How this instrument works
Kelvin measures thermodynamic temperature from absolute zero, the point where a substance holds the least energy physics allows. Celsius measures the same energy but counts from ice's melting point instead, so the two scales share an identical degree size and differ by exactly one number: 273.15. That figure has been fixed since the kelvin's 2019 redefinition tied it to Boltzmann's constant, 1.380649 × 10⁻²³ joules per kelvin, rather than to any physical artefact — subtraction here is definition, not measurement.
Below about 120 K, kelvin stops being an academic preference and becomes the only sensible unit. Liquid nitrogen boils at 77.36 K and is the everyday workhorse of biological sample storage; liquid helium at 4.2 K cools MRI magnets and the outer stages of dilution refrigerators that carry superconducting quantum processors down toward their operating point. A cryostat log sheet reading '7.5 K, still stage' or '1 K pot' is describing real hardware stages on the way to those temperatures, not a rounding exercise — and almost nobody reads those numbers off in Celsius, because at that range Celsius values are all clustered a few hundred degrees below zero and tell you nothing useful at a glance.
Once you cross back above roughly 100 K, the traffic runs the other way. Molecular clouds in interstellar space sit around 10–20 K and get reported in kelvin by astronomers, but a lab technician thawing a −80 °C freezer box, a metallurgist reading a furnace gauge, or anyone checking today's forecast wants Celsius, because that scale's zero and hundred were chosen to bracket water doing something everyone has watched happen.
- Enter the reading from your instrument, log sheet or datasheet into the Kelvin (K) field; it opens at 300, roughly a warm room.
- The Celsius (°C) field updates immediately beneath it, recalculated on every keystroke with no button to press.
- For cryogenic stages, paste the value straight from your controller readout — the field steps in thousandths, so 1.723 K enters cleanly.
- Working the other way from a Celsius spec sheet? Add 273.15 to that figure and enter the result into Kelvin (K) to check your work.
- A negative Kelvin (K) entry is not a very cold sample — it is impossible, since 0 K is the coldest a physical system can be.
Worked example — reading a cryostat stage at 7.5 K
A dilution-refrigerator log records the 'still' stage sitting at 7.5 K partway through a cooldown, well above the millikelvin mixing chamber but far below the 4.2 K liquid-helium bath one stage up. Enter 7.5 into Kelvin (K) and the Celsius (°C) line reads −265.65 °C — a number so far below anything a household thermometer displays that most cryogenic logbooks simply never bother quoting Celsius at all.
Compare that against a warm bench reading: 300 K, a typical climate-controlled lab, converts to 26.85 °C, comfortably within room temperature. And the coldest calibration point most cryostats ever log, 1 K, converts to −272.15 °C — just 1.15 degrees above absolute zero on either scale, which is exactly why kelvin, not Celsius, is what gets written on the equipment.
Questions
Is the 273.15 offset exact or rounded?
Exact. SI defines Celsius temperature as t/°C = T/K − 273.15, so the figure belongs to a definition and carries no measurement uncertainty of its own. Converting in either direction loses nothing; any imprecision in your result came from your original instrument reading, not from this arithmetic.
Why do cryogenics labs report temperature in kelvin instead of Celsius?
Because below about 100 K, Celsius values pile up within a narrow band a few hundred degrees under zero and stop being informative at a glance — a 1 K stage and a 4 K stage look almost identical in Celsius (−272.15 °C versus −269.15 °C) but represent a fourfold difference in absolute temperature, which is what actually governs the physics. Kelvin also ties directly to gas laws, blackbody radiation and Boltzmann statistics without an offset getting in the way.
What is the coldest temperature ever measured?
Laboratory experiments have cooled nuclear spin systems to roughly 100 picokelvin, about one ten-billionth of a degree above absolute zero, using techniques like adiabatic demagnetization — though the bulk material surrounding those spins stayed far warmer. Naturally occurring cold is nowhere close: the Boomerang Nebula, at about 1 K, is thought to be the coldest known natural place in the universe, colder even than the 2.725 K cosmic microwave background that fills all of space.
Do I subtract 273.15 from a temperature difference, too?
No — this is the most common mistake with these two scales. A kelvin and a Celsius degree are identical in size, so a change of 5 K is a change of 5 °C with no offset applied. The 273.15 shift only applies to a point on the scale, not to an interval, a ramp rate, or a coefficient such as W/(m·K), which carries across unchanged.
Can kelvin readings go negative?
Not in the thermodynamic sense used here — 0 K is an absolute floor, the coldest any physical system can be, so a negative Kelvin (K) entry signals a transcription error rather than an unusually cold sample. A separate, more exotic concept called 'negative absolute temperature' exists in statistical mechanics for certain inverted-population systems, but it does not describe anything colder than 0 K; it is a different variable entirely and irrelevant to a cryostat log.
How cold is interstellar space compared to a liquid-helium bath?
Dense molecular clouds, where new stars eventually form, typically sit around 10–20 K — colder than a liquid-nitrogen bath at 77.36 K but noticeably warmer than the 4.2 K boiling point of liquid helium used to cool MRI magnets and cryostats. The near-total vacuum of deep space itself has no temperature in the everyday sense, since temperature describes matter's motion, but the cosmic microwave background radiation that pervades it corresponds to about 2.725 K.