Magnesium as a Battery Enclosure Material

Where the benchmark report asks how Estes Energy Magnus compares numerically to other packs, this one asks a narrower question: is magnesium actually the right enclosure material, on the physics and the evidence published so far. Magnus is used throughout as the worked example because it is the only near-production attempt at this scale; every risk raised below is evaluated against what Estes has and hasn’t disclosed, not against magnesium in the abstract.

This post has been built based on the support of and sponsorship from: illumion, Custom Power, Quarto Technical Services, Eatron Technologies and About:Energy

Enclosure Materials Landscape

Roughly 80% of production EVs use an aluminium enclosure today, with steel accounting for most of the remainder. Composite (SMC, CFRP, glass-fibre thermoplastic) is the fastest-growing challenger, with most major OEMs now running a composite battery-box programme in development or production. Magnesium currently has essentially no production precedent in this application beyond one small-scale motorbike pack: Royal Enfield’s Flying Flea FF.S6 (EICMA 2025, production targeted late 2026), a finned, passively air-cooled magnesium battery case with no coolant channel cast in, sidestepping the water-glycol and coolant-contact-with-structure risks discussed below for Magnus. Beyond the fact of the finned case, Royal Enfield has published no further detail.

Density and market-share figures are current industry-reported figures (SAE, Ducker Carlisle, Novelis, ArcelorMittal, SABIC; see Sources); cost and risk characterisations are this analyst’s synthesis.

Aluminium

A balanced compromise rather than a standout on any one property: decent weight saving over steel, self-passivating corrosion resistance, good thermal conductivity, mature supply chain. Its structural weakness here is a genuine recycling ceiling, iron picked up during end-of-life shredding embrittles high-performance die-cast alloys above ~0.7 wt% Fe, capping them at ~85–90% primary content (independently documented by Oak Ridge National Laboratory, not just an Estes talking point). Aluminium enclosures also generally need an added thermal blanket to meet stringent fire-penetration standards; bare aluminium has been reported to fail within seconds of direct 1,100°C flame exposure without it.

Thin pressed steel

Raw density (7.85 g/cm³, roughly three times aluminium or magnesium) overstates steel’s real mass penalty, because it compares equal thickness, not equal function. What actually governs a battery enclosure is stiffness in some load paths and strength/toughness in others, and steel’s much higher modulus and yield strength let it run thinner. For a bending-stiffness-limited panel (E scaling as t³), steel only needs about 70% of aluminium’s thickness for equal stiffness, closing the mass gap to roughly 2x rather than 3x. For a strength-limited panel, a modest advanced-high-strength-steel grade (~500 MPa) needs only ~55% of aluminium die-cast’s thickness (~150 MPa) for equal bending strength, a mass ratio near 1.6x; a hot-stamped boron steel (~1,300 MPa, the grade ArcelorMittal and others already push for automotive structural parts) needs only ~34% of the thickness, landing at roughly mass parity with aluminium. Which comparison applies depends on which load case actually governs a given panel, underbody impact and puncture resistance (steel’s strongest case) versus overall panel stiffness (where the gap stays wider), so a single density figure was the wrong way to frame it.

Steel also has two advantages this report undersold. First, recycling: it has no equivalent to aluminium’s iron-contamination ceiling (it already is the iron stream), and steel recycling is by far the most mature, highest-volume, highest-recovery-rate loop of any of the four materials, electric-arc-furnace steelmaking from scrap is an enormous, established industry, not an emerging one. Second, thermal-runaway and fire-penetration performance: steel’s melting point (roughly 1,400°C and up) sits well above both a UL 2596-class flame test temperature (1,100°C) and the peak temperatures a cell venting in thermal runaway can reach, whereas aluminium (~570–660°C) and magnesium (~450–650°C) both melt within that range, which is the physical reason bare aluminium has been reported to fail within seconds under direct 1,100°C exposure. Untreated steel should not melt through at that same test condition at all; this report has not found a published UL 2596 result for bare steel to cite directly, but the melting-point gap makes the inference straightforward. That’s a real, underexploited advantage for steel specifically as a fire/thermal barrier layer, independent of its cost and recycling case.

Composite (SMC / CFRP / glass-fibre thermoplastic)

Magnesium’s most direct competitor on the weight-saving pitch: 30–50% mass savings over aluminium are cited in current industry reporting, comparable to or exceeding magnesium’s claimed ~33% density advantage, without the flammability or corrosion baggage. SABIC’s glass-fibre polypropylene, in production on the Honda CR-V PHEV battery cover, passed UL 2596 thermal-runaway testing (pack-side temperature held under 200°C through a 1,100°C flame for five-plus minutes) without a separate thermal blanket, versus aluminium’s reported near-immediate failure unprotected. That part delivered a 10% weight and 10% cost saving over steel-plus-blanket, a concrete shipping precedent magnesium cannot yet match. Weaknesses: poor inherent EMI shielding, immature thermoset recycling, and a track record concentrated in covers rather than load-bearing trays.

Magnesium

The highest-risk, highest-reward option, and the subject of the rest of this report. Genuinely the lightest structural metal (~33% lighter than aluminium by volume), specific stiffness roughly comparable to aluminium’s, decent EMI shielding, and via thixomolding the ability to form very large single-piece thermo-structural bodies. Its liabilities are the three reasons the industry has avoided it for decades: standard alloys ignite at or below their own melting point once molten, bare magnesium corrodes far faster than aluminium and needs a dedicated coating system, and its current life-cycle footprint is worse than aluminium’s unless sourced outside the Pidgeon-process-dominated global supply (~92% of world magnesium). Estes’ pitch is that MgCarbonit™ and Levius™ solve the first two; whether that holds at full pack scale, in the field, over a service life, is not yet demonstrated the way a decade of aluminium production, or even SABIC’s one composite part, has been.

Estes’ own materials lean on three further properties: specific stiffness, EMI shielding, and heat capacity per kilogram. All three numbers check out, but each is framed on the axis that favours magnesium. Specific stiffness (E/rho) for magnesium and aluminium both land around 25-26 GPa cm3/g, magnesium’s lower modulus offset by its lower density, so this is parity with aluminium, not an advantage over it, a fact true of most structural metals rather than something specific to Estes’ alloy. EMI shielding is real but only differentiates magnesium from composite enclosures, both magnesium and aluminium are conductive solids and shield comparably; it says nothing against the aluminium incumbent Magnus is actually replacing. Heat capacity per kilogram is the most selectively framed: magnesium’s specific heat (~1.0 J/g-K) genuinely beats aluminium’s (~0.88-0.90 J/g-K) by mass, but aluminium’s higher density flips that on a volumetric basis, aluminium holds roughly a third more heat capacity per unit volume. For a mass-constrained platform the per-kilogram framing is defensible on Estes’ own terms, but it only speaks to passive, transient heat buffering (absorbing a runaway pulse), the same caveat already raised under Thermal Conductivity below regarding steady-state active-cooling performance.

How Magnesium Shapes the Magnus Design

Magnus is described as a monocast thermo-structural body, a single magnesium casting combining roles conventionally split across enclosure, module trays, cooling plates, and brackets. Estes states the design is cell-to-pack with no module layer, cells in direct thermal contact with the structure; read against its “fewer unique components” claim, the likeliest interpretation is that cell-retention cavities are formed directly into the casting rather than a separate carrier tray, though Estes never confirms this.

That has a direct materials consequence: magnesium is electrically conductive, and cell cans typically sit at a voltage potential, so direct cell-to-structure contact needs a dielectric barrier somewhere in the stack. The Levius coating’s stated “dielectric isolation for high-voltage systems” plausibly does double duty here, both external environmental protection and internal cell isolation, though this is this report’s inference, not a confirmed fact.

Cooling channels are cast directly into the structural walls, coolant contacting the casting itself rather than a bolted-on cold plate, which Estes frames as a deliberate safety choice (no fittings or seals inside the HV boundary). Not every configuration carries active cooling: for “moderate thermal loads” the body is described as acting as a passive heat sink alone, with active channels reserved for higher-performance duty cycles. HV distribution hardware (fuse, contactors, current sensor) appears in Estes’ promotional imagery as a separate module external to the main casting; Estes publishes no fuse rating, contactor spec, or BMS supplier, describing the BMS only as a “distributed architecture” for redundancy.

Why thixomolding, not conventional die casting

Estes’ process choice is thixomolding rather than liquid magnesium high-pressure die casting (HPDC), and the distinction matters for both safety and scale. Thixomolding feeds solid magnesium chips into a heated screw and shapes the alloy semi-solid, roughly 580-600°C at a 30-40% liquid fraction, rather than fully melting it. That avoids the large melt-holding furnace and SF6/SO2-style cover-gas loop that liquid magnesium HPDC typically needs to suppress oxidation, a genuine process-safety and cost simplification independent of any of Estes’ own coating or alloy claims, and well documented in the wider die-casting literature rather than an Estes-specific claim.

Scale has historically been thixomolding’s limit, not its safety case: parts have generally stayed under about 5kg, well short of a full battery enclosure. Some Chinese gigacasting programmes are already moving larger magnesium parts into mass production, and Estes is making a similar bet for a different geography, a US-based giga-thixomolding press targeted at roughly 30kg per shot, one shot per minute, for the first half of 2027, alongside a stated preference for US/allied magnesium feedstock over the Pidgeon-process-dominated global supply chain discussed under Life Cycle Assessment below. Founded April 2024 by engineers from Tesla, Sila, and Cummins and backed by an $11M seed round co-led by BMW i Ventures and Fortescue (July 2025), Estes is a funded, credentialed team, which supports the manufacturing bet being a serious one, but a press of this scale for this material is still an unproven step up rather than a scaled version of an existing production process.

Engineering Risk Areas

Corrosion (salt / galvanic)

Bare magnesium is markedly more vulnerable to chloride-driven and galvanic corrosion than aluminium. Levius (PEO ceramic plus organic topcoat) is validated by 2,000+ hours and a rating of 9 in ASTM B117 salt-fog testing, an intact-surface result. It does not address performance after mechanical damage (stone chips, underbody impact), the materially relevant scenario for underbody-mounted commercial/marine use, and no combined chip-plus-corrosion data (SAE J400 / GM9540P-style) has been published.

Flammability

Standard commodity alloys (AZ91D) ignite at or below their own melting point once molten, the specific reason magnesium has been excluded from battery structures historically. MgCarbonit, via dopants and inline carbon-nanoparticle treatment, is claimed to push the ignition point above the melting point, verified by third-party FAA Ch.25 oil-burner testing (certificates confidential, not reviewed here).

HV insulation, dielectric strength, and pinhole robustness

The only quantitative dielectric claim is that the organic topcoat adds isolation “up to 1500VDC,” a system-compatibility statement, not a material property; no breakdown voltage, coating thickness, or test standard is disclosed, and no insulation-resistance data exists for before vs. after a thermal or mechanical abuse event.

Clarified directly by Estes (CEO Dustin Grace, LinkedIn, August 2026): 1500VDC is the maximum system operating voltage supported, not a withstand or breakdown-voltage test result; actual dielectric withstand is “significantly higher” (no figure given), and current production stays below 1000V. On that reading 1500V comfortably covers both a 400V and an 800V pack; an earlier version of this report misread the claim as a withstand rating and concluded the opposite, that inference is retracted. What remains genuinely undisclosed is unchanged: no breakdown voltage, coating thickness, or test standard has been published, and no insulation-resistance data exists for before vs. after a thermal or mechanical abuse event.

A more fundamental question sits underneath: PEO coatings grow by deliberately driving the substrate above its own dielectric breakdown field, producing thousands of micro-discharges. Characteristic PEO morphology shows donut-shaped growth features with a pore at the centre of each, the literal footprint of the discharge channel; whether that pore self-seals or stays open as a through-thickness defect depends on process control. Porosity is an inherent feature of how PEO coatings form, not an occasional defect, which is almost certainly why Levius pairs PEO with a pore-sealing organic topcoat rather than relying on PEO alone. Published PEO porosity-control data comes from flat test coupons; Magnus is a large single casting with thin walls, cast-in cooling channels, and internal cell-cavity corners, geometries where current density and coverage during PEO processing are known to be harder to control uniformly. Estes has not published anything on coating uniformity across the part, nor the test that would actually settle the question: a holiday/spark test for pinholes, ideally as production-line inspection. ASTM B117 is a corrosion-progression test on samples, not a pinhole-detection test. MgCarbonit’s claimed alloy-level corrosion inhibition would slow attack at a pinhole site, but that mitigates the corrosion consequence, not the dielectric one.

Thermal conductivity

Estes states MgCarbonit runs approximately 60 W/m·K. Common aluminium die-cast structural alloys (A356-T6, A380, Silafont-36-class) run 120–140 W/m·K, roughly double or more. For a design where the wall itself is the heat exchanger, that gap sets the temperature gradient across the wall for a given heat flux; thermal resistance is thickness divided by conductivity, so a thinner magnesium wall (thixomolding reaches ~3.5mm vs. a more typical 4mm+ for aluminium) only partially compensates, a 3.5mm Mg wall at 60 W/m·K still carries roughly double the thermal resistance of a 4mm Al wall at 130 W/m·K. Estes counters low conductivity by citing thermal diffusivity and specific heat capacity, both of which MgCarbonit wins on, but those properties govern passive, transient heat buffering (absorbing a runaway pulse), not the steady-state heat rejection an active cooling channel performs during normal fast-charge or sustained-load operation. The active-cooling conductivity penalty is not directly addressed in what Estes has published.

Water-glycol coolant compatibility

This is the sharpest open issue. Peer-reviewed literature on magnesium engine and transmission components states plainly that most commercial coolant inhibitor packages are formulated for aluminium, cast iron, and copper, and that most existing commercial coolants fail to provide adequate corrosion protection to magnesium alloys without a magnesium-specific inhibitor addition. Corrosion rate is also sensitive to chloride, sulfate, and bicarbonate contamination from hard-water top-offs or coolant aging, the same chloride vulnerability covered under salt corrosion above, except here occurring in continuous, sustained contact for the vehicle’s service life rather than intermittent external spray. Estes’ spec sheet lists cooling simply as “WEG 50/50” with no mention of a magnesium-specific inhibitor formulation, and it is not stated whether Levius, framed throughout Estes’ materials around external/atmospheric exposure, is even applied to the internal coolant-wetted channel walls. If it is not, those surfaces would rely solely on MgCarbonit’s alloy-level inhibition with no topcoat, a materially different and less-tested protection level than the headline salt-fog numbers describe.

Cooling the whole enclosure: a thermal-efficiency trade-off

Because Magnus routes coolant through the structural walls themselves rather than a discrete internal cold plate, the coolant loop’s thermal footprint is effectively the pack’s entire external envelope, not a small contained component. In a conventional pack, the cold plate sits inside the structure near the cells, separated from the outer shell; the coolant-carrying surface and the environment-facing surface are two different parts doing two different jobs, and the outer shell can be left thermally passive or even lightly insulated. In Magnus’s architecture they are the same magnesium wall, so the temperature-controlled coolant is in large-area, direct contact with ambient conditions across most of the pack’s exterior. That drives parasitic heat exchange with the environment, losing conditioning heat to cold ambient air or gaining heat from a hot one, direct sun, road and underbody heat, that works against whatever the thermal management system is trying to hold, a real efficiency cost rather than a serviceability one (these packs are not designed to be serviced regardless of material, so that is not where this design choice actually bites). Mitigating it would call for a separate external thermal shield or insulation layer over the structural body, reintroducing a component to decouple the coolant-carrying structure from ambient, which cuts against the part-count and fitting-elimination simplicity Magnus is pitched on. Estes’ published materials do not mention any external insulation or thermal shield, so it is unclear whether this parasitic loss has been designed around, added as an unaccounted-for extra part, or not yet addressed. It is also a possible tension inside Estes’ own claims: if holding pack temperature costs more auxiliary energy because of ambient coupling across the whole shell, that works against the range benefit of the mass savings claimed elsewhere on the same platform.

Life Cycle Assessment: Magnesium vs. Aluminium

Estes claims a 75% GHG reduction for MgCarbonit sourced from recycled magnesium and 90% from renewable electrolytic production, versus a primary-aluminium baseline. Checked against the independent literature Estes itself cites (Ehrenberger, DLR/German Aerospace Center, 2020 IMA LCA study update), the picture is production-route-dependent. The Pidgeon process, thermal reduction, coal-heavy energy, ~92% of world magnesium supply, comes in at a weighted average of 21.8 kg CO2e/kg cradle-to-gate per DLR (close to Estes’ own cited 24.0), roughly 1.3–2x worse than primary aluminium’s global average (~16 kg CO2e/kg) and meaningfully worse than a European aluminium mix (~8–10 kg CO2e/kg). In DLR’s own worked example, a lighter magnesium cross-car-beam (4kg vs. 5.4kg aluminium) still produced roughly double the cradle-to-gate emissions when sourced from average Pidgeon material, and use-phase fuel savings over 200,000km did not close that gap; DLR’s conclusion is that Pidgeon-sourced magnesium does not break even against aluminium within normal passenger-vehicle lifetimes, only lower-carbon sources do.

Lower-carbon routes exist (Israeli electrolytic, 14.0 kg CO2e/kg; Qinghai China electrolytic byproduct, 5.3–8.5; Brazilian biomass silicothermic, 10.1; Canadian hydrometallurgical/electrolytic on a low-carbon grid, under 5), but none is how the world currently makes magnesium. Estes’ strongest figures depend on sourcing outside the Pidgeon-dominated mainstream via closed-loop scrap and a named partner (“Tidal Metals electrolytic, modeled”) claimed at 0.36 kg CO2e/kg, more than 10x better than even the best real-world low-carbon route documented, and explicitly labelled “modeled” / “Estes estimate,” not measured or audited. DLR’s aircraft case study found the production-emissions penalty for magnesium (any source) repaid within about five flights, since aviation’s use-phase energy intensity is so much higher than a passenger car’s; that is a stronger LCA case than the commercial-vehicle and marine segments where most of Magnus’s targeted volume sits.

Net Assessment

Magnesium’s physical case for this application is real: it is genuinely the lightest structural metal available, with a plausible route to large monocast structures that composite and steel cannot match and aluminium cannot match without a heavier casting. But every one of its historic disqualifiers, flammability, corrosion, and now the coolant-compatibility and dielectric-integrity questions raised in this report, sits on published claims rather than published test data at full pack scale, in the geometry that actually matters.

Composite is currently the more evidenced challenger on the weight-saving pitch specifically, with a named, shipping, third-party-tested precedent (SABIC/Honda) that magnesium cannot yet point to at any comparable scale. None of this means Estes’ approach won’t work; it means the evidence gap between “solved in principle” and “solved in a fielded pack” is still wide, and the gaps line up consistently with the three properties magnesium has always struggled with.

References

1. Estes Energy Solutions, Magnus™ Platform Spec Sheet, Rev. 2026.05 (accessed July 2026). https://estes.energy/product; https://estes.energy/technology
2. Estes Energy Solutions, “How Magnesium Redefines Battery System Architecture,” whitepaper v1, June 2026. https://files.estes.energy/estes_mg_thesis.pdf
3. Ehrenberger, S. (DLR, German Aerospace Center), “Carbon Footprint of Magnesium Production and its Use in Transport Applications,” IMA LCA Study update, October 2020. https://cdn.ymaws.com/www.intlmag.org/resource/resmgr/sustainability/2020-LCA-Study-2021-02-09.pdf
4. Automotive Manufacturing Solutions, “The evolution of EV battery enclosures: balancing optimisation, safety and cost-effectiveness,” 2026 (industry interviews: Novelis, ArcelorMittal, SABIC, Ducker Carlisle, Fraunhofer IWU). https://www.automotivemanufacturingsolutions.com/electrification/the-evolution-of-ev-battery-enclosures-balancing-optimisation-safety-and-cost-effectiveness/529348
5. Industry reporting on composite EV battery enclosure adoption and mass-saving figures (Stratview Research, BatteryTechOnline, ICT Material), 2026. https://www.stratviewresearch.com/3214/composite-EV-battery-enclosures-market.html; https://www.batterytechonline.com/trends/market-trends-propel-composite-ev-battery-enclosure-innovation
6. Krishtal, Yerokhin et al., “Plasma electrolytic oxidation of magnesium and its alloys: Mechanism, properties and applications,” ScienceDirect, 2017. https://www.sciencedirect.com/science/article/pii/S2213956717300075
7. “Plasma Electrolytic Oxidation (PEO) Process, Processing, Properties, and Applications,” PMC, 2021. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8224744/
8. Fattah-alhosseini et al., “Enhancing PEO Coatings on Magnesium Alloys: The Critical Role of Surface Pretreatments,” Advanced Materials Interfaces, 2025. https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202500192
9. Corrosion of magnesium (Mg) alloys in engine coolants, ScienceDirect. https://www.sciencedirect.com/science/article/pii/B9781845697082500111
10. A review on corrosion and corrosion inhibition behaviors of magnesium alloy in ethylene glycol aqueous solution, De Gruyter, 2023. https://www.degruyter.com/document/doi/10.1515/corrrev-2023-0020/html?lang=en
11. Royal Enfield / Flying Flea FF.S6 unveiling, EICMA 2025 (Electrek, 5 Nov 2025). https://electrek.co/2025/11/05/royal-enfield-unveils-flying-flea-s6-scrambler-style-electric-motorcycle-built-for-urban-adventure/
12. UL 2580, Standard for Batteries for Use in Electric Vehicles, dielectric voltage-withstand test methodology (GB/T 18384.3 applies the same convention). https://webstore.ansi.org/standards/ul/ul2580ed2020
13. Greco, L., “Estes Energy To Bring Large Magnesium Thixomolding To America,” The Gigacasting Newsletter / Industry Arsenal, 27 July 2026. https://www.industryarsenal.com/p/estes-energy-to-bring-large-magnesium-thixomolding-to-america
14. Grace, D. (Estes Energy Solutions, CEO), LinkedIn comment thread, August 2026, personal communication, direct correction of the dielectric-withstand interpretation in an earlier draft of this report.

See also: Aluminium