Aluminium versus steel: choosing the right metal

Choosing between aluminium and steel is an engineering decision, not a headline. Steel is about three times as stiff, so it deflects and buckles less for the same section, while aluminium weighs 0.344 of what steel weighs and carries roughly twice the strength per kilogram on a yield basis. A stronger grade changes neither of those facts. The metal that wins depends on the load case, the joint and the environment, and this guide sets out each comparison with numbers so the choice can be made on the design rather than on a slogan.

Stiffness and strength are two different properties

The most common mistake in comparing the two metals is to treat stiffness and strength as one thing. Stiffness is governed by the elastic modulus, and steel is fixed at 210 000 N/mm2 against about 70 000 N/mm2 for aluminium, a ratio close to three. Strength is a separate property, set by the yield or proof stress, and there the picture reverses per unit of weight. A component can be strong and flexible, or stiff and easily overloaded, so the two questions have to be answered separately before a metal is chosen.

The modulus does not depend on the grade

A point that surprises many buyers is that selecting a stronger grade does nothing for stiffness. The Eurocode for steel gives a single modulus of 210 000 N/mm2 for every structural grade from S235 to S460, and separates the grades only by yield and tensile strength. The Eurocode for aluminium does the same, quoting one modulus of 70 000 N/mm2 for all structural alloys and distinguishing 6060-T6 from 6082-T6 by strength, not stiffness. So a move from S235JR to S355J2, or from 6060 to 6082, raises the load a member can carry but leaves its deflection unchanged, because deflection depends on the modulus and the section, and neither of those has moved. Austenitic stainless steel is a partial exception at 200 000 N/mm2, with ferritic grades near 220 000. The grades are set out in our guides to structural steel S235JR, S275JR and S355J2 and EN AW aluminium alloys.

Deflection: aluminium bends three times as much, except under its own weight

For an identical section under the same external load, an aluminium beam deflects three times as far as a steel one, directly because its modulus is a third. That figure, however, applies to an applied load. Deflection under a member’s own weight is not three times greater, because aluminium weighs only 0.344 of what steel weighs, so the two effects almost cancel and the self-weight deflection ratio comes out at about 1.03. An aluminium beam sags under its own mass by practically the same amount as a steel one, which is why aluminium is used for long lightly loaded spans such as walkways and access structures where self-weight dominates.

Buckling: the stronger alloy can perform worse

Buckling follows the modulus, not the strength, so an aluminium column of identical geometry reaches a critical load one third of the steel equivalent, and a stronger alloy does not help. The effect is sharper than it first appears. The reference slenderness that marks the transition to buckling failure is 93.9 for S235 and 76.4 for S355, but 70.2 for 6060-T6 and only 52.6 for 6082-T6. Because relative slenderness is the geometric slenderness divided by that reference value, the same column in the stronger alloy comes out at a higher relative slenderness and loses more of its strength to buckling than the weaker one. Choosing a high-strength alloy to resist buckling is one of the least effective moves an engineer can make.

Strength-to-weight: where aluminium genuinely wins

On strength per unit of weight, aluminium is ahead, but the basis has to be stated or the figure is meaningless. On a yield basis, S355 carries 45.2 kN.m/kg against 92.6 for 6082-T6, an advantage of 2.05 times. On tensile strength the advantage is narrower, around 1.8 to 1.9 times, because steel closes the gap on ultimate strength. The density ratio itself is 2.9 to 1. In bending the advantage grows rather than shrinks, reaching about 2.3 to 2.4 times, because the geometry rewards the lighter metal further when a section is sized for a bending moment. Quoting a single strength-to-weight number without naming the basis is the fastest way to mislead a technical reader.

Equal stiffness in bending: the mass saving is real, with conditions

When a beam is sized for stiffness rather than strength, aluminium can save about half the weight, and the arithmetic is worth showing because the result looks too good otherwise. To match a steel beam’s stiffness, an aluminium beam of the same width needs its depth increased by a factor of 1.44, since second moment of area grows with the cube of depth. That heavier section still ends up at 0.496 of the steel mass, a saving near 50 per cent. The stress check passes with room to spare: the section modulus rises by 2.08 times, so the bending stress falls to about 113 N/mm2, well within 6060-T6 and 6082-T6.

This 50 per cent saving holds only under stated conditions, and each one can erode it. It assumes a solid rectangular section of constant width, and it assumes stiffness governs the design rather than ultimate strength. A deeper thin-walled section may cross into a slender class where local buckling removes part of the gain, and a taller narrower beam is more prone to lateral-torsional buckling. The largest threat is welding, covered below. Note the contrast with pure tension, where the stiffness per unit weight of the two metals is almost identical, 26.8 for steel against 25.9 for aluminium. Aluminium saves weight in bending, not in an axially loaded tie, and confusing the two load cases produces contradictory advice.

The weld takes the advantage away

The single fact that decides many welded designs is that aluminium loses much of its strength beside a weld. For 6082-T6, the proof stress falls from 260 N/mm2 in the parent metal to 125 N/mm2 in the heat-affected zone, a reduction factor near 0.48. That collapses the strength-to-weight comparison: welded 6082-T6 reaches 46.3 kN.m/kg against 45.2 for S355, so the 2.05 times advantage shrinks to 1.02 times, essentially level. Carbon steel grades such as S235 and S355 carry no comparable softening, so for a welded strength-critical structure the aluminium case weakens sharply. The extent depends on temper, not just alloy family: the same 6082 in the T4 condition loses only about 10 per cent, and a 5083 sheet in the O condition loses nothing at all. Design detail for welded aluminium is covered in our guide to joining aluminium profiles.

The strongest metal has no single answer

The question of the strongest metal has no single winner, because it depends on which property is meant. On yield strength, maraging steel leads at roughly 2400 N/mm2. On tensile strength, cold-drawn pearlitic wire reaches about 2940 N/mm2. On scratch hardness, chromium tops the metals at Mohs 8.5, while on indentation hardness tungsten and osmium lead at roughly 350 to 470 HV. On strength per unit of weight, fine steel wire wins, followed by titanium alloys near 205 kN.m/kg. Tungsten carbide, often named here, is not a metal at all but a cermet, and tungsten metal, though it has the highest melting point of any metal at 3422 degrees C, is brittle at room temperature and cannot be used structurally. Five criteria, five winners, and no honest one-word answer.

Fatigue: the Eurocodes do not draw two different curves

A widespread claim holds that steel has a fatigue limit and aluminium does not, so aluminium always fails eventually under cyclic load. That is not how the design codes treat it. The Eurocode fatigue rules for steel and for aluminium share the same construction, both defining a constant-amplitude fatigue limit at five million cycles and a cut-off at 100 million cycles, and both hold the curve flat above that point for the two metals alike. Classical tests to ten million cycles do show a knee in the steel curve that aluminium lacks, but very-high-cycle testing finds steel failing below its supposed limit too, so the difference is not the clean line it is often drawn as. The real difference is the level of the detail category: welded aluminium details sit around 12 to 56 N/mm2 against 36 to 112 N/mm2 for steel, roughly an order lower, which is what matters for a cyclically loaded welded joint.

Thermal expansion and movement

Aluminium expands more than steel with temperature, at 23 against 12 millionths per degree, a ratio of 1.92 to 1 rather than the two-to-one often quoted. Over a 6 m length and a 50 degree rise, steel grows 3.6 mm and aluminium 6.9 mm, a difference of 3.3 mm that expansion joints and sliding fixings have to absorb in facades, roofs and long runs. Steel is not a single number here either: austenitic stainless expands at about 16 millionths per degree, giving 4.8 mm over the same length, while ferritic grades sit close to carbon steel. Treating stainless as if it moved like carbon steel is a common and avoidable error in mixed assemblies.

Galvanic corrosion when the two metals meet

Where aluminium and steel are joined in the presence of moisture, aluminium is the less noble metal and corrodes, and the geometry of the joint decides how badly. The total metal loss is set by the area of the cathode, not the anode, so a large steel surface bolted with a small aluminium fastener drives severe, concentrated attack on that fastener, whereas small steel fasteners in a large aluminium sheet are acceptable because the aluminium loss is spread thin. The rule that follows is that fasteners should be made of the more noble material. This is why the guidance to protect the joint targets the cathode surface: coating the more noble metal removes the driving surface, while a scratch on the anode would only expose a small, fast-corroding area. Galvanised steel is a safe companion to aluminium because the zinc is anodic to both and sacrifices itself in the right direction, a mechanism explained in our guide to galvanised steel DX51D. Ulamex manufactures to EN 1090-3, which for aluminium in contact with other metals requires the faying surfaces to be treated, so in practice both surfaces are protected and the crevice sealed. The corrosion behaviour of aluminium on its own is covered in does aluminium corrode.

Low temperature: crystal structure decides

In the cold, the two metals behave in fundamentally different ways because of their crystal structure. Aluminium has a face-centred cubic lattice and shows no ductile-to-brittle transition, staying tough down to cryogenic temperatures, with 5083 in the O condition gaining about 40 per cent in tensile strength at minus 195 degrees C. Ferritic and carbon steels have a body-centred cubic lattice and can turn brittle as temperature falls, which is why steel is specified by impact grade: quality JR is tested at plus 20 degrees C, J0 at zero, J2 at minus 20, and K2 at minus 20 to a higher energy. The stainless design code divides its grades by crystal structure for exactly this reason, treating austenitic and duplex stainless as adequately tough for service down to minus 40 degrees C. For low-temperature duty the choice is less aluminium against steel than the right family within each.

Fire and elevated temperature

Steel keeps its strength to a far higher temperature than aluminium, but neither metal is safe in a fire without thought. Structural steel retains full strength to about 400 degrees C and falls to half its strength near 600 degrees C, whereas 6082-T6 aluminium is down to half at about 228 degrees C, with 5083 in the O condition the most heat-tolerant aluminium at about 286 degrees C. The design codes reflect this: an aluminium member is assumed to fail at 170 degrees C, against roughly 500 to 700 degrees C for steel depending on how heavily it is loaded. That advantage is real, yet unprotected steel is credited with only about 15 minutes of fire resistance against requirements of 60 to 120 minutes, so protecting steel is the norm rather than the exception. Bare aluminium has a low surface emissivity of 0.3 against 0.7 for steel, which can make a clean aluminium surface heat more slowly than steel in a radiant fire, though paint or soot removes that benefit entirely. Melting point is rhetorically striking and structurally irrelevant: pure aluminium melts at 660.3 degrees C and pure iron at 1538 degrees C, but steel has already lost its useful strength some 900 degrees below its melting point, so the reduction-factor tables, not the melting point, decide the design.

Cost and availability sit on top of the engineering

Once the engineering has narrowed the choice, commercial factors decide between the survivors, and they follow a logic of their own rather than a single price. Steel and aluminium are priced from different raw materials and energy inputs, aluminium is quoted as an exchange price plus regional and conversion premiums, and stainless carries a separate alloy surcharge, so the two metals rarely move together in the market. The layers that build a figure, from grade and form to coating, certification and freight, are set out in our guide to steel and aluminium price drivers, and the mass that drives both freight and handling, where the density difference of 2.9 to 1 between the two metals shows up directly, can be worked out from the metal weight guide. A sound choice weighs the whole-life cost, including coating and maintenance, not the metal figure alone.

A framework for choosing

The decision resolves to a handful of questions, and the table below sets out where each metal has the stronger case across 7 common requirements. Aluminium earns its place where weight is critical, where a component must resist corrosion without a coating, where non-magnetic or cryogenic behaviour is needed, or where a moving part benefits from low mass. Steel earns its place where stiffness or buckling governs, where a welded joint must carry its full strength, where fire performance matters, and where the lowest cost per unit of strength is decisive. In most cases the answer follows directly from which of these requirements dominates.

Both metals from one supplier

Most real projects need a mix, and sourcing carbon steel, stainless and aluminium against one specification avoids the gaps that appear when families are bought separately. Ulamex has traded steel and aluminium since 1988, bringing 38 years of experience across all three families, and manufactures its own structures under EN 1090-1 EXC2 and ISO 3834-2. Material is supplied against the grade, form, dimensions, certificate type and delivery point on the enquiry, with a lead time of 2 to 7 working days on stocked items. The full range is set out in our steel and aluminium supply capabilities.

Send the grades, forms, dimensions and quantities your project calls for, and Ulamex will return a non-binding quotation. Contact the export desk at [email protected] or +48 504 424 761, or write to Zawada 144, 97-200 Tomaszow Mazowiecki, Poland.

Frequently asked questions