General synthesis of high-entropy alloy and ceramic nanoparticles in nanoseconds

High-entropy materials, which include high-entropy alloys and high-entropy ceramics, show promise for their use in many fields, yet a robust synthesis strategy is lacking. Here we present a simple and general approach, laser scanning ablation, to synthesize a library of high-entropy alloy and ceramic nanoparticles. The laser scanning ablation method takes only five nanoseconds per pulse to ablate the corresponding nanoparticle precursors at atmospheric temperature and pressure. The ultrarapid process ensures that dissimilar metallic elements combine regardless of their thermodynamic solubility. As a laser pulse confines energy to the desired microregions, the laser scanning ablation method renders a high-entropy material nanoparticle loading on various substrates, which include thermally sensitive substrates. Applied as electrocatalysts for overall water splitting, the as-prepared high-entropy material nanoparticles can achieve an overpotential of 185 mV @ 10 mA cm–2. This versatile strategy enables the preparation of materials useful for a range of fields, such as biomedicine, catalysis, energy storage and sensors. High-entropy materials are used in a range of applications but their synthesis at the nanoscale remains challenging. Now, a robust and general strategy to prepare high-entropy alloy and ceramic nanoparticles has been developed using laser scanning ablation. This approach takes only five nanoseconds per pulse to ablate precursors at atmospheric temperature and pressure.

H igh-entropy materials (HEMs), which include high-entropy alloys (HEA) and high-entropy ceramics (HEC), are a class of materials that contain at least five near-equimolar principal metal atoms in an amorphous structure or a solid-solution phase 1 . These unconventional compositions and structures offer HEM unprecedented physicochemical properties, such as high strengths 2 , unique electrical and magnetic properties 3 , and promising resistances to wear, oxidation and corrosion 4 . However, it is still a daunting task to integrate multiple elements into HEMs at the nanoscale, which could open up an effective avenue to tune their properties for many applications.
A few synthesis techniques, which include carbothermal shock 5 , moving bed pyrolysis 6 , ultrasonication-assisted wet chemistry method 7 , laser-assisted strategy 8 and electrosynthesis 9 , have been reported for HEA nanoparticle (NP) synthesis (Supplementary  Table 1). However, current methods generally require rigorous conditions, which include high pressure, high temperature and inert atmospheric protection 7,8,10 . These approaches typically produce NPs immobilized on thermally resistant substrates, which can withstand temperatures in excess of 700 K (for example, carbon and Al 2 O 3 ), rather than on thermally sensitive substrates, such as metal, glass, and so on. For synthetic methods that use mild conditions 9,10 , only a small variety of HEA NPs can be produced because the synthesis either requires source materials with the same composition as the NPs (e.g., if we need to synthesize PtIrCuNiCr HEA NPs, bulk source materials of PtIrCuNiCr must be provided) or results in the formation of amorphous NPs. Surprisingly, strategies for the creation of a series of HEC NPs have not yet been explored because research on HEC materials is still in its infancy 11 . Clearly, it is of great significance to develop a general yet robust route to synthesizing both HEA and HEC NPs under mild conditions over a diversity of substrates.
We developed a laser scanning ablation (LSA) approach using a pulsed nanosecond laser to create a library of HEA and HEC NPs at atmospheric temperature and pressure. The laser ablated the corresponding NP precursors in alkanes, which enabled the formation of HEA and HEC (that is, oxide, sulfide, phosphide, nitride and boride) NPs within only a five nanosecond pulse. The ultrarapid process ensured up to nine metallic elements combined uniformly regardless of their thermodynamic solubility. As the laser pulse precisely confines the energy to desired microregions, HEM NPs can form on various substrates, even thermally sensitive ones, such as metals and glass. Substrate-free colloidal HEM NPs can also be created by the LSA technique, which demonstrates the wide applicability of the strategy. Notably, the as-crafted HEM NPs were exploited as high-efficiency, stable bifunctional electrocatalysts with a low noble metal amount for overall water splitting, which exemplifies their utility for energy conversion and storage, among other areas. The HEA NPs of PtIrCuNiCr deliver a current density of 10 mA cm -2 at only 1.42 V, which outperforms state-of-the-art catalysts in the literature.

Results and discussion
HEA NPs on carbon substrates. In a typical LSA process, metal chlorides of equal molar ratio were first loaded onto a substrate. Then, the substrate was transferred to hexane, and irradiated by laser pulses at room temperature (~25 °C), as shown in Fig. 1a,b and Methods. At a constant average power density of 2 × 10 5 W cm -2 , a peak pulse power density of 2 × 10 9 W cm -2 ( Fig. 1c and Supplementary Fig. 1), calculated by pulse energy/pulse width, was used with a pulse width of 5 nm for a rapid HEM fabrication. As carbon nanofibres (CNFs) with surface-bound residual oxygen facilitated the melt metal movement and fission events 11 , we first used CNFs as the substrate ( Supplementary Fig. 2) to support HEA NPs

General synthesis of high-entropy alloy and ceramic nanoparticles in nanoseconds
Bing Wang 1,2,3,4 , Cheng Wang 5 , Xiwen Yu 5 , Yuan Cao 1 , Linfeng Gao 5 , Congping Wu 1,4 , Yingfang Yao 1,3,4,5 ✉ , Zhiqun Lin 2 ✉ and Zhigang Zou 1,3,4,5,6,7 ✉ High-entropy materials, which include high-entropy alloys and high-entropy ceramics, show promise for their use in many fields, yet a robust synthesis strategy is lacking. Here we present a simple and general approach, laser scanning ablation, to synthesize a library of high-entropy alloy and ceramic nanoparticles. The laser scanning ablation method takes only five nanoseconds per pulse to ablate the corresponding nanoparticle precursors at atmospheric temperature and pressure. The ultrarapid process ensures that dissimilar metallic elements combine regardless of their thermodynamic solubility. As a laser pulse confines energy to the desired microregions, the laser scanning ablation method renders a high-entropy material nanoparticle loading on various substrates, which include thermally sensitive substrates. Applied as electrocatalysts for overall water splitting, the as-prepared high-entropy material nanoparticles can achieve an overpotential of 185 mV @ 10 mA cm -2 . This versatile strategy enables the preparation of materials useful for a range of fields, such as biomedicine, catalysis, energy storage and sensors.
with Au, Fe, Co, Cu and Cr elements. HEA products were crafted efficiently on CNFs with a uniform dispersion (Fig. 1d). Energydispersive X-ray spectroscopy (EDS) maps depicted homogeneous distributions of all the five elements throughout the NP. Based on the atomic ratios ( Supplementary Fig. 3a), the configurational entropy (ΔS mix ) of the AuFeCoCuCr NPs was calculated to be 13.3 J mol -1 K -1 , which is classified as HEA 12 with a face-centred cubic structure ( Supplementary Fig. 3b). The Cl element mainly distributed on the NP surface (Fig. 1d), which indicated Cl was excluded during the formation of the HEA NPs. The residue chlorine could be removed by washing the NPs in ethanol to obtain highly purified HEA NPs. Besides hexane, LSA was also successfully applied in other alkanes, such as octane, decane and dodecane ( Supplementary Fig. 4), which suggests the broad liquid-phase applicability of the LSA strategy.
We used a model of photothermal evaporation to understand the formation mechanism of HEA NPs (Supplementary Section 1 and Supplementary Fig. 5). HEA NP synthesis by the LSA entails a number of physicochemical processes (Fig. 2a) 13 . The laser pulse penetrates through the alkane (liquid phase) to the substrate surface loaded with precursors ( Fig. 2a(i)). As a consequence of the intense optical field and high transient temperature in the focus, the precursors are prone to melt or decomposition after irradiation. A high-temperature and high-pressure plume that comprises ions, atoms, clusters and vapours forms at the substrate/liquid interface ( Fig. 2a(ii)). This multielement mixture in the plume enables the formation of solid-solution phases as a result of the large TΔS mix . Recoil from the plume generates emission shockwaves 13 , which cause an ultrasonic expansion of the plume into the surrounding liquid ( Fig. 2a(iii)). The plume rapidly cools down and releases energy to the alkane (Fig. 2a(iii)), which results in HEA nucleation and coalescence of the crystal nucleus ( Fig. 2a(iv)). Meanwhile, the energy released induces the generation of high-pressure and high-temperature bubbles, which expand into the liquid (Supplementary Video 1). During the expansion, the temperature and pressure inside the cavitation bubble reduce rapidly and recover to the original values when the bubble collapses ( Fig. 2a(v)). As a consequence, a large amount of energy is released alongside the formation of the HEA NPs. As the subsequent cooling down period is driven by the ultrafast heat exchange between the ablated area and the surrounding liquid phase, the solid solution phase is maintained without the formation of intermetallic phases.
In this regime, despite different properties (for example, atomic radius, electronegativity and reduction potentials) of the metallic elements (Supplementary Section 2 and Supplementary Tables 2 and 6), a nanosecond metal reduction and alloying process in LSA allows the synthesis of HEA NPs with thermodynamically forbidden compositions and uniform elemental distributions. In the case of AuFeCoCuCr NPs (Fig. 1), the ΔS mix of the Au, Fe, Co, Cu and Cr atoms increases from 0 to 13.3 J mol -1 K -1 after LSA (Fig. 2b). The LSA method also imparted the convenient synthesis of more complex HEA NPs, such as the elements Pt, Au, Pd, Cu, Cr, Sn, Fe, Co and Ni with a face-centred cubic solid-solution phase structure and a calculated ΔS mix of 17.4 J mol -1 K -1 (Fig. 2c,d and Supplementary Figs. 7-10). No element segregation was found even at the atomic scale (Fig. 2c), which represents direct evidence in support of the homogeneous element distribution within NPs.
To control the size and compositional distribution of HEA NPs, we proposed strategies to change the laser scanning times (Fig. 3) and adjust the ablation temperature ( Supplementary Fig.  11). A broader size distribution of HEA NPs was seen on repeated laser scanning. At the first laser scanning, the HEA NPs formed via a direct HEA nucleation from the metal precursors (Fig. 2a). When these NPs were exposed to the subsequent laser pulse, two processes were involved 14 . In addition to the direct HEA nucleation from the metal salts, the other process invoked a continuous enlargement of the existing NPs via fusion with the next incoming species (Fig. 3d). The variation of the NP size and distribution further substantiates that the LSA mechanism in this work follows the photothermal evaporation model over the Coulombic explosion one, as the latter would generate smaller particles with the increasing laser scanning cycles 15,16 . A lower hexane temperature, which provided a lower temperature and pressure environment inside the cavitation bubble, yielded a smaller bubble radius during LSA process. Such a bubble setting reduces metal evaporation loss and thus renders better compositional control ( Supplementary Fig. 11d).
Besides CNFs, we also tried a series of carbon substrates, such as graphene ( Supplementary Fig. 12), carbon nanotubes ( Supplementary Fig. 13) and even carbonized wood ( Supplementary  Fig. 14). HEA particles with sizes from nano-to micrometres were loaded uniformly on these substrates with a high yield and great quality, which reveals that LSA is an effective strategy for large-scale nano-and/or micromanufacturing of HEA particles onto any carbon substrates.
HEA NPs on substrates other than carbon. As a laser pulse can precisely confine energy to the desired microregions without altering the bulk properties of substrates 15 , we then used the LSA method to craft HEA NPs onto substrates other than carbon, such as copper foam ( Supplementary Fig. 15) and glass ( Supplementary Fig. 16). HEA NPs formed on these substrate surfaces, and each element was evenly distributed in the particles, which demonstrates the broad substrate applicability of LSA.
Notably, the size of the HEA NPs under one laser pulse of the same power density was distinct when they were deposited onto different substrates (~11 nm for CNFs (Fig. 3a) and ~700 nm for copper foam ( Supplementary Fig. 15)). The distinction of particle sizes should be mainly caused by different thermal diffusion lengths of substrates, interpreted by the equation L T = √ Dτ (refs. 17  The synthesis of HEA NPs involves the sequential processes of pulse penetration (i), pulse absorption (ii), plume expansion (iii), bubble generation (iv) and bubble collapse (v) during LSA. b, The configurational entropy evolution of HEA NPs with five different elements (labeled from A to E) during the LSA process. The ΔS mix of the Au, Fe, Co, Cu and Cr atoms increased from 0 to 13.3 J mol -1 K -1 after LSA, which indicates the formation of HEA NPs. c, Atomic-scale high-angle annular dark field-scanning transmission electron microscopy (STEM) images and STEM elemental maps for a novenary HEA NP (PtAuPdCuCrSnFeCoNi). The images compare the local concentration distribution of the individual elements for the same region. No element segregation was found even at the atomic scale, which represents the homogeneous element distribution within the NPs. d, X-ray diffraction pattern of the novenary HEA NPs, which implies the face-centred cubic structure of the HEA NPs. a.u., arbitrary units. width). During the HEA solidification ( Fig. 2a(iii),(iv)), the plume went through a transient melting phase before the formation of the HEA. The melting phase flowed along the substrate surface from the hot central region to the rear cool regions (Marangoni effect 19 ) to form melting pools 20 . This process was directly influenced by the thermal diffusion length of the substrates. Carbon materials (that is, graphene, carbon nanotubes and CNFs) have a small thermal diffusivity (~0.1-7.6 × 10 -6 m 2 s -1 ) (refs. [21][22][23], with a small thermal diffusion length (~22-195 nm). All the optical energy was absorbed at the surface without significant thermal diffusion ( Supplementary  Fig. 17). In contrast, a copper substrate has a larger thermal diffusivity (~1 × 10 -4 m 2 s -1 ) and thermal diffusion length (~710 nm) (ref. 24 ). A higher thermal diffusion length leads to a larger area of melting pools. Thus, larger HEA NPs with more melting pool marks were found on the copper foam surface ( Supplementary  Fig. 15). For the glass substrate, a large portion of the laser beam transmitted through the transparent surface, and had no further interaction with the metal precursors. Therefore, few HEA NPs were created on the glass surface.
Colloidal HEA NPs. Colloidal NPs with uniform dimensions have been recognized as unique building blocks for many applications, which include terahertz radiation devices, sensors and catalysts 25 . In this context, we synthesized colloidal HEA NPs by directly irradiating the salt precursors in the presence of oleic acid. Oleic acid functions as a nanodroplet template with an active interface for the HEA NP formation. Furthermore, the presence of oleic acid suppressed the aggregation of HEA NPs due to the effect of steric hindrance. Owing to a high affinity between oleic acid and metals 26 , colloidal HEA NPs of PtIrCuNiCr were successfully produced with a good dispersion and narrow size distribution (Supplementary Fig. 18). These results further demonstrate the wide applicability of the LSA technique. Synthesis of HEC NPs. As the LSA method is a non-equilibrium route (rapid heating and cooling), which can decrease the kinetic barriers to synthesize HEM NPs, we tried to apply LSA to the synthesis of HEC NPs, such as high-entropy sulfides (HES) and highentropy oxides (HEOs) (Fig. 4a and Supplementary Figs. 19 and  20). Amorphous HES (Fig. 4b) and HEO NPs with a homogeneous dispersion of sulfur and oxygen elements (Fig. 4c) were crafted after LSA. To prove generality, we also successfully created other HEC NPs, which included high-entropy borides, phosphides and nitrides by the LSA method ( Fig. 4 and Supplementary Figs. 21 and 22), all of which show amorphous phases. The effects of lattice distortions, a high mixing entropy and sluggish diffusion contribute to the formation of amorphous HEC NPs. The incorporation of non-metallic atoms reduces the crystallite growth rate and expands the atomic space, which leads to lattice distortion. The large lattice distortion results in a high strain energy and in turn raises the overall free energy 27 . Consequently, an amorphous structure with less coordination easily forms to relax the lattice distortion 28 . The high entropy effect enhances the mutual solubility of different atoms in the amorphous structure of HEC NPs, whereas the sluggish diffusion of high-entropy materials imposes a reduced atom mobility, impeding the yield of a viable crystal structure 29 . Both effects (that is, high entropy and sluggish diffusion) synergistically impart the formation of amorphous HEC NPs. In addition, the rapid cooling rate during the LSA process triggers insufficient time and energy for crystallization, which also contributes to their amorphization.   Electrocatalytic water splitting. We demonstrated HEM NPs loaded on graphene as catalysts for electrocatalytic water splitting. HEM NPs are expected to decrease the loading of noble metals without losing their electrocatalytic efficiencies, as noble metals can be highly dispersed in the HEA catalysts. In this work, we chose the low-amount noble metal HEA NPs of PtIrCuNiCr and PtAuPdFeNi as the catalysts for the hydrogen-evolving reaction (HER) and the oxygen-evolving reaction (OER) (Fig. 5 and Supplementary  Fig. 23). As shown in Fig. 5b, the OER requires an overpotentials of only 176 mV for PtIrCuCrNi and 178 mV for PtAuPdFeNi to deliver 10 mA cm -2 , much less than that of the state-of-the-art Ir/C catalyst (248 mV). For HER (Fig. 5c), the PtIrCuNiCr catalyst exhibits remarkable catalytic activity with an onset overpotential of ~0 mV and potentials significantly lower than that of commercial Pt/C at large current densities. The mass activity ( Supplementary  Fig. 24) of PtIrCuNiCr reaches 8.56 A mg -1 at -0.5 V for HER and 6.08 A mg -1 at 1.6 V for OER, which are 10.8-and 11.9-fold higher than those of Pt/C (0.79 A mg -1 for HER) and Ir/C (0.51 A mg -1 for OER), respectively. Given the outstanding OER and HER activities of the PtIrCuCrNi NP, we further utilized it as bifunctional electrocatalysts for overall water splitting. Remarkably, a current density of 10 mA cm −2 was delivered at a combined overpotential of about 185 mV (Fig. 5d), superior to those of most previously reported bifunctional electrocatalysts (Supplementary Table 3). The HEM electrode presented an excellent durability with no noticeable potential augment for more than 72 hours of oxygen release at 200 mA cm −2 (Fig. 5e). In addition, the atom ratios and the characteristic peaks of PtIrCuNiCr did not change significantly before and after the electrocatalytic reaction ( Supplementary Fig. 25), which revealed the outstanding stability of the HEA catalyst for overall water splitting.
To elucidate the mechanism for the high electrocatalytic activity of PtIrCuNiCr, X-ray absorption near-edge spectra of Pt and Ir in the HEA NPs were investigated (Supplementary Fig. 26). The characteristic peaks of the Pt L 3 edge in the metal and HEA NPs agree well with each other, which signifies a similar crystal geometry of the Pt atoms in Pt metal and alloys. The increased white line intensity of the Ir L 3 edge suggests more d-band vacancy induced by the hybridization of the 5d 5/2 state with empty states above the Fermi level on alloying 30 . The unoccupied Ir 5d orbitals can drive the formation of Ir-OOH species, which has been identified as a key intermediate to create molecular oxygen, due to the O 2p-Ir 5d hybridization 31 . Extended X-ray absorption fine structure (EXAFS) spectroscopy provides further structural information about the HEA NPs. The bond length and coordination number of each bond type are summarized in Supplementary Table 4. The available range of Ir data in HEA is rather limited as its signal at a high energy (>11,500 eV) is submerged by the Pt signal. Bonds of the 3d orbital for Cr, Ni and Cu cannot be distinguished from each other as their X-ray scattering is almost the same due to their similar atomic radii. Thus, these bonds are treated as the same during the fitting process. The derived metal-metal bond length in HEA NPs is clearly different from that in pure foils, which indicates the formation of an alloy structure. Compared with the Pt-Pt bond in the foil, the Pt-Pt and Pt-Cu/Ni/Cr bonds in HEA NPs lengthen and shorten, respectively, which results in severe lattice distortion with tensile and compressive strains 32 . The lattice strain further induces a thermodynamically nonequilibrium state, which contributes to a higher potential free energy, and thereby decreases the energy barrier during the catalytic reaction and improves the catalytic performance 33 . In addition, the coordination numbers of the Pt and Ir atoms in HEA NPs are much lower than those in the pure metal foils. The unsaturated coordination around the Pt and Ir sites can increase the initial potential energy of the HEA NPs, which leads to a greatly decreased activation energy barrier 34 . During the electrocatalytic water splitting, these unsaturated metal sites can facilitate the interactions with intermediates, and thereby improve the OER and HER reactions.

Conclusions
We present a simple, general and scalable LSA route to creating HEA and HEC NPs under mild conditions with a broad substrate applicability. Through this strategy, the HEM nanostructure library is substantially advanced towards a higher compositional diversity and structural complexity. As both anode and cathode, the HEA NPs of PtIrCuNiCr manifest a low cell voltage of only 1.42 V at a current density of 10 mA cm -2 in a 1 M KOH electrolyte, and outperform most previously reported bifunctional electrocatalysts. The superior electrocatalytic performance of HEM NPs can be attributed to the severe lattice distortion and unsaturated coordination of the materials. Going beyond electrocatalysis, we envision that by judiciously designing HEM morphology and screening their compositions, HEM NPs may provide tremendous potential for applications in biomedicine, catalysis, energy storage and sensors.

Methods
Synthesis of carbon nanofibres. The electrospinning solution was prepared by dissolving 0.8 g of electrospun polyacrylonitrile fibres (purchased from Macklin) in 10 ml of dimethylformamide at 80 °C for 12 h in a water bath to form a homogeneous solution. The precursor solution was transferred into a syringe before electrospinning at a voltage of 15 kV, a spinning distance of 10 cm and a feed rate of 0.25 ml h -1 . The electrospun fibres were collected on a drum. The polyacrylonitrile nanofibres were first stabilized at 533 K for 5 h in air and then carbonized at 1,073 K for 2 h in nitrogen.

Synthesis of graphene.
Typically, 4 mmol d-glucose, 6 mmol NH 4 Cl and 80 g of metal chloride salts (KCl/NaCl = 51/49 by weight) were thoroughly mixed by a ball-milling treatment, followed by drying at 150 °C for 8 h. Then, the prereacted brown mixture was transferred to a porcelain crucible and pyrolysed at 1,050 °C under N 2 for 1 h with a heating rate of 35 °C min -1 . After naturally cooling to ambient temperature, the blackened products were thoroughly ultrasonically rinsed with distilled water and ethanol several times.
Precursor loading on substrates. Various chloride salts were mixed in ethanol at 0.01 M for each metallic element. The mixed solution was directly dropped onto the wood, glass, copper foam or CNF film with a loading of ~1 ml cm -2 or onto graphene or a nanotube with a loading of ~0.1 ml mg -1 . Then, the loaded substrates were transferred into a vacuum oven for drying at room temperature.
Laser parameters. A pulsed fibre laser (SLT-PTM-100, Jiangsu Yanchang Sunlaite New Energy Co. Ltd) with a fundamental wavelength of 1,064 nm, energy of 14.15 J cm -2 and pulse length of 5 ns was used for the ablation. The laser was operated at a repetition rate of 20 kHz with a travelling speed of 600 mm s -1 . A Gaussian laser beam with a peak power of 20 kW was focused by a set of optical components and supplied at normal incidence to the surface of the substrate, as shown in Fig. 1c. Unless otherwise stated, the laser power used during LSA is set to 20%. Supplementary Fig. 1 depicts the laser scanning strategy. The laser scanning path was set to be back and forth along the x axis and forth along the y axis, which involves two kinds of overlap rates (that is, the scan line overlap and the spot overlap). Based on the scanning speed and the size of the laser spot (30 μm), we calculated the scan line overlap and O S of 0 μm. Namely, the laser spots continuously hit the substrate surface without overlapping. Each point on the substrate received only one laser spot (5 ns) per laser scanning cycle.
Synthesis of HEA NPs on substrates via the LSA method. HEA NPs on the surface of substrates were prepared by pulsed laser ablation of the corresponding precursor-loaded substrates (for example, CNFs, wood, glass or copper foam). The substrate was placed at the bottom of a glass vessel filled with a liquid alkane (for example, hexane, octane, decane or dodecane). The substrates were kept 10 mm below the surface of the liquid alkane. Hexane was the ideal choice because it could keep the precursors on the substrates prior to laser ablation due to the immiscibility between the hexane and the precursors. The oxygen-free structure is beneficial to keep the as-prepared HEA NPs from being oxidized. (It is important to note that to reduce the element loss caused by vapour pressure, a liquid-phase environment in LSA is more favourable than a gas-phase environment.) The substrates were scanned by a pulse laser for several cycles.
For the substrates of graphene and carbon nanotube in a powdered state, we first loaded the metal precursors onto the substrates and then dispersed the modified substrates in hexane at 0.5 mg ml -1 by magnetic stirring. Then, the solution was irradiated under agitation with a fibre laser for 30 min, ensuring that all the substrates interacted with the laser beam.

Synthesis of HEC NPs on substrates via the LSA method.
To synthesize the HES, the salt-precursor-loaded substrates were dipped into a CS 2 solution with the sulfur at 0.1 M to allow a uniform sulfur layer to deposit on the surface of the substrate before laser irradiation. Then, the substrates were transferred into hexane for laser irradiation. Argon gas was continuously passed into hexane during the experiment. The laser parameters used are the same as those used in the preparation of HEA NPs.
To synthesize HEOs, the salt-precursor-loaded substrates were treated by a 0.1 M NaOH solution to convert the metal chlorides into metal hydroxides on the substrate, followed by washing with distilled water to remove the by-product NaCl prior to laser irradiation. Then, the substrates were transferred into hexane for laser irradiation. The laser parameters used are the same as those used in the preparation of HEA NPs.
To synthesize high-entropy borides and nitrides, the salt-precursor-loaded substrates were dipped into 0.1 M sodium borohydride and 0.1 M ammonium chloride aqueous solutions, respectively, before the laser irradiation. Then, the substrates were transferred into hexane for laser irradiation. Argon gas was continuously passed into hexane during the experiment. The laser parameters used are the same as those used in the preparation of HEA NPs.
To synthesize high-entropy phosphides, the salt-precursor-loaded substrates were dipped into a 0.1 M triphenyl phosphine ethanol solution before the laser irradiation. Argon gas was continuously passed into hexane during the experiment. Then, the substrates were transferred into hexane for laser irradiation. The laser parameters used are the same as those used in the preparation of HEA NPs.

Synthesis of HEM NPs without substrates via the LSA method.
Various chloride salts were mixed in ethanol at 0.05 M for each metallic element. The mixed solution (2 ml) was directly dropped into hexane solution (20 ml) that contained oleic acid at 0.05% v/v. Metal salt precursors were located at the bottom of the solution due to their low solubility in hexane, forming a target-like salt layer. The mixture was transferred for laser irradiation. The salt layer was irradiated with the same laser parameters as those used in the preparation of HEA NPs. Finally, the liquid supernatant was collected, washed and centrifugated in methyl acetate (to remove oleic acid) and hexane successively.

Synthesis of high entropy hydroxide via a co-precipitation process.
An aqueous solution (25 ml) that contained metal ions (that is, CuCl 2 , CrCl 3 , FeCl 3 , CoCl 2 and NiCl 2 ) at 0.4 M was added into 50 ml of deionized water under magnetic stirring at 70 °C. Simultaneously, 1 M NaOH was added dropwise into the solution to maintain a pH of around 10. After 8 h, the product was collected by centrifugation, and washed with water and ethanol several times.
Electrocatalytic experiments for water splitting. The as-prepared HEA NPs on graphene, as well as the Pt/C and Ir/C samples, were prepared by ultrasonically mixing 10 mg of the catalyst powder with a mixture of 400 μl of ethanol and 50 μl of 5% Nafion solution for 20 min to form homogeneous inks. For the preparation of the catalytic electrodes, 50 μl of the ink was carefully dropped onto nickel foam (0.5 × 0.5 cm 2 ), which resulted in a HEA NP/graphene loading of 4.4 mg cm −2 on the electrode. Based on the results of inductively coupled plasma mass spectroscopy, the HEA NP loading on graphene was approximately 5.4 wt% with 2.2 wt% Pt and 1.8 wt% Ir. In other words, the actual HEA NP loading was approximately 0.24 mg cm -2 with 0.10 mg cm -2 Pt and 0.08 mg cm -2 Ir loadings. For commercial Pt/C (20%) and Ir/C (5%) catalysts, the Pt and Ir loadings were 0.88 mg cm -2 and 0.22 mg cm -2 , respectively. The electrocatalytic electrode was dried at room temperature naturally.
Electrocatalytic experiments were performed on a CHI 660D electrochemical analyser (CH Instruments, Inc.) with a three-electrode cell system for the OER and the HER. The catalyst ink-loaded nickel foam was used as the working electrode, a Ag/AgCl (saturated KCl) electrode as the reference electrode and a carbon rod as the counter electrode. All the electrochemical experiments were conducted in a 1.0 M KOH aqueous solution at room temperature. All the potentials for OER and HER reported here are referenced to the reversible hydrogen electrode.
The overall water splitting was performed in a two-electrode system with a 1.0 M KOH electrolyte, in which one PtIrCuNiCr-nickel foam electrode served as the negative electrode for the HER and another one acted as the positive electrode for the OER. The durability was assessed at an applied potential to reach a catalytic current density of 200 mA cm −2 for 72 h.
Characterizations. The morphology of the as-prepared samples was examined by field emission scanning electron microscopy (FEI NOVA NanoSEM230) and TEM (Tacnai G2 F20, FEI). The field emission scanning electron microscopy was equipped with an EDS (Oxford Instruments X-Max) and used to obtain the element distribution of the HEA NPs on the wood or nickel foam. The TEM was equipped with an EDS (Elite T EDS System) and employed to record the element distribution of the HEA NPs on graphene, CNTs and CNFs. High-angle annular dark field-STEM analysis (JEM Titan G2 60-300 TEM (FEI)) was used for the characterization at 60-300 kV. To prepare the TEM and STEM specimens, the synthesized samples were dispersed in ethanol by ultrasonic treatment and then transferred onto microgrids. In this process, some HEA NPs are inevitably separated from the substrates. The crystal structures of the samples were measured by a powder X-ray diffractometer (Ultima III, Rigaku Corp.) using Cu Kα radiation (λ = 1.54178 Å, 40 kV, 40 mA). The atomic ratios of the HEM NPs were analysed by a PerkinElmer AVIO500 ICP-MS. The solutions were prepared by digesting the samples in aqua regia followed by dilution with 2% hydrochloric acid. The surface composition of the samples was examined with X-ray photoelectron spectroscopy (XPS, ESCALAB 250) with non-monochromatic Al Kα X-ray as the X-ray source. The binding energy of C 1s (284.6 eV) was used to calibrate the other binding energies. The ultraviolet-visible diffuse reflectance spectra were measured on a PerkinElmer Lambda 950 UV/Vis/NIR spectrometer. Thermogravimetric analysis was measured using a NETZSCH STA 449F3 analyser.

Data availability
All the supporting data are provided in the main text and Supplementary Information. Source data are provided with this paper.