Research progress of amorphous catalysts in the field of electrocatalysis

Noble metal-based amorphous catalysts

Precious metal-based catalysts (such as Pt-based, Ir-based, Ru-based, etc.) are widely used as electrocatalysts; however, their high cost limits their large-scale applications. Optimizing their inherent activity, increasing active sites, and improving operational stability are urgent. Amorphization endows noble metal-based amorphous catalysts with unique surface and active site distribution, improving reaction efficiency and selectivity, thus receiving increasing attention in electrocatalysis. For example, excellent acidic OER performances were measured in amorphous IrO_x (a-IrO_x) thin film catalysts rather than crystalline IrO_x thin films^[50], whose overpotentials at 1 and 10 mA cm^-2 were 190 and 220 mV, respectively. Moreover, Pd₂RuO_x materials^[51] were synthesized by confining atomically thin Pd-PdO nanodomains in amorphous metallic ruthenium oxide (RuO₂) NSs and then performing a reduction treatment [Figure 2A], which exhibited excellent electrocatalytic activity in electrolyzing water. The presence of Pd-PdO heterogeneous interfaces and numerous surface defects in materials facilitates active substance adsorption and thus promotes the electrocatalytic reaction [Figure 2B]. Na-modified defect-rich crystalline/amorphous rutile structure RuO₂ (a/c-RuO₂) catalysts delivered excellent OER catalytic stability [Figure 2C], benefiting from the abundant defects, grain boundaries, and accessibility of active sites^[52].

Figure 2. (A) The synthesis of PdRuO_x materials; (B) HRTEM image of Pd₂RuO_x-0.5 h. (Reproduced with permission^[51]. Copyright 2023, WILEY-VCH). (C) Chronopotentiometry measurement results for a/c-RuO₂ and commercial RuO₂ at 10 mA cm^-2 in 0.1 M HClO₄. (Reproduced with permission^[52]. Copyright 2021, WILEY-VCH). (D) High-magnification SEM images of Pt-a/c-NiCoHPi. (Reproduced with permission^[53]. Copyright 2023, WILEY-VCH ). (E) Schematic illustration of the size-controllable synthesis of a-PdCu nanoparticles. (F) FE of different products; (G) Gibbs free energy diagram of the EOR on a-PdCu and c-PdCu model catalysts in the C1 path. (Reproduced with permission^[54]. Copyright 2022, American Chemical Society).

Given the small reserves and expensive nature of precious metals, reducing their active loading without affecting the overall electrocatalytic efficiency is one of the cost-effective strategies. The Pt-a/c-NiCoHPi catalyst with crystalline/amorphous structure (Pt content is only 2.1%) exhibited excellent HER activity in alkaline solution, with an overpotential of only 19 mV at 10 mA cm^-2 [Figure 2D]^[53]. Wang et al. synthesized amorphous PdCu nanorods with controllable size and morphology [Figure 2E] and used them for ethanol electrooxidation reaction^[54]. The prepared amorphous PdCu nanorods delivered a high Faraday efficiency (FE) of 69.6% for C1 products, which was much higher than that of the corresponding crystalline catalysts and the commercial Pd/C [Figure 2F]. Density-functional-theory (DFT) calculations further revealed that the amorphous structure can significantly reduce the energy barrier for C-C bond breaking on the surface of the PdCu catalysts and promote the kinetic process of the ethanol oxidation reaction [Figure 2G]. Despite the remarkable progress of noble metal-based amorphous catalysts in the field of electrocatalysis, there are still challenges and issues, such as the high cost and stability of noble metal catalysts. Therefore, the development of amorphous catalyst materials with economic benefits and high reactivity is one of the hot spots for future research.

Transition metal-based amorphous catalysts

Thanks to the abundant earth reserves and cost-effectiveness, amorphous catalysts composed of non-noble transition metal elements have been widely studied in electrocatalysis. For instance, Jiang et al.^[55] prepared self-supported FeNiCo-based amorphous catalysts with hierarchical micro/nanopore structures by dealloying amorphous/nanocrystalline precursors [Figure 3A]. The optimized electrocatalyst showed a HER overpotential of 134 mV and an OER overpotential of 206 mV at 10 mA cm^-2 in alkaline solution. The main reason is that amorphous nanoporous frameworks can provide fast reaction kinetics, abundant active sites, and enhanced electrochemically active surface area. The synergistic effect of the structural and compositional advantages of the catalysts finally showed excellent catalytic activity.

Figure 3. (A) Schematic diagram of self-supported hierarchical porous amorphous FeNiCo-based bifunctional catalysts. (Reproduced with permission^[55]. Copyright 2021, Elsevier). (B) TEM and (C) HRTEM image of MoS_x@NCNFs-30 films. (Reproduced with permission^[56]. Copyright 2016, Elsevier). (D) HRTEM diagram of amorphous CoO_x nanoparticles (The insert shows its SAED patterns); (E-G) Free energy diagram for the water oxidation on Co₃O₄, CoOOH, and 1D-CoOx surfaces. (Reproduced with permission^[57]. Copyright 2017, WILEY-VCH). (H) TEM image and SAED pattern (inset) of CoB@TiO₂ nanosheets; (I) NH₃ yields and FE at given potentials. (Reproduced with permission^[59]. Copyright 2022, CCS).

Novel amorphous flower-like molybdenum sulfide@nitrogen doped carbon nanofiber (MoS_x@NCNFs) thin films were synthesized and used to reduce water^[56]. MoS_x NSs immobilized on NCNFs in a unique flower-like structure were detected in the transmission electron microscope (TEM) image [Figure 3B], while the amorphous structure of MoS_x was confirmed by selected area electron diffraction (SAED) (inset of Figure 3B) and Figure 3C. The MoS_x@NCNFs exhibited excellent HER catalytic activity, with an overpotential of 137 mV at 10 mA cm^-2. Amorphous cobalt oxide nanoparticles (ca. 2 nm) acted as water oxidation catalysts with a turnover frequency of up to 8.6 s^-1 in the Ru(bpy)₃²⁺-Na₂S₂O₈ system [Figure 3D], which is superior to that of the most active homogeneous cobalt-based molecules^[57]. Meanwhile, theoretical calculations showed that the obtained structure promotes O-O bond coupling and improves the water oxidation activity compared with crystalline cobalt oxide [Figure 3E-G], further confirming the good activity of amorphous cobalt oxide nanoparticles. Pang et al. used a simple electrochemical dealloying method to prepare a self-supporting amorphous NiFeCoP catalyst with a nanoporous structure (np-NiFeCoP)^[58]. The np-NiFeCoP catalyst exhibited good electrocatalytic activity from its unique structural characteristics, which can expose more active sites and exert synergistic effects between metal elements. In 1.0 M KOH alkaline solution, the overpotentials of HER and OER were 105 and 244 mV for np-NiFeCoP catalyst at 10 mA cm^-2, respectively. Amorphous transition metal boride (TMB) nanomaterials featured numerous defects and dangling bonds, thus providing abundant active sites for the electrocatalytic nitrite reduction reaction (NO₂^-RR). CoB@TiO₂/Ti plate (TP) was prepared by sputtering an amorphous cobalt boride nanofilm on the surface using titanium dioxide (TiO₂) NS arrays as a substrate by Hu et al.[Figure 3H]^[59]. The prepared CoB@TiO₂/TP showed a high NH₃ yield of up to 233.1 μmol h^-1 cm^-2 at -0.7 V and a FE of 95.2% for NO₂^-RR, superior to most reported NO₂^-RR catalysts [Figure 3I].

As one kind of emerging catalyst, transition metal-based amorphous catalysts present many unique properties and application potential. However, its industrialization and commercial applications still face several challenges, such as production cost, catalyst stability, and scale-up preparation. Therefore, future development requires further strengthening of the industrialization technology research and economic feasibility analysis.

Other types of amorphous catalysts

Recently, HEAs and high-entropy oxides (HEOs) composed of five or more elements have become a research hotspot in the field of catalysts due to their tunable electronic structure and wide range of component selection^[60]. Among them, the design of multi-component high-entropy catalysts helps to improve the catalytic activity and stability of the catalysts. In addition, amorphous high-entropy catalysts have disordered local structures, which can provide more reaction active sites and exhibit superior catalytic performance. For example, Wang et al. prepared nanostructured amorphous Fe₂₉Co₂₇Ni₂₃Si₉B₁₂ HEA strips and used them as electrocatalysts for OER [Figure 4A]^[61]. Thanks to their highly disordered configuration, they exhibited excellent catalytic activity in electrocatalytic activity tests, with an overpotential of only 230 mV at a current density of 10 mA cm^-2 [Figure 4B]. The amorphous FeCoNiPB HEOs prepared by a chemical reduction strategy were reported by Wang et al.^[62]. When OER testing was performed in alkaline media, its catalytic performance was superior to quaternary metal oxides such as FeCoPB and the commercial catalyst RuO₂. This can be attributed to its unique amorphous structure and the synergistic effect of multiple components, making it capable of exerting excellent catalytic performance. In order to activate the surface atoms of HEOs, Li et al. constructed low-crystallinity (Fe, Co, Ni, Mn, Zn)₃O₄ (HEO-Origin) NSs with abundant oxygen vacancies [Figure 4C], whose TEM images and the corresponding SAED images showed their amorphous nature [Figure 4D and E]^[63]. Due to its large specific surface area, high activity of surface atoms and the presence of oxygen vacancies, HEO-Origin exhibits excellent OER activity.

Figure 4. (A) XRD patterns of various samples. (B) LSV polarization curves of different catalyst electrodes for the OER in 1.0 M KOH. (Reproduced with permission^[61]. Copyright 2022, Springer). (C) Schematic diagram of HEO-Origin preparation. (D) TEM image and (E) SAED patterns of HEO-Origin. (Reproduced with permission^[63]. Copyright 2023, Elsevier). (F) Schematic diagram of the synthesis of La₂O₃@NP-NS. (G) SAED pattern of La₂O₃@NP-NS. (Reproduced with permission^[64]. Copyright 2020, Springer). (H) Overpotential bar graph and (I) Tafel plot of different catalyst electrodes. (Reproduced with permission^[49]. Copyright 2023, Wiley-VCH GmbH).

Rare earth elements have more unpaired electrons and irregular electron arrangement, which makes rare earth catalysts have high catalytic activity. Rare earth amorphous catalysts are an attractive class of catalysts with abundant active sites and defects in their structures, which provide more reaction sites and further enhance their catalytic activity. Yan et al. reported rare earth metal oxide electrocatalysts composed of ultrathin amorphous La₂O₃ NSs hybridized with uniform La₂O₃ nanoparticles [Figure 4F]^[64]. As shown in Figure 4G, the SAED pattern of the NS surfaces showed broad and diffuse diffraction rings, indicating that La₂O₃@ NP-NS indeed has a large number of amorphous phases with small crystals. Among them, the La₂O₃@NP-NS catalyst with a NS thickness of 2.27 nm exhibited excellent OER activity with an overpotential of 310 mV at 10 mA cm^-2. This can be attributed to the ultra-thin thickness, which shortens the charge diffusion path and promotes electron transfer and enhanced reaction kinetics. For the amorphous Ni-Nb-Y alloy, the catalytic activity of HER can be optimized by controlling the concentration and distribution of Y in the ternary alloy system^[65]. Among them, the Ni_81.3Nb_6.3Y_12.5 catalyst, in which the amorphous phase coexists with a small amount of crystalline phase, exhibits the highest HER activity. Sun et al. obtained amorphous Mn-Ce-O_x, Mn-La-O_x and Mn-Pr-O_x by doping Mn-based catalysts with rare earth metals (Ce, La or Pr). Among them, the Mn-Ce-O_x catalyst has the best oxidation performance^[66]. At a reaction temperature of 239 °C, the maximum oxidation conversion rate of NO is 94.0%. LCO-4 has been proven to have excellent OER activity [Figure 4H and I]^[49]. At a current density of 50 mA cm^-2, the overpotential is only 350 mV, and the extremely low Tafel slope (70 mV dec^-1) further highlights its remarkable activity.

Liquid phase method

The liquid phase method dissolves the desired atoms or molecules in a liquid solution and then forms amorphous catalysts; the synthesis process needs to control parameters such as temperature, concentration, pH, etc.^[67,68]. The atoms or molecules dissolve in the solution and combine to form an amorphous catalyst via chemical reactions or precipitations. This method features the merits of simple operation, easily controlled reaction conditions^[69], diverse catalyst morphology, etc., thus receiving extensive attention. The typical liquid phase method contains the sol-gel method, deposition method, hydrothermal method, etc.

(1) Sol-gel method: The metal alkoxides or metal salts in a solution undergo hydrolysis and condensation to form a sol, followed by gelation and drying to obtain a solid material. This method has the advantages of low-temperature processing, high degree homogeneity, and excellent controllable catalyst size and morphology. For example, Zhang et al. used a sol-gel method to prepare a Prussian blue analog (NiFe-PBA) precursor with a two-dimensional (2D) network and then further annealed it into a Fe₃O₄/NiC_x composite (NiFe-PBA-gel-cal), as a bifunctional electrocatalyst for seawater splitting [Figure 5A]^[70]. It can be seen from the high-resolution TEM (HRTEM) image of the calcined composite material that NiC_x contains both amorphous and crystalline phases [Figure 5B], which is beneficial to the catalytic activity in electrochemical tests. Furthermore, the amorphous Cr₈Co₁-300 catalyst from the sol-gel method was reported^[71], which exhibits a high specific surface area.

Figure 5. (A) Schematic diagram of the synthesis of NiFe-PBA-gel-cal. (B) HRTEM image of NiFe-PBA-gel-cal. (Reproduced with permission^[70]. Copyright 2022, Wiley-VCH). (C) Schematic illustration of CoFe-H structure and its application in electrochemical and photo-electrochemical water oxidation as an OER electrocatalyst. (D) TEM, and (E) HRTEM images of CoFe-H nanosheets (the insert is corresponding FFT). (Reproduced with permission^[72]. Copyright 2017, WILEY-VCH). (F) The fabrication illustration of Ru SAs-MoO_3-x/NF. (G) XRD patterns of Ru SAs-MoO_3-x/NF and Ru SAs-MoO₃/NF. (Reproduced with permission^[78]. Copyright 2023, WILEY-VCH).

(2) Deposition method: Depositing a metal precursor or its compound in a solution on a suitable carrier surface. This method includes the impregnation method, brush coating method, suspension deposition method, etc. Generally, amorphous metal alloys, carbon-based amorphous materials, amorphous nanocomposites, etc., can be prepared via the deposition method. For instance, amorphous cobalt iron hydroxide (CoFe-H) NSs were synthesized by electrodepositing CoFe-H NSs on graphite at 1.42 V [Figure 5C]^[72]. The characterizations of the TEM, HRTEM, and fast Fourier transform (FFT) reveal the amorphous character of CoFe-H NSs [Figure 5D and E]. Moreover, a novel pulsed electrodeposition method was used to dissolve commercially available anodic aluminum oxide (AAO) nano-molds in an alkaline solution that contained Ni-, Pd-, and P-ions^[73]. In the electrodeposition process, amorphous Ni-P and Ni-Pd-P nanowire structures were formed by controlling the pulsed currents and potentials. Furthermore, the electrodeposition method can quickly produce unique three-dimensional (3D) nano-dendrite forests of amorphous α-phase mixed nickel-cobalt hydroxides on stainless steel foil^[74].

(3) Hydrothermal method: A direct method to prepare amorphous catalysts with no need for complex machines or working conditions^[75,76]. Typically, the precursor only interacts with a solvent (often water) for a brief amount of time at a high temperature and high pressure to synthesize catalysts. The reaction conditions of high temperature and high pressure ensure the complete dispersion and dissolution of the solvent and precursor, benefiting the uniform reaction and the even distributed active sites. On the other hand, combining reaction temperature and time can tune the morphology, pore structure, and surface characteristics of amorphous catalysts, finally modifying catalytic performances. For example, the basic La³⁺ from the prepared bulk-phase LaNiO₃ chalcogenide oxides would selectively dissolve in FeCl₃ solution [reaction (1)]^[77]. With the consumption of H⁺, in the reaction process, part Fe³⁺ was hydrolyzed and deposited on the surface of the residual substrate [reaction (2)]. Noticeably, the La³⁺ in LaNiFe hydroxide stemmed from the residual substrate rather than the solution. The successive leaching of the large-sized La³⁺ ions induces the crystalline chalcogenide structure to gradually collapse and turn into an amorphous state, thus generating a novel amorphous Ni-Fe-based OER electrocatalyst. Moreover, the precursor grown on the surface of nickel foam (NF) by hydrothermal process was immersed in RuCl₃ solution and annealed in air to prepare the amorphous molybdenum-based oxide stabilized single-atomic-site (Ru-SAs-MoO_3-x/NF) catalysts [Figure 5F]^[78]. The X-ray diffraction (XRD) of the scraped powder from the Ru-SAs-MoO_3-x/NF surface indicates almost no characteristic peaks [Figure 5G], suggesting its amorphous nature.

Vapor phase method

Amorphous catalysts from a vapor phase approach need to combine atoms or molecules in a specific ratio in an inert atmosphere before causing them to react on the catalyst surface. This method presents significant merits in terms of high purity^[79], homogeneity^[80], controllability, and specific surface area, facilitating the synthesis of amorphous catalysts with complex components. Typically, vapor phase methods can be classified as the physical vapor phase deposition (PVD) and chemical vapor phase (CVD).

Physical Vapor Deposition (PVD): This method refers to depositing thin metal or metal compound films onto a substrate by a physical process, such as sputtering, evaporation, pulsed laser deposition, etc.^[81,82]. The attractiveness is the precise control of deposition conditions, allowing the formation of an amorphous film with unique structures and excellent catalytic properties. PVD shows outstanding advantages in terms of elaborately regulating the thickness and composition of amorphous films. For instance, a thin conformal amorphous IrO_x (a-IrO_x) coating was reported by Smith et al., which was produced at room temperature in an ambient atmosphere by the light-induced breakdown of a photoactive Ir(acac)₃ precursor placed on a conducting substrate [Figure 6A]^[50]. In the XRD pattern, there are no diffraction peaks related to the iridium oxide crystal phase, confirming the amorphous nature of a-IrO_x [Figure 6B]. Moreover, an oxygen-rich vacancy amorphous copper oxide nanofilm (m-CuO_x) catalyst was prepared by a vacuum evaporation method^[83], and its amorphous was indexed from the XRD spectrum.

Figure 6. (A) Schematic diagram of the preparation of amorphous IrO_x. (B) X-ray diffraction patterns for a-IrOx and c-IrOx on FTO. (Reproduced with permission^[50]. (Copyright 2014, American Chemical Society). (C) The general synthetic process of amorphous noble metal NSs. (D) TEM, and (E) Aberration-corrected HAADF-STEM image of amorphous Ir NSs (the inset in E shows the SAED pattern). (Reproduced with permission^[87]. Copyright 2019, Nature).

(2) Chemical Vapor Deposition (CVD): This method means depositing metal atoms on a substrate surface by breaking down the gas-phase precursor compound in an inert gas atmosphere^[84-86]. Generally, this method can be applied to amorphous metal oxides and carbon-based amorphous materials. For example, a variety of amorphous noble metal NSs less than 10 nm thickness was produced from direct annealing, such as amorphous Ir NSs, amorphous IrRu NSs, amorphous IrRhRu NSs, etc. [Figure 6C]^[87]. The annealed medium is a mixture of metal acetylpyruvates and alkali salts. As shown in Figure 6D and E, the amorphous feature of the prepared Ir NSs was confirmed by TEM and aberration-corrected High-Angle Annular Dark Field-Scanning Transmission Electron Microscopy (HAADF-STEM) characterizations.

Solid phase method

This method refers to preparing materials by using solid-state reactions^[88]. In the solid-state reactions for forming amorphous catalysts, the spread and assembly of atoms or molecules can be realized by solid-phase diffusion or physically excited diffusion and aggregation. The solid-phase method, known for its advantages of simple preparation, high controllability, and homogeneity, facilitates the synthesis of amorphous catalysts^[89]. Importantly, this method exhibits remarkable advantage in the synthesis of substances that have a high melting point or are unstable in solution. However, the solid-phase method faces several challenges of lower specific surface area, strong crystallinity, high influence from impurities, and strict preparation requirements. Conventionally, the solid phase method includes melt quenching, sputtering, and ball milling.

Melt-quenching method: The reactants first melt at high temperatures, followed by rapid quenching to room temperature or lower. The rapid cooling benefits to forming an amorphous structure rather than a long-range order structure^[90]. Commonly, this method is conducted to synthesize amorphous metallic alloys by rapidly solidifying the metal-molten mixture. The prepared amorphous alloys own a high degree of compositional flexibility. For instance, amorphous Al₈₀Pd₁₀Cu₅Ni₅ strips with a thickness of 30~50 μm and a width of 2~4 mm [Figure 7A] were reported by Yang et al.^[91]. In detail, an Al₈₀Pd₁₀Cu₅Ni₅ master alloy at Ar atmosphere was first prepared via arc melting. Subsequently, the master alloy in a quartz tube was remelted to a boiling state and sprayed onto a copper roll. Finally, the final amorphous nanoporous PdCuNi-S catalyst was formed by a two-step dealloying method. Moreover, amorphous Ni-Fe alloy strips with a thickness of 25 μm and a width of 2 mm were prepared by using the same synthesis method^[92]. The amorphous catalysts from a melting-dumping strip technique feature changed structure and play a better role in the relevant electrocatalytic reactions. This method features the merits of high-speed solidification, uniformity, controllability, and high production efficiency technology, thus receiving wide applications.

Figure 7. (A) Synthetic procedure of nanoporous PdCuNi-S catalyst. (Reproduced with permission^[91]. Copyright 2019, Elsevier). (B) Schematic diagram of depositing Ir₂₅Ni₃₃Ta₄₂ MG film on Si substrates. (C) HRTEM image and its corresponding SAED pattern (inset) of the as-deposited Ir₂₅Ni₃₃Ta₄₂ MG film. (Reproduced with permission^[93]. Copyright 2019, Wiley-VCH). (D) Schematic illustration of the two-stage ball milling synthesis process for producing amorphous nanoparticles. (E-H) Nanobeam diffraction (NBD) patterns at the surface regions in Ni_74.2Co₅Nb_12.5Y_8.3 nanoparticles. (Reproduced with permission^[95]. Copyright 2020, American Chemical Society).

Sputtering method: This method utilizes energetic particles to bombard the target material surface, resulting in the metal atoms dissociating from the target material and depositing onto the substrate. For example, an amorphous Ir₂₅Ni₃₃Ta₄₂ nano-thick metallic glass (MG) film with low Ir content was grown on a Si substrate by using ion beam deposition (IBD) [Figure 7B]^[93]. The amorphous structure of the deposited Ir₂₅Ni₃₃Ta₄₂ MG thin film was revealed by the HRTEM and related SAED characterizations [Figure 7C]. As a HER electrocatalyst, the obtained amorphous film exhibits high activity and stability in 0.5 M H₂SO₄ solution.

Ball milling method: A typical way to create amorphous catalysts and also known as mechanical alloying. This approach refers to conversing mechanical energy to the collision, friction, and shearing of solid feedstocks in a ball milling vessel; finally, the amorphous catalysts would be produced by mixing and reacting the feedstock powders at a tiny scale^[94]. For instance, a two-stage ball milling process was used to prepare the amorphous Ni_79.2-xCo_xNb_12.5Y_8.3 (x = 0 and 5 at%) OER catalysts^[95], and the precursor was the mixture of nickel (99.9 wt%, -100 mesh), cobalt (99.8 wt%, -100 mesh), niobium (99.99 wt%, -325 mesh), and yttrium (99.9 wt%, -40 mesh) metal element powders [Figure 7D]. The prepared amorphous catalysts deliver excellent structural stability [Figure 7E-H].

Based on the above description, we briefly summarize the preparation methods in Table 1.

Other novel methods

Based on the above analysis, the amorphous catalysts from most traditional methods face the issues of heterogeneity and uncontrollable texture. Hence, numerous novel preparation methods have been developed to obtain higher quality and stable amorphous catalysts. For example, the supersaturated precipitation method was applied to produce amorphous NiFeMo oxide with a mass of up to 515 g^[96]. This method first mixes high-concentration metal chloride precursors (such as NiCl₂·2H₂O and FeCl₃) in water. Then, 1 M Na₂MoO₄·2H₂O aqueous solution was added to the mixture at ultrasonic conditions and reaction for 2 min; the samples were obtained via freeze-drying [Figure 8A]. Moreover, homogeneously mixing the high-temperature calcination-wet chemical synthesized highly crystalline LaCoO₃ (LaCoO₃-Pristine) nanopowders with urea in the mass ratio of 1:10, then heating at 450 °C for 2 h and cooling to room temperature can obtain reduced LaCoO₃ powder (LaCoO₃-Reduce) [Figure 8B]^[97]. The XRD pattern indicates the amorphous state of LaCoO₃-Reduce after reducing by urea [Figure 8C]. The disordered structure of LaCoO₃-Reduce in Figure 8D further demonstrates the amorphous character. Additionally, combining the merits of each preparation method is a new strategy to prepare amorphous catalysts. For example, amorphous TiO₂ nanofibers (a-TiO₂) were produced by combining the preparation methods of precursor solution^[98], electrostatic spinning, and calcination [Figure 8E]. Moreover, combining hydrothermal with phosphating can synthesize MoO₂/Ni₃(PO₄)₂ self-supporting microcolumn electrocatalyst with amorphous MoOx phase at the NF surface (SA-MoO₂/Ni₃(PO₄)₂/NF)^[99].

Figure 8. (A) The synthesis illustration of a-NiFeMo catalyst. (Reproduced with permission^[96]. Copyright 2019, Wiley-VCH). (B) The preparation schematic of crystalline LaCoO₃ and amorphous LaCoO_x nanopowders. (C) XRD patterns of pristine and reduced LaCoO₃ nanopowders. (D) TEM image of reduced LaCoO₃ nanopowders. (Reproduced with permission^[97]. Copyright 2022, American Chemical Society). (E) The schematic diagram for the formation of lotus-root-like a-TiO₂ nanofibers and the atomic-scale deposition of Ga. (Reproduced with permission^[98]. Copyright 2022, American Chemical Society).

Experimental characterizations

A common characterization method of amorphous structures is XRD. The information on crystal structure and lattice parameters can be detected by the incident X-rays. However, the disordered crystal structure of amorphous induces steamed bun peaks occurring in XRD patterns^[100,101], thus making it challenging to provide detailed structural information. Meanwhile, various amorphous catalysts always exhibit similar XRD patterns, resulting in the subtle differences between them indistinguishable via XRD. The micromorphology and fine structure of amorphous catalysts can be observed by TEM. The absence of clear lattice fringes in HRTEM means the disordered and irregular atomic arrangement in amorphous catalysts. The unit cell parameters or crystal face indexes cannot be deduced from HRTEM, which is different from crystalline catalysts. The diffraction ring or fuzzy core spots emerging in SAED can further verify the amorphous feature of catalysts. Similar to the crystalline catalysts, the band energy spectrum analysis from TEM can determine various element distributions in amorphous catalysts. Moreover, Zhan et al. reported that nanobeam electron diffraction (NBED) image of (NbTiZr)_77.8Si_22.2 (at.%) coating showed a broad diffraction ring indicating its amorphous nature [Figure 9A]^[102].

Figure 9. (A) The NBED pattern of (NbTiZr)_77.8Si_22.2 (at.%) coating. (Reproduced with permission^[102]. Copyright 2022, Elsevier). (B) Cr K-edge XANES spectra and (C) Fourier-transforms of k3-weighted Cr K-edge EXAFS spectra in the R-space of H-CrO_x/C-550, MIL-101, commercial Cr₂O₃, and Cr foil. (D) Wavelet transforms for the k3-weighted EXAFS signals of MIL-101, H-CrO_x/C-550, and commercial Cr₂O₃. (Reproduced with permission^[104]. Copyright 2022, American Chemical Society). (E) FTIR spectrum and (F) Raman spectrum of amorphous Fe-Mo-O/NF. (Reproduced with permission^[105]. Copyright 2022, American Chemical Society). (G) N₂ adsorption-desorption isotherms of NG and Cu/Fe-NG. Copyright 2022. (Reproduced with permission^[106]. Copyright 2020, American Chemical Society).

As an important characterization, X-ray absorption near-edge structure (XANES) is derived from X-ray absorption spectroscopy (XAS), which can detect more information in amorphous catalysts. Generally, the elemental oxidized state, coordination number, and chemical environment can be analyzed by the peak shape, position, intensity, and other characteristics of the XANES spectrum. For instance, the valence states of Mn in the original and oxygen-vacant samples were revealed by XANES, and they range between 3+ and 4+^[103]. Compared with XANES, extended X-ray absorption fine structure (EXAFES) can provide more detailed information on interatomic distances and coordination numbers. The structural information on amorphous catalysts can be obtained by analyzing the background noise and vibrational characteristics in the EXAFES spectrum. For example, Pan et al. used XANES and EXAFES to explore the atomic coordination information in amorphous H-CrO_x/C-500 [Figure 9B and C]^[104]. The abundant coordination defects and oxygen vacancies in H-CrO_x/C-550 can be revealed by the Cr-O and Cr-Cr coordination bonds [Figure 9C], the fitting results and wavelet transforms (WT) of Cr K-edge EXAFS [Figure 9D]. Significantly, the intensity and oscillation amplitude decrease in XANES and EXAFS is associated with highly disordered/amorphous structures. The presence of functional groups and short-range ordered structures of amorphous catalysts can be identified by Fourier transform infrared spectroscopy (FTIR) and Raman spectroscopy, respectively. For instance, the peaks at 860 and 685 cm^-1 correspond to Fe(NiMo)-O [Figure 9E], while the peaks at 457, 529, and 664 cm^-1 are associated with the vibration of FeNi-O in Fe-Ni oxide [Figure 9F]^[105], demonstrating the formation of amorphous Fe-Mo-O/NF. The specific surface area of amorphous catalysts can be measured by the N₂ adsorption-desorption curve. As displayed in Figure 9G, the specific surface areas of Cu nanoclusters/FeN₄ composites deposited on N-doped graphene (Cu/Fe-NG) and NG catalysts are 303.7 and 542.6 m² g^-1[106], respectively. Extensive research has been done to experimentally characterize amorphous catalysts, but their detailed crystal structures are still lacking, resulting in the relationship between structure and performance remains unclear. Thus, developing advanced experimental characterizations to detect more information from amorphous catalysts is urgent. X-ray photoelectron spectroscopy (XPS) can characterize the chemical composition, surface chemical state and electronic structure of amorphous catalysts; monitor the surface changes during the reaction process and other aspects, which contributes to an in-depth understanding of amorphous catalysts in terms of their structure, surface reaction properties and catalytic performance. Electron energy loss spectroscopy (EELS) imaging, as a high-resolution, high-sensitivity, and non-destructive catalyst characterization technique, plays an important role in the study of the chemical composition, microstructure, electronic energy level distribution, and surface reactive sites of amorphous catalysts, which contributes to an in-depth understanding of amorphous catalysts. Based on the Cu valence state mapping of EELS spectral imaging, Chen et al. obtained that metallic Cu can be oxidized by SC CO₂ into amorphous Cu_xO, forming a structure with a core of Cu and an amorphous shell of Cu_xO^[107].

Theoretical characterizations

Typically, the complicated properties of amorphous catalysts offer difficulty in capturing full information through experiments. The insufficient experimental characterizations provide a research platform for the development of theoretical characterizing amorphous catalysts. Machine learning (ML) can acquire the most representative feature from numerous data by automatically extracting and selecting features, thus improving its identifiability^[108,109]. Combining numerous experimental data and simulation calculations, a model can be built to predict the structure of amorphous catalysts while designing and optimizing the structure by an optimization algorithm. For instance, the formation of amorphous Cu_xO was analyzed via ML [Figure 10] and can be roughly divided into three steps: surface oxidation, O penetration, and O departing^[107].

Figure 10. (A-F) The simulated mechanism for the amorphization. (Reproduced with permission^[107]. Copyright 2023, Nature).

DFT plays an important role in characterizing amorphous catalysts, and the main roles are summarized as follows. (1) The optimization of the potential energy surface can predict amorphous catalyst structures, such as atomic positions, crystal configuration, local environment, etc.^[110]; (2) The movement of atoms and molecules in amorphous catalysts can be stimulated by molecular dynamics (MD) and other methods^[111], thus revealing the dynamic process of catalytic reactions; (3) The catalytic activity of amorphous catalysts can be predicted by calculating the surface energy, adsorption properties, and energy-band structures; and (4) The electronic principles of amorphous catalysts in catalytic reactions can be revealed by calculating their electronic structural properties, such as electron density, energy band structure, density of states, etc. For example, He et al. modeled the formation mechanism of 2D amorphous PtSe_x by using ab initio MD (AIMD) simulations based on DFT^[112]. The PtSe₂ crystal structure with a PtSe_x surface was obtained by sequentially removing Se atoms from the PtSe₂ monolayer and then calculating the selenium-depleted PtSe_x monolayers to create an amorphous structure via AIMD. Moreover, the combination of ML and DFT can explore the complex structure-activity relationship in binary or ternary multi-metal (hydro)oxide amorphous catalysts^[113].

Structural design

An effective way to tune electrocatalytic properties is to develop innovative structural designs. The designed distinct structures can manipulate the inherent features of catalysts, such as electronic structures and mass transfer. At the same time, amorphous metal oxides with oxygen vacancies offer unique advantages for electrocatalysis. They provide high surface area, enhanced mass transport properties and tunable electronic structure, which facilitates efficient electrocatalytic processes^[114]. For instance, single Ga atoms were distributed in the oxygen vacancies of lotus-root-like a-TiO₂ nanofibers (Ga-SA/a-TiO₂) [Figure 11A-C] to improve their selectivity for NRR^[98]. The inset SAED indicates the amorphous nature of a-TiO₂ and Ga-SA/a-TiO₂, while the electron paramagnetic resonance (EPR) suggests the abundant oxygen vacancies in Ga-SA/a-TiO₂ [Figure 11D]. The obtained Ga-SA/a-TiO₂ delivers a high NH₃ yield (24.47 μg h^-1 mg^-1) and high FE (48.64%) at a very low potential (-0.1 V_RHE) [Figure 11E], exhibiting excellent NRR performances. Moreover, the self-supporting NiFe- layered double hydroxide@Fe-MOF/iron foam-2 (NiFe-LDH@Fe-MOF/IF-2) OER electrocatalyst with a multi-level structure was prepared by in-situ Ni treatment to partially convert Fe-MOF into amorphous NiFe-LDH^[115]. A TEM image [Figure 11F] shows that the structure of the NiFe-LDH@Fe-MOF/IF-2 catalyst is a flower-like structure composed of interconnected NSs. As shown in Figure 11G, amorphous NiFe-LDH was formed on the surface of Fe-MOF. Its unique hierarchical structure is conducive to exposing more active sites, shortening the diffusion path of charge carriers and increasing the rate of electron transfer, thus enhancing its catalytic performance. As displayed in Figure 11H and I, NiFe-LDH@Fe-MOF/IF-2 delivers a low OER overpotential of 183 mV and a Tafel slope of 41.16 mV dec^-1 at 10 mA cm^-2 in 1 M KOH solution.

Figure 11. (A) SEM and (B) TEM images of a-TiO₂. (C) TEM image of Ga SA/a-TiO₂ nanofibers. (D) EPR spectra of Ga SA/a-TiO₂ nanofibers. (E) NH₃ yields and FEs of Ga SA/a-TiO₂ nanofibers at different potentials for 2 h (Reproduced with permission^[98]. Copyright 2022, American Chemical Society). (F) TEM images and SAED pattern of NiFe-LDH@Fe-MOF/IF-2. (G) HRTEM image of NiFe-LDH@Fe-MOF/IF-2. (H) LSV curves of various samples. (I) Corresponding Tafel plots. (Reproduced with permission^[115]. Copyright 2021, Elsevier).

Tuning the self-supported electrode structures can improve their catalytic properties. This can be mainly attributed to (1) Self-supporting system electrocatalytic materials are directly grown or anchored on the conductive substrate through a certain process without the need for adhesives, thus avoiding its inhibitory effect on the conductive ability of the material; (2) Most self-supporting substrates themselves have good electrical conductivity, which can effectively improve the electron transmission of electrocatalytic materials with poor conductivity and promote the effective transmission of carriers during the catalytic process; (3) A strong interaction can be formed between the active material and the substrate. The strong contact interface between the electrocatalytic material and the self-supporting substrate can efficiently transport electrons and reduce the reaction barrier; and (4) Due to the strong contact between the electrocatalytic material and the self-supporting substrate, it avoids the shortcoming of easily falling off during the catalytic process using binders, so it shows good catalytic stability. For example, a self-supported amorphous CoMoO₄ nanowire array on Ti mesh (CoMoO₄ NWA/Ti) was produced by a simple two-step hydrothermal approach^[116], which exhibits excellent HER activity and stability in 1 M KOH solution. Moreover, amorphous SA-MoO₂/Ni₃(PO₄)₂/NF with a self-supported 3D micropillar structure was prepared [Figure 12A and B] to catalytic HER^[99]. The obtained catalysts featured an NF substrate with excellent electrical conductivity and a unique structure that increased the mass transfer rate and electrochemical surface area (ECSA) [Figure 12C], which further enhanced the electrical conductivity and thus accelerated the catalytic reaction. The surface amorphous state endows the SA-MoO₂/Ni₃(PO₄)₂/NF structure with self-repairing ability and improves its corrosion resistance [Figure 12D]. Meanwhile, the surface amorphous state can effectively optimize the Gibbs free energy of hydrogen adsorption and the energy barrier of water decomposition [Figure 12E-H], thus exhibiting excellent electrocatalytic HER performances.

Figure 12. (A) SEM and (B) HRTEM images of SA-MoO₂/Ni₃(PO₄)₂/NF. (C) ECSA estimated by double-layer capacitance measurements. (D) Nyquist plots. DFT calculations: (E) Free energy of hydrogen adsorption. (F) Model of MoO₂ (-111) and (G) model of SA-MoO₂ (-111). (H) Energy barrier of water dissociation. (Reproduced with permission^[99]. Copyright 2023, Wiley-VCH).

Composition tuning

The composition regulation can optimize the electronic structure and surface chemistry of the amorphous catalysts, generating active sites that facilitate electron transfer and improve conductivity, thus improving their catalytic activity and stability. The composition modification strategies include alloying, doping, and compounding. For instance, introducing Co²⁺ to amorphous MoS_X creates a Mo-S-Co structure^[117], which increases active sites from unsaturated Mo atoms and S atoms, making it easier for electrons to migrate between the surface of the catalyst and the active sites, increasing the rate of electron transfer, and then improves its intrinsic electrocatalytic activity. Moreover, a novel amorphous P-doped MoS₂ nanocomposite was prepared by a facile hydrothermal method^[118]. The obtained amorphous P-doped MoS₂ catalyst delivers better catalytic activity and stability than that of the crystalline MoS₂ in acidic electrolytes. The Fe element was doped in amorphous Ni-P-B-O (NPBO) nanocages to prepare Fe-NPBO (FNPBO) catalysts and their amorphous feature was revealed via XRD patterns [Figure 13A]^[119]. The FNPBO, as an OER catalyst, exhibits a smaller overpotential and a smaller Tafel slope than that of NPBO [Figure 13B]. Amorphous NiWO₄ and Fe-doped NiWO₄ powders were prepared by the solution without or with Fe(NO₃)₃·6H₂O, respectively [Figure 13C]^[120]. They show disordered structures [Figure 13D-G], and the electronic structure of NiWO₄ was tuned by Fe³⁺ doping [Figure 13H], improving electron transfer efficiency and further enhancing catalytic performance. The OER overpotential of the synthesized Fe-doped amorphous NiWO₄ exhibits a 90 mV drop than that of amorphous NiWO₄ at 10 mA cm^-2 in 1.0 M KOH solution [Figure 13I]. The specific surface area of amorphous N, P, and F triple-doped CoFe₂O₄ was enlarged by attaching ultrafine MoS₂ nanoclusters on its surface. Since oxygen vacancies can affect the adsorption energy of intermediates, they change the reaction pathways, lower the reaction energy barriers, and improve the catalytic activity and selectivity of electrocatalysts. To further improve their activity, oxygen vacancies were created by using ionic liquids as dopants; thereby, the oxygen-vacancy-rich MoS₂/NPF-CoFe₂O₄ (MNC) catalysts were prepared^[121]. As shown in Figure 13J, the absence of diffraction peaks in the XRD pattern signifies the amorphous feature of MNC. The numerous oxygen vacancies, amorphous structure, large surface area (148.9 m² g^-1), and active sites (such as N-Fe, N-Co, etc.) in MNC contribute to its superior OER performances, showing an overpotential of 250 mV and a Tafel slope of 41 mV dec^-1 at 10 mA cm^-2[Figure 13K-M]. Furthermore, DFT simulations demonstrate that the presence of heteroatoms and MoS₂ can tune the electronic structure and lower the energy barrier of the intermediates during the OER process [Figure 13N].

Figure 13. (A) XRD patterns of the FNPBO products with different ratios of Fe to Ni. (B) LSV curves of different samples. (Reproduced with permission^[119]. Copyright 2019, American Chemical Society). (C) The synthesis of systematic amorphous NiWO₄ and Fe-doped amorphous NiWO₄. (D) HRTEM and (E) SAED of NiWO₄. (F) HRTEM and (G) SAED of Fe-doped NiWO₄. (H) Possible crystal field study of NiWO₄ with or without Fe³⁺ doping. (I) Backward LSV curve of NiWO₄, Fe-doped NiWO₄, annealed NiWO₄, and annealed Fe-doped NiWO₄ at a scan rate of 5 mV/s in 1 M KOH. (Reproduced with permission^[120]. Copyright 2023, American Chemical Society). (J) XRD patterns of the prepared catalysts. (K) N₂ adsorption-desorption isotherm and pore-size distribution (inset) of the MNC. (L) OER polarization curves without iR-compensation. (M) Tafel slopes. (N) The elementary steps for the OER process catalyzed by various catalysts. (Reproduced with permission^[121]. Copyright 2018, WILEY-VCH).

Per the above analysis, composition tuning can significantly improve the catalytic properties of catalysts. On the one hand, the adjustment of the catalyst composition can expand the reaction types, thus realizing various reactions via catalyst design and preparation. Catalysts with varied compositions have distinct chemical characteristics, surface structures, and crystal structures that allow them to be applied to a range of reactants and reaction conditions while enhancing catalytic activity. On the other hand, tuning the ratio and distribution of catalyst components can regulate crystal structure, coordination environment, electronic structure, and other surface properties, thus exposing more active sites while optimizing their adsorption and reaction properties, which, in turn, reduces charge transport paths and losses, accelerates charge transfer efficiency, and improves the conductivity of the material itself. Finally, the tuned composition can limit phase transitions and lattice flaws occurring in catalysts and then reduce catalyst deactivation during the reaction process and exhibit excellent catalytic stability.

Surface modification

The factors of the electronic structure and surface acid-base properties that affect the catalytic reactions of amorphous catalysts can be controlled by surface modification. For example, the microstructure of amorphous FNPBO can be adjusted by various Fe/Ni ratios [Figure 14A-F]^[119]. The microstructure changes from nanocages to NSs as the Fe/Ni ratio in FNPBO increases, thus altering its OER performance. The optimized FNPBO (Fe_6.4Ni_16.1P_12.9B_4.3O_60.2) delivers excellent OER activity of 236 mV overpotential and 39 mV dec^-1 Tafel slope at 10 mA cm^-2 and can continuously operate for 50 h at 236 mV without attenuation [Figure 14G and H]. The amorphous mixed-metal (Ni-Fe) hydroxide nanostructures with uniform Ni/Fe distribution and controllable Ni/Fe ratios were prepared by Gao et al. and used as highly efficient OER electrocatalysts^[122]. The catalytic performances of amorphous Ni-Fe hydroxides and their Ni content exhibit a volcano relationship, indicating the vital for generating excellent alloy hydroxide/oxide OER catalysts is their composition. The highly disordered structures and irregular surfaces endow amorphous Ni_0.71Fe_0.29(OH)_x catalysts with more active site exposure and enhanced charge transfer efficiency. Moreover, a self-assembled amorphous molybdenum oxide-loaded single-atom Ru catalyst (Ru-SAs-MoO_3-x/NF) was reported by Feng et al.^[78]. The obtained Ru SAs-MoO_3-x/NF features a 3D flower-like structure, amorphous features, and uniform distributed Ru monoatoms, which endow the catalyst with a large surface area and numerous active centers, enhancing the charge transfer rate. Finally, it facilitates the mass-transfer kinetics of electrochemical water decomposition.

Figure 14. TEM of FNPBO with different Fe/Ni ratios: (A) Fe_0.3Ni_28.6P_6.8B_8.7O_55.6; (B) Fe_4.0Ni_20.9P_11.6P_4.5O_58.9; (C) Fe_6.4Ni_16.1P_12.9B_4.3O_60.2 (the insert is SAED pattern); (D) Fe_10.2Ni_12.0P_12.3B_4.9O_60.6; (E) Fe_10.0Ni_5.5P_15.5B_5.8O_63.1; and (F) Fe_16.8Ni_6.1P_9.3B_7.7O_60.1. (G) Tafel slope-potential plots of the Ni_33.3B_14.7O_52.0, Ni_33.7P_3.2B_10.2O_52.9, FNPBO catalysts, and commercial RuO₂. (H) Durability test of Fe_6.4Ni_16.1P_12.9B_4.3O_60.2 nanocages at 236 mV. (Reproduced with permission^[119]. (Copyright 2019, American Chemical Society).

Interfacial engineering

The optimization of the interaction between amorphous catalysts and reactants, the improvement of the catalyst surface activity, and the regulation of surface active centers and adsorption sites can be realized by interfacial engineering, thus selectively modulating various reactions. Meanwhile, the microcrystalline regions present in the amorphous catalysts can influence the electrocatalytic performance of the materials, which is due to the fact that the coexistence of the crystalline and amorphous phases can provide a richer surface structure, increase the density of active sites, and improve the reactivity of the catalysts; modulate the surface energy of the catalysts, which makes the reactants easier to be adsorbed and transformed; improve the charge transport performance of the catalysts, reduce the resistance, and increase the electron transport rate, thus accelerating the reaction rate; and increase the stability and resistance to loss of activity, which prolongs the service life of the catalysts. For instance, a new amorphous/crystalline CrO_x-Ni₃N with heterostructure was grown on NF via the combination of hydrothermal and nitridation^[123]. The DFT calculations indicate that the electronic structure modulation at the CrO_x-Ni₃N interface is conducive to optimizing H₂O dissociation (∆G_H-OH) and H adsorption (∆G_H), thus, in turn, contributing to excellent and stable HER properties [Figure 15A]. Meanwhile, CrO_x-Ni₃N delivers a higher electron density near the Fermi energy level than that of Ni₃N, suggesting its improved conductivity [Figure 15B]. The CrO_x-Ni₃N heterostructure shows nanoparticles decorated NS structure [Figure 15C] and exhibits mixed phases; the nanoparticles are crystalline Ni₃N, while the NSs are amorphous CrO_x [Figure 15D]. The abundant oxygen vacancies and amorphous/crystalline interfaces endow CrO_x-Ni₃N catalysts with more active site exposure and increased intrinsic activity, resulting in excellent HER performances. Moreover, the CoFeO_x(OH)_y/CoO_x(OH)_y (CFOH/COH) was prepared by coupling various CoFeO_x(OH)_y with CoO_x(OH)_y^[124], which exhibits a unique amorphous core-shell structure. The overpotential of CFOH/COH catalysts is smaller than A-F and A-C catalysts at the same current density, indicating that the amorphous core/shell structure was indeed conducive to increasing HER activity and accelerating electron transfer rate between the catalyst interface and electrolyte. The optimized CFOH/COH-1 (the ratio of Co/Fe is 1:1) catalyst delivers better catalytic activity than others [Figure 15E]. The nanoflower-like heterostructure amorphous NiFe(OH)_x/crystalline (Ni, Fe)Se on the NF catalyst was designed by Sun et al. and used electrocatalytic water splitting [Figure 15F-J]^[125]. The amorphous NiFe(OH)_x features more active site exposure and a super-hydrophilic interface, facilitating electrolyte contact with the catalyst and the release of gas bubbles. The interfacial electronic interactions between amorphous and crystalline phases can modulate the catalyst's electronic structure and optimize intermediate conversion, increasing the electrical conductivity and thus accelerating catalytic kinetics.

Figure 15. (A) The calculated Gibbs free energy of OH-H and H adsorption for the Ni₃N, CrO_x, and CrO_x-Ni₃N. (B) The total PDOSs of the Ni₃N and CrO_x-Ni₃N. (C and D) TEM characterizations of CrO_x-Ni₃N. (Reproduced with permission^[123]. Copyright 2022, Wiley-VCH). (E) Schematic illustration of the overall water splitting process. Amorphous CFOH/COH catalyst exhibits much better performances than other as-synthesized A-F or A-C catalysts. (Reproduced with permission^[124]. Copyright 2020, Elsevier). (F and G)TEM and HRTEM of NFSe@NFOH/NF. (I and J) Corresponding FFT patterns of the selected regions marked by purple and pink squares, respectively. (H) Lattice spacing corresponds to the selected white region. (Reproduced with permission^[125]. Copyright 2023, Elsevier).

In short, catalyst interfacial engineering can create heterogeneous structures forming microcrystalline regions and significantly increase catalytic activity. The first reason is that the formed heterogeneous structure exposes more active sites and expands the reaction boundary area, thus enhancing the activity at the electrolyte-electrode interfaces. Another reason is that heterostructures can modulate catalyst electronic structures and then improve catalytic activity. The heterostructure interfaces of different materials can provide various lattice structures, change local potentials, modulate electronic states, alter energy levels, and tune catalyst surface active sites, thereby enhancing the charge transfer rate and improving the material conductivity. Finally, the formed amorphous interfaces reduce local stress and dislocation density in crystals, decrease the formation and diffusion of lattice defects, and prevent surface oxidation and corrosion, thus improving the adaptability and long-term stability of catalysts.

Electrochemical modification

The potentials or currents are conducted to electrochemical modulate amorphous catalysts, thus activating the catalyst surface or modulating the electrocatalytic process. The introduction of activatable functional groups, such as hydroxyl and carboxyl groups, on the surface of amorphous catalysts by electrochemical methods can increase the active sites of the catalysts and improve the reactivity; optimize the paths and rates of electron transport and improve the conductivity. For example, the amorphous high-entropy CoNiFeMn phosphorus oxide (CNFMPO) [Figure 16A] was obtained by a simple low-potential electrodeposition method^[126]. The HER-activated CNFMPO undergoes surface reconstruction; during the activation process, the NS layer gradually shrinks and thickens for surface dissolution and redeposition of metal salts. Similarly, OER-activated CNFMPO undergoes surface reconstruction; the surface elements, phosphite, and phosphate at high valence experience dissolution and redeposition, resulting in the NSs gradually transforming into fine nanoparticles and reconstructing into oxides or (oxy)hydroxides. Compared with the reference samples, these structural changes endow the activated CNFMPO catalyst electrode with a faster electron transfer rate, an increased electrochemically active surface area, and more active sites, delivering excellent HER and OER properties in 1 M KOH solution [Figure 16B and C]. Moreover, the amorphous small Sn(HPO₄)₂ (SnP) nanoparticles were prepared by Cheng et al. from a hydrothermal method^[127]. Placing SnP in a CO₂-saturated 1.0 M KOH solution and applying a negative potential (0.7 V_RHE) to conduct in situ electrochemical reconstruction and form a core-shell P-Sn/SnO_x structure electrode [Figure 16D and E]. The presence of amorphous SnO_x layers and surface phosphorus species suppresses the HER and contributes to the current density and stability of CO₂RR. The theoretical results demonstrate that the embedded surface P species and adsorbed phosphoric acid groups effectively contribute to forming formate rather than other products [Figure 16F-I].

Figure 16. (A) TEM images of CNFMPO (inset is SAED). (B) EIS Nyquist plots of HER-activated electrodes. (C) Plots of capacitive current as a function of scan rates for OER-activated electrodes. (Reproduced with permission^[126]. Copyright 2024, Elsevier). (D) A proposed in situ derivatization process. (E) HRTEM images of SnP after CO₂RR. (F-I) Adsorption configurations of CO₂ on the surfaces of SnO (F), SnO-PO₄ (G), and P-O-Sn (H). (I) Free energy diagram of CO₂RR to HCOOH. (Reproduced with permission^[127]. Copyright 2023, Elsevier).

From the above examples, the main reasons for how electrochemical modification improves electrocatalyst catalytic activity can be summarized as follows. (1) Regulating electronic states: The applied voltage can tune the surface-distributed electrons of electrocatalysts, the electron density of active sites, and the adsorption characteristics of reactants. Generally, applying a positive potential can expose more surface active sites and improve the adsorption/activation efficiency of reactants, thus enhancing the catalytic activity; (2) Improvement of charge transport: Modulation of the electron transport properties of amorphous catalysts to reduce resistance, increase electron transport rate, and enhance conductivity; (3) Promoting reaction kinetics: The potential of reactants can be adjusted by applying an external potential, and then changing the activation energy and reaction rate constant. Commonly, the applied proper potential facilitates high-energy intermediates or atomically active states to form on the catalyst surfaces, thus lowering the activation energy of reactants and increasing the reaction rate; and (4) Modulating surface adsorption energy: The exerted applied potential can change the catalyst surface distributed charge and regulate the surface reactant adsorption strength. The increased adsorption sites and moderate adsorption strength can optimize the reactant adsorption and activation states on catalyst surfaces, finally improving catalytic activity.

HER

As a crucial half-reaction in electrocatalytic water splitting, HER plays a pivotal role in sustainable energy conversion and storage fields^[140-142]. This reaction involves a two-electron transfer process, and its pathways can be classified as the Volmer-Heyrovsky process (always used in electrochemical desorption mechanism) and the Volmer-Tafel process (commonly applied in recombinant desorption mechanism). The formation of hydrogen intermediate (H*) is expressed in the following Equations (3-8), which are substantially influenced by the pH of the electrolyte^[141].

In acidic media:

In alkaline media:

The HER mechanism typically contains the steps of electron transfer, atomic rearrangement, intermediate species formation and dissociation, etc. In detail, the hydrogen ions on the cathode surface receive electrons and form hydrogen atoms (H*), and then, two H* integrate to form H₂ molecules^[143]. The occurrence of HER needs to introduce catalysts that can reduce activation energy and accelerate reaction kinetics. In the currently studied electrocatalysts, amorphous catalysts feature the advantages of high activity, uniform loading, low catalyst deactivation rate, and tunability, receiving wide applications in HER.

For example, amorphous/crystalline Rh(OH)₃/CoP heterostructure was constructed to catalyze HER in the entire pH range [Figure 18A]^[144]. In 1.0 M KOH solution, Rh(OH)₃/CoP delivers a lower overpotential of 13 mV and Tafel slope of 24 mV dec^-1 at 10 mA cm^-2, and the HER activity surpasses other comparison samples [Figure 18B and C]. This can be attributed to the unique heterostructure of Rh(OH)₃/CoP, which has special hydrophilic/aerophobic properties, accelerates mass transfer, and provides abundant exposed active sites and appropriate defects. A distinctive fiber-like porous amorphous NF-PVP/Mn_xCo_y was prepared to reduce water^[145], which benefits gas diffusion and the interaction between the catalyst and electrolyte solution. The optimized amorphous NF-PVP/Mn₄Co NSs exhibit a low overpotential of 100 mV at 10 mA cm^-2, showing great potential in HER at a large scale. The HER activity of amorphous RuS₂ is superior to that of crystalline RuS₂, delivering a lower overpotential of 141 mV and a Tafel slope of 65.6 mV dec^-1 at 10 mA cm^-2[146]. Moreover, the ML was conducted to calculate H adsorption Gibbs free energy (ΔG_H), and then screen and analyze the optimal active sites from the established 11,752 models^[147]. The optimal ratio of Pt:Pd:Ni:Cu:P = 0.33:0.17:0.155:0.16:0.185 was acquired for the local chemical environment, confirming the adsorption clusters of Pt, Pd, and Ni atoms are the main factors determining catalytic performances [Figure 18D]. In detail, Pt atoms are favorable to reducing ΔG_H while Pd and Ni atoms are favorable to increasing ΔG_H; they synergistically regulate the catalytic activity of adsorption clusters. The amorphous Fe_0.95S_1.05 (a-FeS) nanorod catalysts were prepared to catalyze HER and showed excellent electrocatalytic activity^[148]. A low overpotential of 67 mV was obtained at 10 mA cm^-2 and can stably operate for 80 h, outperforming other crystal electrocatalysts [Figure 18E]. Moreover, the electrocatalytic HER properties of the recent amorphous catalysts are summarized in Figure 18F^[149-156].

For example, amorphous/crystalline Rh(OH)₃/CoP heterostructure was constructed to catalyze HER in the entire pH range [Figure 18A]^[144]. In 1.0 M KOH solution, Rh(OH)₃/CoP delivers a lower overpotential of 13 mV and Tafel slope of 24 mV dec^-1 at 10 mA cm^-2, and the HER activity surpasses other comparison samples [Figure 18B and C]. This can be attributed to the unique heterostructure of Rh(OH)₃/CoP, which has special hydrophilic/aerophobic properties, accelerates mass transfer, and provides abundant exposed active sites and appropriate defects. A distinctive fiber-like porous amorphous NF-PVP/Mn_xCo_y was prepared to reduce water^[145], which benefits gas diffusion and the interaction between the catalyst and electrolyte solution. The optimized amorphous NF-PVP/Mn₄Co NSs exhibit a low overpotential of 100 mV at 10 mA cm^-2, showing great potential in HER at a large scale. The HER activity of amorphous RuS₂ is superior to that of crystalline RuS₂, delivering a lower overpotential of 141 mV and a Tafel slope of 65.6 mV dec^-1 at 10 mA cm^-2[146]. Moreover, the ML was conducted to calculate H adsorption Gibbs free energy (ΔG_H), and then screen and analyze the optimal active sites from the established 11,752 models^[147]. The optimal ratio of Pt:Pd:Ni:Cu:P = 0.33:0.17:0.155:0.16:0.185 was acquired for the local chemical environment, confirming the adsorption clusters of Pt, Pd, and Ni atoms are the main factors determining catalytic performances [Figure 18D]. In detail, Pt atoms are favorable to reducing ΔG_H while Pd and Ni atoms are favorable to increasing ΔG_H; they synergistically regulate the catalytic activity of adsorption clusters. The amorphous Fe_0.95S_1.05 (a-FeS) nanorod catalysts were prepared to catalyze HER and showed excellent electrocatalytic activity^[148]. A low overpotential of 67 mV was obtained at 10 mA cm^-2 and can stably operate for 80 h, outperforming other crystal electrocatalysts [Figure 18E]. Moreover, the electrocatalytic HER properties of the recent amorphous catalysts are summarized in Figure 18F^[149-156].

Figure 18. (A) Synthesis illustration of the Rh(OH)₃/CoP. Electrocatalytic HER performance in 1 M KOH solution. (B) HER polarization curves, (C) Tafel plots. (Reproduced with permission^[144]. Copyright 2023, Wiley-VCH). (D) H-metal distance contribution analysis as the function of ΔG_H. (Reproduced with permission^[147]. Copyright 2023, Elsevier). (E) LSV polarization curves for the HER of different catalysts. (Reproduced with permission^[148]. Copyright 2021, Elsevier). (F) The electrocatalytic HER properties of the recent amorphous catalysts.

OER

For water splitting, another key half-reaction is OER. However, the four-electron transfer involved in OER leads to a great challenge in improving overall water-splitting kinetics. The commonly accepted OER mechanism at acidic (Equations 9-12) or alkaline (Equations 13-16) conditions includes the following four steps, respectively^[157]. The OER operates in acidic or alkaline solutions, both involving the formation of OH*, O*, and OOH* intermediates. The specific reaction pathways and kinetics of OER highly depend on the binding strength of M-OH*, M-O*, or M-OOH*^[158].

In acidic media:

In alkaline media:

Effective electrocatalysts play a crucial role in driving water splitting. As previously reported electrocatalysts, amorphous catalysts show unparalleled advantages in electrocatalyzing OER and are receiving increasing attention. For instance, hollow multichannel amorphous NiFe nanoreactors (NiFe-ANR) were produced as OER electrocatalysts^[159], and their performances overmatch those of crystalline nanoreactor and amorphous non-nanoreactor catalysts. The amorphous structure enables NiFe-ANR to oxidize and form high-valent reactive species; meanwhile, numerous microcrystalline interfaces provide more active sites. Furthermore, the nanoreactor structure causes a localized high hydroxide concentration in NiFe-ANR and effectively promotes OER. As a result, the overpotential of NiFe-ANR (228 mV) is 118 mV lower than NiFe-CNR and 75 mV lower than NiFe-ANR at 10 mA cm^-2. A pervasive top-down strategy was developed by Liu et al. to prepare amorphous oxide materials by low-temperature treating chalcogenide hydroxides [Figure 19A], which is capable of amorphizing a variety of unitary, binary, and ternary chalcogenide hydroxides^[160]. The current density of the obtained ternary FeCoSn(OH)₆-300 amorphous oxide reaches 178 mA cm^-2 at 1.6 V_RHE and stably operates for 200 h at 100 mA cm^-2 [Figure 19B and C], which surpasses many previously reported OER electrocatalysts. DFT calculations reveal that amorphization strongly promotes the p-d hybridization between Co and Fe sites, significantly lowering the potential barrier of the rate-determining step that converses OH* to O* in OER [Figure 19D and E]. The structural evolution of FeCo-PO_x from crystalline to amorphous can be regulated by the precursor anion etching time [Figure 19F]^[161]. The gained amorphous FeCo-PO_x exhibits a low overpotential of 270 mV at 10 mA cm^-2 and almost no degradation operating for 10 h, suggesting its excellent OER activity and stability [Figure 19G]. The recent reports of amorphous electrocatalyst properties in OER are concluded in Figure 19H^[162-169].

Figure 19. (A) Schematic illustration for the structural evolution of CoSn(OH)₆ during amorphization. (B) The current densities of MCoSn(OH)₆ and MCoSn(OH)₆-300 in O₂-saturated 1.0 M KOH solution. (C) Continuous OER chronopotentiometry test for FeCoSn(OH)₆-300 at 100 mA cm^-2. (D) The PDOS of FeCoSn(OH)₆-300. (E) The energetic trend of OER at U = 0 V. (Reproduced with permission^[160]. Copyright 2022, Nature). (F) XRD patterns of Fe-PO_x 0.5~1.5 h and A-Fe-PO_x. (G) Chronopotentiometric curves of the long-term stability of electrodes at 10 mA cm^-2 without iR compensation (Inset: polarization curves initial and after 2,400 CV). (Reproduced with permission^[161]. Copyright 2023, Elsevier). (H) The recent reports of amorphous electrocatalyst properties in OER.

ORR

The inverse reaction of OER is ORR, which is extremely important in the applications of fuel cells, metal-air batteries, sensors, and other electrochemical devices. The conversion efficiency from chemical to electrical in energy conversion devices strongly depends on ORR. Typically, there are two kinds of reaction pathways of ORR. One is reduced O₂ to H₂O or OH^- via a four-electron transfer pathway in acidic and alkaline media, respectively^[170]. Another pathway is reduced O₂ to H₂O₂ (acidic solution) or HO₂^- (alkaline solution) via a two-electron transfer. Taking a four-electron transfer as an example, ORR can be expressed by the following half-reaction (Equations 17 and 18)^[171].

In acidic media:

In alkaline media:

The common ORR catalysts, such as Pt-group metals and transition metal-based materials, need further optimization to realize practical applications. Amorphous ORR electrocatalysts come into being on this basis and attract marvelous concern. We summarized the recent amorphous electrocatalysts in ORR, as shown in Figure 20A^[172-181]. Oxygen vacancies in amorphous catalysts play an important role in the ORR: they can provide reactive sites and act as electron transport media, participating in the electron transfer and energy level regulation on the catalyst surface, which affects the reaction rate and selectivity; at the same time, the active oxygen vacancies can react with oxygen or oxides to form intermediates of redox reactions and promote the reaction. For example, amorphous mullite SmMn₂O_5-δ nanoparticles with controllable oxygen vacancy concentration were prepared to catalyze ORR^[103]. Moderate oxygen defects endow SmMn₂O_5-δ with an Mn³⁺/Mn⁴⁺ ratio of 1.78 (the valence of Mn is 3.36), demonstrating excellent ORR activity in alkaline solution. Their ORR performances exceed the well-crystallized SmMn₂O₅ (SMO) and are comparable to commercial Pt/C while reducing the undesirable product (HO₂^-) yield from 40% to less than 5%. Amorphous Fe sub-nano-clusters were grown on carbon skeletons (Fe cluster/3D-C) [Figure 20B] as ORR catalysts via a new “amino-induced spatial confinement” strategy^[173]. The amorphous Fe cluster/3D-C loading small Fe (1.5 wt%) delivers a half-wave potential of 0.92 V and a mass activity of 8,600 A g_Fe^-1 [Figure 20C], exceeding the benchmark commercial Pt/C (800 A g_Pt^-1). The current density maintains about 97% operating 24 h, suggesting its high durability for ORR [Figure 20D].

Figure 20. (A) The recent amorphous electrocatalysts in ORR. (B) Schematic of synthesis for Fe subnanocluster/3D-C. (C) Mass activity for Pt/C, Fe NP/3D-C, and Fe sub-nano-cluster/3D-C. (D) methanol tolerance tests for Fe sub-nano-cluster/3D-C. (Reproduced with permission^[173]. Copyright 2019, American Chemical Society). (E) LSV curves of as-prepared catalysts and Pt/C. (F) Computational models, O₂ adsorption configurations, and corresponding adsorption energies for FeN₄ and CuN_x-FeN₄ (X = 2,3,4). (Reproduced with permission^[106]. Copyright 2020, American Chemical Society).

Amorphous Cu/Fe-NG was formed via vacuum annealing and exhibited excellent ORR activity [Figure 20E]^[106]. The Cu/Fe-NG features an initial potential of 1.01 V_RHE, a half-wave potential of 0.88 V_RHE, and a limiting current density of 6.01 mA cm^-2 at 0.4 V_RHE, outperforming that of commercial 20 wt% Pt/C. Highly active bimetallic sites M-N_x (M = Cu, Fe) in amorphous Cu/Fe-NG were revealed via DFT calculations and contributed to enhancing its intrinsic activity [Figure 20F]. A series of amorphous manganese oxide catalysts (MnO_x-PMOD) were produced to reduce O₂. MnO_x-PMOD48 shows poor stability but superior ORR properties than that of crystalline MnO_x and MnO_x/Ti₄O₇^[175]. Ti₄O₇ was further decorated on amorphous catalyst MnO_x/Ti₄O₇-PMOD to improve ORR properties. The ORR half-wave potential of 0.75 V_RHE and limiting diffusion current of 5.21 mA cm^-2 were evaluated in MnO_x/Ti₄O₇-PMOD, surpassing crystalline MnO_x/Ti₄O₇-300 catalyst. Amorphous Pd-Cr nanofibers deliver superior catalytic ORR performances in alkaline solution rather than crystalline Pd-Cr catalysts^[182], such as a higher onset, positive half-wave potential, and smaller charge-transfer resistance, suggesting the excellent ORR properties of amorphous structure catalysts.

CO₂RR

CO₂RR is a transformational process designed to use renewable energy to convert greenhouse gas CO₂ into valuable carbon-based products. This reaction provides a promising way to combat climate change, develop sustainable energy, and achieve carbon-neutral energy systems. However, CO₂RR involves a variety of products, such as C₁ products (CO, HCOOH, CH₄, etc.), C₂ products (C₂H₄, C₂H₆, C₂H₃OH, C₂H₂OH, etc.), and C₃₊ products (C₃H₆, C₃H₈O, etc.). The reaction formulas and related theory potentials for the listed CO₂RR products can be obtained from Equations 19-27^[183].

The multiple products endow CO₂RR with complex mechanisms, which can be ascribed to the formation of various reaction intermediates (*CO, *CCH, etc.) and different reaction pathways. Therefore, it is urgent to develop high-activity and high-selectivity electrocatalysts for CO₂RR. Amorphous CO₂RR electrocatalysts feature the merits of high activity, good selectivity, and stability, thus drawing increasing concern. We concluded the recent CO₂RR properties of amorphous electrocatalysts in Table 2.

For instance, amorphous copper nanoparticles exhibit excellent liquid product selectivity when used as CO₂RR electrocatalysts. As displayed in Figure 21A, when tested under CO₂-saturated 0.1 M KHCO₃ aqueous solution, the total FE based on DFT of the liquid products was up to 59% at 1.4 V_Ag/AgCl (37% and 22% for HCOOH and C₂H₆O, respectively)^[186]. The controllable amorphous copper shell can be created by supercritical CO₂ oxidation and electroreduction, and the related amorphization process is simulated in Figure 10A-F^[107]. The prepared amorphous Cu shell layer in CO₂RR displays a strong selectivity for C₂₊ products [Figure 21B and C]. The FE of C₂₊ products reached 84% at -0.9 V_RHE, and the FE of C₂₊ oxygenates attained 65.3%, which was much greater than that of crystalline Cu. The special dendritic amorphous CuBi-100 electrocatalysts [Figure 21D] were conducted to reduce CO₂ at 0.5 M KHCO₃ solution and exhibited high CH₂O₂ selectivity (FE > 85%) at -0.8~-1.2 V_RHE^[185]. The highest CH₂O₂ selectivity of 94.7% was obtained at -1.0 V_RHE [Figure 21E] in CuBi-100, while the highest CH₂O₂ current density of -12.8 mA cm^-2 was measured at -1.2 V_RHE [Figure 21F]. The minimum R_ct in CuBi-100 promotes the formation of CO₂* intermediates, thus selectively producing CH₂O₂ rather than other products.

Figure 21. (A) The FE of the liquid products of a-Cu and c-Cu at each given potential for 2 h. (Reproduced with permission^[186]. Copyright 2018, WILEY-VCH). (B) The distribution of C₂₊ products over R-8-Cu-12. (C) The plot of C₂₊ oxygenates partial current density vs. maximum C₂₊ oxygenates FE for various catalysts. (Reproduced with permission^[107]. Copyright 2023, Nature). (D) HRTEM (the inset is FFT) image of CuBi-100. (E) FE of formate for CuBi-100 and CuBi-100a. (F) Partial current density of formate for two samples in all potentials range. (Reproduced with permission^[185]. Copyright 2021, Elsevier).

NRR

NRR refers to the chemical process of converting N₂ into organic nitrogen compounds. The main product of NH₃ serves as a vital component in agriculture and industry. However, the thermodynamic stability of the N≡N bond and the competitive reaction of HER limit the NH₃ yield and FE. The development of high-active NRR electrocatalysts is urgent. In the developed electrocatalysts for NRR, abundant surface active sites, high surface area, porous structure, and good electron conductive endow amorphous catalysts with excellent NRR performances. We concluded the recent amorphous electrocatalysts for NRR in Table 3.

For example, amorphous NiSb₂O_6-x nanofibers (a-NiSb₂O_6-x) with Janus structure were proposed via the different but complementary electronic structures between d- and p- block elements [Figure 22A]^[191]. A high NH₃ FE of 23.3% at -0.25 V_RHE was obtained from a-NiSb₂O_6-x [Figure 22B]. Meanwhile, the yield of NH₃ at -0.3 V_RHE (6.8 μg h^-1 mg^-1) is much larger than that of NiO nanofibers (2.4 μg h^-1mg^-1,6.0%), Sb₂O₄ nanofibers (1.5 μg h^-1 mg^-1, 8.7%), and their crystal materials. Furthermore, a-NiSb₂O_6-x nanofibers can stably operate for 12 h [Figure 22C]. Compared to the crystalline counterpart, a-NiSb₂O_6-x exhibits a lower binding energy (-0.27 eV) and a strong adsorption capacity for N₂ by the synergy between Ni and Sb atoms [Figure 22D], thus resulting in high activity for NRR. Amorphous PdCoM (M = Cu, Ag, Fe, Mo) with 2D porous NS architectures [Figure 22E] were prepared to reduce N₂^[190]. The amorphous PdCoCu delivers excellent NRR electrocatalytic activity with an NH₃ yield of 60.68 µg h^-1 mg_cat^-1 and an FE of 42.93% at -0.05 V_RHE [Figure 22F]. Amorphous PdCoCu features a lower energy barrier for *NNH formation rather than NHNH₂* formation, demonstrating that the Cu atoms close to Pd atoms lower the N₂ adsorption barrier and accelerate the reaction kinetics of the crucial hydrogenation step [Figure 22G]. Moreover, MOF-base catalysts with hierarchical porous structure were prepared by Pan et al. for NRR^[104]. In comparison to the solid and crystalline catalysts, the amorphous H-CrO_x/C-550 exhibits excellent NRR activity at room temperature. The NH₃ yield of amorphous H-CrO_x/C-550 is 19.10 μg h^-1 mg_cat^-1, and the FE is 1.4% at -0.7 V_RHE.

Figure 22. (A) Schematic diagram showing the process for the formation of a-NiSb₂O_6-x nanofiber. (B) NH₃ yield and FE of a-NiSb₂O_6-x nanofiber at each potential. (C) Chronoamperometry curve of a-NiSb₂O_6-x nanofiber at 0.3 V_RHE for 12 h. (D) Binding energies of different adsorption configurations of N₂ on a-NiSb₂O_6-x, pc-NiSb₂O_6-x, and c-NiSb₂O_6-x, respectively. (Reproduced with permission^[191]. Copyright 2022, Elsevier). (E) SEM image of PdCoCu NSs. (F) Histograms of average NH₃ yields and corresponding FE of PdCoCu NSs at different potentials. (G) Proposed reaction mechanism for the reduction of N₂ to NH₃. (Reproduced with permission^[190]. Copyright 2023, Wiley-VCH).