ωFQ: Frequency-Dependent Fluctuating Charges

• ωFQ is a model that extends classical fluctuating-charge approaches to include frequency-dependent responses, providing clear insights into nanoscale optical behavior.
• It employs coupled linear systems with Drude-like equations and polarizable force fields to accurately capture both intraband and interband charge dynamics under oscillatory fields.
• The approach enables efficient simulations of plasmonic resonances, dielectric properties, and charge noise in nanostructures with significantly lower computational cost than full TDDFT.

Frequency-dependent fluctuating charges (ωFQ) are a class of classical and quantum-inspired models that generalize the longstanding fluctuating-charge (FQ) or electronegativity-equalization paradigm to explicitly incorporate the dynamical (frequency-dependent) response of atomic or collective charge distributions to external oscillatory fields. These models have emerged as a central tool in describing the optical, dielectric, and charge noise properties of nanostructures, plasmonic particles, molecular systems, and superconducting circuits, with extensions now well established for both atomistic simulations and mesoscopic fluctuation analyses. ωFQ theory allows the efficient and self-consistent calculation of response properties at finite frequencies, amenable to direct comparison with time-dependent quantum approaches and experimental spectroscopies.

1. Theoretical Foundations of ωFQ

Classical FQ models are based on the minimization of a quadratic energy functional—usually parameterized by atomic electronegativities and hardnesses—that yields static partial charges distributed to achieve chemical potential equalization across a molecule or material. In ωFQ, this static paradigm is promoted to the frequency domain: each atomic or site-centered charge $q_i$ becomes a complex, frequency-dependent variable $q_i(\omega)$, whose dynamics are governed by explicit continuity or Drude-like equations.

These models connect to the theory of time-dependent Kohn-Sham DFT and frequency-dependent polarizable force fields. In the frequency domain, the key governing relations for the fluctuations $\mathbf{q}(\omega)$ take the form of a coupled linear system with mass-like (dispersive) terms, allowing for direct calculation of both intraband (free-electron, Drude) and interband (bound-electron, polarizability) contributions (Nicoli et al., 2024, Cheng et al., 2022).

The general ωFQ model, in its atomistic plasmonic context, requires solving a linear system for the set of charges $q_i(\omega)$,

$$-\,i\,\omega\,q_i(\omega) = \sum_{j}\Biggl[ \frac{A_j\,n_j\,[1-f_{ji}(l_{ij})]}{1/\tau_j-i\,\omega} +\frac{A_i\,n_i\,[1-f_{ij}(l_{ij})]}{1/\tau_i-i\,\omega} \Biggr] \frac{\phi_i(\omega)-\phi_j(\omega)}{l_{ij}}$$

where $A_i, n_i, \tau_i$ are atom-specific parameters, $f_{ij}$ encodes quantum-tunneling damping, and $\phi_i(\omega)$ is the total potential at site $i$ (Nicoli et al., 2024). Drude-like denominators, relaxation times, and coupling to dipolar degrees of freedom generalize the energy-minimization procedure, yielding frequency-resolved response functions.

2. Quantum-Inspired Formulations: ACKS2ω and TDDFT Connections

The ωFQ philosophy is extended in the ACKS2ω framework, which derives a block-matrix linear system from the partitioning of the frequency-dependent response function at the DFT level (Cheng et al., 2022). Here, variations of the density are expanded in atom- and multipole-centered bases, leading to a set of coupled equations for frequencies $\omega$: $q_i(\omega)$0 where $q_i(\omega)$1 is the static hardness matrix, $q_i(\omega)$2 is the non-interacting Kohn-Sham response, $q_i(\omega)$3 is the overlap, and $q_i(\omega)$4 enforces charge conservation. The induced atomic monopoles and dipoles $q_i(\omega)$5 are then obtained self-consistently at each frequency (Cheng et al., 2022).

This construction preserves a direct link to full TDDFT linear response while reducing computational cost by orders of magnitude. Accurate absorption spectra and $q_i(\omega)$6 dispersion coefficients are achieved for molecular sets by including only atomic monopole and dipole response, with parameterization directly from a ground-state KS-DFT calculation.

3. Atomistic Implementation and Extensions

The ωFQ methodology has been extensively developed for plasmonic nanoparticles, metallic clusters, and hybrid nanostructures. In these implementations, ωFQ is usually coupled to fluctuating-dipole degrees of freedom (ωFQFμ) to allow both intraband (Drude) and interband (polarization) contributions, capturing the full metallic and dielectric response of the system. The general coupled system,

$q_i(\omega)$7

is solved at each $q_i(\omega)$8, with parameters ($q_i(\omega)$9) fit to reproduce bulk experimental properties and tunneling damping functions enforcing physical locality. Quantum effects such as tunneling are treated by Fermi-like damping functions, and solvation effects are incorporated by coupling to a polarizable classical FQ/Fμ environment (Nicoli et al., 2024).

This approach is fully atomistic, handles systems of thousands of atoms (scaling $\mathbf{q}(\omega)$0), and provides detailed access to structure–property relationships for plasmonic resonances, refractive-index sensitivities, and nanoscale charge dynamics.

4. Fluctuation-Dissipation Approaches to ωFQ in Charge Noise and Impedance

The equilibrium charge fluctuation and noise characteristics at finite frequency are directly governed by the fluctuation–dissipation theorem (FDT). ωFQ models provide practical means to compute and analyze frequency-dependent spectral densities relevant for electronic decoherence, charge noise in superconducting devices, and electrochemical impedance spectroscopy.

For an ensemble of two-level fluctuators (TLFs), as relevant in charge noise for qubits, the naive Markovian Bloch–Redfield approach violates the FDT, failing to capture frequency broadening. The introduction of a spectator-qubit probe and a consistent master equation restores FDT compliance, yielding the ωFQ-consistent spectral density: $\mathbf{q}(\omega)$1 with crossover behavior $\mathbf{q}(\omega)$2 at low $\mathbf{q}(\omega)$3, transitioning to $\mathbf{q}(\omega)$4 above a frequency $\mathbf{q}(\omega)$5, and strict FDT-satisfying relations $\mathbf{q}(\omega)$6 (You et al., 2020).

In nanocapacitor and electrochemical systems, ωFQ principles underpin the computation of dynamic capacitance and impedance from equilibrium charge fluctuations. The complex admittance $\mathbf{q}(\omega)$7 is related to autocorrelations of electrode charge fluctuations $\mathbf{q}(\omega)$8 via the Green–Kubo formula: $\mathbf{q}(\omega)$9 with the frequency-dependent capacitance given by $q_i(\omega)$0 and the impedance $q_i(\omega)$1, all inheriting fluctuation-based physical meaning (2206.13322).

5. Computational Workflows and Validation

Table 1: ωFQ Atomistic Implementation Workflow (Nicoli et al., 2024, Cheng et al., 2022)

Step	Molecular Plasmonics (ωFQFμ)	Polarizable FF (ACKS2ω)
1	Build nanoparticle geometry (ASE)	Geometry optimization (B3LYP)
2	MD sampling with FQ-polarizable solvent	Single-point KS-DFT
3	Extract solvated snapshots	Build basis functions
4	Solve ωFQFμ/FQ equations per snapshot	Assemble ω-dependent matrices
5	Average absorption spectra	Solve linear system at each ω

Validation is performed against TD-DFTB/FQ, Casida TDDFT, and experimental data. For plasmonic nanoparticles, ωFQFμ/FQ accurately reproduces quantum calculations and experimental resonance shifts, capturing subnanometer effects and explicit solvation-induced shifts (Nicoli et al., 2024). ACKS2ω achieves mean absolute percent errors (MAPEs) of ∼2–4% for dispersion coefficients across ∼900 molecular pairs, matching or exceeding TDDFT performance (Cheng et al., 2022).

6. Applications and Limitations

ωFQ models are widely applied to optical response (absorption spectra, plasmonic resonances, solvatochromism), charge and current noise in quantum information devices, dynamic impedance in electrochemical systems, and the calculation of van der Waals dispersion coefficients. Their main advantages include:

• Fully atomistic treatment, capturing shape, size, composition, and explicit solvent structure.
• Treatment of both intraband and interband response channels.
• Orders-of-magnitude lower computational cost compared to TDDFT for large systems.
• Direct satisfaction of fundamental physical relations (FDT).

Limitations include:

• All interactions are classical: genuine exchange/Pauli effects are missing.
• Approximations in parameterization (e.g., FQ solvent models based on small molecules may underestimate shifts).
• No explicit charge-transfer excitations or nonlinear response included.
• Master-equation approaches require weak coupling and Gaussian baths (Nicoli et al., 2024, You et al., 2020).

7. Outlook and Future Directions

Future work aims to extend ωFQ models with empirical or machine-learned hardness and overlap kernels to further improve scaling ($q_i(\omega)$2 or better) and enable application to bulk materials and extended systems (Cheng et al., 2022). There is ongoing research in coupling ωFQ methodologies with explicit treatment of ionic conduction, redox chemistry, and quantum nuclear response for increased spectroscopic fidelity in complex electrolytes (2206.13322). Extensions toward nonlinear optical effects, charge-transfer excitation, and dissipative solvent polarization are under exploration to broaden the applicability spectrum of frequency-dependent fluctuating charge models.