Temperature That ClicksEpisode 1 / Temperature is not a property of matter

In 2019 humanity stopped measuring k_B and fixed its value ── measure temperature in energy and k_B disappears

Temperature is not a property of matter \(k_B\) is not a constant of nature but a unit conversion factor.
Measure temperature in energy and \(k_B=1\). The only thing that matters is the dimensionless ratio \(E/k_BT\) ──
and the single number 25 meV at room temperature explains chemistry, semiconductors and life together.

Tools you'll need: units of energy (J and eV), logarithmic scales, a feel for exponentials The heart of this episode: k_BT(300K) ≈ 1/40 eV

We treat "temperature" as if it were a property of a thing ── this iron is hot, that water is cold. In physics it is not a property of the thing. More than that: the quantity called temperature is surplus. Measure in energy instead. The conversion factor for doing so is \(k_B\), and since the 2019 revision of the International System of Units, \(k_B\) is no longer something we measure but something we decide ── its value is fixed by definition at exactly \(1.380649\times10^{-23}\ \mathrm{J/K}\). Bonus ② of the sister series "Cosmology That Clicks" asked "why did we fix the speed of light?"; this is the temperature version. And once temperature is converted to energy, an astonishing amount becomes visible through a single number ── at room temperature \(k_BT\approx\) 25 meV. Use 25 meV as your ruler and why ice melts but diamond doesn't, why semiconductors work, and why life is possible at room temperature all line up on one diagram.

01The kelvin is a historical leftover

Measure length in metres, time in seconds ── and temperature in kelvin. That is how we learn it. But temperature and length differ in one decisive way.

Temperature can simply be measured in energy. The mean kinetic energy of a gas molecule is \(\frac32 k_BT\). So the thing we call "temperature" is effectively energy. We just call it by a different unit.

The heart of this episode ── k_B is an exchange rate
$$k_B=1.380649\times10^{-23}\ \mathrm{J/K}\quad(\text{a defined value, since 2019})$$

This is not "a measurement of something in nature." It is the exchange rate for converting kelvin into joules ── the same kind of number as the "150" in 1 dollar = 150 yen. Which is why theoretical physics routinely sets \(k_B=1\) in the first line and writes temperature directly as an energy ── and then \(k_B\) never appears again.

On 20 May 2019 the SI was revised: the kelvin abandoned its material-based definition (1/273.16 of the triple point of water) and is now defined by fixing the value of \(k_B\). We stopped measuring and started deciding ── precisely the manoeuvre performed on the speed of light in 1983.

The same argument as "Cosmology That Clicks," Bonus ② That bonus explained the 1983 decision like this ── because the thing we were comparing against got overtaken by light. The moment time measurement using the speed of light became more precise than the length standard, measuring \(c\) lost its point and it became more rational to fix its value and define length with it.
Exactly the same thing happened with temperature. Fixing \(k_B\) and deriving temperature from energy became more precise than a material-dependent standard like the triple point of water. Quantities with units are stage machinery ── the collection's spine at work again.

02Only \(E/k_BT\) matters

Look at any formula where temperature appears. If a state has energy \(E\), the probability of finding it is

Temperature always appears in this shape
$$P\ \propto\ e^{-E/k_BT}$$

The exponent contains nothing but the dimensionless ratio \(E/k_BT\). A dimensionful \(E\) or \(T\) never shows up alone. Therefore ──

"High temperature" has no absolute meaning. There is only whether \(k_BT\) is large or small compared with the energy you care about.

10,000 K is fairly hot as seen by a chemical bond (a few eV) and indistinguishable from absolute zero as seen by a nucleus (MeV). Conversely 1 K is bitterly cold to us but plenty hot for a superconducting gap (below a meV). Temperature means nothing until you name what you are comparing it with.

03There is only one number to memorise

Working out k_BT at room temperature

Put in \(T=300\ \mathrm{K}\) (about 27 °C):

$$k_BT=(1.381\times10^{-23})(300)=4.14\times10^{-21}\ \mathrm{J}$$

Convert to eV (\(1\ \mathrm{eV}=1.602\times10^{-19}\ \mathrm{J}\)):

$$k_BT=\frac{4.14\times10^{-21}}{1.602\times10^{-19}}=0.0259\ \mathrm{eV}\approx \frac{1}{40}\ \mathrm{eV}=25\ \mathrm{meV}$$

Memorise "room temperature = one fortieth of an eV" and everything else is mental arithmetic: \(k_BT\) at temperature \(T\) is \((T/300)\times25\ \mathrm{meV}\). When a physicist says "room temperature," the number 25 meV is what is moving in their head.

Line up the energies around you against that ruler.

ThingEnergy× room-temperature k_BTAt room temperature
A cosmic microwave background photon≈ 0.6 meV0.02lukewarm ── the universe is at 2.7 K
room-temperature k_BT25 meV1the ruler
a hydrogen bond (between water molecules)≈ 0.2 eV8breaking and re-forming ── which is why water is liquid
the band gap of silicon1.1 eV44nearly insulating, but slightly excited ── which is why semiconductors exist
a visible photon2–3 eV80–120heat will never produce one ── which is why room-temperature things don't glow
a C–C covalent bond≈ 3.6 eV140utterly unmoved ── which is why organic molecules survive
nuclear binding≈ 8 MeV300 millioncompletely frozen

That table is the main thing this episode wants to say. One number ── "room temperature" ── and the ratio to the energy of interest is enough to decide, roughly, whether something happens. Hydrogen bonds sit at 8, so water is "almost but not quite breaking" = liquid. Covalent bonds sit at 140, so you do not decompose at room temperature. Life works at room temperature because biological interactions are arranged in exactly the sweet spot between 8 and 140.

04Try it ── move the \(k_BT\) line

The figure below lays various things out on a logarithmic energy axis. Move the temperature slider and the vertical \(k_BT\) line moves.

Reading it is simple ── things to the left of the line (smaller than \(k_BT\)) get broken or shaken by heat; things to the right (larger) are frozen solid. Set it to 300 K and the hydrogen bond sits just to the right of the line while the covalent bond is far away. Push to the solar surface (5800 K) and molecules start coming apart; past \(10^4\) K atoms ionise; at \(10^{10}\) K nuclei break ── which is the history of the universe played backwards.

Figure: energies of various phenomena on a logarithmic axis. The vertical line is k_BT. Move the temperature slider and read off left = broken by heat, right = frozen. Buttons jump to representative temperatures
k_BT (current temperature) broken / excited by heat frozen

05So what is temperature, in the end?

If \(k_B\) is merely a conversion factor, all that is left is "energy." But energy alone is not temperature ── the same 1 J stored in a battery is not the same as 1 J held by a glass of water. What differs?

This episode's answer, and the homework for the next

Temperature is not the amount of energy. It is how that energy is shared out among many degrees of freedom ── the slope of the sharing.

Nor is it "energy per particle" (a bigger system has more total energy at the same temperature). It is something more like "the eagerness to share" ── and next time we will make it stand up in front of you by tossing counters around at random. That will show at once why the form is \(e^{-E/k_BT}\), and why the real quantity is not \(T\) but \(1/k_BT\).

◇ ◇ ◇
The honest line ── the exact reach of "k_B is a conversion factor"

Established: that the 2019 SI revision defines the kelvin through the fixed value \(k_B=1.380649\times10^{-23}\ \mathrm{J/K}\); that unit systems with \(k_B=1\) (temperature expressed as energy) are widely used in theory; that the exponent of the Boltzmann factor \(e^{-E/k_BT}\) is the dimensionless ratio \(E/k_BT\); that \(k_BT(300\,\mathrm{K})=0.0259\) eV; and the energy values in the table (hydrogen bonds, silicon's 1.12 eV band gap, visible photons at 1.6–3.1 eV, C–C bond energies, ≈8 MeV binding per nucleon, typical CMB photon energies) ── all standard values.

Don't overstate it: (1) "\(k_B\) is not a constant of nature" means it is a coefficient linking dimensions, not that \(k_B\) is meaningless. Given that the kelvin is historically entrenched as an independent unit, the number is needed in practice. (2) "Anything bigger than \(k_BT\) is frozen" is a rule of thumb, not a threshold. The Boltzmann factor \(e^{-E/k_BT}\) falls off smoothly, so even at \(E/k_BT=44\) (silicon), with \(10^{22}\) particles around the number excited is far from negligible ── which is exactly why semiconductors work. The "at room temperature" column is a qualitative reading. (3) "Temperature is not a property of matter" means temperature is a state variable of a system, not a constant intrinsic to a substance. Of course "the temperature of this object" is well defined as a state. The final form of this claim ── that temperature depends on the observer's state of motion ── is Episode 7.

Exercises (solvable with this episode's ideas)
  1. Convert \(k_BT\) at 1000 K into eV in your head.
    See the answer
    25 meV at 300 K, so \(25\times(1000/300)\approx 83\) meV = 0.083 eV. It is simply proportional, so mental arithmetic works.
  2. Roughly what temperature is needed to emit visible light (2 eV) thermally? And why don't room-temperature objects glow?
    See the answer
    \(k_BT\sim2\) eV needs \(T\approx 2/0.025\times300\approx 2.4\times10^4\) K. In practice the tail of the distribution makes things glow red at a few thousand K (iron glows around 1000 K). At room temperature \(E/k_BT\approx80\) and \(e^{-80}\approx10^{-35}\) ── essentially no visible photons.
  3. Why does "high temperature" have no absolute meaning?
    See the answer
    Because temperature only ever enters physics through the ratio \(E/k_BT\). The same 10,000 K is hot compared to a chemical bond and indistinguishable from absolute zero compared to a nucleus. Nothing is decided until you name the comparison.
  4. What does the 2019 SI revision have in common with fixing the speed of light in 1983?
    See the answer
    Both stopped measuring a constant and instead decided its value, using it to define the unit. Once fixing the constant beats a material-dependent standard (the triple point of water, the metre bar) in precision, this becomes the rational move. Units are conventions, not discoveries.

Episode 1 summaryTemperature is energy, and k_B is the exchange rate

\(k_B\) is not a number obtained by measuring nature; it is the conversion factor from kelvin to joules, made a defined value in the 2019 SI revision ── we stopped measuring and decided. Which is why theory sets \(k_B=1\), writes temperature as an energy, and never sees \(k_B\) again.

Temperature only ever acts on physics through the dimensionless ratio \(E/k_BT\). So "high temperature" has no absolute meaning; you need something to compare with. There is one number to memorise ── room-temperature \(k_BT\approx\) 25 meV = 1/40 eV. Apply that ruler and the world lines up on one diagram: hydrogen bonds at 8× (almost but not quite breaking = liquid), silicon's band gap at 44× (nearly insulating but slightly excited = a semiconductor), covalent bonds at 140× (unmoved = biomolecules survive).

This document is Episode 1 of the "Temperature That Clicks" series, a reading piece for physics-loving high-schoolers and undergraduates. The 2019 SI definition of the kelvin (fixing \(k_B=1.380649\times10^{-23}\ \mathrm{J/K}\)), the conventional use of \(k_B=1\) unit systems in theory, that the Boltzmann exponent is the dimensionless ratio \(E/k_BT\), \(k_BT(300\,\mathrm{K})\approx0.0259\) eV, and the energy values in the table (hydrogen bond, silicon band gap, visible photon, C–C bond, binding energy per nucleon, CMB photon) are all established facts and standard values. That "\(k_B\) is not a constant of nature but a conversion factor" means a coefficient linking dimensions; that "anything above \(k_BT\) is frozen" is a rule of thumb, the Boltzmann factor falling off smoothly so that excitation is not negligible at large \(E/k_BT\) when particle numbers are huge; and that the final form of "temperature is not a property of matter" is Episode 7 ── all spelled out in the body's "honest line." The figure places representative values on a logarithmic energy axis; each value indicates an order of magnitude. ── To print, use your browser's "Print" and "Save as PDF" (in the print version the slider and answers are frozen and hidden). Next: Episode 2, Temperature is how energy gets shared out / Contents / sister series Cosmology That Clicks (Bonus ②, "Why did we fix the speed of light?").

Print / make a PDF: ⌘+P (Ctrl+P on Windows). On screen, the temperature slider moves the k_BT line so you can read what breaks and what freezes. Buttons jump to representative temperatures. "See the answer" opens each solution.