Sodium, which has long been regarded as one of the simplest metals, displays a great deal of structural, optical, and electronic complexities under compression. We compressed pure Na in the body-centered cubic structure to 52 GPa and in the face-centered cubic structure from 64 to 97 GPa, and studied the plasmon excitations of both structures using the momentum-dependent inelastic X-ray scattering technique. The plasmon dispersion curves as a function of pressure were extrapolated to zero momentum with a quadratic approximation. As predicted by the simple free-electron model, the square of the zero-momentum plasmon energy increases linearly with densification of the body-centered cubic Na up to 1.5-fold. At further compressions and in face-centered cubic Na above 64 GPa, the linear relation curves progressively toward the density axis up to 3.7-fold densification at 97 GPa. Ab initio calculations indicate that the deviation is an expected behavior of Na remaining a simple metal.high pressure | free-electron gas | generalized gradient approximation | alkali metal W ith a single conduction s electron per atom, solid sodium exemplifies the free-electron (FE) gas as the oldest and most studied model for describing metals (1). Under compression, Na continues to manifest how the FE metal evolves under extreme densification of the conduction electrons (2-5). Na is one of the most compressible materials. From 1 bar to 200 GPa, its volume reduces by fivefold, and its Wigner-Seitz radii, r s ¼ ½3∕ð4πn e Þ 1∕3 , reduces from 3.93 to 2.33, where n e is the conduction electron density (6, 7). During such a large range of electronic densification, Na exhibits numerous surprising behaviors, including melting at 118 GPa and 300 K (3), transitions to nine high-pressure phases including some truly exotic, complex crystal structures with as many as 512 atoms per unit cell (4,8,9), and eventually to an optically transparent, insulating solid (6, 10). These structural and optical studies indirectly revealed the unique physics of FE under compression (2). An interesting question is how and at what point Na loses its simple metallic behavior through this long journey.Scattering experiments measure the dynamic structure factor, Sðq;ωÞ, of electrons as a function of the momentum, q, and frequency, ω (or energy E ¼ ℏω). The theoretical solution of Sðq;ωÞ is one of the most studied many-body problems in condensedmatter physics, and has been approached by the random phase approximation (RPA) with various modifications (11, 12). However, RPA is a good approximation only for hypothetical, denseelectron metals with r s < 1. For real metals with r s ¼ 2-6, the exact theoretical solution of Sðq;ωÞ, except for possibly near q ¼ 0, remains uncertain; direct experimental constraints are essential. High pressure covers a very large tuning range for the r s of Na, providing an ideal testing ground for the basic physics of electronic dynamics and RPA theories.At low pressures, Sðq;ωÞ for most metals has been determined by electron energy-loss spectro...