Impact of Electrical Conductivity on the Electrochemical Performances of Layered Structure Lithium Trivanadate (LiV3-x M x O8, M= Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo, x = 0.01-0.1) as Cathode Materials for Energy Storage.

Pristine trivanadate (LiV3O8) and doped lithium trivanadate (LiV3-x M x O8, M = Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo, x = 0.01/0.05/0.1 M) compounds were prepared by a simple reflux method in the presence of the polymer, Pluronic P123, as the chelating agent. For comparison, pristine LiV3O8 alone was also prepared in the absence of the chelating agent. The Rietveld-refined X-ray diffraction patterns shows all compounds to exist in the layered monoclinic LiV3O8 phase belonging to the space group of P21/m. Scanning electron microscopy analysis shows the particles to exhibit layers of submicron-sized particles. The electrochemical performances of the coin cells were compared at a current density of 30 mA/g in the voltage window of 2-4 V. The cells made with compounds LiV2.99Zr0.01O8 and LiV2.95Sn0.05O8 show a high discharge capacity of 245 ± 5 mA h/g, with an excellent stability of 98% at the end of the 50th cycle. The second cycle discharge capacity of 398 mA h/g was obtained for the compound LiV2.99Fe0.01O8, and its capacity retention was found to be 58% after 50 cycles. The electrochemical performances of the cells were correlated with the electrical properties and the changes in the structural parameters of the compounds.

Introduction

Secondary lithium ion batteries with high energy density, power density, and long cycling stability are considered as the heart of the growing mobile technology and future electric vehicles. For intensive commercial applications, drawbacks such as costly battery materials and the safety issues associated with them have to be addressed. Layered structured cathode materials such as lithium trivanadate (LiV3O8), vanadium pentoxide (V2O5), and other vanadium derivatives have good scope for lithium ion battery technology because of their low cost, high energy density, and better safety features.[1−5] LiV3O8 (LVO) is a nonlithiated cathode material in which the lithium ions can freely intercalate into and deintercalate from the LVO’s layered structure, exhibiting better structural stability than V2O5 and other vanadium derivatives.[6−9] It has a layered monoclinic structure with a space group of P21/m, reported by Wadsley in 1957.[10] It consists of V–O layers arranged one above the other made up of two interconnected units by corner-sharing oxygen atoms, such as the VO6 octahedra unit and VO5 trigonal bipyramid units, providing structural stability.[11−13]

However, practical batteries with LVO electrodes show capacity degradation and poor rate performance because of their low electrical conductivity.[14] This was significantly overcome by means of cation doping such as Na, K, Cu, and Ag in the lithium site and Mo, Si, Mn, Ni, Cr, and so forth in the vanadium sites.[15−22] But the clear role of doping on the electrical property of the compound LiV3O8 and its influence on the electrochemical performance of the cell have not been reported in any of the literature. Research on the inter-relation between the oxidation states of metal ions, the role of ionic and electronic conductivities of the compound, and the electrochemical performance of the battery is the need of the hour for the betterment of battery materials. Recently, our group reported on the correlation between the electrical and electrochemical performance of the cathode materials, LiNi0.5Mn0.5O2 and LiNi1/3Co1/3Mn1/3O2.[23,24] In the present work, lithium trivanadate and LiV3–MO8 (M = Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo) compounds with dopants of different oxidation states were prepared at different levels of doping x = 0.01/0.05/0.1 M. The changes in the electrical properties of the compound upon doping were explored. The materials were prepared by a very simple and cost-effective Pluronic P123-assisted reflux method. Studies on the doping effect of Zn/Fe/Sn/Nb at the vanadium site of LiV3O8 have not been reported elsewhere.

Results and Discussion

Structure and Morphology

X-ray diffraction (XRD) patterns of LVO-P, LVO-WP, and doped compounds (0.05 M level of doping) are shown in Figure a,b (Figure S1a,b for 0.01 and 0.1 M level doped compounds, respectively). The XRD patterns of all compounds correspond to the layered monoclinic crystalline phase with space group P21/m.[13−15] The lattice parameter values for LVO-P and LVO-WP were found to be a = 6.700 Å, b = 3.625 Å, and c = 12.096 Å and a = 6.619 Å, b = 3.578 Å, and c = 11.939 Å, respectively. The lattice parameters “a”, “b,” and “c” were found to be in the range of 6.516–6.772, 3.485–3.677, and 11.835–12.241 Å, respectively, for doped compounds and are given in Table . The lattice parameter values of all compounds were found closely matched with JCPDS card no. 72-1193.

Figure 1

(a) Rietveld-refined XRD patterns of LVO-P and LVO-WP. (b) XRD pattern of LiV2.95M0.05O8 (M = Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo) compounds.

Table 1

Lattice Parameters from Rietveld-Refined XRD Patterns

doping level	compound	a (Å)	b (Å)	c (Å)
	LVO-P	6.700	3.625	12.096
	LVO-WP	6.619	3.578	11.939
0.01 M	LiZn0.01V2.99O8	6.654	3.603	12.051
	LiCo0.01V2.99O8	6.667	3.607	12.034
	LiFe0.01V2.99O8	6.642	3.591	11.980
	LiSn0.01V2.99O8	6.655	3.600	12.011
	LiTi0.01V2.99O8	6.581	3.561	11.879
	LiZr0.01V2.99O8	6.593	3.564	11.891
	LiNb0.01V2.99O8	6.651	3.596	12.000
	LiMo0.01V2.99O8	6.649	3.599	12.012
0.05 M	LiZn0.05V2.95O8	6.597	3.563	11.878
	LiCo0.05V2.95O8	6.663	3.603	12.016
	LiFe0.05V2.95O8	6.571	3.550	11.835
	LiSn0.05V2.95O8	6.576	3.554	11.852
	LiTi0.05V2.95O8	6.583	3.559	11.872
	LiZr0.05V2.95O8	6.603	3.567	11.886
	LiNb0.05V2.95O8	6.664	3.605	12.024
	LiMo0.05V2.95O8	6.607	3.581	11.944
0.1 M	LiZn0.1V2.9O8	6.561	3.496	11.765
	LiCo0.1V2.9O8	6.772	3.677	12.241
	LiFe0.1V2.9O8	6.541	3.569	11.878
	LiSn0.1V2.9O8	6.516	3.485	11.862
	LiTi0.1V2.9O8	6.581	3.559	11.883
	LiZr0.1V2.9O8	6.591	3.560	11.874
	LiNb0.1V2.9O8	6.585	3.562	11.880
	LiMo0.1V2.9O8	6.627	3.594	11.988

On comparison with LVO-P, the shift in the peak corresponding to the (100) plane either toward the higher or lower 2θ angle (Figure ) and significant changes in the intensity of the (100) plane was observed for all doped compounds. Shifting of the peak corresponding to the (100) plane toward lower 2θ angle indicates the increase in the interlayer distance d100 and vice versa (Table S1). The increase in d100 could be better for the Li+ ion insertion and reinsertion, which will lead to better electrochemical performance. On comparison with LVO-P, the shift in the peak corresponding to the (100) plane toward the lower 2θ value was observed for the compounds LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, and LiV2.99Fe0.01O8, as shown in Figure d. The interlayer distance value (d100) was found to be 6.352, 6.356, 6.366, and 6.370 Å for the compounds LVO-P, LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, and LiV2.99Fe0.01O8, respectively. The XRD pattern of the compound LiV2.9Co0.1O8 shows a very low interlayer distance value (d100 ≈ 5.951 Å), which may result in poor electrochemical performance. All compounds show minor peaks at 2θ ≈ 22°, corresponding to the LiV2O5 phase (5–10%), which is commonly observed in the LiV3O8 compound prepared through any of the reported methods,[25−30] and are considered as an active phase in LiV3O8, which helps in improving the electrochemical performance of the battery.[13] No other impurity phases were observed at any of the doping level.

Figure 2

Peak shift corresponding to the (100) plane in (a) LiV2.99M0.01O8, (b) LiV2.95M0.05O8, (c) LiV2.9M0.1O8 (M = Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo) compounds, and (d) LVO-P, LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, and LiV2.99Fe0.01O8 compounds.

Figure shows the morphology of the selected compounds, LVO-P, LVO-WP, LiV2.99Zr0.01O8, LiV2.99Fe0.01O8, and LiV2.95Sn0.05O8, and other 0.05 M level doped compounds. Scanning electron microscopy (SEM) analysis shows the particles to have submicron-sized particles. Rod-shaped structures can be clearly observed for the compound LVO-P and Zn, Zr, Sn, Fe-doped compounds ranging in length from 4 to 10 μm. The irregular-shaped particles were observed for LVO-WP. Changes in particle sizes were observed upon doping. The BET surface area and pore volume were found to be 0.68, 0.57, 1.20, 1.11, and 1.01 m2/g and 0.16, 0.13, 0.28, 0.26, and 0.23 cm3/g for the compounds LVO-P, LVO-WP, LiV2.99Zr0.01O8, LiV2.99Fe0.01O8, and LiV2.95Sn0.05O8, respectively. The morphologies of 0.01 and 0.1 M level of doped compounds were also found to have rod-shaped submicron-sized particles (images not shown).

Figure 3

SEM images of bare and doped LiV3O8 compounds.

Raman Spectroscopy and Electrical Studies

Raman and electrical studies were carried out for the parent and selected doped compounds based on their better electrochemical performance (to be discussed under galvanostatic studies).

Raman analysis is considered as a more suitable method to investigate the effect of dopants on the LiV3O8 crystal lattice. The Raman spectra of the select compounds are shown in Figure a,b. For LVO-P and LVO-WP, the observed band at 773 cm–1 is related to the atomic motions of the corner-sharing oxygen among the VO6, VO5, and LiO6 polyhedrons of LiV3O8.[31] The Raman bands at 990 and 969 cm–1 could be attributed to the V–O stretching vibrations of VO5 pyramids. The weak vibrations of the B1g and Ag symmetry species of LiV3O8 were observed at 540 and 477 cm–1, respectively.[32]

Figure 4

Raman spectra for (a) LVO-P and LVO-WP and (b) compounds LVO-P, LiV2.99Zr0.01O8, LiV2.99Fe0.01O8, and LiV2.95Sn0.05O8.

The changes in the peak position and relative intensities between the peaks for doped compounds shown in Figure b indicate doping effect. The minor peak corresponding to the LiV2O5 phase was observed for all compounds (indicated by * symbol), as discussed earlier in the XRD section. The vibrations of the B1g and Ag symmetry species were shifted to a lower wave number region in the order of LiV2.95Sn0.05O8, LiV2.95Zr0.01O8, and LiV2.95Fe0.01O8. This indicated the rapid formation of VO5 and VO6 polyhedrons and thus a little distortion of the crystal lattice upon doping[32] because the ionic radii of doping ions are greater than that of V5+. Moreover, the intensity ratio between the two peaks at 990 and 773 cm–1 were found to be in the order of LiV2.99Zr0.01O8 > LiV2.99Fe0.01O8 > LiV2.95Sn0.05O8 > LVO-P compound. This behavior suggests that the formation rate of VO5 and VO6 polyhedrons of V3O8– layers was much faster than the superposition rate of V3O8– layers linked by the interlayered Li+ ions in LiV2.95Zr0.01O8.[31,32]

Electrical Studies

On the basis of the electrochemical performances (to be discussed later), electrical studies were conducted for the selected compounds LVO-P, LVO-WP, LiZr0.01V2.99O8, LiSn0.05V2.95O8, LiFe0.01V2.99O8, and other doped LiV2.95M0.05O8 (M = Zn/Co/Fe/Ti/Zr/Nb/Mo) compounds by means of ac impedance, four-probe method, and transference number studies.[23,24]Figure a represents the complex impedance plot for LVO-P, LVO-WP, LiZr0.01V2.99O8, LiSn0.05V2.95O8, and LiFe0.01V2.99O8 taken at room temperature (303 K) in the frequency range of 42 Hz to 1 MHz. The Nyquist plot shows the overlapping of two semicircles related to the parallel combination of a resistor and a capacitor. High-frequency semicircle is attributed to the grain, and low-frequency semicircle is related to the grain boundary effect.[23] The bulk resistance (Rb) of the sample was calculated by the intercept of the high-frequency semicircle on the x-axis. The electrical conductivity was calculated using the conductivity formula σ = l/RbA, where “l” and “A” are the thickness and area of the pellet, respectively (Table ) and the values were verified with the conductivity values calculated from the mid-frequency plateau region in the conductivity spectra (Figure S2). For comparison, the dc four-probe conductivity value was also measured from the V/I spectra (Figure S3) and matched with the impedance analysis. This validates our measurements. The compound LVO-P shows 1 order of higher electrical conductivity (10–6 S/cm) than the compound LVO-WP (10–7 S/cm). The electrical conductivity of doped compounds was in the order of ∼10–6 S/cm.

Figure 5

(a) Nyquist plot (Z′ vs −Z″) and (b) transference number measurements for compounds LVO-P, LVO-WP, LiV2.99Zr0.01O8, LiV2.99Fe0.01O8, and LiV2.95Sn0.05O8.

Table 2

Data from Electrical Studies of the As-Prepared Compounds

	conductivity (S/cm) (±0.1)			transference number (±0.02)
samples	Nyquist plot	conductivity spectra	dc four-probe	ionic (ti)	electronic (te)
LVO-P	7.93 × 10–6	8.12 × 10–6	7.59 × 10–6	0.81	0.19
LVO-WP	6.49 × 10–7	6.58 × 10–7	6.67 × 10–7	0.77	0.23
LiV2.99Zr0.01O8	7.28 × 10–6	7.14 × 10–6	6.97 × 10–6	0.60	0.40
LiV2.99Fe0.01O8	5.16 × 10–6	5.14 × 10–6	5.59 × 10–6	0.83	0.17
LiV2.95Sn0.05O8	6.68 × 10–6	6.54 × 10–6	6.49 × 10–6	0.61	0.39
LiV2.95Zn0.05O8	4.98 × 10–6	5.17 × 10–6	5.03 × 10–6	0.76	0.24
LiV2.95Co0.05O8	2.98 × 10–6	3.11 × 10–6	2.91 × 10–6	0.74	0.26
LiV2.95Fe0.05O8	4.76 × 10–6	4.91 × 10–6	5.04 × 10–6	0.75	0.25
LiV2.95Ti0.05O8	4.67 × 10–6	4.74 × 10–6	4.83 × 10–6	0.78	0.22
LiV2.95Zr0.05O8	5.27 × 10–6	5.34 × 10–6	5.47 × 10–6	0.79	0.21
LiV2.95Nb0.05O8	4.35 × 10–6	4.44 × 10–6	4.39 × 10–6	0.76	0.24
LiV2.95Mo0.05O8	4.68 × 10–6	4.74 × 10–6	4.85 × 10–6	0.69	0.31

Simple and well-known wagner polarization method was used to find the ionic and electronic transference numbers (ti and te) for all as-prepared compounds.[23,24] The change in current (I) was measured by applying a steady dc potential of 50 mV across the pellets. It was found that the initial conductivity values calculated from the observed current was decreasing with increasing time, before getting saturated. Significant differences were observed in the case of the percentage of ionic and electronic conduction, as calculated from the transference number measurements (Figure b) using the formulawhere “σ0” and “σα” are the initial and saturated conductivity values, respectively. The difference between the initial and saturated conductivity values gives the value of the ionic transference number.

Whereas the compound LVO-P showed 81% of ionic and 19% of electronic conductivity, the compound LVO-WP was found to have 74% of ionic and 26% of electronic conductivity. A very good mixed conductivity was observed in the case of compounds LiV2.99Zr0.01O8 and LiV2.95Sn0.05O8 with ∼60% of ionic and ∼40% of electronic conductivity. For all other doped compounds, LiV2.95M0.05O8 (M = Zn, Co, Fe, Ti, Zr, Nb, and Mo), mixed conductivity was observed, as given in Table .

Electrochemical Analysis

Cyclic Voltammetry (CV)

Cyclic voltammetry is one of the well adopted technique to understand redox potentials and complement the galvanostatic cycling profiles of cathode materials.[33−35]Figure a shows the second cycle CV plots for the compounds LVO-P and LVO-WP. During intercalation, broad peaks were observed at 2.9 and 3.0 V for the compounds LVO-P and LVO-WP, respectively. The shift in the anodic peak toward the lower potential side in the case of LVO-P indicates the better kinetics for lithium deintercalation. Whereas the cathodic peak 2.7 V refers to single-phase reaction of lithium intercalation at tetrahedral sites, the peak at 2.4 V refers to lithium intercalation at octahedral sites resulting in two-phase transformation between Li3V3O8 and Li4V3O8.[36−38] The cathodic peak at 2.3 V expected due to the slow kinetic reaction as seen in many reports was not observed in the present work. This indicates the absence of any V3+ ion dissolution in the electrolyte. The cathodic peak at 2.6 V, usually observed due to the presence of Li0.3V2O5, was not observed for both the compounds.[13] The anodic/cathodic peak around 3.4/3.2 V observed in the case of LVO-P shows a trace of Li0.3V2O5 because of different energy sites for lithium ions.

Figure 6

Comparative second cycle CV plots of cells made with (a) LVO-P and LVO-WP, (b) LVO-P, LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, and LiV2.99Fe0.01O8 (scan rate: 0.058 mV/s).

In Figure b, the comparison of the second cycle CV plots (for compounds selected based on their high specific-discharge capacity and better cycling stability, to be discussed in the galvanostatic charge–discharge section) shows that the anodic peaks were shifted to lower potentials in the order of LVO-P, LiV2.95Sn0.05O8, LiV2.99Fe0.01O8, and LiV2.99Zr0.01O8. The cathodic peaks were found to shift to higher potentials in the same order as mentioned above. This indicates the compound LiV2.99Zr0.01O8 to have better lithium diffusion kinetics. The CV of compound LiV2.99Zr0.01O8 shows an anodic peak around 2.5 V, which indicates easy accessibility of lithium sites during deintercalation.[39] As seen in Figure a, the anodic peak at 2.5 V was suppressed in the case of 0.05 and 0.1 M level of zirconium doping. Irrespective of dopants, a similar trend was observed for all compounds (Figure a–h). This indicates the accessibility of lithium ion sites to be affected by the level of doping. The anodic and cathodic peaks around ∼3.6 V in the case of all doped samples show deintercalation and intercalation of lithium ions, respectively, in the octahedral lithium site of the LiV3O8 host structure.[36−38]

Figure 7

Second cycle CV plots of cells made with (a) LiV3–ZrO8, (b) LiV3–SnO8, (c) LiV3–FeO8, (d) LiV3–ZnO8, (e) LiV3–CoO8, (f) LiV3–TiO8, (g) LiV3–NbO8, and (h) LiV3–MoO8 (x = 0.01/0.05/0.1 M); scan rate: 0.058 mV/s.

In the case of LiV3–ZrO8 compounds, the anodic and cathodic peaks were shifted to lower and higher potential sides, respectively, in the case of 0.01 and 0.05 M levels of doping, when compared to 0.1 M level of doping (Figure a). This indicates that doping more than 0.05 M of zirconium at the vanadium site results in poor lithium diffusion kinetics. The cathodic peaks at ∼3.2 and 3.4 V were observed in the CV plot of LiV2.99Zr0.01O8 (indicated by arrow marks in Figure a), but not seen in the case of other two levels of doping. The above peaks were due to the presence of the Li0.3V2O5 phase, holding different energy sites for lithium ions,[40−42] as discussed in XRD (indicated by * symbol in Figure b). Similar cathodic peaks belonging to the Li0.3V2O5 phase were also observed in the case of other doped samples, depending on the type of doping and doping level (Figure b–h). For LiV3–SnO8 compounds (Figure b) and in the case of other doped samples as in Figure , expected anodic/cathodic peaks were observed, indicating good lithium diffusion at all levels of doping.

Figure 8

(a) Second cycle charge–discharge profiles and (b) cycling stability for LVO-P and LVO-WP.

Charge–Discharge Studies

Figure a shows the second cycle charge–discharge curves of cells made with LVO-P and LVO-WP. During the discharge, a clear two-step intercalation process was observed. In the case of LVO-P, the plateau regions around 2.7 and 2.4 V during the discharge refer to the lithium intercalation at tetrahedral and octahedral sites, respectively, of the LiV3O8 structure and were in accordance with the cathodic peaks observed in CV.[36−39] Cycling stability could be observed in Figure b. The 2nd and 50th cycle discharge capacity values are given in Table . The compound LVO-P was found to deliver the highest discharge capacity of 342 mA h/g at the 2nd cycle with a capacity retention of 71% at the end of the 50th cycle, whereas the compound LVO-WP delivered a comparatively low discharge capacity of 206 mA h/g at the 2nd cycle and showed a capacity retention of 82% at the end of the 50th cycle.

Table 3

Discharge Capacity of All Compounds

		discharge capacity (mA h/g) (±3 mA h/g)
doping level	compound	2nd cycle	50th cycle	capacity retention at 50th cycle	irreversible capacity loss
	LVO-P	342	254	71	41
	LVO-WP	206	169	82	12
0.01 M	LiV2.99Zn0.01O8	343	234	60	54
	LiV2.99Co0.01O8	303	238	78	47
	LiV2.99Fe0.01O8	398	236	59	78
	LiV2.99Sn0.01O8	289	251	86	10
	LiV2.99Ti0.01O8	325	196	60	51
	LiV2.99Zr0.01O8	250	247	98	25
	LiV2.99Nb0.01O8	340	234	68	18
	LiV2.99Mo0.01O8	292	205	70	18
0.05 M	LiV2.95Zn0.05O8	239	225	94	5
	LiV2.95Co0.05O8	154	150	97	8
	LiV2.95Fe0.05O8	211	206	97	7
	LiV2.95Sn0.05O8	245	241	98	9
	LiV2.95Ti0.05O8	225	211	93	15
	LiV2.95Zr0.05O8	242	190	75	23
	LiV2.95Nb0.05O8	192	164	85	3
	LiV2.95Mo0.05O8	202	187	92	4
0.1 M	LiV2.9Zn0.1O8	201	185	91	3
	LiV2.9Co0.1O8	113	91	81	12
	LiV2.9Fe0.1O8	196	160	82	6
	LiV2.9Sn0.1O8	210	174	82	10
	LiV2.9Ti0.1O8	172	162	94	25
	LiV2.9Zr0.1O8	189	179	94	5
	LiV2.9Nb0.1O8	192	120	62	2
	LiV2.9Mo0.1O8	203	135	66	3

The second cycle charge–discharge curves for 0.01, 0.05, and 0.1 M level doped LiV3O8 compounds are shown in Figure a–c. The second cycle discharge capacity values are given in Table . For the doped compounds, irrespective of the dopants, a high capacity was observed in the case of 0.01 M level of doping. This was in accordance with the CV plots, showing broader peaks for 0.01 M level of doping, as discussed above. The highest discharge capacity of 398 mA h/g was observed in the case of the LiFe0.01V2.99O8 compound at the end of the second cycle, which was the highest discharge capacity among all compounds. When compared to LVO-P, a close value of the second cycle discharge capacity was observed for the compounds LiV2.99Zn0.01O8, and LiV2.99Nb0.01O8. Among doped compounds, very low second cycle discharge capacity was observed for compounds LiV2.95Co0.05O8 and LiV2.9Co0.1O8.

Figure 9

Second cycle charge–discharge profiles of (a) LVO-P and LiV2.99M0.01O8, (b) LVO-P and LiV2.95M0.05O8, and (c) LVO-P and LiV2.9M0.1O8 (M = Zn/Co/Fe/Sn/Ti/Zr/Nb/Mo) compounds.

Cycling Stability of Doped Compounds

As seen in Figure a–h, better cycling stability was observed for LiV3–MO8 (M = Zn, Co, Fe, Sn, Ti, Zr, Nb, and Mo) compounds with x = 0.05 M level of doping. Only in the case of zirconium, 0.01 M level of doping was found to be better than 0.05 and 0.1 M doped compounds, in terms of cycling stability. Among the zirconium-doped compounds, LiV2.99Zr0.01O8 delivers a discharge capacity of 250 and 247 mA h/g at the end of the 2nd and 50th cycles, respectively, with an excellent capacity retention of 98%, whereas further doping of zirconium degraded the electrochemical performance. Ren et al.[9] reported on LiV3–ZrO8 (x = 0.00, 0.02, 0.04, 0.06, and 0.08) prepared by the citrate sol–gel method, and concluded the x = 0.06 M level of doping to be the best. They reported a discharge capacity of 269 and 246 mA h/g at the end of the 2nd and 50th cycles, respectively, for 0.1 C rate at the voltage window of 1.8–4.0 V. This indicates a stability of 92%. The present work shows the best capacity and stability for 0.1 C rate at the 0.01 M level of doping and is degraded for higher level of doping. This indicates that the preparation method chosen plays a key role in determining the optimized doping level to get the best electrochemical performance of the cells. Among the tin-doped compounds, LiV2.95Sn0.05O8 delivers a discharge capacity of 245 and 241 mA h/g at the end of the 2nd and 50th cycles, respectively, with an excellent capacity retention of 98%. For LiV2.99Sn0.01O8, although the cycling stability was only 86%, still it shows a better discharge capacity of 251 mA h/g at the 50th cycle. However, the performance was degraded in the case of high level of doping LiV2.9Sn0.1O8.. Therefore, the compounds, LiV2.99Zr0.01O8, and LiV2.95Sn0.05O8, could be concluded as the best in terms of capacity and stability among the doped compounds. The comparative cycling stabilities of the compounds LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, LVO-P, and LiV2.99 Fe0.01O8 are shown in Figure . Song et al.,[39] reported on the 0.15 M of molybdenum-doped LiV3O8 nanosheets with a surface area of 24.8 m2/g to deliver a capacity of 217 and 206 mA h/g at the initial and 100th cycle, respectively, for a current density of 300 mA/g. The electrical conductivities of the bare and 0.15 M of molybdenum-doped LiV3O8 nanosheets were reported to be 3.52 × 10–6 and 2.89 × 10–5 S/cm, respectively. Nevertheless in our case, molybdenum doping was not found to improve the performance of LiV3O8, at any of the doping level.

Figure 10

Cycling stability of cells made with (a) LiV3–ZrO8, (b) LiV3–SnO8, (c) LiV3–FeO8, (d) LiV3–ZnO8 (e) LiV3–CoO8 (f) LiV3–TiO8, (g) LiV3–NbO8, and (h) LiV3–MoO8 (x = 0.01/0.05/0.1 M).

Figure 11

Cycling stability for the selected compounds LVO-P, LiV2.99Zr0.01O8, LiV2.95Sn0.05O8, and LiV2.99 Fe0.01O8.

Overall, as discussed under electrical studies, though the electrical conductivity of doped compounds were in the same order as that of LVO-P, the percentage of ionic and electronic conductivity values observed from transference number studies highly affected the electrochemical performance. The higher the percentage of ionic conductivity, the higher was the discharge capacity, as in the case of compound LiV2.99Fe0.01O8. However, its lower percentage of electronic conductivity resulted in a poor cycling stability. The compounds, LiV2.99Zr0.01O8 and LiV2.95Sn0.05O8, with a good mixed conduction nature showed good cycling stability, as discussed above.

Conclusions

LiV3O8 was prepared via the most simple and cost-effective reflux method. The presence of polymer P123 in the preparation was found to affect the percentage of ionic and electronic conduction, which in turn influences the capacity and cycling stability of the cells, respectively. The effect of different levels of doping with different elements on the structure, electrical, and electrochemical performance of the compound was investigated. Among all doped compounds, LiV2.99Zr0.01O8 and LiV2.95Sn0.05O8 was found to deliver the best electrochemical performance both in terms of higher discharge capacity (∼250 mA h/g) and cycling stability (∼98%). The above compounds exhibited a mixed conducting nature with ∼60% of ionic and ∼40% of electronic conductivity. Though the compound LiV2.99Fe0.01O8 was found to exhibit the best discharge capacity of 398 mA h/g for a current density of 30 mA/g because of its high percentage of ionic conductivity of ∼83%, its cycling stability was found to be degraded because of its poor percentage of electronic conductivity of ∼17%. We conclude that, the type and level of doping will successfully enhance the electrochemical properties of the parent compound only when it exhibits a mixed conducting nature.

Experimental Section

Pristine and doped lithium trivanadate compounds were prepared by the Pluronic P123-assisted reflux method. LVO-P: Initially 6 mM polymer (Pluronic P123, Aldrich) was dissolved in 100 mL of distilled water, and a homogenous viscous solution was obtained. To this, 0.1 M lithium acetate and 0.3 M ammonium metavanadate were added and refluxed for 2 h in a constant temperature oil bath at a temperature of 120 °C. The resultant solution was evaporated on a hot plate, and the dry residue was collected. The fine powder was then calcinated at 550 °C for about 10 h in air at a heating rate of 3 °C/min. A similar procedure was followed for all doped compounds. The raw materials for different doped LiV3–MO8 (M = Zn, Co, Fe, Sn, Ti, Zr, Nb, and Mo) compounds were zinc acetate, cobalt acetate, iron citrate, tin chloride, titanium oxysulfate, zirconyl nitrate, niobium pentoxide, and molybdenum oxide, respectively. Doping was made at three different levels of x = 0.01/0.05/0.1 M. LVO-WP: For comparison, LiV3O8 was also prepared by refluxing the required amount of lithium acetate and ammonium metavanadate in the absence of the polymer P123, and the fine powder was calcinated in the same way as done for LVO-P.

Material Characterization

The structure of the pristine and doped compounds were examined by an X-ray diffractometer (Bruker D8) equipped with Cu Kα-radiation, and the lattice parameter values were obtained by TOPAS software version 4.2. The morphology of the compounds was analyzed by a scanning electron microscope (JEOL JSM-6700F). Raman spectrometer (model LabRAM HR Evolution, HORIBA Scientific) with a wavelength of 514 nm, and a power of 100 mW was used for the Raman studies.

For electrical studies, pellets with a diameter of 1 cm at a pressure of ten tons were made using hydraulic pellet press, and the pressed pellets were sandwiched between aluminum blocking electrodes and subjected to electrical studies. The ac impedance study was performed using a computer controlled CH Instrument (CHI-600E) with a fixed amplitude voltage of 10 mV. Four-probe measurements made by SES Instruments (India) Pvt. Ltd were used to find the dc conductivity. The percentages of ionic and electronic conductivities of the compounds were found by transference number measurements carried out using a Keithley 4001 source meter.

For electrochemical testing, slurries were prepared by mixing the as-prepared active material (LVO), Super P carbon black, and poly(vinylidene fluoride) (Kynar 2801) binder (70:15:15 in wt %) dissolved in N-methyl pyrrolidone under stirring for 24 h. The electrodes were made by coating the slurry onto the Al foil. The composite electrodes were kept in a vacuum oven for 24 h at 70 °C. The CR2016 coin-type cells were assembled with the above cathodes, lithium metal as the anode (Kyokuto Metal Co., Japan), and Celgard 2502 film as the separator. A solution of 1 M LiPF6 in ethylene carbonate and dimethyl carbonate (1:1, v/v) (Merck) was used as the electrolyte. The coin cells were assembled in an Ar-filled glovebox at an O2 and H2O level of less than 1 ppm. The cyclic voltammogram was made in the voltage range of 2–4 V (at a scan rate of 0.058 mV/S) using a MacPile II (BioLogic, France) system. Galvanostatic cycling was carried out at room temperature (25 °C) in the voltage range of 2–4 V at a current rate of 0.1 C by using Bitrode multiple battery testers (model SCN, Bitrode, USA).