Sodium-ion (Na-ion) chemistry has moved from laboratory curiosity to commercial stationary product, and Eleven Energy’s residential system is among the first to put it in front of UK installers. For specifiers and investors, the relevant question is not whether sodium can store charge, which has been established for decades, but whether the combination of material economics, thermal behaviour and degradation characteristics justifies displacing lithium iron phosphate (LFP) as the default domestic chemistry. This teardown works from first principles in crystal-structure physics, electrolyte chemistry and battery management system (BMS) estimation, and then maps those fundamentals onto cost and installation consequences. Where a figure describes Eleven Energy’s own product, it is the manufacturer’s stated characterisation; where it describes the wider chemistry, it is a representative engineering range and should be checked against the datasheet of any specific unit.
Why Sodium Behaves Differently from Lithium: Ionic Radius and Thermodynamic Potential
Sodium and lithium are both alkali metals and both form singly charged cations, which is why the intercalation architecture of Na-ion cells closely mirrors that of lithium-ion. The differences begin with size and mass. The Shannon ionic radius of Na+ is approximately 1.02 Å, against 0.76 Å for Li+, a volume ratio of roughly 2.4:1. Sodium is also three times heavier per mole (23.0 g against 6.94 g). Both facts matter for energy density: a larger, heavier charge carrier requires a more open host lattice and contributes more dead mass per unit of charge transferred.
The standard electrode potential shifts too. The Na/Na+ couple sits at about −2.71 V versus the standard hydrogen electrode, compared with −3.04 V for Li/Li+. Sodium is therefore approximately 0.33 V less reducing than lithium, so, all else equal, a Na-ion full cell has a lower electromotive force than an equivalent lithium cell. Because specific energy is the product of capacity and voltage, this alone costs several per cent of energy density before any structural effects are counted. Combined with the lattice volume required to host the larger cation, it explains the 20 to 30 per cent volumetric energy-density deficit cited for Na-ion cells relative to LFP.
The larger radius has a compensating benefit in solvation. The charge density of Na+ is lower, so its Coulombic attraction to solvent molecules is weaker and its solvation shell is looser. The desolvation energy required for the ion to shed solvent molecules at the electrode surface, which is a dominant contributor to charge-transfer resistance, is lower than for Li+ in the same carbonate solvents. This is the physical origin of the favourable rate and low-temperature behaviour examined later in this article.
Anode Physics: Hard Carbon and the Failure of Graphite
Graphite, the standard lithium-ion anode, does not store sodium effectively in carbonate electrolytes. The thermodynamics of sodium-graphite intercalation compounds are unfavourable because the binding of Na to graphene layers is weaker than the cohesive energy of metallic sodium, so sodium tends to plate rather than intercalate. The practical anode is therefore hard carbon: a non-graphitising, disordered carbon synthesised by pyrolysis of precursors such as biomass, resins or sugars, typically at 1,000 to 1,600°C.
The structure of hard carbon is a turbostratic arrangement of small, randomly oriented graphene-like domains separated by nanopores. Its sodium storage proceeds through several overlapping mechanisms, described by the “adsorption–intercalation–pore filling” model. At higher potentials (above roughly 0.1 V versus Na/Na+) sodium adsorbs on defect sites and edges, producing the sloping region of the voltage curve. At lower potentials, sodium inserts between expanded graphene layers (interlayer spacing above 0.37 nm, wide enough for Na+) and then fills closed nanopores as quasi-metallic clusters, producing the low-voltage plateau that contributes most of the usable capacity. Typical reversible specific capacities are 250 to 350 mAh/g.
Two engineering consequences follow. First, the plateau sits very close to the sodium plating potential at 0 V versus Na/Na+, a narrower safety margin than graphite offers lithium, which makes charge-rate control and low-temperature charge management a BMS responsibility, not an assumption of the chemistry. Second, hard carbon has a lower first-cycle Coulombic efficiency than graphite, often 80 to 90 per cent, because irreversible sodium is consumed forming the solid-electrolyte interphase (SEI) on a high-surface-area, defective carbon. Cell manufacturers manage this through surface coating, pre-sodiation or cathode over-sizing, each of which carries a cost.
Cathode Physics: Layered Oxides and Prussian Blue Analogues
Two cathode families dominate commercial Na-ion development, and they have opposite structural logic.
Layered transition-metal oxides (NaxMO2). These consist of sheets of edge-sharing MO6 octahedra (M typically Ni, Fe, Mn, Cu, sometimes Ti or Mg) between which sodium resides. Sodium layered oxides are classified by Delmas notation according to the coordination of sodium and the stacking sequence: O3-type (octahedral sodium, three layers per repeat) and P2-type (prismatic sodium, two layers per repeat). O3 phases begin with a full sodium inventory and give higher initial capacity, but undergo multiple phase transitions on desodiation (O3 to O′3 to P3), accompanied by sheet gliding and volume change. P2 phases offer wide, open prismatic diffusion channels and lower diffusion barriers, producing excellent rate capability and cycle stability, but start sodium-deficient and so need sacrificial sodium sources. Because the large Na+ sits in sites with a pronounced preference for either octahedral or prismatic coordination, these lattices are more prone to phase transitions on cycling than their lithium counterparts, and the transitions are the principal origin of mechanical lattice strain.
Prussian Blue analogues (PBAs). These adopt a face-centred cubic framework of transition-metal ions (commonly iron, manganese or nickel) bridged by cyanide ligands, AxM[Fe(CN)6]. The open framework contains large interstitial sites and three-dimensional channels that accommodate Na+ with minimal lattice distortion, giving very small volume change on cycling and potentially long life. The challenges are manufacturing quality: crystal water and vacancies in the lattice reduce capacity and Coulombic efficiency, and rigorous drying and synthesis control are required. PBAs generally operate at 3.0 to 3.4 V and use abundant iron and manganese, with no nickel or cobalt requirement, which aligns with the supply-chain logic of the chemistry.
Whichever cathode is used, the full-cell nominal voltage is generally 2.8 to 3.2 V, slightly below LFP’s 3.2 V. For domestic systems assembled at 48 V nominal, this means a few additional cells in series, and the cell and pack designers adjust accordingly.
Electrolyte Chemistry and the Interphase
Most Na-ion cells use a carbonate solvent blend (ethylene carbonate with diethyl or dimethyl carbonate, or propylene carbonate) with sodium hexafluorophosphate (NaPF6) as the salt at roughly 1 mol/L, with additives such as fluoroethylene carbonate to stabilise the SEI. NaPF6 shares the moisture sensitivity of LiPF6: hydrolysis yields hydrofluoric acid, so dry-room manufacturing remains necessary. Alternative salts, such as sodium bis(fluorosulfonyl)imide (NaFSI) and sodium perchlorate in research cells, offer improved thermal stability and conductivity but pose their own cost, corrosion and safety questions.
The SEI on hard carbon is thicker and more soluble than the lithium SEI on graphite, because sodium salts of organic decomposition products (sodium alkyl carbonates, Na2CO3, NaF) have higher solubility in carbonate solvents. This continually dissolving and re-forming interphase is a significant calendar-ageing mechanism and is the principal reason electrolyte additive packages are the subject of intensive proprietary development.
Manufacturing Material Costs: Current Collectors and Scarcity
One of the structural advantages of Na-ion is the current collector. In a lithium-ion cell the anode potential falls to near 0 V versus Li/Li+, at which aluminium alloys with lithium and would be destroyed, so the anode must use copper foil. Sodium does not alloy with aluminium at these potentials, so aluminium foil can serve as the current collector on both electrodes. This is a genuine and unique material benefit of the chemistry.
The numbers are worth stating precisely, because they are often overstated. Copper has a density of 8.96 g/cm3 against 2.70 g/cm3 for aluminium, so for a given foil thickness, an aluminium collector is about 70 per cent lighter. Copper trades at roughly three to four times the price of aluminium per kilogram, with materially higher volatility in spot pricing. Anode copper foil in an LFP cell is typically in the range of 0.5 to 0.8 kg per kWh and represents in the order of 4 to 8 per cent of cell material cost. Replacing it with aluminium foil therefore saves a modest but real few pounds per kWh, plus the weight reduction and the removal of exposure to copper price cycles. It is a useful contributor to the cost case, but not the decisive one on its own.
The larger economic argument concerns scarcity and price volatility of the active materials. Lithium carbonate prices have moved by an order of magnitude in the past five years, from under $10,000 per tonne to above $70,000 per tonne at the peak, then back to levels that have compressed margins across the industry. Sodium, supplied as soda ash (sodium carbonate) at a few hundred dollars per tonne, is about two orders of magnitude cheaper per tonne and geographically ubiquitous. Na-ion cathodes can also avoid nickel and cobalt entirely, and PBA or iron-manganese layered oxide cathodes draw on abundant elements.
It would be misleading to assert that Na-ion cells are already cheaper than LFP in every market. LFP has benefited from more than a decade of gigafactory scale, and recent cell prices have fallen below $60 per kWh in volume. Sodium-ion cells currently benefit from lower materials cost but suffer from lower production scale, lower energy density (more cell, casing and separator per kWh), hard-carbon cost and yield learning. The consensus projection is for cost parity or a modest advantage as volume rises, and a stronger advantage whenever lithium spot prices rebound. For a UK installer, the practical consequence in 2026 is that the retail cost per kWh of Na-ion storage is likely to be in the same range as LFP, with the price lever largely about future supply security. Investors should treat claims of an immediate large discount with caution.
Comparative Matrix: Cell Variants in Stationary Storage
| Electrochemical Storage Cell Variant | Nominal Cell Voltage (V) | Volumetric Energy Density (Wh/L) | Thermal Operating Window Range | Cycle Life Longevity (to 80% SoH) | Projected Capital Cost per kWh (UK Market 2026) |
|---|---|---|---|---|---|
| Sodium-ion, layered oxide cathode with hard-carbon anode | Approximately 3.0–3.1 | Approximately 200–300 | Approximately −20°C to 55°C discharge; low-temperature charge retained with managed rates | Approximately 3,000–5,000 cycles | Approximately £450–£700 installed; early-volume pricing at or slightly above LFP |
| Sodium-ion, Prussian Blue analogue cathode with hard-carbon anode | Approximately 3.0–3.3 | Approximately 150–250 | Approximately −20°C to 60°C; wide window, moisture control critical in manufacture | Approximately 3,000–6,000 cycles (low lattice strain) | Approximately £450–£700 installed; scale-dependent |
| Lithium iron phosphate (LFP) with graphite anode | 3.2 | Approximately 300–400 | Approximately −20°C to 55°C discharge; charging typically restricted below 0°C | Approximately 4,000–8,000 cycles | Approximately £400–£650 installed; lowest-cost established benchmark |
| Lithium nickel manganese cobalt oxide (NMC) with graphite anode | 3.6–3.7 | Approximately 500–700 | Approximately −20°C to 55°C discharge; charging typically restricted below 0°C | Approximately 1,500–3,000 cycles | Approximately £550–£800 installed; more common in mobility than domestic storage |
Matrix notes: cost figures are indicative installed ranges for a typical 5 to 15 kWh domestic system including inverter, installation and VAT treatment appropriate to the scheme, and are not quotations. Cycle life depends on depth of discharge, temperature and C-rate and should be taken from warranted values in practice.
Thermal Kinetics and Sub-Zero Charge Execution
In freezing conditions, the rate-limiting steps in a lithium-ion cell are the slowing of ion transport in the electrolyte, the desolvation of Li+ at the anode surface, and the sluggish diffusion of lithium within graphite. As temperature falls, the ionic conductivity of carbonate electrolytes declines roughly exponentially following Arrhenius behaviour, while the charge-transfer resistance at the graphite interface rises sharply. When charging current is applied, the anode potential is driven below 0 V versus Li/Li+ and metallic lithium deposits on the surface instead of intercalating. The deposit forms needle-like dendrites that can pierce the separator and cause internal short circuits, and also consumes cyclable lithium irreversibly. This is why LFP storage systems generally prohibit or restrict charging below about 0°C, commonly using heating elements to bring the cells into range before accepting power.
Sodium-ion cells start from a better position for two reasons. First, the lower desolvation energy and higher mobility of the weakly solvated Na+ in the electrolyte means that conductivity and interfacial kinetics degrade less steeply with falling temperature. Second, the porous, disordered structure of hard carbon provides fast ion-access pathways, and sodium diffusion coefficients in the electrolyte and in the layered oxide host remain usable at −20°C. Manufacturers report the retention of 80 to 90 per cent of room-temperature capacity at −20°C for well-optimised cells. Eleven Energy’s stated figure is 85 per cent capacity retention at −20°C, which is consistent with this range.
It is important not to overstate the case. Hard carbon’s plateau lies only about 0.1 V above the sodium plating potential, so sodium metal plating is not impossible; it is a lower-probability failure at a given temperature and rate than lithium plating on graphite, not an eliminated one. Sodium plating also behaves differently from lithium plating: sodium metal is highly reactive and forms a more soluble SEI, which can reduce the persistence of deposits, but the underlying risk of capacity loss and internal short circuit from deposition remains if charge current is applied excessively at low temperature. Safe design therefore relies on the combination of chemistry and control. The BMS should apply temperature-dependent charge current limits, cell-level temperature sensing and, where the installation location warrants it, modest heating. For uninsulated UK garages, lofts and outbuildings, where winter temperatures can approach −10°C, a chemistry that continues to accept charge with a small derating, rather than refusing it outright, is a meaningful practical benefit, particularly for systems that must absorb surplus generation in cold, bright weather.
Safety and the 0 V Transport State
A distinctive property of Na-ion cells is that they can be discharged to 0 V without damage. In lithium-ion cells, over-discharge drives the anode potential up to the point where the copper current collector dissolves, at roughly 3.0 V versus Li/Li+, leaving copper ions that later deposit as dendritic shorts on recharge. Because the aluminium collector on a Na-ion anode is stable across the whole accessible potential window, with no equivalent dissolution mechanism, a fully discharged Na-ion cell can be held at 0 V with both terminals shorted and still recover capacity when recharged.
The logistical implications are significant. Lithium-ion batteries must typically be shipped at around 30 per cent state of charge as Class 9 dangerous goods under UN 3480 and 3481, because a charged cell stores energy that can drive thermal runaway if damaged. A cell at 0 V stores essentially no electrochemical energy, so the hazard profile in transport and storage is lower, and pack-level shipping, warehousing and installation-day handling are simplified. Regulations still apply to Na-ion cells and packs, and certification is required, so claims of exemption from dangerous-goods rules should be verified against the current UN Manual of Tests and Criteria and national guidance. The safer handling is a real benefit, but the regulatory classification is a separate matter.
In operation, Na-ion cells also display favourable thermal runaway behaviour: onset temperatures for exothermic decomposition tend to be higher than those of NMC and broadly comparable to LFP, and heat release is typically lower for layered oxide and PBA cathodes. As always, abuse tolerance is a function of cell design and not only chemistry, so installers should rely on UL 9540A, IEC 62619 and equivalent test evidence for the specific product.
Degradation Vectors: Calendar Ageing versus Cycle-Induced Lattice Strain
Over a 10-year horizon, two degradation families operate in parallel. They respond to different stressors and require different mitigation strategies.
Calendar ageing proceeds whenever the cell exists, and is accelerated by high temperature and high state of charge. In Na-ion cells, the dominant mechanisms are SEI dissolution and reformation on hard carbon (consuming sodium inventory and electrolyte), transition-metal dissolution from the cathode with deposition on the anode, and, for PBAs, slow loss of structural water and cyanide-framework defect growth. A domestic system that sits at high state of charge in a warm cupboard will age faster than one that is cycled within a mid-window at stable, cool temperatures.
Cycle-induced mechanical strain arises from repeated insertion and extraction of the large Na+ ion. Hard carbon swells modestly, typically below 10 per cent, which is acceptable. Layered oxide cathodes are the greater concern, because phase transitions between O3, P3 and related structures involve sliding of oxide layers and abrupt changes of the c-axis lattice parameter, producing particle-level cracking, loss of electrical contact and exposure of fresh surface to electrolyte. PBA cathodes experience much smaller lattice change and tend to be more durable mechanically. Designers manage layered-oxide strain through multi-element substitution (Ti, Mg, Cu, Li) which suppresses phase transitions and stabilises the structure through high-entropy design, and through limiting the upper cut-off voltage.
For a domestic use case of one cycle per day, ten years equals roughly 3,650 cycles, and a daily depth of discharge of 70 to 90 per cent is typical. A Na-ion cell warranted for 3,000 to 5,000 cycles to 80 per cent state of health will therefore finish the decade at or slightly above the end-of-life threshold, which is why the combination of warranty terms, throughput limits and operating profile deserves careful reading. LFP’s higher cycle ratings provide more headroom in high-throughput applications such as time-of-use arbitrage with two cycles a day.
Battery Management System Estimation Algorithms
The BMS is where the chemistry-specific physics becomes an engineering problem, and Na-ion presents specific challenges. State-of-charge (SoC) estimation in LFP is notoriously difficult because the open-circuit voltage (OCV) curve is flat across most of the capacity range. Hard-carbon/layered-oxide Na-ion cells have the opposite character: the sloping voltage profile of hard carbon and of layered oxides means that OCV is a strong function of SoC across much of the range, which makes voltage-based correction more practical. That sloping behaviour is a diagnostic advantage that may allow simpler SoC estimators to achieve the same accuracy as more complex ones for LFP.
The practical approach combines coulomb counting with periodic OCV recalibration at rest, typically within an extended Kalman filter (EKF) or unscented Kalman filter whose state vector includes SoC and polarisation voltages from an equivalent circuit model (ECM) with one or two resistor-capacitor branches. Parameters of the ECM (ohmic resistance, charge-transfer resistance and diffusion time constants) vary strongly with temperature, so they are stored as lookup tables or parameter-adaptive models, with recursive least squares updating to track ageing. Hysteresis in hard carbon and in some layered oxides, caused by different charge and discharge paths in the OCV curve, must also be modelled to avoid SoC drift, particularly under partial cycling typical of solar self-consumption.
State-of-health (SoH) estimation uses capacity-based and resistance-based indicators, increasingly supplemented by incremental capacity analysis (dQ/dV) and differential voltage analysis, where phase transitions in layered oxides produce distinctive peaks that shift and decay as the cell ages. Temperature-dependent charge current limiting, the core of cold-weather protection, uses the estimated cell temperature and internal resistance to compute the maximum current that keeps the anode potential above a plating margin, often 20 to 50 mV. In a well-implemented BMS, this means charge is permitted in cold weather at tapered current, rather than being refused or being permitted at damaging rates. Cell balancing is standard, but Na-ion’s tolerance of deep discharge provides additional margin for passive balancing strategies and for recovery from low-voltage events.
Implications for Specifiers, Investors and Consultants
For residential electrical specifiers, the key practical differences are that Na-ion permits installation in colder, uninsulated locations with a smaller derating, offers lower stored-energy hazard during transport and handling, and offers a nominal voltage and energy density that requires only minor changes to the system design. Physical volume per kWh will be larger by roughly a quarter to a third; check cupboard, wall-bracket and garage dimensions during survey, and ensure that any inverter compatibility and communications protocol are confirmed.
For clean-tech investors, the thesis rests on supply-chain security, cost trajectory and technology risk. The advantages are structural and increase when lithium prices rise, but the near-term cost advantage is smaller than headline claims, LFP continues to improve, and the cycle-life evidence base for Na-ion is thinner. A reasonable position is to treat sodium-ion as a hedge and a complement to LFP, with the strongest early use cases in cold-climate, space-tolerant, low-cycling or safety-sensitive installations. For micro-generation consultants, the recommendation is to evaluate each proposal against the warranted throughput (MWh), temperature range, round-trip efficiency and installed cost per kWh of usable capacity, and not to rely on chemistry labels alone.
The balanced conclusion is that sodium-ion is a credible and in some respects superior chemistry for stationary domestic storage. It is not yet a clear cost winner, it does not eliminate the need for careful low-temperature charge management, and its long-term field data remain limited. Its real strengths are abundant materials, aluminium current collectors, benign low-temperature behaviour and a 0 V transport state. Those are meaningful asset-optimisation advantages that justify serious evaluation alongside, not in place of, the LFP baseline.