Most of this material list cannot answer the question
The stiffness of a part is set by the elastic modulus E, and its weight by the density. Divide one by the other and the ratio answers exactly one question: for a given stiffness, which material weighs least. The unit is MJ/kg, and it is the same arithmetic as E in GPa divided by density in g/cm³ — a gigapascal is a joule per cubic millimetre and a gram per cubic centimetre is a tonne per cubic metre, so the two conversions cancel.
Every material on this site has a density. Only 33 of the 48 records also have a modulus:
| Family | Records | With a modulus |
|---|---|---|
| Metals | 11 | 6 |
| Plastics | 1 | none |
| Rubber | 5 | none |
| Composites | 3 | 3 |
| Wood | 28 | 24 |
The gaps are not missing data entry, they are gaps in what the row is. The metal rows without a modulus are family entries — a family has no single alloy, so it has no single modulus, and the grade rows underneath it carry the number. The engineered wood products are product classes rather than species: the number belongs to the mill's own sheet, and this database does not hold per-product panels. Rubber has no Young's modulus in the sense the other rows use, because its stress–strain curve bends from the first percent — any single number would need the strain it was read at attached to it, and what this database carries for rubber instead is a hardness and a temperature window. Nylon appears with a density and no modulus, and no number is invented here to fill the cell.
Divide by density and the six metals land in one band
| Metal | E (GPa) | Density (g/cm³) | E/ρ (MJ/kg) |
|---|---|---|---|
| Stainless Steel A4 (316) | 193.0 | 8.00 | 24.13 |
| Stainless Steel A2 (304) | 193.0 | 7.90 | 24.43 |
| Carbon Steel 1018 | 200.0 | 7.87 | 25.41 |
| Carbon Steel A36 | 200.0 | 7.85 | 25.48 |
| Aluminium 6061-T6 | 68.9 | 2.70 | 25.52 |
| Aluminium 6063-T5 | 69.0 | 2.70 | 25.56 |
E runs from 68.9 to 200.0 GPa — a factor of 2.9. Density runs from 2.70 to 8.00 g/cm³ — a factor of 3.0. Divide one by the other and both factors cancel: 24.13 to 25.56 MJ/kg, which is ±2.9% about 24.8. The whole spread of the ratio, 1.1 times, is smaller than the spread of either number that went into it.
Aluminium against carbon steel is the case worth holding onto. E is 2.90 times lower, density is 2.91 times lower, and the ratio is therefore the same to within 0.4%. Swapping steel for aluminium in anything limited by stiffness and by mass buys a fraction of a percent — the weight saving in an aluminium part comes from the strength and the shape, not from the stiffness.
That is not an accident of these six records. All of them are metals, which means all of them share metallic bonding, and both E and density measure the same thing from two sides: how tightly the atoms are packed and how stiff the bonds between them are. Move to another bond type and the ratio moves, which is what the last two sections are about.
The division is legal because E is read before anything yields
Three columns on these pages look interchangeable and are not. E is the slope of the first straight part of the stress–strain line — a small-strain elastic constant. Modulus of rupture is a bending stress at which a specimen breaks. Yield strength is a 0.2% offset, and tensile strength is the maximum. The last three are all measured by breaking something, by three different tests, and putting them in one column would be a category error; the wood article is where that distinction is worked out in full.
E is the one that survives the crossing. A modulus read in bending and a modulus read in tension are the same number for the same material, because both are read while the material is still behaving elastically. So a wood figure tabulated as a static-bending MOE on small clear specimens at 12% moisture and a metal figure tabulated as a handbook modulus are the same kind of quantity. They are not measured by the same test, and how much the two tests differ is not in this database — but it cannot be what separates the rows below, which differ by up to a factor of 2.0, and this page will not put a number on it.
One thing that does not cross: the minimum/typical basis printed on every metal page describes the strength columns. No standard guarantees a modulus — it is a physical constant, so every entry here carries it as a typical value, and the 6 grades above sit in 5 different supply conditions with the ratio still in one band. Nothing in the table mixes a guaranteed number with an average one.
The density to divide by is not the specific gravity on the page
Each wood page prints two different numbers that both look like density. The one at the top is the working density at 12% moisture, which is what a piece of timber weighs. The one further down is the USDA specific gravity, defined as oven-dry weight over volume at 12% moisture — so it sits below the working density, and the page says so in its own note.
The modulus is a 12%-moisture value, so the density that goes with it is the 12% one. Divide by the specific gravity instead and every wood in the table comes out higher by a median of 11.4% — which is about half the total spread of the whole wood family. A reader who computes the ratio from the number that happens to be labelled specific gravity gets a wood ranking that is internally consistent and comparable with nothing on this page.
Twenty-four woods spread over a factor of two
| Species | E (GPa) | Density (g/cm³) | E/ρ (MJ/kg) |
|---|---|---|---|
| Poplar (Yellow) | 10.9 | 0.42 | 25.9 |
| Fir (Douglas) | 13.4 | 0.53 | 25.4 |
| Spruce (Sitka) | 10.8 | 0.43 | 25.2 |
| Basswood | 10.1 | 0.42 | 24.0 |
| Hemlock (Western) | 11.2 | 0.47 | 23.9 |
| Pine (Southern Yellow) | 12.3 | 0.55 | 22.4 |
| Larch | 12.9 | 0.58 | 22.2 |
| Redwood | 9.2 | 0.42 | 22.0 |
| Ipe | 21.6 | 1.00 | 21.6 |
| Balsa | 3.4 | 0.16 | 21.1 |
| Cedar (Western Red) | 7.7 | 0.37 | 20.7 |
| Cypress (Bald) | 9.9 | 0.48 | 20.7 |
| Birch (Yellow Birch) | 13.9 | 0.69 | 20.1 |
| Mahogany (Honduran) | 10.3 | 0.55 | 18.8 |
| Cherry (Black) | 10.3 | 0.56 | 18.3 |
| Hickory | 14.9 | 0.82 | 18.2 |
| Ash (White) | 12.0 | 0.67 | 17.9 |
| Maple (Hard Maple) | 12.6 | 0.70 | 17.9 |
| Walnut (Black) | 11.6 | 0.66 | 17.6 |
| Teak | 10.7 | 0.63 | 17.0 |
| Beech (American) | 11.9 | 0.72 | 16.5 |
| Oak | 12.3 | 0.75 | 16.4 |
| Rosewood (East Indian) | 12.3 | 0.85 | 14.4 |
| Cedar (Eastern Red) | 6.1 | 0.47 | 12.9 |
Best to worst is a factor of 2.0 — 25.9 MJ/kg for poplar (yellow) against 12.9 for cedar (eastern red), with a median of 20.4. That is wider than the 1.1 times the metals span, and it is the price of the same rule: the metals are all one bond type, and twenty-four species of tree are twenty-four different cell structures.
The lightest species in the table is balsa at 0.16 g/cm³, and it is not the least stiff per kilogram: at 21.1 MJ/kg it sits above the median and well above cedar (eastern red). What makes a light wood light is the empty space inside the cells, and empty space carries neither load nor mass, so E and density fall together and the ratio hardly notices. A wood is a bundle of tubes with the stiffness coming from the walls; take tubes away and you take both numbers away.
The only way out is a direction
| Composite | Along fibre E (GPa) | Along fibre E/ρ | Across fibre E (GPa) | Across fibre E/ρ |
|---|---|---|---|---|
| Glass Fibre Composite (GFRP) | 39.0 | 18.6 | 8.6 | 4.1 |
| Carbon Fibre Composite (CFRP) | 177.0 | 110.6 | 10.8 | 6.8 |
| Aramid Fibre Composite (AFRP) | 87.0 | 63.0 | 5.5 | 4.0 |
Along the fibre, carbon composite is 4.4 times steel on this index. Across the fibre the same laminate is 6.8 MJ/kg — lower than every metal and every wood in this database, and lower than cedar (eastern red). Nothing changed between those two rows except which way the load points, and the composites article is where that pair of numbers gets its full treatment.
That is the whole exception. The index is roughly constant because stiffness and mass both come from how much material is packed into a volume — align long covalent chains in one direction and you get stiffness without density, so the ratio rises to 110.6 for carbon, 63.0 for aramid and 18.6 for glass. Point the load the other way and you are loading the matrix, which is a polymer, so the ratio falls to a polymer's.
The index is only as good as the member you name
The number in the tables above is the mass of an equally stiff tie rod: a member loaded along its axis, with the cross-section free to be whatever the load needs. Bolts are tie rods, which is why this index is the one to quote for a fastener. It is not the index for a beam, and the reason is algebra, not opinion — for a beam of a given stiffness, the section has to grow as E falls, so mass goes as ρ divided by a fractional power of E. Which power depends on what the drawing lets you change:
- Tie rod, area free: mass ∝ ρ / E, so the index is E / ρ.
- Beam, section scaled in proportion: the second moment goes as area squared, mass ∝ ρ / √E.
- Beam, width fixed and depth free: the second moment goes as area cubed, mass ∝ ρ / ∛E.
| Material | E/ρ (MJ/kg) | Tie rod | Beam, similar section | Beam, depth free |
|---|---|---|---|---|
| Carbon Steel 1018 | 25.4 | 1.00 | 1.00 | 1.00 |
| Stainless Steel A2 (304) | 24.4 | 1.04 | 1.02 | 1.02 |
| Aluminium 6061-T6 | 25.5 | 1.00 | 0.58 | 0.49 |
| Oak | 16.4 | 1.55 | 0.38 | 0.24 |
| Spruce (Sitka) | 25.2 | 1.01 | 0.23 | 0.14 |
| Balsa | 21.1 | 1.20 | 0.16 | 0.08 |
| Carbon Fibre Composite (CFRP) | 110.6 | 0.23 | 0.22 | 0.21 |
Read the aluminium row twice. In a tie it is 1.00 — 0.4% lighter than steel and nothing more. In a beam whose depth is free it is 0.49, about half. Same two numbers from the first table, a different member, and the answer changes from rounding to real.
Oak goes the other way. As a tie rod it is 1.55 times the mass of steel; as a beam with free depth it is 0.24. Balsa is 1.20 as a tie and 0.08 as a beam — a factor of 15 between two answers that both come from the same pair of material numbers. Quoting a specific stiffness without saying which member it is for is not a simplification; it is a different question being answered.
Where this stops
- Strength. Nothing here says whether the part survives. A material can win the stiffness index and fail at a quarter of the load — the grade article is where strength gets its numbers, and the two indices pick different winners.
- Cost and availability. Not a column in this database. The best ratio on this page is a unidirectional laminate, and what it costs to buy is not something these files can tell you.
- The section has to be allowed to grow. Two of the three beam indices above assume the drawing lets the depth increase. In a fixed envelope the section cannot grow, the index does not apply, and the comparison you need is a different one.
- Moisture and temperature. The wood numbers are 12% moisture values and nothing here says what wet timber does. The modulus of a polymer or a rubber moves with temperature well before it moves with load.
- One direction only. Every composite figure on this page is a unidirectional laminate along the fibre. A woven or quasi-isotropic plate of the same fibre is a different material with different numbers.
The materials library is the way into all 48 records, and each page prints the source of its modulus and its density so the division above can be redone by hand.