betaFeedback

WorksPhysics and Astronomy

Quantum Mechanical Assessment of Optimal Photovoltaic Conditions in Organic Solar Cells

Abstract

Recombination losses contribute to reduce JSCJ_{SC}, VOCV_{OC} and the fill factor of organic solar cells. Recent advances in non-fullerene organic photovoltaics have shown, nonetheless, that efficient charge generation can occur under small energetic driving forces (ΔEDAΔE_{DA}) and low recombination losses. To shed light on this issue, we set up a coarse-grained open quantum mechanical model for investigating the charge generation dynamics subject to various energy loss mechanisms. The influence of energetic driving force, Coulomb interaction, vibrational disorder, geminate recombination, temperature and external bias are included in the analysis of the optimal photovoltaic conditions for charge carrier generation. The assessment reveals that the overall energy losses are not only minimized when ΔEDAΔE_{DA} approaches the effective reorganization energy at the interface but also become insensitive to temperature and electric field variations. It is also observed that a moderate reverse bias reduces geminate recombination losses significantly at vanishing driving forces, where the charge generation is strongly affected by temperature.

Schematic representation of the donor–acceptor system and external electric field

Figure 4.

Efficient charge generation has always been a major concern for organic photovoltaics (OPV) due to the strong exciton binding energy in organic materials.[1, 2, 3, 4, 5] Thus, fullerene derivatives have been used as prototypical electron acceptor materials, because of their remarkable electron-accepting capabilities. However, the efficiency of fullerene-based organic solar cells (OCS) has stagnated around 13% as a consequence of the excessive energy losses undergone to achieve electron-hole (e-h) charge separation, and the very poor optical characteristics of the C60_{60} molecule that does not contribute to charge generation. Meanwhile, a new class of non-fullerene acceptor (NFA) materials have been developed, which revealed a new paradigm for OPV, one of efficient charge generation and high output voltage despite lower energetic gradients at the heterojunction. Nowadays, the conversion efficiency of the NFA devices is approaching the barrier of 20%,[2] as a result of the high photo-absorption of non-fullerene small molecules; the push-pull design of the NFA molecules that facilitate the electron-hole pair separation upon photoexcitation; the complementary light-harvesting roles played by large band-gap donor materials and low-bandgap NFA materials; besides the beneficial morphological properties of the NFA molecules (e.g., miscibility, stability, and planarity). For OSCs the output voltage is given by Vout=ECT/q−ΔVrecr−ΔVrecnrV_{out} = E_{CT}/q - \Delta V_{rec}^{r} - \Delta V_{rec}^{nr},[6, 7, 8] where ECT/qE_{CT}/q is related to the energetic driving force[7, 9] and the voltage losses due to recombination comprise both radiative (ΔVrecr\Delta V_{rec}^{r}) and non-radiative (ΔVrecnr\Delta V_{rec}^{nr}) pathways. In NFA-OSCs the voltage loss mechanisms have been empirically reduced whereas the interplay among the relevant physical processes at the D:A interface remains a subject of study.[6, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]

In this paper we combine the Ehrenfest and Redfield methods in a coarse-grained open quantum mechanical model to investigate, from a fundamental point of view, the energy loss effects that occur during the free charge carrier generation process at the D:A interface. We study the influence of the energetic driving, temperature and external bias on the charge generation process. The simulation results demonstrate that the overall energy losses are minimized at the activationless regime of the charge separation, where driving force matches the effective reorganization energy of the heterojunction, corroborating experimental studies.[20, 21, 22] Moreover, the optimal photovoltaic condition is insensitive to variations of temperature and reverse bias.

We partition the total Hamiltonian, responsible for the charge generation and recombination processes, as

H=HS+HB+HSB.(1)H = H_{S} + H_{B} + H_{SB}. \tag*{(1)}

The system Hamiltonian, HSH_S, comprises the degrees of freedom of the electron and the hole and a subset of vibrational reorganization modes that are described as classical modes within the framework of the Ehrenfest self-consistent method. Thus, HSH_S is a mixed quantum-classical Hamiltonian (details provided in Support Information). The bath Hamiltonian, HBH_B, accounts for the degrees of freedom of the environment, namely the vibrational degrees of freedom and fluctuations of the dielectric background, which we describe as an ensemble of quantum harmonic oscillators. The system-bath coupling (HSBH_{SB}) is described within the framework of the Redfield theory. Using the adiabatic representation for the system Hamiltonian, HS∣φa⟩=Ea∣φa⟩H_S|\varphi_a\rangle= E_a|\varphi_a\rangle, the Redfield equation for the reduced density matrix σ\sigma reads

∂σab∂t=−iωabσab+∑c,dRabcdσcd(2)\frac{\partial\sigma_{ab}}{\partial t} = -i\omega_{ab}\sigma_{ab} + \sum_{c,d} R_{abcd}\sigma_{cd} \tag*{(2)}

where ωab=(Ea−Eb)/ℏ\omega_{ab} = (E_a-E_b)/\hbar designates the eigenfrequencies of the electronic system and RabcdR_{abcd} is the Redfield relaxation tensor.[23, 24] The first term on the right-hand-side (RHS) of Eq. (2) describes the coherent quantum dynamics of S, the second term describes the interaction of S with the environment. The implementation of the Redfield equations for this model has been described elsewhere.[25] Later on we will incorporate a term for the recombination effects into Eq. (2).

The D:A interface is modelled as a two-dimensional (2D) coarse-grained lattice, with energy profile corresponding to a staggered (type II) interface, as described in Figure 1 (PDF p. 4). The spatial arrangement of the 2D-D:A interface is shown at the bottom of the figure, where each site of the lattice represents a molecular site in the coarse-grained model. Periodic boundary conditions are applied along the transverse direction. In the simulations, we consider the photoexcitation of a molecular site of the donor material, but the obtained results are equally valid for a photoexcitation in the acceptor material. After charge separation, the electron and hole are collected at the respective collecting layers at the border of the 2D lattice (see Figure 1 (PDF p. 4)). The electron and hole collecting layers in the model do not represent the actual cathode and anode terminals of the device, they are simply used as a theoretical tool to avoid the reflection of the wavepackets at the left and right boundaries; they are positioned below ECB(A)E_{CB}(A) and above EVB(D)E_{VB}(D), respectively.

Model for the donor-acceptor interface

Figure 1 (PDF p. 4). Model for the donor-acceptor interface. The figure on top shows the energetics of the D:A interface, including the electron and hole collectors. The gaussian envelopes depict the electron-hole excitation dynamics and charge separation. The upward arrow indicates the photoexcitation of the electron-hole pair (black) whereas the downward dashed arrows describe overall radiative and non-radiative recombination decays. At the bottom, the spatial configuration of the two-dimensional model of the D:A interface. Each molecule is described as a site in a 2D lattice.

The coarse-grained Hamiltonian HSH_S, for either the electron and hole photoexcited particles, is

HS≡Hel/hl=∑iN{Eiel/hl+εi(t)−Φiel/hl(t)}∣i⟩⟨i∣+∑i≠jNVij(t)∣i⟩⟨j∣,(3)H_S \equiv H^{\mathrm{el/hl}} = \sum_{i}^{N} \left\{ E_i^{\mathrm{el/hl}} + \varepsilon_i(t) - \Phi_i^{\mathrm{el/hl}}(t) \right\} |i\rangle\langle i| + \sum_{i\ne j}^{N} V_{ij}(t)|i\rangle\langle j| , \tag*{(3)}

where NN denotes the total number of molecular sites in the lattice. The orthogonal basis set {∣i⟩}\{|i\rangle\} consists of diabatic states associated with each of the molecular sites. In addition, Eiel=EiCBE_i^{\mathrm{el}} = E_i^{\mathrm{CB}} and Eihl=EiVBE_i^{\mathrm{hl}} = E_i^{\mathrm{VB}} are the on-site energies of the electron (conduction band) and hole (valence band); εi\varepsilon_i is the confinement energy associated with site ii; and Φiel/hl\Phi_i^{\mathrm{el/hl}} describes the electron-hole electrostatic interaction, given in the time-dependent mean-field approximation by

Φiel/hl(t)=ξbind(∑jPjhl/el1+dij).(4)\Phi_i^{\mathrm{el/hl}}(t) = \xi_{\mathrm{bind}} \left( \sum_j \frac{P_j^{\mathrm{hl/el}}}{1+d_{ij}} \right) . \tag*{(4)}

where ξbind\xi_{\mathrm{bind}} is the on-site electron-hole binding energy (refer to Table I for a complete list of model parameters), PjelP_j^{\mathrm{el}} (PjhlP_j^{\mathrm{hl}}) is the time-dependent electron (hole) population on site jj and dij=∣R⃗i−R⃗j∣d_{ij} = |\vec{R}_i-\vec{R}_j| is the distance between molecular sites ii and jj. The treatment accounts for the decrease of the el-hl Coulomb barrier induced by the charge delocalization, besides describing the influence of temperature in the el-hl separation, as shown elsewhere.[25] The last term in Eq. (13),

valuedescription
NN84number of sites
ℓo\ell_o0.5 nmconfinement radius for the empty molecular site
VnnV^{nn}−40-40 meVnearest-neighbor electronic coupling
dijd_{ij}1.3 nmdistance of nearest neighbor sites
ξbind\xi_{\mathrm{bind}}0.2 eVelectron-hole coupling constant
λ\lambda50 meVreorganization energy
ωc\omega_c1334 cm−1^{-1} (165 meV)cutoff frequency
Ω\Omega272 cm−1^{-1} (34 meV)classical normal mode frequency
γ−1\gamma^{-1}0.1 psphonon-phonon relaxation time
κ\kappa1 ps−1^{-1}molecular site decay rate constant

Table I. Values and description of the parameters used in calculations (refer to SI for details).

Vij=V0Fij=V0[2ℓiℓjℓi2+ℓj2exp⁡(−dij2ℓi2+ℓj2)](5)V_{ij}=V_0 F_{ij}=V_0\left[\frac{2\ell_i\ell_j}{\ell_i^2+\ell_j^2}\exp\left(\frac{-d_{ij}^2}{\ell_i^2+\ell_j^2}\right)\right] \tag*{(5)}

describes the tunneling energy between lattice sites, where V0V_0 is the bare electronic coupling and the form factor FijF_{ij} is the time-dependent overlap between lattice sites ii and jj, with ℓ\ell being the confinement radius of the molecular site (see SI for details). The time dependence of the energy parameters comprising Eq. (13), namely εi(t)\varepsilon_i(t) and Vij(t)V_{ij}(t) gives rise to the intra-molecular and inter-molecular vibrational effects, respectively, that we associate with the Holstein and Peierls couplings.[25] The treatment includes dynamic disorder, which is taken into account by the intra-molecular and inter-molecular electron-phonon couplings in the system Hamiltonian (HSH_{\mathrm{S}}). The dynamic disorder is also taken into account in the system-bath Hamiltonian (HSBH_{\mathrm{SB}}) that gives rise to the Redfield relaxation tensor and, lastly, through the action of the classical Berendsen thermostat that acts on the molecular sites (describe in SI). These interactions give rise to localization of the electron and hole wavepackets,[26] polaron formation,[25] and temperature dependence of the charge generation rate. The model does not include static disorder.

Recombination is the main efficiency loss mechanism in photovoltaics, even more so in OSCs due to the convoluted charge separation process in organic devices. Recombination processes can be separated in two types: geminate and non-geminate (including bimolecular and trap-assisted processes). The former designates the annihilation of a correlated electron-hole pair generated by the same photoexcitation event, or the recombination of a charge separated (CT) state before complete dissociation. Experiments have shown that geminate recombination is more important at short-circuit and low illumination conditions.[27] Bimolecular recombination, on the other hand, involves fully dissociated carriers generated by independent photoabsorption events. It is more relevant at open-circuit and/or high excitation conditions. A quantum mechanical model that is well suited for describing the geminate recombination in molecular systems is the Haberkorn model. It was initially proposed to describe recombination in radical pairs,[28] but later was used to describe the trapping and exciton recombination in the Fenna-Matthews-Olson (FMO) protein complex[29] and OSC models.[26] Herein we include the vibrational dynamics, in addition to the electron-hole dynamics to extend the plain Haberkorn formalism. Thus, consider a quantum state ∣Ψη(t)⟩=∑iCiη(t)∣i⟩|\Psi^\eta(t)\rangle=\sum_i C_i^\eta(t)|i\rangle that comprises the reduced density matrix ση\sigma^\eta, where η=el,hl\eta=el,hl des-

Equation not extracted

ignates the type of charge carrier. As in the Haberkorn model, the mutual annihilation of electrons and holes is given by the dynamical equation

ddt∣Ψη(t)⟩=−Γ^Rη′∣Ψη(t)⟩(6)\frac{d}{dt}|\Psi^{\eta}(t)\rangle=-\hat{\Gamma}_{R}^{\eta'}|\Psi^{\eta}(t)\rangle \tag*{(6)}

where η′\eta' is the carrier type opposite to η\eta. To include the vibrational effects into the Haberkorn recombination model, we use the operator Γ^Rη′\hat{\Gamma}_{R}^{\eta'}, defined as

Γ^Rη′=∣Ψη′(t)⟩ κF^ ⟨Ψη′(t)∣,(7)\hat{\Gamma}_{R}^{\eta'}=|\Psi^{\eta'}(t)\rangle\ \kappa\hat{F}\ \langle\Psi^{\eta'}(t)|, \tag*{(7)}

where the symmetric operator κF^=∑iκFii∣i⟩⟨i∣+∑i≠jκFij∣i⟩⟨j∣\kappa\hat{F}=\sum_i\kappa F_{ii}|i\rangle\langle i|+\sum_{i\ne j}\kappa F_{ij}|i\rangle\langle j| includes the molecular site vibrations, as defined in Eq. (14). The parameter κ\kappa in Eq. (20) (see Table I) is a molecular decay rate constant. Coropceanu and collaborators[10, 14] associate κFii\kappa F_{ii} with the Einstein spontaneous emission rate for the radiative recombination term whereas the Marcus-Levich-Jortner (MLJ) formula was used to define the non-radiative decay rate κFij\kappa F_{ij}; herein we use the same parameter κ\kappa for both processes. A comparison between the vibronic Haberkorn model and the three-state model[10, 12, 14, 16, 17] is presented as SI. Substitution of Eq. (20) into Eq. (6) yields the following equation for the coefficients of the wavepacket in the molecular site basis

ddtCiη=−κ(∑k,l[Ckη′]∗FklClη′)Ciη′.(8)\frac{d}{dt}C_i^\eta=-\kappa\left(\sum_{k,l}[C_k^{\eta'}]^*F_{kl}C_l^{\eta'}\right)C_i^{\eta'}. \tag*{(8)}

Rewriting Eq. (8) in terms of the time derivative of the reduced density matrix, σijη=Ciη[Cjη]∗\sigma_{ij}^{\eta}=C_i^\eta[C_j^\eta]^*, we get

ddtση∣rec=−κ(ση′F^ση+σηF^ση′).(9)\frac{d}{dt}\sigma^\eta\big|_{\mathrm{rec}}=-\kappa\left(\sigma^{\eta'}\hat{F}\sigma^\eta+\sigma^\eta\hat{F}\sigma^{\eta'}\right). \tag*{(9)}

The quantum mechanical recombination rate given by Eq. (9), for geminate recombination, can be associated with the classical law of mass action for the rate of bimolecular recombination R=knpR=knp, where nn and pp are the concentrations of free electrons and holes, respectively, and kk is an effective recombination constant given, for instance, by the Langevin model.[30] It is also worth noting that Eq. (9) gives the recombination losses during the exciton diffusion and for the CT state at the D:A interface.

Before adding this recombination rate term to the Redfield Master Equation, Eq. (2), it is transformed to the adiabatic representation of basis states {∣φa⟩}\{|\varphi_a\rangle\} by σ˙recη→(U^η)−1σ˙recηU^η\dot{\sigma}_{\mathrm{rec}}^\eta\rightarrow(\hat{U}^\eta)^{-1}\dot{\sigma}_{\mathrm{rec}}^\eta\hat{U}^\eta, where U^η\hat{U}^\eta is the unitary operator that diagonalizes H^η\hat{H}^\eta. The overall charge neutrality of the heterojunction is conserved with Eq. (9). Based on Eq. (9), the energy lost by a molecular site via recombination is calculated as Wirec=σiirec(t)EigapW_i^{rec} = \sigma_{ii}^{rec}(t)E_i^{gap} where σiirec\sigma_{ii}^{rec} is the population lost by recombination and EigapE_i^{gap} is the corresponding CB-VB gap. The total energy loss due to recombination is calculated as ∑iWirec(tf)\sum_i W_i^{rec}(t_f), where the sum runs over all the molecular sites and tft_f is the final simulation time. Therefore, to obtain the voltage loss due to geminate recombination (ΔVrec\Delta V^{rec}) during the charge separation process we calculate

ΔVrec=1e∑iWirec(tf)(10)\Delta V^{rec} = \frac{1}{e}\sum_i W_i^{rec}(t_f) \tag*{(10)}

where ee is the electronic charge. Notice that this method accounts only for the geminate electron-hole recombination. Bimolecular recombination processes become important at open circuit conditions and high excitation intensities, which are not described by the present simulations.

The model parameters used in the simulations are presented in Table I. The set of coupled differential equations produced by combining Eqs. (2) and (9) are solved numerically by a fourth order adaptive step size Runge-Kutta method. We simulate the process of free charge generation, including exciton photoexcitation, excitation diffusion, CT state formation and separation, for a single el-hl pair in the presence of geminate recombination, during a time frame of ten picoseconds.

We consider that a Frenkel exciton is photoexcited in the center of the donor region, as shown in Figure 1 (PDF p. 4). The coarse-grained model consists of a 2D-square lattice with nearest neighbor (nn) distance a=1.3a = 1.3 nm and equal D and A domain lengths of approximately 7.5 nm. The effective electronic coupling between nearest neighbors is set to be Vnn=−40V^{nn} = -40 meV, as given by Eq. (14), which is in agreement with other molecular crystal models.[26, 31, 32, 33, 34, 35, 36] The low frequency vibration mode Ω\Omega is treated classically. It modulates the radii ℓi(t)\ell_i(t) of the molecular sites and, thereby, the electronic coupling VijV_{ij}, as given by (S4). The classical vibrations are self-consistently coupled to the el-hl quantum states by semiclassical Ehrenfest forces (Supporting Information). The high energy quantum vibrations are incorporated into the Redfield relaxation tensor, up to the cutoff frequency ωc=1334\omega_c = 1334 cm−1^{-1} that corresponds to the localized C=C stretching mode. For the system-bath coupling we use a reorganization energy parameter of λ=50\lambda= 50 meV for either the electron and hole density matrices. The magnitude of λ\lambda is commonly found to be in the range of 100-300 meV for organic blends.[37] However, it has been shown that the values of λ\lambda obtained from experiments tend to be larger than the bare values obtained by static calculations. Vanderwal et al.[37] derived the empirical relationship λexp=λcalc+Δ\lambda_{exp} = \lambda_{calc} + \Delta, where Δ≈95\Delta\approx95 meV is a correction due to dynamic disorder (disregarding the static disorder). Additionally, McGehee et al.[30] derived another relation in which λexp=λcalc+σ2/(2kT)\lambda_{exp} = \lambda_{calc} + \sigma^2/(2kT), where σ\sigma is the static energetic disorder parameter and σ2/(2kT)≈108\sigma^2/(2kT) \approx108 meV at room temperature, assuming σ=75\sigma= 75 meV. λexp\lambda_{exp} can be associated with the CT emission spectral broadening that is caused by disorder. In our model we ascribe the effects of disorder fully to the molecular vibrations. Therefore, the effective reorganization energy affecting the el-hl dynamics in our simulations can be estimated as λeff≈50\lambda_{eff} \approx50 meV +100+ 100 meV =150= 150 meV. This is in good agreement with the optimal charge generation conditions obtained in our simulations, which is given EDA≈λeffE_{DA} \approx\lambda_{eff} that occurs between 100 and 200 meV, as shown ahead. As for the recombination rate constant, κ\kappa, its value is defined based on the characteristics of the coarse-grained model. Experiments have determined that the exciton diffusion length in organic semiconductors and polymers is within the range of 5-20 nm,[38, 39] and the diffusion length is limited by the exciton lifetime. In our coarse-grained model the length of the donor (D) domain is about 7.5 nm, which should correspond approximately to an exciton lifetime. To follow this constraint and to evince the effects of the recombination within the span of the simulation time (10 ps), we use κ=1\kappa= 1 ps−1^{-1}. For the sake of comparison, Figure S4 in the Supporting Information presents simulation results with and without recombination. Finally, in accordance with the model (Figure 1 (PDF p. 4)), we define the energetic driving force of the D:A interface as EDA≡ECB(D)−EVB(A)E_{DA} \equiv E^{CB}(D) - E^{VB}(A).

Calculations of time-dependent charge collection and energy collection with and without recombination

Figure S4. Calculations of time-dependent charge collection, a) and b), and energy collection, c) and d). Left-hand-side (LHS) panel: no recombination effects. Right-hand-side (RHS) panel: with recombination effects. Calculations performed at TB=300T_B = 300 K for various LUMO level offsets, EDAE_{DA}, as indicated by the legend.

The evaluation of the activation energy (EaE_a) required for el-hl pair separation has been open to debate,[6, 9, 11, 20, 21, 40, 41] given that the obtained values may vary depending on the technique and the time scale of the measurements. From the theoretical viewpoint, considering the kinetic models, EaE_a is an effective parameter that is influenced by Coulombic dissociation and recombination rates.[9, 11, 40] Figure 2 shows the collected electronic population after recombination, for various bath temperatures (TBT_B) and energetic driving forces (EDAE_{DA}), that is Ne(EDA,TB)N_e(E_{DA}, T_B), without electric field bias. The simulations show, as a whole, that the charge generation rate increases with temperature for all D:A offsets. In addition, the effect of TBT_B on NeN_e reveals three distinct electron separation regimes, as depicted in the inset of Figure 2: a) the normal electron separation for negligible EDA=0E_{DA} = 0 eV; b) the activationless regime for EDA≈0.1E_{DA} \approx0.1 eV; and c) the inverted Marcus regime for EDA=0.3E_{DA} = 0.3 eV. The highest internal quantum efficiency (IQE) is obtained for the activationless regime, EDA≈0.1E_{DA} \approx0.1 eV, whereby the charge generation process is almost independent of the temperature. The maximum IQE is obtained for EDA≈λeffE_{DA} \approx\lambda_{eff}, where λeff\lambda_{eff} corresponds to the effective reorganization energy at the heterojunction, corroborating experimental studies that show the existence of such an optimal driving force.[20, 21, 22] However, for vanishing energetic driving force, EDA≈0E_{DA} \approx0, and no electric fields (to be discussed next), the charge generation is very sensitive to the temperature. Lastly, if the energy offset becomes excessive, i.e. EDA≫λE_{DA} \gg\lambda, the charge generation process enters the inverted region described by the Marcus-Jortner-Levich theory[42, 43, 44], where charge transfer is assisted by quantum vibronic effects. A detailed account of the impact produced by vibronic effects (such as polaron formation) and internal electric fields on the free charge carrier generation is was presented by Andermann and Rego elsewhere.[25] Figure S2 and Figure S3 of the SI evince the importance of electron-hole interaction and vibrational dynamics to recombination, respectively.

Collected electronic population subjected to recombination losses for various bath temperatures and donor-acceptor energy offsets

Figure 2. Collected electronic population (NeN_e) subjected to recombination losses, for various bath temperatures (TBT_B) and donor-acceptor energy offsets: a) EDA=0E_{DA} = 0 eV, b) EDA=0.1E_{DA} = 0.1 eV and c) EDA=0.3E_{DA} = 0.3 eV. The illustration at the top-right corner is a reference to the Marcus Theory, depicting electron transfer in the normal regime (a), the activationless regime (b), and the inverted regime (c) as the energetic driving force EDAE_{DA} increases.

Effect of electron-hole interaction on geminate recombination during charge separation

Figure S2. Effect of electron-hole interaction Φiel/hl(t)∣i⟩⟨i∣\Phi_i^{\mathrm{el/hl}}(t)|i\rangle\langle i|, from Equation (4) of the paper, on the geminate recombination rate during charge separation. The bath temperature is TB=300T_B = 300 K. The energetic driving force EDAE_{DA} is: 0 (a), 0.1 eV (b), 0.2 eV (c), 0.3 eV (d), and 0.4 eV (e).

Effect of vibrational dynamics on the geminate recombination rate during charge separation

Figure S3. Effect of vibrational dynamics (also considered as dynamical disorder) on the geminate recombination rate during charge separation. The bath temperature is TB=300T_B = 300 K. The energetic driving force EDAE_{DA} is: 0 (a), 0.1 eV (b), 0.2 eV (c), 0.3 eV (d), and 0.4 eV (e).

In addition to the energetic driving force for charge separation, the voltage losses in the OSCs also include the contributions from recombination processes, both radiative and non-radiative. Figure 3 shows the effect of driving force (EDAE_{DA}), temperature (TBT_B) and electric field bias (E\mathcal{E}) on the voltage loss due to geminate recombination, ΔVrec(EDA,TB,E)\Delta V_{rec}(E_{DA}, T_B, \mathcal{E}), as given by Eq. (10). The simulations show that ΔVrec\Delta V_{rec} is practically insensitive to temperature and reverse bias field if the energetics of the heterojunction is set to the activationless regime of charge separation, in this case EDA=0.1E_{DA} = 0.1 eV. Out of this point the losses due to geminate recombination at the D:A interface increase as the temperature decreases, particularly so for vanishing D:A offset in the absence of reverse bias (Figure 3-a). The influence of temperature on the geminate recombination losses during charge carrier-generation has not been extensively investigated because it is difficult to measure geminate recombination at low illumination intensities, since parasitic leakage currents can affect the results.[45] To study this issue, Gao et al.11`` used low excitation intensities to observe a decreasing VOCV_{OC} at low temperatures, which was ascribed to reduced carrier generation. Here, it is important to distinguish the nonequilibrium carrier generation process that is observed at low illumination intensities from the quasi-equilibrium conditions attained at open-circuit with moderate-to-high illumination. In the first case, the efficiency of charge generation is determined by the competition between photogeneration of free charge carriers and geminate recombination, as described in Figure 3. In the second case, bimolecular recombination kinetics is the main cause of energy loss, so that at the steady-state qVOC=Eg−kBTln⁡(NcNv/np)qV_{\mathrm{OC}}=E_{\mathrm{g}}-k_{\mathrm{B}}T\ln(N_{\mathrm{c}}N_{\mathrm{v}}/np), where nn and pp are the electron and hole stationary free carrier densities.[6, 11, 30] Thus, at high excitation densities nn and pp are nearly temperature independent and VOCV_{\mathrm{OC}} decreases linearly with temperature.[11, 46] At low excitations, however, the density of free carriers become temperature dependent, n(T)n(T) and p(T)p(T),[11, 15, 22, 46] thereby changing the behavior of VOC(T)V_{\mathrm{OC}}(T). Chan et al.{sup}15 reported that as the blends P3TEA:SF-PDI2 and PM6:Y6 are cooled, the reduced charge separation rate leads to increasing charge recombination, either directly at the D:A interface or via the LE state. Thus, in this case $V_{\mathrm{OC}}$ can exhibit a sublinear dependence on temperature,{citep}22,46,47 or even increase with temperature{citep}11` if the carrier-generation process dominates VOCV_{\mathrm{OC}}. Notice, though, that the application of a reverse bias at the heterojunc-

Geminate recombination induced voltage-loss as a function of charge-transfer driving force, temperature, and bias field

Figure 3. a) Geminate recombination induced voltage-loss, ΔVrec(EDA,TB,E)\Delta V^{\mathrm{rec}}(E_{\mathrm{DA}},T_{\mathrm{B}},\mathcal{E}), as calculated with Eq. (10), without bias field (E=0\mathcal{E}=0), for TB=100T_{\mathrm{B}=100}, 200, 300 and 400 K. b) Same as in a) with a reverse bias field E=−10\mathcal{E}=-10 V/μ\mum. The dashed lines indicate the approximate position of the optimal charge generation regime, which does not change appreciably under the influence of temperature or reverse bias. c) Temperature and bias field dependence of geminate recombination voltage-loss without D:A driving force, ΔVrec(EDA=0,TB,E)\Delta V^{\mathrm{rec}}(E_{\mathrm{DA}}=0,T_{\mathrm{B}},\mathcal{E}).

tion should promote the free carrier generation,[48] thereby decreasing the losses by geminate recombination and improving VOCV_{\mathrm{OC}}, as shown in Figure 3-b. The effect of a reverse bias is particularly strong in the absence of driving force (EDA=0E_{\mathrm{DA}}=0), as shown in Figure 3-c, where it is shown that a moderate bias reduces recombination losses substantially.

The various recombination pathways contribute to reduce JSCJ_{\mathrm{SC}}, VOCV_{\mathrm{OC}} and the fill factor (FF) of the OSCs, thereby limiting its power conversion efficiency. In regard to free carrier generation, this work indicates that the best device performances should be realized at the optimal offset EDA≈λeffE_{\mathrm{DA}} \approx\lambda_{\mathrm{eff}}, which produces low overall energy losses for free carrier generation and superior JSCJ_{\mathrm{SC}} and FF.

Acknowledgements. AMA is grateful for the financial support from FAPESC (Fundação de Amparo à Pesquisa do Estado de Santa Catarina). LGCR acknowledges support by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior Brasil (CAPES) - Finance Code 001, by the Brazilian National Counsel of Technological and Scientific Development (CNPq) and the National Institute for Organic Electronics (INEO). The authors acknowledge support by allocation of supercomputer time from Laboratory for Scientific Computing (LNCC/MCTI, Brazil).

References

Supporting Information for Quantum Mechanical Assessment of Optimal Photovoltaic Conditions in Organic Solar Cells

Semiclassical Electron-Phonon Hamiltonian

The paraboloids depicted in Figure S1-a) describe the local confinement potential felt by the electron and the hole at a given molecular site. We assume that the parabolic confinement potential undergoes a vibrational reorganization whenever charge is transferred in or out of the molecular site. This mechanism gives rise to coherent vibrational effects that we describe within the framework of the Ehrenfest method. The density profile of the electron and hole in the molecular site is given by the gaussian

Schematics of the electronic relaxation and semiclassical electron-phonon system

Figure S1. a) Schematics of the electronic relaxation. The parabola describes the confinement potential felt by the electron (or hole) with its gaussian wavefunction profile. On the left-hand side we show an empty molecular site, with confinement energy εo\varepsilon_{\mathrm{o}} and the corresponding confinement length ℓo\ell_{\mathrm{o}}. On the right-hand side we show an occupied molecular site, with on-site energy ε\varepsilon and ℓ\ell. The electronic relaxation energy is Δε=εo−ε\Delta\varepsilon= \varepsilon_{\mathrm{o}} - \varepsilon. b) The semiclassical electron-phonon system is equivalent to a spring-damper hybrid quantum-classical system.

g(r⃗−R⃗)=2πℓ2exp⁡[−(r⃗−R⃗ℓ)2](11)g(\vec{r}-\vec{R})=\sqrt{\frac{2}{\pi\ell^{2}}}\exp\left[-\left(\frac{\vec{r}-\vec{R}}{\ell}\right)^{2}\right] \tag*{(11)}

with R⃗\vec{R} designating the position of the molecular site, and ℓ\ell is the confinement radius of the molecular site, given by

ℓ=2ℏ2meε(12)\ell=\sqrt{\frac{2\hbar^{2}}{m_{\mathrm{e}}\varepsilon}} \tag*{(12)}

for a gaussian wavepacket of on-site energy ε\varepsilon.

To write the system Hamiltonian, HSH_{S}, we assume an orthogonal basis set comprised of diabatic electronic states {∣i⟩}\{\lvert i\rangle\} associated with the molecular sites, as described in the paper,

HS≡Hel/hl=∑iN{Eiel/hl+εi(t)−Φiel/hl(t)}∣i⟩⟨i∣+∑i≠jNVij(t)∣i⟩⟨j∣.(13)H_{S}\equiv H^{\mathrm{el/hl}}=\sum_{i}^{N}\left\{E_{i}^{\mathrm{el/hl}}+\varepsilon_{i}(t)-\Phi_{i}^{\mathrm{el/hl}}(t)\right\}\lvert i\rangle\langle i\rvert+\sum_{i\ne j}^{N}V_{ij}(t)\lvert i\rangle\langle j\rvert. \tag*{(13)}

The coupling VijV_{ij} between lattice sites is given by

Vij=V0F(ℓi,ℓj,dij)=V0[2ℓiℓjℓi2+ℓj2exp⁡(−dij2ℓi2+ℓj2)],(14)V_{ij}=V_{0}F(\ell_{i},\ell_{j},d_{ij})=V_{0}\left[\frac{2\ell_{i}\ell_{j}}{\ell_{i}^{2}+\ell_{j}^{2}}\exp\left(\frac{-d_{ij}^{2}}{\ell_{i}^{2}+\ell_{j}^{2}}\right)\right], \tag*{(14)}

where V0V_{0} is the bare electronic coupling and the form factor FF results from the overlap between gaussian wavepackets located at lattice sites ii and jj

F(ℓi,ℓj,dij)=∫gi(r⃗−R⃗i)gj(r⃗−R⃗j)dxdy.(15)F(\ell_{i},\ell_{j},d_{ij})=\int g_{i}(\vec{r}-\vec{R}_{i})g_{j}(\vec{r}-\vec{R}_{j})dxdy. \tag*{(15)}

The time dependence of εi(t)\varepsilon_{i}(t) and Vij(t)V_{ij}(t), which comprise Eq. (13), gives rise to *intra-*molecular and *inter-*molecular vibrational couplings that we associate with the Holstein and Peierls couplings, respectively.

We describe the coherent coupling between the electron-hole states and the fundamental intra-molecular vibrational mode via the Ehrenfest method. In this model the semiclassical electron-phonon system is equivalent to a spring-damper hybrid quantum-classical system, as shown in Figure S1-b) The classical Hamiltonian hih_i describing the vibrational mode of site ii is given by

hi=pi22μ+μΩ22(ℓi−ℓo)2+Weh+Qthermo,(16)h_i = \frac{p_i^2}{2\mu} + \frac{\mu\Omega^2}{2}(\ell_i-\ell_{\mathrm{o}})^2 + W^{\mathrm{eh}} + Q^{\mathrm{thermo}}, \tag*{(16)}

where ℓi\ell_i is the vibrational coordinate associated with the molecular site ii and pip_i is the conjugate momentum. In addition, ℓo\ell_{\mathrm{o}} is the confinement length for an empty (that is, neutral) site, μ\mu is the effective mass of the vibrational mode and Ω\Omega is the frequency of the relevant normal mode. Refer to Table 1 in the paper for a complete list of model parameters. The energy WehW^{\mathrm{eh}} is the work exchanged with the electronic degrees of freedom via the Ehrenfest force FehF^{\mathrm{eh}} and QthermoQ^{\mathrm{thermo}} is the heat exchanged with the classical thermostat. A similar hamiltonian has been considered by Egorova et al.[49, 50] for quantum intra-molecular degrees of freedom. The Ehrenfest force that acts on the vibrational coordinate of a given molecular site kk is

Fkeh=Fkel+Fkhl=−∂Uel∂ℓk−∂Uhl∂ℓk,(17)F_k^{eh}=F_k^{\mathrm{el}}+F_k^{\mathrm{hl}}=-\frac{\partial U^{\mathrm{el}}}{\partial\ell_k}-\frac{\partial U^{\mathrm{hl}}}{\partial\ell_k}, \tag*{(17)}

with Uel=Tr[σelHel]U^{\mathrm{el}}=Tr[\sigma^{\mathrm{el}}H^{\mathrm{el}}] and Uhl=Tr[σhlHhl]U^{\mathrm{hl}}=Tr[\sigma^{\mathrm{hl}}H^{\mathrm{hl}}]. Solving equation (S7) we obtain[25]

Fkeh=−(σkkel+σkkhl)dεkdℓk−∑i≠j(σijel+σijhl)∂Vij∂ℓk.(18)F_k^{eh}=-(\sigma_{kk}^{\mathrm{el}}+\sigma_{kk}^{\mathrm{hl}})\frac{d\varepsilon_k}{d\ell_k}-\sum_{i\ne j}(\sigma_{ij}^{\mathrm{el}}+\sigma_{ij}^{\mathrm{hl}})\frac{\partial V_{ij}}{\partial\ell_k}. \tag*{(18)}

The first term on the right-hand side (RHS) of Equation (S8) is responsible for the relaxation of the molecular site due to its charge occupation; this is the classical work exchanged between electronic and vibrational degrees of freedom. The second term gives rise to coherent vibronic effects, for it is proportional to the off-diagonal elements of σ\sigma and couples the electronic and nuclear coordinates via ∂Vij/∂ℓk\partial V_{ij}/\partial\ell_k.[51]

The classical equations of motion derived from Eq. (16) are solved with the velocity Verlet algorithm, coupled to a classical bath, described by the Berendsen thermostat,[52] with relaxation constant γ−1=0.1\gamma^{-1}=0.1 ps. The perturbing forces change the equilibrium configuration of the parabolic confining potential, which reacts with a restoring force Fres=−μΩ2(ℓ−ℓo)F^{\mathrm{res}}=-\mu\Omega^2(\ell-\ell_o).

Vibronic Recombination Model

In order to compare the vibronic recombination model with the three-state recombination model,[10, 12, 14, 16, 17] let us consider the dynamical equation for recombination, Eq. (6), in the paper

ddt∣Ψη′(t)⟩=−Γ^Rη′∣Ψη′(t)⟩,(19)\frac{d}{dt}|\Psi^{\eta'}(t)\rangle=-\hat{\Gamma}^{\eta'}_R|\Psi^{\eta'}(t)\rangle, \tag*{(19)}

where

Γ^Rη′=∣Ψη′(t)⟩κF^⟨Ψη′(t)∣,(20)\hat{\Gamma}^{\eta'}_R=|\Psi^{\eta'}(t)\rangle\kappa\hat{F}\langle\Psi^{\eta'}(t)|, \tag*{(20)}

so that we obtain the following set of coupled equations

ddt∣Ψel(t)⟩=−(⟨Ψhl(t)∣κF^∣Ψel(t)⟩)∣Ψhl(t)⟩(21)\frac{d}{dt}|\Psi^{\mathrm{el}}(t)\rangle=-\left(\langle\Psi^{\mathrm{hl}}(t)|\kappa\hat{F}|\Psi^{\mathrm{el}}(t)\rangle\right)|\Psi^{\mathrm{hl}}(t)\rangle \tag*{(21)}
ddt∣Ψhl(t)⟩=−(⟨Ψel(t)∣κF^∣Ψhl(t)⟩)∣Ψel(t)⟩.(22)\frac{d}{dt}|\Psi^{\mathrm{hl}}(t)\rangle=-\left(\langle\Psi^{\mathrm{el}}(t)|\kappa\hat{F}|\Psi^{\mathrm{hl}}(t)\rangle\right)|\Psi^{\mathrm{el}}(t)\rangle. \tag*{(22)}

For the sake of simplifying the argument we are assuming here that the electron and hole states are described by wavefunctions. However, in our treatment the electron and hole states are described by reduced density matrices because of the system-bath interaction. So, let us consider the matrix element ⟨Ψhl(t)∣κF^∣Ψel(t)⟩\langle\Psi^{hl}(t)|\kappa\hat{F}|\Psi^{el}(t)\rangle that gives rise to the recombination of electrons due to the operator

κF^=κ∑iFii∣i⟩⟨i∣+κ∑i≠jFij∣i⟩⟨j∣.(23)\kappa\hat{F} = \kappa\sum_i F_{ii}|i\rangle\langle i| + \kappa\sum_{i\ne j} F_{ij}|i\rangle\langle j| . \tag*{(23)}

Substituting Eq. (23) into the previous matrix element and writing the electron and hole wavepackets in the molecular site basis, ∣Ψη(t)⟩=∑iCiη(t)∣i⟩|\Psi^\eta(t)\rangle= \sum_i C_i^\eta(t)|i\rangle, we get

⟨Ψhl(t)∣κF^∣Ψel(t)⟩=κ∑i[Cihl(t)]∗FiiCiel(t)+κ∑i≠j[Cihl(t)]∗FijCjel(t)(24)\langle\Psi^{hl}(t)|\kappa\hat{F}|\Psi^{el}(t)\rangle= \kappa\sum_i [C_i^{hl}(t)]^* F_{ii} C_i^{el}(t) + \kappa\sum_{i\ne j} [C_i^{hl}(t)]^* F_{ij} C_j^{el}(t) \tag*{(24)}
=κ∑i[Cihl(t)]∗Ciel(t)+κ∑i(∑j(j≠i)[Cihl(t)]∗FijCjel(t)),(25)= \kappa\sum_i [C_i^{hl}(t)]^* C_i^{el}(t) + \kappa\sum_i \left( \sum_{j}^{(j\ne i)} [C_i^{hl}(t)]^* F_{ij} C_j^{el}(t) \right) , \tag*{(25)}

where Fii=1F_{ii}=1 as given by Eq. (5) of the paper. This expression can be compared with the three-state model. The first term describes the radiative recombination taking place on molecular site ii, and κ\kappa can be associated with the radiative decay constant, since it does not depend on temperature or vibrational motion. The second term couples distinct molecular sites. For instance, a hole in the donor site ii is coupled to electrons in both donor and acceptor sites via FijF_{ij}, which is the vibronic coupling term that depends on the temperature. This term is a generalization of the LE-CT hybridization considered in the three-state model. However, the choice of a single recombination parameter κ\kappa, associated with the el-hl lifetime, for both contributions is clearly an oversimplification. Previous studies have treated κ\kappa differently for radiative and non-radiative recombination processes. For instance, Coropceanu and collaborators[10, 14] associate κr\kappa_r with the Einstein spontaneous emission rate for the radiative recombination term whereas the Marcus-Levich-Jortner (MLJ) formula was used to define the non-radiative decay rate κnr\kappa_{nr}. The interplay of charge transfer and non-radiative decay was investigated by atomistic non-adiabatic molecular dynamics simulations in dye-sensitized semiconductor interfaces,[53] where it was revealed a rate competition between the two mechanisms.

Finally, it is worth noting that Eq. (25) describes the recombination losses during the exciton diffusion and during the life-time of the CT state at the D:A interface.

Supporting Simulation Results

Electron-Hole Interaction and Recombination

Vibrational Dynamics and Recombination

Charge Generation with and without Recombination

By comparing the LHS (no recombination) with the RHS (with recombination) panels of Figure S4, we observe that both the charge separation and energy conversion efficiencies decrease as a consequence of recombination for all values of the driving force EDAE_{DA}. For each case, the photovoltaic energy conversion decreased by: 19% (ΔErec≈190\Delta E^{rec} \approx190 meV) for EDA=0E_{DA} = 0 (black), 12% (ΔErec≈120\Delta E^{rec} \approx120 meV) for EDA=0.1E_{DA} = 0.1 eV (red), 16% (ΔErec≈140\Delta E^{rec} \approx140 meV) for EDA=0.2E_{DA} = 0.2 eV (green), 21% (ΔErec≈160\Delta E^{rec} \approx160 meV) for EDA=0.3E_{DA} = 0.3 eV (blue) and 24% (ΔErec≈150\Delta E^{rec} \approx150 meV) for EDA=0.4E_{DA} = 0.4 eV (magenta). The results show that for the activationless regime (EDA=100E_{DA} = 100 meV ≈λ\approx\lambda) the losses due to recombination are minimized, since charge separation is more efficient in this regime.[25] As the driving force increases, for EDA=0.2E_{DA} = 0.2, 0.3 and 0.4 eV, and the system enters the Marcus inverted regime for charge separation, the losses due to recombination increase. We ascribe this effect to reduced charge separation, as the driving force exceeds to the reorganization energy at the interface. Here the combination of big energetic driving forces and increased recombination renders the overall energy loss for EDA=0.4E_{DA} = 0.4 eV the highest of all cases, followed by EDA=0.3E_{DA} = 0.3 and 0.20.2 eV, respectively. However, the biggest energy loss due exclusively to recombination (in absolute value) occurs for EDA=0E_{DA} = 0, because of the much slower electron-hole pair separation rate without the energetic driving force. Internal electric fields should improve considerably the power conversion efficiency at vanish driving forces. The simulation results indicate that the condition EDA≈λE_{DA} \approx\lambda has the effect of minimizing the effects of the geminate recombination.

References

  1. [1]T. M. Clarke and J. R. Durrant, Chem. Rev. 110, 6736 (2010).
  2. [2]A. Armin, W. Li, O. J. Sandberg, Z. Xiao, L. Ding, J. Nelson, D. Neher, K. Vandewal, S. Shoaee, T. Wang, H. Ade, T. Heumüller, C. Brabec, and P. Meredith, Adv. Energy Mater. 11, 2003570 (2021).
  3. [3]P. Cheng, G. Li, X. Zhan, and Y. Yang, Nat. Photonics 12, 131 (2018).
  4. [4]M. Hallermann, S. Haneder, and E. Da Como, Appl. Phys. Lett. 93, 290 (2008).
  5. [5]S. M. Menke and R. J. Holmes, Energ. & Environ. Sci. 7, 499 (2014).
  6. [6]K. D. Rosenthal, M. P. Hughes, B. R. Luginbuhl, N. A. Ran, A. Karki, S.-J. Ko, H. Hu, M. Wang, H. Ade, and T.-Q. Nguyen, Adv. Energy Mater. 9, 1901077 (2019).
  7. [7]J. Liu, S. Chen, D. Qian, B. Gautam, G. Yang, J. Zhao, J. Bergqvist, F. Zhang, W. Ma, H. Ade, O. Inganäs, K. Gundogdu, F. Gao, and H. Yan, Nat. Energy 1, 16089 (2016).
  8. [8]J. Hou, O. Inganäs, R. H. Friend, and F. Gao, Nat. Mater. 17, 119 (2018).
  9. [9]L. Perdigón-Toro, H. Zhang, A. Markina, J. Yuan, S. M. Hosseini, C. M. Wolff, G. Zuo, M. Stolterfoht, Y. Zou, F. Gao, D. Andrienko, S. Shoaee, and D. Neher, Adv. Mater. 32, 1906763 (2020).
  10. [10]V. Coropceanu, X.-K. Chen, T. Wang, Z. Zheng, and J.-L. Brédas, Nat. Rev. Mater. 4, 689 (2019).
  11. [11]F. Gao, W. Tress, J. Wang, and O. Inganäs, Phys. Rev. Lett. 114, 128701 (2015).
  12. [12]F. D. Eisner, M. Azzouzi, Z. Fei, X. Hou, T. D. Anthopoulous, T. J. S. Dennis, M. Heeney, and J. Nelson, J. Am. Chem. Soc. 141, 6362 (2019).
  13. [13]A. Classen, C. L. Chochos, L. Lüer, V. G. Gregoriou, J. Wortmann, A. Osvet, K. Forberich, I. McCulloch, T. Heumüller, and C. J. Brabec, Nat. Energy 5, 711 (2020).
  14. [14]X.-K. Chen, D. Qian, Y. Wang, T. Kirchartz, W. Tress, H. Yao, J. Yuan, M. Hülsbeck, M. Zhang, Y. Zou, Y. Sun, Y. Li, J. Hou, O. Inganäs, V. Coropceanu, J.-L. Bredas, and F. Gao, Nat. Energy 6, 799 (2021).
  15. [15]C. C. S. Chan, C. Ma, X. Zou, Z. Xing, G. Zhang, H.-L. Yip, R. A. Taylor, Y. He, K. S. Wong, and P. C. Y. Chow, Adv. Funct. Mater. 31, 2107157 (2021).
  16. [16]J. Benduhn, K. Tvingstedt, F. Piersimoni, S. Ullbrich, Y. Fan, M. Tropiano, K. A. McGarry, O. Zeika, M. K. Riede, C. J. Douglas, S. Barlow, S. R. Marder, D. Neher, D. Spoltore, and K. Vandewal, Nat. Energy 2, 17053 (2017).
  17. [17]D. Qian, Z. Zheng, H. Yao, W. Tress, T. R. Hopper, S. Chen, S. Li, J. Liu, S. Chen, J. Zhang, X.-K. Liu, B. Gao, L. Ouyang, Y. Jin, G. Pozina, I. A. Buyanova, W. M. Chen, O. Inganäs, V. Coropceanu, J.-L. Bredas, H. Yan, J. Hou, F. Zhang, A. A. Bakulin, and F. Gao, Nat. Mater. 17, 703 (2018).
  18. [18]J. Hofinger, C. Putz, F. Mayr, K. Gugujonovic, D. Wielend, and M. C. Scharber, Mater. Adv. 2, 4291 (2021).
  19. [19]X.-K. Chen, V. Coropceanu, and J.-L. Brédas, Nat. Commun. 9, 5295 (2018).
  20. [20]D. C. Coffey, B. W. Larson, A. W. Hains, J. B. Whitaker, N. Kopidakis, O. V. Boltalinа, S. H. Strauss, and G. Rumbles, J. Phys. Chem. C 116, 8916 (2012).
  21. [21]A. J. Ward, A. Ruseckas, M. M. Kareem, B. Ebenhoch, L. A. Serrano, M. Al-Eid, B. Fitzpatrick, V. M. Rotello, G. Cooke, and I. D. W. Samuel, Adv. Mat. 27, 2496 (2015).
  22. [22]B. P. Rand, D. P. Burk, and S. R. Forrest, Phys. Rev. B 75, 115327 (2007).
  23. [23]A. G. Redfield, IBM J. Res. Dev. 1, 19 (1957).
  24. [24]A. Redfield, in Advances in Magnetic and Optical Resonance, Vol. 1 (Elsevier, 1965) pp. 1–32.
  25. [25]A. M. Andermann and L. G. C. Rego, J. Chem. Phys. 156, 024104 (2022).
  26. [26]G. Candiotto, A. Torres, K. T. Mazon, and L. G. C. Rego, J. Phys. Chem. C 121, 23276 (2017).
  27. [27]D. Credgington, F. C. Jamieson, B. Walker, T.-Q. Nguyen, and J. R. Durrant, Adv. Mater. 24, 2135 (2012).
  28. [28]R. Haberkorn, Mol. Phys. 32, 1491 (1976).
  29. [29]P. Rebentrost, M. Mohseni, and A. Aspuru-Guzik, The Journal of Physical Chemistry B 113, 9942 (2009), pMID: 19603843.
  30. [30]T. M. Burke, S. Sweetnam, K. Vandewal, and M. D. McGehee, Adv. Energy Mater. 5, 1500123 (2015).
  31. [31]M. H. Lee, J. Aragón, and A. Troisi, J. Phys. Chem. C 119, 14989 (2015).
  32. [32]V. Coropceanu, J. Cornil, D. A. da Silva Filho, Y. Olivier, R. Silbey, and J.-L. Brédas, Chem. Rev. 107, 926 (2007).
  33. [33]E. Mozafari and S. Stafström, J. Chem. Phys. 138, 184104 (2013).
  34. [34]A. A. Kocherzhenko, D. Lee, M. A. Forsuelo, and K. B. Whaley, J. Phys. Chem. C 119, 7590 (2015).
  35. [35]A. De Sio and C. Lienau, Phys. Chem. Chem. Phys. 19, 18813 (2017).
  36. [36]A. Troisi and G. Orlandi, J. Phys. Chem. A 110, 4065 (2006).
  37. [37]K. Vandewal, J. Benduhn, K. S. Schellhammer, T. Vangerven, J. E. Rückert, F. Piersimoni, R. Scholz, O. Zeika, Y. Fan, S. Barlow, D. Neher, S. R. Marder, J. Manca, D. Spoltore, G. Cuniberti, and F. Ortmann, J. Am. Chem. Soc. 139, 1699 (2017).
  38. [38]O. V. Milknenko, H. Azimi, M. Scharber, M. Morana, P. W. M. Blom, and M. A. Loi, Energy Environ. Sci. 5, 6960 (2012).
  39. [39]Y. Firdaus, V. M. Le Corre, S. Karuthedath, W. Liu, A. Markina, W. Huang, S. Chattopadhyay, M. M. Nahid, M. I. Nugraha, Y. Lin, A. Seitkhan, A. Basu, W. Zhang, I. McCulloch, H. Ade, J. Labram, F. Laquai, D. Andrienko, L. J. A. Koster, and T. D. Anthopoulos, Nat. Commun. 11, 5220 (2020).
  40. [40]Y. Dong, H. Cha, J. Zhang, E. Pastor, P. S. Tuladhar, I. McCulloch, J. R. Durrant, and A. A. Bakulin, J. Chem. Phys. 150, 104704 (2019).
  41. [41]T. Unger, S. Wedler, F.-J. Kahle, U. Scherf, H. Bässler, and A. Köhler, J. Phys. Chem. C 121, 22739 (2017).
  42. [42]J. Jortner, J. Chem. Phys. 64, 4860 (1976).
  43. [43]P. F. Barbara, T. J. Meyer, and M. A. Ratner, J. Phys. Chem. 100, 13148 (1996).
  44. [44]S. Chaudhuri, S. Hedström, D. D. Méndez-Hernández, H. P. Hendrickson, K. A. Jung, J. Ho, and V. S. Batista, J. Chem. Theory Comput. 13, 6000 (2017).
  45. [45]C. M. Proctor and T.-Q. Nguyen, Appl. Phys. Lett. 106, 083301 (2015).
  46. [46]D. Rauh, A. Wagenpfahl, C. Deibel, and V. Dyakonov, Appl. Phys. Lett. 98, 133301 (2011).
  47. [48]C. M. Proctor, S. Albrecht, M. Kuik, D. Neher, and T.-Q. Nguyen, Adv. Energy Mater. 4, 1400230 (2014).
  48. [49]D. Egorova, A. Kühl, and W. Domcke, Chem. Phys. 268, 105 (2001).
  49. [50]D. Egorova, M. Thoss, W. Domcke, and H. Wang, J. Chem. Phys. 119, 2761 (2003).
  50. [51]K. Sachtleben, K. T. Mazon, and L. G. C. Rego, Phys. Rev. Lett. 119, 090601 (2017).
  51. [52]H. J. Berendsen, J. v. Postma, W. F. van Gunsteren, A. DiNola, and J. R. Haak, J. Chem. Phys. 81, 3684 (1984).
  52. [53]L. G. C. Rego and G. Bortolini, J. Phys. Chem. C 123, 5692 (2019).

Paper details

vlnc:W_a6fxtn56
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents