Renormalized Hopping Parameters

• Renormalized hopping parameters are emergent, scale-dependent effective amplitudes derived by integrating out high-frequency degrees of freedom in quantum lattices and classical stochastic processes.
• They capture the suppression or enhancement of particle mobility in disordered and interacting systems via techniques such as RG, DMFT, and strong-disorder analyses.
• Their scaling behavior under renormalization links microscopic dynamics to macroscopic transport properties, serving as key indicators of phase transitions.

Renormalized hopping parameters are emergent, scale-dependent effective hopping amplitudes or rates that result from integrating out high-energy (or high-frequency, short-scale, or non-ergodic) degrees of freedom in interacting quantum lattice models and classical stochastic processes. They appear in renormalization group (RG), mean-field, and strong-disorder treatments of hopping and transport phenomena, capturing the suppression or enhancement of particle mobility due to disorder, interactions, statistical constraints, or quantum fluctuations. Contemporary formulations span configuration space RG for many-body localization, real-space strong-disorder RG for hopping in energy-disordered lattices, dynamical mean-field theory (DMFT) for correlated Fermi liquids, Gutzwiller-projected mean-field approaches in strong Mott systems, and hierarchical/lattice RG for models with complex long-range structure.

1. Renormalized Hopping in Configuration Space: Many-Body Localization

Monthus and Garel (Monthus et al., 2010) define the renormalized hopping $V_L$ between two many-body Fock space configurations $\mathcal{C}_A$ and $\mathcal{C}_B$ at distance $L$ (minimal particle moves) using an exact Aoki-type configuration-space RG. Starting from a Hamiltonian

$$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$

the RG sequentially decimates away all configurations except $\mathcal{C}_A$ and $\mathcal{C}_B$, updating couplings via the Schur-complement recursion

$$V_{ij}^{\rm(new)} = V_{ij} + \frac{V_{i0} V_{0j}}{E - V_{00}}$$

for each eliminated configuration $0$ at fixed energy $E$. The final effective off-diagonal matrix element is the renormalized $\mathcal{C}_A$0 between the endpoint configurations.

Statistically, in the localized (many-body insulating) phase ($\mathcal{C}_A$1), the typical value $\mathcal{C}_A$2 decays exponentially:

$\mathcal{C}_A$3

with a localization length diverging at the transition, $\mathcal{C}_A$4 ($\mathcal{C}_A$5). In the delocalized phase ($\mathcal{C}_A$6), $\mathcal{C}_A$7 remains finite as $\mathcal{C}_A$8, displaying an essential singularity

$\mathcal{C}_A$9

with $\mathcal{C}_B$0. The RG-defined $\mathcal{C}_B$1 correlates directly with real-space two-point functions; in the localized phase, both the configuration-space renormalized hopping and mid-spectrum real-space correlators decay with the same length scale.

2. Real-Space Renormalization of Hopping: Strong-Disorder and Hierarchical Lattices

Disordered classical and quantum lattices exhibit a hierarchy of time and energy scales, permitting a real-space RG treatment of hopping rates. In energy-disordered systems (Velizhanin et al., 2013), the Miller–Abrahams hopping rates

$\mathcal{C}_B$2

are renormalized via successive clustering ("trap" merging) based on local landscape minima and escape rates. The resulting RG recursions yield new cluster ("super-site") free energies and effective prefactors:

$\mathcal{C}_B$3

Iterating this, the distribution of rates $\mathcal{C}_B$4 and activation barriers evolves, and after several steps, only the slowest system-wide processes (macroscopic transport) remain, characterized by highly renormalized hopping prefactors and barriers.

For hierarchical models with power-law decaying hopping (1311.0780), the RG yields an exact recursion for the hopping amplitude:

$\mathcal{C}_B$5

leading to $\mathcal{C}_B$6. Disorder induces competing fixed points, with Cauchy or delta-function distributions of effective (renormalized) on-site potentials, depending on the hopping decay exponent.

3. Dynamical Mean-Field and Gutzwiller-Renormalized Hopping: Fermi Liquid and t-J Models

Within dynamical mean-field theory (DMFT) for correlated electrons (Hewson, 2016), the Fermi-liquid self-energy structure produces a low-frequency quasiparticle weight

$\mathcal{C}_B$7

leading to a rigid rescaling of the bare hopping:

$\mathcal{C}_B$8

All kinetic terms, and thus the bandwidth, are reduced by $\mathcal{C}_B$9, which collapses to zero at the Mott transition, driving the system insulating.

In projected t–J models, e.g., for La$L$0Ni$L$1O$L$2 (Tian et al., 2024), Gutzwiller-renormalized mean-field theory (RMFT) multiplies each hopping amplitude $L$3 by a factor $L$4 derived from local densities and order parameters. For uniform hole density $L$5 and zero magnetization, $L$6, vanishing in the Mott limit, approaching one in the noninteracting limit. All kinetic and superexchange terms thus become

$L$7

with local $L$8 and $L$9 factors controlled by the self-consistent mean-field order parameters.

4. RG Flow and Scaling of Hopping in Quantum and Classical Systems

Renormalization of hopping amplitudes/rates encapsulates a physical flow from microscopic to emergent effective models, leading to crossovers, fixed points, and criticality.

In classical and quantum RG, the running hopping amplitude $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$0 typically obeys

$$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$1

as in bosonized 1D superconductors with power-law hopping (Lobos et al., 2012), where $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$2. When $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$3, the hopping is relevant: RG drives it to strong coupling, quenching quantum fluctuations and restoring phase coherence. In hierarchical RG for the Anderson model, the flow equations similarly determine whether hopping decays to zero (localized, strong disorder) or remains robust (delocalized, weak disorder) (1311.0780).

In generalized Euclidean random matrix models with spatially decaying hopping (Kutlin et al., 2020), the RG produces an energy-dependent suppression of hopping tails:

$$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$4

with localization when the effective exponent exceeds the dimension, i.e., $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$5.

5. Statistical Interpretation, Phase Structure, and Macroscopic Transport

The statistical properties of renormalized hopping parameters provide direct signatures of phases and phase transitions:

• In many-body localization, the exponential decay of $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$6 in configuration space quantifies ergodicity breaking; the divergence of $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$7 signals the transition (Monthus et al., 2010).
• In classical hopping with energy disorder, the renormalized distribution of prefactors and barriers yields the long-time macroscopic diffusion coefficient

$$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$8

where $$H_{\rm many} = \sum_{\mathcal{C},\mathcal{C}'} V_{\mathcal{C},\mathcal{C}'} |\mathcal{C}\rangle\langle \mathcal{C}'|$$9 and $\mathcal{C}_A$0 are the terminal RG-renormalized parameters (Velizhanin et al., 2013).

• In quantum Fermi liquids, the vanishing of $\mathcal{C}_A$1 and of $\mathcal{C}_A$2 tracks the Mott transition, with universal ratios of renormalized interactions to bandwidth (Hewson, 2016).

Tables below summarize the scaling of renormalized hopping in various contexts:

Model	RG-renormalized hopping	Scaling/Behavior
MBL, config. RG (Monthus et al., 2010)	$\mathcal{C}_A$3	Exponential decay/infinite random variable, phases split
t–J model, Gutzwiller RMFT (Tian et al., 2024)	$\mathcal{C}_A$4	$\mathcal{C}_A$5 (homogeneous, PM)
Hubbard, DMFT (Hewson, 2016)	$\mathcal{C}_A$6	$\mathcal{C}_A$7 at Mott, $\mathcal{C}_A$8 at weak interaction
Energy-disordered hop (Velizhanin et al., 2013)	$\mathcal{C}_A$9	Strongly suppressed at low $\mathcal{C}_B$0
Hierarchical Anderson (1311.0780)	$\mathcal{C}_B$1	$\mathcal{C}_B$2

6. Applications and Relevance to Measurement and Inference

Renormalized hopping parameters are central for connecting microscopic models to experimentally accessible bulk quantities:

• The RG-derived $\mathcal{C}_B$3 and its scaling dictate the many-body mobility edge and transport thresholds in interacting disordered quantum wires (Monthus et al., 2010).
• In stochastic models, renormalized (coarse-grained) hopping rates are extracted from noisy and blurred trajectory data by moment or covariance-based estimators, facilitating unbiased measurement of underlying transport coefficients and their effective anisotropy (Mishima, 2020).
• In strongly correlated electronic systems, the renormalized hopping computed by DMFT, RMFT, or RPT sets the bandwidth and effective mass governing observables such as resistivity, charge/spin susceptibility, and pairing properties (Hewson, 2016, Tian et al., 2024, Pandis, 2014).

In each of these applications, neglect of renormalization may lead to misinterpretation of physical responses, incorrect effective-medium theories, or failure to predict emergent localization/transport phenomena.

7. Perspectives and Research Frontiers

Ongoing work targets characterizing nonperturbative regimes (absence of small parameter), universal scaling of renormalized hopping across model classes, and the role of rare-region and non-ergodic effects. The interplay between real-space and configuration space RG, emergence of nontrivial fixed points (Cauchy/delta), and the explicit construction of coarse-grained transport models from raw trajectory data represent important methodological advances (Monthus et al., 2010, Kutlin et al., 2020, Mishima, 2020). Application to multi-band, multi-channel, and non-equilibrium systems, as well as automated extraction of renormalized couplings in experimental time-resolved data, defines a key direction for future inquiry.