CODATA Releases

The CODATA task group on fundamental constants publishes updated recommended values every four years. AtomicAndPhysicalConstants.jl ships constants from every release since 2002 and lets you choose which one to use at runtime.

Supported releases

Release	Notes
2002	earliest supported; `M_TRITON` is not tabulated in this release (`CODATA2002.M_TRITON == NaN`)
2006
2010	first release with `G_HELION_NUCLEAR`, `ANOMALY_ELECTRON`, `ANOMALY_MUON`
2014
2018
2022	default

Checking the active release

using AtomicAndPhysicalConstants
RELEASE_YEAR   # Int – the currently active release year

Changing the release

using AtomicAndPhysicalConstants

set_release(year = "2014")
# [ Info: The default CODATA release is now 2014.
#         Restart your Julia session for this change to take effect.

The setting is stored via Preferences.jl and persists across Julia sessions until changed again.

Restart required

Constants are resolved at package load time, so a Julia session restart is required for the new release to take effect.

Unsupported releases

Only the years listed under Supported releases above are valid. If the stored preference points at an unsupported release (for example after hand-editing LocalPreferences.toml), the package raises an error at load time listing the valid options rather than silently falling back to the default.

To revert to the default (2022):

set_release()          # no argument → defaults to "2022"

Direct access to release structs

Each release is also exported as a named CODATA_release struct. You can inspect or use individual values without changing the global default:

CODATA2018.M_ELECTRON    # electron mass from the 2018 release
CODATA2014.C_LIGHT       # speed of light from the 2014 release

Constants not in CODATA

The pion masses (M_PION_0, M_PION_CHARGED) are not published in any CODATA release. They are taken from Particle Data Group (PDG) tables and remain constant regardless of the selected release year.

CLASSICAL_RADIUS_FACTOR is derived rather than tabulated: it is computed as R_ELECTRON * M_ELECTRON from the active release. It is therefore not a field of the CODATA_release structs — CODATA2018.CLASSICAL_RADIUS_FACTOR does not exist, only the exported top-level constant.

Constants that differ from the tabulated release value

G_DEUTERON, G_HELION, and G_TRITON are pulled from the NIST/CODATA tables but are rescaled before export. NIST tabulates these g-factors against the nuclear magneton $\mu_N = e\hbar/2m_p$; the package converts them to the particle's own magneton so that the gyromagnetic anomaly $a = (|g|-2)/2$ comes out correct:

G_DEUTERON == CODATA2022.G_DEUTERON_NUCLEAR * M_DEUTERON / M_PROTON   # active release

The release structs name these fields G_DEUTERON_NUCLEAR, G_HELION_NUCLEAR, and G_TRITON_NUCLEAR — the suffix flags the nuclear-magneton normalization, so there is no field whose name matches the exported constant and whose value differs from it:

CODATA2022.G_DEUTERON_NUCLEAR   # 0.8574382335  — as published by NIST
CODATA2022.G_DEUTERON           # ERROR: no such field
G_DEUTERON                      # 1.7140254606… — renormalized for a = (g-2)/2

Both the mass and the g-factor come from the active release, so the conversion stays self-consistent when you call set_release. The active release's unscaled values are also kept as the unexported _G_DEUTERON, _G_HELION, and _G_TRITON. See Physical Constants for the full explanation.

The other g-factors (G_ELECTRON, G_MUON, G_PROTON, G_NEUTRON) are exported exactly as tabulated.

Release coverage caveats

The gyromagnetic anomalies (ANOMALY_ELECTRON, ANOMALY_MUON) and the helion g-factor (G_HELION_NUCLEAR) were not individually tabulated by CODATA until the 2010 release. Using a pre-2010 release will still define these symbols; worth verifying their values against the source data if you are doing precision work prior to that release.