Filament Materials

Part of Lighting

The evolution from carbon to osmium to tungsten filaments, why tungsten won, and what properties make an incandescent filament material viable.

Why This Matters

The incandescent lamp works by heating a wire so hot it glows. The material of that wire determines how bright, how efficient, and how long-lasting the lamp will be. The history of filament materials is a story of progressive discovery of what properties matter, what materials have them, and why tungsten was the winning choice among all elements investigated.

For a rebuilding civilization, this knowledge matters in two ways. First, it explains why replicating incandescent lamp manufacturing is difficult — tungsten requires extremely high processing temperatures and specialized metallurgy. Second, it provides the principles needed to evaluate alternative materials for improvised incandescent sources when manufactured bulbs are unavailable.

Properties Required of a Filament Material

To function as an incandescent filament, a material must satisfy several requirements simultaneously:

High melting point: the filament glows because it is heated to incandescence (above ~1,000 K). More light, and whiter light, comes from higher temperature. The practical maximum operating temperature is limited by the material’s melting point. The material must withstand its operating temperature for hundreds or thousands of hours.

Low vapor pressure at operating temperature: even solid materials evaporate slowly. A filament that evaporates rapidly will thin at hot spots, leading to breakage. High vapor pressure at operating temperature means short filament life. This requirement rules out many high-melting materials that sublime readily.

Drawability into fine wire: the filament must be drawn into wire 20–70 micrometers thick (thinner than a human hair). The material must be ductile enough to survive this process. Brittle high-melting materials (ceramics, many carbides) cannot be drawn into wire.

Adequate electrical resistance: the filament must have a resistance high enough at operating temperature that the correct power is dissipated in the design voltage and lamp geometry.

Reasonable cost and availability: a material that requires exotic mining and processing may be impractical despite excellent properties.

Carbon: The First Filament

Edison’s 1879 lamp used carbonized plant fibers — cotton thread, bamboo strips, paper — as the filament. These gave a black carbon filament with a high melting point (carbon sublimes at 3,600°C) but high vapor pressure at operating temperatures. Carbon evaporates rapidly in a vacuum at 2,000 K, blackening the bulb and thinning the filament within hours at high temperature.

The solution was to operate carbon filaments at lower temperatures (1,200–1,500 K), producing a yellowish-red glow with efficacy of only 1.5–4 lm/W. At these temperatures, evaporation was slow enough for 200–300 hours of life. But efficiency was poor — most of the electrical energy radiated as infrared heat, not visible light.

Squirted carbon filaments (1880s): Mann and others developed a process of squirting cellulose acetate solution through a fine nozzle, then carbonizing the resulting fiber. These were more uniform in cross-section and somewhat more consistent in resistance than hand-prepared plant fibers.

Edison’s bamboo filaments (1880): Edison found that carbonized bamboo fibers, with their very regular cellular structure, gave the most uniform carbon filaments of natural materials available. Japanese bamboo varieties were imported for lamp production. This gives some sense of the manufacturing challenge — global supply chains were established to find a good filament material.

The Metals Between Carbon and Tungsten

Osmium filament (c. 1898, Auer von Welsbach): osmium was the first metal to be drawn into fine wire and used as an incandescent filament. Osmium has a melting point of 3,033°C and could be operated at higher temperatures than carbon, producing 6–10 lm/W. But osmium is extremely rare (one of the rarest stable elements) and its compound osmium tetroxide, formed by oxidation, is intensely toxic. Osmium lamps had brief commercial life.

Tantalum filament (1903, von Bolton): tantalum (melting point 2,996°C) could be drawn into wire and operated at higher temperatures than carbon, giving 5–7 lm/W. Tantalum was somewhat less rare than osmium and the lamps were commercially significant for a few years. However, tantalum filaments were fragile and difficult to manufacture at consistent quality.

Nernst lamp (1897, Nernst): not a metal filament but a ceramic (mixed rare earth oxides) glower. The Nernst lamp required external preheating to start (the ceramic is an insulator when cold, conductor when hot) but operated in air without a bulb vacuum. Efficacy of 5–6 lm/W. Used commercially in Europe and the US briefly but was complex to manufacture and operate.

Tungsten: The Optimal Choice

Tungsten has the highest melting point of all elements: 3,422°C. Its vapor pressure at 2,500 K is extremely low — much lower than carbon or the earlier metal alternatives. It can be operated at high temperature (2,500–2,900 K) with acceptable evaporation rate, producing white light at 10–15 lm/W.

Ductile tungsten was first produced in 1909 by Coolidge at General Electric, by adding small amounts of potassium and aluminum oxide to tungsten powder before sintering. These dopants prevent grain boundary migration at high temperature, which would cause the normally brittle tungsten to become ductile and drawable. This “non-sag” tungsten wire could be wound into coils and shaped reliably.

Coiled-coil filament (1934): winding the filament into a coil, then coiling that coil again, dramatically reduces heat loss by radiation from adjacent coil sections — the coils radiate heat back and forth among themselves. This increases the effective operating temperature and improves efficacy to 12–15 lm/W. The coiled-coil configuration is standard in all modern tungsten incandescent lamps.

Gas filling (nitrogen, then argon): vacuum bulbs allowed carbon and early metal filaments to operate without oxidation, but the filament still evaporated and thinned. Filling with an inert gas (nitrogen first, then argon at lower cost) slows evaporation by convective return of evaporated atoms to the filament surface. Argon-filled lamps at 0.1–0.2 atm pressure achieved better life than vacuum lamps at the same temperature.

Halogen Regenerative Cycle

Tungsten-halogen lamps (1950s onward) extend filament life significantly by using a halogen gas (iodine, bromine) that reacts with evaporated tungsten vapor: tungsten vaporizes, migrates toward the cooler bulb wall, reacts with iodine to form tungsten iodide (gaseous), migrates back toward the hot filament, and deposits tungsten back on the filament. This “halogen cycle” keeps the bulb wall clean and re-deposits evaporated tungsten, extending life to 2,000–4,000 hours and allowing higher operating temperatures (20–25 lm/W).

The cycle requires the bulb wall to stay above 250°C (otherwise tungsten iodide condenses on the wall and the cycle fails) — which is why halogen lamps use quartz bulbs that can tolerate this temperature, and why they must not be touched with bare hands (skin oils on quartz create thermal stress failures at high temperature).

Improvised Incandescent Sources

For a rebuilding civilization that cannot manufacture tungsten wire, improvised incandescent sources have very limited lifespan:

Carbon rods from batteries, operated in vacuum or inert atmosphere, produce a reddish glow at 1,000–1,200 K. Life measured in tens of hours if operated conservatively. Useful for emergency illumination.

Fine carbon threads from charred plant fibers (bamboo, rattan, grass stems) in a sealed glass vessel with evacuated air. The glass must be sealed airtight to prevent oxidation. This replicates Edison’s basic experiment. Efficacy is very low but light quality is warm and usable. Life is short but the material is renewable.

The lesson from filament history: the difficulty of achieving high efficacy from incandescent sources drove investment in discharge lamp technology (fluorescent, sodium, mercury) and eventually LED technology. A rebuilding civilization that can access salvaged LEDs should prioritize those over attempting to replicate tungsten wire manufacturing.