Phantom crossing or dark interaction?

Technical Summary: Phantom Crossing or Dark Interaction?

Problem Statement
Recent cosmological analyses combining DESI BAO, Planck CMB, and Pantheon+ supernova data have suggested the possibility of a "phantom" phase in the expansion of the Universe, where the dark energy equation of state ($w_{de}$) crosses the phantom divide ($w_{de} < -1$). In General Relativity, phantom dark energy is generally considered unphysical due to instabilities. This paper investigates whether this apparent phantom behavior is an intrinsic property of dark energy or an effective phenomenon arising from a non-gravitational interaction between dark matter and dark energy. A primary challenge in this field is the degeneracy between the interaction term $Q(z)$ and the dark energy equation of state $w_{de}(z)$; background observations alone cannot distinguish between the two without additional assumptions.

Methodology
To break the degeneracy between the interaction $Q(z)$ and the intrinsic equation of state $w_{de}(z)$, the authors adopt a specific modeling strategy:

1. Intrinsic Dark Energy: They assume a physically motivated "thawing quintessence" form for the intrinsic equation of state, $w_{de}(z)$, described by an analytic function (Eq. 2.12) dependent on a present-day parameter $w_q$. Crucially, they do not impose a prior forcing $w_q \geq -1$, allowing the data to determine if the intrinsic sector is phantom or non-phantom.
2. Effective Dark Energy: They assume the effective equation of state, $w_{de}^{eff}(z)$, which is what observations actually constrain, follows the standard $w_0, w_a$ parametrization (Eq. 2.13).
3. Interaction Reconstruction: By specifying both $w_{de}(z)$ and $w_{de}^{eff}(z)$, the interaction term $Q(z)$ and the effective dark matter equation of state $w_{dm}^{eff}(z)$ are uniquely determined via the modified conservation equations (Eqs. 2.2–2.8).
4. Data Analysis: The model is constrained using a combination of DESI DR2 BAO measurements, Planck 2018 CMB compressed likelihoods, and the Pantheon+ Type Ia supernova compilation. The authors perform a Bayesian analysis to derive posterior distributions for cosmological parameters ($w_q, w_0, w_a, \Omega_m, H_0$, etc.) and reconstruct the redshift evolution of the interaction.
5. Model Comparison: The interacting model (IntModel) is compared against the standard flat CPL parametrization (fCPL) using multiple statistical criteria: $\chi^2$, Bayesian evidence ($\ln Z$), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Deviance Information Criterion (DIC).

Key Results

• Intrinsic vs. Effective Behavior: The data favors a non-phantom intrinsic dark energy equation of state ($w_q > -1$) without imposing a non-phantom prior. The best-fit value is $w_q = -0.885 \pm 0.058$. Conversely, the effective equation of state ($w_{de}^{eff}$) shows a phantom-like behavior at higher redshifts ($z \gtrsim 1$), but this is not statistically significant (well within $1\sigma$). At low redshifts ($z \lesssim 1$), the effective behavior is non-phantom with moderate significance (up to $3\sigma$).
• Interaction Dynamics: The interaction function $Q(z)$, normalized as $A(z) = Q/(3H\rho_{de})$, shows a sign change. The interaction is negative (energy flow from dark energy to dark matter) at high redshifts ($z \gtrsim 0.5$) and positive (energy flow from dark matter to dark energy) at low redshifts.
• Significance of Interaction: While the sign change at high redshift is not significant, a non-zero interaction is favored at more than $3\sigma$ at $z \sim 0.3$. At this redshift, energy flows from dark matter to dark energy.
• Dark Matter Properties: There is no significant evidence for a non-zero effective dark matter equation of state ($w_{dm}^{eff}$), except for a $\sim 2\sigma$ indication of $w_{dm}^{eff} < 0$ at $z \sim 0.3$.
• Model Selection: The interacting model (IntModel) provides a fit comparable to the standard fCPL model. While the BIC strongly prefers fCPL (due to the penalty for the extra parameter $w_q$), other criteria (AIC, DIC, and Bayesian evidence) show no significant preference for fCPL over IntModel. The authors conclude that the data does not show significant evidence for a preference of the standard model over the interacting one.

Significance and Claims
The paper claims that the apparent "phantom crossing" suggested by recent data may not be a physical reality of dark energy but rather an effective signature of a dark sector interaction. By reconstructing the interaction $Q(z)$ using a physically motivated intrinsic thawing quintessence model, the authors demonstrate that:

1. The intrinsic dark energy remains non-phantom, resolving the theoretical unphysicality of phantom energy in General Relativity.
2. A non-zero interaction between dark matter and dark energy is statistically favored at $z \sim 0.3$ ($>3\sigma$).
3. The energy transfer direction reverses around the epoch of dark matter-dark energy equality ($z \sim 0.5$), flowing from dark matter to dark energy at late times and vice versa at early times.

The authors conclude that interacting dark energy models are viable alternatives to both the standard $\Lambda$CDM model and phantom dark energy scenarios, offering a consistent explanation for current observational tensions without requiring unphysical intrinsic dark energy properties.