High-Entropy Spinel Oxides Produced via Sol-Gel and Electrospinning and Their Evaluation as Anodes in Li-Ion Batteries

: In the last few years, high-entropy oxides (HEOs), a new class of single-phase solid solution materials, have attracted growing interest in both academic research and industry for their great potential in a broad range of applications. This work investigates the possibility of producing pure single-phase HEOs with spinel structure (HESOs) under milder conditions (shorter heat treatments at lower temperatures) than standard solid-state techniques, thus reducing the environmental impact. For this purpose, a large set of HESOs was prepared via sol-gel and electrospinning (by using two different polymers). Ten different equimolar combinations of ﬁve metals were considered, and the inﬂuence of the synthesis method and conditions on the microstructure, morphology and crystalline phase purity of the produced HESOs was investigated by a combination of characterization techniques. On the other hand, the presence of speciﬁc metals, such as copper, lead to the formation of minority secondary phase(s). Finally, two representative pure single-phase HESOs were pre-liminarily evaluated as active anode materials in lithium-ion batteries and possible strategies to enhance their rate capability and cyclability were proposed and successfully implemented. The approaches introduced here can be extensively applied for the optimization of HEO properties targeting different applications


Introduction
The development of advanced materials with tailorable properties is a cornerstone of materials science and new technologies. Very recently, high-entropy materials (HEMs), a new class of materials obtained by using a configurational entropy-based design approach inherited from that of high-entropy alloys, have undergone an explosion of interest thanks to their unique structure and their novel and often unexpected functional properties [1][2][3][4]. HEMs are single-phase solid solutions stabilized in a multi-component system (generally, five or more elements in equimolar or near-equimolar ratio) by high configurational entropy. Numerous high-entropy compounds have been successfully synthesized and reported, preparation/fibrous film deposition and heat treatment) that, in an industrial production chain, can progress simultaneously. In syntheses via electrospinning, a technique particularly suitable for the manufacturing of fibrous nanostructured materials with high aspect-ratio and surface-area [41], two polymers (polyvinylpyrrolidone and polyacrylonitrile) are used to investigate the influence of the polymer on the formation of fibers and on their microstructure. Ten different equimolar combinations are considered, each containing five metals selected from Mg, Ti, Cr, Mn, Fe, Co, Ni, Cu and Zn. Their choice is operated by selecting cations with a similar ionic radius and a good chemical compatibility. The produced HESOs are analyzed by complementary techniques to investigate how the synthesis method and conditions influence their structure, morphology and crystalline phase. Since the results of their preliminary evaluation as active anode materials in LIBs indicate that the materials need to be improved by optimizing their preparation conditions, herein, two strategies are proposed to enhance the rate capability and cyclability of two representative samples. The first strategy is to add lithium into the HESO lattice, as the multi-valence nature of the HESO constituent elements promotes the formation of oxygen vacancies, known to favor Li + transport [32]. The second (conventional) strategy is to embed the HESO particles in a carbonaceous matrix [42][43][44]. Both the strategies are successfully implemented.

Reagents and Chemicals
The following reagents were utilized as metal sources for the preparation of HESOsamples: magnesium (II) acetate tetrahydrate, Mg(CH 3  Citric acid monohydrate, HOC(COOH)(CH 2 COOH) 2 ·H 2 O (purity: 98%; CAS No. 5949-29-1, Sigma Aldrich), was used as a complexing agent for the preparation of HESOsamples via the sol-gel method.
All chemicals were used without any further purification. Table 1 reports the selected equimolar combinations of five metals and the codes of the HESO-samples produced. The suffix sg marks the HESOs prepared by the sol-gel method. In the case of the electrospun HEOs, the suffixes a and b identify the samples prepared by using polyvinylpyrrolidone (PVP) or polyacrylonitrile (PAN) as a polymer, respectively. In the syntheses of HESO samples via the sol-gel (SG) method, stoichiometric amounts of the selected five metals were dissolved in 30 g water, one at a time ( Figure S1a). After magnetic stirring at 350 rpm for 1 h, 3.5 g citric acid was added and the resulting mixture was further stirred (350 rpm) at 90 • C until a gel was formed (about 1 h). The as-obtained gel was dried at 80 • C overnight and, subsequently, calcined in static air at 350 • C for 2 h, at 500 • C for 2 h, at 800 • C for 2 h and at 900 • C for 2 h (Figure S1(c1)), finally obtaining powders consisting of HESO particles ( Figure S1(d1)), having spinel structure as a unique/primary phase ( Figure S1e). Note that the treatment at the highest temperature (lower than in the conventional solid reaction [16,[33][34][35]) lasts only 2 h, i.e., a much shorter time than that necessary to produce HESOs by the frequently utilized solid-state approach (typically 12-24 h [16,[33][34][35]45,46]). The same consideration applies to the case of electrospun samples.

Syntheses of HESO-Samples
In the case of the lithiated sample (Li-HESO), Li(CH 3 COO) was dissolved in an aqueous solution containing the selected five metals salts. Then, the same preparation steps as for the unlithiated samples were carried out.

Syntheses by Electrospinning
In syntheses via electrospinning (ES), two different polymer/solvent solutions were evaluated, namely PAN/N,N-dimethylformamide (DMF) and PVP/DMF+ethanol (EtOH). In the latter case, the spinnable solution was obtained by dissolving PVP (25 wt%) in EtOH and the metal salts were separately dissolved in DMF. Both solutions were magnetically stirred at 350 rpm at room temperature (RT) until they became clear. Then, they were mixed under continuous stirring (350 rpm) for 2 h. The PAN/DMF solution was prepared by dissolving PAN (7.1 wt%) in DMF. Then, proper amounts of the metallic salts were added ( Figure S1a), one at a time, under continuous stirring (350 rpm) until a homogeneous spinnable solution containing the selected equimolar combination of five metals was obtained (about 1 h). Regardless of the metal combination selected, the total amount of metals was kept constant (38.5 wt% relative to PAN and 24.7 wt% relative to PVP). After slow mixing overnight at RT, the solution was ready for deposition.
Each homogeneous spinnable solution obtained was loaded into a 20 mL syringe equipped with a 40 mm long 0.8 mm gauge stainless-steel needle. Electrospinning (Figure S1c2) was carried out, at 22 ± 1 • C and 40% relative air humidity, by using a CH-01 Electro-spinner 2.0 (Linari Engineering s.r.l.). The solution feeding rate, applied DC voltage and needle-collector distance were fixed at 23.5 µL min −1 , 15 kV and 11 cm, respectively. The pristine polymer/metal salts nanofibers (NFs) were ejected from the needle and collected over an aluminum foil to form a non-woven fibrous membrane.
After drying at RT overnight to remove the solvent residuals, the as-spun membrane was peeled from the collector and calcined in static air ( Figure S1(c2)). A four-step heating process was operated in the case of PVP-derived HESO-samples (350 • C for 2 h, 500 • C for 2 h, 800 • C for 2 h and 900 • C); PAN-derived HESO-samples were obtained through a two-step heat-treatment (700 • C for 2 h and 900 • C for 2 h). The obtained powders consisted of porous architectures, generally built up by HESO-NFs ( Figure S1(d2)) with spinel structure as a unique/primary phase ( Figure S1e).
In addition to pure HEO-samples, a composite carbon/HESO sample was prepared. For this purpose, the same PAN/DMF-based solution as for the corresponding pure HESO was used, but after the as-spun membrane had been detached from the collector, it was first stabilized in air at 225 • C [42] for 2 h, and subsequently carbonized under inert atmosphere through the same two-step heat-treatment as for pure HESO samples, so that the oxide grains formed in the composite NFs experienced the same temperatures as in those of pure oxide.

Characterization
After thermal processing, the produced HESO-samples were analyzed by means of a number of complementary techniques. A Phenom Pro-X scanning electron microscope, equipped with an energy-dispersive X-ray (EDX) spectrometer, was used to investigate the texture and morphology of HESO-samples by means of scanning electron microscopy (SEM). High-resolution transmission electron microscopy (HRTEM), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), selected-area electron diffraction (SAED) and energy dispersive X-ray spectroscopy (EDX) elemental mappings were carried out on a FEI Talos F200S scanning/transmission electron microscope, operated at 200 kV.
The crystalline phases of the HESO-samples were identified by X-ray diffraction (XRD) and micro-Raman spectroscopy (MRS). A Bruker D2 Phaser diffractometer, equipped with a Ni β-filtered Cu-Kα radiation source, was used to record XRD patterns, whereas JCPDS database of reference compounds allowed for the diffraction-peak identification. A NT-MDT NTEGRA Spectra SPM spectrometer, equipped with MS3504i 350 mm monochromator and ANDOR Idus CCD, was used to measure Raman scattering excited by a solid-state laser operating at 532 nm. Raman measurements were carried out in reflection mode using a Mitutoyo high numerical aperture 100× objective to excite and collect the scattered signal from the sample surface. The acquired signal was dispersed by an optical grating (600 lines/mm) and, finally, detected by the cooled CCD. The use of a very low laser power (250 µW at the sample surface) prevented local heating of the samples and annealing effects.

Electrochemical Characterization
Electrochemical properties of HEOs were evaluated using CR2032-type coin cells, which were assembled in an Ar-filled glovebox. First, 70% active materials (HESO), 20% conductive carbon black (Super P, Timcal, Billerica, MA, USA) and 10% polyvinylidene fluoride (PVDF, Alfa Aesar, Karlsruhe, Germany) were homogeneously mixed with Nmethyl-2-pyrrolidone (NMP, anhydrous 99.5%, Sigma-Aldrich, Steinheim, Germany). The resulting slurry was uniformly cast on Cu foil (Goodfellow, Hamburg, Germany) with a doctor blade apparatus and then dried in a vacuum oven at 60 • C for 1 h. After a cold-laminating step, electrodes with a diameter of 12 mm were punched out and dried overnight at 120 • C under vacuum using a Büchi glass oven. The mass loading of active material was ≈ 1.5 mg cm −2 . Lithium metal foil was used as both counter and reference electrodes. 1 M LiPF 6 (99.9% battery grade, ABCR, Karlsruhe, Germany) solution in a mixture of ethylene carbonate (EC, 99.9%, ABCR), diethyl carbonate (DEC, 99.9%, ABCR) and dimethyl carbonate (DMC, 99.9%, ABCR) with a 1:1:1 volume ratio was used as the electrolyte. A glass microfiber filter (Whatman, Steinheim, Germany) was used as a separator.
Galvanostatic charge-discharge (GCD) cycling was carried out at room temperature using a CT2001 A battery testing system (Landt Instruments, Wuhan, China). Cyclic voltammograms (CVs) were measured on a Bio-Logic (Seyssinet-Pariset, France) VMP3 multichannel potentiostat/galvanostat with a built-in electrochemical impedance spectroscopy (EIS) analyzer. EIS was tested at open-circuit potential (OCP), in the frequency range of 100 kHz to 0.1 Hz with an amplitude of 10 mV.  Figure S2 shows the NF diameter distributions as automatically estimated by the analysis software of the SEM microscope. In PVP-derived samples, the distributions are centred on values increasing in the order 5MK-a (580 nm) < 5ME-a (630 nm) < 5ML-a (770 nm). In PAN-derived samples, the distribution centre values increase in the order 5MG-b (380 nm) < 5MF-b (510 nm) < 5MD-b (520 nm) < 5ME-b (660 nm) < 5ML-b (720 nm). The comparison of the results relative to samples 5ME-a and 5ML-a with those relative to samples 5ME-b and 5ML-b suggests a possible role of the metal combination in determining the variation of the NF diameter.

Morphology
Regardless of the sample morphology, preparation method and conditions, the dispersion of the composing elements is always spatially uniform within the samples, as revealed by SEM/EDX analysis ( Figure S3). Figure 2 displays some representative HRTEM images of the investigated samples. Images with lower magnification are shown in Figure S4a-d. Regardless of their composition, samples produced by the SG route consist of aggregates of crystalline oxide particles (see the SAED pattern in the inset of Figure S4a), having polyhedral shape and size ranging between a few tens of nm to hundreds of nm (Figure 2a,e). The electrospun PAN-derived NFs are micrometers long and have diameters ranging between 150 and 450 nm. Independent of their composition, they are built-up by interconnected crystalline oxide grains (see the SAED patterns in the insets to Figure S4b-d) with similar shape and size to that of particles synthesized by the SG method ( Figure 2b-d,f-h). For both synthesis routes, the STEM/EDX elemental maps ( Figure S5) prove the homogeneity of the spatial distribution of metals and oxygen at the nanometer scale.        [45,[47][48][49][50][51][52].

Phase of the Oxide
In most samples, no additional reflections are detected, which proves that the spinel phase is the unique crystalline phase formed. On the contrary, in samples 5MA-sg, 5MKsg, 5ML-sg (Figure 3a), 5MKa, 5ML-a (Figure 3b), 5ML-b and 5MD-b (Figure 3c), the detection of weaker additional reflections reveals the formation of secondary phases. Interestingly, this seems to be a common feature of the diffractograms of samples having the same nominal composition (e.g., 5ML-sg, 5ML-a and 5ML-b). The reflections, detected in the XRD patterns of 5ML-sg, 5ML-a and 5ML-b at 32.6°, 38.6°, 48.6°, 61.6°, 66.0° and 67.9° (owing to the scale, only the most intense of them are visible in Figure 3), are from  [53,54]. The presence of Cu among the five selected metals promotes the In most samples, no additional reflections are detected, which proves that the spinel phase is the unique crystalline phase formed. On the contrary, in samples 5MA-sg, 5MK-sg, 5ML-sg (Figure 3a), 5MKa, 5ML-a (Figure 3b), 5ML-b and 5MD-b (Figure 3c), the detection of weaker additional reflections reveals the formation of secondary phases. Interestingly, this seems to be a common feature of the diffractograms of samples having the same nominal composition (e.g., 5ML-sg, 5ML-a and 5ML-b). The reflections, detected in the XRD patterns of 5ML-sg, 5ML-a and 5ML-b at 32.  [53,54]. The presence of Cu among the five selected metals promotes the formation of CuO as a secondary phase also in sample 5MA-sg and 5MH-sg. In 5MI-sg and 5MJ-sg, a weak reflection from the 112 planes of α-Mg is detected at 66 • [55,56].
The behavior of sample 5MD-b differs from that of the samples with the same nominal composition. Reflections 222, 400 and 440 seem to be split owing to the appearance of very close signals at ca. 36.8 • , 42.7 • and 62.2 • . Their detection might be indicative either of the existence of two spinel sublattices or of the segregation of some metal(s) in a secondary phase. In the latter case, the reflections might originate from 111, 200 and 220 planes of nickel oxide (NiO, JCPDS No. 04-0835) [59,60] and/or cobalt oxide (CoO, JCPDS No. 78-0431) [61], both characterized by the same fcc crystalline structure.
Rietveld refinements from XRD data ( Figure S6) were carried out to assess the relative amount of the secondary phase(s). The results obtained are reported in Table S1. In samples produced via the SG method, secondary phases reach 12.4%, while in electrospun samples their percentage is 7.3-8.8%.
For all samples, the mean sizes of the HESO crystallites are calculated from the most intense reflection through the Scherrer's equation [60]. The values obtained, reported in Figure 3, range from 19.2 to 38.7 nm for HESOs produced via the SG method. For electrospun samples, the mean sizes of the crystallites vary in the ranges 20.5-44.5 nm and 15.9-27.4 nm, for PVP-and PAN-derived HESOs, respectively.
For a given metal combination, the observed variations of crystallite sizes do not follow any general trend or dependence on the preparation method. As an example, in the 5MD-series, the mean crystallite size increases in the order 5MD-sg (22.9 nm) < 5MD-a (43.7 nm) < 5MD-b (43.9 nm), whereas in the 5MF-series, it varies in the order 5MF-sg (38.7 nm) > 5MF-b (22.5 nm) > 5MF-a (20.5 nm).
Additional information on the phase-purity of the prepared samples and its spatial homogeneity is inferred by measuring Raman scattering from several random locations in each specimen ( Figure S7). No appreciable differences in the relative intensities and in the positions of the bands at the probed locations are observed in pure single-phase samples ( Figure S7a,b). On the contrary, changes in the spectral profile at a random location indicate local deviations from the average chemical composition and segregation in a secondary phase ( Figure S7c). The obtained results by these measurements are in agreement with the indications emerged from the XRD analysis. Figure 4 shows the averaged Raman spectra of pure single-phase samples, obtained from the spectra acquired at random locations in each sample. These spectra provide a global picture of the entire sample. The vibration modes theoretically predicted for the spinel-group [62,63] include five Raman-active phonon modes, namely three triply degenerate modes having F 2g symmetry, one doubly degenerate mode with E g -symmetry and an A 1g -symmetry mode. The lowest frequency F 2g (1) mode might be due to the translation motion of the BO group [64] or originate from a complete translation of the AO 4 unit within the spinel lattice [65,66]. The E g mode is associated with symmetric bending of O anion with respect to octahedral B cation [67,68]. Higher frequency F 2g (2) and F 2g (3) modes arise from the asymmetric stretching and asymmetric bending of oxygen, respectively [65,67,69]. Finally, the A 1g mode is attributed to the vibration along the direction joining an oxygen atom to the tetrahedral A cation. In the A 2+ B 3+ 2O4 spinel structure, cations occupy both octahedral and tetrahedral sites. Depending on the sites occupied by the A and B cations, normal, inverse or partial inverse spinel structures can be formed [65]. Frequencies and intensities of the Raman bands are sensitive to the type of cations, A 2+ or B 3+ , that occupy octahedral and tetrahedral sites. For example, the spectrum of chromate spinels (ACr2O4) changes with changing A cations [65]. All predicted Raman active modes are detected in the spectrum of ZnCr2O4, whereas the lowest wavenumber band F2g(1) is absent in that of MgCr2O4. By comparing the spectra of FeCr2O4 and MgCr2O4, noticeable changes in the relative intensity of bands are observed. Since in the chromate spinels, A 2+ and Cr 3+ cations are ordered in the tetrahedral and octahedral coordinated sites respectively, the observed variations in the Raman spectra can be attributed to the different divalent cations [65]. The same considerations hold if the trivalent cation changes. This may even cause the appearance of new vibration modes, as reported for FeAl2O4 and FeCr2O4 [65]. In addition, the presence of defects that generate distortions of the crystalline lattice contributes to shift the Raman bands and change their relative intensities [66].
However, in the case of HESOs, assessing the type of sites occupied by cations in the lattice is made even more difficult than in common spinels due to the inherent complexity of five-cation systems. In the present samples, the F2g(1) mode is detected between 160 and 220 cm -1 , whereas the very weak Eg mode is located between 250 and 420 cm -1 . Their relative intensities and frequency positions do not appreciably vary with the metal combinations (compare spectra shown in Figure 4). This finding points to the involvement of cations with oxidation state 3+ (namely Mn 3+ and Fe 3+ ), present in the samples of all series (5MD/E/F/G). The higher frequency region of the spectra (>450 cm -1 ), where the F2g(2), F2g(3) and A1g modes are detected, seems to be more sensitive to the changes in metal combinations (compare the spectra of samples produced by the same method in Figure 4a). Conversely, samples produced by different methods or polymers, but having the same nominal composition, exhibit quite similar spectral profiles (compare the spectra of samples 5MD-sg, 5MD-a and 5MD-b). The case of the series 5ME represents the only exception and deserves further discussion. In the spectra of the electrospun samples, similar to each other, a weak shoulder is observed on the higher frequency side of the A1g mode (at 625 cm -1 ), as shown for sample 5ME-a (Figure 4a). The splitting of the A1g mode into two branches is more sharply visible in the spectrum of sample 5ME-sg (Figure 4a), which closely resembles that of (Co,Cr,Fe,Mn,Ni)3O4 HESO synthesized via the solid-state reaction route [47]. This suggests that inversion may occur to some extent, owing to the replacement of some trivalent (divalent) cations by divalent (trivalent) cations in the octahedral (tetrahedral) sites [47,65]. In the A 2+ B 3+ 2 O 4 spinel structure, cations occupy both octahedral and tetrahedral sites. Depending on the sites occupied by the A and B cations, normal, inverse or partial inverse spinel structures can be formed [65]. Frequencies and intensities of the Raman bands are sensitive to the type of cations, A 2+ or B 3+ , that occupy octahedral and tetrahedral sites. For example, the spectrum of chromate spinels (ACr 2 O 4 ) changes with changing A cations [65]. All predicted Raman active modes are detected in the spectrum of ZnCr 2 O 4 , whereas the lowest wavenumber band F 2g (1) is absent in that of MgCr 2 O 4 . By comparing the spectra of FeCr 2 O 4 and MgCr 2 O 4 , noticeable changes in the relative intensity of bands are observed. Since in the chromate spinels, A 2+ and Cr 3+ cations are ordered in the tetrahedral and octahedral coordinated sites respectively, the observed variations in the Raman spectra can be attributed to the different divalent cations [65]. The same considerations hold if the trivalent cation changes. This may even cause the appearance of new vibration modes, as reported for FeAl 2 O 4 and FeCr 2 O 4 [65]. In addition, the presence of defects that generate distortions of the crystalline lattice contributes to shift the Raman bands and change their relative intensities [66].
However, in the case of HESOs, assessing the type of sites occupied by cations in the lattice is made even more difficult than in common spinels due to the inherent complexity of five-cation systems. In the present samples, the F 2g (1) mode is detected between 160 and 220 cm −1 , whereas the very weak E g mode is located between 250 and 420 cm −1 . Their relative intensities and frequency positions do not appreciably vary with the metal combinations (compare spectra shown in Figure 4). This finding points to the involvement of cations with oxidation state 3+ (namely Mn 3+ and Fe 3+ ), present in the samples of all series (5MD/E/F/G). The higher frequency region of the spectra (>450 cm −1 ), where the F 2g (2), F 2g (3) and A 1g modes are detected, seems to be more sensitive to the changes in metal combinations (compare the spectra of samples produced by the same method in Figure 4a). Conversely, samples produced by different methods or polymers, but having the same nominal composition, exhibit quite similar spectral profiles (compare the spectra of samples 5MD-sg, 5MD-a and 5MD-b). The case of the series 5ME represents the only exception and deserves further discussion. In the spectra of the electrospun samples, similar to each other, a weak shoulder is observed on the higher frequency side of the A 1g mode (at 625 cm −1 ), as shown for sample 5ME-a (Figure 4a). The splitting of the A 1g mode into two branches is more sharply visible in the spectrum of sample 5ME-sg (Figure 4a), which closely resembles that of (Co,Cr,Fe,Mn,Ni) 3 O 4 HESO synthesized via the solid-state reaction route [47]. This suggests that inversion may occur to some extent, owing to the replacement of some trivalent (divalent) cations by divalent (trivalent) cations in the octahedral (tetrahedral) sites [47,65].

Electrochemical Properties
The superior lithium-storage properties of rock-salt (Mg,Co,Ni,Cu,Zn) HEO, produced by solid-state reaction from an equimolar mixture of MgO, CoO, NiO, CuO and ZnO [11], as anode material have been pointed out by many groups [16,31]. For example, it was shown that such an HEO can deliver a reversible capacity of 920 mAh g −1 after 300 cycles at 100 mA g −1 and an excellent rate capability (600 mAh g −1 at 2 A g −1 ) [16]. Furthermore, electrospun (Mg 0.2 Co 0.2 Ni 0.2 Cu 0.2 Zn 0.2 )O NFs exhibit excellent cycling stability (390 mAh g −1 after 300 cycles at 500 mA g −1 ) [36].
Here, two representative HESOs with the same nominal composition, (Cr 0.2 Mn 0.2 Fe 0.2 -Co 0.2 Zn 0.2 ) 3 O 4 , are considered: 5MF-sg, produced by SG route and; 5MF-b, prepared by ES using PAN as a polymer. Their electrochemical behavior as active anode materials in LIBs is preliminarily evaluated by carrying out measurements in the potential window of 0.01-3.0 V (versus Li/Li + ). Figure 5 displays the results obtained.
Here, two representative HESOs with the same nominal composition, (Cr0.2Mn0.2Fe0.2Co0.2Zn0.2)3O4, are considered: 5MF-sg, produced by SG route and; 5MF-b, prepared by ES using PAN as a polymer. Their electrochemical behavior as active anode materials in LIBs is preliminarily evaluated by carrying out measurements in the potential window of 0.01-3.0 V (versus Li/Li + ). Figure 5 displays the results obtained.   Figure 5a,b shows the cyclic voltammetry (CV) curves of the electrodes based on (Cr,Mn,Fe,Co,Zn) HESO NPs (5MF-sg) and (Cr,Mn,Fe,Co,Zn) HESO NFs (5MF-b) in the initial five cycles at a scan rate of 0.1 mV s −1 . These electrode materials display similar CV patterns, suggesting similar electrochemical processes of the two HESOs. The multiple anodic and cathodic peaks are associated with the Li + insertion-deinsertion process. In the first cycle, there is a strong cathodic peak located at around 0.4 V; this can be ascribed to the formation of solid electrolyte interphase (SEI) due to the decomposition of electrolyte [32]. The weaker cathodic peak at higher potential arises from the reduction of the HESO cations [71]. In the following anodic scan, the oxidation peaks may refer to the re-oxidation of metal elements into oxides. In the second cycle, the cathodic peak shifts to a higher potential and its intensity is significantly reduced, which might be caused by the irreversible reaction of the first lithiation process and pulverization of the electrode materials [16,72]. In the subsequent cycles, both the reduction and oxidation peaks are not overlapped very well, suggesting that the HESO anodes do not exhibit good stability and reversibility for the Li + insertion and extraction reactions.
To evaluate the capacity of the considered HESOs, galvanostatic charge-discharge tests were carried out at a current density of 20 mA g −1 (Figure 5c,d). In the first lithiation process, there is a clear potential plateau at around 0.45 V, corresponding to the reduction reaction of transition metals and the formation of the SEI layer, which is consistent with CV curves. The initial lithiation/delithiation capacities of anodes 5MF-b and 5MF-sg are 1297/950 and 1171/785 mAh g −1 , which correspond to initial Coulombic efficiencies (CE) of 73.2% and 60.5%, respectively. Higher capacities pertain to the HESO NFs. This is because electrospun fibrous structures provide direct current pathways along their axis, favoring the enhancement of capacity [36]. The initial capacity loss is associated with the SEI formation due to the decomposition of the electrolyte and structural rearrangement [48,72,73]. In the following cycles, all three samples show slopes rather than obvious plateaus during the discharge process, which could be associated with the different reaction voltage of multiple metals and structure rearrangement [45].
The rate capabilities of anodes 5MF-b and 5MF-sg were also evaluated, as shown in Figure 5e,f. With increasing current density, both samples show a fast capacity decay, implying a poor stability. The results of the long-term stability evaluation further prove that their structure is not stable (Figure 5g), possibly due to the pulverization of the HESO grains induced by the large volume changes occurring during the lithiation/delithiation process. The evaluation of the other pure single-phase samples (not shown for briefness) leads to similar results.

Incorporating Lithium in the HESO Lattice
Lökçu et al. [74] have recently evaluated the electrochemical performance of (MgCoNi-Zn) 1−x Li x O, prepared by solid state reaction from milled MgO, CoO, NiO, ZnO, and Li 2 O oxides, as anode material in LIBs. They have reported that the incorporation of lithium in the HEO lattice generates oxygen vacancies (OVs) and the higher the OV concentration in the anode, the higher its specific capacity. In particular, (MgCoNiZn) 0. 65 Li 0.35 O anode still delivers 610 mAh g −1 after 130 cycles at 1 A g −1 [74].
Here, a lithiated sample with nominal composition Li 1.8 (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O x , coded as Li5MF-sg, is produced by the SG route by adding Li salt to the same solution utilized to prepare sample 5MF-sg. The incorporation of lithium in the HESO lattice leads to the formation of greater oxide particles (compare SEM, TEM and HRTEM images of samples 5MF-sg and Li5MF-sg in Figures S8 and S9), without affecting the spatial distribution of the elemental constituents ( Figure S9e), the oxide phase (compare XRD patterns of the two samples in Figure S10a) and its spatial uniformity, as evaluated by MRS ( Figure S11). Nevertheless, by comparing the micro-Raman spectra of the samples (Figure S10b), some differences come into view. In Li5MF-sg, a shoulder appears at 479 cm −1 , while the F 2g (1) and A 1g modes downshift (the apparent peak positions move from 189 and 654 cm −1 to 181 and 634 cm −1 , respectively). Both these spectral changes are indicative of the generation of OVs induced by the incorporation of Li in the oxide lattice [75][76][77]. Figure 6 compares the electrochemical properties of anodes 5MF-sg and Li5MFsg. Overall, CV and charge-discharge curves of Li 1.8 (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O x NPs (Figure 6b,d) do no not substantially differ from those of (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O 4 NPs (Figure 6a,c), which is indicative of similar electrochemical processes. However, the strong reduction peak at the first cycle shifts to higher potential, and the peak intensity is stronger in the following cycles than that of 5MF-sg, which could be caused by Li incorporation in the HESO lattice. As reported in literature [46], the electrochemical process can be tailored by the incorporation of different elements in high-entropy oxide structures. The initial lithiation/delithiation capacities of anode Li5MF-sg (1112/733 mAh g −1 ) are slightly lower than for anode 5MF-sg (1171/785 mAh g −1 ), probably due to the larger particle size. Nonetheless, the corresponding initial CE is higher (66% against 60.5%); this value is comparable to that reported for (FeCoNiCrMnZnLi) 3 O 4 prepared using the standard solid-state method (65%) [46]. Appl. Sci. 2022, 12, 5965 13 of 19 cm −1 , while the F2g(1) and A1g modes downshift (the apparent peak positions move from 189 and 654 cm −1 to 181 and 634 cm −1 , respectively). Both these spectral changes are indicative of the generation of OVs induced by the incorporation of Li in the oxide lattice [75][76][77]. Figure 6 compares the electrochemical properties of anodes 5MF-sg and Li5MF-sg. Overall, CV and charge-discharge curves of Li1.8(Cr0.2Mn0.2Fe0.2Co0.2Zn0.2)3Ox NPs ( Figure  6b,d) do no not substantially differ from those of (Cr0.2Mn0.2Fe0.2Co0.2Zn0.2)3O4 NPs ( Figure  6a,c), which is indicative of similar electrochemical processes. However, the strong reduction peak at the first cycle shifts to higher potential, and the peak intensity is stronger in the following cycles than that of 5MF-sg, which could be caused by Li incorporation in the HESO lattice. As reported in literature [46], the electrochemical process can be tailored by the incorporation of different elements in high-entropy oxide structures. The initial lithiation/delithiation capacities of anode Li5MF-sg (1112/733 mAh g −1 ) are slightly lower than for anode 5MF-sg (1171/785 mAh g −1 ), probably due to the larger particle size. Nonetheless, the corresponding initial CE is higher (66% against 60.5%); this value is comparable to that reported for (FeCoNiCrMnZnLi)3O4 prepared using the standard solid-state method (65%) [46]. Figure 6e,f compares the results of the rate capability test carried out on anodes 5MFsg and Li5MF-sg. With increasing current densities, the capacities gradually decrease because of the increased electrode polarization. Compared with 5MF-sg, Li5MF-sg Figure 6e,f compares the results of the rate capability test carried out on anodes 5MFsg and Li5MF-sg. With increasing current densities, the capacities gradually decrease because of the increased electrode polarization. Compared with 5MF-sg, Li5MF-sg achieves larger capacities and superior rate capability. When tested at 20, 50, 100, 200, 500 and 1000 mA g −1 , the Li5MF-sg electrode delivers good lithiation specific capacities of 702, 501, 388, 300, 213, 169 and 146 mAh g −1 , respectively. Additionally, when this sample is cycled at 500 mA g −1 for 300 cycles (Figure 6g), a reversible capacity as high as 484 mAh g −1 is reached, which is much higher than that of 5MF-sg (150 mAh g −1 ) and close to that reported for (FeCoNiCrMnZnLi) 3 O 4 prepared using the standard solid-state method (522 mAh g −1 ) after 100 cycles at the same rate [46]. The slightly increasing trend of specific capacity seems to be peculiar to materials with large particles [46].
The improved cycling stability could be ascribed to the increased configurational entropy deriving from the presence of the sixth metal in the combination and OVs generated, which induces the phase stabilization effect in the HESO [52,78]. The enhanced rate capability can be attributed to the improved ionic conductivity associated with the existence of a large number of OVs [46]. Figure 6h shows the results of EIS measurements performed to further evaluate Li + transfer behavior of the two anodes. Nyquist plots show a semicircle in the high-to-medium frequency region related to charge transfer resistance at the electrode/electrolyte interface; the straight line in the low-frequency region is associated with the Warburg impedance, which depends on the solid-state diffusion of Li + -ions in the bulk electrode material. For anode Li5MF-sg, the semicircle diameter is much smaller than that of 5MF-sg, while the slope of the straight line is higher. These findings indicate greatly reduced charge transfer resistance and faster Li + transport in

Incorporating the Oxide in a Carbonaceous Matrix
Incorporating the active oxide particles in a carbonaceous matrix is a commonly adopted strategy to improve the electrochemical performance of anode materials prepared using different methods [43,44]. The carbon matrix favours the electrical transport and withstands the large volume changes, preventing the pulverization of the oxide particles.
Here, a composite carbon/HESO sample, coded as C5MF-b, is produced by ES followed by stabilization and carbonization. The sample, with nominal composition C/(Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O 4 , exhibits a fibrous structure, similar to pure HESO (Figures S12 and S13). TEM/EDX analysis ( Figure S14) reveals that the sample contains both C-poorer (Figure S13a,c) and C-richer NFs ( Figure S14b,d). In the former, the oxide grains are generally larger ( Figure S13a), whereas in the latter, they are finer ( Figure S13b-d), in agreement with the well assessed ability of carbon to reduce the size of oxide grains and prevent their agglomeration [42][43][44]. In the micro-Raman spectrum of sample C5MF-b ( Figure S15), the Raman fingerprints of both amorphous carbon [35,36] and (Cr,Mn,Fe,Co,Zn) HESO particles are detected. Figure 7 compares the electrochemical properties of anodes 5MF-b and C5MF-b. Looking at their CV profiles (Figure 7a,b), it can be seen that the peak intensity and CV curve area of C5MF-b are higher than that of 5MF-b, implying that C/HESO NFs have better electrochemical kinetics [70]. Moreover, the second and third charge/discharge curves of anode C5MF-b are overlapped very well (Figure 7d), unlike those of 5MF-b (Figure 7c). This indicates that the composite NFs have better stability and reversibility for Li + insertion and extraction reactions than the pure HESO NFs [43]. Due to the specfic capacity of carbon (372 mAh g −1 ) [42][43][44], the C/HESO NFs exhibit lower initial lithiation/delithiation capacities compared to HESO NFs (1159/777 against 1297/950 mAh g −1 ). Figure 7e,f compares the results of the rate capability test carried out on anodes 5MF-b and C5MF-b. The measured delithiation capacities of C/HESO NFs are 834, 835, 784, 727, 610, 526, 471 and 345 mAh g −1 at 20, 50, 100, 200, 500, 1000 and 2000 mA g −1 , respectively. These values are higher than those measured in pure HESO NFs, especially at high current density, as expected [42][43][44]. In the long-term cyclability test (Figure 7g), the capacity decreases during the first 100 cycles and then increases. These capacity changes can be explained by the activation process and structural reconstruction associated with the decreases during the first 100 cycles and then increases. These capacity changes can be explained by the activation process and structural reconstruction associated with the active material [31]. After 400 cycles at a rate of 500 mA g −1 , C/HESO NFs deliver a capacity of around 545 mAh g −1 , which is much higher than that of 5MF-b (265 mAh g −1 ). The results of EIS measurements (Figure 7h) evidence that the charge transfer resistance of C5MF-b is significantly lower than that of 5MF-b, confirming that the improvement in the electrochemical performance can be ascribed to the enhanced eletronic conductivity and mechanical stability promoted by carbon rather than to changes in the ionic transport.

Conclusions
This work demonstrates the possibility of producing pure single-phase HEOs with spinel structure (HESOs) under milder conditions than the standard solid-state techniques (i.e., with reduced environmental impact). Both XRD and Raman analyses highlight that the presence of copper or titanium in the equimolar combinations of the five metals considered favors the formation of secondary phase(s). SEM analysis shows that in ES syntheses, greater selectivity towards the NF formation is obtained using PAN as the polymer.
Pure single-phase HESOs are evaluated as active anode materials in lithium-ion batteries. They exhibit great potential (e.g., very high initial specific capacities) but poor rate Active material [31]. After 400 cycles at a rate of 500 mA g −1 , C/HESO NFs deliver a capacity of around 545 mAh g −1 , which is much higher than that of 5MF-b (265 mAh g −1 ). The results of EIS measurements (Figure 7h) evidence that the charge transfer resistance of C5MF-b is significantly lower than that of 5MF-b, confirming that the improvement in the electrochemical performance can be ascribed to the enhanced eletronic conductivity and mechanical stability promoted by carbon rather than to changes in the ionic transport.

Conclusions
This work demonstrates the possibility of producing pure single-phase HEOs with spinel structure (HESOs) under milder conditions than the standard solid-state techniques (i.e., with reduced environmental impact). Both XRD and Raman analyses highlight that the presence of copper or titanium in the equimolar combinations of the five metals considered favors the formation of secondary phase(s). SEM analysis shows that in ES syntheses, greater selectivity towards the NF formation is obtained using PAN as the polymer.
Pure single-phase HESOs are evaluated as active anode materials in lithium-ion batteries. They exhibit great potential (e.g., very high initial specific capacities) but poor rate capability and long-term stability. Hence, two possible strategies are proposed and successfully implemented to improve their electrochemical performance, as shown for two representative samples. The common strategy of incorporating carbon is demonstrated in the case of electrospun (Cr,Mn,Fe,Co,Zn) HESO, while the incorporation of lithium in the HESO lattice is applied to the case of (Cr,Mn,Fe,Co,Zn) HESO produced via the sol-gel method.
The addition of Li enhances entropy and generates oxygen vacancies, thus improving stability and ionic transport. After 300 cycles at 0.5 A g −1 , Li 1.8 (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O x NPs deliver a much higher capacity than (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O 4 NPs (484 against 150 mAh g −1 ). The addition of C improves the electronic transport and mechanical stability. At 2 A g −1 rate, in composite C/(Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O 4 NFs, the measured delithiation capacity increases (from 15) to 345 mAh g −1 . Furthermore, after 400 cycles at a rate of 0.5 A g −1 , they still deliver 545 mAh g −1 against 265 mAh g −1 of pure (Cr 0.2 Mn 0.2 Fe 0.2 Co 0.2 Zn 0.2 ) 3 O 4 NFs.
Despite the presented results indicating the need to better tune the preparation conditions to improve the rate capability and cyclability, this preliminary study provides a further contribution to the understanding of HEOs. This is of great importance since these materials can potentially be utilized for the development of all components of rechargeable Li + and Na + ion batteries. Moreover, thanks to their generality, the approaches shown can be extensively applied to the optimization of HEO properties, even aimed at different applications.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/app12125965/s1, Figure S1: Schematic description of the experimental procedure followed to synthesize the HEO-samples; Figure S2: NF diameter distributions; Figure S3: Results of the SEM/EDX analysis on the most representative samples; Figure S4: TEM images; Figure S5: Results of the STEM/EDX analysis; Figure S6: Rietveld refinements; Table S1: results of Rietveld refinements; Figure S7: Micro-Raman spectra of representative samples; Figure S8: SEM and TEM images of samples 5MF-sg and Li5MF-sg; Figure S9: HRTEM images of samples 5MFsg and Li5MF-sg; Figure S10: XRD and MRS images of samples 5MF-sg and Li5MF-sg; Figure S11: Micro-Raman spectra of samples 5MF-sg and Li5MF-sg; Figure S12: TEM images of sample C5MFb. Inset: SAED pattern; Figure S13: HRTEM images of sample C5MF-b; Figure S14: Results of the STEM/EDX analysis on sample C5MF-b; Figure S15: Micro-Raman spectra of samples 5MF-b and C5MF-b.

Data Availability Statement:
The raw data and processing aids, required to reproduce the findings of this work, cannot be shared at this time as they form the basis for an ongoing study.