Field of the Invention

The present invention relates to V-Nb-Mo oxide and carbon-MOx (M = V-Nb-Mo) composite materials for use as high performance (high power, high energy) anodes in lithium ion or sodium ion batteries and hybrid capacitors.

Background

The present invention has been devised in light of the above considerations.

Due to increasing interest in electrification of transport and the development of smart grids, flexible and stable high energy and ultra-high power energy storage devices are required. ¹ Some of that interest has focused on the development of supercapacitors that can deliver superior power densities than batteries, but with higher energy densities than traditional electric double layer capacitors. ^2,3 The applications for such devices include regenerative breaking in Electric Vehicles (EVs) or for fast charge, power hungry energy storage device applications. ⁴ In an effort to develop high to moderate energy and high power devices, a hybrid supercapacitor cell architecture consisting of a low potential transition metal oxide electrode and a high potential carbon based capacitor electrode, is often employed. ⁵ These hybrid cell designs can benefit from the use of high surface area metal oxide anodes that can exhibit pseudocapacitive charge storage; this enables higher energy densities at high power densities to be delivered due to fast Faradaic redox reactions at the electrode/electrolyte interface. ^6-9

In the design of next generation anodes for batteries and supercapacitors, it is desirable to i) improve the intrinsically poor electronic conductivity of the anode and, ii) offset any capacity fading as a result of volume changes associated with conversion or intercalation reactions during cycling; if these challenges can be solved at scale, then they could lead to even greater performance capabilities for hybrid supercapacitors. ¹⁰ Strategies applied to solve these issues include nanosizing of the active materials and or the use of composite electrodes in which the active phase performance may be somehow buffered by a non-active or less active phase. In the latter case, the two materials can display synergistic effects such as the creation of built-in electric fields or new charge storage sites or promotion of different active species, all of which could result in improved ionic and electrical conductivity and electrochemical performance. ¹¹ For a high energy anodes, charge is typically stored via intercalation, conversion or pseudocapacitive processes. ⁶ In a conversion type electrode for a charge storage device, a metal oxide active material typically reversibly reacts with lithium ions to give the metallic state and an oxidised lithium species. Molybdenum oxides as M0O3 and M0O2 have previously been reported as being able to deliver high capacities due to conversion reactions such as those proposed by W. Tang et al. ¹² shown below:

Mo0 ₂ H — Li ⁺ H~ e- <-> Li ₀₉₈ MoO ₂ Li ₀₉₈ MoO ₂ H — 3 Li ⁺ H ^~ 3e ^_ <-> Mo + 2 Li ₂ 0

However these reactions have large volume changes associated with them during cycling at low potentials which can cause structural damage to the electrode and lead to a loss in capacity over time. ¹³ In contrast to conversion electrodes (or classical intercalation electrodes), charge can also be stored via pseudocapacitive fast Faradaic processes that occur on the surface of the active material. ¹⁴ Vanadium oxides such as VO ₂ are of interest as inexpensive pseudocapacitive electrodes that can show multiple oxidation states but are not stable due to structural deformations during cycling. ^15-17 Finally, Niobium oxides, which can contain both Nb ⁴⁺ and Nb ⁵⁺ are of interest as they can also display high levels of pseudocapacitive charge storage. ^{18 19} Whilst the individual metal oxides of V, Nb and Mo offer both pros and cons for energy storage, a number of binary oxide combinations such as VO _x /MoO _y oxides have been shown to offer significantly improved performance, e.g. 3D hierarchically ordered VO _x /MoO _y ²⁰ oxide anodes made in gram scale amounts were able to deliver a stable capacity of 440 mAh g ^-1 over 400 cycles at 4 A g ^-1 . Another example of the benefits of mixing vanadium and molybdenum oxides was shown by S-Q. Wang et al. ¹¹ Nanorods of VO _x @Mo0 ₃ were synthesised via electrochemical deposition onto carbon electrodes and were found to give a very high capacitance of 1166 mF cm ^-2 at the high current density of 100 mA cm ^-1 , this was double the performance of the same electrodes with only M0O3 deposited.

Furthermore, supercapacitors and more recently hybrid capacitors have been reported as energy storage devices that could be utilised to great effect in electric vehicles and large scale grid energy systems. ^51,52 The use of pseudocapacitive materials has received a lot of focus as a way to improve the energy density of capacitors, however they still suffer from poor performance at high current densities due to low conductivity and poor lithium diffusion kinetics of the materials. ³³ Methods such as nanosizing particles, incorporating porosity and doping of transition metals into the structure have all been reported as avenues to improve the capacity retention at high C -rates. ^54-56 The synergistic interaction between different metal ions on the surface of materials has been demonstrated to improve the conductivity of a metal oxide, as well as increasing the number of charge storage sites, generation of electric fields and altering the reactivity of metal ions through the distribution of local electron density in the metal-oxygen bonds. ⁵⁷ Molybdenum oxides, such as M0O2, have been shown to have great potential as candidates for anode materials due to their high theoretical capacity of 838 mAh g ^_1 , high levels of pseudocapacitance, multi electron transfer reactions, low cost of materials and ease of fabrication. ⁵⁸ One of the major issues surrounding the widespread use of molybdenum oxides is the conversion reaction of molybdenum oxide and lithium to molybdenum metal and lithium oxide. ^59'60 This provides the majority of the charge storage but suffers from long term stability degradation from the volume changes that occur during cycling, as well as the retention of capacity at high current densities. Formation of a vanadium/molybdenum oxide composite material was demonstrated to have improved electrochemical performance due to favourable interactions between the metals. ⁵⁹

In the synthesis of nanomaterials Continuous Hydrothermal Flow Synthesis (CHFS) processes have emerged as being desirable as they can rapidly generate varying compositions, produce metastable or intimately mixed oxides and are scalable. ^21,22

Summary of the Invention

Herein, we report the V-Nb-Mo oxide phase diagram library synthesised via CHFS as a method of materials discovery to identify high capacity materials that could be used as anodes in lithium- ion batteries and hybrid ion capacitors and batteries. Figure la shows the schematic of the CHFS system that was used to synthesise all the nominal compositions depicted in Figure lb. It was decided that this study would focus on the central hexagon area of the ternary plot to investigate the interaction between the three metals oxides. Samples are labelled according to the solution nominal ratio of V to Nb to Mn e.g. VNM622 as atomic ratio 6:2:2 etc.

Herein, we also report a method to improve the electrochemical performance of mixed vanadium-niobium-molybdenum oxide particles. Through a continuous hydrothermal flow synthesis method, nanoparticles of V/Nb/Mo oxide (referred to as VNM herein) and nanoparticles of VNM embedded into a carbon matrix were fabricated and subjected to heat treatment to form graphitic composite materials for use as high performance anodes in lithium- ion batteries (LIBs), and hybrid supercapacitors (HSCs) (referred to as CVNM). The presence of the graphitic carbon increased the stability of the active material during cycling and improved the capacity retention at high C-rates.

In one aspect the present proposals relate to an electrode material comprising, or consisting of, a V-Nb-Mo oxide wherein V _x Nb _y Mo _z represents the x:y:z molar ratio of V:Nb:Mo and wherein x is in the range 0.1 to 0.5; y is in the range greater than 0 to 0.3; and z is in the range 0.5 to 0.9, and wherein x+y+z = 1. Preferably x is in the range 0.2 to 0.4; y is in the range greater than 0 to 0.2; and z is in the range 0.6 to 0.8. Importantly, the material comprises all three elements V, Nb and Mo. Preferred V-Nb-Mo oxides are selected from the following VNM compositions: V _0.3 Nb _0.1 Mo _0.6 , V _0.3 Nb _0.2 Mo _0.5 , V _0.2 Nb _0.2 Mo _0.6 , V _0.2 Nb _0.1 Mo _0.7 , V _0.3 NbtraceMo _0.7 , V _0.4 NbtraceMo _0.6 , Vo.4Nbo.iMoo.5.

In one aspect the present proposals the electrode material comprises V-Nb-Mo oxide as defined herein and a carbon material. The carbon material is preferably present in an amount in the molar ratio range of carbon: V-Nb-Mo oxide of up to 3:1 (i.e. up to 3 parts carbon to 1 part V- Nb-Mo oxide), preferably between 0.5:1 and 2:1; preferably between 1:1 and 2:1; preferably about 2:1. In some cases the amount of carbon material present in the active material is a weight % of up to 5 wt%, preferably between about 0.5 and 5 wt%, preferably between about 0.5 and 2.5 wt.%, preferably between about 0.75 and 1.75 wt.%, preferably between about 1 and 1.75 wt.%, preferably between about 1 and 1.5 wt.%, preferably between about 1 and 1.25 wt.%, preferably about 1.12 wt%.

The present proposals also relate to partially or fully lithiated or sodiated, preferably lithiated, electrode materials as defined herein.

The electrode material as defined herein comprising or consisting of V-Nb-Mo oxide and optionally including a carbon material as defined herein, preferably has a surface area as defined by BET measurements (as described herein) greater than 10 rrfg ^-1 , preferably greater than 12 rrfg ^-1 , preferably greater than about 20 rrfg ^-1 , preferably greater than 50 rrfg ^-1 . This high surface area is thought to contribute to rapid charge/discharge characteristics of energy storage devices formed using these materials in an electrode (e.g. anode). Materials including the carbon material may demonstrate higher specific surface areas than those without the carbon material, in some cases up to 80 nfg ^-1 or more.

A half-cell formed using an electrode comprising the electrode material as defined herein preferably demonstrates a capacity of greater than about 350 mAh g ^-1 , preferably greater than 400 mAh g ^_1 , preferably greater than 500 mAh g ^-1 , preferably greater than 600 mAh g ^_1 , preferably greater than 700 mAh g ^_1 , at a specific current of less than 2 Ag ^-1 , preferably less than 1 Ag ^-1 , preferably less than 0.5 Ag ^-1 , preferably less than 0.2 Ag ^-1 , preferably about 0.1 Ag ^-1 .

The carbon material may be any organic material, and is preferably a carbohydrate or pure carbon (e.g. carbon at greater than 95wt.% purity). For example, the carbon material may be selected from sugars, e.g. sucrose. The carbon material may be initially present as, e.g. a carbohydrate, which is subsequently heated under vacuum or in an inert atmosphere to give pure carbon (e.g. carbon at greater than 95wt.% purity). In some cases this carbon may be selected from graphite, amorphous carbon, graphene or other carbon forms, and is preferably graphite. The presence of the carbon material, in particular graphitic carbon, increased the stability of the active material during cycling and improved the capacity retention at high charge-rates.

In some aspects the V-Nb-Mo oxide as defined herein is present in particulate form, having a coating of carbon material as defined herein and/or embedded in a matrix of carbon material as defined herein.

The present proposals also relate to an electrode comprising the electrode material as defined herein. The electrode may be an anode or cathode but is preferably an anode.

The present proposals also relate to an electrochemical cell, half cell, battery (e.g. lithium-ion, or sodium-ion; preferably lithium-ion, battery), capacitor, or hybrid capacitor comprising either an electrode (preferably an anode) as defined herein, or comprising an electrode material as defined herein.

The present proposals also relate to a hybrid capacitor or supercapacitor having an anode comprising or consisting of an electrode material as defined herein; optionally also having a cathode comprising or consisting of a carbon material, preferably activated carbon. The electrode materials as defined herein may provide beneficial high specific capacitance properties.

The present proposals also relate to the use of an electrode material as defined herein in the formation of an electrochemical cell, half cell, battery (e.g. lithium-ion, or sodium-ion; preferably lithium-ion, battery), capacitor, or hybrid capacitor.

The invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or expressly avoided.

Summary of the Figures

Embodiments and experiments illustrating the principles of the invention will now be discussed with reference to the accompanying figures in which:

Figure 1. a) schematic showing the set-up of the CHFS reactor used for the synthesis of the mixed metal oxide materials, b) Ternary plot used to calculate the nominal compositions of the mixed materials, compositions synthesised are shown as dark dots, c) Ternary plot identifying region of primary interest, d) Expanded region of interest with compositions A-G identified: A V _0.3 Nb _0.2 Mo _0.5 , B V _0.2 Nb _0.2 Mo _0.6 , C V _0.2 Nb _0.1 Mo _0.7 , D V _0.3 NbtraceMo _0.7 , E V04NbtraceMo _0.6 , F V _0.4 Nb _0.1 Mo _0.5 , G V _0.3 Nb _0.1 Mo _0.6 .

Figure 2. XRD patterns of monoclinic VO ₂ ref. 25 (black) VNM172 (green), VNM613 (orange) and VNM217 (blue) in the range 2 theta range of 5 to 40 degrees.

Figure 3. a) Ternary heat plots showing specific capacity vs. nominal composition at 0.1 A g ^_1 , for half cells vs. Li metal. The phase information for each material is also shown, amorphous by a circle with a line through it, materials with more than one phase present by a plain circle and single-phase materials as a bold circle. b) Ternary heat plots showing specific capacity vs. nominal composition at 5 A g ^-1 , for half cells vs. Li metal. The phase information for each material is also shown, amorphous by a circle with a line through it, materials with more than one phase present by a plain circle and single-phase materials as a bold circle.

Figure 4. XRD patterns of VNM316, VNM406, monoclinic VO ₂ (ref 26) and monoclinic M0O2 (ref. 25) in the 2 theta range 5 to 40 degrees.

Figure 5. The fitted XPS spectra of VNM406 for the a) V2p peaks and b) Mo3d peaks, c) the fitted XPS spectra of V2p of VNM316 and d) the fitted XPS spectra of Mo3d peaks of VNM3 16. e) Layered EDS map of VNM316 with f) individual EDS maps of the constituent metals and oxygen of VNM316.

Figure 6. CV curves of a half cell of VNM316 showing a) the first three cycles at a scanrate of lmV.s ^-1 and b) CV curves of a half cell of sample VNM316 at increasing scan rates.

Figure 7. 1 ^st cycle CV curves of half cells of both VNM406 (blue dash) and

VNM3 16 (green solid) at a) 1 mVs ^-1 and b) 20 mVs ^-1 . Power Law charge separation calculations showing the pseudocap acitive contribution as a shaded curve within the 1 mVs ^'1 CV curve of c) VNM3 16 and d) VNM406.

Figure 8. Charge discharge curves of a half-cell of VNM316 showing a) the 2nd, 6th and 10th cycles at 0.1 A g ^-1 b) charge discharge curves at increasing specific currents from 0.1 to 10 A g ^'1 . c) the rate performance of VNM316 (blue) and VNM406 (red) at increasing specific currents, d) long term cycling performance and coulombic efficiency of VNM316 at 1 A g ^-1 .

Figure 9. a) CV of the VNM316//AC full cell at a scanrate of 5 mV s ^_1 . b) GCD curves of the VNM316//AC full cell at 0.1 (black), 0.5 (red) and 1 A g ^-1 (blue).

Figure 10. Ragone plot showing the gravimetric energy and power of the hybrid supercapacitor cell of VNM316//AC compared to other reported devices utilising anode materials such as anatase TiO ₂ (ref 23), M0O ₂ and MoO ₂ /TiO ₂ (46).

Figure 11. Ternary plots of composition showing the specific capacity as a heat map for the specific currents: a) 0.2 A g ^-1 ; b) 0.5 A g ^-1 ; c) 1 A g ^-1 ; d) 2 A g ^-1 ; e) 5 A g ^-1 .

Figure 12. Graph showing the deconvoluted diffusion limited and pseudocapacitive contributions to the total charge stored during a CV test of a selection of the materials synthesised. Figure 13. XRD patterns of the VNM (red) and CVNM-1 (grey) materials pre-heat treatment showing them to be a match to the monoclinic Mo02 reference pattern in orange.

Figure 14. a) TEM image of CVNM-6 particles b) HRTEM of CVNM-6 with lattice spacings highlighted c) HRTEM image of CVNM-6 particles with the carbon layer thickness highlighted.

Figure 15. TEM image of CVNM-6 particles and EDS scans of vanadium (green), niobium (orange), molybdenum (blue), oxygen (yellow) and carbon (red).

Figure 16. Graph of a) BJH pore diameter of the CVNM materials, b) Raman scattering of pre-HT CVNM-6 (red) and post-HT CVNM-6 (black) with the D and G bands highlighted. Figure 17. XPS high resolution Mo3d scans of different CVNM materials with the doublets for Mo(VI) (blue), Mo(V) (yellow) and Mo(IV) (red) highlighted.

Figure 18. GCD graphs of a) CVNM-1 (green) b) CVNM-8 (orange), c) rate performance of all CVNM materials at increasing specific currents, d) rate performance graph of CVNM-6 at increasing specific currents and e) long term cycling capacity retention of CVNM-6 at 5 A g ^_1 . Figure 19. a) CV curve of CVNM-6 at 1 mV s ^_1 with the redox peaks labelled, b) graph showing the 1 ^st , 10 ^th and 50 ^th CV curves from long term cycling ay 1 mV s ^_1 c) CV scans of CVNM-6 at increasing scan rates from 0.1 to 10 mV s ^_1 , d) CV curve of CVNM-6 at 1 mV s ^_1 (black) with the capacitive contribution superimposed (red shaded area) and e) capacitive (shaded) and diffusion-limited (clear) charge storage contributions for CVNM-6 (red) and VNM316 (blue).

Figure 20. XRD patterns of VNM pre- (grey) an post-heat treatment (red) exhibiting a phase change to a mixed valence crystal phase

Figure 21. PXRD patterns of all CVNM materials synthesised with the monoclinic M0O2 reference pattern shown (orange). Figure 22. BET isotherm of CVNM-6

Figure 23. XPS high resolution V2p scan of CVNM, with the peaks for the V(5+) (blue) oxidation state highlighted.

Figure 24. XPS high resolution Mo3d scan of VNM.

Figure 25. Graph showing the theoretical capacity of a hybrid device with different AGCVNM ratios. Figure 26. graphs of AC//CVNM-6 showing a) GCD curves at 0.5 (black) and 1 (red) A g ^-1 , b) GCD curves at 2 (blue) and 5 (green) A g ^-1 , c) nyquist plot and d) frequency vs complex active (red) and reactive (black) power with the time relaxation constant quoted next to the resonance frequency.

Figure 27. Ragone plot of specific energy and power of reported hybrids of composite carbon-metal oxide materials. Data compiled from sources [89-91]

Detailed Description of the Invention

Aspects and embodiments of the present invention will now be discussed with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference.

In the search for high capacity pseudocapacitive active anode materials for hybrid supercapacitors (with activated carbon as the counter electrode), a 48-sample nanoparticle V-Nb-Mo oxide phase diagram library was made via continuous hydrothermal synthesis and each sample investigated electrochemically in a device. When viewed across the compositional space, both the electrochemical performance and physical characteristics were found to vary with nominal composition. The sample with nominal composition Vo _. 3Nbo _.i Moo.6 was found to be predominantly monoclinic crystal phase and delivered a specific capacity of 747 mAh g ^-1 at 0.1 A g ^-1 and 180 mAh g ^-1 at 5 A g ^-1 . XPS studies suggested metal -oxygen-metal interactions all three metals as well as an increased ratio of V ⁴⁺ to V ⁵⁺ compared to the niobium free analogue composition. When used in a hybrid supercapacitor with activated carbon, the device displayed high energy density at low power (101.8 Wh kg ^-1 at 365 W kg ^-1 ) and moderate energy density at very high power (10.3 Wh kg ^-1 at 18.3 kW kg ^-1 as well as long-term cycling stability (93 % capacity retention after 450 cycles at 1 A g ^-1 ).

The use of a continuous hydrothermal flow synthesis method was investigated as a way of creating carbon coated mixed vanadium, niobium and molybdenum oxides. The carbon/metal oxide materials were subjected to a heat treatment under inert atmosphere to graphitise the carbon. It was found by PXRD that the carbon additive protected the monoclinic M0O2 crystal phase from the heat treatment. TEM imaging identified small particles embedded into a carbon matrix with varying levels of coating depending on the carbon/metal ratio. Raman spectroscopy confirmed the conversion of carbon precursor to graphitic carbon, which was attributed to the large increase in BET surface area measured. Electrochemical analysis indicated well-defined redox peaks as well as a pseudocapacitive charge storage. CVNM-6 was found to have a pseudocapacitive contribution of 72, 92 and 97 % at 0.1, 1 and 20 mV s ^_1 . The material CVNM-6 was found to have the best cycling performance, exhibiting specific capacities of 505, 546, 584, 567, 506, 366, and 197 mAh g ^_1 at specific currents of 0.1, 0.2, 0.5 1, 2, 5 and 10 A g ^_1 . Further to this it was observed that the CVNM materials all displayed an increase in capacity over cycling, this was attributed to activation of the electrode through increasing access of electrolyte into the porous structure. When constructed into hybrid capacitors with activated carbon as the cathode, the Li-ion device was found to have a specific energy of 81 Wh kg ^-1 at a specific power of 1270 W kg ^-1 and at a high specific power of 8500 W kg ^-1 the specific energy was found to be 11 Wh kg ^-1 . The electrochemical performance of the carbon-vanadium, niobium, molybdenum oxide materials identifies them as useful as anodes in both lithium ion batteries and hybrid capacitors and their application to larger energy storage devices.

* * *

The features disclosed in the foregoing description, or in the following claims, or in the accompanying drawings, expressed in their specific forms or in terms of a means for performing the disclosed function, or a method or process for obtaining the disclosed results, as appropriate, may, separately, or in any combination of such features, be utilised for realising the invention in diverse forms thereof.

While the invention has been described in conjunction with the exemplary embodiments described above, many equivalent modifications and variations will be apparent to those skilled in the art when given this disclosure. Accordingly, the exemplary embodiments of the invention set forth above are considered to be illustrative and not limiting. Various changes to the described embodiments may be made without departing from the spirit and scope of the invention.

For the avoidance of any doubt, any theoretical explanations provided herein are provided for the purposes of improving the understanding of a reader. The inventors do not wish to be bound by any of these theoretical explanations.

Any section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described.

Throughout this specification, including the claims which follow, unless the context requires otherwise, the word “comprise” and “include”, and variations such as “comprises”,

“comprising”, and “including” will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps. It must be noted that, as used in the specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and/or to “about” another particular value. When such a range is expressed, another embodiment includes from the one particular value and/or to the other particular value. Similarly, when values are expressed as approximations, by the use of the antecedent “about,” it will be understood that the particular value forms another embodiment. The term “about” in relation to a numerical value is optional and means for example +/- 10%.

Examples

Chemicals and Consumables

Precursors for Materials Synthesis: Ammonium metavanadate (NH ₄ VO ₃ ), ammonium heptamolybdate tetrahydrate [(NH ₄ ) ₆ Mo ₇ O ₂₄ ● ₄ H ₂ O], ammonium niobate oxalate hydrate (C ₄ H ₄ NNbO ₉ ^● 4.8H ₂ O), oxalic acid, ascorbic acid and N-methyl-2-pyrrolidone (NMP) were purchased from Sigma Aldrich, Dorset, UK.

Chemicals or Consumables for Analysis and Testing: for electrochemical testing, CR2025 coin cells were used with the active material electrodes as the working electrode and lithium metal foil discs (PI-KEM, Staffordshire, UK) as the counter electrode. Glass microfiber separators (Whatman, GF/D, Buckinghamshire, UK) were drenched in 1M L1PF6 in 1:1 v/v mixture of ethylene carbonate (EC) and ethyl methyl carbonate (EMC) as the electrolyte (Sigma Aldrich, Dorset, UK). Hybrid cells were assembled using activated carbon (YP50F, Kuraray Chemical Co., Japan) printed electrodes (supplied by Warwick Manufacturing Group at Warwick University, UK) as the cathode. Super P™ conductive carbon powder was purchased from Alfa Aesar, Heysham, UK, polyvinylidene fluoride (PVDF) (PI-KEM, Staffordshire, UK). TEM studies were conducted using 300-mesh copper film grids (Agar Scientific, Stansted, UK).

Continuous Hydrothermal Flow Synthesis (CHFS) of Nanopowders

Heterometallic V-Nb-Mo oxide nanomaterials were synthesised using a CHFS lab scale process. Ammonium metavanadate and oxalic acid were dissolved in an aqueous solution in the ratio 1:2 and stirred overnight until the colour changed from orange to dark blue. Ammonium heptamolybdate tetrahydrate and ascorbic acid were pre-mixed in the ratio 4: 1 to obtain a solution of MO ⁴⁺ . The vanadium and molybdenum precursor solutions and an aqueous solution of ammonium niobate oxalate hydrate were mixed together in a beaker with the nominal ratios determined by the ternary plot. The total metal concentration in each precursor mixture was 0.1M. Synthesis involved three diaphragm pumps (Primeroyal K, Milton Roy, Pont-Saint-Pierre, France) which were used to deliver feeds to a Confined Jet Mixer (CJM, patent GB1008721) as follows; Pump 1 (P1) was used to deliver a stream of hot deionized water (was preheated in line from room temperature to 450 °C). Pump 2 (P2) and pump 3 (P3) were used to deliver a metal salt solution to the CMJ, where it was mixed with the superheated water feed from PI, resulting in the formation of nanoparticles. Pump 4 (P4) was used to deliver ambient temperature deionised water (quench) feed into a second CJM that was located 43 cm above the initial CJM. The nanoparticle slurry from this second CMJ was then cooled using a pipe-in-pipe heat exchanger before passing through a back-pressure regulator (BPR) valve. The slurries were collected and then cleaned by dialysis in ion permeable membrane bags (Medicell Membranes Ltd, London, UK) in deionised water until the conductivity of the supernatant was <100 μS as determined by a conductivity probe (Hanna Instruments, model H198311). Powders were obtained by centrifuging the slurries to a wet sludge and then freezing this to -40°C under a vacuum of <10 Pascals, this was then heated to 40°C over 24h using a Freeze drier (Virtis Genesis 35XL, Biopharma process systems, Winchester, UK) All materials were collected as black powders.

Synthesis of Carbon-Vanadium/Niobium/Molybdenum oxides

The CVNM nanomaterials were synthesised using a continuous hydrothermal flow method (CHFS) with a duel mixer set up and a quench. Typically, ammonium metavanadate (NH ₄ VO ₃ ) was dissolved in H ₂ O with Oxalic acid in the molar ratio 1 :2 and stirred overnight until the colour had changed from orange to blue. This was mixed with pre-dissolved aqueous solutions of ammonium niobate oxalate hydrate (C ₄ H ₄ NNbO ₉ ●4.8H ₂ O), ammonium heptamolybdate tetrahydrate [(NH ₄ ) ₆ Mo ₇ O ₂₄ ● ₄ H ₂ O] and sucrose (C12H22O11) to give a total volume of 1 L in which the molar ratio was 3 : 1 :6 of metals to a range of sucrose concentrations with a total metal molarity of 0.3 M (all chemicals were from Sigma Aldrich, UK). After the solution has passed through the CHFS reactor, it was collected as a slurry of nanoparticles that were cleaned by dialysis in deionised water and freeze-dried to give a brown powder. The materials were then heat treated in a nitrogen atmosphere at 750°C for 3 hours to give the final black powders.

Physical Characterisation of Materials

The collected materials were initially characterised by Powder X-ray Diffraction (PXRD) on a STOE StadiP diffractometer., using Mo-Kα radiation, λ = 0.7107 Å, in the 2θ range from 2 to 40 ^° with a step size of 0.5 ^° and a step time of 10 seconds. The materials were measured using transmission through foils. Identification of the crystal phases present in the XRD data was performed using Match! Software (Crystal Impact.com, Version 1.1 lh). X-ray photoelectron spectroscopy (XPS) was performed to ascertain the valance states of the metal ions present. Samples were prepared by affixing each sample to a carbon film and collecting data in automatic mode. A Thermo Scientific K-alpha™ spectrometer using A1-Kα radiation was used to perform high resolution regional scans for vanadium, molybdenum and oxygen, these were conducted at 50 eV. The data was processed by CasaXPS™ software (version 2.3.16) and calibrated using the carbon C Is peak at a binding energy of 284.8 eV. Metal content analysis was performed by X-ray fluorescence (XRF), using an Episilon 1 Benchtop Spectrometer.

BET (Brunauer-Emmett-Teller) surface area measurements were performed using a Micrometries Tristar II with liquid Nitrogen. All samples were degassed for 12 hours under N2 gas flow at 120 °C before measurements were taken.

Transmission Electron Microscopy (TEM) was performed on the as-synthesised nanomaterials on a JEOL JEM 2100 TEM using a LaB ₆ filament. The TEM images were analysed to determine the size and morphology of the particles. Image capture was performed by a Gatan Orius digital camera, measurements of particles and d-spacing was performed on Gatan Microscopy Suite software. Samples were dispersed in methanol and ultrasonicated, then pipetted onto 300-mesh copper film grid.

EDS was performed using the JEOL JEM 2100 that was also used for TEM imaging.

Raman spectroscopy was performed using a Renishaw inVia™ Microscope with 514 nm wavelength laser. Scans were collected in the range 3200 to 100 cm ^-1 .

Preparation of Electrodes, Half Cells and Hybrid Capacitor Coin Cells.

For the manufacture of anodes; each of the metal oxide powders were mixed with SuperP™ conductive carbon powder and PVDF polymer binder in the form of a pre-dissolved 10 wt.% in NMP solution, in the mass ratio 80:10:10, respectively. This slurry was ball milled at 1000 rpm for lh (Pulverisette 7, Fritsch, Germany) and spread onto 9 pm thick copper foil (PI-KEM, Staffordshire, UK), then dried on a hotplate at ca. 130 °C for 10 minutes and then cut into 15 mm diameter disc. The electrodes were dried under vacuum at 90 °C overnight and then transferred to argon filled glovebox for cell assembly (O ₂ and H ₂ O kept below 0.1 and 3 ppm respectively).

Cell assembly; half cells were assembled with the active material electrodes as the working electrode and Li metal foil as the counter electrode in a coin cell. Glass microfiber separators were drenched in 1M LiPF ₆ in 1:1 v/v mixture of EC/EMC as the electrolyte. Hybrid capacitor cells were made using the same procedure as described above, but using printed activated carbon electrodes as the counter electrode, the preparation of which is described elsewhere. ²³ Electrochemical Measurements

Galvanostatic cycling was performed on half cells in the potential range 0.05 to 3.00 V vs Li/Li ⁺ at a range of specific currents from 0.1 to 10.0 A g ^_1 on an Arbin Instrument (Model BT-2000 battery tester). Cyclic voltammetry (CV) testing was performed in the same potential window, at increasing scan rates in the range 0.05 to 100 mV.s ^_1 .

During electrochemical testing, Galvanostatic Charging and Discharging (GCD) and CV testing was performed on hybrid capacitor cells in the voltage window 0.50 to 4.15 V. The specific capacitance was calculated from the GCD curve according the equation below:

Where / is the current, AV/At is the change in potential over time and m is the combined mass of active material on both electrodes. The energy density (E) and power density (P) with units W h kg ^-1 and W kg ^-1 respectively for hybrid capacitor cells were determined by:

Where C is the capacitance, m is the combined active mass of both electrodes, V is the operating potential window and R is the equivalent series resistance, which was calculated from the IR drop in the galvanostatic discharge curve of the hybrid device.

Electrochemical Characterisation of CVNM Materials

Experiments were performed on CR2032 type coin cells at room temperature. The working electrodes were prepared by mixing the CVNM materials with Super P carbon black and polyvinylidene fluoride binder in the ratio 80:10:10. The coin cells were assembled in an argon filled glovebox, and consisted of the working electrode, a glass microfiber separator (Whatman), and the electrolyte (3:7 v:v ethyene carbonate (EC)/dimethyl carbonate (DMC)). Galvanostatic charge/discharge tests were performed by an Arbin Instrument (BT-2000) at different specific currents in the voltage range 0.05-3 V vs Li/Li ⁺ . Cyclic voltammetry and electrochemical impedance spectroscopy (EIS) were performed on a Gamry Interface 1000 potentiostat over the same voltage range and over a frequency range of 100 kHz to 10 mHz respectively.

Physical Characterisation of Nominal Compositions

Figure 2 shows the XRD patterns for three representative compositions across the range synthesised. All the materials with high proportions of vanadium and niobium showed very broad peaks, as highlighted by VNM172 and VNM613 (figure 2). This is suggestive of the particles exhibiting small and amorphous nature. The broad nature of the high niobium materials has been previously attributed to the poorly crystalline nature of as-prepared NbiOs made via CHFS ²⁴ . The high molybdenum content materials were found to have more defined patterns, the identified diffraction peaks at -12.7, -16.7 and -24.0 ° were assigned the monoclinic VO ₂ (M0O2 style) crystal structure (ICDS no. 34033 with the space group Phi _/ ci) ²⁵ . The BET surface areas were found to vary across the compositional space, the materials VNM163, VNM145, VNM262, VNM226, VNM451, VNM433, VNM415 and VNM631 were found to be ca. 218, 230, 182, 39, 97, 5, 3 and 15 m ² g ^_1 , respectively.

Electrochemical analysis of nominal compositions

The specific capacities of the as-synthesised materials were determined by Galvanostatic C rate tests of half cells at a range of specific currents in the potential window 0.05 to 3.00 V vs Li/Li ⁺ and plotting the data as ternary “heat maps” to show the relationship between specific capacity and nominal composition (Figure 3). The phase information for each sample is also shown in figures 3a and 3b, with amorphous samples shown as a circle with a line through it, samples that were not single phases as a plain circle, with single phase compositions being represented as a bold circle. The bright red spot in figure 3a shows a VNM316 nominal composition that has the highest capacity, delivering an impressive 662 mAh g ^_1 (at 0.1 A g ^_1 ). At 5 A g ^_1 the VNM316 material had a capacity of 120 mAh g ^_1 , again far outperforming any other VNM oxide material in the library (as shown in figure 3b) and identifying it as a composition for further characterisation. The electrochemical performance at other specific currents are shown in the Figures 11a, lb, 11c, l id, l ie).

When comparing the electrochemical results in half cells to both the nominal composition and the phase information of each material, it can be observed that the materials that showed the highest capacities were also identified as being single phase and being located on the periphery of compositional space that was explored, particularly in the corner of the ternary plot with low nominal amounts of niobium present, that was also found to be fairly crystalline. ²⁴

Physical Characterisation of VNM316

Figure 4 shows the XRD plot for the as-synthesised materials VNM316 and VNM406. Both materials show peaks at 2Q = 12.7, 16.9 and 24.7 °, which can be associated with both monoclinic M0O2 (ICSD no. 80830, space group P1 ₂₁ / ₁ ) ²⁶ and monoclinic VO ₂ (ICSD no. 34033). The peak shift seen for both materials is placed halfway between the reference peaks of Mo and VO ₂ suggesting that these materials have a crystal structure that is somewhere between the two. Both reference materials have the same space group and almost identical PXRD patterns except for two small peaks at 2θ = 19 and 25.5 ^° seen in VO ₂ . These two small peaks (indicated by a star) are seen in the PXRD patterns for VNM316 and VNM406 ²⁵ ’ ²⁷ . The presence of Niobium caused significant peak broadening for the VNM316 with respect to sample VNM406 (suggesting phase separation of a Nb ₂ O ₅ -like poorly crystalline phase) as well as an observable peak shift in the (011) peak to a lower 2Q value by 0.33 degrees, possibly suggesting the doping of Nb into the monoclinic VO ₂ like crystal structure ²⁸ .

Whole pattern fitting using a Le Bail method was performed for both VNM406 and VNM316. Both patterns were fitted to the space group P1 _21/c1. The reported lattice parameters for the reference patterns and the two materials fitted are shown in table 1. The addition of the niobium was found to increase the size of parameters a, b, angle b and the cell volume slightly supporting the idea that the niobium had inserted into the crystal structure rather than being isolated within the material.

Table 1 : Table showing the reported lattice parameters for monoclinic VO ₂ (ref 26) and M0O2

(ref 25) compared to the calculated lattice parameters for VNM406 and VNM316.

The oxidation states of the metals present were characterised by XPS. The survey spectrum confirmed the presence of V, Mo, Nb and O. As seen in Figure 5a and 5c, the high-resolution V 2p of both VNM406 and VNM316 are deconvoluted by fitting peaks for both V ³⁺ and V ⁴⁺ oxidation states ²⁹ . In the V2p _3/2 peak for VNM406 the peaks at 517.3 and 516.4 eV are assigned to V ⁵⁺ and V ⁴⁺ respectively in the ratio 14: 1, for the material VNM316 the prominence of V ⁴⁺ was increased by a factor of ten with the ratio of the fitted V2p _3/2 peaks found to be 1.4:1. The XPS spectra also suggests the presence of Nb-O-V interactions as the V2p _3/2 peaks were shifted by 0.1 eV to a lower binding energy. This could be explained by the marginally lower electronegativity of Vanadium compared to Niobium pulling electron density away from Nb and towards V. ¹¹ The Mo3d spectra for both materials are shown in Figure 5b and 5d; for VNM406 the Mo3ds _/2 and Mo3d _3/2 peaks at 232.4 and 235.6 eV were a good match for Mo (VI) ³⁰ . The location of the Mo3d peaks for VNM316 was found to shift to higher binding energies of 232.7 and 235.8 eV respectively. This shift also suggests the occurrence of Nb-O-Mo interactions, which due to the lower electronegativity of Nb, pulls electron density away from the Molybdenum ions. The XPS spectra of Niobium in sample VNM316, confirmed the presence of Nb ⁵⁺ in the material.

The homogeneity of the metals in the VNM406 and VNM316 samples was further investigated by EDS/TEM. Figure 5e and 5f shows a layered EDS/TEM image of a particle of sample VNM316 with the scans for V, Mo, Nb and O shown in red, green, orange and blue respectively. As can be seen, the particle has a homogeneous mixture of metals present, further supporting the suggestion of V-O-Mo-O-Nb interactions. The ratio of V:Nb:Mo was estimated by EDS analysis on the surface to be 4:1:5 which is not as would be expected for the bulk composition, however this is not surprising given it is a surface method only. Bulk elemental analysis was performed by XRF analysis and found to be in good agreement with the nominal composition. The BET surface area of VNM406 and VNM316 was determined to be ca. 12 and 23 m ² g ^-1 respectively; the increased surface area of particles of the latter was presumed to play a beneficial role in rapid charge/discharge cycling.

Electrochemical Characteristics of VNM316 and VNM406

The electrochemical performance of a half cell of sample VNM316 was analysed by cyclic voltammetry (CV) and galvanostatic charge discharge measurements (GCD). Figure 6a shows the CVs of the first 3 cycles at a scan rate of 1 mV s ^-1 . The material shows a pair of reversible redox peaks at 1.6/1.5 and 1.1/1.0 V vs Li/Li ⁺ , both of which have been reported to be assigned to the reversible phase transition of monoclinic M0O2 to orthorhombic partially lithiated Li _x MoCh and then to monoclinic Li _0.98 MoO ₂ . As the potential of the cell drops there is a large peak centred at 0.25 V vs Li/Li ⁺ , which is characteristic of the conversion reaction of Li _0.98 MoO ₂ to Mo metal, a process that can store large amounts of energy usually at slow scan rates but is associated with large volume changes and reduced cycling stability. There is also a very broad peak visible on the forward scan centred at 2.3 V vs Li/Li ⁺ ; due to the broadness of the CV on the backwards scan it is not possible to say with any certainty if the process is reversible and is just exceptionally broad or if it is a non-reversible process ^31-34 . Overall the very broad nature of the CV curves as well as the propensity to remain unchanged over the three cycles indicates a high nature of pseudocapacitive charge storage occurring as well as good cycling stability, this was mirrored in the CVs at higher scanrates which retained a broad overall shape (Fig 6b) ³⁵ .

Figure 7a shows the CV curves of VNM406 (blue dashed) compared to VNM316 (green solid) at a scanrate of 1 mV s ^-1 , the overall smaller size and sharper nature of the redox peaks is immediately apparent for the former. This difference in curve shape is mirrored at the faster scanrate 20 mV. s ^{' 1} (Figure 7b). The contribution of pseudocapacitance to the charge storage was calculated by the relationship between the power laws for capacitive and diffusion limited currents ³⁶ . By plotting a Cartesian equation for a range of scan rates of equation 4 for both the lithiation and the delithiation scans, the slope and y intercept gives values for the parameters ai and a2, which are then used to calculate the relative contribution from each mechanism of charge storage for a current response at a range of potentials on the CV curve. The results are shown for the half-cell of sample VNM316 (Fig. 7c) and sample VNM406 (Fig. 7d) as a purely pseudocap acitive CV curve (shaded area). The half-cell of VNM316 was found to have a higher pseudocap acitive contribution to the overall capacity compared to VNM406 at 1 mV s ^-1 , 82.7 to 74.6 % respectively. The pseudocapacitive contributions of a range of different compositions is shown in Figure 12. Figure 8a shows GCD curves for the VNM316 half-cell on the 2 ^nd (red), 6 ^th (blue) and 10 ^th (green) cycles at a specific current of 0.1 A g ^-1 ; as can be seen the smooth shape of the curves does not change much as the cell is cycled. The lack of observable plateaus in the GCD curves supports the CV results that pseudocapacitive charge storage is dominant ³⁷ . On the second cycle for the VNM316 half-cell, the specific capacity was 731.3 mAh g ^-1 (C.E. = 90 %) and this was found to increase to 786.9 mAh g ^-1 (C.E. = 93 %) by the 10 ^th cycle. The gradual increase in specific capacity and C.E. was observed to continue as the material was cycled at higher specific currents (Fig. 8b), e.g. at 0.5 A g ^-1 the specific capacity remained high at 761.3 mAh g ^-1 and the C.E. increased to 96 %. This activation behaviour has been previously reported for other transition metal oxides that exhibit porosity and was attributed to an increase in access of the electrolyte into the electrode therefore, allowing more active material to be utilised and faster lithium diffusion kinetics through the material ^38-40 . Figure 8c shows the rate performance of sample VNM316 compared to VNM406 at increasing specific currents from 0.1 to 10 A g ^-1 .

For the half-cell of VNM316 as the specific current increased from 0.1, 0.2, 0.5, 1, 2, 5, to 10 A g ^-1 , the capacities decreased in the order 747, 761, 658, 542, 338, 185 to 83 mAh g ^-1 , respectively. Once the specific current was returned to 0.1 A g ^-1 , the specific capacity returned to a slightly higher level than the start (up to 829 mAh g ^-1 ), highlighting the excellent rate capability and stability. In comparison, the rate performance of the VNM406 half-cell showed lower capacities at all specific currents and no gradual activation behaviour as seen for the VNM316 cell. The cycling stability of VNM316 is further displayed in figure 8d, where over 450 cycles at 1 A g ^-1 , with steady activation behaviour observed in the first 200 cycles, thereafter for another 250 cycles there was a slow decrease of the specific capacity. Overall after 450 cycles the material retained ca. 94 % of original capacity whilst maintaining an excellent C.E. The reversible capacity of the VNM316 half-cell was thus significantly higher than that of commercially available graphite (theoretical capacity of 372 mAh g ^-1 ) under similar conditions. ⁴³ Electrochemical Performance of the Hybrid Supercapacitor

A hybrid supercapacitor was assembled using VNM316 as the anode and activated carbon as the cathode. To achieve the best performance of the cell the respective specific capacities of the anode and cathode were mass balanced, the ideal ratio of anode to cathode was calculated to be 1 : 16 from the method set out by H-Y. Wei et al. ⁴⁴ Fig. 9a and Fig. 9b show the CV curves at a range of scanrates and the GCD curves at increasing specific currents of the hybrid cell. The overall quasi- rectangular shape of the CVs shows typical capacitive behaviour, the symmetric shape of the GCD curves further highlights this ⁴⁵ . The hybrid cell was found to give specific capacitances of 29.8, and 13.7 F g ^-1 at specific currents of 0.5 and 1 A g ^-1 . The VNM316:AC hybrid cell exhibited a high energy density of 127 Wh kg ^-1 at a power density of 365 W kg ^-1 , and 10.3 Wh kg ^-1 at the high-power density 18.25 kW kg ^-1 , as calculated using the method described in section 2.5. These results compare very favourably to other reported high power anode half-cells and supercapacitors that contain Molybdenum oxide or Titania based active materials synthesised via similar hydrothermal flow methods ^23,46 and are shown in the Ragone plot in Figure 10.

Discussion - VNM

The major part of a V-Mo-Nb oxide ternary phase diagram was rapidly synthesised via CHFS with most of the as-synthesised materials exhibiting a monoclinic VO ₂ or monoclinic MoCE-like crystal structure phase except for those with high proportions of niobium oxide, which were found to be less crystalline (with high specific surface areas up to 230 m ² g ^-1 ). The as-synthesised active materials were made into anodes for half-cells and subjected to electrochemical testing to map performance trends. Sample VNM316 (nominal composition Vo _. 3Nbo _.i Moo _, 6) was found to have the best overall performance at high and low power; its cell displayed a high specific capacity of 747 mAh g ^-1 at low specific current of 0.1 A g ^-1 and 181 mAh g ^-1 at the high specific current of 5 A g ^-1 and showed good long-term cycling with a ca. 94 % capacity retention over 450 cycles. Analysis of cyclic voltammetry suggested that pseudocapacitive charge storage made up the major contribution (83%) at a scan rate of 1 mV s ^-1 . The significantly increased capacity (and pseudocapacitive contribution) may be attributed to a balance between several factors such as multiple active redox centres were present, interactions between Nb-O-V and Nb-O-Mo, which were suggested by both XPS and EDS. This sample also had the highest V ⁴⁺ content on the surface compared to the material VNM406, a relatively large surface area of 23 m ² g ^-1 (favouring pseudocapacitive charge storage) and small particle size. Consequently, the hybrid supercapacitor cell made using sample VNM316 achieved an energy density of 101.8 Wh kg ^-1 at low power (365 W kg ^-1 ), and an energy density of 10.3 Wh kg ^-1 at very high power (18250 W kg ^-1 ). Remarkably, the mixed vanadium, molybdenum and niobium oxide materials as defined herein also display good long-term cycling performance, which may be related to the small particle size and/or excellent homogeneous mixing of the three constituent metals within the samples made via CHFS.

Structure and Morphology of the Carbon-mixed metal oxide nanoparticles

The CVNM and VNM materials collected immediately after synthesis were brown and black powders respectively. Both materials, pre and post heat treatment were investigated by XRD, TEM, EDS, BET, BJH, Raman, XPS and XRF. Prior to the heat treatment in argon atmosphere, both the VNM metal oxide and the CVNM composite materials were found by XRD to be a good match for monoclinic M0O2 ICSD no. 80830, with a slight peak shift to lower angles due to the presence of V and Nb in the structure (figure 13). ^64,65 Post heat treatment at 750°C, the VNM metal oxide underwent a phase change to a mixed vanadium/molybdenum oxide phase with multiple oxidation states present (figure 20). ⁶⁶ Post heat treatment, the colour of the CVNM materials was a black powder. The presence of the carbon in the CVNM composite was found to protect all materials synthesised and retain the monoclinic M0O2 crystal phase (figure 21). XRF analysis of the CVNM materials identified good reliability between the concentration of aqueous metal-salt precursors pre-CHFS, the initial CHFS product and the post-heat treatment carbon-metal oxide products.

The morphology of the resultant VNM and CVNM materials was examined by transmission electron microscopy (TEM) and energy dispersive x-ray spectroscopy (EDS). After the introduction of the carbon precursor and post-CHFS heat treatment, TEM of the CVNM materials showed metal oxide particles embedded in carbon matrix (figure 14a). Lattice fringe orientations in the high resolution TEM image (figure 14b) show clear lattice spacings of d = 0.34 nm were measured, these correspond to the (1 -1 -1) plane of monoclinic M0O ₂ . ⁶⁴ It is further observed that the individual particles are coated in a thick layer of carbon with an average thickness of 5.8 nm (figure 14c). The small particle size measured suggests the presence of the carbon layer inhibits particle growth through agglomeration during the graphitisation reaction. ⁶⁷ Elemental mapping of the CVNM materials through EDS identified all three metals, oxygen and carbon present in the particles. Figure 15 shows the intimate and homogenous mixing of the three metal oxides with each other and with the carbon coatings, this matched the elemental mapping of VNM particles and is typical of mixed materials synthesised by CHFS. ⁶⁸

To further characterise the carbon coated particles, BET and BJH analysis were performed, figure 22 shows the isotherm of CVNM-1. The specific surface area was found to increase for CVNM-6 due to the heat treatment process from 26 to 85 m ² g ^_1 . This can be explained by the transformation of large amorphous carbonaceous particles converting to distinct carbon coatings on nanoparticles, thus increasing the surface area. This trend was consistent across all the carbon to metal oxide ratios investigated, however there was no significant difference in surface area as the carbon content was increased. BJH pore size analysis identified a narrow distribution of pore sizes, with the average diameter being 3.7 nm for all materials, this is indicative that the heat treatment process is central to the formation of the pores (figure 16a). ⁶⁹ A good interaction between the active material and the electrolyte can be achieved because of the large BET surface area and average pore size, thus improving the electrochemical performance. ⁷⁰ ’ ⁷¹ Raman spectroscopy was performed to investigate the nature of the amorphous carbon coating. Figure 16b shows the spectra for CVNM-6 collected prior to heat treatment, in which the broad feature at 1400 cm ^-1 and the broad peak at 1600 cm ^-1 were assigned to the D and G bands which correspond respectively to disorder along the c-axis of graphitic carbon materials and of the C-C stretch along the plane of carbon sheets. ⁷⁰ Post heat treatment, the sharpness of the peaks increases suggesting the formation of more graphitic carbon within the material, the intensity of the G band is greater than the D band suggesting a higher proportion of sp ² hybridised carbon present in the material. A high ratio of G:D band intensity has been shown to correspond to improved electrochemical performance of the composite materials. ^{72, 73}

XPS was used to investigate the oxidation state of the metals present and the electronic interactions. The survey scan of CVNM revealed peaks for vanadium, niobium, molybdenum, oxygen and carbon. The high resolution scan for V2p showed two peaks at 523.7 and 516.4 eV with a spin-orbit splitting of D = 7.33 eV, these were assigned to the V2p _3/2 and V2pi _/2 peaks respectively (Figure 23). These were matched to the V ⁵⁺ oxidation state. ⁷⁴ The high resolution scans of Mo3d showed three peaks that were fitted with three sets of doublets each with a spin- orbit splitting of D = 3.15 eV (figure 17). The Mo3d _5/2 peaks of the doublets were located at 229.8, 231.4 and 232.6 eV, these were assigned to the Mo ⁴⁺ , Mo ⁵⁺ and Mo ⁶⁺ oxidation states respectively. ⁷⁵ Analysis of the peak area gave a composition of oxidations states of 33%, 20% and 47% respectively. The high resolution scan for Nb3d gave a doublet of peaks at 209.9 and 207.2 with a spin-orbit splitting of D = 2.58 eV, these were assigned to the Nb3d _3/2 and Nb3d _5/2 peaks respectively for the Nb ⁵⁺ oxidation state. ⁷⁶ In comparison the high resolution scan of Mo3d for VNM identified only two peaks (figure 24). These were a good fit for a doublet with a spin-orbit coupling of D = 3.15 eV, the peaks were located at 232.7 eV and 235.8 eV and were assigned to the Mo3d _5/2 and Mo3d _3/2 peaks respectively for the oxidation state Mo ⁶⁺ . This suggests that the carbon coating of CVNM is providing partial protection from oxidation. ⁷⁷ The binding energies of vanadium and molybdenum both suggest the presence of Nb-O-V and Nb-O-Mo interactions as the binding energies of the V2p and Mo3d scans were shifted to lower and higher values respectively compared to reference literature. This could be explained by the differences in electronegativity between the three metals allowing for electron density to be pulled away from molybdenum ions towards niobium, and away from niobium towards vanadium ions. ⁵⁷ This supports the EDS analysis of intimately mixing between the metal oxides within the same particles. The quantification of the elemental ratios was performed by XRF. It was revealed that the atomic percentage of vanadium, niobium and molybdenum in the CVNM materials were a good match to the nominal concentrations of pre-cursors prior to the CHFS process, CVNM-6 was found to have the composition V:Nb:Mo 28%, 12% and 60% respectively.

CHN analysis was performed to quantify the carbon content of the composite materials, it was found that CVNM-6 had a carbon content of 1.12 %.

In summary the material properties measured have confirmed sp ² hybridised carbon coated nanoparticles of mixed V, Nb and Mo in the molecular structure of monoclinic MoCh, matching the crystal structure of the best performing VNM oxide previously reported.

Electrochemical Lithium Ion Storage Performance

The insertion/extraction performance of CVNM and VNM anodes were analysed in half cell configurations against lithium metal counter electrodes. Figure 18a and 18b show the initial GCD curves for CVNM-1 and CVNM-8 respectively. The initial discharge and charge capacities for CVNM-1 were 871 and 450 mAh g ^_1 , which was marginally greater than the values for the higher carbon content sample CVNM-8 of 743 and 397 mAh g ^_1 . The first cycle coulombic efficiency of CVNM-1 and 8 were found to be 52 and 53 % respectively. All the CVNM materials exhibited similar initial charge and discharge capacities and first cycle efficiencies with little variation as the carbon content increased. All of the CVNM materials tested were found to display the same smooth shape GCD curves with two small plateaus at 1.35 and 1.70 V vs Li/Li ⁺ and a larger plateau below 0.25 V vs Li/Li ⁺ . The large capacity loss on the first cycle is attributed to the decomposition of electrolyte involved in the formation of the solid electrolyte interface layers (SEI) on the surface of the active material. ⁷⁸ The rate performance of the CVNM materials was tested over increasing specific currents from 0.1 to 10 A g ^_1 (figure 18c). Overall the performance of the CVNM materials increased with carbon content and then decreased, indicating an optimum carbon/metal oxide molar ratio of 2:1 and an optimum carbon percentage of 1.12 %. At the lower currents of 0.2, 0.5 and 1 A g ^_1 , CVNM-4 performed the best with reversible capacities of 638, 666 and 646 mAh g ^_1 . However, at the high currents of 5 and 10 A g ^_1 CVNM-6 was found to outperform the other materials with reversible capacities of 366 and 197 mAh g ^_1 respectively (figure 18d). All CVNM materials were noted to display an increase in capacity over cycling at low currents, CVNM-6 increased from 522 to 604 mAh g ^_1 over 10 cycles at 0.2 A g ^_1 . This is attributed to an activation of the conversion reaction of molybdenum oxide and lithium to molybdenum metal and lithium oxide, as well as an increase in the number of active sites available for lithium storage. ^79'81 when the capacity is returned to 0.1 A g ^_1 the reversible capacity is far higher at 803 mAh g ^_1 . The long term cycling stability at the high specific current of 10 A g ^_1 was investigated for CVNM-6. After 200 cycles it retained an impressive 74 % capacity, after 500 cycles this had slowly decreased and plateaued at 66 %. The rate performance results of CVNM-6 compare very well to other carbon/metal oxide materials reported in literature (table 1). The suitability of CVNM-6 as an anode material for lithium ion batteries is further supported when taking the scalability of the synthesis method into account.

Analysis of Electrochemical Dynamics

To elucidate the mechanisms and kinetics of the electrochemical energy storage mechanisms occurring in the CVNM materials that enable the excellent rate performance, the materials were investigated by cyclic voltammetry. Figure 19a shows the CV curve of CVNM-6 at a scanrate 1 mV s ^_1 . Two pairs of well-defined cathodic/anodic peaks located at ~ 1.2/1.45 and 1.5/1.8 V vs. Li/Li ⁺ were assigned to the reversible phase transition of partially lithiated molybdenum oxide (Li _x MoO ₂ ) from monoclinic-to-orthorhombic-to-monoclinic. ⁸² The large peak below 0.5 V vs. Li/Li ⁺ was attributed to the high capacity conversion reaction of molybdenum oxide and lithium to Mo° and LEO. ⁸³ Figure 19b shows the 1 ^st , 10 ^th and 50 ^th cycles of CVNM-6 at 1 mV s ^_1 , as can be seen the shape of the curve on the first cycle (black) has a smaller overall area and larger peak at 1.17 V vs Li/Li ⁺ on the backward scan that could be attributed to reactions involving the formation of the SEI. The shape of the redox peaks on the forward scan between cycle 1 and 10 do not change. The shape of the CVs between cycle 10 and cycle 50 remains very uniform with the shape of the CV changing very little and the redox peaks only decreasing in size slightly over cycling. This highlights the stability of the CVNM material during the high volume fluctuation conversion reactions occurring. ^{65, 84, 85} Figure 19c shows the CV curves of CVNM-6 at increasing scan rates from 0.1 to 100 mV s ^_1 . The shape of the CV remains consistent as the scan rate is increased, with a slight increase in peak polarisation for the redox peaks observed.

The area under the CV curve has been shown to be representative of the total charge stored from both diffusion-limited processes and capacitive processes. ⁸⁶ The overall large area of the CV curves for CVNM-6 are indicative of pseudocapacitive charge storage. ⁶⁵ The relationship between measured current [i(V)\ and scan rate (v) at a certain potential is described by:

This allows for the determination of the proportion of charge stored by capacitive and diffusion- limited process. Figure 19d shows the CV curve at 1 mV s ^_1 with the purely capacitive curve superimposed (red shaded area). Figure 19e shows the capacitive contribution of CVNM-6 (red) compared to the non-carbon coated VNM316 (blue) at increasing scan rates. Using a method outlined by Trasatti et al. ⁸⁷ the proportion of capacitive charge storage at increasing scan rates is also quantified. Both materials show increasing contributions from capacitive charge storage methods, however CVNM-6 shows greater capacitive contributions over VNM316 at every scan rate above 0.1 mV s ^_1 . Hybrid Ion Capacitor Performance

The electrochemical performance of CVNM-6 indicates it could perform well as an anode material in both LIBs and HICs. To investigate this, a coin cell hybrid lithium-ion capacitor was constructed with activated carbon as the cathode. In order to achieve the best performance, the two electrodes were mass balanced with the optimal cathode to anode ratio determined by figure 25 as 2.5:1. ⁸⁸ The AC//CVNM-6 hybrid cell was investigated by galvanostatic charge and discharge cycling, with figure 26a and 26b showing the results for 0.5 and 1 A g ^'1 , and 2 and 5 A g ^'1 respectively. The shape of the GCD curves display triangular shaped capacitor-like behaviour with only a small IR drop observed on the discharge for 0.5 and 1 A g ^'1 . As the current is increased to 2 and 5 A g ^'1 the size of the IR drop increases, and the discharge time decreases. To investigate the conductivity of the AC//CVNM-6 device, EIS was performed and the analysis of complex active and reactive power was applied. Figure 26c shows the nyquist plot of the AC//CVNM-6 device. A value for the charge transfer resistance was calculated using an equivalent circuit to be 12.56 ohms and the value for the ESR (resistivity of the device) was calculated to be 10 ohms. Figure 26d shows a plot of complex active and reactive power vs. frequency. The resonance frequency was determined from the crossover of the two plots and was used to calculate a value of t ₀ = 0.29 s for the time relaxation constant. The GCD data was converted into specific power and energy values and plotted on the Ragone plot (figure 27) against other reported carbon/metal oxide composites. ^91'93 The AC//CVNM coin cell HIC exhibited a specific energy of 81 Wh kg ^-1 at a power of 1270 W kg ^-1 . At a high power of 8500 W kg ^-1 the energy was 11 Wh kg ^-1 .

Discussion - CVNM

In Summary, the use of a continuous hydrothermal flow synthesis method was investigated as a way of creating carbon coated mixed vanadium, niobium and molybdenum oxides. As noted above, the synthesis is also effective form non-carbon coated materials. The carbon/metal oxide materials were subjected to a heat treatment under inert atmosphere to graphitise the carbon. It was found by PXRD that the carbon additive protected the monoclinic M0O2 crystal phase from the heat treatment. TEM imaging identified small particles embedded into a carbon matrix with varying levels of coating depending on the carbon/metal ratio. Raman spectroscopy confirmed the conversion of carbon precursor to graphitic carbon, which was attributed to the large increase in BET surface area measured. Electrochemical analysis indicated well-defined redox peaks as well as a pseudocapacitive charge storage. CVNM-6 was found to have a pseudocapacitive contribution of 72, 92 and 97 % at 0.1, 1 and 20 mV s ^_1 . The material CVNM-6 was found to have the best cycling performance, exhibiting specific capacities of 505, 546, 584, 567, 506, 366, and 197 mAh g ^-1 at specific currents of 0.1, 0.2, 0.5 1, 2, 5 and 10 A g ^-1 . Further to this it was observed that the CVNM materials all displayed an increase in capacity over cycling, this may be attributed to activation of the electrode through increasing access of electrolyte into the porous structure. When constructed into hybrid capacitors with activated carbon as the cathode, the Li-ion device was found to have a specific energy of 81 Wh kg ^-1 at a specific power of 1270 W kg ^-1 and at a high specific power of 8500 W kg ^-1 the specific energy was found to be 11 Wh kg ^-1 . The electrochemical performance of the carbon-vanadium, niobium, molybdenum oxide materials identifies them as candidates for use as anodes in both lithium ion batteries, sodium-ion batteries and hybrid capacitors and their application to larger energy storage devices.

References

A number of publications are cited above in order to more fully describe and disclose the invention and the state of the art to which the invention pertains. Full citations for these references are provided below. The entirety of each of these references is incorporated herein.

1 J. B. Goodenough and Y. Kim, Chem. Mater., 2010, 22, 587-603.

2 P. Simon and Y. Gogotsi, Nat. Mater., 2008, 7, 845-854.

3 M. Armand and J. M. Tarascon, Nature, 2008, 451, 652-657.

4 R. Kotz and M. Carlen, Electrochim. Acta, 2000, 45, 2483-2498.

5 W. S. V. Lee, E. Peng, M. Li, X. Huang and J. M. Xue, Nano Energy, 2016, 27, 202-212.

6 V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597-1614.

7 M. Liibke, J. Shin, P. Marchand, D. Brett, P. Shearing, Z. Liu and J. A. Darr, J. Mater. Chem. A, 2015, 3, 22908-22914.

8 M. Liibke, P. Marchand, D. J. L. Brett, P. Shearing, R. Gruar, Z. Liu and J. A. Darr, J. Power Sources, 2016, 305, 115-121.

9 M. Liibke, N. Ding, M. J. Powell, D. J. L. Brett, P. R. Shearing, Z. Liu and J. A. Darr, Electrochem. commun., 2016, 64, 56-60.

10 C. C. Subrahmanyam Goriparti, Ermanno Miele, Francesco De Angelis, Enzo Di Fabrizio, Remo ProiettiZaccaria, J. Power Sources, 2014, 257, 421,443.

11 S.-Q. Wang, X. Cai, Y. Song, X. Sun and X.-X. Liu, Adv. Funct. Mater., 2018, 28, 1803901.

12 W. Tang, C. X. Peng, C. T. Nai, J. Su, Y. P. Liu, M. V. V. Reddy, M. Lin and K. P. Loh, Small, 2015, 11, 2446-2453.

13 H. S. Kim, J. B. Cook, H. Lin, J. S. Ko, S. H. Tolbert, V. Ozolins and B. Dunn, Nat. Mater., 2017, 16, 454-462.

14 V. Augustyn and B. Dunn, Energy Env., 2014, 7, 1597-1614.

15 N. A. Chernova, M. Roppolo, A. C. Dillon and M. S. Whittingham, J. Mater. Chem.,

2009, 19, 2526.

16 W. Chen, J. Peng, L. Mai, H. Yu and Y. Qi, Chem. Lett., 2004, 33, 1366.

17 P. Poizot, S. Laruelle, S. Grugeo and L. Dupont, Ionics (Kiel)., 2000, 6, 321-330.

18 E. Lim, H. Kim, C. Jo, J. Chun, K. Ku, S. Kim, H. I. Lee, I. S. Nam, S. Yoon, K. Kang and J. Lee, ACS Nano, 2014, 8, 8968-8978.

19 M. Lubke, A. Sumboja, I. Johnson, D. Brett, P. Shearing, Z. Liu and J. A. Darr, Electrochim. Acta, 2016, 192, 363-369.

20 D. Zhao, J. Qin, L. Zheng and M. Cao, Chem. Mater., 2016, 28, 4180-4190.

21 J. A. Darr, J. Zhang, N. M. Makwana and X. Weng, Chem. Rev., 2017, 117, 11125—

11238.

22 X. Weng, J. K. Cockcroft, G. Hyett, M. Vickers, P. Boldrin, C. C. Tang, S. P. Thompson, J. E. Parker, J. C. Knowles, I. Rehman, I. Parkin, J. R. G. Evans and J. A. Darr, J. Comb. Chem. , 2009, 11, 829-834.

23 D. Bauer, A. J. Roberts, N. Matsumi and J. A. Darr, Nanotechnology , 2017, 28, 195403- 195413.

24 M. Liibke, A. Sumboja, I. D. Johnson, D. J. L. Brett, P. R. Shearing, Z. Liu and J. A. Darr, Electrochim. Acta , 2016, 192, 363-369.

25 J. M. Longo and P. Kierkegaard, Acta Chem. Scand. , 1970, 24, 420-426.

26 A. A. Bolzan, B. J. Kennedy and C. J. Howard, Aust. J. Chem., 1995, 48, 1473-1477.

27 M. Ghedir, H. Vincent, M. Marezio and J. C. Launay, J. Solid State Chem., 1977, 22, 423-438.

28 J. Papavasiliou, G. Avgouropoulos and T. Ioannides, Appl. Catal. B Environ., 2007, 69, 226-234.

29 J. Kasperkiewicz, J. A. Kovacich and D. Lichtman, J. Electron Spectrosc. Relat. Phenom., 1983, 32, 128.

30 N. S. McIntyre, D. D. Johnston, L. L. Coatsworth, R. D. Davidson and J. R. Brown, Surf. Interface Anal., 1990, 15, 265.

31 D. Wu, R. Shen, R. Yang, W. Ji, M. Jiang, W. Ding and L. Peng, Sci. Rep., 2017, 7, 1-9.

32 D. Bauer, A. J. Roberts, C. L. Starkey, R. Vedarajan, D. Brett, P. Shearing, N. Matsumi and J. A. Darr, Int. J. Electrochem. Sci., 2018, 13, 5120-5140.

33 V. Pralong, V. Gopal, V. Caignaert, V. Duffort and B. Raveau, Chem. Mater., 2012, 24, 12-14.

34 A. Bhaskar, M. Deepa and R. Narasinga, ACS Appl. Mater. Interfaces, 2013, 5, 255-2566.

35 A. Gonzalez, E. Goikolea, J. A. Barrena and R. Mysyk, Renew. Sustain. Energy Rev., 2016, 58, 1189-1206.

36 J. Wang, J. Polleux, J. Lim and B. Dunn, J. Rhys. Chem. C, 2007, 111, 14925-14931.

37 A. . Dylla, G. Henkelman and K. . Stevenson, Acc. Chem. Res., 2013, 46, 1104-1112.

38 L. Guo and Y. Wang, J. Mater. Chem. A, 2015, 3, 4706-4715.

39 X.-Y. Xue, Z.-H. Chen, L.-L. Xing, S. Yuan and Y.-J. Chen, Chem. Commun., 2011, 47, 5205.

40 Y. Luo, J. Luo, J. Jiang, W. Zhou, H. Yang, X. Qi, H. Zhang, H. J. Fan, D. Y. W. Yu, C. M. Li and T. Yu, Energy Environ. Sci., 2012, 5, 6559.

43 L. Wan, J. Shen, Y. Zhang and X. Li, J. Alloys Compd. , 2017, 708, 713-721.

44 H. Y. Wei, D. S. Tsai and C. L. Hsieh, RSCAdv., 2015, 5, 69176-69183.

45 R. B. Rakhi, W. Chen, D. Cha and H. N. Alshareef, Nano Lett., 2012, 12, 2559-2567.

46 H. Kim, Adv. Energy Mater., 2013, 3, 1500-1506.

51 J. Ding, W. Hu, E. Paek and D. Mitlin, Chem. Rev., 2018, 118, 6457-6498.

52 P. Simon and Y. Gogotsi, Nat. Mater., 2008, 7, 845-854.

53 V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597-1614.

54 H. Park, H. Bin Wu, T. Song, X. W. Lou and U. Paik, Adv. Energy Mater., ,

DOI: 10.1002/aenm.201401945.

55 F. Zoller, K. Peters, P. M. Zehetmaier, P. Zeller, M. Doblinger, T. Bein, Z. Sofer and D. Fattakhova-Rohlfmg, Adv. Funct. Mater., 2018, 28, 1-10.

56 M. Liibke, A. Sumboja, F D. Johnson, D. J. L. Brett, P. R. Shearing, Z. Liu and J. A. Darr, Electrochim. Acta, 2016, 192, 363-369.

57 S.-Q. Wang, X. Cai, Y. Song, X. Sun and X.-X. Liu, Adv. Funct. Mater., 2018, 28, 1803901.

58 W. Tang, C. X. Peng, C. T. Nai, J. Su, Y. P. Liu, M. V. V. Reddy, M. Lin and K. P. Loh, Small, 2015, 11, 2446-2453.

59 D. Bauer, T. E. Ashton, D. J. L. Brett, P. R. Shearing, N. Matsumi and J. A. Darr, Electrochim. Acta, 2019, 322, 134695.

60 D. Wu, R. Shen, R. Yang, W. Ji, M. Jiang, W. Ding and L. Peng, Sci. Rep., 2017, 7, 1-9.

61 E. Lim, H. Kim, C. Jo, J. Chun, K. Ku, S. Kim, H. I. Lee, F S. Nam, S. Yoon, K. Kang and J. Lee, ACS Nano, 2014, 8, 8968-8978.

62 C. Zhang, X. Peng, Z. Guo, C. Cai, Z. Chen, D. Wexler, S. Li and H. Liu, Carbon N. Y., 2012, 50, 1897-1903.

63 Z. Chen, V. Augustyn, J. Wen, Y. Zhang, M. Shen, B. Dunn and Y. Lu, Adv. Mater.,

2011, 23, 791-795.

64 A. A. Bolzan, B. J. Kennedy and C. J. Howard, Aust. J. Chem., 1995, 48, 1473-1477.

65 Z. Hao, Q. Chen, W. Dai, Y. Ren, Y. Zhou, J. Yang, S. Xie, Y. Shen, J. Wu, W. Chen and G. Q. Xu, Adv. Energy Mater., 2020, 10, 1-8.

66 L. . Plyasova, L. . Soloveva, G. . Kryukova, V. . Zabolotnyi and I. . Olenkova, Zhurnal Strkturnoi Khimii, 1991, 32, 110-115.

67 G. Zhu, Q. Li, Y. Zhao and R. Che, ACS Appl. Mater. Interfaces, 2017, 9, 41258-41264.

68 J. A. Darr, J. Zhang, N. M. Makwana and X. Weng, Chem. Rev., 2017, 117, 11125—

11238.

69 M. Zhi, C. Xiang, J. Li, M. Li and N. Wu, Nanoscale, 2013, 5, 72-88.

70 L. Shen, X. Zhang, E. Uchaker, C. Yuan and G. Cao, Adv. Energy Mater . , 2012, 2, 691- 698.

71 X. Li, M. Qu and Z. Yu, Solid State Ionics, 2010, 181, 635-639.

72 D. Bresser, E. Paillard, R. Kloepsch, S. Krueger, M. Fiedler, R. Schmitz, D. Baither, M. Winter and S. Passerini, Adv. Energy Mater., 2013, 3, 513-523.

73 M. M. Doeff, J. D. Wilcox, R. Yu, A. Aumentado, M. Marcinek and R. Kostecki, J. Solid State Electrochem., 2008, 12, 995-1001.

74 G. Silversmit, D. Depla, H. Poelman, G. B. Marin and R. De Gryse, J. Electron Spectros. Relat. Phenomena, 2004, 135, 167-175.

75 J. Baltrusaitis, B. Mendoza- Sanchez, V. Fernandez, R. Veenstra, N. Dukstiene, A. Roberts and N. Fairley, Appl. Surf. Sci., 2015, 326, 151-161.

76 N. Ozer, M. D. Rubin and C. M. Lampert, Sol. Energy Mater. Sol. Cells, 1996, 40, 256.

77 M. Bystrzejewski, J. Solid State Chem., 2011, 184, 1492-1498.

78 L. Liu, P. Guan and C. Liu, J. Electrochem. Soc., 2017, 164, A3163-A3173.

79 Y. Luo, J. Luo, J. Jiang, W. Zhou, H. Yang, X. Qi, H. Zhang, H. J. Fan, D. Y. W. Yu, C.

M. Li and T. Yu, Energy Environ. Sci., 2012, 5, 6559.

80 L. Guo and Y. Wang, J. Mater. Chem. A, 2015, 3, 4706-4715.

81 W. Ai, J. Jiang, J. Zhu, Z. Fan, Y. Wang, H. Zhang, W. Huang and T. Yu, Adv. Energy Mater., 2015, 5, 1-8.

82 W. Tang, C. X. Peng, C. T. Nai, J. Su, Y. P. Liu, M. V. V. Reddy, M. Lin and K. P. Loh, Small, 2015, 11, 2446-2453.

83 K. Cao, T. Jin, L. Yang and L. Jiao, Mater. Chem. Front., 2017, 1, 2213-2242.

84 J. Ding, W. Hu, E. Paek and D. Mitlin, Chem. Rev., 2018, 118, 6457-6498.

85 L. Yu, H. Hu, H. Bin Wu and X. W. D. Lou, Adv. Mater., ,

DOI: 10.1002/adma.201604563.

86 J. Wang, J. Polleux, J. Lim and B. Dunn, J. Phys. Chem. C, 2007, 111, 14925-14931.

87 S. Trasatti and G. Buzzanca, J. Electroanal. Chem. Interfacial Electrochem, 1971, 29, 1- 5.

88 B. Li, J. Zheng, H. Zhang, L. Jin, D. Yang, H. Lv, C. Shen, A. Shellikeri, Y. Zheng, R. Gong, J. P. Zheng and C. Zhang, Adv. Mater., 2018, 30, 1-19.

91 S. Shen, W. Guo, D. Xie, Y. Wang, S. Deng, Y. Zhong, X. Wang, X. Xia and J. Tu, J. Mater. Chem. A, 2018, 6, 20195-20204.

92 X. Wang and G. Shen, Nano Energy, 2015, 15, 104-115.

93 D. Bauer, A. J. Roberts, N. Matsumi and J. A. Darr, Nanotechnology, 2017, 28, 195403- 195413.