In Situ Reconstruction of Helical Iron Borophosphate Precatalyst toward Durable Industrial Alkaline Water Electrolysis and Selective Oxidation of Alcohols

Iron‐based (pre)catalysts have attracted enormous attention for various electrooxidation reactions due to the low cost, high abundance, and multiple accessible redox states of iron. Herein, a well‐defined helical iron borophosphate (LiFeBPO) is developed as an electro(pre)catalyst for the oxygen evolution reaction (OER) and selective alcohol oxidation. When deposited on nickel foam (NF), LiFeBPO exhibits an exceptional OER performance at ambient conditions attaining a current density of 100 mA cm−2 at ≈276 mV overpotential in 1 m KOH. Notably, this anode sustains durable alkaline water electrolysis at 500 mA cm−2 for over 330 h under industrial conditions (6 m KOH and 85 °C). In –situ and ex situ investigations reveal a deep reconstruction of LiFeBPO during OER, which transforms into a 3D open porous skeleton assembled by ultrasmall, low‐crystalline α‐FeOOH nanoparticles (interfacing with NiOOH of NF). This structure contributes to exposing accessible surface active sites, as well as accelerating mass transport and bubble detachment. Moreover, this electrode also catalyzes the electrooxidation of alcohols (methanol, ethylene glycol, and glycerol) to formic acid (FA) with high selectivity and full conversion. This study provides promising solutions for designing suitable anodes for the simultaneous production of green hydrogen fuel and value–added FA from electrooxidation reactions.


Introduction
[7] The cathodic hydrogen evolution reaction (HER) is rather uncomplicated, but its anodic oxygen evolution reaction (OER) involving fourelectron/proton transfer steps is kinetically sluggish, thus limiting the overall efficiency of electrochemical water splitting.10] The current commercial alkaline water electrolyzers (AWEs) utilize Raney Ni alloy as the anode electrocatalyst.However, the performance of such a catalyst is still moderate and must be improved significantly to meet the challenging expectation for upscaling under industrial-relevant reaction conditions. [11,12]Keeping this in mind, in the last few years, numerous efforts have been dedicated to developing transition metal (TM)-based electrocatalysts that are inexpensive, contain earth-abundant elements, and are environmentally benign, as well as highly active for alkaline OER. [5,13]etal borophosphates (BPOs) are well known for their rich structural chemistry with complex anionic structures.The geometric configuration in the structure of BPOs is highly dependent on the B/P ratios, that is, anionic BPOs with isolated species, oligomers, rings, and layers, as well as cagelike structures, can be achieved by varying the B/P ratios. [14,15]enefiting from their structural flexibility and uniqueness, metal BPOs have been explored in many catalytic applications, which have shown admirable activity.For example, Cr 2 B(PO 4 ) 3 , which consists of oligomeric [BP 3 O 12 ] 6− anions linked by the Cr 2 O 9 octahedral dimers could be utilized as an efficient catalyst for benzyl alcohol oxidation. [16,17]3D open frameworklike copper BPO (Na 5 KCu II 3 [B 9 P 6 O 33 (OH) 3 ]•H 2 O) containing trigonal-planar BO 2 (OH), tetrahedral BO 4 , and PO 4 groups presented the high catalytic capability for the degradation of chitosan. [18]Interestingly, the analog [Na(H 2 O)][K(H 2 O)]{Na 4 -Cd 3 [BO 2 (OH)] 3 B 6 P 6 O 27 } with similar 1D wheel-shaped BPO anions has also recently been demonstrated to be an efficient photocatalyst for the degradation of reactive brilliant red (X3B). [19]][22][23] Helical BPOs comprise 1D infinite loop-branched helical channels built by alternative borate (BO 4 ) and phosphate (PO 4 ) units and linked by TMO 4 (OH 2 ) 2 octahedra with the formula AM II (H 2 O) 2 [BP 2 O 8 ]•H 2 O (A = alkali metal, M II = divalent TM).Within the structure, the alkali metal ions (Li, Na, Rb, Cs) are located at the free threads of helical ribbons, irregularly surrounded by surrounding oxide oxygen sites, while crystal water molecules occupy their inner walls stabilizing the overall structure. [14,21]Previous results have shown that the distinct crystal structure of metal BPOs undergoes an OER-driven reconstruction process with severe leaching of alkali metals, boron, and phosphorus forming an electrolyte permeable amorphous bulk active layered double hydroxides (LDHs) phases with enhanced active electrochemical surface area, as well as abundant edge active sites and defect sites (which are beneficial to adsorbing oxygen-containing intermediates during OER). [14,20,21,23,24]Such restructured LDHs engender tremendous advantages over the traditional active TM LDH electrocatalysts displaying remarkable activity and even long-term stability for months.
Considering Fe possesses the largest earth-crust abundance compared with other water oxidation-active TMs (e.g., Co, Ni, and Mn), as well as Fe-based compounds usually display rich coordination chemistry and flexible redox states, [25][26][27] developing helical Fe-based BPOs as OER catalysts are highly desirable from the aspects of preparation cost, source availability, and catalytic performance.More importantly, although helical TM-based BPOs have been developed as efficient electro(pre)catalysts for alkaline water electrolysis under ambient conditions, that is, in 1 m KOH under room temperature (pH = 13.89), [14,20,21,23,28]their potentials in the industrial-grade wa-ter electrolyzers have never been explored, obscuring their practical prospects.It should be noted herein that in spite of numerous reported TM-based OER catalysts with impressive performance, very few of them could maintain long-term excellent catalytic activity under industrially relevant conditions, that is, efficiently delivering 200-400 mA cm −2 current density in the concentrated 6 m KOH at a high temperature of 50-80 °C for at least hundreds of hours (where the corresponding pH values of KOH electrolytes are 13.70-14.49), [28]][31] Therefore, it is of enormous significance and necessity to search for low-cost and environmentally benign anodes that could successfully be applied for industrial AWEs with competitive activity and long-lasting durability.
On the other hand, analogical to water oxidation, the anodic electrooxidation of alcohols, which avoids the cost-expensive and environmentally-hazardous oxidation processes, also follows a multi-step electron transfer process. [32,33]Particularly, when utilized as anodes in direct-alcohol fuel cells, electric energy is typically generated accompanied by high-end valueadded chemicals. [34]Among them, C 1 formic acid (FA) presents enormous potential in the chemical industry, such as functioning as fuel in fuel cells, raw materials for organic chemicals, metal surface treatment agents, rubber additives and industrial solvents. [32,35][38] In this regard, partially oxidative C 2 and C 3 products from polyols, or the target FA mixed with other intermediates and CO 2 , are more frequently obtained products, [39][40][41][42] consequently giving rise to additional technical, economic, and environmental concerns.Based on these points, a reliable alcohol oxidation catalyst that efficiently and expeditiously yields FA with high selectivity is of particular interest, yet it has rarely been reported.Since TM-based BPOs have been demonstrated to be able to efficiently catalyze electrochemical OER and oxygenation process, [14,21] thus a question arises here as to whether they can also efficiently drive the oxidation of alcohols, which is also a multi-electron transfer process similar to OER.
Motivated by the abovementioned challenges, in the present work, a distinct Fe-based BPOs catalyst, LiFe II (H 2 O) 2 [BP 2 O 8 ]•H 2 O (LiFeBPO), was developed, which was then deposited on nickel foam (LiFeBPO/NF) through a binder-free electrophoretic deposition (EPD) method.The as-obtained LiFeBPO/NF displayed exceptional OER capability in 1 m KOH at room temperature, yielding 100 mA cm −2 current densities at merely ≈276 mV overpotential, which significantly outmatched NF-supported noble metal-based (IrO 2 ) and various directly synthesized Fe-based (oxy)hydroxides under the same conditions.Applying in situ Raman spectroscopy coupled with various ex situ characterizations, the origins of this excellent performance could be identified as follows: i) With the leaching of Li, B, and P, under mild working conditions in alkaline media, a deep surface reconstruction of the LiFeBPO electro(pre)catalyst into an active -FeOOH occurs, contributing to the facile formation of active structure and further reduction of OER overpotentials; [43,44] ii) the evolved ultrasmall low-crystalline -FeOOH nanoparticles were eventually interconnected as a 3D open porous architecture, beneficial for exposing surface catalytic sites with high electrolyte accessibility, as well as concurrently enhancing the mass transport and bubbles escape; [31,45,46] iii) the surface oxidation of NF substrate during OER leads to the formation of NiOOH, which generates an interface with in situ reconstructed -FeOOH, synergistically improving the OER performance.Consequently, the LiFeBPO/NF anode can enable the assembled quasi-industrial AWE (6 m KOH and 85 °C) to yield 500 mA cm −2 current density at merely ≈1.79 V cell voltage, and further maintain such an industrial-grade current density for over 330 h with negligible decay.Besides, the bifunctionality of the anodic oxidation catalysis for LiFeBPO/NF can be validated by oxidizing various alcohols (including methanol, ethylene glycol, and glycerol) fully into FA without any detectable by-products.We believe that our work could pave a new avenue to design and fabricate low-cost and excellent BPO anodes for the electrochemical oxidation of water and organics to enable the simultaneous, green, industrial-scale production of H 2 and value-added chemicals.

Structural Characterization of LiFeBPO
The LiFeBPO single-crystals were attained by the mildhydrothermal synthesis and its phase purity was verified by the Rietveld refinement of the powder X-ray diffraction (PXRD), and scanning electron microscopy (SEM) coupled with energy dispersive X-ray (EDX) characterizations (Figure 1a-e and Figure S1 and Table S1, Supporting Information).The helical LiFeBPO single-crystals present the hexagonal bipyramid feature and crystallize in the chiral space groups P6 5 22 or P6 1 22 (for crystal structure details, see Figure S2, Supporting Information).In order to obtain a good quality of the catalyst film deposited on substrates via EPD, the as-synthesized single-crystal samples were downsized through mild wet ball milling prior to the electrochemical tests (all LiFeBPO hereafter in this work were based on ball-milled samples except where specifically stated).The ball-milled samples still preserved the well-crystalline BPO phase with unaltered elemental composition and distribution (Figures S3-S6, Supporting Information), which was further confirmed by inductively coupled plasma-atomic emission spectroscopy (ICP-AES), transmission electron microscopy (TEM), as well as the associated selected area electron diffraction (SAED), and X-ray photoelectron spectroscopy (XPS) results (Figure 1f and Figure S7 and Table S2, Supporting Information).Based on these characterizations, the overall scheme of well-defined LiFeBPO serving as an efficient electro(pre)catalyst for anodic oxidation reaction is illustrated in Figure 1g.

Electrocatalytic OER Activity Under Ambient Conditions
To explore its electrochemical performance, we deposited the as-grounded well-defined LiFeBPO on NF (LiFeBPO/NF) using the binder-free EPD method (Figures S8-S10, Supporting Information; see deposition details in Experimental Section).As expected, the downsized LiFeBPO/NF behaved distinctly better OER than that of NF-supported un-milled LiFeBPO single crystals (Figure S11, Supporting Information).More importantly, as revealed by linear sweep voltammetry (LSV) curves measured at a scan rate of 5 mV s −1 in 1 m KOH at room temperature, LiFeBPO/NF exhibited an excellent OER activity, which can drive 100 mA cm −2 current density at only ≈276 mV overpotential, superior to the commercial noble metal-based reference IrO 2 deposited on NF (IrO 2 /NF) with the same mass loading, as well as the bare NF (Figure 2a).Note that almost the same OER overpotentials were required for LiFeBPO/NF when performing the LSV curve at a lower scan rate of 1 mV s −1 (all other test conditions were identical), further signifying its excellent catalytic activity (Figure S12, Supporting Information). [47][50][51] It is important to mention that nearly all Fe-based compounds are regarded as precatalysts and they transform into the corresponding Fe-based (oxy)hydroxides during alkaline OER. [25,27,52]Therefore, the OER performance of LiFeBPO/NF was further compared with directly synthesized Fe(OH) 3 , -FeOOH, and Fe  S4, Supporting Information).Furthermore, the OER kinetics of LiFeBPO/NF was examined by its Tafel slope (41 mV dec −1 ), which was similar to the previously reported efficient Fe-and NiFe-based OER catalysts (Figure S17, Supporting Information). [27,52]However, LiFeBPO/NF possessed significantly larger electrochemically active surface area (ECSA) than those of the directly synthesized Fe-and NiFe-based (oxy)hydroxide reference samples, which can be verified by their double-layer capacitance (C dl ) (Figure 2c and Figure S18, Supporting Information).Note that the determination method of C dl we employed might not represent the real ECSA of our Fe-based catalysts.Ideally, an accurate C dl is obtained by operating cyclic voltammetry (CV) or impedance spectroscopy within a potential window accommodating non-faradaic current, and thereby, a potential region below the Fe redox chemistry or OER catalytic current is chosen to perform CV cycles. [24,53,54]On the other hand, Fe-based (oxy)hydroxides are usually non (or low)-conductive in such potential regions.[55] As there are no techniques that can measure the C dl precisely, we have adopted the CV cycling method in our current work.Moreover, the LSV curves of LiFeBPO/NF and all Fe-based reference samples were normalized against ECSA to further disclose the superior inherent activity of LiFeBPO.As C dl is proportional to ECSA, we directly utilized C dl values for normalization (Figure S19, Supporting Information).In addition, the chronopotentiometry (CP) measurement at 100 mA cm −2 for LiFeBPO/NF displayed a sustained 24 h working stability with negligible decay in activity (Figure 2d).After such an OER CP, we measured the C dl of the post-OER LiFeBPO/NF, and it showed a nearly 1.5 times increase than that of the one before the reaction, which already suggested an obvious reconstruction process leading to the exposure of higher active surface area (Figure 2c and Figure S20, Supporting Information).This is probably influenced by the decreased conductivity which was induced with the transformation from LiFeBPO into (oxy)hydroxides (especially low conductivity of -FeOOH, whose in situ formation in our case will be discussed in the next section), [56] thereby exhibiting a stable catalytic performance for LiFeBPO/NF behaved during OER catalysis.The excellent OER capability of LiFeBPO/NF was further substantiated by its Faradaic efficiency (FE) up to ≈98% (Figure S21, Supporting Information).In addition, LiFeBPO/NF anode measured in an ambient condition in Fe-free 1 M KOH was inferior to that measured in commercially available 1 m KOH (Figure S22, Supporting Information).According to the previous reports, this is probably because of the synergistic effect of Fe impurities on the NF substrate and the interaction between NF and the as-deposited Fe-based catalysts. [57,58]To further evaluate the inherent catalytic activity of LiFeBPO, the ball-milled LiFeBPO was also deposited on the electrochemically inert fluorine-doped tin oxide glass substrate (LiFeBPO/FTO, Figures S23 and S24, Supporting Information).Similar to LiFeBPO/NF, the LiFeBPO/FTO also showed high intrinsic OER activity which outmatched considerable TM-based catalysts supported on FTO (Figure S25 and Table S5, Supporting Information) including a stable OER CP at 10 mA cm −2 for 24 h (Figure S26, Supporting Information).

Determination of Reconstructed Structures via Ex Situ and In Situ Characterizations
In this section, ex -situ recorded Raman spectra were compared with quasi-in situ derived Raman investigations to shed more light on the OER-driven reconstruction of LiFeBPO.For this pur-pose, we utilized LiFeBPO/FTO electrodes in order to avoid any possible spectral interference/overlaps with the NF substrate during the measurement.][61] Remarkably, in the quasi-in situ Raman spectra, obtained from freeze-quenched electrodes at −196 °C, of post-OER LiFeBPO/FTO after 24 h CP at 10 mA cm −2 in the ambient alkaline media, almost all the abovementioned bands related to pristine LiFeBPO disappeared.Only traces of phosphate and borate, which corresponded to the residual LiFeBPO species, were detected at ≈1061 cm −1 .Concomitantly, bands at ≈300 (A g ), ≈389 (A g ), ≈485 (A g ), ≈555 (A g ), and ≈688 (B 2g ) cm −1 emerged, which can be assigned to the pristine goethite (-FeOOH) phase. [27,62,63]This indicates a deep conversion from LiFeBPO to -FeOOH under mild working scenarios (1 m KOH and room temperature).Such facile formation of real active structures with sufficient exposed surface reactive sites is beneficial to lowering the OER overpotentials. [43,44]Note that the band at ≈635 cm −1 originated from the FTO substrate. [52]Besides, the morphology of this post-OER LiFeBPO was also examined by SEM and the results showed a porous surface structure that was basically interconnected by a number of ultrasmall nanoparticles (Figure S27, Supporting Information).The associated SEM-EDX and ICP-AES characterizations reveal that around 60-80% of Li, B, and P atoms leached out and possibly dissolved into the electrolyte, while Fe species were well-preserved (Figures S28 and S29 and Table S6, Supporting Information).This observation was further supported by TEM characterizations in which numerous ultrasmall nanoparticles surrounding LiFeBPO bulk after 24 h ambient OER at 10 mA cm −2 , were interconnected as a 3D open skeleton-like architecture (Figure S30, Supporting Information, and Figure 3b).Focusing on these reconstructed nanoparticles, despite their low crystallinity, a highresolution TEM (HRTEM) image together with its fast Fourier transformed (FFT) pattern was obtained, which identified the distance of the presented lattice fringe (≈0.244 nm), corresponding to the (111) facet of -FeOOH (PDF# 29-0713) (Figure 3c).This signified that the bulk LiFeBPO precatalyst reconstructed into ultrasmall low-crystallinity -FeOOH nanoparticles during OER.The high-angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) and the associated EDX elemental mapping results (Figure 3d-g) further disclosed the deep surface reconstruction of LiFeBPO after ambient OER CP at a low current density of 10 mA cm −2 with the significant leaching of surface P (also for Li and B, as revealed by ICP-AES) atoms.Correspondingly, this post-OER LiFeBPO exhibited porous surface morphology.These results can also be corroborated by the EDX findings, that is, the atomic ratio of Fe and P for the near-edge area of post-OER LiFeBPO was around 1:0.1 (Figure 3d).Notably, such an as-formed 3D open porous skeleton surface promoted mass transport and gas diffusion, [12,31] while the ultrasmall size and low-crystalline -FeOOH was conducive to exposing surface active sites which were accessible to the electrolyte, [45,46]  synergistically contributing to the superior OER performance.In addition, the significant surface transformation of LiFeBPO into -FeOOH can also be substantiated via the post-OER XPS characterizations (Figure S31, Supporting Information).Based on the above analysis, we can conclude that a deep surface reconstruction of LiFeBPO was achieved under mild working conditions (ambient environments and low current density), which finally existed as 3D open porous architecture interconnected by the in situ evolved ultrafine -FeOOH nanoparticles.This reconstruction process can be furnished to a deeper degree by altering the reaction conditions, for example, under the industrial environment, which is elaborated in the next section.Besides, it is worth mentioning that for NF-supported LiFeBPO, apart from the enhancement of conductivity, NF substrate further contributed to the overall activity probably by constructing an interface of the in situ formed NiOOH (oxidation of NF during OER) and -FeOOH (Figure S32, Supporting Information).The formation of such OER-active species has also been uncovered by previous work on Fe-catalysts supported by NF during water oxidation. [64]According to the literature focusing on OER reaction mechanisms, the possible one interpreting our case well might be the bifunctional mechanism, which was slightly different from the conventional adsorbates evolution mechanism (AEM).In brief, to begin, with the increase of the applied potential during OER, the as-deposited LiFeBPO and the surface of NF substrate are oxidized into -Fe III OOH coupled with Ni III OOH support.Then, the Fe atoms in -Fe III OOH are converted into electrophilic higher-valence Fe=O through a proton-coupled electron transfer process, followed by the nucleophilic attack of external OH − on such Fe=O centers.6][67][68][69]

Industrially Relevant Alkaline Water Electrolysis
Encouraged by the promising OER performance, we further utilized LiFeBPO/NF as both anode and cathode to assemble an AWE, that displayed superior activity compared to the one integrated by LiFeBPO/NF anode and commercial NF-supported Pt/C cathode activity in 1 m KOH and room temperature (Figure S33, Supporting Information).When substituting the cathode using the NF-supported NiPt 3 @NiS (the previously developed efficient HER catalyst in our group, [70] NiPt 3 @NiS/NF), the asobtained LiFeBPO/NF (+) ǁ NiPt 3 @NiS/NF (−) cell performed even more efficiently in overall water splitting (Figure S33, Supporting Information), which inspired us to employ this NiPt 3 @NiS/NF as the cathode of our assembled AWE applied under much harsher industrially relevant conditions (6 m KOH and 85 °C).Remarkably, LiFeBPO/NF (+) ǁ NiPt 3 @NiS/NF (−) can afford 500 mA cm −2 current density at only ≈1.79 V cell voltage (Figure 4a).Interestingly, such an AWE could maintain a quite sustained industrial-grade current density for over 330 h (Figure 4b).It is worth noting that although some of the anodes have proven to be efficient under elevated industrial conditions, their long-term operational stabilities for AWE were still a matter of great concern under the premise of concurrently satisfying all the essential points of a practical AWE including large current densities, high working temperatures, and increased electrolyte concentrations (Table S7, Supporting Information), [12,71] thus the LiFeBPO/NF overcame these barriers and provided a promising practical potential.More importantly, motivated by its exciting working stability, we additionally characterized the LiFeBPO/NF anode after over 330 h electrolysis under industrially relevant conditions.As shown in Figures S32, S34, and S35, and Table S8, Supporting Information, the corresponding quasi-in situ Raman, SEM-EDX, and ICP-AES findings of LiFeBPO/NF after over 330 h industrial CP suggest an almost complete transformation of LiFeBPO into -FeOOH (reaching surface stability) under harsh working conditions (higher current density, longer electrolysis duration, greater pH value and elevated temperature of electrolyte), in which -FeOOH coupled well with the NiOOH (derived from the in situ surface reconstruction of NF substrate), being responsible for the anodic water oxidation process.Compared with the post-OER characterizations under ambient conditions (Figure 3 and Figures S27-S31 and Table S6, Supporting Information), this also affirms that the OER-driven surface oxidation can be dynamically intensified upon strengthening the catalytic conditions, during which the thickness of the reconstructed surface was accordingly enhanced.

Merits of Developing LiFeBPO/NF as the Anode for Alkaline Water Electrolysis
Based on the as-obtained findings of LiFeBPO/NF serving as the anode in alkaline water electrolysis, as well as the natures of other helical TM-based BPOs precatalysts applied in alkaline water oxidation, the merits of utilizing our LiFeBPO/NF anode can be summarized from its superiority as precatalysts and fabrication costs.
On one hand, compared with other commonly used Fe-based precatalysts (such as MOFs, phosphides, phosphates/borates, and intermetallics), [52,[72][73][74][75] LiFeBPO possesses very distinct helical crystal structures accommodating thermally unstable atoms (Li, B, and P) (elaborated in the Introduction part and Figure S2, Supporting Information).During alkaline OER, it rapidly and deeply (even completely when strengthening the reaction conditions) reconstructed into the real active disordered (oxy)hydroxide phases (-FeOOH in our case).This was systematically uncovered by several post-OER characterizations (both under ambient and industrial conditions), including the findings of quasi-in situ Raman, (HR)TEM, SEM-EDX, ICP-AES, HAADF-STEM together with the associated EDX and elemental mapping, and XPS (Figure 3 and Figures S27-S32, S34, and S35, Supporting Information, as well as Tables S6 and S8, Supporting Information).However, for most Fe-based precatalysts, the deep or complete reconstruction into (oxy)hydroxides is difficult to be achieved; [72,73] Remarkably, combining the low crystallinity feature of the reconstructed -FeOOH in our case with the observations on the other helical metal BPOs used as OER precatalysts, [14,21,23] it can be concluded that the unique structure and composition features in LiFeBPO enable this kind of precatalysts to show some additional features as compared to most reported Fe-based electrocatalysts.In specific, within the structure of the low-crystallinity -FeOOH in situ formed from LiFeBPO precatalyst, the small domains of edge-sharing TMO 6 octahedra were stacked in an unordered manner, which thereby presented a large active electrochemical surface area with rich edge active sites and lattice defects, and more importantly, the feature of electrolyte permeability.As a result, this transformed phase behaved with satisfactory bulk activity with abundant active sites. [14,21,23]Conversely, the reconstructed (oxy)hydroxides from most other Fe-based precatalysts tend to be only active in their (near-)surface regions. [76]Another point worth mentioning in our case is that with a sharp size reduction where the bulk LiFeBPO reconstructed into ultrasmall -FeOOH nanoparticles, which were interconnected as a 3D open porous skeleton (Figure 3b-d, Figures S27 and S30, Supporting Information), conducive to the accessible surface active sites exposure, as well as the mass transport and bubble detachment.This OER-favorable property of the in situ formed (oxy)hydroxides was also barely reported by other Fe-based precatalysts.
On the other hand, the fabrication process of LiFeBPO/NF anode is economically desirable, particularly when compared with that of Raney Ni alloys which commercially serve as the anode for industrial AWEs.Specifically, the hydrothermal synthesis and ball milling for LiFeBPO were conducted under facile and mild reaction conditions (e.g., moderate temperature and synthesis time, easy and safe implementation).In comparison, Raney Ni alloys are typically fabricated through mechanical and high-temperature treatment, as well as strong alkali leaching.[79][80] It is also worth pointing out that the step of downsizing the as-prepared bulk materials is usually inevitable. [78,79]Besides, during the deposition process, polymeric binders are often used, which gives rise to the additional increment of preparation cost. [12,81,82]Conversely, the EPD method that was used in our case is binder-free, which enables the successful preparation of an efficient and robust selfsupported electrode in an inexpensive way.Therefore, the fabrication of LiFeBPO/NF is economically competitive in respect of industrial production.

Selective Electrooxidation of Alcohols
[85] Inspired by the promising water oxidation activity of LiFeBPO as elaborated above, we further extended the application of this catalyst for selective alcohol oxidation, including methanol and polyols (ethylene glycol and glycerol), which undergoes a similar anodic proton/electron transfer process.
To begin, LSV curves were recorded in a three-electrode system where LiFeBPO/NF or its reference sample, bare NF, served as a working electrode in 1 m KOH with and without 0.1 m ethylene glycol.In Figure 4c, compared to the situation in pure KOH media, the ethylene glycol oxidation current density of LiFeBPO/NF was significantly boosted from ≈1.35 V (vs RHE), indicating its outstanding oxidation ability.Moreover, such a current density was much higher than that of bare NF electrode, which can be further verified by their bulk electrolysis measurements through the chronoamperometry (CA) in 1 m KOH with 0.1 m ethylene glycol electrolyte (Figure 4d).After merely 150 min, 579 C charges were passed at the LiFeBPO anode, which showed both the conversion and FE of ethylene glycol to formate (the form of FA presented in alkaline media) were as high as ≈99%, while bare NF undergoes the same reaction at a yield of ≈40% and a FE of ≈73% (317 C passed charges).The results were qualitatively and quantitatively determined by the 1 H nuclear magnetic resonance ( 1 H NMR) spectroscopy of these as-obtained reaction solutions before and after CA (Figure 4e and Figure S36, Supporting Information), manifesting that LiFeBPO/NF anode can robustly and rapidly produce FA via ethylene glycol oxidation with excellent activity and selectivity.Moreover, the excellent continuous electrooxidation performance of LiFeBPO/NF was also verified by bulk electrolysis in a larger volume of reaction solutions (Figure S37, Supporting Information).Besides, the bulk electrolysis of ethylene glycol with a concentration gradient of 0.1, 0.3, and 0.5 m catalyzed by LiFeBPO/NF unveiled that 0.3 m was probably the maximum concentration to produce FA with high efficiency and selectivity (details in Figures S36, S38, and S39, Supporting Information).To further validate its versatility for alcohol to FA conversion, we investigated the performance of LiFeBPO/NF for the electrooxidation of methanol and glycerol.Surprisingly, within very short periods of bulk electrolysis, nearly full conversion (≈99%) of methanol (386 C passed charges within 130 min) and glycerol (772 C passed charges within 105 min) into FA were achieved, accompanied by a high FE (≈99% for both alcohol substrates) (Figure 4f and Figures S40-S42, Supporting Information).Alternatively, we also conducted 13 C NMR characterizations for the post-CA reaction solutions of ethylene glycol, methanol, and glycerol catalyzed by LiFeBPO/NF, proving the only product that could be detected was formate, while other detectable carbon-based intermediates and CO 2 (presented as carbonates in alkaline media) were absent or in a negligible amount, meaning an outstanding formate selectivity of LiFeBPO for alcohol oxidation (Figure S43, Supporting Information).All the above results demonstrate the high efficiency and selectivity of LiFeBPO/NF to promote the electrooxidation of alcohols into formate products, as summarized in Figure 4g.

Conclusion
In conclusion, we have reported a helical iron BPO deposited on NF (LiFeBPO/NF), serving as an efficient bifunctional electrode for anodic alkaline water and alcohol oxidation.Impressively, LiFeBPO/NF can deliver a current density of 100 mA cm −2 at only ≈276 mV in an ambient environment (1 m KOH and room temperature), much superior to NF-supported noble metal IrO 2 and most of the previously reported Fe-based OER catalysts.This exceptional OER performance is credited to the deep structural transformation of LiFeBPO even under very mild working conditions, thus realizing the facile formation of active -FeOOH species.Such in situ emerged ultrasmall low-crystallinity -FeOOH eventually existed as a 3D open porous skeleton structure by interconnecting each other, beneficial to the exposure of surface active sites with high electrolyte accessibility.Meanwhile, the mass transport and gas bubble detachment were also substantially elevated.Moreover, benefiting from integrating the reconstructed -FeOOH with NiOOH derived from the surface oxidation of NF substrate, LiFeBPO/NF can further serve as the robust anode in an assembled AWE under industrial conditions, that is, 6 m KOH and 85 °C, delivering the large current density of 500 mA cm −2 at a low cell voltage of ≈1.79 V, which can be well maintained for over 330 h.In addition, LiFeBPO/NF can selectively facilitate the oxidation of various alcohols, including methanol, ethylene glycol, and glycerol, in which they were almost fully oxidized into formic acid within very short periods.We believe that the LiFeBPO anode developed in the current work presents highly promising practicality and applicability for the industrial production of green hydrogen fuel and valuable chemicals.For the synthesis of LiFeBPO single crystals, a mild hydrothermal approach was employed which was similar to previous works. [14,21]First, 1.44 g of FeC 2 O 4 • 2 H 2 O and 2.718 g Li 2 B 4 O 7 were added to 20 mL distilled water and the mixture was stirred for around 10 min.Then, 5.558 g H 3 PO 4 (85%) was added dropwise, followed by the addition of 2 mL HCl (37%).After being thoroughly stirred, the mixture was sealed into a 50 mL Teflon-lined autoclave (filling degree: <50%) and heated under 180 °C for 3 days.After cooling down naturally to room temperature, the as-obtained reaction product was centrifuged and washed with distilled water (three times) and acetone (two times) and dried under the ambient environment overnight.To realize the EPD, the as-prepared LiFeBPO single crystal was downsized from around 10-20 μm to 1-2 μm through mild wet ball milling.In detail, around 400 mg LiFeBPO single crystal samples were filled into an agate beaker (a volume of 12 mL) in a filling of ≈60% (the diameter of agate milling balls was 10 mm), and simultaneously the ethanol was utilized as the solvent.The ball milling was operated at 750 rpm for 10 min with a short break of 10 min, which was repeated ten times.
Electron Microscopy: SEM characterizations were performed on an LEO DSM 982 microscope combined with EDX (EDAX, Apollo XPP).Also, TEM was operated on an FEI Tecnai G2 20 S-TWIN transmission electron microscope (FEI Company, Eindhoven, Netherlands) based on a LaB 6 source at 200 kV acceleration voltage.For the measurement of the sample after electrocatalysis, LiFeBPO/FTO was peeled off the FTO substrate (to exclude interference from the NF substrate) and then transferred onto a carbon-coated Cu grid.EDX was performed with an EDAX r-TEM SUTW detector (Si (Li) detector), and the images were captured using a GATAN MS794 P CCD camera.The SEM and TEM characterizations were finished at the Zentrum für Elektronenmikroskopie (ZELMI) of TU Berlin.
XPS Measurements: The XPS measurements were performed on an ESCALAB 250Xi spectrometer (Thermo Scientific, USA) with a pass energy of 30 eV at a power of 100 W (10 kV and 10 mA) and a monochromatized AlK X-ray (h = 1486.65eV) source.All samples were measured under a pressure <1 × 10 −9 Pa.Spectra were collected using the Avantage software (Version 5.979) with a step of 0.05 еV.The analysis of the as-obtained highresolution XPS spectra was carried out using Casa XPS (Casa Software Ltd.).
Raman Spectroscopy: The ex situ and quasi-in situ Raman spectroscopic measurements were performed with the 407 nm emission from a Krypton ion laser (Innova 70, Coherent) which was used for excitation, while a confocal Raman spectrometer (Lab Ram HR-800 Jobin Yvon) coupled with a liquid-nitrogen cooled charge-coupled device (CCD) camera was employed to record the spectroscopic data.The sample was implemented into a Linkam Cryostage THMS600 cryostat.Throughout the entire measurements, the applied laser power and temperature of the sample were set to 1 mW and 80 K, respectively.Spectra were recorded at three different spots on the sample surface. [88]For the quasi-in situ characterizations, the LiFeBPO/FTO and LiFeBPO/NF electrodes after different CP runs were rapidly freeze-quenched using liquid N 2 under vigorous Ar gas flow, and subsequently stored in liquid N 2 .
Electrode Preparation: First, EPD was utilized to deposit powder catalysts on NF and FTO substrates.A potential difference of −10 V was applied in a mixture of iodine and acetone.The EPD details can be referred to in the previous report. [52,89]The sample loading on each NF and FTO was ≈2 and ≈0.8 mg cm −2 , respectively.
Electrochemistry Tests: In the case of ambient conditions, the electrochemistry tests were conducted in a standard three-electrode system in 1 m KOH under room temperature at a potentiostat (SP-200, BioLogic Science Instruments) equipped with the EC-Lab v10.20 software.The asdeposited catalysts on various substrates by EPD served as the working electrodes, while a commercially available Pt wire (0.5 mm diameter × 230 mm length; A-002234, BioLogic) and a Hg/HgO (CH Instruments, Inc.) were used as the counter electrode and reference electrode, respectively.An 85% iR compensation was applied to LSV, CA, and CP.The scan rate of both LSV and CV was set as 5 mV s −1 (or 1 mV s −1 for comparison).The applied potential was converted into the values against RHE by the equation: E(RHE) = E(Hg/HgO) + 0.098 V + (0.059 × pH) V. [28] The steady-state measurements were used to obtain Tafel slopes which were realized by performing CA at a couple of fixed potentials (5 min at each potential with an increasing stepwise of 15 mV). [47]To estimate C dl , CV was run at a non-Faradaic voltage window.Considering that C dl was linearly proportional to ECSA, and to investigate the intrinsic activity of the catalysts, the current density was normalized to their C dl . [25]The EIS was recorded at 1.5 V (vs RHE) for all OER samples supported on NF to obtain the Nyquist diagrams.The amplitude of the sinusoidal wave was defined within a frequency range of 100 kHz to 1 mHz.The FE was determined by comparing the amount of experimentally evolved O 2 and that of theoretically calculated one, following the Faradic equation: FE = (4 × V × F)/-(V m × Q).V, F, V m , and Q represent the volume of experimentally evolved O 2 (mL), Faraday constant (96 485 C mol −1 ), molar volume (24.5 L mol −1 , RT), and the total amount of electrical charge (C).The drainage method was utilized to measure the amount of evolved gas at a constant current of 100 mA for 3600 s. [7] In addition, to investigate whether the Fe impurity of 1 m KOH which was commercially available would influence the presented OER performance of LiFeBPO/NF, such an electrolyte was specially purified using the methods mentioned in the previous work. [90]To simulate the quasi-industrial AWE, LiFeBPO/NF served as the anode, which was coupled with the previously reported NiPt 3 @NiS/NF, assembling into a two-electrode cell without membrane.The working condition was set to fulfill the industrial requirements, that is, 85 °C electrolyzer temperature and 6 m KOH.The electrolyte was kept stirred at 500 rpm during electrochemistry.All LSV and CP curves of this AWE were obtained after an 85% iR compensation.When conducting a long-term CP test under industrial conditions, the water of the electrolyte evaporated severely, increasing KOH concentration.Therefore, when the electrolyte was significantly reduced, distilled water was added to the electrolyzer to recover its height level to the original value.
Bulk Electrolysis of Alcohols: The LSV and bulk electrolysis measurements were conducted in a three-electrode setup in analogy to the one used for OER tests. [90]The bulk electrolysis was investigated using the CA technique at a constant potential of 1.46 versus RHE in 10 mL KOH and 0.1 m alcohol.A bare NF and one with electrophoretically deposited LiFeBPO were used as the anode, platinum wire as the cathode, and Hg/HgO as the reference electrode.To investigate the performance of continuous electrooxidation of alcohol performance, the bulk electrolysis was performed at a constant potential of 1.46 versus RHE in 300 mL KOH and 0.1 m alcohols.To investigate the effects of alcohol concentration on the performance of producing FA by electrooxidation, bulk electrolysis (all 150 min CA at 1.46 V vs RHE) was conducted using LiFeBPO/NF in 10 mL 1 m KOH with a concentration gradient of ethylene glycol (0.1, 0.3, and 0.5 m).

NMR Analysis of the Organic Compounds:
The oxidation reaction solutions were characterized using 1 H and 13 C NMR spectroscopy conducted at a Bruker AV400 instrument.The NMR sample was prepared by mixing a 150 μL aliquot of the reaction solution, 0.1 m dimethyl sulfoxide (DMSO, internal standard), and 450 μL D 2 O solvent.MestReNova software was used to process and plot the spectra.At 4.7 ppm, a sharp peak can be observed, which represented the H 2 O from the aqueous reaction mixture.This peak was also used as a reference to reflect the chemical shifts of the other proton signals.
Calculation of the FE for the Oxidation of Alcohols: The FE was calculated based on the mol of formed formate and the associated passed charge.The product identification and conversion were determined by 1 H NMR spectroscopy using the DMSO as the internal standard of known concentrations.The FE of formate formation was calculated using the following equation FE (%) = mol of formate formed × F × n e total charge passed (Q) × 100 (1)   where F is the Faraday constant (96 485 C mol −1 ), n e is the number of electrons required for the oxidation process, which is 4 for methanol oxidation, 6 for ethylene glycol oxidation, and 8 for glycerol oxidation, respectively.

Figure 1 .
Figure 1.a) Rietveld refinement of PXRD pattern of the LiFeBPO single crystals, in which Rwp, Rp, and S represent the weighted profile parameter, unweighted profile parameter, and goodness-of-fit indicator, respectively.b) Representative LiFeBPO single crystal with a hexagonal bipyramidal feature, as well as its associated EDX mapping for c) Fe, d) P, and e) O.Such a homogenous elemental distribution, suggested the phase purity of LiFeBPO singlecrystals.f) TEM image with its corresponding SAED pattern (inset) verifying the highly crystalline nature of LiFeBPO (after ball milling).g) Schematic demonstration of LiFeBPO serving as an efficient electro(pre)catalyst for anodic oxidation reactions, where the magenta, blue, cyan, dark yellow, red, light pink, and dark brown spheres represent Li, Fe, B, P, O, H, and C atoms, respectively.

Figure 2 .
Figure 2. a) The LSV curves of LiFeBPO/NF, IrO 2 /NF, directly synthesized Fe-and NiFe-based (oxy)hydroxides deposited on NF, and bare NF normalized against geometric area under ambient conditions.b) Nyquist plots (obtained from electrochemical impedance spectroscopy, EIS, fitting to an equivalent circuit, which is shown in the inset) and c) C dl values of directly synthesized Fe-and NiFe-based (oxy)hydroxides supported on NF, as well as LiFeBPO/NF before and after d) 24 h CP at a constant current density of 100 mA cm −2 under ambient conditions.
2 O 3 catalysts supported on NF, as well as Fe x Ni 3-x O 4 (considering the possible influence of Ni from NF) under identical conditions (Figures S13-S16, Supporting Information), where LiFeBPO/NF displayed distinctly lower overpotential and better charge-transfer resistance (R ct = 1.117) (Figure 2a,b and Table

Figure 3 .
Figure 3. a) Ex situ Raman spectrum of LiFeBPO taken after electrophoretically deposited (EPD) on FTO (LiFeBPO/FTO) in comparison to the quasi-in situ Raman spectrum of post-OER LiFeBPO/FTO, which was freeze-quenched after 24 h OER CP (at 10 mA cm −2 in 1 m KOH and room temperature).b) TEM and c) the corresponding HRTEM images (inset: the associated FFT pattern), as well as d) HAADF-STEM image of LiFeBPO together with the elemental mapping of e) Fe, f) P (severe leaching can be found), and g) O after 24 h OER CP (at 10 mA cm −2 in 1 m KOH and room temperature).

Figure 4 .
Figure 4. a) The geometric area-normalized LSV curves of LiFeBPO/NF (+) ǁ NiPt 3 @NiS/NF (−) under ambient and industrial conditions.b) The CP curve of LiFeBPO/NF (+) ǁ NiPt 3 @NiS/NF (−) at a constant current density of 500 mA cm −2 under industrial conditions with an inset of the corresponding illustration.c) The LSV curves of LiFeBPO/NF and bare NF in 1 m KOH electrolyte with and without 0.1 m ethylene glycol; d) The bulk electrolysis (CA) curves of LiFeBPO/NF and bare NF in 1 m KOH electrolyte with 0.1 m ethylene glycol at a constant potential of 1.46 V (vs RHE); e) The 1 H NMR spectra of 0.1 m ethylene glycol electrolyte before bulk electrolysis, and the ones after electrolysis (at 1.46 vs RHE for 150 min) catalyzed by LiFeBPO/NF and bare NF; f) The LSV curves of LiFeBPO/NF in 1 m KOH electrolyte with 0.1 m methanol and 0.1 m glycerol; g) The electrooxidation scheme of various alcohols into formates catalyzed by LiFeBPO/NF (EG and MeOH represent ethylene glycol and methanol, respectively).