Molecular Electron Transfer in Optical Cavities: From Excitonic to Vibronic Polaritons Takumi Hidaka, Tomohiro Fukushima, Nguyen Thanh Phuc Abstract Strong coupling between molecular excitations and quantized electromagnetic fields in optical cavities provides a powerful means to control the physical and chemical properties of molecular systems. Here, we study electron transfer (ET) dynamics in cavity-coupled molecules using the numerically exact hierarchical equations of motion (HEOM) method, which captures nonperturbative and non-Markovian effects beyond standard perturbative theories. We identify distinct resonance and collective effects associated with polariton formation and show that the ET rate saturates in the strong-coupling regime, a feature not captured by perturbative approaches. We further extend the cavity-modified ET model by incorporating the nuclear-coordinate dependence of molecular electric dipole moments, which gives rise to a three-body interaction involving molecular electronic and vibrational degrees of freedom and cavity photons. This vibronic polariton formation leads to non-monotonic, oscillatory dependencies of the ET rate on the light-matter coupling strength and cavity frequency, which we attribute to quantum interference among multiple transfer pathways. These findings establish cavity-modified electron transfer as a multichannel quantum process governed by the interplay of electronic, vibrational, and photonic degrees of freedom. I Introduction Recent advances in cavity quantum electrodynamics have established that confining molecules within tailored electromagnetic environments can strongly couple their internal excitations to quantized light fields, opening new routes to control material properties and chemical reactivity Vidal21 . In this strong-coupling regime, the hybridization of optical cavity modes with molecular excitations gives rise to light-matter eigenstates known as molecular polaritons Ebbesen16 ; Nagarajan21 . Depending on whether the cavity couples predominantly to electronic or vibrational molecular transitions, one distinguishes exciton polaritons and vibrational polaritons, respectively. The latter, formed under vibrational strong coupling, has been reported to influence ground-state chemical reactivity Thomas16 ; Lather19 ; Thomas19 ; Ahn23 and has been linked to enhanced intermolecular vibrational energy transfer Xiang20 and accelerated intramolecular vibrational relaxation Chen22 . Electronic strong coupling and the resulting exciton polaritons have been widely explored as platforms to reshape exciton migration, nonadiabatic dynamics, photochemical transformations Bhuyan23 ; Pandya21 ; Pandya22 ; Cheng23 ; Xu23 ; Balasubrahmaniyam23 ; Zeng23 ; Peng23 ; Lee24 ; Amin24 , and spin transport Phuc23 . Among the processes amenable to cavity control, condensed-phase electron transfer (ET) is of particular importance, as it underpins charge separation and transport in organic electronics, redox chemistry, and numerous biological functions MayKuhnTextbook ; NitzanTextbook . In condensed-phase ET, the transfer rate is governed by the interplay among three key elements: (i) the electronic coupling between donor and acceptor states; (ii) the energetic driving force and activation barrier, shaped by solvent and intramolecular reorganization; and (iii) the thermal fluctuations and dissipation induced by the molecular environment. Optical cavities can, in principle, modify each of these elements. Hybridizing electronic states with a cavity mode creates polaritonic manifolds that reshape energy gaps and effective couplings. Additional radiative and dissipative channels introduced by cavity leakage alter the competition between coherent and incoherent dynamics. Furthermore, collective light-matter coupling in molecular ensembles can renormalize effective interactions and change how disorder and fluctuations impact reaction kinetics. Together, these mechanisms make cavity-modified ET a rich arena for controlling charge transport and chemical reactivity in nanophotonic environments Hayashi24 . Despite growing theoretical interest in cavity-modified condensed-phase ET Semenov19 ; Phuc20 ; Mandal20 ; Chowdhury22 ; Wei22 ; Saller22 ; Saller23 ; SallerJCP23 ; Koessler25 , most existing studies have relied on perturbative rate expressions, typically within the Fermi’s golden rule (FGR) framework, which are rigorously justified only in the linear response regime. However, experimentally relevant cavity QED settings can involve strong light-matter coupling and structured vibrational environments, where nonperturbative and non-Markovian effects may be essential for accurate quantitative predictions and mechanistic understanding. In this work, we employ the hierarchical equations of motion (HEOM) Tanimura20 to simulate the quantum dynamics of cavity-modified ET beyond the perturbative regime. Effective transfer rates are extracted by exponential fitting of the time-evolving donor and acceptor populations computed from HEOM. We organize our investigation as follows. We begin with the minimal model of cavity-modified ET, in which the nuclear-coordinate dependence of the molecular electric dipole moments is neglected, so that the cavity couples exclusively to the electronic degrees of freedom. Within this model, we characterize ET rate modification as a function of the driving force and reorganization energy for two distinct coupling channels: direct transition coupling and energy fluctuation coupling. We examine the effect of strong light-matter coupling on coherence dynamics by comparing the system’s time evolution with and without cavity coupling. Resonance and collective effects arising from polariton formation are investigated by varying the cavity frequency and the number of molecules coupled to a single cavity mode, respectively. The role of cavity loss is assessed by augmenting the HEOM with a Lindblad-form dissipation term. We then generalize the model to include the nuclear-coordinate-dependent molecular dipole moments, which introduces a three-body interaction coupling the molecular electronic and vibrational degrees of freedom to the cavity photons. The influence of the vibronic polariton formation on ET dynamics is investigated for environments characterized by both Drude-Lorentz and underdamped spectral densities, revealing a rich interplay of quantum interference and nonlinear coupling effects. II System Model We consider N_{\text{mol}} identical molecules in the condensed phase strongly coupled to a single mode of an optical cavity. In the dipole gauge, the light-matter interaction is described by the Hamiltonian Semenov19 : \displaystyle\hat{H}= \displaystyle\sum_{n=1}^{N_{\text{mol}}}\left(\hat{H}_{\text{M}}^{(n)}+\hat{H}_{\text{B}}^{(n)}+\hat{H}_{\text{MB}}^{(n)}\right)+\hbar\omega_{\text{c}}\hat{a}^{\dagger}\hat{a} \displaystyle-\sum_{n=1}^{N_{\text{mol}}}\hat{\boldsymbol{\mu}}^{(n)}\cdot\hat{\mathbf{E}}+\frac{1}{2\mathcal{V}\epsilon_{0}}\left(\sum_{n=1}^{N_{\text{mol}}}\hat{\boldsymbol{\mu}}^{(n)}\cdot\boldsymbol{\epsilon}\right)^{2}, (1) where \hat{H}_{\text{M}}^{(n)} , \hat{H}_{\text{B}}^{(n)} , and \hat{H}_{\text{MB}}^{(n)} represent the molecular, bath (environment), and molecule-bath interaction Hamiltonians, respectively, of the n -th molecule. The cavity mode, characterized by frequency \omega_{\text{c}} , polarization unit vector \boldsymbol{\epsilon} , and effective mode volume \mathcal{V} , is described by photonic creation and annihilation operators \hat{a}^{\dagger} and \hat{a} . Light-matter interaction enters through the coupling between the total molecular electric dipole moment \hat{\boldsymbol{\mu}}^{(n)} and the cavity electric field operator \displaystyle\hat{\mathbf{E}}=i\boldsymbol{\epsilon}\sqrt{\frac{\hbar\omega_{\text{c}}}{2\mathcal{V}\epsilon_{0}}}\left(\hat{a}-\hat{a}^{\dagger}\right), (2) where \epsilon_{0} is the vacuum permittivity. The final term in Eq. ( 1 ) is the dipole self-energy, which ensures gauge invariance and becomes significant in the ultrastrong coupling regime where the light-matter coupling strength is comparable to the characteristic molecular energy scales and the cavity frequency FriskKockum19 . This dipole gauge Hamiltonian is obtained via the Power-Zienau-Woolley unitary transformation from the minimal coupling Hamiltonian in the Coulomb gauge TannoudjiTextbook . We first consider a single molecule ( N_{\text{mol}}=1 ) strongly coupled to a single cavity mode. To model electron transfer, we project the molecular Hamiltonian onto the two relevant electronic states: the donor state |\text{D}\rangle and acceptor state |\text{A}\rangle . The environment is represented as a collection of harmonic oscillators with frequencies \nu_{j} and annihilation operators \hat{b}_{j} , yielding the combined molecular and bath Hamiltonian: \displaystyle\hat{H}_{\text{M}}+\hat{H}_{\text{B}}+\hat{H}_{\text{MB}} \displaystyle= \displaystyle E_{\text{D}}|\text{D}\rangle\langle\text{D}|+\left[E_{\text{A}}+\sum_{j}\lambda_{j}(\hat{b}_{j}+\hat{b}_{j}^{\dagger})\right]|\text{A}\rangle\langle\text{A}| \displaystyle+\sum_{j}\hbar\nu_{j}\hat{b}_{j}^{\dagger}\hat{b}_{j}+H_{\text{DA}}|\text{D}\rangle\langle\text{A}|+H_{\text{AD}}|\text{A}\rangle\langle\text{D}|, (3) where E_{\text{D}} and E_{\text{A}} are the donor and acceptor energies, H_{\text{DA}} and H_{\text{AD}} characterize the electronic coupling between states, and \lambda_{j} represents the vibronic coupling strength of the j -th bath mode. Throughout our numerical calculations, we set H_{\text{DA}}=H_{\text{AD}}=30\,\text{cm}^{-1} . Expressing the full Hamiltonian in the donor-acceptor basis in terms of dipole matrix elements yields Semenov19 \displaystyle\hat{H}= \displaystyle E_{\text{D}}|\text{D}\rangle\langle\text{D}|+\left[E_{\text{A}}+\sum_{j}\lambda_{j}(\hat{b}_{j}+\hat{b}_{j}^{\dagger})\right]|\text{A}\rangle\langle\text{A}| \displaystyle+\sum_{j}\hbar\nu_{j}\hat{b}_{j}^{\dagger}\hat{b}_{j}+H_{\text{DA}}|\text{D}\rangle\langle\text{A}|+H_{\text{AD}}|\text{A}\rangle\langle\text{D}|+\hbar\omega_{\text{c}}\hat{a}^{\dagger}\hat{a} \displaystyle+\hbar\omega_{\text{c}}\left(\hat{a}-\hat{a}^{\dagger}\right)\Big(g_{\text{D}}|\text{D}\rangle\langle\text{D}|+g_{\text{A}}|\text{A}\rangle\langle\text{A}| \displaystyle+t_{\text{DA}}|\text{D}\rangle\langle\text{A}|+t_{\text{AD}}|\text{A}\rangle\langle\text{D}|\Big)+\hbar\omega_{\text{c}}\Big[|g_{\text{D}}|^{2}|\text{D}\rangle\langle\text{D}| \displaystyle+|g_{\text{A}}|^{2}|\text{A}\rangle\langle\text{A}|+|t_{DA}|^{2}-\left(g_{\text{D}}+g_{\text{A}}\right)\Big(t_{\text{DA}}|\text{D}\rangle\langle\text{A}| \displaystyle+t_{\text{AD}}|\text{A}\rangle\langle\text{D}|\Big)\Big], (4) where the coupling parameters are defined by projections of the molecular dipole operator onto the donor-acceptor basis: \displaystyle g_{\text{D}}= \displaystyle i\sqrt{\frac{1}{2\hbar\omega_{\text{c}}\mathcal{V}\epsilon_{0}}}\boldsymbol{\mu}_{\text{DD}}\cdot\boldsymbol{\epsilon}, (5) \displaystyle g_{\text{A}}= \displaystyle i\sqrt{\frac{1}{2\hbar\omega_{\text{c}}\mathcal{V}\epsilon_{0}}}\boldsymbol{\mu}_{\text{AA}}\cdot\boldsymbol{\epsilon}, (6) \displaystyle t_{\text{DA}}= \displaystyle i\sqrt{\frac{1}{2\hbar\omega_{\text{c}}\mathcal{V}\epsilon_{0}}}\boldsymbol{\mu}_{\text{DA}}\cdot\boldsymbol{\epsilon} (7) with \boldsymbol{\mu}_{\text{DD}}=\langle\text{D}|\hat{\boldsymbol{\mu}}|\text{D}\rangle , \boldsymbol{\mu}_{\text{AA}}=\langle\text{A}|\hat{\boldsymbol{\mu}}|\text{A}\rangle , and \boldsymbol{\mu}_{\text{DA}}=\langle\text{D}|\hat{\boldsymbol{\mu}}|\text{A}\rangle denoting the diagonal (permanent) and off-diagonal (transition) dipole matrix elements. In the minimal model formulation, we treat the dipole matrix elements as constants, independent of nuclear coordinates. This approximation is valid when the Born-Oppenheimer separation holds and the electronic charge distribution is only weakly modulated by vibrational motion. We will relax this constraint in the generalized model presented later, where nuclear-coordinate dependence of the dipole moments gives rise to a three-body electronic-vibrational-photonic interaction. The final term in Eq. ( 4 ) accounts for dipole self-energy contributions. This formulation reveals two distinct coupling channels through which the cavity mode influences electron transfer. The first channel, characterized by g_{\text{D}} and g_{\text{A}} , corresponds to cavity-induced fluctuations of the donor and acceptor state energies (energy-fluctuation coupling). The second channel, characterized by t_{\text{DA}}=-t_{\text{AD}}^{*} , couples the cavity field directly to the electronic transition between donor and acceptor states (direct transition coupling). The influence of the environment on the system dynamics is encoded in the spectral density J(\omega)=(2\pi/\hbar)\sum_{j}\lambda_{j}^{2}\delta(\omega-\nu_{j}) , with the overall system-environment coupling strength quantified by the reorganization energy \lambda_{0}=\sum_{j}\lambda_{j}^{2}/(\hbar\nu_{j}) . Unless otherwise stated, we employ a Drude-Lorentz spectral density representing a collection of low-frequency intermolecular vibrational modes: \displaystyle J_{\text{DL}}(\omega)=\frac{4\lambda_{0}W\omega}{\omega^{2}+W^{2}}, (8) where the cutoff frequency W=\tau_{\text{b}}^{-1} is inversely related to the bath relaxation time, which we set to \tau_{\text{b}}=100\;\text{fs} . To capture the system’s quantum dynamics, we employ the hierarchical equations of motion (HEOM) Tanimura20 , which provides numerically exact solutions for open quantum systems across arbitrary coupling strengths. The HEOM approach proceeds by first expanding the bath correlation function \displaystyle C(t)=\int_{0}^{\infty}\text{d}\omega J(\omega)\left[\coth\left(\frac{\beta\omega}{2}\right)\cos\omega t-i\sin\omega t\right], (9) into a sum of exponential functions, exploiting the structure of the spectral density J(\omega) and temperature-dependent factors. After formally integrating out the bath degrees of freedom, one obtains a hierarchical set of coupled differential equations governing not only the reduced density matrix of the system but also a hierarchy of auxiliary density operators (ADOs). These ADOs encode the complete memory of past system-bath interactions at different time depths. Each level of the hierarchy captures correlations extending further back in time, with higher-level ADOs representing increasingly long-lived memory effects. The coupled evolution of the system density operator and the ADO hierarchy enables exact treatment of non-Markovian dynamics, capturing memory effects that persist when the bath relaxation timescale is comparable to or longer than the characteristic timescales of the system’s evolution. We solve the HEOM to obtain the time evolution of the donor and acceptor populations, p_{\text{D}}(t) and p_{\text{A}}(t)=1-p_{\text{D}}(t) , with initial conditions corresponding to the molecule in the donor state and the environment in thermal equilibrium: p_{\text{D}}(0)=1 and p_{\text{A}}(0)=0 . Effective transfer rates are extracted by fitting the computed population dynamics to the kinetic rate equations: \displaystyle\frac{\text{d}p_{\text{D}}}{\text{d}t}= \displaystyle-k_{\text{DA}}p_{\text{D}}+k_{\text{AD}}p_{\text{A}}, (10) \displaystyle\frac{\text{d}p_{\text{A}}}{\text{d}t}= \displaystyle k_{\text{DA}}p_{\text{D}}-k_{\text{AD}}p_{\text{A}}, (11) where k_{\text{DA}} and k_{\text{AD}} denote the forward and backward transfer rates. The solution to these equations yields \displaystyle p_{\text{D}}(t)=\frac{k_{\text{AD}}}{k_{\text{AD}}+k_{\text{DA}}}+\frac{k_{\text{DA}}}{k_{\text{AD}}+k_{\text{DA}}}e^{-(k_{\text{AD}}+k_{\text{DA}})t}, (12) from which k_{\text{DA}} is determined via exponential fitting. III Results and Discussion III.1 Minimal Model: Cavity-Electronic Coupling Figure 1 displays the relative change in electron transfer rate, k_{\text{cav}}/k_{\text{non-cav}} , as a function of the reorganization energy \lambda_{0} and the driving force -\Delta G=E_{\text{D}}-E_{\text{A}} . Panel (a) shows results when light-matter interaction occurs exclusively through the direct transition coupling channel ( t_{\text{DA}}\neq 0 , g_{\text{D}}=g_{\text{A}}=0 ). Here we set \hbar\omega_{\text{c}}=100\,\text{cm}^{-1} and |t_{\text{DA}}|=0.5 . The cavity coupling produces substantial rate enhancement across a broad parameter space, with increases exceeding four-fold under optimal conditions. This pronounced effect arises because the cavity field couples directly to the transition dipole moment, effectively opening a new coherent pathway for electron transfer. Panel (b) presents the corresponding results when interaction occurs solely through the energy-fluctuation coupling channel ( t_{\text{DA}}=0 , g_{\text{D,A}}\neq 0 ), with parameters \hbar\omega_{\text{c}}=100\,\text{cm}^{-1} , |g_{\text{D}}|=0.6 , and |g_{\text{A}}|=0.1 . While this channel also yields rate enhancement, the effect is less pronounced than for the t_{\text{DA}} channel. Here, the cavity modulates the energies of the donor and acceptor states rather than directly facilitating the electronic transition, resulting in a qualitatively different enhancement mechanism with generally smaller magnitude. Figure 1: Relative change k_{\text{cav}}/k_{\text{non-cav}} in electron transfer rate as a function of reorganization energy \lambda_{0} and driving force -\Delta G for a single molecule ( N_{\text{mol}}=1 ). (a) Direct transition coupling only ( t_{\text{DA}}\neq 0 , g_{\text{D}}=g_{\text{A}}=0 ). (b) Energy-fluctuation coupling only ( t_{\text{DA}}=0 , g_{\text{D,A}}\neq 0 ). To examine how cavity coupling affects the temporal dynamics of electron transfer, we plot the time evolution of the donor population in Fig. 2 for the direct transition coupling channel with \lambda_{0}=50\,\text{cm}^{-1} and -\Delta G=100\,\text{cm}^{-1} . Comparison between cavity-coupled (red solid) and uncoupled (black dashed) cases reveals two distinct effects of the light-matter interaction. First, the cavity accelerates the transfer process, as evidenced by the faster decay of the donor population toward its equilibrium value. Second, and equally important, the cavity-coupled system exhibits pronounced oscillatory behavior that persists over significantly longer timescales, indicating an extended coherence time compared to the bare molecular system. This prolonged coherence can be understood through the dynamic polaron decoupling effect Phuc21 ; Phuc25 ; Spano15 ; Herrera16 ; Phuc19a ; Takahashi20 . When light-matter coupling is sufficiently strong, the formation of delocalized polariton states effectively reduces the local system-environment interaction. The polariton, being a hybrid light-matter excitation distributed across both photonic and molecular components, is less susceptible to environmental fluctuations that would otherwise rapidly dephase purely molecular transitions. This decoherence protection mechanism allows coherent dynamics to survive longer, manifesting as sustained oscillations in the population dynamics. Figure 2: Time evolution of the donor population p_{\text{D}}(t) for the direct transition coupling with \lambda_{0}=50\,\text{cm}^{-1} and -\Delta G=100\,\text{cm}^{-1} . Red solid line: cavity-coupled system ( \hbar\omega_{\text{c}}|t_{\text{DA}}|=50\,\text{cm}^{-1} ). Black dashed line: uncoupled system. Cavity coupling accelerates transfer and extends coherence time through the dynamic polaron decoupling effect. To assess the validity of perturbative approaches, we compare our numerically exact HEOM results with analytic expressions derived from Fermi’s golden rule (FGR) Semenov19 . FGR provides a widely used perturbative framework for calculating transition rates, treating the donor-acceptor coupling to second oder. For cavity-modified ET, the FGR expressions take different forms depending on the coupling channels. For the energy-fluctuation coupling channel ( g_{\text{D,A}} channel), the FGR rate is \displaystyle k_{\text{DA}}^{\text{FGR}}= \displaystyle\sqrt{\frac{\pi}{\hbar^{2}k_{\text{B}}T\lambda_{0}}}H_{\text{DA}}^{2}e^{-|g_{\text{D}}-g_{\text{A}}|^{2}}\sum_{m=0}^{\infty}\frac{|g_{\text{D}}-g_{\text{A}}|^{2m}}{m!} \displaystyle\times\exp\left\{-\frac{(-\Delta G-\lambda_{0}-m\hbar\omega)^{2}}{4k_{\text{B}}T\lambda_{0}}\right\}, (13) where the summation over photon number states m reflects the cavity-induced modulation of the activation barrier. For the direct transition coupling channel ( t_{\text{DA}} channel), the FGR rate is \displaystyle k_{\text{DA}}^{\text{FGR}}= \displaystyle\sqrt{\frac{\pi}{\hbar^{2}k_{\text{B}}T\lambda_{0}}}\Bigg[H_{\text{DA}}^{2}\exp\left\{-\frac{(-\Delta G-\lambda_{0})^{2}}{4k_{\text{B}}T\lambda_{0}}\right\} \displaystyle+|\hbar\omega t_{\text{DA}}|^{2}\exp\left\{-\frac{(-\Delta G-\lambda_{0}-\hbar\omega)^{2}}{4k_{\text{B}}T\lambda_{0}}\right\}\Bigg], (14) where the two terms represent the bare molecular transfer and the cavity-mediated transfer pathway, respectively. Figure 3 compares these FGR predictions (solid lines) with HEOM results (dots) as a function of driving force -\Delta G for \lambda_{0}=50\,\text{cm}^{-1} . Although FGR reproduces the overall dependence of the ET rate on the driving force, the HEOM results exhibit systematically narrower profiles and a noticeable shift of the peak position relative to the FGR curves. These discrepances stem from the perturbative nature of FGR. In contrast, HEOM fully captures nonperturbative couplings as well as non-Markovian memory effects that arise when the bath relaxation timescale is comparable to the characteristic timescale of electron transfer dynamics. These memory effects manifest as correlations between the system’s past and present states, leading to rate modifications and spectral narrowing that cannot be predicted by time-local, perturbative theories. Figure 3: Comparison of ET rates as a function of driving force -\Delta G for (a) the direct transition coupling channel ( t_{\text{DA}} ) and (b) the energy-fluctuation coupling channel ( g_{\text{D,A}} ), with \lambda_{0}=50\,\text{cm}^{-1} . Solid lines: Fermi’s golden rule predictions. Dots: numerically exact HEOM results. While FGR captures qualitative trends, it fails to reproduce the quantitative rates and spectral widths due to neglect of nonperturbative and non-Markovian effects. The limitations of perturbative theory become most pronounced when examining the dependence of the ET rate on light-matter coupling strength. According to Eq. ( 14 ), FGR predicts a quadratic scaling of the cavity-induced rate enhancement: \delta k_{\text{DA}}\propto|t_{\text{DA}}|^{2} , where \delta k_{\text{DA}}=k_{\text{cav}}-k_{\text{non-cav}} represents the change in transfer rate upon cavity coupling. Figure 4 compares this FGR prediction (blue curve) with HEOM results (red dots) as a function of the coupling strength |t_{\text{DA}}| . The quadratic scaling holds reasonably well only in the coupling regime |t_{\text{DA}}|\lesssim 0.3 (corresponding to \hbar\omega_{\text{c}}|t_{\text{DA}}|\lesssim 30\,\text{cm}^{-1} ). Beyond this threshold, the HEOM results deviate systematically from the perturbative prediction: the rate enhancement saturates rather than continuing to grow quadratically. This saturation signals entry into the ultrastrong coupling regime, where the light-matter interaction energy becomes comparable to other characteristic energy scales in the system. Notably, even within the moderate coupling window where quadratic scaling approximately holds, the HEOM-predicted coefficient differs quantitatively from the FGR value. This discrepancy reflects the fact that FGR neglects non-Markovian bath correlations and higher-order virtual processes that contribute to the effective rate even at modest coupling strengths. Figure 4: Cavity-induced change in ET rate, \delta k_{\text{DA}}=k_{\text{cav}}-k_{\text{non-cav}} (normalized by k_{\text{non-cav}} ) as a function of direct transition coupling strength |t_{\text{DA}}| for \lambda_{0}=50\,\text{cm}^{-1} and -\Delta G=100\,\text{cm}^{-1} . Red dots: HEOM results. Blue solid curve: FGR prediction (Eq. ( 14 )). The quadratic scaling \delta k_{\text{DA}}\propto|t_{\text{DA}}|^{2} breaks down beyond |t_{\text{DA}}|\sim 0.3 , where the rate saturates as the system enters the ultrastrong coupling regime. III.2 Resonance, Loss, and Collective Effects An important feature of cavity-modified ET is the resonance behavior that emerges when the cavity frequency is tuned relative to the molecular and environmental energies. Figure 5 shows the ET rate as a function of cavity frequency \omega_{\text{c}} for both coupling channels. To isolate the effect of cavity frequency from changes in coupling strength, we rescale the coupling parameters according to t_{\text{DA}},g_{\text{D,A}}\propto 1/\sqrt{\omega_{\text{c}}} (see Eqs. ( 5 )–( 7 )), keeping the effective light-matter interaction fixed at constant cavity mode volume \mathcal{V} . This normalization ensures that variations in the ET rate arise solely from energetic resonance conditions rather than from trivial changes in coupling magnitude. Both coupling channels exhibit clear resonance peaks where the ET rate is maximized, but the optimal frequencies differ significantly between the two mechanisms. For the direct transition coupling channel [panel (a)], the resonance occurs when the cavity frequency matches the effective energy gap between the donor and acceptor states. At this condition, the cavity mode is resonant with the electronic transition, enabling efficient cavity-mediated transfer through coherent emission and reabsorption of virtual photons. In contrast, for the energy-fluctuation coupling channel [panel (b)], the resonance condition is more subtle, as it depends on the energetic overlap between cavity-shifted donor and acceptor potentials modulated by the reorganization energy and thermal fluctuations. The cavity effectively renormalizes the energetics of both states, and the optimal frequency reflects a balance between activation barrier reduction and thermal accessibility. These distinct resonance behaviors highlight the qualitatively different mechanisms by which the two coupling channels modify electron transfer: the t_{\text{DA}} channel opens a new coherent transfer pathway through direct cavity-mediated coupling, while the g_{\text{D,A}} channel reshapes the free energy landscape governing the transfer process. Figure 5: Resonance behavior of the ET rate as a function of cavity frequency \omega_{\text{c}} with coupling parameters rescaled as t_{\text{DA}},g_{\text{D,A}}\propto 1/\sqrt{\omega_{\text{c}}} at fixed mode volume. (a) Direct transition coupling channel ( t_{\text{DA}} ): cavity-transition resonance. (b) Energy-fluctuation coupling channel ( g_{\text{D,A}} ): resonance reflects optimal cavity-induced modulation of the activation barrier. The different resonance frequencies underscore the distinct physical mechanisms of the two coupling channels. In realistic experimental implementations, optical cavities have finite quality factors due to photon leakage into external electromagnetic modes. This cavity loss, characterized by a photon lifetime \tau_{\text{c}} , causes spectral broadening of the cavity resonance and introduces dissipation into the light-matter dynamics. To capture these effects, we extend our model by incorporating Lindblad dissipation into the quantum master equation for the reduced density operator \hat{\rho} : \displaystyle\frac{\text{d}\hat{\rho}}{\text{dt}}=-\frac{i}{\hbar}\left[\hat{H},\hat{\rho}\right]+\frac{\kappa}{2}\left(2\hat{a}\hat{\rho}\hat{a}^{\dagger}-\hat{a}^{\dagger}\hat{a}\hat{\rho}-\hat{\rho}\hat{a}^{\dagger}\hat{a}\right), (15) where \kappa=\tau_{\text{c}}^{-1} is the cavity loss rate and the Lindblad term describes Markovian photon leakage from the cavity mode. This dissipator models the irreversible decay of cavity photons while preserving the positivity and trace of the density operator, as required for a physically valid quantum evolution. Figure 6 reveals a non-monotonic dependence of the ET rate enhancement on the cavity loss rate \kappa . This behavior reflects a competition between two opposing effects. In the low-dissipation regime, increasing cavity loss actually enhances the ET rate. Here, dissipation plays a constructive role by broadening the cavity resonance, which relaxes stringent energy-matching requirements and enables fluctuation-assisted transport. This mechanism is analogous to environment-assisted quantum transport observed in other contexts, where moderate decoherence can facilitate incoherent hopping between states that would otherwise remain energetically misaligned Rebentrost09 ; Phuc18 ; Phuc19b ; Phuc23Floquet . The photon leakage effectively provides an additional pathway for energy dissipation that can help the system overcome activation barriers. However, in the high-dissipation regime, excessive cavity loss becomes detrimental. Strong damping suppresses the coherent light-matter dynamics necessary for polariton formation and cavity-mediated transfer. The rapid photon decay prevents the buildup of coherent cavity excitation, essentially reverting the system toward the bare molecular limit. This inhibition resembles a quantum Zeno-like effect: frequent dissipative events disrupt the coherent evolution before it can facilitate electron transfer. The optimal cavity loss rate, corresponding to the peak in Fig. 6 , represents a balance between these competing mechanisms: sufficient dissipation to enable flexible energy matching, but not so much as to destroy the coherent polariton dynamics underlying cavity-enhanced transfer. Figure 6: ET rate enhancement k_{\text{cav}}/k_{\text{non-cav}} as a function of cavity loss rate \kappa for the direct transition coupling channel ( t_{\text{DA}} ) with \lambda_{0}=50\,\text{cm}^{-1} and -\Delta G=100\,\text{cm}^{-1} . The non-monotonic behavior reflects a competition between fluctuation-assisted transport in the low-dissipation regime and coherence suppression in the high-dissipation regime. The optimal enhancement occurs at intermediate loss rates where cavity-mediated coherent dynamics and dissipative broadening are balanced. A distinctive feature of cavity QED is the emergence of collective effects when multiple emitters couple to a shared cavity mode. Whether such collective coupling can influence electron transfer—a fundamentally local process occurring at individual molecular sites—is a non-trivial question, especially when the per-molecule light-matter coupling strength is modest. The answer depends on whether collective polariton formation modifies the energy landscape and decoherence dynamics in ways that align favorably or unfavorably with the conditions for efficient ET. To investigate collective effects systematically, we compare systems with N_{\text{mol}}=1 and N_{\text{mol}}=2 molecules while maintaining fixed per-molecule coupling strength. This ensures that the light-matter interaction per molecule remains constant, allowing us to isolate purely collective phenomena arising from cooperative coupling to the same cavity mode. All molecules are initialized in their donor states with the cavity in the vacuum state. In the absence of cavity coupling, ET occurs independently at each molecule, and the rate is naturally independent of N_{\text{mol}} . We neglect the dipole self-energy term in this analysis to focus on the essential collective physics. Figure 7 displays the relative ET rate enhancement k_{\text{cav}}/k_{\text{non-cav}} for both N_{\text{mol}}=1 and N_{\text{mol}}=2 across three different driving forces -\Delta G . The results reveal that cavity-modified ET rates depend sensitively on the number of coupled molecules, confirming the presence of collective effects. However, the the nature of this collectivity is not universal—it can be either enhancing or suppressing depending on system parameters. We identify three distinct regimes of collective behavior depending on the driving force: (i) Positive collective effect ( -\Delta G=-100\,\text{cm}^{-1} , center panel): Enhancement increases with molecular number, following k_{\text{non-cav}}