Sodium-Ion Battery Electrolyte Salt Selection
sodium-ion battery electrolyte salt selection is one of those decisions that looks like a purchasing detail and turns out to define cycle life, low-temperature behavior, safety margins and cost. I have watched teams tune a hard carbon anode and a layered oxide cathode for months, then lose most of the gain because the salt and solvent system was inherited from a lithium battery program unchanged. The chemistry is related but not identical, and the difference shows up in the first formation cycle and again at cell 800.
What the salt actually controls in a sodium-ion cell
The salt supplies the charge carrier, but its anion also decides three things that matter more than the conductivity datasheet. First, it sets the composition of the solid electrolyte interphase on hard carbon, which is where the irreversible sodium inventory is consumed during formation. Second, it decides whether the aluminium cathode current collector stays passive or slowly dissolves. Third, it sets how much hydrofluoric acid forms when trace water gets in, which drives cathode metal dissolution and slow fade.
One piece of physics surprises people arriving from lithium. The sodium ion is larger, roughly 1.02 angstrom against 0.76 angstrom, but it is a weaker Lewis acid and holds its solvation shell loosely. The solvated sodium ion is therefore often no bigger than solvated lithium in carbonate solvents, and the desolvation penalty at the electrode is smaller. That is why a 1.0 M sodium electrolyte matches or beats a comparable lithium electrolyte on conductivity, typically 7 to 9 mS/cm at 25 degrees Celsius for NaPF6 in EC:DMC and 8 to 12 mS/cm in propylene carbonate or ether rich blends.
Salt candidates and the trade you make with each one
NaPF6: the default for a reason
Sodium hexafluorophosphate is the production standard for most cells shipping today. It gives good conductivity, passivates the aluminium collector with a fluoride layer that holds to roughly 4.3 to 4.5 V against sodium metal, and has a mature supply base. The cost is moisture sensitivity. Like its lithium cousin, NaPF6 hydrolyses to HF and phosphoryl fluoride, and the HF attacks the cathode and the SEI. I hold electrolyte water below 20 ppm and preferably below 10 ppm, and I treat the dry room dew point, usually minus 40 degrees Celsius or colder, as a process control rather than a facility specification.
NaClO4: cheap, conductive, and not going into a product
Sodium perchlorate has excellent conductivity and oxidative stability and is a favourite in academic literature because it is cheap and easy to dry. It is also a strong oxidiser, handled as a class 5.1 hazardous material in most jurisdictions, and the consequence of a perchlorate residue inside a drying oven is not something a plant wants. I use it only for screening work in a fume hood with dedicated glassware.
NaFSI and NaTFSI: high performance with an aluminium problem
The imide salts give high conductivity, good thermal stability and reach 12 to 15 mS/cm in ether solvents. The catch is that they do not passivate the aluminium cathode foil. The aluminium salt formed is soluble, so the collector keeps corroding and anodic current on an aluminium electrode climbs from as low as 3.5 to 4.0 V against sodium. To use an imide above 4.0 V you need a mitigation: a few tenths of a molar of NaPF6 as a co-salt to lay down the fluoride layer, sodium difluoro(oxalato)borate as an additive, or a hard cathode voltage limit. I have seen all three work, and all three fail when someone changed the ratio without re-running the corrosion test.
Borates: NaBOB, NaDFOB and film formation
The borate family brings oxalate groups that decompose into a robust boron-rich interphase and generate little or no HF. Used as additives at 0.5 to 1 percent rather than as the main salt, they improve high-temperature cycle life and reduce gas. They cost more per kilogram and solubility is limited, so treat them as a finishing tool.
Solvent systems that sodium allows and lithium does not
This is the most under-used advantage of sodium-ion chemistry. Propylene carbonate is nearly unusable in a graphite lithium cell because it co-intercalates and exfoliates the anode. Hard carbon does not fail that way, so PC becomes available, bringing a melting point near minus 49 degrees Celsius, a dielectric constant around 65 and a wide liquid range. A PC blend is the first thing I reach for when a customer asks for discharge at minus 30 or minus 40 degrees Celsius. The trade is viscosity and poor room-temperature rate, so the practical answer is a ternary blend with a low-viscosity linear carbonate such as EMC.
Ether solvents, diglyme and DME in particular, also deserve mention. They support sodium co-intercalation into hard carbon, giving excellent rate and low-temperature behaviour at conductivities above 12 mS/cm. The limitation is oxidation: ethers break down above roughly 4.0 V against sodium. That rules them out for a layered oxide cathode charged to 4.2 V and makes them a good match for a polyanionic cathode with a 3.4 V plateau.
Concentration is a cost decision before it is a performance one
Conductivity peaks where carrier number times mobility peaks, which for most carbonate systems sits near 1.0 to 1.2 M. Going to 2 or 3 M improves oxidative stability and aluminium passivation, but viscosity rises steeply, wetting of a thick electrode suffers, and you pay for salt you cannot use. Salt is a major line item in cost. Run the standard concentration first and only accept a higher one if data justifies it.
Additives and the hard-carbon SEI
Hard carbon loses a large share of the sodium inventory to SEI formation on the first cycle, and the additive package is the cheapest lever on that number. Fluoroethylene carbonate at 2 to 5 percent by weight is the standard fix, moving first cycle coulombic efficiency from the 75 to 85 percent range into the upper 80s or low 90s depending on the carbon and the formation protocol. Vinylene carbonate at 1 to 2 percent does similar work with a different decomposition profile.
The caveat is that FEC is not free. It is consumed over life, generates carbon dioxide at elevated temperature and high voltage, and when hot can defluorinate into HF. If a cell swells after a 45 degree Celsius storage test, the FEC level and the cathode cut-off voltage are the first two things I look at. Sodium difluoro(oxalato)borate at half to one percent often lets you drop FEC while holding formation efficiency, which suits a high-temperature product.
Prussian blue cathodes raise the dryness bar
If the cathode is a Prussian blue analogue, water is a different problem. These materials carry coordinated and zeolitic water in the lattice that does not all come out in electrode drying. It migrates into the electrolyte and hydrolyses salt into HF. For those cells I take the lowest water specification the supplier will commit to, add borate, and validate with a 60 degree Celsius storage test measuring HF rise and gas volume, not just capacity retention.
Specifying, handling and qualifying an electrolyte lot
On a certificate of analysis I look at water by Karl Fischer, HF by titration, chloride and metals by ICP. My working acceptance limits are water below 20 ppm, HF below 50 ppm, chloride below 5 ppm and each of iron, nickel and copper below 1 ppm. Colour and free acid drift are cheap early warnings that a lot sat too long or got warm. None appear on a datasheet; you have to ask.
Mixing and storage are where good electrolyte gets ruined. Salt addition into carbonate is exothermic, so dissolve under temperature control, usually below 30 to 40 degrees Celsius, with enough agitation to avoid local hot spots. Filtration to 0.2 to 1 micron removes gel particles that later show up as shorts. Store cool, between 5 and 25 degrees Celsius, in HDPE or PTFE lined containers. Never glass for a fluoride salt, because HF etches glass and leaches silicates. Shelf life runs around 12 months for a base solution but only a few months once FEC is in, so date the drums and use them in order.
What a change of salt or solvent costs you
Any change to the salt source, the salt ratio, the solvent blend or an additive level is a formulation change, and that invalidates the qualification you already paid for. I re-run formation and first cycle efficiency in coin or single-layer pouch cells, a rate and temperature matrix, a 45 to 60 degree Celsius storage test with gas measurement, and an aluminium corrosion check on foil. If the cell is already in a pack, the transport evidence under UN 38.3 and the safety case under IEC 62619 or UL 1973 need review, because a different electrolyte changes the thermal and venting behaviour those tests recorded. IEC 62133-2 applies only to portable sealed cells and batteries; a rack module is assessed under IEC 62619.
Dual sourcing is worth the effort. Sodium-ion is early enough in its cost curve that salt supply can be lumpy, and a single-source salt is a single point of failure for the line. What makes it workable is a tight incoming specification and a written change-control agreement, so a supplier process tweak triggers a conversation rather than a surprise in field returns.
FAQ
Why is NaPF6 the default salt for sodium-ion cells?
It combines adequate conductivity with a fluoride layer that keeps the aluminium collector passive to roughly 4.3 to 4.5 V against sodium. The imide salts are more conductive but corrode aluminium, and the perchlorate is a strong oxidiser that plants avoid. NaPF6 is the workable compromise, provided water stays below 20 ppm.
Is sodium-ion electrolyte less conductive than lithium electrolyte?
Not necessarily. The sodium ion is larger but weakly solvated, so its solvated radius is comparable to lithium and its desolvation penalty is lower. A 1.0 M NaPF6 carbonate formulation typically measures 7 to 9 mS/cm at 25 degrees Celsius.
Can propylene carbonate be used in a sodium-ion cell?
Yes, and this is a real advantage over lithium-ion with graphite anodes. Hard carbon does not exfoliate in PC the way graphite does, so PC becomes available for its low melting point near minus 49 degrees Celsius and high dielectric constant. Blend it with a linear carbonate such as EMC to recover rate capability.
How much FEC should a hard carbon formulation contain?
Two to five percent by weight is the normal range, and it lifts first cycle coulombic efficiency from the 75 to 85 percent band into the upper 80s or low 90s. Watch for carbon dioxide and HF at elevated temperature; if a cell swells in hot storage, revisit the FEC level and the cut-off voltage together.
Why do imide salts need NaPF6 as a co-salt?
NaFSI and NaTFSI form soluble aluminium corrosion products, so the collector never passivates and anodic current climbs from roughly 3.5 to 4.0 V against sodium. Adding a few tenths of a molar of NaPF6 lays down an insoluble fluoride film that protects the foil while keeping the imide conductivity benefit.
What water specification should I put on an electrolyte purchase order?
Water below 20 ppm by Karl Fischer, targeting 10 ppm, plus HF below 50 ppm, chloride below 5 ppm and individual metals below 1 ppm. Ask for these on the certificate of analysis rather than accepting a datasheet, and re-test on receipt if the shipment sat in a warm warehouse.
Which safety standards apply after a formulation change?
Transport evidence under UN 38.3 and, depending on product form, cell and pack safety assessment under IEC 62619 or UL 1973 all need review, because a new electrolyte changes thermal and venting behaviour. IEC 62133-2 covers portable sealed cells and batteries, not rack modules.
Further Reading
References
