Geometric criticality in the driven Jaynes-Cummings model Ken Chen, Jia-Hao Lü, Hao-Long Zhang, Fan Wu, Wen Ning, Zhen-Biao Yang, Shi-Biao Zheng Abstract When the photonic mode in the Jaynes-Cummings model is driven by an external classical field, the system can undergo the photon-blockade breakdown phase transition at a critical point. Such a phase transition has been detailedly investigated, but the critical properties of the eigenstates remain largely unexplored so far. We here study the geometric criticality associated with these eigenstates. The amplitude and phase of the drive serve as the control parameter of the governing Hamiltonian. We find the quantum metric and Berry curvature tensors for each eigenstate display divergent behaviors in the critical region. More importantly, the divergence associated with bright eigenstates is much more pronounced than that for the unique dark state. Our theoretical results can be experimentally confirmed in circuit quantum electrodynamics systems, where the driven Jaynes-Cummings model has been realized. I INTRODUCTION In recent years, increasing interest has been paid to quantum geometry Törmä et al. 2022 ; Peotta et al. 2025 ; Törmä 2023 ; Liu et al. 2024 , quantified by the quantum geometric tensor (QGT) Provost and Vallee 1980 , which describes the inherent geometry of the eigenstate space of the Hamiltonian governing the behaviors of a quantum state. The real part of the QGT, which is symmetric and referred to as quantum metric, measures the distance between nearby eigenstates of the Hamiltonian. The quantum metric plays a critical role in understanding important physical phenomena in solid-state physics, exemplified by flat-band superconductivity Liu et al. 2024 . Its imaginary part, which is antisymmetric and known as Berry curvature Berry 1984 , provides information about phase change of the eigenstates Törmä 2023 . The integral of the Berry curvature over a closed manifold yields the Chern number, which quantifies the topology of the manifold. The Berry curvature can be thought as the fictitious magnetic field emanated by a topological monopole, and is responsible for intriguing electronic transport phenomena, e.g., quantum anomalous Hall effect Yu et al. 2010 ; Chang et al. 2013 ; Checkelsky et al. 2014 . In recent years, the QGT has been measured with different types of pseudospins, including NV centers in diamond Yu et al. 2020 , ultracold atoms Chen et al. 2022 , and superconducting qubits Roushan et al. 2014 ; Tan et al. 2019 , polariton systems Fläschner et al. 2016 , and Bloch electrons in solids Gianfrate et al. 2020 . Apart from laying the foundation for topological physics, the QGT provides a new insight to quantum phase transitions without resorting to order parameters Kim et al. 2025 ; Liao et al. 2021 ; Zanardi et al. 2007 ; Campos Venuti and Zanardi 2007 ; Rezakhani et al. 2010 . In the proximity of a quantum phase transition, the eigenstates of the Hamiltonian undergo a dramatic change even if the control parameter is only slightly changed. As such, it is expected that the QGT associated with each eigenstate exhibits divergent behaviors near the critical point. Such critical behaviors have been theoretically investigated in different systems, including the parametrically-driven nonlinear resonator Zhang et al. 2024 , Dicke model Zhu et al. 2024 , and parametrically-driven Tavis-Cummings model Lü et al. 2023 . These systems bear a common feature that the Hamiltonian is symmetric under the parity transformation. The corresponding phase transition originates from the spontaneous parity breaking of the parity. Despite these theoretical advancements, experimental observations of the critical geometric phenomena are still lacking. In this paper, we study the critical behaviors of entangled eigenstates of the driven Jaynes-Cummings model (JCM) Alsing et al. 1992 ; Carmichael 2015 ; Curtis et al. 2021 ; Fink et al. 2017 ; Lü et al. 2026 , where photonic mode is coupled to a qubit and driven by an external classical field. This drive breaks the U(1) symmetry of the Jaynes-Cummings model. Neither does the driven Jaynes-Cummings model have the parity symmetry. Such a system possesses a unique dark eigenstate and infinitely many bright doublets. The quasienergy splittings of the doublets vanish after crossing the critical point, where the drive strength is equal to half of the coupling strength between the qubit and the photonic mode. The critical behavior of the qubit in the dark state has been exploited for realizing robust quantum metrology Lü et al. 2026 . However, the critical geometric features have not been touched upon so far. We here investigate quantum geometry for such a critical system. The calculated quantum metric and Berry curvature for each eigenstate diverge at the critical point. The critical features associated with bright eigenstates are much stronger than those for the dark state. As neither the thermodynamical limit nor the scaling limit is required to observe the critical behaviors of the driven JCM, this is in distinct contrast with other qubit-photon systems Zhu et al. 2024 ; Lambert et al. 2004 ; Ashhab 2013 ; Hwang et al. 2015 . As such, our theoretical results can be experimentally demonstrated with presently available techniques, which would shed new light on critical phenomena in fully quantum-mechanical light-matter systems. II eigenstates and eigenenergies of the driven JCM We begin by considering the driven JCM Alsing et al. 1992 ; Dutra et al. 1994 ; Feng et al. 2015 , which describes a two-level system resonantly coupled to a quantum field mode while simultaneously driven by an on-resonance signal field. The interaction Hamiltonian in the interaction picture is given by ( \hbar=1 is set) \displaystyle H={} \displaystyle\Omega[a^{\dagger}|g\rangle\langle e|+a|e\rangle\langle g|+\eta(a^{\dagger}e^{-i\phi}+ae^{i\phi})/2], (1) where |g\rangle and |e\rangle are the ground and excited state of the two-level system, a^{\dagger} and a are the creation and annihilation operators for the quantized field mode, \Omega represents the coupling strength, and \eta characterizes the rescaled amplitude of the driven field with the phase \phi . Next, we solve the eigenvalue problem with \displaystyle H|\psi_{E}\rangle=E|\psi_{E}\rangle. (2) Here we expand the eigenstate as \displaystyle|\psi_{E}\rangle={} \displaystyle|\psi_{E}^{+}\rangle|e\rangle+|\psi_{E}^{-}\rangle|g\rangle, (3) where |\psi_{E}^{+}\rangle and |\psi_{E}^{-}\rangle are field states normalized, satisfying the condition of \langle\psi_{E}|\psi_{E}\rangle=1 . Therefore, the eigenvalue problem can be expressed as \displaystyle\Big[\frac{\Omega\eta}{2}(a^{\dagger}e^{-i\phi}+ae^{i\phi})-E\Big]|\psi_{E}^{+}\rangle+\Omega a|\psi_{E}^{-}\rangle={} \displaystyle 0, (4a) \displaystyle\Omega a^{\dagger}|\psi_{E}^{+}\rangle+\Big[\frac{\Omega\eta}{2}(a^{\dagger}e^{-i\phi}+ae^{i\phi})-E\Big]|\psi_{E}^{-}\rangle={} \displaystyle 0. (4b) Here we multiply Eq. ( 4a ) and Eq. ( 4b ) on the left by a^{\dagger} and a , respectively, which gives rise to \displaystyle\left\{-\left[\frac{\eta}{2}(a^{\dagger}e^{-i\phi}+ae^{i\phi})-\frac{E}{\Omega}\right]^{2}+a^{\dagger}a\right\}|\psi_{E}^{-}\rangle{} \displaystyle-\frac{\eta}{2}e^{i\phi}|\psi_{E}^{+}\rangle={} \displaystyle 0, (5a) \displaystyle\left\{-\left[\frac{\eta}{2}(a^{\dagger}e^{-i\phi}+ae^{i\phi})-\frac{E}{\Omega}\right]^{2}+aa^{\dagger}\right\}|\psi_{E}^{+}\rangle{} \displaystyle+\frac{\eta}{2}e^{-i\phi}|\psi_{E}^{-}\rangle={} \displaystyle 0. (5b) By combining Eq. ( 5a ) and Eq. ( 5b ), we obtain \displaystyle\mathcal{A}_{p}(E)\mathcal{A}_{m}(E)|\psi_{E}^{-}\rangle={} \displaystyle 0, (6) where \displaystyle\mathcal{A}_{p}(E)={} \displaystyle\mathcal{A}(E)+\frac{1}{2}\sqrt{1-\eta^{2}}, (7a) \displaystyle\mathcal{A}_{m}(E)={} \displaystyle\mathcal{A}(E)-\frac{1}{2}\sqrt{1-\eta^{2}}, (7b) \displaystyle\mathcal{A}(E)={} \displaystyle-\left[\frac{\eta}{2}(a^{\dagger}e^{-i\phi}+ae^{i\phi})-\frac{E}{\Omega}\right]^{2}+\frac{a^{\dagger}a+aa^{\dagger}}{2}, (7c) where \mathcal{A}_{p}(E) and \mathcal{A}_{m}(E) satisfy the commutation relation, i.e., [\mathcal{A}_{p}(E),\mathcal{A}_{m}(E)]=0 . Therefore, the general solution may be written as \displaystyle|\psi_{E}^{-}\rangle={} \displaystyle c_{p}|\psi_{E,p}^{-}\rangle+c_{m}|\psi_{E,m}^{-}\rangle, (8) where |\psi_{E,p}^{-}\rangle and |\psi_{E,m}^{-}\rangle are solutions to the equations \displaystyle\mathcal{A}_{p}(E)|\psi_{E,p}^{-}\rangle={} \displaystyle 0, (9a) \displaystyle\mathcal{A}_{m}(E)|\psi_{E,m}^{-}\rangle={} \displaystyle 0. (9b) Using Eq. ( 5a ) and Eqs. ( 7 ), we can write \displaystyle|\psi_{E}^{+}\rangle={} \displaystyle(\frac{\eta}{2}e^{i\phi})^{-1}\{\mathcal{A}_{p,m}-\frac{1}{2}[1\pm\sqrt{1-\eta^{2}}]\}|\psi_{E}^{-}\rangle, (10) where the \pm notation indicates two alternative forms with + corresponding to the p-subscripted form and - to the m-subscripted form. According to Eqs. ( 9 ) and Eq. ( 10 ), we have \displaystyle|\psi_{E}^{+}\rangle=-(\eta e^{i\phi})^{-1}\{{} \displaystyle c_{p}[1+\sqrt{1-\eta^{2}}]|\psi_{E,p}^{-}\rangle (11) \displaystyle+c_{m}[1-\sqrt{1-\eta^{2}}]|\psi_{E,m}^{-}\rangle\}. To solve Eqs. ( 9 ), we use squeezing and displacement transformations to convert Eqs. ( 9 ) into eigenvalue equations for a harmonic oscillator. We multiply Eqs. ( 9 ) on the left by D^{\dagger}(\alpha)S^{\dagger}(\xi) and transform the operators \mathcal{A}_{p}(E) and \mathcal{A}_{m}(E) using \displaystyle D^{\dagger}(\alpha)aD(\alpha)={} \displaystyle a+\alpha, (12a) \displaystyle S^{\dagger}(\xi)aS(\xi)={} \displaystyle a\cosh r-a^{\dagger}e^{i\theta}\sinh r, (12b) where D(\alpha)=e^{\alpha a^{\dagger}-\alpha^{*}a} and S(\xi)=e^{-\frac{1}{2}(\xi a^{\dagger 2}-\xi^{*}a^{2})} . So when we choose \displaystyle\alpha={} \displaystyle-\eta(1-\eta^{2})^{-3/4}\frac{E}{\Omega}e^{-i\phi}, (13a) \displaystyle\xi={} \displaystyle re^{-2i\phi},\quad r=\frac{1}{4}\ln\left(1-\eta^{2}\right), (13b) Eqs. ( 9 ) can be substituted by the following equations \displaystyle a^{\dagger}a|\tilde{\psi}_{E;p}^{-}\rangle={} \displaystyle\left[(\frac{E}{\Omega})^{2}(1-\eta^{2})^{-3/2}-1\right]|\tilde{\psi}_{E;p}^{-}\rangle, (14a) \displaystyle a^{\dagger}a|\tilde{\psi}_{E;m}^{-}\rangle={} \displaystyle(\frac{E}{\Omega})^{2}(1-\eta^{2})^{-3/2}|\tilde{\psi}_{E;m}^{-}\rangle, (14b) where |\tilde{\psi}_{E;p}^{-}\rangle=D^{\dagger}(\alpha)S^{\dagger}(r)|\psi_{E;p}^{-}\rangle and |\tilde{\psi}_{E;m}^{-}\rangle=D^{\dagger}(\alpha)S^{\dagger}(r)|\psi_{E;m}^{-}\rangle . The solutions of Eqs. ( 14 ) are Fock states if and only if the constants on the right-hand sides are restricted to the discrete set of non-negative integers. From Eqs. ( 14 ), we obtain the discrete eigenenergies as \displaystyle E_{0}={} \displaystyle 0, (15a) \displaystyle E_{n,\pm}={} \displaystyle\pm\sqrt{n}\Omega(1-\eta^{2})^{3/4}\quad(n=1,2,...). (15b) The corresponding states are \displaystyle|\tilde{\psi}_{E;p}^{-}\rangle={} \displaystyle|n-1\rangle,\quad(n=1,2,...) (16a) \displaystyle|\tilde{\psi}_{E;m}^{-}\rangle={} \displaystyle\left\{\begin{aligned} &|0\rangle\\ &|n\rangle,\quad(n=1,2,...).\end{aligned}\right. (16b) Furthermore, we can obtain \displaystyle|\psi_{E;p}^{-}\rangle={} \displaystyle S(r)D(\alpha)|n-1\rangle=|r,\alpha,n-1\rangle, (17a) \displaystyle|\psi_{E;m}^{-}\rangle={} \displaystyle\left\{\begin{matrix}S(r)D(\alpha)|0\rangle=|r,\alpha,0\rangle\\ S(r)D(\alpha)|n\rangle=|r,\alpha,n\rangle.\end{matrix}\right. (17b) The ratio of constants c_{p} and c_{m} in Eq. ( 8 ) is determined by imposing the constraint from Eq. ( 4b ), while their separate values are fixed through the normalization \displaystyle\langle\psi_{E}|\psi_{E}\rangle=\langle\psi_{E}^{+}|\psi_{E}^{+}\rangle+\langle\psi_{E}^{-}|\psi_{E}^{-}\rangle=1. (18) The calculation, while conceptually straightforward, involves considerable algebraic manipulation. We therefore present only the final results as follows \displaystyle|\psi_{0}\rangle={} \displaystyle S(\xi)|0\rangle|\phi_{0}\rangle, (19a) \displaystyle|\psi_{n,\pm}\rangle={} \displaystyle S(\xi)D(\alpha_{n,\pm})(|n-1\rangle|\phi_{1}\rangle\pm|n\rangle|\phi_{0}\rangle)/\sqrt{2}, (19b) where S(\xi) and D(\alpha_{n,\pm}) are the squeezing and displacement operators for the photonic field, respectively, with \xi=\frac{1}{4}e^{-2i\phi}\ln\left(1-\eta^{2}\right) and \alpha_{n,\pm}=\mp\sqrt{n}\eta e^{-i\phi} . The expressions of |\phi_{0}\rangle and |\phi_{1}\rangle are given by \displaystyle|\phi_{0}\rangle={} \displaystyle c_{+}|g\rangle-e^{-i\phi}c_{-}|e\rangle, (20a) \displaystyle|\phi_{1}\rangle={} \displaystyle c_{+}|e\rangle-e^{i\phi}c_{-}|g\rangle, (20b) with c_{\pm}=\left(1\pm\sqrt{1-\eta^{2}}\right)^{1/2}/\sqrt{2} . III critical QGT Figure 1: The quantum metric \mathcal{G}_{\eta\eta} (a), \mathcal{G}_{\phi\phi} (c) and the Berry curvature \mathcal{B}_{\eta\phi} (b), \mathcal{B}_{\phi\eta} (d) with respect to \eta for \phi=0 . The eigenstates |\psi_{i}\rangle\ (i=0,1,2,3,4,5) correspond to n=0,1,2,3,4,5 , respectively. The quantum metric \mathcal{G}_{\eta\phi} , \mathcal{G}_{\phi\eta} and the Berry curvature \mathcal{B}_{\eta\eta} , \mathcal{B}_{\phi\phi} are equal to zero, and not shown here. Here, we apply the QGT to investigate the singular behavior of the driven JCM in the vicinity of the critical points. We consider a nondegenerate quantum state |\psi(\lambda)\rangle in a parameter space spanned by N dimensionless parameters \lambda(\lambda_{1},\lambda_{2},...,\lambda_{N}) , the QGT is defined as Provost and Vallee 1980 \displaystyle Q_{\mu\nu}={} \displaystyle\langle\partial_{\mu}\psi(\lambda)|[1-\mathcal{P}(\lambda)]|\partial_{\nu}\psi(\lambda)\rangle (21) \displaystyle={} \displaystyle\mathcal{G}_{\mu\nu}-\frac{i}{2}\mathcal{B}_{\mu\nu}, where \{|\psi(\lambda)\rangle\} forms a complete orthonormal basis space and \mathcal{P}(\lambda)=|\psi(\lambda)\rangle\langle\psi(\lambda)| is the Hilbert eigenspace projection operator. Its real component defines the quantum metric, \mathcal{G}_{\mu\nu}={\rm Re}\{Q_{\mu\nu}\}=\mathcal{G}_{\nu\mu} , while imaginary component is related to the Berry curvature, \mathcal{B}_{\mu\nu}=-2{\rm Im}\{Q_{\mu\nu}\}=-\mathcal{B}_{\nu\mu} . A remarkable feature of the QGT is that its divergent behavior in the vicinity of the critical point Campos Venuti and Zanardi 2007 ; Tan et al. 2019 ; Gutiérrez-Ruiz et al. 2021 . To elucidate this behavior, we derive a perturbative expression for QGT. Substituting the identity operator \mathbb{I}=\sum_{m}|\psi_{m}\rangle\langle\psi_{m}| into Eq. ( 21 ) and using \displaystyle\langle\psi_{m}|\partial_{\nu}\psi_{n}\rangle=\frac{\langle\psi_{m}|\frac{\partial H}{\partial\nu}|\psi_{n}\rangle}{E_{n}-E_{m}}\quad(m\neq n), (22) which is determined by the eigenvalue equation H|\psi_{n}\rangle=E_{n}|n\rangle , we reformulate Eq. ( 21 ) as follows GU 2010 : \displaystyle Q_{n,\mu\nu}={} \displaystyle\sum_{m\neq n}\frac{\langle\psi_{n}|\frac{\partial H}{\partial\mu}|\psi_{m}\rangle\langle\psi_{m}|\frac{\partial H}{\partial\nu}|\psi_{n}\rangle}{(E_{m}-E_{n})^{2}}. (23) This expression confirms that the QGT components—the quantum metric tensor and Berry curvature—become singular at stationary points of the QPT, marked by ground-state level crossings. We consider the QGT space formed by the parameters \eta and \phi , as shown in Fig. 1 . Fig. 1 (a) and Fig. 1 (c) show the real parts of the QGT, i.e., \mathcal{G}_{\eta\eta} and \mathcal{G}_{\phi\phi} , with respect to \eta , over the ground state |\psi_{0}\rangle and the forefront five low-lying excited states |\psi_{i}\rangle\ (i=1,2,3,4,5) ; while Fig. 1 (b) and Fig. 1 (d) show the imaginary parts of the QGT, i.e., \mathcal{B}_{\eta\phi} and \mathcal{B}_{\phi\eta} , as the function of \eta , over the same states. It is clearly shown that, all \mathcal{G}_{\eta\eta} , \mathcal{G}_{\phi\phi} , \mathcal{B}_{\eta\phi} and \mathcal{B}_{\phi\eta} , in the excited states |\psi_{i}\rangle\ (n=1,2,3,4,5) , exhibit significantly greater divergence compared to the ground state |\psi_{0}\rangle . Furthermore, the behavior of the divergence becomes stronger as the energy level separation of |\psi_{i}\rangle\ (n=1,2,3,4,5) from |\psi_{0}\rangle becomes larger, which can also be seen from Eq. ( 23 ). Figure 2: The Bures metric g_{\eta\eta} (a) and g_{\phi\phi} (b) as a function of \eta . Here |\psi_{i}\rangle , |\psi_{iq}\rangle and |\psi_{ir}\rangle ( i=0,1,2 ) represent the eigenstates of the composite system, the two-level subsystem and the quantum field mode, respectively. The g_{\eta\phi} and g_{\phi\eta} are equal to zero, and not shown here. In quantum information geometry, the Bures metric defines a natural Riemannian structure on the space of density matrices, which is derived from the Bures distance. For two density matrices \rho_{1} and \rho_{2} , the Bures distance is defined via the Uhlmann fidelity \mathcal{F}\left(\rho_{1},\rho_{2}\right)=\left(\text{tr}\sqrt{\sqrt{\rho_{1}}\rho_{2}\sqrt{\rho_{1}}}\right)^{2} Uhlmann 1976 as \displaystyle dS^{2}\left(\rho_{1},\rho_{2}\right)=2\left(1-\sqrt{\mathcal{F}\left(\rho_{1},\rho_{2}\right)}\right). (24) The corresponding Bures metric g_{ij} is given by \displaystyle\sum_{i,j}{g}_{ij}(\textbf{x})dx_{i}dx_{j}=dS^{2}\left(\rho_{\textbf{x}},\rho_{\textbf{x}+d\textbf{x}}\right), (25) which quantifies the distinguishability between two neighboring density matrices, \rho_{\mathbf{x}} and \rho_{\mathbf{x}+d\mathbf{x}} , in a parameter space \mathbf{x} . As shown in Fig. 2 , the Bures metric components g_{\eta\eta} and g_{\phi\phi} become divergent in the proximity of the critical point. The eigenstates of the composite system are denoted by |\psi_{i}\rangle ( i=0,1,2 ), which describe the product state of the eigenstates of the two-level subsystem ( |\psi_{iq}\rangle ) and the quantum field mode ( |\psi_{ir}\rangle ), i.e., |\psi_{i}\rangle=|\psi_{iq}\rangle\otimes|\psi_{ir}\rangle . Moreover, it is shown that Bures metric components g_{\eta\eta} and g_{\phi\phi} diverge much more strongly in the excited states |\psi_{i}\rangle(i=1,2) than in the ground state |\psi_{0}\rangle . Interestingly, this divergent behavior is monotonically enhanced as the energy-level distance of |\psi_{i}\rangle\ (i=1,2) from the ground state |\psi_{0}\rangle increases. It’s noted that the Bures metric components g_{\eta\eta} and g_{\phi\phi} of the composite system are dominated by contributions from the quantum field mode eigenstates |\psi_{ir}\rangle ( i=0,1,2 ). IV Experimental feasibility The difficulty in the experimental preparation of a target quantum state stems from the fact that, although it is theoretically well-defined, its direct initialization is often prohibitively challenging. To circumvent this, a widely employed strategy leverages the quantum adiabatic theorem Lü et al. 2026 ; Albash and Lidar 2018 ; Farhi et al. 2000 ; Born and Fock 1928 ; Kato 1950 ; Sarandy and Lidar 2005 . Such an approach initializes the system in an easily preparable eigenstate at an initial parameter value, subsequently applying a slow, continuous drive to evolve the Hamiltonian towards the final parameter value. Within our framework, the dynamics of the driven JCM can be governed by the time-dependent critical parameter \eta=\sqrt{1-[(kt)^{4/3}+1]^{-1}} Curtis et al. 2021 , where k determines the ramping velocity. Initial conditions at t=0 are set to |0\rangle|g\rangle for the ground state and to the superposition state (|n-1\rangle|e\rangle\pm|n\rangle|g\rangle)/\sqrt{2} for the i th eigenstates, respectively. However, the experimental feasibility of this approach is fundamentally constrained by the adiabatic condition, which requires that the driving rate ( \sim 1/t , t is the evolution time) be much smaller relative to the spectral gaps along the parameter trajectory Polkovnikov et al. 2011 ; Kolodrubetz et al. 2017 . This requirement becomes particularly stringent near quantum critical regions, where the protocol’s driving speed must be drastically reduced. Notably, the requisite slow-down is even more pronounced for the excited states. This substantial prolongation of adiabatic evolution time introduces a fundamental limitation for implementing robust quantum information protocols under realistic experimental conditions. To overcome this challenge, a possible solution resorts to so-called counterdiabatic driving, which is also known as the adiabatic shortcut del Campo 2013 ; Sels and Polkovnikov 2017 ; Finžgar et al. 2025 ; Morawetz and Polkovnikov 2025 . By introducing auxiliary control fields that actively suppress non-adiabatic excitations, it enables, in principle, a transitionless quantum evolution within an arbitrarily short duration, offering a pathway toward robust finite-time quantum state engineering in critical systems. Furthermore, a recent study has demonstrated an excited state preparation technique that exploits the physical coupling between photons and electrons to directly convert between ground and excited states Burton and Filip 2025 . V Conclusions In conclusion, we have studied the quantum geometry associated with the eigenstates of the driven JCM. Our results reveal that both the quantum metric and Berry curvature exhibit critical behaviors near the photon-blockade breakdown phase transition. The divergent features of these quantum geometric properties for bright eigenstates are much more pronounced than those for the dark eigenstate. Our work provides a direct insight into critical phenomena in fully quantum-mechanical light-matter systems from the view of differential geometry. The predicted geometric features can be observed with present techniques in superconducting circuits. VI