Total: 1
The condensation of photons into a macroscopically populated ground state, a superradiant phase transition (SRPT), is one of the most striking predictions of cavity quantum electrodynamics (cQED), yet has resisted experimental realization in equilibrium for over fifty years. Whether such a transition can survive at all in light-matter coupled systems is still debated, with a widely established \textit{No-Go} theorem ruling it out across several models. Graphene cyclotron transitions ultrastrongly coupled to terahertz (THz) cavities have been at the heart of this debate. At leading order, the linear Dirac dispersion of electrons does not generate the diamagnetic $\vec{A}^2$ term that enforces the theorem, making graphene the cleanest candidate for a No-Go-evading phase transition. Here, we present the first THz spectroscopic measurements of an encapsulated monolayer graphene flake ultrastrongly coupled to a single complementary split-ring resonator. By tuning the carrier density we sweep the system from weak coupling into the ultrastrong regime, reaching a normalized coupling $Ω_\text{R}/ω_\text{cav} \approx 0.4$. This is well into the range where a second order SRPT would manifest by a softening of the lower polariton branch, which we do not observe. The full polariton dispersion is instead quantitatively reproduced by a Hopfield Hamiltonian derived from first principles using a near-field model that accounts for the sub-wavelength character of the cavity. Our results establish an experimental baseline for predictions of cavity-driven phase transitions in two-dimensional Dirac systems, and rule out a No-Go-evading SRPT in graphene Landau polaritons up to the strongest couplings accessible today, with direct implications for proposals invoking vacuum-induced order in solid-state cQED.