Electrolyte contact changes nano-Li4Ti5O12 bulk properties via surface polarons

It is of general interest to combine the faradaic processes based high energy density of a battery with the non-faradaic processes based high power density of a capacitor in one cell. Surface area and functional groups of electrode materials strongly affect these properties. For the anode material Li4Ti5O12 (LTO), we suggest a polaron based mechanism that influences Li ion uptake and mobility. Here we show electrolytes containing a lithium salt induce an observable change in the bulk NMR relaxation properties of LTO nano particles. The longitudinal 7Li NMR relaxation time of bulk LTO can change by almost an order of magnitude and, therefore, reacts very sensitively to the cation and its concentration in the surrounding electrolyte. The reversible effect is largely independent of the used anions and of potential anion decomposition products. It is concluded that lithium salt containing electrolytes increase the mobility of surface polarons. These polarons and additional lithium cations from the electrolyte can now diffuse through the bulk, induce the observed enhanced relaxation rate and enable the non-faradaic process. This picture of a Li+ ion equilibrium between electrolyte and solid may help with improving the charging properties of electrode materials.


Supplementary Note 2: Sodium bis(trifluoromethanesulfonyl)imide (NaTFSI)
In case of the sodium salt, a small shift of 0.09ppm (49Hz) is observed when comparing the stock solution with the LTO-stock solution mixture ( Supplementary Fig. 2). The T1 relaxation time constant shows only a minor difference between both samples, suggesting a surface interaction independent mobility. Even a further measurement of the frozen sample at -10°C did not reveal an additional peak for a surface adsorbed species (Supplementary Fig. 3). Further analysis of the 19 F T1 relaxation data by Inverse Laplace Transformation supports the result of hardly adsorbed species by a main single relaxation time (Supplementary Figs. 4,5). Figure 2: 19 F NMR spectrum of NaTFSI in DMC (stock solution) measured at 12°C (blue), and LTO plus 2µmol NaTFSI in 1.5mL DMC (red). In case of the lithium salt, a 19 F NMR shift of 0.18ppm (103Hz) is observed and additionally a second, broader component with a shift of 0.66ppm (374Hz) is clearly visible ( Supplementary Fig. 6). These signals are temperature dependent. The signals start to merge with increasing temperature, proving exchange between two reservoirssurface-interacting anions and anions of the inter-LTO-particle solution. The broader component has a shorter T1 value, suggesting a stronger interaction with the surface, which supports the charge compensation on the solid (Supplementary Fig. 8) This proves the picture of a double layer that is formed when using lithium salts, and not for sodium salts. This is a further demonstration of the selective intercalation of lithium cations into the material.

Supplementary Note 4: Estimation of polaron concentration
The concentration of polarons was estimated by quantitative NMR of the adsorbed anions.
The amount (n) of 4µmol LiTFSI was dissolved in 1.5mL DMC (Volume A) and was filled in a Young type NMR tube. This volume corresponds to a height (HA) of 10.4cm in the Young type NMR tube. A quantitative 19 F NMR spectrum was recorded. 100mg LTO was added, shaken and waited until the powder was settled (24h). The volume increased to a height of 10.6cm in the NMR tube. The settled LTO (Volume B) had a height (HB) of 3.9cm in the tube and the supernatant solvent (Volume C) had a height (HC) of 6.7cm. The integrals IA, IB, IC of the quantitative 19 F NMR spectra of volumens A, B, and C were determined, with IA, IB, IC proportional to the concentration of fluorine nuclei in the detection volume. Volumes B and C could be distinguished by shifting the position of the NMR tube such that the volume of interest resided within the coil used for 19 F NMR detection. The coil had a hight on the order of 1 cm, allowing for a differentiation of the two volumes.
The amount X of anions (in an artificial unit of cm) in a volume X can be correlated to the signal intensity IX and height Hx of that volume in the NMR tube, The amount of anions in A should be the same as in B+C combined, since no additional fluorine was provided when LTO was added. We expect

Supplementary Note 5: Arrhenius plot of relaxation data
Arrhenius plot representation of the 7 Li T1 relaxation data ( Figure S9). Although the temperature range for the measurements of the two-phase system was limited, some important conclusions can be drawn. The T1 measurements of a partially lithiated LTO material (such as Li4.1Ti5O12) can be interpreted in terms of the Curie-Weiss relaxation as paramagnetically induced relaxation caused by mobile polarons in the bulk material. In case of an unlithiated sample, polarons are barely present in the bulk, and dipol-dipol or quadrupolar coupling induced relaxation are expected to be the main relaxation pathways. The decreasing slope from 0M to 1M to 2M is, therefore, best interpreted as transition from a dipoldipol/quadrupolar relaxation dominated process to a process with paramagnetic as well as dipoldipol/quadrupolar contributions. A credible calculation of an activation barrier is therefore not possible, but the qualitative statement that additional negative charges in the bulk cause the observed changes in relaxation rate is supported.
Supplementary Figure 9: Arrhenius plot representation of the 7 Li T1 relaxation data (mean of distribution A sample of 10mg pure and dry LTO powder (called LTO_pure) is compared to a sample of LTO+Li(TFSI). Electrochemical impedance spectra (EIS) were performed in a Swagelok cell between two steel stamps on the potentiostat VSP-300 from Bio-Logic. A frequency range from 1MHz to 0.5Hz at a temperature of 22°C was used. The EIS measurement shows a tendency of faster components in the case of LTO+Li(TFSI). This is part of ongoing work.