UV-complete and stable Quintom Dark Energy models in the light of DESI DR2

Abstract

We propose that Quintom dark energy, the simplest framework allowing crossing of the cosmological-constant boundary, admits a natural UV completion in a 5D anisotropic orbifold lattice: the Non-Perturbative Gauge-Higgs Unification (NPGHU) model. In this setup, a bulk 5D SU(2) gauge field projects on the 4D boundary to a complex scalar and a U(1) gauge field, identified with the dynamical dark-energy sector, while the Standard Model and dark matter remain localized in four dimensions. At late times, bulk-induced dimension-6 higher-derivative operators generate both physical and phantom scalar and gauge degrees of freedom. We show that the resulting 4D effective action is a modified Quintom model whose background equation of state can naturally realize Quintom-B behavior. A crucial contribution arises from the massive gauge ghost, allowing an excellent fit to DESI data with negligible fine-tuning, unlike standard Quintom scenarios. We further show that the inherited properties of the NPGHU construction e.g. absence of fundamental ghost instabilities, absence of potential terms and a finite low-energy cutoff $Λ$ associated with approximate Lorentz invariance, play a central role in the consistency of the effective theory under linear perturbations and vacuum decay. For the most natural regime, $Λ\approx {\cal O}(10)H_0$, the model remains robust despite the presence of IR phantom modes. Our results provide a natural and predictive framework in which Quintom dark energy can be consistently embedded in a fundamental theory.

Figures (6)

• Figure 1: The 5d anisotropic orbifold lattice spacetime, constracted for the first time in Irges:2006hgKnechtli:2007ea, where we live on the 4d $U(1)$ brane at the fixed point $n_5 = 0$. At each point on the extra dimension, $n_5 = \{1, ..., N_5 -1 \}$, lies a 4d brane with an SU(2) gauge field on it.
• Figure 2: A 2d cartoon picture of the phase diagram of our model. There is the Confined, the Hybrid and the Higgs phase, all of them separated by a 1st order phase transition. Each point on the blue-line corresponds to a finite cut-off. In this work we are interested only in the Higgs-Hybrid transition part which lies after the triple point (green dot) with $\gamma \ll 1$. There the purple-dot represents the energy scale of the localization 1st order phase transition, at the finite cut-off $\mu = \Lambda$, while the orange-dot denotes the energy scale of our universe today ($\mu = H_{\rm m, 0}$). The black dashed line suggests that today the universe lives in the vicinity of the phase transition in the Higgs phase but not far from $\Lambda$.
• Figure 3: Left panel: The evolution of the R-ghost $\chi_{0,n}$ remains always subdominant compared to the phantom $\varphi_{2,n}$ for $|\lambda_\chi||\varphi_{2,\alpha}|^2/H^2_{\rm m,0} = {\cal O}(1)$ and $\chi_{0,n, \alpha} = 0$. Right panel: Given $\lambda_\chi < 0$, increasing the quartic coupling we reach a number above which $\chi_{0,n}$ surpasses the phantom at the very late redshift and dominates the EoS.
• Figure 4: The evolution of the scalar and gauge physical and phantom fields as a function of the $N = \ln a/a_0$ under the mass-ratio $m_{A_2}^2/H_{\rm m,0}^2 = m_{\phi_2}^2/H_{\rm m,0}^2 = 3$ and the democratic initial conditions $|\varphi_{1,n, \alpha}| \approx {A}_{2,n,\alpha} \approx |\varphi_{2,n, \alpha}| = 1$. As we have explained below Eq. (\ref{['freeparameters']}) the evolution of $A_{i,1,n}$ is negligible compared with the others.
• Figure 5: The evolution of $w_q$ (cyan line) as a function of the normalized scale factor for two benchmark initial conditions and fixed ghost field masses, $m_{A_2}^2/H_{\rm m,0}^2 = m_{\phi_2}^2/H_{\rm m,0}^2 \approx 3$. The red line corresponds to the cosmological boundary $w_{\rm \Lambda CDM} = -1$. Left panel: A Quintom-B-like model is realized for the initial kinetic term hierarchy $|\dot\varphi_{1,n, \alpha}| > |\dot\varphi_{2,n, \alpha}|$ and $e^{-2N_\alpha}(\dot A_{2, n, \alpha})^2/3 <|\dot\varphi_{1,n, \alpha}|^2 - |\dot\varphi_{2,n, \alpha}|^2 < e^{-2N_\alpha} (\dot A_{2, n, \alpha})^2/ 2$. Right panel: A quintessence-like behavior is obtained for the inverse hierarchy, $|\dot\varphi_{1,n, \alpha}| < | \dot\varphi_{2,n, \alpha} |$.