2·2.5H2O or Cr(NO3)3·9H2O.

Synthesis of precursor NiCoFeCdCrLaSnZn-OH

40 mg of the as-prepared Cu2O nanocubes was dispersed into the mixture solution of 40 mL ultra-pure water and 40 mL ethanol. Afterward, 666.6 mg PVP-K30, 3.6 mg NiCl2·6H2O, 3.6 mg CoCl2·6H2O, 4.2 mg FeSO4·7H2O, 3.4 mg CdCl2·2.5H2O, 6.0 mg Cr(NO3)3·9H2O, 4.9 mg La(NO3)3·nH2O, 3.4 mg SnCl2·2H2O and 4.3 mg ZnSO4·7H2O were dissolved in the above solution under stirring for 15 min. Finally, 32 mL of 1.0 M Na2S2O3 was added dropwise into the solution system. After being stirred for 0.5 h, the NiCoFeCdCrLaSnZn-OH were centrifugally separated and dried by vacuum freezing. The NiCoFeCdCrLa-OH and NiCoFeCdCrLaSn-OH were prepared with the same method, just without SnCl2·2H2O and ZnSO4·7H2O. The NiCoFeCdCrLaSnZnOHNaOH were synthesized by the same synthetic route, with Na2S2O3 replaced by NaOH and without Cu2O and PVP.

Synthesis of HEOs

The dried HE-OH was annealed in a tube oven at 600 °C for 2 h with a programmed heating rate of 5 °C min−1. Additionally, the NiCoFeCd-O, NiCoFeCr-O, and NiCoFe-O were synthesized by the same synthetic route without adding Cr or Cd. The NiCoFeCdCrLa-O, NiCoFeCdCrLaSn-O, and NiCoFeCdCrLaSnZn-O were synthesized by the same synthetic route with the addition of La(NO3)3·nH2O (4.9 mg), SnCl2·2H2O (3.4 mg), and ZnSO4·7H2O (4.3 mg). The NiCoFeCdCrONaOH was synthesized by the same synthetic route, with Na2S2O3 replaced by NaOH.

Characterization

X-ray diffraction (XRD) patterns were tested from Bruker D8 Advance (scan range of 10–80°). Scan electron microscopy (SEM) images were carried out on a JEOL TSM-7500F field emission scanning electron microscope. The transmission electron microscopy (TEM), High Angle Annular Dark Field-Scanning Transmission Electron Microscopy (‌HADDF-STEM), and corresponding elemental mapping images were obtained from the FEI Talos F200S emission scanning electron microscope and operating at an accelerating voltage of 200 kV. The X-ray photoelectron spectroscopy (XPS) measurements were performed using an Escalab 250Xi electron spectrometer equipped with Mg Kα radiation. The sample preparation methodology for XPS comprises the following procedures: Initially, a minimal amount of the powdered specimen is precisely deposited onto an aluminum foil substrate. Subsequently, a segment of double-sided Scotch tape (~5 × 5 mm2 in area) is precisely excised and securely grasped with fine-tipped tweezers to facilitate sample mounting. The tape-sample assembly is then uniformly compressed using a tablet press under controlled pressure until the sample surface achieves optimal planarity and is completely devoid of surface contaminants. Finally, the prepared sample is mounted onto the sample stage for subsequent analysis. All acquired spectra are calibrated using the C 1 s peak (with a binding energy of 284.8 eV) as the reference standard. Metal element content images were obtained from inductively coupled plasma-optical emission spectrometer (ICP-OES). The high-resolution transmission electron microscopy (HRTEM) imaging was acquired on a JEOL JEM-ARM200CF microscope operated at 200 kV with a Schottky cold-field emission gun at Wuhan University. Ultraviolet-visible Spectrophotometer (UV-Vis) spectroscopy measurements were carried out using a Shimadzu 3600 Plus. X-ray Absorption Fine Structure (XAFS) spectra, Synchrotron radiation X-ray diffraction (SXRD), and Synchrotron radiation X-ray diffraction Pair Distribution Function (PDF) were acquired under ambient conditions in transmission mode at beamlines BL14W1, BL14B1, and BL13SSW of the Shanghai Synchrotron Radiation Facility.

Catalytic reduction of p-nitrophenol

Typically, NaBH4 (0.1898 g) was dissolved into 20 mL ultrapure water, and then NaBH4 aqueous solution was added into the 20 mL of 2 mM 4-NP to make the concentration of 4-NP to be 1 mM in the catalytic system. Then, after stirring for 5 min, 20 μL of catalyst suspension (2 mg/mL) was injected quickly into the system of 4-NP and NaBH4 to cause the hydrogenation reaction. Subsequently, 0.5 ml solution from the system was diluted to 10 ml during the reaction process at regular intervals (30 s) and then monitored through a UV–vis spectrometer. The catalyst was separated by the filtration device and dried after the reduction reaction had finished, and then the performance test was carried out again with other experimental conditions that remained unchanged for recyclability experiments with ten successive runs.

$${{\mathrm{ln}}}\left(\frac{{C}_{t}}{{C}_{0}}\right)={{\mathrm{ln}}}\left(\frac{{A}_{t}}{{A}_{0}}\right)=-{kt}$$
(4)

Where k is the apparent rate constant, t is the reaction time, Ct (At) and C0 (A0) are defined as the concentrations (absorbances) of 4-NP at time t and t = 0, respectively.

$$K=\frac{k}{{m}_{{metal}}}$$
(5)

Where K is the ratio rate constant, mmetal is the weight of the metal in the used catalyst.

$${{\rm{Conversion}}}=\frac{{C}_{0}-{C}_{t}}{{C}_{0}}\times 100\%$$
(6)

Here, C0 is the initial absorption and Ct is the final absorption at a desired interval of time, t.

The activation energy for the reduction of p-nitrophenol could be calculated according to the Arrhenius equation:

$${{\mathrm{ln}}}\,k=-\frac{{E}_{a}}{{RT}}+{{\mathrm{ln}}}\,A$$
(7)

Where R is the ideal gas constant (equal to 8.314 J mol-1 K-1), A is the index factor, and the slope of the logarithmic plot of rate constant (k) and T-1 can be indicated as -Ea/R.

Computational detail

All the calculations are performed in the framework of the density functional theory with the projector augmented plane-wave method, as implemented in the Vienna ab initio simulation package (VASP6.2.0)74,75. The generalized gradient approximation (GGA) proposed by Perdew, Burke, and Ernzerhof (PBE) is selected for the exchange-correlation potential76,77. The long-range van der Waals interaction is described by the DFT-D3 approach78. The cut-off energy for the plane wave is set to 480 eV. The energy criterion is set to 10−5 eV in the iterative solution of the Kohn-Sham equation. A vacuum layer of 20 Å is added perpendicular to the sheet to avoid artificial interaction between periodic images. The Brillouin zone integration is performed using a 5 × 3 × 1 k-mesh. All the structures are relaxed until the residual forces on the atoms have declined to less than 0.02 eV/Å. Data analysis and visualization are carried out with the help of the VASPKIT79 code and VESTA80.

The adsorption energy Eads is expressed as:

$$\Delta {E}_{{abs}}={E}_{A+B}-{E}_{A}-{E}_{B}$$
(8)

Where EA+B is the total energy of slab A model with B adsorption, EA is the energy of a A slab, and EB is that for a B molecule.