When you think about what makes a lithium-ion battery work well, the conversation usually lands on cathode materials or anode designs. But buried in the chemistry, less visible but no less important, is the electrolyte. More specifically, how the solvent molecules that make up that electrolyte behave at the molecular level.
A study from Yamaguchi University digs into exactly that question by comparing two sulfite-based solvents with the same chemical composition but fundamentally different shapes: dimethyl sulfite (DMS), a flexible linear molecule, and ethylene sulfite (ES), a rigid cyclic one. What they found challenges some assumptions about how lithium ions interact with their surrounding environment.
The solvent shape problem
Most conventional lithium-ion battery electrolytes rely on carbonate solvents like ethylene carbonate and dimethyl carbonate. These work, but they come with tradeoffs. Ethylene carbonate has a high melting point and high viscosity, which limits low-temperature performance. Dimethyl carbonate flows more easily but doesn’t form stable solid-electrolyte interphase (SEI) layers on graphite anodes.
Sulfite solvents have emerged as alternatives because they offer some of the best qualities of carbonates without some of the downsides. Ethylene sulfite, the cyclic variant, can form stable SEI films on graphite electrodes and remains a low-viscosity liquid at room temperature. That combination is genuinely useful.
But the Yamaguchi team asked a different question: what happens when you take the same chemical formula, break open the ring, and give the molecule room to flex?
Dimethyl sulfite is essentially ethylene sulfite unchained. Instead of being locked into a ring, its bonds can rotate, producing multiple conformers. These are different shapes the same molecule can adopt depending on its energy state. In bulk solution, DMS exists as a mixture of gauche-gauche (GG) and gauche-trans (GT) forms, which sit very close in energy. But there’s a third arrangement, the trans-trans (TT) conformer, that’s about 12 kJ/mol higher in energy. Unfavorable on its own.
That TT conformer, it turns out, matters more than anyone expected.
What Raman spectroscopy revealed about lithium coordination
Using Raman spectroscopy paired with density functional theory (DFT) calculations, the researchers mapped out how lithium ions organize themselves in DMS-based electrolytes across a wide concentration range. They used lithium bis(fluorosulfonyl)amide (LiFSA) as the lithium salt.
In dilute solutions (up to about 2.0 M), the Raman data showed a clear trend. As lithium salt concentration increased, the peak corresponding to free DMS decreased while a new peak for bound DMS appeared at a higher frequency. An isosbestic point, the spectroscopic signature of a clean two-state equilibrium, confirmed that lithium ions were pulling DMS molecules out of the solvent pool and into their solvation shells.
Quantitative analysis gave a solvation number of 2.9 ± 0.3. That means each lithium ion in dilute DMS solution coordinates with roughly three solvent molecules.
For comparison, ethylene sulfite gives a solvation number of 4.2, the conventional four-coordinate tetrahedral arrangement that lithium ions almost always adopt in solution.
So the flexible solvent wraps fewer molecules around each lithium ion. But how does it manage to achieve a stable coordination with only three molecules?
The DFT calculations provided the answer. The lowest-energy Li(DMS)₃⁺ complex isn’t just three DMS molecules all doing the same thing. It’s a mixed arrangement of two monodentate DMS molecules (each donating one oxygen to the lithium) and one bidentate DMS molecule (donating two oxygens). That bidentate molecule? It’s the TT conformer, the one that’s unstable in bulk solution.
The electrostatic field generated by the lithium ion is strong enough to pull that otherwise unfavorable conformer into a stable arrangement. The lithium ion creates a microenvironment where the normal energy rules don’t apply, and the flexible DMS molecule can adopt a shape that rigid ES simply cannot.
The binding energy for Li(DMS)₃⁺ came out to −444.9 kJ/mol, roughly 10 kJ/mol less stable than Li(ES)₄⁺. That difference, while modest in absolute terms, has outsized consequences for how the battery actually performs.
Highly concentrated electrolytes: where shape matters less
At the other end of the concentration spectrum (3.0 to 4.9 M), the picture changes. The Raman spectra no longer show a clean isosbestic point. Instead, peaks shift progressively toward higher frequencies, indicating that lithium ions are now sharing solvent molecules and being bridged by FSA anions into extended ionic networks.
By 4.9 M, the solvation number drops to 1.3. Essentially every DMS molecule in the system is tied up coordinating with lithium, and the structure looks nothing like the discrete solvation complexes seen in dilute solutions.
At these high concentrations, the DMS and ES systems become nearly indistinguishable. Both form ionically ordered structures where lithium ions are interconnected through FSA anions. The molecular flexibility that created such a stark difference in dilute solutions effectively disappears when the system is crowded enough that anion bridges dominate the coordination environment.
This convergence has practical implications. At high concentrations, the choice between DMS and ES might matter less than the choice of anion. That shifts the design problem in a different direction.
Activation energy: the hidden bottleneck
One of the most important metrics for battery performance, especially for fast charging, is the activation energy for lithium-ion insertion at the graphite electrode. This value reflects how much energy is required to strip solvent molecules off the lithium ion (desolvation) before the bare ion can wedge itself between graphite layers.
The researchers measured this using temperature-dependent electrochemical impedance spectroscopy, extracting the low-frequency resistance at different temperatures and fitting it to an Arrhenius model.
The results were striking:
| Electrolyte | Concentration | Activation Energy (kJ/mol) |
|---|---|---|
| DMS | 1.0 M | 22.0 ± 1.2 |
| ES | 1.0 M | 49.0 ± 3.8 |
| DMS | 4.9 M | 37.0 ± 1.8 |
| ES | 4.9 M | 34.6 ± 2.0 |
| Conventional carbonate | 1.0 M | 53 to 59 |
In dilute solutions, DMS offers an activation energy less than half that of conventional carbonate electrolytes. That’s a dramatic difference. Lithium ions can shed their solvent shell and enter the graphite electrode far more easily, which translates directly to faster charge rates and better low-temperature performance.
The mechanism is straightforward. Fewer solvent molecules in the coordination shell, and those complexes are less tightly bound. Breaking apart Li(DMS)₃⁺ takes less energy than dismantling Li(ES)₄⁺ or the four-coordinate carbonate complexes used in today’s batteries.
In concentrated solutions, that advantage narrows. Both DMS and ES converge to around 35 to 37 kJ/mol because the dominant interaction is no longer between lithium and solvent. It’s between lithium and the FSA anions. Solvent flexibility becomes a secondary factor.
The SEI problem: kinetics vs. stability
Low activation energy sounds great, but it’s only one piece of the puzzle. The other piece is what happens at the electrode surface during the first few charge cycles.
When the researchers ran charge-discharge cycling tests on graphite half-cells, the dilute 1.0 M DMS electrolyte delivered an initial discharge capacity close to the theoretical graphite maximum of 372 mAh/g. But after 50 cycles, only 5.5% of that capacity remained. That’s not a gradual fade. It’s catastrophic failure.
Cyclic voltammetry revealed why. During the first cathodic scan, a reduction current appeared around 2.0 V vs. Li/Li⁺, corresponding to reductive decomposition of DMS itself. The resulting SEI film, formed from decomposed DMS, was unstable and couldn’t protect the electrode surface from continued solvent breakdown. Every cycle, more DMS decomposed, more resistance built up, and capacity plummeted.
The highly concentrated 4.9 M DMS electrolyte told a completely different story. After 50 cycles, it retained 86.7% of its initial capacity with coulombic efficiency near 100%. The difference: at high concentrations, the reduction event shifted from 2.0 V to around 1.0 to 1.5 V, and the SEI that formed came from FSA anion decomposition rather than DMS decomposition. FSA-derived SEI films hold up better.
So there’s a fundamental tension at work. Dilute DMS electrolytes give you excellent kinetics but terrible cycling stability. Concentrated DMS electrolytes sacrifice some of that kinetic advantage but gain the SEI quality needed for long-term operation.
What this means for battery design
The most practical takeaway from this work is the suggestion of mixed-solvent electrolytes. If you combine DMS with ES or with a conventional carbonate, you might be able to engineer a system where DMS provides its kinetic advantage while the more stable solvent component ensures a robust SEI film.
Whether that works depends on understanding the preferential solvation behavior in mixed systems. Would lithium ions preferentially coordinate with DMS because it’s easier to shed? Or would ES dominate the solvation shell because its complexes are more thermodynamically stable? The answer determines whether you get the kinetic benefit of DMS or just dilute it into irrelevance.
Beyond mixed solvents, this work points to molecular flexibility as a design parameter that most electrolyte screening studies don’t explicitly consider. When researchers evaluate new solvent candidates, they typically look at donor number, dielectric constant, viscosity, and electrochemical window. The ability of a solvent molecule to adopt multiple conformations, and how those conformations change under the strong electric field near a lithium ion, is rarely quantified.
It probably should be.
The bigger picture
Lithium-ion battery electrolyte design has been iterating on the same basic chemistry for decades. Carbonate solvents with lithium hexafluorophosphate or lithium tetrafluoroborate are the default, and most research focuses on additive packages or marginal solvent tweaks.
This study suggests that the real gains might come from thinking about solvent molecules not as static containers for lithium ions, but as dynamic participants that change shape, reorganize, and respond to local electric fields. A solvent’s ability to flex, its conformational freedom, might matter just as much as its dielectric constant or viscosity in determining whether a battery charges fast, cycles long, or works in cold weather.
The sulfite system studied here is unlikely to replace carbonate electrolytes in commercial cells anytime soon. But the design principle, using molecular flexibility to tune the balance between solvation stability and desolvation kinetics, could apply to a wide range of solvent families.
For anyone working on next-generation electrolytes, whether for conventional lithium-ion, lithium-metal, or solid-state batteries, the takeaway is the same: the shape of your solvent matters, and how easily that shape can change matters even more.