Is Dark Matter Really Matter?

Abstract

In the standard model of cosmology, it is assumed that dark matter is pressureless with equation of state $w=0$ and dark energy has $w=-1$. We test these assumptions jointly using DESI DR2 distance measurements, including the recent Lyman-$\alpha$ full-shape Alcock-Paczynski (AP) information, DES supernovae, and two complementary CMB treatment. When constant $w_{dm}$ and $w_{de}$ are varied together, we find $w_{dm}=0.000968^{+0.000501}_{-0.000496}$ and $w_{de}=-0.9380^{+0.0259}_{-0.0262}$ (68%). With an alternative CMB treatment that marginalizes over the lensing spectrum, the corresponding constraints are $w_{dm}=0.000870^{+0.000408}_{-0.000410}$ and $w_{de}=-0.9353^{+0.0258}_{-0.0254}$. Both standard $\Lambda$CDM values are disfavored at approximately $2\sigma$ in the joint extension. Neither parameter departs significantly from its standard value when only that parameter is varied. This behavior arises because late-time distances favor $w_{de}>-1$, while maintaining the early-Universe physical matter density requires a compensating positive $w_{dm}$, which changes the mapping to the matter density today. Allowing dynamical dark energy clarify further on complexity of the situation: phantom crossing for dark energy makes $w_{dm}=0$ consistent with the data, whereas a positive $w_{dm}$ preference persists when crossing is forbidden. Interestingly, the Pad'e-$w$ parameterization that provides a flexible description of a class of quintessence models (with no phantom crossing), along with $w_{dm}$ free, is even mildly favored over the phantom-crossing $w_0w_a$ model according to both the best-fit $\chi^2$ and the DIC under both CMB treatments. One can conclude that the apparent preference for phantom crossing may instead reflect deviations in the dark-matter sector rather than dark-energy dynamics alone. [abridged]