Many-body quantum geometric effects and entanglement at the 3D metal-insulator quantum phase transition

Technical Summary: Many-body quantum geometric effects and entanglement at the 3D metal-insulator quantum phase transition

Problem Statement
Quantum geometry is a unifying concept in condensed matter physics, typically formulated within band theory to describe phenomena ranging from nonlinear topological responses to flat-band superconductivity. However, its applicability to disordered, interacting systems remains less explored. A central challenge is identifying experimental probes that can characterize the ground-state quantum geometry and multipartite entanglement across a three-dimensional metal-insulator quantum phase transition (MIQPT). Specifically, there is a need to distinguish between the correlation length ($\xi$), which diverges at the critical point, and the local quantum geometric length ($\ell$), which characterizes the spatial spread of local wavefunctions. In phosphorus-doped silicon (Si:P), a model system for the MIQPT, the behavior of these length scales and their relationship to the diverging dielectric constant and the century-old Herzfeld metallization criterion requires clarification.

Methodology
The study utilizes phosphorus-doped silicon (Si:P) samples with doping concentrations ranging from 32% to 103% of the critical concentration ($x_c \approx 3.5 \times 10^{18} \text{ cm}^{-3}$). The experimental approach combines:

1. Time-domain THz spectroscopy (TDTS): Measuring complex transmission in the 0.2–2 THz range to directly obtain complex conductivity.
2. Fourier transform infrared (FTIR) reflectivity: Measuring reflectivity in the far-infrared (3–19.5 THz) and mid-infrared (21–210 THz) ranges.
3. Global Kramers-Kronig consistent fitting: A variational fitting procedure combining TDTS and FTIR data using a dense ensemble of Lorentz oscillators to reconstruct the full frequency-dependent optical conductivity, $\sigma(\omega)$.

The core theoretical tool is the first negative moment sum rule of the optical conductivity ($S^{-1}_{-1}$), defined as:
$$S^{-1}_{-1} = \frac{2}{\pi} \int_0^\infty \frac{\text{Re}\,\sigma(\omega)}{\omega} d\omega = \frac{2e^2 n_e}{\hbar} \langle \ell_a \ell_b \rangle = \frac{F_Q(T \to 0)}{2}$$
where $\langle \ell_a \ell_b \rangle$ is the localization tensor (related to the second cumulant of the electron distribution) and $F_Q$ is the quantum Fisher information (QFI), which bounds multipartite entanglement. The authors extract the quantum geometric length $\ell$ from this integral and compare it against the correlation length $\xi$ derived from the dielectric constant scaling.

Key Results

1. Distinct Length Scales: The study establishes that the quantum geometric length $\ell$ and the correlation length $\xi$ are distinct scales near the MIQPT. While $\xi$ diverges continuously as the system approaches the critical concentration (consistent with the divergence of the dielectric constant), $\ell$ remains finite and does not diverge continuously.
2. Behavior of $\ell$: Far from the transition, $\ell$ coincides with the effective Bohr radius of the phosphorus donors ($\sim 1.3$–$1.8$ nm). As the doping concentration increases toward $x_c$, $\ell$ enhances by a factor of approximately 2.4 but stays finite. At the critical point, $\ell$ jumps discontinuously to infinity, rather than diverging continuously.
3. UV Dominance: The lack of continuous divergence in $\ell$ is attributed to the ultraviolet (UV) domination of the sum rule in three dimensions. The integral is dominated by high-frequency optical weight, rendering it insensitive to the critical softening and low-frequency fluctuations that drive the divergence of the dielectric constant and $\xi$.
4. Quantum Geometric Correction to Clausius-Mossotti: The enhancement of $\ell$ is interpreted as a "puffing" of the donor polarizability volume due to intersite dipole fluctuations. By scaling the single-dopant polarizability $\alpha_D$ by the ratio $[\ell(x)/\ell(0)]^3$, the authors derive a "quantum geometry corrected" Clausius-Mossotti relation. This corrected model significantly improves the prediction of the critical concentration $x_c$, reducing the error from an order of magnitude (in the standard model) to approximately 20%.
5. Entanglement and QFI: The QFI, bounded by $S^{-1}_{-1}$, is shown to contain a separable single-particle contribution and a non-separable inter-site component. The latter reflects cooperative enhancement and entanglement between donors, providing a quantum mechanical foundation for the Herzfeld metallization criterion.

Significance and Claims
The paper claims to provide an experimental probe of many-body quantum geometry in disordered interacting systems via the first negative moment of optical conductivity. The primary significance lies in demonstrating that:

• The quantum geometric length $\ell$ characterizes the local wavefunction size and polarization fluctuations but is not the critical length scale of the 3D MIQPT.
• The divergence of the dielectric constant in 3D systems is driven by critical fluctuations that do not affect the UV-dominated sum rule governing $\ell$.
• Incorporating quantum geometric corrections (via the scaling of $\ell$) into mean-field descriptions like the Clausius-Mossotti relation yields a more accurate description of the dielectric response and the metallization threshold.
• The work unifies concepts from quantum geometry, many-body localization, and quantum information, extending geometric approaches from translationally invariant models to complex, correlated materials.

The authors conclude that while $\ell$ is not the critical length of the MIQPT, its enhancement reflects a quantum mechanical "puffing" of donor orbitals, offering a robust route to characterizing ground-state properties in interacting disordered systems.