Ruby is an alumina crystal doped with chromium ions. Under laser excitation these ions emit an intense red luminescence made of two sharp lines, R1 near 694.3 nm and R2 near 692.9 nm.

Compressing the crystal modifies the crystal field around the ion: the doublet shifts to the red in a reproducible way, by about 0.365 nm per GPa near ambient. Measuring the position of R1 therefore measures the pressure — for fifty years the most widely used secondary pressure gauge in diamond anvil cells.

For measurements away from ambient temperature, resistive heating or a cryostat, the SrB₄O₇:Sm²⁺ borate sensor is offered as an alternative: its 685.4 nm line is almost insensitive to temperature, where the ruby line needs a correction.

A pressure scale is a published calibration linking the line shift to pressure, established against a primary standard. Our software implements eight of them.

• Ruby: AIRAPT 2020, the international recommendation and the advised choice; Sokolova 2013; Dewaele 2008, quasi-hydrostatic helium up to about 150 GPa; Mao 1986 quasi-hydrostatic and Mao 1978 non-hydrostatic.
• Borate: Rashchenko 2015, Jing 2013 for non-hydrostatic media, and Datchi 1997 with its built-in correction above 500 K.

Below about 50 GPa all the ruby scales agree to better than 2 percent: what matters is quoting the scale used in your publications. Above that they diverge and the choice becomes significant.

The ruby R1 line also shifts to the red as temperature rises, in the same direction as the pressure effect. Near ambient the sensitivity is about +0.00746 nm per kelvin, which biases the pressure by roughly 0.02 GPa per kelvin: an error of 50 K is worth about 1 GPa.

The standard published correction is applied symmetrically to the measurement and to the zero reference, so that just after a zero calibration the pressure stays exactly zero whatever the temperature — only the difference matters. Benches with their own temperature calibration can enter their own coefficients per temperature zone.

Each reading carries an estimated uncertainty combining the uncertainty of the scale itself and the uncertainty on the line position, about half a detector pixel converted into GPa through the local slope of the scale.

This figure is an instrumental repeatability estimate, not a complete uncertainty budget: it includes neither the uncertainty of the entered temperature, nor the pressure gradients inside the cell, nor the departure from hydrostaticity of the transmitting medium — contributions the experimenter must assess.

Hydrostaticity indicator: fitting the R2 line separately gives the R1–R2 splitting, about 1.4 nm at ambient. Its evolution, together with the broadening of the lines, signals the build-up of non-hydrostatic stresses in the cell.