Kelvin to Celsius Converter

The conversion is one subtraction

Celsius temperature is kelvin temperature minus 273.15. Written as a formula, °C = K - 273.15, and the reverse is K = °C + 273.15. There is no fraction and no offset beyond that single number, because the two scales use a degree of the same size.

The 273.15 is not a measured constant that might drift. The BIPM definition of the kelvin fixes the Boltzmann constant at 1.380649 x 10^-23 J/K, and the Celsius scale is defined by placing 0 °C at 273.15 K. That makes the offset exact, so a converter can print as many decimals as the input supports.

Work an example. 300 K minus 273.15 is 26.85 °C. Multiply by 9/5 and add 32 to reach Fahrenheit: 26.85 x 1.8 + 32 = 80.33 °F. The same 300 K is also 540 °R on the Rankine scale, which uses Fahrenheit sized degrees from absolute zero.

The subtraction is easy enough to do by hand, and the errors come from the direction. Adding 273.15 to a kelvin value produces a number that is 546.3 too high, which looks plausible on a thermometer and is wrong by a mile.

How this page differs from the Celsius to Fahrenheit converter

The Celsius to Fahrenheit converter handles a two step change: multiply by 9/5, then add 32. This page handles a one step change between two scales that share a degree size, so the arithmetic is an offset and the interesting questions are different ones.

Three of them come up often. How far above absolute zero is this reading, and is a negative kelvin value even possible. How many decimals should survive the conversion when a lab reports 0.01 K and a weather app reports 0.1 °C. And what the reference points are, from liquid nitrogen at 77.4 K to body temperature at 310.15 K, so a result can be sanity checked against something physical.

This page answers those. It converts in both directions, prints the Fahrenheit value as a cross check, shows how far the reading sits above absolute zero, and flags input below 0 K instead of quietly returning a colder than possible number.

Absolute zero and why negative kelvin is not a cold temperature

0 K is -273.15 °C. At that point a system holds the minimum thermal energy quantum mechanics allows, and no amount of cooling removes more. Laboratories reach temperatures within a billionth of a kelvin of absolute zero, but they do not cross it.

So a reading of -5 K is not cold, it is impossible, and the converter says so rather than returning -278.15 °C. The one place the phrase negative temperature appears in physics is a special case: in some spin systems, a population inversion produces a negative thermodynamic temperature that behaves as hotter than any positive temperature, because heat flows from it to everything else. That is a laboratory curiosity in isolated quantum systems, not a cold object, and it does not make -5 K a valid input on a thermometer.

For ordinary work the floor is the floor. A cryogenic lab may work at 4.2 K, the boiling point of helium, or at 77.4 K with liquid nitrogen. Both are positive values and both convert with the same subtraction. The coldest outdoor temperature recorded in Antarctica, near -89 °C, is about 184 K, which is still 184 degrees above the floor.

Precision: 0.1, 1 or 10 degrees

Kelvin and Celsius carry the same size degree, so the precision of the input is the precision of the output. The subtraction adds no uncertainty and removes none. What changes is how many of those digits mean anything for the job at hand.

Rounding to the nearest 10 K takes 26.85 to 30, an error of 3.15 K. That is acceptable when the question is whether a furnace is in the 850 to 900 K band and useless when the question is whether a sample is at 26.85 or 26.90. Match the decimals to the instrument, not to the calculator.

One trap sits between 0.1 and 0.01. A reading given as 300.0 K has four significant figures, so 26.9 °C is the honest answer. Printing 26.85 implies the input was known to 0.01 K, which it was not. The converter will show the extra digit if you ask for it, but the input sets the limit.

Reference temperatures in both scales

These are the values worth memorizing, because they let you check a conversion in your head. The cryogenic figures come from the NIST cryogenics reference table, which lists normal boiling points at atmospheric pressure.

Liquid nitrogen is often quoted as 77 K or 77.15 K. The measured normal boiling point is 77.4 K, which is why the table uses that value; the difference is 0.25 K, below the precision most dewars are read at.

Two of these are definitions rather than measurements. 0 K and 273.15 K are fixed by the scale definitions, and 373.15 K is the boiling point only at one atmosphere of pressure. At 1,500 meters water boils near 368 K, or 95 °C, which matters for cooking times and for anyone calibrating a thermometer in Denver.

Where each scale gets used

Kelvin is the working scale of thermodynamics and of most laboratory science, because absolute temperature appears in gas laws, radiation laws and reaction rates. A gas at 300 K has twice the absolute temperature of a gas at 150 K, and the pressure doubles at fixed volume. The same statement in Celsius, 27 °C against -123 °C, hides that ratio completely.

Celsius owns daily life almost everywhere except the United States. Weather, ovens, refrigerators and medical thermometers report in °C in most countries, and a forecast of 30 °C means a hot day while 300 K means nothing to a listener. The conversion matters when a lab result, a sensor log or a product spec arrives in kelvin and has to be read by someone thinking in Celsius.

Industry mixes the two. A furnace may be specified at 1,173 K for a 900 °C process, a cryogenic freezer at 77 K, and a chiller at 278.15 K for 5 °C. The numbers look unrelated until you subtract 273.15 and see the Celsius value underneath. Lighting is the odd case: a 2700 K lamp label describes color, not temperature, and converting it to Celsius produces a meaningless number.

Sensor logs make the scale choice concrete. A logger recording at 1 Hz writes 86,400 rows a day, and every row carries the same 273.15 offset once it is converted to Celsius. Doing that subtraction in the export step, rather than inside a downstream formula, keeps one unit in the file and removes a whole class of double conversion bugs.

Mistakes: 273 instead of 273.15, and adding instead of subtracting

The two errors that show up most often are a rounded offset and a reversed sign. Both produce results that look reasonable at a glance, which is why they survive review.

Subtracting 273 instead of 273.15 is a 0.15 K error. It is harmless for a weather report and unacceptable for a calibration certificate. The reversed sign is worse because 573 °C is a plausible temperature for an oven or a soldering iron, so nobody questions it. If a conversion of a room temperature reading lands above 500 °C, the sign is the first thing to check.

A third error is unit confusion in software. Some APIs return kelvin and label the field temp, others return Celsius with the same field name, and a pipeline that assumes the wrong one is off by 273.15 everywhere. Store the unit next to the number, and convert once at the edge of the system rather than in the middle of a calculation.