Testing signatures of phantom crossing through full-shape galaxy clustering analysis

TL;DR

This work investigates late-time phantom crossing in dynamical dark energy by applying a full-shape (FS) galaxy clustering analysis to BOSS DR12, augmented with DESI-DR2 BAO and SN Ia data, in a framework largely independent of CMB constraints. It tests two DE parameterizations—the CPL $w_0$–$w_a$ form and a minimal $w_{\dagger}$CDM transition model—within the EFTofLSS, including AP distortions and a bias expansion, to extract growth- and geometry-related information from FS multipoles. Across data combinations, the analysis reveals strong evidence for deviations from ΛCDM, with a phantom crossing indicated around $z\sim0.4$–$0.5$ (CPL) and negative transition amplitudes in $w_{\dagger}$CDM; the strength of the evidence grows with added SN Ia data, yielding up to ~5σ in some cases, while S_8 remains lower than Planck-based expectations and H_0 hovers near 67 km/s/Mpc. Importantly, these results demonstrate that FS clustering provides a powerful, independent probe of late-time DE dynamics, maintaining robustness against the absence of CMB data and offering complementary insights to geometric probes for testing beyond-ΛCDM scenarios.

Abstract

Recent observations of baryon acoustic oscillations (BAO) from the Dark Energy Spectroscopic Instrument (DESI) survey, when combined with measurements of the cosmic microwave background (CMB) and Type Ia supernovae (SNIa), provide compelling evidence for a phantom crossing at late times, along with statistically significant deviations from the standard $Λ$CDM model. In this work, we investigate the role of redshift-space galaxy clustering data by employing the pre-reconstruction full-shape (FS) galaxy power spectrum from the Baryon Oscillation Spectroscopic Survey (BOSS) data release 12 (DR12) sample. This dataset is analyzed in combination with BAO measurements from DESI data release 2 (DR2) and various SNIa samples. Our analysis demonstrates that the joint combination of these datasets can yield deviations from $Λ$CDM at a significance level of up to $\sim 5σ$, suggesting strong indications that the dark energy equation of state parameter $w(z)$ may have crossed the phantom divide ($w = -1$) in the redshift range $z \sim 0.4$-$0.5$. The precise location and strength of this crossing depend on the adopted theoretical parameterizations. Importantly, our results reveal that this trend persists even in the absence of CMB data, underscoring the robustness of the FS power spectrum as a powerful and independent probe for testing dark energy models and for distinguishing between competing cosmological scenarios.

Figures (7)

• Figure 1: Galaxy power spectrum multipoles: monopole ($\ell = 0$, solid lines) and quadrupole ($\ell = 2$, dashed lines). We vary the values of the DE parameters $w_0$ and $w_a$, as indicated in the legend, showing that different choices lead to significant changes in the galaxy power spectrum.
• Figure 2: Left panel: 68% and 95% confidence contours for $w_a$ and $w_0$ from different data sets (see the legend) in the $w_0 w_a$CDM model. Right panel: Same contours for $\Delta$ and $z_{\dagger}$ in the $w_{\dagger}$CDM model.
• Figure 3: Statistical reconstruction of the DE EoS for the $w_0w_a$ and $w_{\dag}$ parametrizations, showing 1$\sigma$ and 2$\sigma$ confidence levels from the combined FS+DESI+DESY5 dataset analyses.
• Figure 4: Whisker plot displaying 68% confidence intervals on $S_8$, as inferred from a range of recent analyses under different cosmological assumptions. Square markers indicate results obtained using the CPL parameterization: $S_8 = 0.775^{+0.047}_{-0.047}$ from FS+DESI-DR2 [This work], $S_8 = 0.758^{+0.045}_{-0.045}$ from FS+DESI-DR2+DESY5 [This work], $S_8 =0.711^{+0.028}_{-0.028}$ from FS+DESI-DR1+PP Chen:2024vuf, $S_8 =0.841^{+0.012}_{-0.012}$ from CMB+DESI-DR2 Mirpoorian:2025rfp, $S_8 =0.831^{+0.025}_{-0.025}$ from CMB$\kappa\times$LRG+DESI-DR2 Sabogal:2025jbo, and $S_8 =0.772^{+0.034}_{-0.034}$ from LOW-Z$\times$(KiDS+FS+CMB$_{\kappa,\mathrm{ISW}}$) Reeves:2025axp. Diamond markers correspond to standard $\Lambda$CDM analyses: $S_8 =0.679^{+0.035}_{-0.031}$ and $S_8 =0.749^{+0.048}_{-0.048}$ from FS with simulation and conservative priors Ivanov:2024xgb, $S_8 =0.808^{+0.017}_{-0.017}$ from CMB$\kappa$+ACT-DR6+DESI-DR1 (QSO+OLG) Maus:2025rvz, $S_8 =0.775^{+0.022}_{-0.022}$ from CMB$\kappa$+ACT-DR6 Sailer:2024jrx, $S_8 =0.922^{+0.059}_{-0.073}$ from CMB$\kappa\times$DESI-DR1 (QSO) deBelsunce:2025qku, and $S_8 =0.792^{+0.022}_{-0.022}$ from LOW-Z$\times$DES/KiDS Lange:2023khv. Circular markers denote alternative scenarios, yielding $S_8 =0.627^{+0.103}_{-0.127}$ from HSC-Y3+ACT-DR6+CMB+DESI-DR1 Terasawa:2025fpf and $S_8 =0.775^{+0.026}_{-0.026}$ from DESI-DR1+PP+RSD+CC Sabogal:2024yha. The salmon and hatched light blue bands indicate the $1\sigma$ and $2\sigma$ confidence intervals from Planck CMB and the latest KiDS results, respectively.
• Figure 5: 1D posteriors and 2D marginalized probability contours for the bias parameters $b_1^{(i)}$, $b_2^{(i)}$, and $b_{G_2}^{(i)}$ ($i=1,\ldots,4$), corresponding to the four independent subsamples defined by redshift and sky region (LOWZ/CMASS in NGC/SGC). Results are shown for all cosmological models considered in this work, as indicated in the legend.