Enhanced Electrochemical Performance of Li 4 Ti 5 O 12 by Niobium Doping for Pseudocapacitive Applications

: Niobium-doped nanocrystalline Li 4 Ti 5 O 12 (LTO) is synthesized by the solid-state reaction method, and the inﬂuence of dopant concentration ( x = 2–10 mol%) on microstructural and electrochemical properties is studied. The X-ray diffraction and Raman patterns assessed the cubic spinel structure of Li 4 Ti 5 − x Nb x O 12 phase in all samples. Marginal changes in the lattice parameters, unit cell volume and dislocation density of LTO are observed with Nb substitution. The higher ionic radius of Nb induces a lattice expansion, which may be favorable for more ion intercalation/deintercalation. The SEM and TEM images display uniformly distributed nano-sized cubical particles. The repre-sented ( hkl ) orientations of the SAED pattern and d-spacing (0.46 nm) between bright fringes conﬁrm the well-crystallized LTO phase. The EDS and elemental mapping results demonstrate that Nb elements are uniformly doped in LTO with a proper stoichiometric ratio. The optimized 8%Nb-doped LTO electrode exhibits pseudocapacitive behavior and delivers a high speciﬁc capacitance of 497 F g − 1 at a current density of 1 A g − 1 with 92.3% of speciﬁc capacitance retention even after 5000 cycles. and electrochemical properties of LTO by doping with Nb 5+ . The Nb-doped Li 4 Ti 5 O 12 nanopowders were successfully synthesized by a simple solid-state reaction method. Li 4 Ti 5 − x Nb x O 12 samples with composition in the range 0.0 ≤ x ≤ 0.1 are investigated. Structural and mor-phological characteristics are investigated by X-ray diffraction (XRD), Raman scattering (RS) spectroscopy, scanning electron microscopy (SEM), and high-resolution transmis-sion electron microscopy (HRTEM), selected area electron diffraction (SAED), respectively. Electrochemical properties were investigated by means of cyclic voltammetry (CV) galvanostatic charge–discharge (GCD) titration, and electrochemical impedance spectroscopy (EIS). The inﬂuence of Nb in the 2–10% doping range on electrochemical properties are studied, and their suitability for use as electrodes for pseudocapacitors is demonstrated.


Introduction
The demand for electrochemical energy storage (EES) devices has grown day by day considering their potential applications in various electronic products ranging from toys to submarines. Among these devices, supercapacitors have been established as one of the proficient energy storage systems due to their high-power density, fast charge/discharge capability, and comprehensive cycle life properties [1,2]. The electrochemical activity of the supercapacitors relies effectively on the charge conservation of active electrode materials (carbon, metal oxides, and conducting polymers) [3][4][5][6]. Essentially, the transition-metal oxides (TMOs) progressed as favorable electrode materials due to excellent structural stability during the electrochemical reaction, as well as the fact that they are plentiful and less expensive. Moreover, several researchers have stated that TMO electrodes exhibit either capacitive/faradaic reactions or both and are ideal for the application of electrochemical double layer capacitors (EDLCs) or pseudocapacitors (PCs) [7][8][9][10][11]. Besides supercapacitors and batteries, the hybrid devices are an interesting combination of both devices [12].
Among all TMOs, the cubic spinel Li 4 Ti 5 O 12 (LTO) is one of the potential candidates for EES applications due to its outstanding cycle life and zero strain property during the charging/discharging process. It delivers a high theoretical capacity of 175 mAh g −1 since three or more Li + ions are accommodated per one formula unit. Moreover, a flat discharge profile at 1.55 V (vs. Li + /Li) demonstrates that it avoids the formation of SEI film. However, the LTO is a large bandgap (3.54 eV) semiconductor, in which the electrons are transferred from the 3d state of Ti to the 2p states of O, which leads to low electrical conductivity (10 −13 -10 −9 S cm −1 ) and limits the electrochemical activity [13][14][15][16]. Several
Li 4 Ti 5−x Nb x O 12 powders (0.0 ≤ x ≤ 0.1) were synthesized by the solid-state reaction method. The stoichiometric amounts of Li 2 CO 3 , TiO 2 and Nb 2 O 5 were used as precursor materials. The mixture of raw materials was ground for 3 h with a sufficient amount of ethanol poured into the powder. The obtained slurry was placed in a micro-oven and dehydrated at 80 • C for 5 h. Further, the powders were sintered in a furnace at an optimized temperature of 900 • C for 5 h in air. The crystallographic structures of prepared Nb-doped Li 4 Ti 5 O 12 powders were analyzed by X-ray diffractometry (XRD, Siefert-303 TT diffractometer). The XRD spectra were recorded with the Cu-Kα radiation source (λ = 0.15406 nm) in the angular 2θ range from 10 • to 70 • at a scan rate of 1 • min −1 . The structural analysis of Li 4 Ti 5−x Nb x O 12 powders was also performed by Raman spectroscopy utilizing a LabRAM-HR 800 confocal apparatus (Horiba, Kyoto, Japan) equipped with a He Ne laser (632 nm as excitation wavelength). The microstructure and morphology of synthesized powders were characterized by HRTEM and SEM using a HRTEM-FEI microscope (TECHNAIG2-30S-twin D905) and a SEM microscope (Carl Zeiss EVO50), respectively. The elemental composition of the synthesized Li 4 Ti 5 O 12 powder was analyzed by an energy dispersive spectrometer attached in the TEM column.
The electrochemical properties of the synthesized Nb-doped LTO samples were studied using a three-electrode aqueous glass cell composed of platinum plate (>99.95%, 10 mm × 20 mm × 0.5 mm, Latech) as the counter electrode, Ag/AgCl (+0.195 V vs. RHE, Allum Co., Kissimmee, FL, USA, model RE-1S) as the reference electrode and Nb-doped Li 4 Ti 5 O 12 as the working electrode. The working electrode loading was~1.5 mg cm −2 . The working electrode was made by the doctor blade method using the active material (wt. 80%) mixed with a conducting agent, i.e., acetylene black (wt. 10%), and a binder, i.e., polyvinylidene fluoride (wt. 10%), which are used to enhance the electronic conductivity [43]. The mixture was added with a few drops of N-methyl-2-pyrrolidone to form a slurry and deposited on the Ni substrates. The electrode was dried at 100 • C for 1 h on a hot plate, then placed in a furnace at 200 • C for 3 h. The electrochemical performance of the prepared samples was conducted in the 1 mol L −1 saturated lithium sulfate (electrolyte) aqueous solution. The cyclic voltammetry and galvanostatic charge-discharge were conducted at room temperature with a CHI 608C electrochemical workstation in the potential window ranging from −0.3 V to +0.5 V vs. Ag/AgCl. Moreover, EIS data were recorded using a CHI 608C electrochemical workstation in the frequency from 1 MHz to 1 Hz at corresponding open-circuit potentials. Figure 1a shows the X-ray diffraction patterns of Li 4 Ti 5−x Nb x O 12 (0.0 ≤ x ≤ 0.1). The diffraction peaks of pure and Nb-doped LTO powders are ascribed to the cubic spinel structure of Li 4 Ti 5 O 12 (JCPDS No. 49-0207) with the Fd3m space group [44,45]. No reflections due to impurities can be observed in the doping range 2-8%, indicating that Nb 5+ ions are successfully incorporated into the lattice of Li 4 Ti 5 O 12 without rescinding the structure or producing a new phase when x < 0.1. However, several impurity phases, such as Li 2 Ti 3 O 7 and TiO 2 , can be observed at x = 0.1, which can thus be considered as the limit of solubility of Nb in LTO. The predominant (111) reflection and other typical planes of the spinel phase specify the crystalline nature of the synthesized powders. Figure 1b shows the enlarged view of the (111) crystalline plane in the range 17.5 • ≤ 2θ ≤ 19.5 • , giving evidence of a slight shift of (111) orientation to lower diffraction 2θ angles with increasing the molar ratio of Nb +5 from 0 to 8%, which infers that Nb +5 ions are introduced in the LTO lattice. This result confirms no modification in the structure but a change of the cubic parameter. With the increasing Nb doping molar ratio, the calculated lattice parameters (Table S1 in Supplementary Materials) are slightly higher than that of the pristine LTO.

Structure and Morphology of Nb-Doped LTO
Li2Ti3O7 and TiO2, can be observed at x = 0.1, which can thus be considered as the limit of solubility of Nb in LTO. The predominant (111) reflection and other typical planes of the spinel phase specify the crystalline nature of the synthesized powders. Figure 1b shows the enlarged view of the (111) crystalline plane in the range 17.5° ≤ 2θ ≤ 19.5°, giving evidence of a slight shift of (111) orientation to lower diffraction 2θ angles with increasing the molar ratio of Nb +5 from 0 to 8%, which infers that Nb +5 ions are introduced in the LTO lattice. This result confirms no modification in the structure but a change of the cubic parameter. With the increasing Nb doping molar ratio, the calculated lattice parameters (Table S1 in Supplementary Materials) are slightly higher than that of the pristine LTO. The Rietveld refinement profile of 8%Nb-doped LTO is shown in Figure 1c and the atomic site occupancies are listed in Table S2 in Supplementary Materials. XRD analysis and Rietveld refinement demonstrate that traces of the doped Nb 5+ ions have successfully entered the lattice structure of the bulk LTO and the Nb doping does not change the spinel structure of LTO. Best fit was obtained by substituting Nb for Ti into the octahedral 16d sites. The cubic spinel Nb-doped LTO structure can be described in terms of the notation of [Li3]8a[LiTi4.92Nb0.08]16d[O12]32e within the Fd3m space group.
From previous literatures, M 5+ doped (M = Nb, Ta) had been widely proved for improving physical and chemical properties of LTO owing to the significant effects of the metal ion dopants on their inner electronic and crystalline structure. Nb dopants with a valence 5+ and similar ionic radii (0.69 Å ) as Ti 4+ (0.68 Å ) are considered to be reasonable solubility in Ti sites [30].
Crystallite size and dislocation density were extracted from XRD patterns (Table S1). The crystallite size (L) was calculated using the Debye-Scherrer formula (1) [46]: where K is a dimensionless shape factor (0.9 for quasi-spherical particles), λ is the wavelength of the X-ray source, B is the full width at half maximum (FWHM), and θ is the Bragg angle. The dislocation density (d) indicating the density of defects in the films was calculated using the following equation (2) [46]: The Rietveld refinement profile of 8%Nb-doped LTO is shown in Figure 1c and the atomic site occupancies are listed in Table S2 in Supplementary Materials. XRD analysis and Rietveld refinement demonstrate that traces of the doped Nb 5+ ions have successfully entered the lattice structure of the bulk LTO and the Nb doping does not change the spinel structure of LTO. Best fit was obtained by substituting Nb for Ti into the octahedral 16d sites. The cubic spinel Nb-doped LTO structure can be described in terms of the notation of [Li 3 ] 8a [LiTi 4.92 Nb 0.08 ] 16d [O 12 ] 32e within the Fd3m space group.
From previous literatures, M 5+ doped (M = Nb, Ta) had been widely proved for improving physical and chemical properties of LTO owing to the significant effects of the metal ion dopants on their inner electronic and crystalline structure. Nb dopants with a valence 5+ and similar ionic radii (0.69 Å) as Ti 4+ (0.68 Å) are considered to be reasonable solubility in Ti sites [30].
Crystallite size and dislocation density were extracted from XRD patterns (Table S1). The crystallite size (L) was calculated using the Debye-Scherrer formula (1) [46]: where K is a dimensionless shape factor (0.9 for quasi-spherical particles), λ is the wavelength of the X-ray source, B is the full width at half maximum (FWHM), and θ is the Bragg angle. The dislocation density (d) indicating the density of defects in the films was calculated using the following equation (2) [46]: The surface morphology of pure and Li 4 Ti 5−x Nb x O 12 (0.00 ≤ x ≤ 0.1) nanopowders was recorded using SEM and presented in Figure 2a-f. It can be clearly observed that the prepared samples are well crystallized with well-faced grains. Moreover, the morphology of the images is maintained in cubical shape with a random size distribution even at high doping concentration. The particle size distributions of the synthesized pure and Nb-doped LTO samples are illustrated in statistical histograms and shown in Figure 3a-f. All the LTO samples show a heterogeneous or random size distribution of particles in the nanoscale range. The size distribution of pure LTO powders is in the range 200-900 nm, with a maximum at 450-500 nm. Only a few particles agglomerate, giving a finite size distribution above 750 nm. By increasing the Nb-doping from 2% to 8%, the maximum is shifted to smaller size, at 250 nm for x = 0.08.
The surface morphology of pure and Li4Ti5-xNbxO12 (0.00 ≤ x ≤ 0.1) nanopowders was recorded using SEM and presented in Figure 2a-f. It can be clearly observed that the prepared samples are well crystallized with well-faced grains. Moreover, the morphology of the images is maintained in cubical shape with a random size distribution even at high doping concentration. The particle size distributions of the synthesized pure and Nbdoped LTO samples are illustrated in statistical histograms and shown in Figure 3a-f. All the LTO samples show a heterogeneous or random size distribution of particles in the nanoscale range. The size distribution of pure LTO powders is in the range 200-900 nm, with a maximum at 450-500 nm. Only a few particles agglomerate, giving a finite size distribution above 750 nm. By increasing the Nb-doping from 2% to 8%, the maximum is shifted to smaller size, at 250 nm for x = 0.08.  The chemical composition of the Li4Ti4.92Nb0.08O12 electrode was characterized by energy dispersive X-ray analysis as depicted in Figure 4. The oxygen, titanium, and niobium elemental peaks appear at energy levels of 0.154, 4.536, and 2.167 keV, respectively. The peak of the lithium element cannot appear from the EDAX spectrum because it has very low emission energy. The observed atomic percentages of Ti, O, and Nb elements of the (8%Nb-doped LTO) sample are 29.04, 69.84 and 1.12, in agreement with the chemical formula. The elemental mapping of Li4Ti4.92Nb0.08O12 is presented in Figure 5, showing that The surface morphology of pure and Li4Ti5-xNbxO12 (0.00 ≤ x ≤ 0.1) nanopowders was recorded using SEM and presented in Figure 2a-f. It can be clearly observed that the prepared samples are well crystallized with well-faced grains. Moreover, the morphology of the images is maintained in cubical shape with a random size distribution even at high doping concentration. The particle size distributions of the synthesized pure and Nbdoped LTO samples are illustrated in statistical histograms and shown in Figure 3a-f. All the LTO samples show a heterogeneous or random size distribution of particles in the nanoscale range. The size distribution of pure LTO powders is in the range 200-900 nm, with a maximum at 450-500 nm. Only a few particles agglomerate, giving a finite size distribution above 750 nm. By increasing the Nb-doping from 2% to 8%, the maximum is shifted to smaller size, at 250 nm for x = 0.08.  The chemical composition of the Li4Ti4.92Nb0.08O12 electrode was characterized by energy dispersive X-ray analysis as depicted in Figure 4. The oxygen, titanium, and niobium elemental peaks appear at energy levels of 0.154, 4.536, and 2.167 keV, respectively. The peak of the lithium element cannot appear from the EDAX spectrum because it has very low emission energy. The observed atomic percentages of Ti, O, and Nb elements of the (8%Nb-doped LTO) sample are 29.04, 69.84 and 1.12, in agreement with the chemical formula. The elemental mapping of Li4Ti4.92Nb0.08O12 is presented in Figure 5, showing that The chemical composition of the Li 4 Ti 4.92 Nb 0.08 O 12 electrode was characterized by energy dispersive X-ray analysis as depicted in Figure 4. The oxygen, titanium, and niobium elemental peaks appear at energy levels of 0.154, 4.536, and 2.167 keV, respectively. The peak of the lithium element cannot appear from the EDAX spectrum because it has very low emission energy. The observed atomic percentages of Ti, O, and Nb elements of the (8%Nb-doped LTO) sample are 29.04, 69.84 and 1.12, in agreement with the chemical formula. The elemental mapping of Li 4 Ti 4.92 Nb 0.08 O 12 is presented in Figure 5, showing that the Ti, O, and Nb elements are uniformly distributed in the compound. Since this is the highest level of doping below the limit of solubility of Nb, attention in the following will be focused on this composition. the Ti, O, and Nb elements are uniformly distributed in the compound. Since this is the highest level of doping below the limit of solubility of Nb, attention in the following will be focused on this composition.     the Ti, O, and Nb elements are uniformly distributed in the compound. Since this is the highest level of doping below the limit of solubility of Nb, attention in the following will be focused on this composition.       the Ti, O, and Nb elements are uniformly distributed in the compound. Since this is the highest level of doping below the limit of solubility of Nb, attention in the following will be focused on this composition.        Figure 7 presents the Raman spectra of pure and Li4Ti4.92Nb0.08O12 samples recorded in the 100-1000 cm −1 spectral range. LTO spinel exhibits six Raman active modes (3F2g + 1Eg + 2A1g) [47], which are observed for both samples and located at 228, 253, 366, 418-430, 670, and 755 cm −1 , respectively. The high peak in the low frequency at 228 cm −1 corresponds to the bending vibrations of O-Ti-O bonds (F2g mode). A weak Raman peak observed at 253 cm −1 is associated to the F2g mode. In the mid-frequency region, two peaks located at 366 and 418-430 cm −1 can be ascribed to the stretching vibrations of the Li-O bonds in LiO4 (F2g mode) and LiO6 polyhedra (Eg mode), respectively. The peak at 670 cm −1 (A1gmode) with a shoulder at 755 cm −1 (A1g mode) is attributed to the vibrations of Ti-O bonds in TiO6 octahedra. However, a shift (from 418 to 430 cm −1 ) of peaks associated to Li-O bonds in LiO6 polyhedra was observed when the doping concentration of Nb 5+ increases from 0 to 8%. The red shift of the peak represents a weakening of the Li-O bond in Li4Ti5-xNbxO12 samples, linked to the increase in the lattice parameters in the crystal. Moreover, no change in the peak position of 228 cm −1 corresponding to the O-Ti-O symmetric stretching mode is observed, which gives evidence of the structural stability of Nb-doped Li4Ti5O12. The electrical conductivity of 8%Nb-doped LTO is found to be 2.37 × 10 −6 S cm −1 , which is higher than that of the pure LTO (4.6 × 10 −9 S cm −1 ). This motivated the study of the Li4Ti4.92Nb0.08O12 as an electrode for the pseudocapacitive application.

Electrochemical Performance
The significant variances in the electrochemical behavior of the Nb-doped LTO electrodes are clearly observed by the shape of the CV curves in the potential window of −0.3 to 0.5 V at a scan rate of 100 mV s −1 (Figure 8). With increasing Nb substitution from 2 to 8%, the typical charge-discharge curves change from a rectangular shape without redox peaks to a quasi-rectangular shape with anodic/cathodic peaks. Moreover, current densities are increased with increasing the Nb concentration in the LTO framework. When the doping concentration increases to 10%, the anodic/cathodic peaks are suppressed. This negative impact on the lithiation process may be the conduction band filled with more ntype Nb 5+ ions and the less electroactive Ti 4+ centers. In addition, the shape of chargedischarge curves shows a shift from pseudocapacitive to a battery-like profile induced by the Nb doping, associated to an increase in the electronic conductivity. The 8%Nb-doped LTO presents good oxidation-reduction curves with broad anodic/cathodic peaks of higher current densities and lower peak potential separation (0.33 V) than the other electrodes, confirming that the optimized Li4Ti4.92Nb0.08O12 electrode displays the highest electrochemical activity. The electrical conductivity of 8%Nb-doped LTO is found to be 2.37 × 10 −6 S cm −1 , which is higher than that of the pure LTO (4.6 × 10 −9 S cm −1 ). This motivated the study of the Li 4 Ti 4.92 Nb 0.08 O 12 as an electrode for the pseudocapacitive application.

Electrochemical Performance
The significant variances in the electrochemical behavior of the Nb-doped LTO electrodes are clearly observed by the shape of the CV curves in the potential window of −0.3 to 0.5 V at a scan rate of 100 mV s −1 (Figure 8). With increasing Nb substitution from 2 to 8%, the typical charge-discharge curves change from a rectangular shape without redox peaks to a quasi-rectangular shape with anodic/cathodic peaks. Moreover, current densities are increased with increasing the Nb concentration in the LTO framework. When the doping concentration increases to 10%, the anodic/cathodic peaks are suppressed. This negative impact on the lithiation process may be the conduction band filled with more n-type Nb 5+ ions and the less electroactive Ti 4+ centers. In addition, the shape of charge-discharge curves shows a shift from pseudocapacitive to a battery-like profile induced by the Nb doping, associated to an increase in the electronic conductivity. The 8%Nb-doped LTO presents good oxidation-reduction curves with broad anodic/cathodic peaks of higher current densities and lower peak potential separation (0.33 V) than the other electrodes, confirming that the optimized Li 4 Ti 4.92 Nb 0.08 O 12 electrode displays the highest electrochemical activity. Figure 9a presents the cyclic voltammograms of the Li4Ti4.92Nb0.08O12 electrode in the voltage region of −0.3 to 0.5 V at different scan rates in the range 5-100 mV s −1 . All CV curves show a nonrectangular shape with broad anodic/cathodic peaks corresponding to a pseudocapacitive charge storage process [48]. The integral area of the CV curve proportionally increases with the increasing scan rate, and the positions of the anodic/cathodic peaks are marginally shifted. These results suggest an improvement in the reversibility of the redox reactions from the electrode. The charge storage mechanism on the surface of the electrodes influenced by the crystallite size, phase, and composition can be deduced from the exponent α of the power law =  , which relates the peak current (ip) of the CV curve to the scan rate (ν): α = 0.5 indicates a diffusion charge storage control, and α = 1 means a pure capacitive storage mechanism [49]. The value of α ≈ 0.63 deduced from the plot of log ip vs. log ν (Figure 9b) gives evidence that the charge storage is pseudocapacitive with both diffusion-controlled and capacitive-storage processes [50]. The galvanometric charge-discharge (GCD) voltage curves of the Nb-doped LTO electrodes were studied using the chronopotentiometry technique at a current density of 2 A g −1 in the potential range of −0.3 to + 0.5 V vs. Ag/AgCl. The results are shown in Figure 10a-e, in which the dashed lines represent the charge/discharge behavior of an ideal capacitor. It is evidenced that LTO electrodes doped with 2%, 4%, and 10% Nb show an almost linear discharge profile. In contrast, the lineament of the GCD curves of 6% and 8%Nb-doped samples indicates that charges are partially intercalated/de-intercalated into/from the surface of the electrode like pseudocapacitive charge storage reactions,  All CV curves show a nonrectangular shape with broad anodic/cathodic peaks corresponding to a pseudocapacitive charge storage process [48]. The integral area of the CV curve proportionally increases with the increasing scan rate, and the positions of the anodic/cathodic peaks are marginally shifted. These results suggest an improvement in the reversibility of the redox reactions from the electrode. The charge storage mechanism on the surface of the electrodes influenced by the crystallite size, phase, and composition can be deduced from the exponent α of the power law i p = Kν α , which relates the peak current (i p ) of the CV curve to the scan rate (ν): α = 0.5 indicates a diffusion charge storage control, and α = 1 means a pure capacitive storage mechanism [49]. The value of α ≈ 0.63 deduced from the plot of log i p vs. log ν (Figure 9b) gives evidence that the charge storage is pseudocapacitive with both diffusion-controlled and capacitive-storage processes [50]. Figure 9a presents the cyclic voltammograms of the Li4Ti4.92Nb0.08O12 electrode in the voltage region of −0.3 to 0.5 V at different scan rates in the range 5-100 mV s −1 . All CV curves show a nonrectangular shape with broad anodic/cathodic peaks corresponding to a pseudocapacitive charge storage process [48]. The integral area of the CV curve proportionally increases with the increasing scan rate, and the positions of the anodic/cathodic peaks are marginally shifted. These results suggest an improvement in the reversibility of the redox reactions from the electrode. The charge storage mechanism on the surface of the electrodes influenced by the crystallite size, phase, and composition can be deduced from the exponent α of the power law =  , which relates the peak current (ip) of the CV curve to the scan rate (ν): α = 0.5 indicates a diffusion charge storage control, and α = 1 means a pure capacitive storage mechanism [49]. The value of α ≈ 0.63 deduced from the plot of log ip vs. log ν (Figure 9b) gives evidence that the charge storage is pseudocapacitive with both diffusion-controlled and capacitive-storage processes [50]. The galvanometric charge-discharge (GCD) voltage curves of the Nb-doped LTO electrodes were studied using the chronopotentiometry technique at a current density of 2 A g −1 in the potential range of −0.3 to + 0.5 V vs. Ag/AgCl. The results are shown in Figure 10a-e, in which the dashed lines represent the charge/discharge behavior of an ideal capacitor. It is evidenced that LTO electrodes doped with 2%, 4%, and 10% Nb show an almost linear discharge profile. In contrast, the lineament of the GCD curves of 6% and 8%Nb-doped samples indicates that charges are partially intercalated/de-intercalated into/from the surface of the electrode like pseudocapacitive charge storage reactions, The galvanometric charge-discharge (GCD) voltage curves of the Nb-doped LTO electrodes were studied using the chronopotentiometry technique at a current density of 2 A g −1 in the potential range of −0.3 to + 0.5 V vs. Ag/AgCl. The results are shown in Figure 10a-e, in which the dashed lines represent the charge/discharge behavior of an ideal capacitor. It is evidenced that LTO electrodes doped with 2%, 4%, and 10% Nb show an almost linear discharge profile. In contrast, the lineament of the GCD curves of 6% and 8%Nb-doped samples indicates that charges are partially intercalated/de-intercalated into/from the surface of the electrode like pseudocapacitive charge storage reactions, which is consistent with the CV profiles ( Figure 8 and Figure S1 in Supplementary Materials). GCD profiles display a shift from pseudocapacitive to battery-like behavior induced by Nb doping. The specific capacity of the Li 4 Ti 5−x Nb x O 12 pseudocapacitive electrodes are 39.4, 57.5, 95.8, 111.1 and 60.2 mAh g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.10, respectively, at a given current density of 2 A g −1 . Additionally, the 8%Nb-doped LTO delivered specific capacities of 138 and 75 mAh g −1 at current densities of 1 and 4 A g −1 . The coulombic efficiency of the prepared 2%, 4%, 6%, 8%, and 10%Nb-doped LTO electrodes was found to be 55.7%, 69%, 79.2%, 98.5% and 42.3%, respectively (see Figure S2 in Supplementary Materials).
Micro 2021, 1, FOR PEER REVIEW 9 which is consistent with the CV profiles ( Figure 8 and Figure S1 in Supplementary Materials). GCD profiles display a shift from pseudocapacitive to battery-like behavior induced by Nb doping. The specific capacity of the Li4Ti5-xNbxO12 pseudocapacitive electrodes are 39.4, 57.5, 95.8, 111.1 and 60.2 mAh g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.10, respectively, at a given current density of 2 A g −1 . Additionally, the 8%Nb-doped LTO delivered specific capacities of 138 and 75 mAh g −1 at current densities of 1 and 4 A g −1 . The coulombic efficiency of the prepared 2%, 4%, 6%, 8%, and 10%Nb-doped LTO electrodes was found to be 55.7%, 69%, 79.2%, 98.5% and 42.3%, respectively (see Figure S2 in Supplementary Materials). It appears that the Li4Ti4.92Nb0.08O12 electrode is discharged over a longer time than other electrodes, proving the higher electrochemical capability of the charge storage. The calculated specific capacitances of the Li4Ti5-xNbxO12 electrodes are found to be 142, 207, 345, 400, and 217 F g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.1, respectively, at a given current density of 2 A g −1 (Figure 11), which illustrates the improvement in the electrochemical behavior of the electrodes with increasing the niobium concentration from 0.02 to 0.08. Figure 11. Comparison of the specific capacitances of Nb-doped LTO electrodes at 2 A g −1 . Figure 12a presents the GCD profiles of the Li4Ti4.92Nb0.08O12 electrode carried out at current densities of 1, 2 and 4 A g −1 . At lower current, more charges participate in the electrochemical reactions, while a low number of charges contribute at higher current densities due to time constraints. The calculated specific capacitance of the electrode was found to be 497, 400, and 270 F g −1 at applied current densities of 1, 2, and 4 A g −1 , respectively. Correspondingly, the cycling stability of the Li4Ti4.92Nb0.08O12 electrode was studied over 5000 charge-discharge cycles as shown in Figure 12b. The capacitance retention is 92.3% at a current density of 1 A g −1 after 5000 cycles. It appears that the Li 4 Ti 4.92 Nb 0.08 O 12 electrode is discharged over a longer time than other electrodes, proving the higher electrochemical capability of the charge storage. The calculated specific capacitances of the Li 4 Ti 5−x Nb x O 12 electrodes are found to be 142, 207, 345, 400, and 217 F g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.1, respectively, at a given current density of 2 A g −1 (Figure 11), which illustrates the improvement in the electrochemical behavior of the electrodes with increasing the niobium concentration from 0.02 to 0.08.
Micro 2021, 1, FOR PEER REVIEW 9 which is consistent with the CV profiles ( Figure 8 and Figure S1 in Supplementary Materials). GCD profiles display a shift from pseudocapacitive to battery-like behavior induced by Nb doping. The specific capacity of the Li4Ti5-xNbxO12 pseudocapacitive electrodes are 39.4, 57.5, 95.8, 111.1 and 60.2 mAh g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.10, respectively, at a given current density of 2 A g −1 . Additionally, the 8%Nb-doped LTO delivered specific capacities of 138 and 75 mAh g −1 at current densities of 1 and 4 A g −1 . The coulombic efficiency of the prepared 2%, 4%, 6%, 8%, and 10%Nb-doped LTO electrodes was found to be 55.7%, 69%, 79.2%, 98.5% and 42.3%, respectively (see Figure S2 in Supplementary Materials). It appears that the Li4Ti4.92Nb0.08O12 electrode is discharged over a longer time than other electrodes, proving the higher electrochemical capability of the charge storage. The calculated specific capacitances of the Li4Ti5-xNbxO12 electrodes are found to be 142, 207, 345, 400, and 217 F g −1 for x = 0.02, 0.04, 0.06, 0.08 and 0.1, respectively, at a given current density of 2 A g −1 (Figure 11), which illustrates the improvement in the electrochemical behavior of the electrodes with increasing the niobium concentration from 0.02 to 0.08. Figure 11. Comparison of the specific capacitances of Nb-doped LTO electrodes at 2 A g −1 . Figure 12a presents the GCD profiles of the Li4Ti4.92Nb0.08O12 electrode carried out at current densities of 1, 2 and 4 A g −1 . At lower current, more charges participate in the electrochemical reactions, while a low number of charges contribute at higher current densities due to time constraints. The calculated specific capacitance of the electrode was found to be 497, 400, and 270 F g −1 at applied current densities of 1, 2, and 4 A g −1 , respectively. Correspondingly, the cycling stability of the Li4Ti4.92Nb0.08O12 electrode was studied over 5000 charge-discharge cycles as shown in Figure 12b. The capacitance retention is 92.3% at a current density of 1 A g −1 after 5000 cycles.  Figure 12a presents the GCD profiles of the Li 4 Ti 4.92 Nb 0.08 O 12 electrode carried out at current densities of 1, 2 and 4 A g −1 . At lower current, more charges participate in the electrochemical reactions, while a low number of charges contribute at higher current densities due to time constraints. The calculated specific capacitance of the electrode was found to be 497, 400, and 270 F g −1 at applied current densities of 1, 2, and 4 A g −1 , respectively. Correspondingly, the cycling stability of the Li 4 Ti 4.92 Nb 0.08 O 12 electrode was studied over 5000 charge-discharge cycles as shown in Figure 12b. The capacitance retention is 92.3% at a current density of 1 A g −1 after 5000 cycles. The charge storage mechanism of the electrode significantly depends on the electrode internal resistance, electrolyte resistance, and interface resistance between electrode and electrolyte. These are determined from the Nyquist plots of the synthesized electrodes shown in Figure 13 along with the equivalent circuit composed of a series of resistances (Rs), which corresponds to the internal resistance of the cell (i.e., electrolyte and electrodes contribution), a charge transfer resistance (Rct) in parallel with a constant phase element (CPE) and a Warburg element (inset of Figure 13). In the high-frequency region, Rs and Rct of the electrodes were observed from the intercept of the x-axis and the diameter of the semicircle, respectively, while the Warburg impedance (Zw ≈ σwω −1/2 ) of the element W is determined by the slope in the low-frequency region. The obtained Rs, Rct and σw values for pristine and Nb-doped LTO electrodes are listed in Table 1. Values are decreased when increasing the doping percentage of niobium from 2 to 8% and then increased. This may be due to an increase in lattice parameter by increasing the doping percentage and hence improves electron mean free path and contact area between the electrode and electrolyte. Note the slight difference between Rs values (31.0 Ω vs. 18.1 Ω) is due to the high ohmic resistivity of pristine LTO compared with the doped electrode. The Warburg part shows an identical slope for x = 8% and x = 2%, which suggests that the diffusion coefficient of Li is not affected by the doping. It, therefore, seems that the effect of doping is only the increase in electrical conductivity, which is not surprising, due to the small variation of the crystalline parameter and the absence of lithium vacancies.   The charge storage mechanism of the electrode significantly depends on the electrode internal resistance, electrolyte resistance, and interface resistance between electrode and electrolyte. These are determined from the Nyquist plots of the synthesized electrodes shown in Figure 13 along with the equivalent circuit composed of a series of resistances (R s ), which corresponds to the internal resistance of the cell (i.e., electrolyte and electrodes contribution), a charge transfer resistance (R ct ) in parallel with a constant phase element (CPE) and a Warburg element (inset of Figure 13). In the high-frequency region, R s and R ct of the electrodes were observed from the intercept of the x-axis and the diameter of the semicircle, respectively, while the Warburg impedance (Z w ≈ σ w ω −1/2 ) of the element W is determined by the slope in the low-frequency region. The obtained R s , R ct and σ w values for pristine and Nb-doped LTO electrodes are listed in Table 1. Values are decreased when increasing the doping percentage of niobium from 2 to 8% and then increased. This may be due to an increase in lattice parameter by increasing the doping percentage and hence improves electron mean free path and contact area between the electrode and electrolyte. Note the slight difference between Rs values (31.0 Ω vs. 18.1 Ω) is due to the high ohmic resistivity of pristine LTO compared with the doped electrode. The Warburg part shows an identical slope for x = 8% and x = 2%, which suggests that the diffusion coefficient of Li is not affected by the doping. It, therefore, seems that the effect of doping is only the increase in electrical conductivity, which is not surprising, due to the small variation of the crystalline parameter and the absence of lithium vacancies. The charge storage mechanism of the electrode significantly depends on the electrode internal resistance, electrolyte resistance, and interface resistance between electrode and electrolyte. These are determined from the Nyquist plots of the synthesized electrodes shown in Figure 13 along with the equivalent circuit composed of a series of resistances (Rs), which corresponds to the internal resistance of the cell (i.e., electrolyte and electrodes contribution), a charge transfer resistance (Rct) in parallel with a constant phase element (CPE) and a Warburg element (inset of Figure 13). In the high-frequency region, Rs and Rct of the electrodes were observed from the intercept of the x-axis and the diameter of the semicircle, respectively, while the Warburg impedance (Zw ≈ σwω −1/2 ) of the element W is determined by the slope in the low-frequency region. The obtained Rs, Rct and σw values for pristine and Nb-doped LTO electrodes are listed in Table 1. Values are decreased when increasing the doping percentage of niobium from 2 to 8% and then increased. This may be due to an increase in lattice parameter by increasing the doping percentage and hence improves electron mean free path and contact area between the electrode and electrolyte. Note the slight difference between Rs values (31.0 Ω vs. 18.1 Ω) is due to the high ohmic resistivity of pristine LTO compared with the doped electrode. The Warburg part shows an identical slope for x = 8% and x = 2%, which suggests that the diffusion coefficient of Li is not affected by the doping. It, therefore, seems that the effect of doping is only the increase in electrical conductivity, which is not surprising, due to the small variation of the crystalline parameter and the absence of lithium vacancies.

Discussion
In this study, we have shown that niobium doping in Li 4 Ti 5−x Nb x O 12 in the compositional range 0.02 ≤ x ≤ 0.10 does not affect the structural properties of the LTO electrode. All compounds have crystallite size of~40 nm and crystallize with cubical morphology. The partial substitution of Nb for Ti only slightly increases the unit cell volume and enhances the electrical conductivity of the spinel lattice as expected. The electrochemical capacitance, cycle stability, and Nyquist data confirm a decrease in the charge transfer resistance and an increase in the electrical conductivity. The small lattice expansion probably does not create large free 3D channels for enhancing the lithium ions' transport during the insertionextraction reactions. However, a significant increase in the electrical conductivity (by almost three orders of magnitude) has been observed. Wolfenstine and Allen [51] reported that the DC electronic conductivity of Li 4 Ti 5 O 12 increased from~8 × 10 −10 to~1 × 10 −9 S cm −1 when doped with 5% Ta. Similarly, Zhang et al. [52] show that the substitution of W 6+ ions for Ti 4+ ions in the structure, i.e., Li 4 Ti 5-x W x O 12 increases the conductivity by several orders of magnitude from σ ≈ 10 −13 S cm −1 (pristine LTO) to σ = 1.5 × 10 −1 S cm −1 (Li 4 Ti 4.9 W 0.1 O 12 ) but with a slight sacrifice of capacity at low C-rate. Tian et al. [41] shows that the conductivity of Li 4  In the current study, the effect of Nb 5+ doping in LTO materials was investigated for high-performance supercapacitors in depth. The pentavalent donor-type doping is needed to improve the electrical conductivity of the material, and Nb 5+ ions can effortlessly substitute Ti 4+ in the lattice of Li 4 Ti 5 O 12 in an extensive series of concentrations. From the experimental results, the crystal size, phase purity, inter-planar distance, etc., are parameters, which are effectively dependent on the doping level of Nb. As a result, the 8%Nb-doped LTO had better electrochemical activity than others due to its small particle size, high conductivity, and good ion diffusivity associated with the increased lattice parameters. It is obvious that 10%-Nb doping in LTO has a negative impact on the lithiation process since the conduction band is filled with more n-type Nb 5+ ions and the number of electroactive Ti 4+ centers are diminished. According to the results, adding a small amount of Nb to the substance could theoretically improve its overall performance by enhancing its metallic-like character, while having slight effects on its electrochemical properties. Contrastingly, a higher concentration of Nb may lead to poor electrochemical efficiency. From the CV and GCD studies, it can be concluded that the 8%Nb-doped LTO demonstrated both faradaic and capacitive storage mechanisms. On the surface of the electrodes, faradaic and capacitive processes occurred at Nb-doped centers and Ti sites, respectively. As a result, Nb 5+ ions increase the electrical conductivity and charge storage of the working electrode.
According the EIS results from Figure 13, the comparison of the low-frequency range corresponding to the ionic conductivity does not show a significant change between the 8%Nb-doped LTO and pristine LTO sample. This indicates that the concentration of Li vacancies (as expected for an increase in the Li-ion diffusion) is not changed in the doped-LTO structure. The large increase in the electronic conductivity from 4.6 × 10 −9 S cm −1 (pure LTO) to 2.37 × 10 −6 S cm −1 (Li 4 Ti 4.92 Nb 0.08 O 12 ) is due to a charge compensation by a Ti vacancy given in Equation (3) 12 have been investigated. The specific capacitance of the 8%Nb-doped LTO anode material is 497 F g −1 after 5000 charge-discharge cycles at a current rate of 1 A g −1 (or 6C rate) and a specific discharge capacity of 111.1 mAh g −1 at current density of 2 A g -1 . A comparison of our results with the electrochemical properties of doped LTO electrodes (i.e., energy density, power density) reported in the literature is reported in Table 2. In terms of capacity (mAh g −1 ) or capacitance (F g −1 ), there electrodes include AC/LTO (~45 mAh g −1 ) [54], 80LFP-20AC/LTO (~70 mAh g −1 ) [54], AC/LTO (27 mAh g −1 ) [37], LTO/graphene (207 F g −1 ) [39], sol-gel prepared LTO (103 mAh g −1 at 0.1C) [34], pre-lithiated LTO (79 mAh g −1 at 0.2C) [35], porous LTO granules (43 F g −1 at 0.8 A g −1 rate) [36]. As a result, the discharge rate of the 8%Nb-doped LTO is faster than that of the nanostructured LTO materials reported so far, emphasizing a great promise of the Nb-doped intercalation-type material for high-energy and high-power asymmetric supercapacitor applications.

Conclusions
Niobium-doped (2-10 mol%) nanocrystalline Li 4 Ti 5 O 12 electrodes were successfully synthesized by solid-state reaction with a sintering temperature of 900 • C. The microstructural and elemental compositions of the synthesized electrodes were analyzed using XRD, SEM, HRTEM, Raman, and EDS, respectively. A single spinel phase (cubic Fd3m structure) remains in Li 4 Ti 5−x Nb x O 12 doped with Nb in the range 0.02 ≤ x ≤ 0.08. The (111) peak shift to lower 2θ angles indicates a marginal increase in the lattice parameter and elementary volume. SEM and TEM images displayed homogeneously distributed cubical grains with size of nanometer range. The particle size of 8%Nb-doped LTO electrodes is found in the range of 200-250 nm. The EDS spectra and elemental mappings show the presence of Ti, O, and Nb with stoichiometric ratios in the prepared electrodes. The Raman spectrum of 8%Nb-doped LTO displayed six Raman active modes with a small red shift in Li−O and no change in O−Ti−O bonds. The electrochemical performance of the synthesized Nb-doped LTO electrodes studied using CV, GCD, and EIS techniques shows that the optimized Li 4 Ti 4.92 Nb 0.08 O 12 sample exhibited a pseudocapacitive behavior. This electrode delivers a specific capacitance of 497 F g −1 at a current density of 1 A g −1 with 92.3% capacitance retention after 5000 cycles. Hence, Li 4 Ti 4.92 Nb 0.08 O 12 is a promising material as an electrode for pseudocapacitive applications.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/micro1010004/s1, Table S1: The lattice parameters of the Li 4 Ti 5−x Nb x O 12 samples synthesized by solid-state reaction at 900 • C; Table S2: Results of the Rietveld refinement for pristine LTO and 8%Nb-doped LTO; Figure S1