Physical Constants
A table of 14 physical constants with symbol, unit and provenance, marking which values are exact by definition and which are measured.
Reference
| Symbol | Constant | Value | Unit | Status | Notes |
|---|---|---|---|---|---|
| c | Speed of light in vacuum | 299,792,458 | m/s | Exact by definition | Exact since 1983, when the metre was redefined as the distance light travels in 1/299,792,458 of a second. Measuring c more precisely now refines the metre, not the speed. |
| h | Planck constant | 6.62607015 × 10⁻³⁴ | J·s | Exact by definition | Fixed in 2019, which is what defines the kilogram. The Kibble balance realises a mass from this number rather than from a metal cylinder kept near Paris. |
| ℏ | Reduced Planck constant | 1.054571817… × 10⁻³⁴ | J·s | Exact, derived | h divided by 2 pi. Exact as a definition but irrational as a decimal, so any printed form is truncated. |
| e | Elementary charge | 1.602176634 × 10⁻¹⁹ | C | Exact by definition | Fixed in 2019, which defines the ampere. Also the conversion from electronvolts to joules, since one electronvolt is the energy one elementary charge gains across one volt. |
| k | Boltzmann constant | 1.380649 × 10⁻²³ | J/K | Exact by definition | Fixed in 2019, which defines the kelvin. It converts a temperature into an energy per particle, which is why it appears wherever thermal energy meets a single molecule. |
| NA | Avogadro constant | 6.02214076 × 10²³ | mol⁻¹ | Exact by definition | Fixed in 2019, which defines the mole. The mole is now a count rather than a mass of carbon-12, so this number carries no uncertainty. |
| R | Molar gas constant | 8.31446261815324 | J/(mol·K) | Exact, derived | Exactly N_A times k, so it inherited their exactness in 2019. Watch the unit: this is the value for pascals and cubic metres, and 0.082057 L·atm/(mol·K) is the same constant. |
| F | Faraday constant | 96,485.33212… | C/mol | Exact, derived | Exactly N_A times e, the charge on a mole of electrons. Converts between moles of electrons and coulombs in any electrochemical calculation. |
| σ | Stefan-Boltzmann constant | 5.670374419… × 10⁻⁸ | W/(m²·K⁴) | Exact, derived | Built from k, h and c, so exact. The fourth power is why a modest temperature rise makes a large difference to radiated power. |
| b | Wien’s displacement constant | 2.897771955… × 10⁻³ | m·K | Exact, derived | Peak wavelength times absolute temperature. Divide by a temperature to get the wavelength a black body radiates most strongly at. |
| g₀ | Standard gravity | 9.80665 | m/s² | Exact by definition | A convention, not a measurement, and not the value anywhere in particular. Local gravity runs from about 9.78 m/s² near the equator to 9.83 m/s² at the poles. |
| atm | Standard atmosphere | 101,325 | Pa | Exact by definition | Exactly 101,325 Pa by definition. Distinct from the IUPAC standard pressure of 1 bar, which is 100,000 Pa, and the two give different molar volumes. |
| T₀ | Ice point | 273.15 | K | Exact by definition | Zero degrees Celsius in kelvin, exactly. The offset every gas law needs, and one that cancels out of any temperature difference, so q = mcΔT never needs it. |
| G | Gravitational constant | 6.67430(15) × 10⁻¹¹ | m³/(kg·s²) | Measured | The least precisely known constant on this page, and the only measured one in the table. CODATA gives 6.67430(15), so the uncertainty is in the fifth digit: about 1 part in 45,000, which is enormous next to everything above. |
Left out of the table because their digits moved between the 2018 and 2022 CODATA adjustments: the vacuum permittivity ε₀, Coulomb’s constant kₑ and the vacuum permeability μ₀. Each has its own page with its value and uncertainty: CODATA 2022’s for ε₀ and μ₀, and for kₑ, which CODATA does not list, the figure that follows from CODATA’s ε₀.
Citing this tool
Last updated . Add the date you accessed it as well, which a citation of a page that can change asks for.
What exact by definition actually means
Most of the values in this table carry no uncertainty at all. Not a small
uncertainty: none. The speed of light is 299,792,458 m/s because
the metre is defined as the distance light travels in
1/299,792,458 of a second, so that figure is a unit conversion
rather than a result. A better experiment refines the metre. It cannot move
c.
That reads backwards at first, because a constant sounds like something you go out and measure. The resolution is that every measurement compares a quantity against a unit, and the unit has to come from somewhere. Fixing the constant and letting the unit follow is the more stable arrangement, because a defined number cannot drift and a physical object can.
The 2019 revision turned the definitions round
Until 2019 the kilogram was a platinum-iridium cylinder held near Paris, and the
Planck constant was whatever experiment said it was in terms of that cylinder.
The revision inverted that dependency. h is now fixed at
6.62607015 × 10⁻³⁴ J·s and the kilogram is whatever mass is
consistent with it, realised on a Kibble balance rather than by comparison
with metal. The same move fixed three further constants and pinned three more
units to them:
- The elementary charge
edefines the ampere. - The Boltzmann constant
kdefines the kelvin. -
The Avogadro constant
N_Adefines the mole, which became a count rather than a mass of carbon-12.
Exactness is inherited
A constant assembled only from exact constants is itself exact. The molar gas
constant is the clearest case: R is exactly
N_A × k, so when those two were fixed in 2019 R
stopped being a measured quantity and became
8.31446261815324 J/(mol·K), every digit of it. The Faraday
constant is exactly N_A × e for the same reason, and the
Stefan-Boltzmann constant is built from k, h and
c.
Exact does not mean short, and it does not mean unambiguous.
Watch the units rather than the digits. The reduced Planck
constant is h/2π, exactly defined and irrational, so any printed
decimal form of it is truncated. R is listed here for pascals and
cubic metres, and the same constant reads
0.082057 L·atm/(mol·K) in the units a gas law question usually
arrives in.
The gravitational constant is the one exception
Exactly one row here is a measurement. CODATA gives G as
6.67430(15) × 10⁻¹¹ m³/(kg·s²), so the uncertainty sits in the
fifth digit: about 1 part in 45,000. Next to constants with
no uncertainty whatsoever that is an enormous gap, and it is there because
gravity is desperately weak at bench scale. There is no way to screen an
apparatus from the mass of everything else in the room.
The electromagnetic constants are left out of the table on purpose. Several of
them moved digits between the 2018 and 2022 CODATA adjustments, and a stale
digit in a row of bare values is worse than no row. Each has a page of its own
instead, which gives its value with its uncertainty and says where the value comes
from: the CODATA 2022 adjustment for the permittivity and the permeability, and for
Coulomb’s constant, which CODATA does not list, the CODATA permittivity it is
derived from. Standard gravity is worth one caution of its own: at
9.80665 m/s² it is a convention rather than a measurement, and not
the local value anywhere in particular. Real surface gravity runs from about
9.78 m/s² near the equator to 9.83 m/s² at the poles.
Common questions
What does it mean for a physical constant to be exact?
It means the value is fixed by definition and has no experimental uncertainty, so no future measurement can change it. The speed of light is exact because the metre is defined from it, and since 2019 the Planck constant, elementary charge, Boltzmann constant and Avogadro constant are exact in the same way. Constants built only from those, such as the molar gas constant, are exact as a consequence. A more precise experiment now sharpens the unit rather than the constant.
Why is the gravitational constant known so much less precisely than the others?
Because it is still measured rather than defined, and gravity is extraordinarily weak at the scale of a laboratory. CODATA quotes 6.67430(15), which puts the uncertainty in the fifth digit, roughly 1 part in 45,000. Separate experiments using different apparatus have disagreed by more than their stated error bars, which is why the accepted uncertainty stays comparatively wide.