V2O5-anchored Carbon Nanotubes for enhanced electrochemical energy storage M. Sathiyaa, A. S. Prakasha,*, K. Rameshaa, J-M. Tarasconb and A. K. Shuklac CSIR Central Electrochemical Research Institute-Chennai Unit, CSIR-Madras Complex, Taramani, Chennai-600 113, India. b Laboratoire de Reactivite et Chimie des Solides, CNRS UMR 6007, 33, rue Saint Leu – Universite de Picardie Jules Verne, 80039 Amiens, France c a Solid State and Structural Chemistry Unit, Indian Institute of Science, Bangalore-560 012, India. 1. Supporting information S-1 Figure S-1, Structure of (a) crystalline V2O5 and (b) V2O5 xerogels.

In crystalline V2O5, single layers of V2O5 are arranged in orderly manner whereas in V2O5 xerogel, bilayers of single V2O5 layers are arranged as stacks along the c-axis of monoclinic unit cell. Oxygen coordination of vanadium resembles a square pyramid S1 in both structures. Oxygen atoms shown between the layers represent oxygen of water molecules. 2. Supporting information S-2 1. 81 A (062)-V2O5 3. 4 A CNT- (002) (062)-V2O5 1. 92 A (-114) V2O5 Figure S-2, High resolution Transmission electron micrograph showing growth of (-114) plane of V2O5 parallel to (002) planes of CNT lattice.

The (062) planes of V2O5 which are perpendicular to (-114) planes are also shown. S2 3. Supporting information S-3 Figure S-3, Cyclic voltammogram of crystalline V2O5 (sigma Aldrich) at a scan rate of 0. 1mV/ sec. Cyclic voltammogram of crystalline V2O5 reveal four reduction peaks at ~3. 25, 3. 05, 2. 2 and 1. 5V in the first cathodic sweep. These peaks are attributed to phase transition of ? – V2O5 to ? , ? , ? and ? phase which is in good agreement with previous reports1. Formation of ? phase is irreversible and ? phase is cycling reversibly from second cycle onwards. S3 4. Supporting information S-4 Figure S-4.

Plot showing linear relationship of log? vs log i for cathodic (discharge) and anodic (charge) sweeps of cyclic voltammogram. According to Cottrell equation, i= nFAC0jvD0/ v(? t). 2 It can be simplified as i= at-1/2 where ‘a’ is the collection of constants. Further, (Scan rate)1/2 can be used in place of t-1/2. Hence, current response for the voltammetric sweep follows the power law relationship, i= a? b, where a and b are adjustable parameters. Thus log i= log a+ b log ? , value of b can be calculated from the slope of straight obtained by plotting log i vs. log ?. Further, b= 0. 5 for diffusion limited processes (i= a? /2) and is unity for non diffusion limited processes (i= a? ). Thus CV experiments were carried out at different scan rates of 0. 1 to 5 mV/ sec and current S4 values at different potentials were plotted as a function of scan rate. From the slope of the straight line obtained, b- value is calculated during cathodic and anodic sweeps. 5. Supporting information S-5 Figure S-5. Dependence of slope ‘b’ (derived from linear fit of log i vs log ? ) as a function of cell voltage. As explained in previous section (supporting information Fig. S4), value of b is calculated and is plotted as a function of voltage V.

Slope b is comparatively lower at peak potentials indicating the dominance of diffusion limited intercalation. Whereas it is near to 1 at other potentials indicating more of capacitive contribution. S5 6. Supporting information S-6 Figure S-6. The plots of ? 1/2 vs i/? 1/2 used for calculating constants a1 and a2 at different potentials. According to power law relationship, i= a? for non diffusion limited processes and i= a? 1/2 for diffusion limited processes. Thus, total current i= a? + a? 1/2 and i(V)/ ? 1/2 = a1? 1/2 + a2. Current values at different potentials were calculated from cyclic voltammogram at different scan rates of 0. to 5mV/ sec. Plots of i/? 1/2 vs. ?1/2 have been drawn at different potentials and from the straight line obtained value of a1 (slope) and a2 (intercept) are calculated. S6 7. Supporting information S-7 (a) (b) Figure S-7. Dependence of stored charge vs sweep rate. The Trasatti procedure involves analysis of dependence of voltammetric charge as a function of sweep rate. This allows discriminating charge storage due to easily accessible surface S7 (outer) and not easily accessible surface or inner surface. The voltammetric charge (q*) is a measure of total charge exchanged between the electrode and the electrolyte.

This includes the amount of Li+ diffusing to the inner surface, cracks, pores, grain boundaries etc. As the sweep rate increases, the diffusion of Li+ in to the inner surface becomes more difficult. As the sweep rate, ? > ? , all the subsurface regions are excluded and q* tends to q *outer where q *outer is amount of charge stored at outer and easily accessible active surface. On the other hand, as the sweep rate ? > 0, the access to subsurface region is fully allowed and q* tends to qt* which is the total charge stored due to whole of active surface including inner and outer surface.

The difference between q*total and q *outer gives q*inner. To calculate the charge storage at ? >0 and ? >? , experiments have been carried out at various scan rates of 0. 05,0. 1……….. 50mV/ sec with Ohmic drop compensation. The q*total (qt ) derived from the intercept of straight line 1/ q* vs. ?1/2 is 2987C/g (Fig. S-7(a)). Similarly, q *outer derived from the intercept of plots q* vs. ?-1/2 (Fig. S-7 (b)) is 1720C/g. * 8. Supporting information S-8 0. 12 0. 10 slope ‘b’ 1. 0 0. 8 0. 6 0. 4 1. 0 0. 8 0. 6 0. 4 Discharge Current/ mA. mg -1 0. 08 0. 06 0. 04 0. 2 0. 00 -0. 02 -0. 04 -0. 06 Charge 1. 5 2. 0 2. 5 3. 0 + 0. 2 3. 5 4. 0 Voltage/ V Vs. Li / Li -0. 08 1. 0 1. 5 2. 0 2. 5 3. 0 + 3. 5 4. 0 Voltage/ V Vs. Li / Li Figure S-8. Potentiostatic cycling to understand the charge storage mechanism in V2O5 xerogel electrodes. Cyclic voltammogram of V2O5 xerogels ( scan rate= 0. 1mV/ sec) prepared by the same way but without the addition of CNTs (a) Inset is the plot of slope b as S8 a function of voltage V. Slope b is calculated by plotting log i vs. log v at various scan rates as described for V2O5- CNT composites. b) Capacitive and diffusion controlled charge storage separation with cyclic voltammogram of 0. 1mV/sec scan. The shaded portion of the CV corresponds to capacitive contribution of about 15%. Separation of diffusion controlled and non diffusion controlled processes have been carried out as explained in supporting information S-4 and S-6. Bulk V2O5 xerogel materials showed capacity of 200mAh/ g out of which only 15% is from non diffusion controlled processes. It showed comparatively lesser intercalation capacity and double layer and redox pseudocapacitance than V2O5 anchored CNT.

This shows the unique advantage of anchoring thin films of V2O5 on CNT surfaces. It provides huge surface area, proper pore distribution for electrolyte percolation, etc which help in improving the double layer capacitance as well as intercalation. 9. Supporting information S-9 Figure S-9. X-ray photoelectron spectrum of V2O5/CNT electrode after electrochemical reduction to 1. 5 V. S9 Survey spectrum shows peaks of Li-1S along with V, O and C originating from the sample. P and F peaks are coming from electrolyte and SEI component. 10. Supporting information S-10 Figure S-10.

X-ray photoelectron spectrum of Li 1S region of lithium intercalated V2O5/CNT electrode at 1. 5V. Li 1S spectra appeared around 56. 1eV. S10 11. Supporting information S-11 Table S-1. Charge-storage contributions as function of sweep rate. On increasing the voltammetric sweep rate the time for lithium ion diffusion into the host lattice decreases and hence capacity due to intercalation decreases. Whereas change in charge storage due to surface properties is comparatively lesser making the redox pseudocapacitive behavior to dominate over intercalation capacity at higher rates and vice versa.

Total stored charge Scan rate (cathodic sweep) in C/g 0. 05 2893 Contribution from double layer+ redox pseudo capacitance 64. 65% Lithium ion intercalation capacity (mA. h/g) 284. 4 (1. 93 lithium intercalation) 255 (1. 7 lithium intercalation) 132. 6 (0. 9 lithium intercalation) 66. 3 (0. 45 lithium intercalation) 0. 1 2780 67% 1 2410 80 2 2150 89 S11 12. Supporting information S12 Table S-2. XPS peak fit profile for the as prepared V2O5 anchored CNT samples and after discharging the sample up to 1. 5V.

Peak fit is carried out such that separation between V2p3/2 and V2p1/2 is 7. 3V and the areas of the peaks are in the ratio of 2:1. Separation (eV) SI. No Element Position (eV) Area FWHM As prepared V2O5- CNT 1. 2. V2p3/2 V2p1/2 517. 19(4) 524. 49(4) 7. 3 7. 3 45100. 7 22550. 36 1. 78(5) 3. 65(4) After Li-intercalation up to 1. 5V 3. 4. 5. 6. 7. V2p3/2 V2p1/2 V2p3/2 V2p1/2 V2p1 516. 25(1) 523. 55(2) 517. 33(4) 524. 63(4) 520. 95(1) 7. 3 7. 3 7. 3 7. 3 67121. 98 33650. 99 13191. 81 6595. 91 7390. 93 2. 20(9) 2. 74(8) 1. 25(1) 2. 6(0) 2(0) S12