Kelvin

Kelvin (symbol K, not °K) is the SI base unit of temperature. The zero point is absolute zero — the theoretical minimum temperature at which all molecular motion ceases — equal to −273.15°C or −459.67°F.

Kelvin uses the same step size as Celsius: a temperature change of 1 K is identical to a change of 1°C. The two scales differ only in offset:

• Water freezes at 273.15 K (0°C)
• Water boils at 373.15 K (100°C)
• Body temperature: ~310 K
• Room temperature: ~293-298 K

Used in: scientific work universally, colour temperature (warm white lights ~2700 K, daylight ~5500 K, blue sky ~10000 K), thermodynamic equations where absolute temperature matters (gas laws, Stefan-Boltzmann), and any context where ratios of temperatures need to make physical sense (a ratio of Celsius temperatures across 0°C is meaningless because Celsius has an arbitrary zero).

Since 2019, Kelvin is defined via the Boltzmann constant rather than via the triple point of water. The definition change preserved every practical value to within metrology precision; nothing on a thermometer reads differently because of it.

Why no degree symbol? Until 1968, kelvin temperatures were written “°K” like Celsius and Fahrenheit. The 13th CGPM dropped the degree symbol that year, reasoning that kelvin is a measure of absolute thermodynamic quantity rather than a position on an arbitrary scale — the same logic that gives metre, joule, and ampere no symbol. Modern scientific writing strictly uses “K” without “°”. The 1948 capitalisation rule (the unit is “kelvin” lowercase but the symbol is “K” uppercase, like volt/V and pascal/Pa) was specifically adopted to honour William Thomson, Lord Kelvin, while keeping the unit a regular noun.

Practical kelvin ranges nobody quotes: liquid nitrogen boils at 77 K (−196°C) — cheap enough that universities pour it into bowls for chemistry demos. Helium liquefies at 4.2 K, the temperature at which most superconducting magnets operate (MRI machines, particle accelerators). The cosmic microwave background sits at 2.725 K — the literal temperature of the universe, measured from any direction. The lowest temperatures ever produced in a laboratory are in the picokelvin range (10⁻¹² K), achieved by laser-cooled rubidium atoms; the absolute zero of 0 K remains unreachable by the third law of thermodynamics. Related: temperature converter, colour temperature. Reference: BIPM SI Brochure — Kelvin definition.

Worked example

Convert a comfortable room temperature of 22°C to kelvin and back. Kelvin = Celsius + 273.15, so 22 + 273.15 = 295.15 K. To compute the radiant power per square metre emitted by an object at that temperature using the Stefan-Boltzmann law j = σT⁴ (σ = 5.670374419×10⁻⁸ W·m⁻²·K⁻⁴), you must use kelvin: j = 5.67e-8 × 295.15⁴ ≈ 430 W/m². Plugging the Celsius value (22) into T⁴ would yield ~13 W/m² — wrong by a factor of 33, because T⁴ is meaningless on a scale whose zero is arbitrary. Now do the inverse for a 3000 K incandescent bulb filament: 3000 − 273.15 = 2726.85°C — that’s the literal filament temperature, not the colour-temperature label printed on the box (though they happen to agree here because the bulb radiates as a near-blackbody).

When and why it matters

You need kelvin any time the physics involves an absolute temperature: gas-law calculations (PV = nRT), thermal radiation (Stefan-Boltzmann, Wien’s displacement law), reaction-rate Arrhenius equations, noise-floor calculations for radio receivers (kTB noise), and selecting LED/screen white points (a 6500 K display matches D65 daylight; 2700 K mimics warm tungsten). The mistake people make: subtracting two Celsius temperatures and treating the difference as a ratio. A jump from 10°C to 20°C is not “twice as hot” — in kelvin it’s 283.15 → 293.15 K, only a 3.5% change in absolute thermal energy. Reference: NIST — SI units: temperature.