Aerozith: design of a new lightweight alloy
Fable 5.1 designed an aluminium-lithium alloy that could be 42 % lighter than titanium and beat Ti-6Al-4V in strength per weight. Status: unverified.
Status: Unverified. A model designed this alloy. Nobody has simulated, melted or tested it yet. All property values are estimates.
Aerozith is the design of an aluminium-lithium alloy with an estimated density of 2.6 g/cm³ and a yield strength of 600–700 MPa. That would make it about 42 % lighter than titanium, stronger than pure titanium, and better than Ti-6Al-4V in strength per weight.
The question
Design an alloy that does not exist yet, lighter and stronger than titanium, with a recipe for 100 g and a name.
"Stronger than titanium" depends on which titanium:
- Pure titanium (Grade 2): about 4.5 g/cm³, yield strength 350–450 MPa.
- Ti-6Al-4V (Grade 5, the aerospace standard): about 4.43 g/cm³, yield strength 830–900 MPa.
The targets were therefore: density well below 3 g/cm³, yield strength above pure titanium, and specific strength above Ti-6Al-4V. Beating Ti-6Al-4V in absolute strength was ruled out as unrealistic for a light-metal base.
Choosing the base
| Base | Density (g/cm³) | Yield strength (MPa) | Why not |
|---|---|---|---|
| Magnesium alloys | 1.7–1.9 | 150–350 | too soft, corrodes, hard to process |
| Beryllium | 1.85 | 300–400 | toxic, very expensive, brittle |
| Titanium aluminides (TiAl) | 3.8–4.0 | 400–600 | barely lighter than titanium, brittle at room temperature |
| Aluminium-lithium (chosen) | 2.5–2.7 | 500–700 | – |
Aerozith combines three known strengthening mechanisms that today's alloys use separately:
- Precipitation hardening like the 7xxx series (zinc, magnesium, copper), as in Al 7075.
- Nanoscale Al₃(Sc,Zr) precipitates that keep the grain ultra-fine, as in Scalmalloy.
- Lithium for lower density and higher stiffness, as in Al-Li alloys such as Al 2099.
Recipe for 100 g
Short formula: Al-5Zn-4Mg-3Li-2Cu-0.7Sc-0.3Zr-0.1Er (weight percent).
| Element | Amount (g) | Role |
|---|---|---|
| Aluminium (Al) | 84.9 | matrix, light, easy to work |
| Zinc (Zn) | 5.0 | precipitation hardening via MgZn₂ |
| Magnesium (Mg) | 4.0 | solid-solution hardening, partner of zinc, lowers density |
| Lithium (Li) | 3.0 | about 9 % lower density, about 15 % higher stiffness |
| Copper (Cu) | 2.0 | extra strength, better strength when warm |
| Scandium (Sc) | 0.7 | Al₃Sc nanoprecipitates, fine grain |
| Zirconium (Zr) | 0.3 | keeps the Al₃(Sc,Zr) particles from coarsening |
| Erbium (Er) | 0.1 | thermal stability, cheaper partial substitute for scandium |
Estimated properties
| Material | Density (g/cm³) | Yield strength (MPa) | Specific strength (kN·m/kg) | Young's modulus (GPa) |
|---|---|---|---|---|
| Aerozith (estimate) | 2.6 | 600–700 | 230–270 | 80–85 |
| Ti-6Al-4V | 4.43 | 830–900 | 190–200 | 114 |
| Pure titanium Grade 2 | 4.51 | 350–450 | 80–100 | 105 |
| Scalmalloy | 2.67 | 500–520 | 190–195 | 70 |
| Al 7075-T6 | 2.81 | 500 | 180 | 72 |
Comparison values are typical datasheet values (approximate). The model's own error estimate is ±150 MPa for yield strength and ±0.05 g/cm³ for density.
How it was made
- Model: Claude Fable 5.1 (Anthropic), in the Claude app.
- Effort: three chat messages, about 4 minutes, on 20 October 2026.
- Tools: none. No web search, no literature search, no simulation. Only the model's trained knowledge of metallurgy.
Risks
| Risk | Effect | Rating |
|---|---|---|
| Lithium oxidises in the melt | melting only under argon or vacuum, high cost | high |
| Hot cracking during casting | ingots crack at 5 % Zn + 3 % Li + 2 % Cu | high |
| Stress corrosion cracking | known weakness of 7xxx alloys, possibly worse with lithium | medium |
| Unwanted phases (e.g. Al₂CuLi) | brittleness, lower toughness | medium |
| Scandium cost | 0.7 % Sc makes it more expensive than titanium | medium |
| Lower strength than estimated | the mechanisms add up less than assumed | medium |
Review by a second model
Step 3 of our method: a second model looks for errors. This review was done by Claude Opus 5.5, also from trained knowledge only, without a literature search. Its numbers are approximate.
- 3 % lithium is a lot. Most commercial Al-Li alloys contain about 1–2.7 % lithium. More lithium often makes the alloy brittle.
- About 15 % alloying elements in total is very high. At that level, coarse and brittle phases form easily.
- 600–700 MPa is optimistic. The best Al-Li alloys in use today reach roughly 500–600 MPa.
- Novelty is uncertain. Al-Mg-Li alloys already exist (for example the Russian alloy 1420), and Al-Zn-Mg-Cu-Li alloys have been studied in research. Only a literature search can show whether this exact composition is new.
- The density of about 2.6 g/cm³ is plausible.
Verdict: an interesting starting point, but the estimated strength should be read as an upper bound. The review does not change the status.
Next steps
- Literature search on Al-Zn-Mg-Li and Al-Li-Sc systems, and a patent check.
- CALPHAD calculation (Thermo-Calc or Pandat) of phase stability at 20–200 °C.
- Lab melt of 100–500 g under argon, tensile test after T6 heat treatment.
- If it works: test it for laser powder-bed 3D printing, where Scalmalloy is used today.
Until steps 2 and 3 are done, Aerozith stays at status 1: unverified.
Links