Mott–Hubbard Metal–Insulator Transition

• The Mott–Hubbard metal–insulator transition is a quantum phase change where strong local Coulomb repulsion shifts materials from a metallic, itinerant state to an insulating, localized state.
• It is characterized by the suppression of quasiparticle coherence and a coexistence region with distinct critical endpoints, often analyzed using DMFT and its cluster extensions.
• Real-world examples in compounds like VO₂ and V₂O₃ highlight the transition's role in advancing our understanding of strongly correlated materials and potential applications in Mottronics.

The Mott–Hubbard metal–insulator transition (MIT) is a paradigmatic quantum phase transition in which electronic correlations—predominantly local Coulomb repulsion—drive a system from a metallic, itinerant state to an insulating, localized state, even without a corresponding breaking of crystal or magnetic symmetry. The transition, first formulated in the context of the Hubbard model, is a hallmark of strongly correlated electron systems, including classic compounds such as V₂O₃ and VO₂, and is foundational for understanding complex oxides, organic materials, and various engineered cold atom systems. The nature, universality, and microscopic mechanisms underlying the Mott–Hubbard MIT have been a focus of intense theoretical and experimental study; recent advances have further elucidated the transition's entropic, topological, and non-equilibrium properties, as well as its role in mottronics and quantum materials.

1. Models and Formalism

The canonical framework for the Mott–Hubbard transition is the Hubbard model,

$$H = -t \sum_{\langle ij \rangle,\sigma} (c^\dagger_{i\sigma} c_{j\sigma} + \text{h.c.}) + U \sum_{i} n_{i\uparrow} n_{i\downarrow},$$

where $t$ is the nearest-neighbor hopping (kinetic energy), $U$ is the on-site Coulomb repulsion, $c^\dagger_{i\sigma}$ ($c_{i\sigma}$) creates (annihilates) an electron of spin $\sigma$ at site $i$, and $n_{i\sigma}$ is the occupation operator. Extensions include multi-orbital models, explicit inclusion of Hund's coupling $J$, spin-orbit interaction $\zeta$, spatial disorder, coupling to phonons (Hubbard–Holstein), and spatially modulated or multi-layer hybridizations.

In multi-orbital systems and real materials such as VO₂, the minimal two-orbital Hubbard Hamiltonian with intra-dimer hopping ($t$0), intra- and inter-orbital interactions ($t$1, $t$2), and Hund’s coupling ($t$3), is essential for capturing the competition between Mott and dimerization-driven localization, as well as the role of electronic correlations in driving the MIT (Nájera et al., 2016).

Dynamical mean-field theory (DMFT) has become the central theoretical tool for studying the Mott–Hubbard MIT. DMFT maps the lattice many-body problem onto a self-consistent quantum impurity model, capturing local dynamical correlations exactly while neglecting spatial fluctuations. Extensions to cluster DMFT (CDMFT), statistical DMFT (statDMFT), and plaquette approaches allow incorporation of short-range and spatially inhomogeneous correlations critical for describing transitions in lower dimensions, disorder, or non-uniform systems (Yanagi et al., 2014, Walsh et al., 2018, Suárez-Villagrán et al., 2019).

2. Mechanism and Critical Behavior of the Transition

The key physical mechanism of the Mott–Hubbard MIT is the correlation-induced suppression of quasiparticle coherence, expressed formally via the quasiparticle residue: $t$4 where $t$5 is the self-energy. As $t$6 increases at fixed $t$7, $t$8 decreases and vanishes discontinuously at a critical value $t$9 (in DMFT), marking the metal–insulator boundary. The system displays a coexistence region, bounded below by $U$0, where both metallic and insulating DMFT solutions exist and the transition is strongly first order at low temperatures (Nájera et al., 2016, Niu et al., 2021).

The (U,T) phase diagram exhibits a characteristic "Mott dome": at $U$1 (critical endpoint), the MIT is first order with hysteresis; at $U$2 the two spinodals merge and the transition endpoint displays critical scaling akin to the Ising universality class. Above $U$3 the system undergoes a sharp but continuous crossover along the Widom line, identified by peaks in thermodynamic susceptibilities (compressibility, entropy, information-theoretic measures) (Walsh et al., 2018).

Cluster and multi-orbital extensions break the universality; for example, in the modified periodic Anderson model, the T=0 transition is strictly continuous (second order), and non-local correlations modestly shift but do not fundamentally alter the critical behavior (Majumder et al., 19 Jun 2025). In cases with strong spatial inhomogeneity, e.g., layered interactions or disorder, the transition is percolative, and the first-order character is smeared in the thermodynamic limit (Suárez-Villagrán et al., 2019).

3. Realizations and Experimental Signatures

The Mott–Hubbard transition is realized in a diverse set of correlated electron materials:

• VO₂ and V₂O₃: In VO₂, a two-orbital Hubbard model with intra-dimer hopping, solved by DMFT, reproduces the first-order transition, mid-infrared optical conductivity peaks, and phase coexistence observed experimentally. The splitting of heavy quasiparticle bands in the metallic region gives rise to characteristic spectroscopic signatures, and the transition is shown to be dominated by electronic correlations (Mott physics), not only by dimer (Peierls) formation (Nájera et al., 2016). In V₂O₃, LDA+DMFT captures the critical U/W for the transition and its sensitivity to Cr-doping and applied pressure, revealing microscale phase separation near the MIT (Hansmann et al., 2013).
• Three-band (Emery) model and Charge-Transfer Systems: In cuprates, the three-band Hamiltonian shows a quantum phase transition from an antiferromagnetic insulator to a paramagnetic metal as the charge-transfer energy $U$4 is reduced below a critical value ($U$5 eV). The nature of the gap (Mott-Hubbard vs. charge transfer) depends on the interplay of $U$6, $U$7, and orbital occupancy (Vitali et al., 2018).
• Layered and Quasi-One-Dimensional Systems: Anisotropic models with weak interchain coupling or spatially modulated interactions display two-stage transitions: a deconfinement (1D insulator to 2D metal) and a subsequent 2D Mott-Hubbard transition at finite $U$8, in sharp contrast to predictions from weak-coupling Hartree–Fock, which always opens a gap at infinitesimal $U$9 (Moukouri et al., 2011, Mondaini et al., 2013).
• Disordered and Multicomponent Systems: In the presence of disorder, statDMFT shows that the local first-order Mott physics is preserved on finite clusters, but becomes a highly inhomogeneous, percolation-like transition in the thermodynamic limit. Metallic puddles nucleate within an insulating matrix, in natural accord with nano-imaging experiments (Suárez-Villagrán et al., 2019). In mass-imbalanced systems, the MIT remains first-order for any finite hopping of the heavier species, and both components share a single Kondo temperature (Philipp et al., 2016).

The essential experimental signatures include abrupt changes in resistivity, disappearance of Drude peaks, development of Hubbard bands in photoemission, phase coexistence revealed by nano-imaging, and latent heat or entropy jumps at the transition (Nájera et al., 2016, Hansmann et al., 2013, Feng et al., 2010).

4. Variants, Extensions, and Topological Interpretation

The universality of the Mott–Hubbard transition is enriched by competing physics:

• Spin-Orbit and Hund’s Coupling: In 5d transition metal oxides, incorporating SOC and Hund’s interactions yields phase diagrams with metallic, band-insulating, and Mott-insulating regions. Spin-orbit coupling reduces degeneracy and bandwidth, enhancing the insulating tendency, and leads to nontrivial interaction-renormalized SOC (Du et al., 2012).
• Charge Transfer vs. Mott Physics (ZSA Diagram): The Zaanen-Sawatzky-Allen classification emphasizes that the nature of the gap switches from $c^\dagger_{i\sigma}$0-dominated (Mott-Hubbard insulator, $c^\dagger_{i\sigma}$1) to $c^\dagger_{i\sigma}$2-dominated (charge-transfer insulator, $c^\dagger_{i\sigma}$3), and negative charge transfer ($c^\dagger_{i\sigma}$4) can drive metallicity even with large $c^\dagger_{i\sigma}$5 (Jana et al., 2021, Vitali et al., 2018).
• Topological Recasting: The Mott transition in the infinite-dimensional Hubbard model can be exactly mapped to a topological transition in an auxiliary Su-Schrieffer-Heeger (SSH) chain; the insulating phase corresponds to the topological SSH phase with a robust zero mode in the self-energy, associated with the opening of the Mott gap (Sen et al., 2020).

5. Thermodynamic, Entropic, and Dynamical Aspects

Quantitative analysis of the transition includes calculation of:

• Double Occupancy and Compressibility: Discontinuities in double occupancy, entropy, and charge compressibility are hallmark signatures of the first-order transition and the critical endpoint (Walsh et al., 2018, Niu et al., 2021).
• Entropy and Information Measures: The single-site von Neumann entropy and mutual information serve as precise identifiers of the transition, with entropy dropping discontinuously in the insulator, and mutual information peaking at the critical point and along the Widom line—a demonstration that the Mott–Hubbard transition is also detectable via quantum information-theoretic probes (Majumder et al., 19 Jun 2025, Walsh et al., 2018).
• Kinetics and Nucleation: Non-equilibrium dynamics of the transition show that the transformation proceeds by nucleation and growth of Mott droplets with classical Kolmogorov-Johnson-Mehl-Avrami kinetics at early times, followed by avalanche-like behavior in the intermediate regime, even without quenched disorder. Structural two-point correlations display universal scaling consistent with sharp interface growth and percolation physics (Chern, 2019).

6. Coexistence, Inhomogeneity, and Control

The phase coexistence and intrinsic inhomogeneity at the MIT are central to the physical picture:

• The coexistence region in the (U, T) plane, bracketed by spinodals $c^\dagger_{i\sigma}$6 and $c^\dagger_{i\sigma}$7, is characterized by mixed metallic and insulating domains, observable in nano-imaging and reproduced theoretically via effective medium theory and statDMFT (Hansmann et al., 2013, Suárez-Villagrán et al., 2019).
• In lower dimensions and/or with disorder, Imry–Ma arguments and numerical studies establish that the first-order character is rounded and replaced by a smooth, percolative crossover, in contrast to mean-field or high-dimensional cases (Suárez-Villagrán et al., 2019).
• In engineered or cold-atom systems with spatially patterned interactions, exotic transitions, including reentrant anisotropic metallic states at large U, are possible by decoupling non-interacting pathways from correlated regions (Mondaini et al., 2013).

The electronic, rather than structural, nature of the Mott–Hubbard MIT enables ultrafast and non-volatile switching in candidate Mottronics devices, with thresholds and timescales that can be quantitatively predicted by DMFT-based models and controlled via external fields, gating, or strain (Nájera et al., 2016).

7. Broader Significance and Outlook

The Mott–Hubbard MIT forms the organizing principle for the phase behavior of a vast range of strongly correlated materials. Its theoretical description—now unified across DMFT, information theory, and topological language—provides tools for predicting, interpreting, and controlling metal–insulator transitions in compounds from vanadates and nickelates to organics, artificial lattices, and strongly interacting Fermi gases.

Current frontiers include elucidating universal scaling in low dimensions, uncovering the role of nonlocal and longer-range correlations, exploring kinetics far from equilibrium, and realizing control and device applications (“Mottronics”) by leveraging the electronic origin of the transition. The field continues to provide insights into quantum criticality, entanglement, and the interplay of charge, spin, orbital, lattice, and topological degrees of freedom in correlated systems (Nájera et al., 2016, Sen et al., 2020, Majumder et al., 19 Jun 2025).