Low‐Temperature Atomic Layer Deposition Synthesis of Vanadium Sulfide (Ultra)Thin Films for Nanotubular Supercapacitors

Herein, the synthesis of vanadium sulfide (VxSy) by atomic layer deposition (ALD) based on the use of tetrakis(dimethylamino) vanadium (IV) and hydrogen sulfide is presented for the first time. The (ultra)thin films VxSy are synthesized in a wide range of temperatures (100–225 °C) and extensively characterized by different methods. The chemical composition of the VxSy (ultra)thin films reveals different vanadium oxidation states and sulfur‐based species. Extensive X‐ray photoelectron spectroscopy analysis studies the effect of different ALD parameters on the VxSy chemical composition. Encouraged by the rich chemistry properties of vanadium‐based compounds and based on the variable valences of vanadium, the electrochemical properties of ALD VxSy (ultra)thin films as electrode material for supercapacitors are further explored. Thereby, nanotubular composites are fabricated by coating TiO2 nanotube layers (TNTs) with different numbers of VxSy ALD cycles at low temperature (100 °C). Long‐term cycling tests reveal a gradual decline of electrochemical performance due to the progressive VxSy thin films dissolution under the experimental conditions. Nevertheless, VxSy‐coated TNTs exhibit significantly superior capacitance properties as compared to the blank counterparts. The enhanced capacitance properties exhibited are derived from the presence of chemically stable and electrochemically active S‐based species on the TNTs surface.


Introduction
Metal sulfides (MSs) are semiconducting materials composed of sulfur anions associated with one or more metal cation(s) that depending on the element composition are usually classified into binary (CdS), ternary (CuFeS 2 ), and quaternary MSs (Cu 2 ZnSnS 4 ).[3][4] These properties have attracted in the last years substantial attention from a wide range of applications bringing in MSs into the nanotechnology revolution.Thus, MSs nanomaterials are a prospective alternative in multidisciplinary applications, demonstrating a promising performance in energy conversion (e.g., CO 2 reduction, water splitting, or N 2 reduction) [3,5] and energy storage devices (various metal ion batteries, supercapacitors [SCs]), [1,2,6] sensors, [7] photodetectors, [8] and biomedical applications, [9] among others.
A wide variety of methods have been reported for the synthesis of MSs nanomaterials.Among them, chemical and physical methods are the main approaches.Chemical methods mostly include hydrothermal, [10] solvothermal, [11] electrodeposition, [12] chemical bath deposition, [13] and microwave-assisted processes, [14] while physical methods span from laser ablation technique [15] to physical vapor deposition [16] along with other vacuum deposition methods, such as chemical vapor deposition [17] and atomic layer deposition (ALD). [18,19]ALD in particular is a well-established thin film deposition technique based on gas phase chemical precursors exposed separately to the surface of the substrate followed by a purging step.Accordingly, the precursors react alternatively with the active sites of the surface by self-limiting gas-surface (precursor-substrate) reactions resulting in a layer-by-layer growth, while during the purging step the excess of precursor and reaction by products are removed from the chamber.
The conformality and the nanometer scale thickness control are the result of the self-limited chemical reactions and the consequent layer-by-layer growth, respectively. [20]Nowadays, the advances in nanofabrication techniques allow the rational design and development of complex nanostructured materials with high surface-to-volume ratio that can act as active supporting materials in hybrid composites.The functionalization with secondary nanomaterials for the fabrication of synergetic hybrid nanostructures is an effective approach to fully exploit the high active surface area benefits and optimize the nanomaterial's performance.It is here where ALD emerges as the unique deposition technique enabling conformal and homogenous deposition of secondary materials on nanostructures with complex shape while most of the aforementioned deposition methods are utterly limited. [21,22]n the last years, the ALD features have attracted enormous interest and stimulated an intense research work devoted to the design and synthesis of new precursors, which has greatly extended the portfolio of materials deposited by ALD.In the case of MSs, there is an excellent review where the synthesis of a large number of MSs by ALD has been summarized. [19]Among the wide range of MSs, vanadium-based materials exhibit outstanding electrical conductivity and capacitance together with an excellent redox reversibility and electrochemical properties. [23]anadium is an abundant element that exhibits multiple oxidation states (V 2þ , V 3þ , V 4þ , V 5þ ) which provides a rich and versatile chemical reactivity.Consequently, vanadium sulfides include a wide number of phases such as VS, VS 2 , VS 4 , VS 6 , V 2 S 3 , V 3 S, and V 2 S 5 .Some of them, mostly VS 2 and VS 4 , have been explored as promising materials in energy storage applications such as electrode material for SCs and different metal ion batteries. [24]SCs are electrochemical energy storage devices featured by offering high energy density and power density, fast charge/discharge rate, and long lifetime. [25]Regarding the main limitation of the SCs, it is (compared to batteries) the lower energy density.Different strategies have been developed to mitigate such limitation by the fabrication of electrode materials including oxides, sulfides, carbides, or nitrides, among others, through rationally designed nanostructures. [26,27]o the best of our knowledge, there is only one work reporting the growth of vanadium sulfide by ALD, where the authors extensively characterized VS 4 thin films deposited using tris(N,N 0 -diisopropylacetamidinate) vanadium (III) and H 2 S as vanadium and sulfur source, respectively. [28]This work is the first work that reports the ALD synthesis of vanadium sulfide (V x S y ) based on the use of tetrakis(dimethylamino) vanadium (IV) (TDMAV), which has been used for the synthesis of vanadium oxide, [29] yet not combined so far with hydrogen sulfide for the synthesis of vanadium sulfide.Herein, the novel ALD synthesis of V x S y using TDMAV and H 2 S in a wide range of temperatures (from 100 to 225 °C) is presented for the first time.The synthesis conducted at temperatures as low as 100 °C is highly interesting because that is compatible with temperature-sensitive substrate materials, e.g., flexible polymeric materials.The resulting vanadium sulfide thin films were extensively characterized confirming the V x S y chemical composition, morphology, and structure, verifying thus that the chemical reactions between TDMAV and H 2 S follow ALD principles.Second, the capacitance properties of V x S y as electrode material for SCs application were further explored.Thus, the active supporting material of choice was TiO 2 nanotube layers (TNTs), which have exhibited outstanding potential as nanostructured electrodes for energy storage applications, including fuel cells and SCs. [30]Thereby, we fabricated nanotubular composites by coating such TNTs with different numbers of V x S y ALD cycles, namely, 25, 50, 100, and 200.Different electrochemical techniques were used to characterize the capacitance properties of those nanotubular composites as electrode material for SCs application.

Results and Discussion
The ALD growth of V x S y thin films was studied at different temperatures, namely, 100, 150, 190, and 225 °C, applying the following ALD conditions: TDMAV pulse (600 ms)-N 2 purge (15 s)-H 2 S pulse (1000 ms)-N 2 purge (15 s) as described in the Experimental Section and illustrated in Figure S1, Supporting Information, by a flowchart.Figure 1a shows the thickness of the V x S y films deposited on Si wafer.The results showed in Figure 1b revealed similar growth per cycle values, indicating an ALD window within the range of deposition temperatures explored.The thicknesses of the V x S y films deposited on Si wafer as a function of the deposition temperature were measured by cross-sectional scanning electron microscope (SEM) images (see Figure 1c).To confirm that the V x S y growth follows the ALD principle, the self-saturated nature of the surface reactions was evaluated by varying one precursor pulse duration keeping the rest of the parameters of the process fixed.Thereby, the effect of the pulse duration of TDMAV and H 2 S on the V x S y growth rate was studied varying the TDMAV pulse duration using a constant H 2 S (500 ms) pulse duration and varying H 2 S pulse duration along with a constant TDMAV dose (600 ms).The results, shown in Figure S2, Supporting Information, revealed that the reaction between H 2 S and TDMAV follows to the ALD principle for all the H 2 S pulse durations and for a TDMAV pulse duration ≥ 300 ms, exhibiting a growth rate value of ≈0.085 nm cycle À1 .
We further demonstrated the conformal and homogeneous deposition of V x S y thin films into TNTs.Figure S3, Supporting Information, shows top view SEM images of the as-deposited V x S y thin films on TNTs coated with 25, 50, 100, 200, and 400 ALD cycles at 100 °C using a dose of 300 ms for TDMAV and 500 ms for H 2 S. Top view SEM cross-sectional images at different locations (top, middle, and bottom) of the TNTs coated after 400 V x S y ALD cycles allowed to demonstrate the uniformity of the V x S y thin film thickness deposited all along the TNTs, as shown in Figure S4, Supporting Information.The ALD nature of the V x S y growth was further confirmed by the evaluation of the V x S y thickness deposited on TNTs as a function of the number of ALD cycles, namely, blank, 100, 200 and 400, by top view cross-sectional SEM images, as shown in Figure S5a-d, Supporting Information.Figure S5e, Supporting Information, shows that the V x S y thickness is linearly dependent on the number of ALD cycles while the linear extrapolation to the origin indicated no nucleation delay during the V x S y growth.These TNTs coated at 100 °C with different numbers of V x S y ALD cycles were used for the assessment of the capacitance properties as described in detail ahead.
Figure S6, Supporting Information, shows the atomic force microscope (AFM) images obtained by scanning an area of 2 μm Â 2 μm on silicon wafers after 300 ALD cycles.The ALD V x S y was found to grow as a smooth thin film with root mean square roughness values (R q ) within 0.21 and 0.41 nm as shown in Table S1, Supporting Information.
The X-ray diffraction (XRD) results in Figure S7, Supporting Information, showed that even the V x S y thin film deposited at the highest temperature, i.e., 225 °C, exhibited an amorphous structure.A postannealing treatment of as-deposited V x S y thin films was carried out at 400 °C for 2 h under vacuum (≈2 mbar) in N 2 atmosphere as a strategy to 1) enhance the crystalline degree and 2) reduce the content of bulk O.While the postannealing treatment did not induce any increase in the degree of crystallinity of the V x S y thin film as shown in the corresponding XRD pattern in Figure S7, Supporting Information, it did cause to a diminution in the bulk O content as described by X-ray photoelectron spectroscopy (XPS) results ahead.
An extensive and detailed analysis by XPS allowed to study the surface chemical composition of the V x S y thin films deposited on TNTs as a function of different ALD parameters such as deposition temperature, dosing time of TDMAV and H 2 S, and the number of cycles.Figure S8, Supporting Information, shows the XPS survey spectra of the V x S y obtained after 200 ALD cycles at different deposition temperatures (100, 150, 190, and 225 °C) revealing the presence of V, S, C, O, and traces of N. The elemental composition of the V x S y deposited at different temperatures obtained from XPS analysis is shown in Table S2, Supporting Information.The samples were not exposed to the atmosphere and the precursors are O-free, so that the content of bulk O cannot be ascribed to surface oxidation but can be present due to residual H 2 O or O 2 in the ALD reactor during the ALD process.Nevertheless, XPS analysis revealed primarily a diminution in the bulk O content close to half of initial after the postannealing treatment (as shown in Table S3, Supporting Information), which, in turn, led to a chemical reduction of the vanadium illustrated by an increase of the content of V 2þ and V 3þ (see Figure S9, Supporting Information).Figure 2 presents the peak-fitted high-resolution V 2p, O 1s, and S 2p XPS spectra of the V x S y deposited at different temperatures, namely, 100, 150, 190, and 225 °C.XPS results described ahead indicated very similar chemical composition regardless of the deposition temperature.The fitting of the V 2p spectra revealed four spin-orbit doublets for V 2p 3/2 and V 2p 1/2 , suggesting the presence of different vanadium oxidation states, V 2þ (peaks located at ≈513.6 and ≈521.1 eV), [31][32][33] V 3þ (peaks located at ≈514.8 and ≈522.5 eV), [31,32] V 4þ (peaks located at ≈516.3 and ≈524.0 eV), [31,34] and V 5þ (peaks located at ≈517.3 and ≈525.1 eV). [33]The presence of V 3þ and V 2þ indicated the partial reduction of the V 4þ (the oxidation state of V in the precursor molecule TDMAV) to lower oxidation states during the ALD growth, giving V x S y species where vanadium adopted multiple oxidation states.Regarding the S 2p spectra, the corresponding fitting suggested three spin-orbit doublets for S 2p 3/2 and S 2p 1/2 , indicating the presence of different sulfur species.The peaks at ≈161.4 and ≈162.6 eV were subscribed to the presence of sulfide S 2À [32] while the content of disulfide S 2 2À (─S─S─) 2À was ascribed to the peaks at ≈162.1 and ≈163.3 eV. [35]he peaks at ≈163.8 and ≈164.9 eV were related to surface thiol group (S─H). [33]he evolution of the chemical composition of the V x S y thin film as a function of the number of ALD cycles was also analyzed by XPS using the postannealed V x S y TNTs obtained after applying 25, 50, 100, and 200 ALD cycles at a deposition temperature of 100 °C. Figure S10, Supporting Information, presents the respective peak-fitted high-resolution V 2p, O 1s, and S 2p XPS spectra.Interestingly, the TNTs coated with ultrathin V x S y layer after only 25 ALD cycles (≈2 nm calculated) enabled a valuable insight into the interface built up between TiO 2 and the V x S y at the very early stage of the V x S y growth process.The V 2p high-resolution XPS spectrum displayed two peaks corresponding to the spin-orbit splitting of V 2p 3/2 and V 2p 1/2 at binding energies of ≈517.0 and ≈524.7 eV, respectively.The fitted spectrum showed a mixture of vanadium oxidation states V 4þ and V 5þ .Given that the vanadium oxidation state in the TDMAV precursor is þ4, the oxidation state þ5 suggests a partial oxidation of V 4þ to V 5þ upon the vanadium precursor reaction with the hydroxyl groups of the TiO 2 surface at the very early ALD cycles, forming V─O bond as indicated by the peak at ≈530.4 eV.This peak is also related to the Ti─O bond from the TNTs as is denoted in Figure 2. The small spin-orbit doublet at ≈516.3 and ≈523.9 eV was attributed to V 4þ , suggesting the formation of VS 2 over the interfacial vanadium oxide.The curve fitting of the O 1s XPS spectrum allowed to discern that the main contribution of the peak is due to the Ti─O bond originated from the TiO 2 substrate, along with minor contributions from hydroxyl surface groups.Regarding the peak-fitted S 2p high-resolution XPS spectrum, it showed a set of peaks corresponding to different sulfur species: 1) two peaks at ≈164.0 and ≈165.2 eV related to surface thiol S─H group (or sulfhydryl group); and 2) two peaks at binding energies of ≈162.3 and ≈163.4 eV ascribed to disulfide S 2 2À , which could be originated from the sulfur excess as bridging sulfide moieties.The TNTs deposited with a higher number of ALD cycles (50, 100, and 200) showed significant differences both in the V 2p and S 2p spectra, as compared to the TNTs coated with 25 ALD cycles.The fitting of the V 2p and S 2p spectra was very similar to those shown in Figure 2. Therein, the presence of V 2þ (peaks located at ≈513.6 and ≈521.1 eV), V 3þ (≈514.8 and ≈522.5 eV), V 4þ (≈516.3 and ≈524.0 eV), and V 5þ (≈517.3 and ≈525.1 eV) and sulfide S 2À (≈161.4 and ≈162.6 eV), disulfide S 2 2À (≈162.1 and ≈163.3 eV), and thiol S─H groups (≈163.8 and ≈164.9 eV) was confirmed.As one can expect, the intensity of the peak associated with V─O and Ti─O bonds related to the interface between TNTs and V x S y diminished along with the increasing thickness of the V x S y film deposited.Importantly, XPS analysis also shed light on the effect of both different H 2 S and TDMAV dosing time on the chemical composition of V x S y thin film.In the former case (for a fixed TDMAV dose of 300 ms), shown in Figure S11, Supporting Information, lower H 2 S dosing times (100 and 250 ms) resulted in V x S y thin films where V 5þ was the dominant vanadium oxidation state, while higher H 2 S dosing times (500 and 1000 ms) caused a substantial increase in the presence of lower vanadium oxidation states as V 4þ , V 3þ , and V 2þ .In the latter case (for a fixed H 2 S dose of 500 ms) where various TDMAV dosing times were applied, vanadium exhibited low and high oxidation states for every TDMAV dosing time, though a greater content of the lower oxidation states V 3þ and V 2þ as compared to V 5þ and V 4þ (see Figure S12, Supporting Information) was observed for higher TDMAV dose.On the basis of the XPS results presented, we hypothesize that the origin of the V 3þ and V 2þ species could be as follows.Under certain synthesis conditions including sufficiently high H 2 S dosing times (500 ms or higher), S 2À monosulfide species (oxidation state -2) act as reducing agent during the V x S y growth and oxidize to disulfide S 2 2À species (oxidation state -1), while V 5þ species (acting as an oxidant agent) originate lower vanadium oxidation states V 4þ , V 3þ , and V 2þ .The growth of the V x S y film on TNTs after 50 ALD cycles was characterized by scanning transmission electron microscopy X-ray energy-dispersive spectroscopy (STEM-EDX) due to the extremely low thickness of the V x S y film (≈4 nm). Figure 3a shows a STEM high-angle annular dark-field (STEM-HAADF) image of a single TNT decorated with 50 ALD cycles of V x S y .The respective EDX spectra confirmed the presence of S and V (see Figure S13, Supporting Information), while Figure 3b-f represents STEM-EDX elemental maps of S, V, Ti, O, and S-V overlay corresponding to the single TNT fragment shown in Figure 3a, and identified the distribution of S and V on the TNT surface.These results revealed the homogeneous distribution of V and S over the whole surface of the TNTs, where also exist deposits of S and V as denoted in EDX maps (Figure 3b,c, respectively).Figure 3g shows a high-resolution TEM (HR-TEM) image where the ultrathin V x S y coating (≈4 nm thick in excellent agreement with the estimated growth rate per cycle) is illustrated at the outer TNT wall.The properties of the V x S y thin films as electrode material for SCs application were evaluated by postannealed TNTs coated at 100 °C with different V x S y ALD cycles, namely, 25, 50, 100, and 200.These V x S y -coated TNTs were electrochemically characterized as described in the Experimental Section by means of cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS).The electrolyte of choice was Na 2 SO 4 due to its neutral pH and the fact that V x S y electrodes revealed a large potential window in Na 2 SO 4 . [36]In order to evaluate properly the capacitance properties of V x S y , it is required to identify a potential window with non-Faradic current contribution from TNTs (what enables to ascribe any faradic current contribution to redox processes experienced by V x S y ).Thereby, the electrochemical behavior of the blank TNTs was characterized by CVs in a wide potential range (between þ1 and À1 V). Figure S14a, Supporting Information, shows the CV curve of the blank TNTs where a suitable potential window from À0.3 to þ1 V was identified, i.e., with no anodic and cathodic peaks related to redox reactions (non-Faradic current contribution).Likewise, Figure S14a, Supporting Information, shows the CV curve obtained from TNTs coated with 100 V x S y ALD cycles where two anodic peaks appeared associated with V x S y oxidation processes between ≈ þ0.2 and ≈ þ0.65 V. Based on these results, the potential window from À0.3 to þ0.5 V was selected to assess the capacitive properties of the V x S y .The stability of blank TNTs was assessed and confirmed by the application of 500 CVs, as illustrated in Figure S14b, Supporting Information.CV measurements were conducted applying different scan rates (20, 40, 60, 80, 100, and 150 mV s À1 ) in the potential window from þ0.5 to À0.3 V. Figure S15, Supporting Information, shows the CV curves obtained at different scan rates from TNTs coated with different number of V x S y ALD cycles: 25, 50, 100, and 200.The presence of an anodic peak observed for those TNTs coated with 50, 100, and 200 ALD cycles indicated an additional pseudocapacitive effect whose current density exhibited a gradual increase along with higher number of ALD cycles.The origin of the anodic peak was ascribed to the content of V 2þ and V 3þ (verified by the XPS results shown in Figure 2) and the related oxidation process(es).Interestingly, the lack of a cathodic peak suggests the irreversibility of the oxidation process.Unlike, the samples with no content of V 2þ and V 3þ (i.e., TNTs coated with 25 V x S y ALD cycles) showed no apparent anodic peak and the CV curves adopted a quasirectangular shape, indicating an electrical double-layer capacitor behavior with no pseudocapacitive contribution.Figure 4a compares the CV curves obtained from blank TNTs and counterparts coated with 25, 50, 100, and 200 V x S y ALD cycles and illustrates the trend described above.Therein, TNTs coated with V x S y exhibited larger areas under the CV curves as compared to the blank TNTs, indicating a significant enhanced electrochemical performance and the benefits of V x S y coating on the capacitive properties.The study of the cycling stability was explored by the application of 500 CVs as shown in Figure 4b, where the resulting CV curves (after 500 CVs) exhibited a decline of the electrochemical performance.In Figure S16, Supporting Information, the CV curves (1st, 100th, 250th, and 500th) at 20 mV s À1 depicted the gradual degradation of the electrochemical performance by the decrease of the area under CV curve in all the cases.In parallel, the anodic peak exhibited by as-deposited TNTs during the first cycle attenuated until it disappeared after certain number of cycles.Regarding the postperformance CV curves (i.e., the 500th CV curves), they did exhibit an electrical double-layer capacitor behavior along with slightly pseudocapacitive behavior, which was ascribed to the redox reaction occurring between thiol/disulfide redox couple.The enlarged CV area corresponding to those TNTs coated with greater number of V x S y ALD cycles indicated a higher ability to store an electrical charge.
In order to shed light on the cause(s) behind the degradation of the electrochemical performance, the postperformance TNTs were characterized by various means.SEM images in Figure S17, Supporting Information, compare the as-deposited and postperformance TNTs decorated with 100 and 200 V x S y ALD cycles.Therein, a drastic reduction of the V x S y thin film thickness was observed after electrochemical measurements.To describe such extreme reduction of the V x S y thin film thickness, it is crucial to determine the V x S y chemical stability under the experimental conditions, i.e., in the potential window applied (from À0.3 to þ0.5 V) and in the aqueous solution with a neutral pH (Na 2 SO 4 is a neutral salt pH ≈ 5.9).The multiple oxidation states that V can adopt (V 2þ , V 3þ , V 4þ , and V 5þ ) make its chemistry complex and the dissolution process certainly difficult to be fully elucidated.Nevertheless, the dissolution of the V x S y thin film could be described as multistep process as follows: 1) V 2þ oxidizes easily to V 3þ within the potential window applied originating the anodic peak at ≈ þ0.25 V.The gradual attenuation of the anodic peak along the number of CVs indicates the decrease in the V 2þ content until all the V 2þ has been oxidized to V 3þ (no anodic peak).2) At the same time, V 3þ and V 4þ at the electrode surface in direct contact with the electrolyte dissolve into the electrolyte solution by hydrolytic and/or oxidation reactions (V 4þ has a high affinity for oxygen) forming solvated and/or hydrated complexes. [37]Overall, it results in a constant and gradual dissolution of the V x S y thin film and the corresponding degradation of the electrochemical performance as illustrated by the continuous reduction of the area under the CV curves, as observed in Figure S16, Supporting Information.
EIS measurements were conducted as described in the Experimental Section in order to obtain the information about the interface charge-transfer resistance and the double-layer capacitance. [38]The electrochemical evaluation of the electrode material was conducted within the frequency range from 10 mHz to 100 kHz using a three-electrode setup.The Nyquist plots obtained from as-deposited and postperformance V x S ycoated TNTs (i.e., after 500 CVs) are shown in Figure 4c,d.The equivalent electric circuit applied to the as-deposited V x S ycoated TNTs (Circuit 1) is shown in Figure 4e.However, it must be noted that the severe dissolution of the V x S y thin film observed after electrochemical performance brought about a change in the electrical characteristics making Circuit 1 not suitable for a reliable fitting of the experimental values obtained from the postperformance V x S y -coated TNTs.Consequently, a different equivalent electric circuit named Circuit 2 (see Figure 4f ) was designed considering postperformance V x S y -coated TNTs as noncoated TNTs (with no V x S y coating). [39]The Circuit 1 employed to calculate the fitting curves for the as-deposited V x S y -coated TNTs was composed by the following elements: the resistance of the solution (R s ) connected in series with (R ct || ϕ dl ), representing the V x S y -coated TNTs, and (R c || ϕ c Z d ) representing the interface between the double layer capacitance and the V x S y thin film.The ϕ dl and ϕ c signify the constant phase element of the double layer and the compact layer, respectively; Z d represents the diffusion impedance and it is related to the ionic transport through the double-layer capacitance and the V x S y thin film, and the R ct and R c are the charge transfer resistance and compact layer resistance, respectively.In the case of the Circuit 2, the component Z d was removed due to the drastic dissolution of the V x S y thin film.The fitting curves shown in Figure 4c,d exhibited 1) a semicircle at high-mid frequency range related to the interfacial charge transfer resistance and double-layer capacitance (R ct || ϕ dl ), and 2) a linear curve at lower frequency range related to ion diffusion process (R c || ϕ c Z d for Circuit 1) and (R c || ϕ c for Circuit 2).The Nyquist plots exhibited a significant increment of the interfacial charge transfer resistance between the electrolyte and postperformance V x S y -coated TNTs, assigned to the massive loss of the V x S y thin film.On account of the continuously lowering of V x S y mass due to its gradual dissolution, the capacitance was not determined applying gravimetric methods but using EIS.Thereby, the capacitance (C) of the as-deposited and postperformance V x S y -coated TNTs was determined by the use of Equation ( 1) applying the electrical parameters of the equivalent circuit (i.e., R ct , ϕ) whose values were calculated by the fitting curves: where R ct is the transfer charge resistance, n is the exponential factor (0 ≤ n ≥ 1), showing that in this case the values of n were left free, and Q is the pseudocapacitance obtained from the electrical element ϕ dl .The calculated capacitance values (see Table 1) showed that the postperformance V x S y -coated TNTs exhibited lower capacitance as compared to as-deposited counterparts, confirming the degradation of the electrochemical performance observed in the decreasing area under the CV curve.Although direct comparison of the electrochemical capacitance performance is difficult due to the high number of different parameters involved, it was found that TNTs decorated with different materials such as NiO, [40] MnO 2 , [41] MoO 3 , [42] TiO 2 , [43] and Bi 2 O 3 [44]   have reported similar capacitance values in the order ten(s)/ hundred(s) of mF cm À2 (see Table S4, Supporting Information) as those presented herein by the postperformance V x S y -coated TNTs (see Table 1).XPS analyses of the postperformance TNTs after 500 CVs revealed meaningful identification of the electrochemically active species on the electrode surface.Figure 5 presents the peak-fitted high-resolution V 2p, O 1s, and S 2p XPS spectra for TNTs decorated with 25, 50, 100, and 200 V x S y ALD cycles after 500 CVs.Remarkably, the V 2p high-resolution XPS spectra were very similar in all the cases and exhibited a drastic intensity decline of the peaks related to V, indicating a severe drop of the content of V as compared to as-deposited TNTs.The V 2p spectra presented two peaks corresponding to the spin-orbit splitting of V2p 3/2 and V2p 1/2 at binding energies of ≈517.0 and ≈524.5 eV, respectively.The respective curve-fitted spectrum revealed a mixture of vanadium oxidation states V 4þ and V 5þ , and the absence of V oxidation states V 2þ and V 3þ .The fitting of the O 1s XPS spectra present the main peak at 530.0 eV assigned to the Ti─O bond originated from the TiO 2 substrate and V─O from the V 2 O 5 at the TNT interface, along with lower contributions from hydroxyl surface groups.In the case of the fitted S 2p highresolution XPS spectra, TNTs decorated with 25 V x S y ALD cycles showed two peaks corresponding to different S species: 1) a peak at ≈164.0 eV related to surface thiol S─H groups (or sulfhydryl groups) and 2) a peak at a binding energy of ≈168.0 eV due to sulfate groups (SO 4 ) 2À . [33]In addition to thiol S─H and sulfate (SO 4 ) 2À groups, the fitted S 2p spectra from TNTs decorated with 50, 10, and 200 V x S y ALD cycles revealed the presence of sulfide S 2À by the peaks at 162.0 and 163.2 eV, and sulfoxide S─O ascribed to the peaks at 165.4 and 166.5 eV, enriching the content of S species.Notably, XPS results in Figure 5 revealed a lack of disulfide S 2 2À species, as compared to as-deposited samples (see Figure S10, Supporting Information, please) confirming that they were reduced to thiol S─H groups during electrochemical cycling.In fact, the XPS peaks related to thiol S─H species obtained from all the postperformance V x S y -decorated TNTs samples exhibit a higher intensity, as compared to as-deposited V x S y -decorated TNTs samples, indicating a greater amount of thiol groups.The XPS elemental composition analysis (see Table S5, Supporting Information) of the postperformance TNTs exhibited interesting results.Remarkably, it was observed a dissimilar content of S as a function of the number of V x S y ALD cycles, i.e., TNTs coated with a higher number of V x S y ALD cycles contained greater content of S. Contrarily, the content of V can be considered very similar for all the postperformance TNTs regardless of the number of V x S y ALD cycles.
These results drove to consider the postperformance V x S ydecorated TNTs as S-doped TNTs after the massive loss of V (by dissolution).The postperformance TNTs were further characterized by means of STEM-EDX.Figure S18a, Supporting Information, shows a STEM-HAADF image of a fragment of a single postperformance TNT decorated with 50 ALD cycles of V x S y where no morphological changes in the nanotubular structure were observed.The STEM-EDX elemental maps corresponding to the TNT fragment of Figure S18a, Supporting Information, exhibit the chemical distribution of S, V, Ti, O, and S-V overlay (see Figure S18b-f ).It must be noted that the STEM-EDX elemental map revealed the presence of S-based nanoislands at the TNT surface which is in good agreement with the XPS results (see Figure 5), revealing the presence of different S-based species.The HR-TEM image of the TNT outer wall free of a deposit (Figure S18g, Supporting Information) illustrates the drop of the V x S y coating thickness after 500 CVs, also observed in the aforementioned SEM images of TNTs coated with 100 and 200 V x S y ALD cycles (Figure S17, Supporting Information).
In order to describe the origin of the electrochemical activity and the differences in the electrochemical performance, i.e., different capacitance values, exhibited by the postperformance V x S y -decorated TNTs, we must consider the STEM-EDX and XPS results described above.XPS elemental composition analysis results indicated the content of different S-based species as the most apparent difference in the surface chemical composition among the postperformance V x S y -decorated TNTs (see Table S5, Supporting Information).Noticeably, TNTs loading a higher content of S-based species exhibited an enlarged CV area, indicating the storage of a higher amount of charge, that is, greater capacitance values as illustrated in Figure 4b.It suggested the presence of different S-based species, i.e., sulfoxide S─O, sulfide S 2À , and thiol S─H groups on the TNTs surface as the origin of the enhanced electrochemical activity exhibited by the postperformance TNTs (as compared to blank counterparts).Actually, the benefits of sulfur doping on electrochemical properties toward SC applications have been widely reported, mostly in carbon-based materials [45][46][47][48][49] as reviewed in a recent work. [50]The superior electrochemical performance exhibited by V x S y -decorated TNTs, as compared to the blank counterpart, can be ascribed to different energy storage factors: 1) under the effect of an electric field, the polarizable nature of S and the unpaired electrons (in the 3p orbital) intensifies the polarization of the media and eventually the dielectric constant of the electrolyte, leading to larger capacitance [45] ; 2) pseudocapacitive contribution originated from the redox reactions occurring between thiol S─H and disulfide S 2 2À redox couple; 3) the presence of high-polarized S species on the TNTs walls, such as sulfate (SO 4 ) 2À , sulfoxide S─O, sulfide S 2À , and thiol S─H groups, acting as active sites for the charge storage, substantially increase the electrodeÀelectrolyte interaction and attract more electrolyte ions enlarging the charge-storage density; 4) the presence of S species on the TNTs surface results in enhanced electrical conductivity and surface wettability facilitating the electrolyte ions diffusion transport at the TNTs, ultimately enhancing the double-layer capacitive behavior.In consequence, we can consider the S-based species electrochemically active on the surface of the TNTs as the cause of the increasing the electrical charge store, and hence the superior capacitance properties.Accordingly, it is expected that a higher content of S-based species on the TNTs surface drives to superior electrochemical performance, as it was observed herein.

Conclusion
In summary, we present the novel synthesis of V x S y by ALD based on the use of TDMAV and H 2 S as precursors in a wide range of temperatures (100-225 °C).The chemical reaction mechanism between both precursors was proved to follow the ALD principles showing self-limited gas-surface reactions and a linear dependence of the V x S y thickness on the number of ALD cycles.The resulting V x S y thin films were extensively characterized by different techniques revealing amorphous smooth V x S y thin films where V adopted several oxidation states (V 2þ , V 3þ , V 4þ , V 5þ ) and the S was present in the form of different species.The effect on the chemical composition of the V x S y thin films of different ALD parameters (deposition temperature, different precursors dosing, and number of cycles) was studied by a comprehensive XPS analysis.The rich chemistry properties expected from the variety of oxidation states displayed by vanadium stimulated us to study the V x S y thin films as electrode material for SC application.The capacitance properties of V x S y thin films were thus explored by the fabrication of high surface area nanotubular composites by coating TiO 2 nanotube layers with different number of V x S y ALD cycles at 100 °C.Under the electrochemical experimental conditions, V x S y thin films showed a poor chemical stability causing after cycling test a drastic reduction of the V x S y thickness.It came along with massive loss of the V content, eventually resulting in S-doped TiO 2 nanotube layers, yet exhibiting superior electrochemical properties as compared to the blank counterparts.The presence of electrochemically active S-based species on the TNTs surface significantly contributed enhanced the electrochemical properties of the postperformance TNTs by increasing the electrical charge stored.

Experimental Section
Sample Fabrication Procedures: The deposition of V x S y thin films was carried out on substrates of different nature, such as p-type Si wafers (100) and TNTs (220 nm average inner diameter/5 μm average thickness ≈22 average aspect ratio) via ALD using the Beneq TFS-200 system.TDMAV (Strem; min.95%) and hydrogen sulfide (H 2 S, Messer, 99.5%) were used as V and S precursors, respectively.The V precursor was heated up to 60 °C to increase its vapor pressure.N 2 (99.9999%) was used as carrier gas at a flow rate of 500 standard cubic centimeters per minute (sccm) in a continuous flow process.Under these conditions, one ALD growth cycle was defined by the following sequence: TDMAV pulse (600 ms)-N 2 purge (15 s)-H 2 S pulse (1000 ms)-N 2 purge (15 s).ALD processes at different temperatures, namely, 100, 150, 190 and 225 °C were conducted applying 300 ALD cycles in order to determine the ALD window and the effect of the deposition temperature on the V x S y growth rate.Additionally, the evaluation of the self-saturated nature of the reactions to verify that the surface reactions follow the ALD principles was assessed by varying one precursor pulse length while keeping the rest of the parameters of the process fixed.The doses applied for TDMAV were 100, 300, and 600 ms (keeping H 2 S dose constant at 500 ms), while the doses for H 2 S were 100, 250, 500, and 1000 ms (keeping TDMAV constant at 600 ms).Regarding the fabrication of V x S y -coated TNTs, it was carried out at 100 °C applying different number of V x S y ALD cycles (25, 50, 100, and 200) and the following ALD conditions: TDMAV pulse (300 ms)-N 2 purge (15 s)-H 2 S pulse (500 ms)-N 2 purge (15 s).
Electrochemical anodization of Ti foils was employed for the fabrication of the self-organized TNTs as described in our previous works. [51]The asprepared amorphous TNTs were converted into anatase phase upon postannealing in a muffle oven at 400 °C for 1 h. [52]ample Characterization Methods: The study of the morphology and structure of the V x S y thin films on TNTs and Si wafer substrates as well as the cross-sectional evaluation of the V x S y deposited on Si wafers were conducted by field-emission SEM JEOL JSM 7500F.
A HR-TEM imaging and a STEM-EDX chemical mapping of V x S y decorated with 50 ALD cycles were conducted at nanoscale at single TNTs before and after CV measurements by using an image C S -corrected TEM TITAN Themis 60-300 (Thermo Fisher Scientific) operated at 300 kV and equipped with STEM HAADF imaging detector and super-X spectrometer.STEM-EDX maps were performed in net intensities from energy lines of elements of interest, i.e., O-K, S-K, Ti-K, and V-K.All TEM data were acquired and postprocessed in software Velox.
The topology of the V x S y thin films deposited on Si wafers was determined in air by an NTEGRA (NT-MDT) AFM using tapping mode with a HA-HR tip (ScenSans)and a step of 8 nm.The roughness values were obtained as the mean value of 3 measurements of a scanned area of 2 Â 2 μm 2 .
The chemical surface composition of V x S y was characterized by XPS (ESCA2SR, Scienta-Omicron) using a monochromatic Al Kα (1486.7 eV) X-ray source operated at 250 W. The binding energy scale correction was carried out using the C 1s adventitious carbon at 284.8 eV.The data analysis was performed with CasaXPS program (Casa Software Ltd.).The quantitative analysis was performed using the elemental sensitivity factors provided by the manufacturer.
XRD measurements were carried out using Panalytical Empyrean with Cu tube and Pixcel3D detector.The patterns were recorded in the range of 5-65°, using a step size of 0.026°and an incident angle of 1°.
Electrochemical Testing Methods: All electrochemical tests were carried out using a standard three-electrode setup with a carbon rod as counter electrode and Ag/AgCl (3 M KCl) as reference electrode employing an electrochemical system (ZAHNER, Messsysteme PP211, Thales 4.05 software) in an aqueous 0.1 M Na 2 SO 4 solution with an exposed area of 1cm 2 .Linear sweep CV measurements were conducted within the potential window from À0.3 to þ0.5 V (from positive to negative potentials) at different scan rates: 20, 40, 60, 80, 100, and 150 mV s À1 .Cycling stability test was performed by applying 500 cycles between À0.3 and þ0.5 V at a scan rate of 20 mV s À1 .EIS measurements were recorded at open-circuit potential in the range of frequencies from 100 kHz to 10 mHz with sine amplitude of 10 mV.The fitting of the EIS results was performed using an equivalent electrical circuit (described and showed in the Section 2).Electrical parameter values of the V x S y thin films were calculated using Gamry software.The capacitance values for the as-deposited and postperformance V x S ycoated TNTs were calculated by the use of Equation (1) (described and showed in the Section 2) applying the electrical parameters values of the equivalent circuit determined by the fitting curves.

Figure 1 .
Figure 1.a) Thickness dependence of the V x S y thin films on the deposition temperature after 300 ALD cycles on Si wafer.b) Dependence of thickness growth per cycle on the deposition temperature.c) Representative cross-sectional SEM image of a V x S y thin film with a thickness of ≈26 nm on a Si wafer after 300 cycles at a temperature of 150 °C, where the V x S y film is indicated by dashed lines.

Figure 2 .
Figure 2. High-resolution XPS spectra of V 2p and O 1s (top row) and S 2p (bottom row) obtained from TNTs deposited with 200 V x S y ALD cycles at different deposition temperatures.

Figure 3 .
Figure 3. a) HAADF-STEM image of a fragment of a single TNT decorated at 100 °C with 50 ALD cycles of V x S y .Corresponding STEM-EDX elemental maps exhibiting the distribution of b) S, c) V, d) Ti, and e) O species.f ) STEM-EDX elemental maps showing a color-mix overlapping of S-V maps.g) HR-TEM image illustrates the interface (denoted by the dashed blue line) between the TNT and the V x S y coating along the TNT wall.

Figure 4 .
Figure 4. a) CV curves obtained in the range from À0.3 to +0.5 V with a scan rate of 60 mV s À1 for blank TNTs and as-deposited counterparts coated with 25, 50, 100, and 200 V x S y ALD cycles.b) CV curves obtained at 60 mV s À1 after 500 CVs for blank TNTs and counterparts coated with 25, 50, 100, and 200 V x S y ALD cycles.Nyquist plot from c) as-deposited TNTs and d) postperformance TNTs coated with 25, 50, 100, and 200 V x S y ALD cycles.Equivalent circuit applied to e) as-deposited TNTs and f ) postperformance TNTs coated with 25, 50, 100, and 200 V x S y ALD cycles.

Table 1 .
Interfacial charge transfer resistance (R ct ) and capacitance values calculated for as-deposited and postperformance (after 500 CVs) V x S y -coated TNTs.