Tungsten and tantalum compared with lead, gold and other metals

At a glance

Which of these metals are both dense and high-melting?

Two: tungsten and tantalum. Our test is a density above 15 g/cm³ and a melting point above 2,000 °C, the usual cut-off for a refractory metal (general engineering knowledge, not in our data). Both lines are ours, not a standard. Tungsten (19.3 g/cm³, 3,422 °C) and tantalum (16.6 g/cm³, 3,017 °C) pass both. Molybdenum, niobium and rhenium are also refractory but are not in our data.

None of the others passes both. Gold is the only one that is dense: it matches tungsten at 19.3 g/cm³ but melts at 1,065 °C. Lead (11.3 g/cm³), palladium (12 g/cm³), nickel (8.89 to 8.91 g/cm³), copper (8.9 to 8.96 g/cm³) and silver (10.5 g/cm³) are under the density line and melt below 2,000 °C. Palladium melts highest of them, at 1,555 °C; lead melts at 328 °C. Nickel (1,455 °C) and copper (1,083 to 1,085 °C) serve as reference metals.

Tungsten carbide is a cobalt-bonded composite with no melting point in our data, and its practical limit is the binder (general engineering knowledge, not in our data). Tungsten figures here are for the pure metal, not heavy alloy.

What limits tungsten and tantalum in a hot part?

Oxidation, hydrogen and chemistry limit them, not the melting point. A melting point is not a use limit (general engineering knowledge, not in our data).

Tungsten oxidises in air at elevated temperature and must be protected; the profile gives no temperature. In general it starts to oxidise at several hundred °C and its oxide turns volatile at higher temperature (general engineering knowledge, not in our data). It is also brittle and hard to form or machine. The sourced hot use is a rocket nozzle throat insert (NASA report CR-66759), chosen for the melting point.

Tantalum forms a dense Ta2O5 oxide layer in air at room temperature that protects the metal. That skin does not carry over to heat: above a few hundred °C tantalum absorbs oxygen, nitrogen and hydrogen and embrittles, so it is a vacuum or inert-gas metal at temperature (general engineering knowledge, not in our data). Hydrofluoric acid is the main chemical exception, hot sulfuric acid and hot alkaline solutions also corrode it, and in hot acids hydrogen can enter the lattice (hydrogen embrittlement); the profile records no threshold and no corrosion rate.

Palladium melts at 1,555 °C, but hydrogen limits it. Below about 293 °C and under 20 bar of hydrogen the hydride phase swells the lattice by more than 10 percent, and pure palladium membranes warp and fracture.

Lead is the opposite case, because low melting is the feature. Tin-lead solder Sn63Pb37 melts at 183 °C, high-lead Sn10Pb90 over 268 to 302 °C and lead-free SAC305 over 217 to 220 °C. Those are solder alloys, not lead metal.

Which metal is best for electrical contacts and conductors?

On resistivity silver is lowest, then soft copper, gold and tungsten. The chart shows thermal conductivity because every charted metal stores it; resistivity is given here in words because copper has none. Silver is 1.6e-8 Ω·m, gold 2.2e-8 Ω·m and tungsten 5.4e-8 Ω·m. Copper's datasheet gives 58 MS/m and 100 %IACS for soft C10200 strip only, so cold-worked tempers conduct less. Converted (our arithmetic, not a stored value), soft copper falls between silver and gold. Palladium is 16 %IACS.

Contact surfaces fail by chemistry, so conduction is only half the choice. Gold is not attacked by oxygen or most acids at room conditions, and a 2025 NASA workshop tutorial on NASA-STD-8739.11 gives a minimum contact gold plating of 50 microinches. Gold has two traps. In tin-lead solder, above roughly 5% gold AuSn4 crystals can embrittle the joint; that threshold comes from a 1972 JPL flight-connector study, and common workmanship rules are stricter (general engineering knowledge, not in our data). On aluminium, gold can form brittle gold-aluminium compounds in wire bonds, known as purple plague (general engineering knowledge; our gold profile holds no sourced bond-wire data).

Silver tarnishes to Ag2S in sulphur-bearing air. In low Earth orbit atomic oxygen turns exposed silver into a black oxide that cracks and flakes, so spacecraft bury it, as in silvered Teflon blankets. Palladium's supplier contact resistance is 6 mΩ at 30 gf, unchanged after 1,000 h at 150 °C, and 10 mΩ after 720 h in flowers of sulfur at 50 °C; that is one supplier test. Copper forms Cu2O, then CuOHCl and CuS in rural air. In a separate atomic-oxygen exposure on LDEF, the thickness of a sputtered copper film rose from 74.2 to 106.7 nm as oxides formed.

How strong and hard are these metals?

Strength figures exist for only some of these metals, in different forms, so no chart ranks them. Annealed palladium is 280 MPa tensile and 270 MPa yield from one supplier, against 170 to 210 MPa and 60 to 120 MPa from another with the temper not stated. Nickel 200 annealed rod and bar shows 105 to 210 MPa yield and 380 to 520 MPa tensile as nominal ranges, not minimums. Soft copper C10200 strip is 69 MPa yield.

Silver bar (99.95%) is 330 MPa tensile hard and 172 MPa soft, with a tensile modulus of 82.7 GPa. Gold has only wire figures, 342.7 MPa for 50 µm hard wire and 135 MPa for 750 µm annealed wire, which depend on diameter and temper and are not bulk gold. Lead has no tensile figure; its Brinell hardness is 38.3 MN/m2 and its modulus 16 GPa.

Tungsten and tantalum have hardness and modulus but no strength. Tungsten shows Vickers 3,430 MN m-2 and modulus 411 GPa; tantalum shows Vickers 873 MN/m2 and modulus 186 GPa. Neither states test conditions. Cemented carbide grades with 6 to 10 percent cobalt are 1,200 to 1,605 HV30 in one supplier's table of typical values.

The units differ (MN/m², HV, HV30, kgf/mm²) and test loads are unstated, so we give a read-across, not a ranking. An MN/m² figure is about ten times its HV number (general engineering knowledge, not in our data). That puts tungsten at a few hundred HV and tantalum under a hundred, against 40 HV for annealed palladium and 95 for hard silver.

Which metals shield radiation, and what does their mass cost?

Lead is better per gram and tantalum per centimetre of thickness at both 0.1 and 1 MeV, the two energies where our data holds both. NIST total mass attenuation coefficients at 0.1 MeV are 5.549 cm²/g for lead and 4.302 for tantalum, against 0.1704 for aluminium as a light-metal baseline. At 1 MeV they are 0.07102, 0.06567 and 0.06146, so the three are close. Lead's coefficient jumps from 1.910 to 7.683 cm²/g at its K edge near 88 keV, and between tantalum's K edge and lead's, tantalum beats lead per gram too (general engineering knowledge, not in our data).

Thickness depends on density, so the order reverses per centimetre. Tantalum (16.6 g/cm³) is 1.5 times as dense as lead (11.3 g/cm³), a larger factor than lead's per-gram lead (5.549 against 4.302 at 0.1 MeV, 0.07102 against 0.06567 at 1 MeV). Multiplying each coefficient quoted here by density (our arithmetic, not a stored value) puts tantalum ahead per centimetre at both energies. The sourced spacecraft example is tantalum: Europa Clipper has a vault plate about 1 mm thick and 18 by 28 cm protecting its electronics, with no dose reduction in our data.

A 1973 NASA review recommends a tungsten alloy for gamma shield layers near a space reactor and calls tungsten the most efficient gamma attenuator per unit weight, though costly. Tungsten is denser than tantalum, but we hold no numeric tungsten attenuation, so it cannot be set against lead and tantalum.

Which of these metals are toxic, corrode or are hard to buy?

Lead is a reproductive toxicant in our data: 84.4% of notifying companies give H360, and the WHO fact sheet says there is no known safe blood lead concentration. OSHA's limit is 0.050 mg/m³ over 8 hours. The EU RoHS directive caps lead at 0.1% in most new electronics, and lead has been a REACH Candidate List substance since 27 June 2018. Yet NASA asks for at least 3% lead in tin finishes on space electronics unless a mitigation is approved, because pure tin grows whiskers.

The others carry hazards too. Nickel is a skin sensitiser (H317, 99.9% of 4,778 reports), and the US National Toxicology Program lists the metal as a reasonably anticipated human carcinogen. Cemented carbide dust brings cobalt hazards: the Sandvik data sheet for 3 to 10% cobalt carries the signal word Danger and lists H350i (may cause cancer by inhalation) and H334 (breathing difficulties if inhaled), while tungsten carbide itself is not classified. Powders of tungsten (73.4% of 741 reports), tantalum (56% of 455) and palladium (31.5% of 616) are notified as flammable solids. For gold, 99.2% of 506 reports say the metal does not meet GHS criteria.

For corrosion, Nickel 200 is described as unexcelled in caustic alkalies and excellent in flowing sea water, but stagnant sea water under deposits and nitric acid attack it. In a 20-year air test at a railroad-industrial site it corroded at 0.0056 mm/a.

Supply is concentrated for the refractory metals. China mined 78.8% of tungsten in 2025 (computed from USGS tonnes) and added export controls on selected tungsten items in February 2025. Congo (Kinshasa) mined 52% of tantalum in 2025 (computed from USGS tonnes against a rounded world total). Indonesia mined 66.7% of nickel, and Russia and South Africa 39.5% and 37.9% of palladium (2024 estimate). Producer shares are smaller for copper (Chile, 23%), gold (China, 11.5%) and silver (Mexico, 24.2%).

Terms used

Refractory metal

a metal that melts above about 2,000 °C, the usual convention, such as tungsten or tantalum. Melting point alone does not set the safe service temperature (general engineering knowledge, not in our data).

Mass attenuation coefficient

how strongly a material absorbs X-rays or gamma rays per gram, in cm²/g, at one photon energy. Multiply by density for the effect per centimetre of thickness (general engineering knowledge, not in our data).

%IACS

electrical conductivity as a percentage of annealed copper's. Soft oxygen-free copper is 100.

Vickers and Brinell hardness

indentation tests whose numbers depend on load and unit. A value in MN/m² is about ten times its HV number (general engineering knowledge, not in our data).

K edge

the photon energy at which a metal's absorption jumps because K-shell electrons can first be removed. Below it the metal is more transparent (general engineering knowledge, not in our data).

Annealed

heated and slowly cooled so the metal is soft. Cold working then raises strength.

Hydrogen embrittlement

hydrogen entering a metal's lattice and making it brittle or distorted. Our data records it for palladium and for tantalum in hot acids.

Cemented carbide

tungsten carbide grains held together by a cobalt binder, also called hardmetal. It is a composite, not the pure metal.

H-statement

a coded hazard phrase, for example H360 for possible damage to fertility. Percentages here are shares of company reports to a database, not a legal classification.

What to check before you order

1. Ask for the grade, purity and form together, for example C10200 soft strip, Nickel 200 annealed bar, 99.95% silver bar or a named carbide grade with its cobalt content.
2. Ask whether each strength or hardness figure is a minimum, typical or nominal value, and in which unit.
3. For hot parts, ask for the maximum use temperature in your atmosphere (air, vacuum or inert gas) and, for tantalum, the oxygen and hydrogen pick-up limit.
4. For shielding, ask for attenuation at your photon energies and the required mass per area. Do not size a shield from this article.
5. Get the safety data sheet for the exact form of lead, nickel, tungsten carbide and any powder, and check RoHS and REACH limits for lead before design.
6. For contacts, ask for plating thickness, underplate and solder compatibility, and test in your own sulfur or atomic-oxygen environment.

What do we not know yet?

• No profile stores a maximum service temperature or a creep figure, tungsten's oxidation onset or tantalum's gas pick-up when hot, so melting point is all we hold for hot-part limits.
• Tungsten, tantalum and lead have no tensile or yield strength. Hardness is stored in MN/m², HV and HV30 without test loads, so we give a rough read-across and no hardness chart.
• Tungsten has no numeric attenuation, lead and tantalum have photon coefficients at a few energies only, and no profile holds electron, proton or neutron shielding data or a spacecraft dose result.
• Copper has no resistivity entry, and thermal expansion is stored under different property names, so a blank cell in the at-a-glance table means a property is stored under another name or not at all.
• We hold no tungsten heavy-alloy, molybdenum, niobium or rhenium profile. Prices are stored only for tungsten (APT) and nickel (LME), in different forms, and not for lead or tantalum, so there is no cost ranking.
• Corrosion is mostly qualitative: tungsten, tantalum, lead and gold have no corrosion rate, and tungsten carbide has no wear rate or transverse rupture strength.

Questions people ask