Short communicationUnderstanding the characteristics of high-voltage additives in Li-ion batteries: Solvent effects
Introduction
Overcharge events are an urgent safety issue in the lithium-ion battery (LIB) industry. Presently, electrolyte additives are widely employed for the overcharge protection of commercial-grade LIBs. Among them, a redox shuttle (RS) bypasses the surplus current via a repetitive redox reaction between the cathode and the anode [1], [2], [3], [4], [5]. The RS compounds have their own redox potentials, selected to prevent self-discharge or shuttling during charging, which should be 0.3–0.4 V higher than the end-of-charge potential of the cathode. Aromatic compounds are another class of electrolyte additives used for overcharge protection. Polymerizable monomers such as cyclohexyl benzene (CHB) and biphenyl (BP) are electrochemically oxidized at the overcharged cathode to form a passive polymeric film at the cathode|electrolyte interface [6], [7], [8], [9], [10], [11], [12]. In addition, non-polymerizable aromatic compounds such as tert-butyl benzene act as redox mediators, which facilitate oxidative decomposition of electrolyte solvents that leads to gas evolution [6].
Regardless of the mechanism of overcharge protection, the electrolyte additives should remain inert during the normal operation of LIBs and become active only at overcharge events; this requires the additives to have an oxidation potential that is 0.3–0.4 V higher than the cathode potential. For instance, LiCoO2 reaches the fully charged state near 4.2 V (vs. Li/Li+) and therefore requires additives with an oxidation potential higher than 4.5 V. Thus, determination of the oxidation potential is a crucial step in evaluating novel electrolyte additives.
By means of density functional theory calculations, the present work has obtained the oxidation potentials of 108 organic molecules, which are possible candidates for LIB additives. Calculation accuracy is assessed through comparison with experimental data. Understanding the relationship between the oxidation potential and the molecular characteristics is considered to be very useful for developing novel electrolyte additives with specified oxidation potentials.
Density functional theory (DFT) has become a popular method for calculating oxidation potentials for a vast array of organic molecules used in LIBs [13], [14], [15], [16]. The ground-state structures of the molecules have been fully optimized within C1 symmetry by means of DFT methods. Spin-restricted and unrestricted schemes have been employed for even- and odd-numbered electron systems, respectively. All DFT geometry optimizations were obtained using DMol3 numerical-based density-functional computer software [17], [18] implemented in the Materials Studio Modeling 3.2 package from Accelrys, Inc. The Kohn–Sham equation was calculated with the gradient-corrected BPW91 functional: the 1988 exchange functional of Becke [19] with the correlation functional of Perdew and Wang [20]. A numerical basis set of double-zeta plus polarization (DNP) quality was employed. Utilizing the geometries obtained at the BPW91/DNP level, single-point energy calculations were performed at the B3PW91/6-31 + G(d,p) level of theory. The functional includes a three-parameter adiabatic connection exchange term [21], i.e., a linear combination of exact Hartree–Fock exchange, Slater exchange [22], and B88 gradient-corrected exchange [19].
This study employs the conductor-variant polarized continuum model (CPCM) [23], which places the solute in a molecular-shaped cavity imbedded in a continuum dielectric medium. In the CPCM models, the variation of the free energy when going from vacuum to solution is composed of the work required to build a cavity in the solvent (cavitation energy) together with the electrostatic (solute–solvent interaction and solute polarization) and non-electrostatic work (dispersion and repulsion energy). A dielectric constant of 31.9 is adopted and is a weighted average value between the dielectric constants of ethylene carbonate (EC: 89.2) and ethyl methyl carbonate (EMC: 2.9), because EC/EMC = 1/2 solution is used as the solvent in cyclic voltammetry experiments. The solvation energies obtained from the CPCM calculations are free energies. In the self-energy of the solute, the entropy and zero-point contributions are omitted due to molecular modes, assuming that they would make only a relatively minor contribution to the desired oxidation potentials [24], [25]. All of the single-point DFT and CPCM calculations are carried out by using the program package Gaussian03 [26].
Section snippets
Experimental details
Battery-grade 1 M LiPF6 in EC/EMC (1/2, v/v) was chosen as the base electrolyte solution. Using the Karl–Fisher titration method, the water content in the electrolytes was determined to be less than 20 ppm. The free-acid content in the electrolytes was less than 50 ppm.
The oxidation potentials of the benzene derivatives were determined by linear sweep voltammetry (LSV). A Pt disc (area = 0.02 cm2) was used as the working electrode. The reference electrode was Li foil and the counter electrode was Pt
Results and discussion
The observed oxidation potential (Eox) values, together with the calculated ionization potential (IP) and Eox values, for 108 molecules are listed in Table 1S (see supplementary data). All potentials are reported vs. Li/Li+. The calculated oxidation potentials are in good agreement with the experimental data, where the root-mean-square (RMS) deviation is 0.08 V and the maximum deviation is 0.15 V. The correlation coefficients (R2) of IP and Eox(calc.) with respect to Eox(exp.) are 0.84 and 0.98,
Conclusions
We have calculated their IP and Eox values for comparison purposes with experimental values. The calculated potentials are in close agreement with the experimental values, where the RMS deviation is 0.08 V and the maximum deviation is 0.15 V. Molecules exhibiting high Eox (>4.5 V) show one of the following two features: (1) IP > 7.70 eV or (2) IP < 7.70 eV with a relatively large molecule size. Consideration of bulk solvent effects is important to describe fully the experimental variation in Eox; in
References (30)
J. Power Sources
(2006)- et al.
Electrochem. Commun.
(2007) - et al.
Electrochem. Commun.
(2007) - et al.
Electrochim. Acta
(2004) - et al.
J. Power Sources
(2006) - et al.
J. Power Sources
(2007) - et al.
Chem. Phys. Lett.
(2003) - et al.
J. Electrochem. Soc.
(2005) - et al.
J. Electrochem. Soc.
(2005) - et al.
Electrochemistry
(2003)
J. Appl. Electrochem.
J. Appl. Electrochem.
Electrochem. Solid-State Lett.
J. Phys. Chem.
J. Electrochem. Soc.
Cited by (56)
Cathode solid electrolyte interface's function originated from salt type additives in lithium ion batteries
2016, Electrochimica ActaCitation Excerpt :The same hydrolysis of LiDFOB is speculated from Fig. 4. Therefore, the SEI consists of the trivalent boron and OG as coincided with peaks observed by Xu and Han et al. [14,21]. However, there was no information of the concrete chemical structure of SEI component from our observations, since the detected information could not be bridged in reasonable way.
Distinct Reaction Characteristics of Electrolyte Additives for High-Voltage Lithium-Ion Batteries: Tris(trimethylsilyl) Phosphite, Borate, and Phosphate
2016, Electrochimica ActaCitation Excerpt :In addition, TMSPa can be applied in various kinds of cathode materials, such as LiNi0.5Mn1.5O4 and LiNi0.4Co0.2Mn0.4O2, because CEI films effectively suppress the Mn and Ni dissolution from the cathode surface [31,32]. The importance of the development of high-voltage electrolyte additives has been highlighted in recent publications [16–39]. Note that most of the studies, however, have intensively focused on investigating the effectiveness of CEI films and HF-scavenging performances, as they are highly responsible for ensuring a high level of interfacial stability for various kinds of cathode materials.
Computational screening of organic molecules as redox active species in redox flow batteries
2016, Current Applied PhysicsCitation Excerpt :These criteria were devised by calculating the oxidation potential, reduction potential, and solvation energy of quinoxaline derivatives. We selected 106 organic molecules [28] as candidates for redox active materials, calculated their oxidation potential, reduction potential, and solvation energy, and, based on these results, identified appropriate candidates that satisfy the criteria of a wide redox window and high solubility. Furthermore, we will present two examples of a new material design that takes account of the characteristics specified in the screening criteria.