59th Annual Report on Research 2014

The major goal of this proposal was to improve the safety and lifetime of Li-ion batteries, specifically protecting batteries from overcharge, simply defined as the condition of a battery being charged past the end-of-charge potential of the cathode material. Overcharge can be combated using additives called redox shuttles, which act as an internal shunt to prevent electrolyte oxidation. A major obstacle in using redox shuttles for overcharge protection in commercial batteries relates to the number of cycles of overcharge protection that most redox shuttles can offer. In this project, we sought to increase the stability of radical cations of redox shuttles in lithium-ion battery electrolyte in order for these additives to be utilized in long-term cycling for commercial applications to prevent batteries from overcharging, a dangerous condition that can lead to capacity fade, build-up of internal pressure, and elevated temperatures.

Our initial studies focused on testing the stability of radical cations of some already reported redox shuttles to determine if the radical cation lifetimes in aprotic organic solvents correlated with the extend of overcharge protection observed in lithium-ion batteries. The lifetimes were measured using UV-vis and EPR spectroscopy. For a series of related compounds, specifically fused aromatic rings with heteroatoms incorporated into the central ring positions, we found that compounds whose radical cations showed negligible degradation over at least 5 hours showed extensive (> 100 cycles) overcharge protection. Those that showed more rapid decay survived only a few cycles of overcharge protection. In one case, changes in spectral shape indicate a reaction has occurred, and in this case, the compound may have formed a new product capable of shuttling current and thus allowing effective overcharge.

We then synthesized and studied derivatives of the more stable phenothiazine class of compounds, creating versions with different cores including phenoxazine, diphenylamine, and fluorene. The N-position was functionalized with ethyl groups to prevent dimerization, forming new N-N bonds. Also the positions para to the N atom were modified with methyl, fluoro, chloro, bromo, cyano, and/or trifluoromethyl groups in hopes of preventing dimerization, which has been shown to occur in phenothiazine and triarylamines. Of the cores we explored, we found phenothiazine to be the most stable, and out of the substituents, methyl, chloro, and trifluomethyl resulted in the least decomposition and most extended overcharge protection. We also studied a series of N-substituted phenothiazine derivatives in which the N atom was functionalized with a variety of alkyl groups, carbonyls, or aryl groups and found that primary, secondary, and aromatic substituents led to greatest stability and most extended overcharge protection. Interestingly the most and least-substituted groups – methyl and tert-butyl – were the least stable of the series.

We have also studied the decomposition products of the formation of radical cations to determine why certain cores or substituents are more or less favorable. Radical cations were generated by reaction with chemical oxidants or in bulk electrolysis experiments, and the byproducts were analyzed by a variety of spectroscopic techniques. We found that for substituents on the N-position of phenothiazine, methyl, tert-butyl, and acyl derivatives fragmented, forming the unsubstituted phenothiazine radical cation. We propose that the methyl group is easily attacked in an S_N2 reaction, and DFT calculations of free energies of reaction for the series support this pathway as being lowest for this derivative. DFT calculations are also supportive of an S_N1 decomposition pathway for the tert-butyl-substituted derivative, forming the tert-butyl cation and the radical of phenothiazine. In all cases, we have observed the formation of phenothiazine and dimers and higher oligomers of these derivatives. Because ethyl, iso-propyl, and aryl groups are less prone to substitution reactions, these compounds survive longer in their oxidized states.

In regard to substitutents at the para position to the N atom, we have made some progress in identifying radical cation byproducts. Bromine substituents lead to particularly unstable radical cations, which is not surprising given that the C-Br bond is the weakest of the series. The radical cations of these derivatives are not particularly unstable, though, and we believe that brominated phenothiazines decomposed in lithium-ion batteries by forming radical anions at the anode/electrolyte interface. We found that this compound is reduced by lithium naphthalide, which is similar in reactivity to a lithiated graphitic anode, and decomposes rapidly in its radical anion form. We also analyzed the contents of lithium-ion batteries post-cycling and found that phenothiazines without bromine atoms remaining. While this is not the most easily reduced compound of the series that we have been studying, it is the least stable in its radical anion form.

In summary, we have found that UV-vis can be used to screen potential redox shuttles in that unstable radical cations do not effectively protect batteries from overcharge. We have identified substituents that lead to extended overcharge protection and have identified decomposition pathways by analyzing the byproducts of radical cation formation and/or by analyzing the components of lithium-ion batteries post-cycling.