How stiff are engineering materials? Modulus from plastic to tungsten

Which common engineering material is the stiffest?

Among the materials compared here, the stiffest are ceramics and tungsten. The best-documented values are silicon carbide (395 and 410 GPa for sintered grades, 421 GPa for a polycrystalline entry), Tungsten (W) at 411 GPa, Alumina (Al2O3) at 370 GPa and silicon nitride at 180 to 320 GPa. The highest single number, tungsten carbide at 530 to 700 GPa, comes from a compound page for pure WC with no cobalt binder, so it is not a figure for a cutting-tool or mining grade, and the profile holds no modulus for a cobalt-bound grade.

Among structural metals the gap between families is large. Tungsten (W) is 2 times as stiff as 1018 steel, which has 190 to 210 GPa across its entries. 316 stainless (190 to 205 GPa), Ti-6Al-4V (114 GPa) and Copper (Cu) (117 GPa) follow.

How stiff are plastics, and how do light metals, glass and fibres compare?

Steel is 3 times as stiff as 6061 aluminium at 68.9 GPa, and aluminium is 1.5 times as stiff as AZ31B magnesium. Fused silica shows 72.7 GPa, 1.1 times the aluminium value, and borosilicate glass is printed at 64 kN/mm2, which is numerically equal to GPa (general engineering knowledge, not in our data). Kevlar para-aramid fibre yarn and single-fibre tests show 70.5 to 112 GPa.

Plastics are far less stiff, and their datasheets often print the value in MPa where metals use GPa. PEEK (polyetheretherketone) is printed at 4000 MPa, which is 4 GPa, for the Victrex 450G grade at 23 °C. ABS shows 2400 MPa for the SABIC MG94F grade, Polypropylene (PP) shows 1150 MPa for a random copolymer, and polycarbonate shows 2400 MPa in its tensile test. Convert everything to one unit before ranking.

Moisture moves these numbers more than it moves metals. The PA-66 profile holds two BASF grades, Ultramid A3K Q601 (dry) with a tensile modulus of 3 GPa and Q603 (conditioned) at 1.2 GPa. They carry different grade codes, so they are not strictly one resin. The profile gives the Q601 moisture uptake at equilibrium as 2.8 % at 50 % relative humidity; the moisture content of Q603 is not stated.

General engineering knowledge, not in our data: plastics creep, meaning they keep deforming under a steady load, so a short-test modulus overstates long-term stiffness, and they soften near the glass transition, which the PEEK profile puts at a midpoint of 150 °C. Glass-filled grades are much stiffer; the PA-66 profile lists a 33 % glass-filled Vydyne R533H grade but holds no modulus for it. An ultrasonic (dynamic) modulus also differs from a tensile (static) one.

Is stiffness the same as strength?

No. Stiffness is how far a part deflects under load while it springs back, and strength is the load at which it deforms for good or breaks. The two are separate properties (general engineering knowledge, not in our data), and our data shows it. Ti-6Al-4V has a modulus of 114 GPa and Copper (Cu) has 117 GPa, nearly the same stiffness, but the highest titanium yield strength is 15.9 times the soft-temper copper value.

For a buyer the practical point is that a stronger grade does not make a stiffer part. General engineering knowledge, not in our data: steels all sit near the same modulus and aluminium alloys near theirs, whatever their strength. Our data fits this. 1018 steel shows a yield strength of 240 to 370 MPa depending on whether the bar is cold drawn or hot rolled, while its modulus stays in 190 to 210 GPa. 316 stainless has almost the same modulus range, 190 to 205 GPa, but a yield strength of 205 MPa in the datasheet and product-standard rows, with an annealed plate measured at 295 K giving 193 to 213 MPa.

Ceramics and glasses are stiff but flaw-sensitive. The issue is low toughness more than low strength: Alumina (Al2O3) has a modulus of 370 GPa and a fracture toughness of 4 to 5 MPa·√m, and its profile notes a tensile strength of 262 MPa, lower than its flexural strength of 379 MPa. Borosilicate glass shows flexural strengths of 7 to 25 MPa, the low end being the design value for safety use, and the fused silica profile notes that a damaged window pane can fail suddenly under tension.

A fibre can show the opposite of a stiff ceramic. Kevlar para-aramid fibre yarn and single-fibre tests give a tensile strength of 2,920 to 3,040 MPa at a modulus of 70.5 to 112 GPa, while 1018 steel bar shows 430 to 440 MPa at 190 to 210 GPa. A fibre and a bar are not like-for-like, but the fibre has much higher strength at a lower modulus.

Which stiff materials are also light?

It depends on how the part is loaded. For pulling and compression, the density chart shows steel, titanium, aluminium and magnesium along nearly the same line from the origin. Stiffness per unit of density, taken as the largest modulus over the largest density, is 27 GPa per g/cm³ for 1018 steel, 26 GPa per g/cm³ for Ti-6Al-4V, 26 GPa per g/cm³ for 6061 and 25 GPa per g/cm³ for AZ31B. The ceramics are far higher: 133 GPa per g/cm³ for the polycrystalline silicon carbide entry and 95 GPa per g/cm³ for AD-995 Alumina (Al2O3). Tungsten (W) gives 21 GPa per g/cm³, and Kevlar para-aramid fibre fibre gives 78 GPa per g/cm³.

For bending of a beam or panel the answer changes. General engineering knowledge, not in our data: the bending stiffness of a part scales with its thickness cubed, so a thicker section of a lighter metal can match a thinner steel section at lower weight. That is why 6061, at 68.9 GPa and 2.7 g/cm³, can beat steel for a stiff panel. AZ31B is the lightest structural metal listed here at 1.8 g/cm³ and the least stiff.

Composites are missing from the chart on purpose. The CFRP profile gives 1.3 to 1.8 g/cm³ as a resin density and a fibre density but no laminate density, so we cannot place it. General engineering knowledge, not in our data: a laminate loaded along its fibres is a stiff and light option, and one loaded across them is not.

How do I read a modulus number: tensile or flexural, and under which conditions?

Check four things: the loading mode, the grade, the direction and the conditions. The tensile modulus comes from pulling a bar and the flexural modulus from bending one, and for plastics the two give close but not identical numbers. Our data shows small gaps within a grade: dry PA-66 has 3 GPa tensile and 2.9 GPa flexural, PEEK gives 4000 MPa tensile and, in a note, 3800 MPa flexural (ISO 178, 23 °C), and polycarbonate gives 2400 MPa tensile at 1 mm/min and 2350 MPa flexural at 2 mm/min for one Makrolon grade. Do not build a pair from different grades. 6061 lists 68.9 GPa as the average of tension and compression, with compression about 2 % greater, and Ti-6Al-4V is labelled as a tension value.

Grade and process change the number. Silicon nitride runs 180 to 320 GPa: the lowest, 180 GPa, is a reaction-bonded grade with a density of 2.53 g/cm3, and the highest, 320 GPa, is a hot isostatically pressed grade. Silicon carbide shows 395 to 421 GPa across a NIST ultrasonic measurement, a polycrystalline entry and a direct-sintered grade, plus a generic reaction-bonded entry at 370 GPa with no named source.

Direction and test form decide for composites and fibres. CFRP gives 11.7 to 230 GPa: 11.7 GPa across the fibres of a unidirectional laminate, 135 GPa for a T700S fibre-in-resin composite at 60 % fibre volume, 164 GPa along the fibres of an IM7 laminate and 230 GPa for the bare T700S fibre, which is not a laminate. Kevlar para-aramid fibre shows 70.5 to 112 GPa between standard-modulus yarn, high-modulus yarn and a single fibre.

Conditions are often missing. The Tungsten (W) profile notes that the source gives no temperature for modulus, and Copper (Cu) lists its modulus at 20 °C. General engineering knowledge, not in our data: modulus falls as temperature rises, more in polymers than in metals, and porosity lowers it in ceramics. Units add one more trap: PEEK and ABS are printed in MPa, PA-66 in both GPa and MPa and borosilicate glass in kN/mm2.

Terms used

Young's modulus (elastic modulus)

the stress a material carries per unit of strain, meaning per unit of fractional stretch, while it still springs back. A higher number means a stiffer material. Datasheets print it in GPa, MPa, psi or kN/mm2, so convert before comparing.

Stiffness of a part

how little a finished part deflects under a load. It depends on the modulus of the material and on the shape of the part, such as its thickness and cross section.

Yield strength

the stress where a metal or plastic starts to deform for good. It is a strength, not a stiffness, and it can differ a lot between two materials with the same modulus.

Tensile modulus

the modulus found by pulling a test bar and reading stress against strain in the early, nearly linear part of the curve. Plastics have no truly straight part, so standards read a slope between two set strain points.

Flexural modulus

the modulus found by bending a bar on two supports and measuring load against deflection. It can differ a little from the tensile value for plastics.

Anisotropic

having different properties in different directions. Fibre composites such as CFRP (carbon-fibre reinforced epoxy) are strongly anisotropic, so one material has several moduli.

Conditioned

a polyamide sample that has taken up moisture from air before testing. Conditioned and dry nylon give different moduli, so the state must be printed with the number.

Brittle

breaking with little or no permanent deformation. Ceramics and glasses are stiff but brittle, so they crack under tension or impact instead of bending.

What to check before you order

1. Ask for the test method behind the modulus (tensile, flexural, compression or ultrasonic), the test standard and speed, and the temperature of the test against your part's range.
2. Ask for the measured modulus with its tolerance, not only a typical value. The NIST silicon carbide entry prints 395 GPa with a tolerance of plus or minus 10 GPa.
3. For polyamides, ask whether the value is dry-as-moulded or conditioned at a stated relative humidity. For any plastic, ask about creep under your steady load.
4. For fibre composites, ask for the modulus in the fibre direction and across it, the fibre volume fraction and the layup. A fibre value is not a laminate value.
5. For ceramics, ask for the grade, process route, density or porosity of that exact batch, and ask about price and machining. We hold no price data. The silicon carbide profile cites H350i (may cause cancer by inhalation) for the powder, so read the supplier's safety data sheet before machining.
6. For cemented carbide, ask for the cobalt content and grain size with the modulus.
7. Check whether a number is typical, minimum or an estimate, and convert all units to GPa before comparing.

What do we not know yet?

• The only tungsten carbide modulus in our data is a pure-compound figure from an aggregator page with no primary reference. We hold no modulus for cobalt-bound cemented carbide grades, and no effect of binder content or grain size.
• PEEK, ABS, polypropylene and PA-66 record tensile or flexural modulus under names without a standard unit, and the polycarbonate profile mixes tensile and flexural tests under one name. The PEEK flexural modulus of 3800 MPa sits in a note, not in a property row, and the ABS and polypropylene tensile and flexural values come from different grades, so there is no matched pair for either.
• No profile holds a stiffness-per-mass benchmark; the per-density figures in the density section are the largest modulus divided by the largest density of each profile, so they mix entries for some ceramics. CFRP has no laminate density. We hold no modulus for glass-filled plastics, creep, or dynamic against static testing.
• Temperature and humidity dependence of modulus is not in these profiles apart from the dry and conditioned polyamide pair. Test method is printed for few moduli, for example the NIST ultrasonic value for silicon carbide.
• The borosilicate glass modulus is stored under a non-standard property name in kN/mm2, so it is not in the charts. AZ31B magnesium has a modulus from one supplier sheet listing and no yield strength, so it is absent from the strength chart.
• We hold no cost or price data for the ceramics, and no machining data beyond the silicon carbide dust warning. Materials outside this set, such as graphene or lithium, are not compared.

Questions people ask