Ultrafast hole transfer mediated by polaron pairs in all-polymer photovoltaic blends

The charge separation yield at a bulk heterojunction sets the upper efficiency limit of an organic solar cell. Ultrafast charge transfer processes in polymer/fullerene blends have been intensively studied but much less is known about these processes in all-polymer systems. Here, we show that interfacial charge separation can occur through a polaron pair-derived hole transfer process in all-polymer photovoltaic blends, which is a fundamentally different mechanism compared to the exciton-dominated pathway in the polymer/fullerene blends. By utilizing ultrafast optical measurements, we have clearly identified an ultrafast hole transfer process with a lifetime of about 3 ps mediated by photo-excited polaron pairs which has a markedly high quantum efficiency of about 97%. Spectroscopic data show that excitons act as spectators during the efficient hole transfer process. Our findings suggest an alternative route to improve the efficiency of all-polymer solar devices by manipulating polaron pairs.

and electron (right) polarons. EA represents a transition of polaron to higher level states.
Note that the absorption spectra of hole and electron polarons are similar.          Figure 4e). The P2 transition is the mid-infrared range, exceeding the spectral coverage in our measurements. In this case, we can assign the ESA features at 1.45 eV and 2.4 to 2.6 eV to iPPs. The ESA feature at 1.12 eV is then assigned to EXs.
The spatial separation between the electron and hole for an exciton is much shorter than that for an iPP, so that the recombination rate of an exciton is generally faster than that of an iPP 2 . The signal at 1.12 eV decays faster than the signal probed at 1.45 eV if there is, should be captured by TRFL spectroscopy. In addition, the decay dynamics of EX emission is longer than that characterized by TA spectroscopy, which is caused by the limited temporal resolution of TRFL measurement and different states contributed to TA and fluorescence signals.

Supplementary Note 3. Electron transfer process in J51/N2200
Electron transfer contributes significantly to the device performance. We perform TA experiments to study the dynamics of electron transfer in the J51/N2200 blend film.

Supplementary Note 4. Quantification of the generation yield of iPPs
To quantify the yield of iPP generation, the chemical doping approach 7 has been adopted (Supplementary Methods). The doping induced absorption can be acquired by subtracting the absorption spectrum of the pure N2200 solution from that of the doped-N2200 solution, which is plotted in Supplementary Figure 11. The spectral feature of polaron absorption derived from chemical doping approach show a slight blueshift to that derived from spectro-electrochemistry measurements, which is induced by absorption differences of N2200 in the forms of solution and film. The infrared absorption peak can be assigned to P1 band of the doping induced hole polarons 5,7 which is further to calculate the polaron cross-section. The molar absorption coefficient 1 P ε of positively charged polarons at P1 band in polymer N2200 can be calculated as: The cross section P σ + can be calculated as In the pristine film of N2200, the generation yield is estimated to be ~ 36%.

Supplementary Note 5. Theoretical simulation
The SSH theory is the benchmark for understanding the elementary excitations in To this end, we consider the celebrated SSH model to study the iPPs and EXs in the polymer chains 8,10,11 . The Hamiltonian is written as where E H is the Hamiltonian for the electrons with the form being where j denotes the site index for the thiophene unit and σ for the spin index, j ε the on-site electronic energy, 0 t the hopping integral, α the vibronic coupling strength and j u the displacement of the j-th site. The many-body interaction between electrons is neglected since the separation distance between positive and negative charges in the iPP state is relatively large.
L H is the Hamiltonian of lattice which reads where K is the elastic constant and M is the mass of the thiophene. In our computations, we take 480 sites on the lattice with 240 sites denoting for J51 (j < 0) and the other 240 for N2200 (j > 0). The 160 sites (-80 ≤ j ≤ 80) in the middle of the lattice represent the interfacial regime. The on-site energy j ε is thus set to -7.80eV for j < -80 (the HOMO energy of J51) and -8.34eV for j > 80 (the HOMO energy of N2200), and in between the on-site energy slowly varies following a hyperbolic tangent function. 0 t is calculated to be 0.09eV for both molecules. α and K can be obtained by taking the vibrational spectrum into account. Supplementary Figure 13 displays the spectral densities of HOMO energies for both J51 and N2200. We take 1769 cm -1 as the primary bare optical phonon mode Q ω , so that K can be calculated by The reorganization energy calculated from the spectrum is 0.018eV 12 , so that α is determined to be 3.4 eV/Å as we have calculated that the average The nonadiabatic dynamics simulation method is employed to calculate the dynamics of the EXs and polarons in polymer chains 13,14 . In detail, the temporal evolution of the lattice site is expressed as with η being the damping coefficient. Herein, is the relevant element of the electronic density matrix which is defined as being the wave function of k-th orbit at j-th site whose temporal evolution follows the standard  Figure 14). The equivalent binding energies for EXs and iPPs can be calculated to be in the order of 500 meV and 30 meV in this work, which are comparable to the calculated values in literatures. 15 It is also worth noting that, since the existence of iPP state demands relatively long conjugation length (e.g. 10 nm), the quantum yield of iPP state is however much lower than that of EX. J51 and N2200 do not seem to suffer from this problem so that they can act as high-efficiency materials for charge generation.

Supplementary Note 6. Hole transfer dynamics in different OPV blends.
Hole transfer also exists in the polymer/fullerene blends 16  For TA data recorded from the blend film, we adopted a model with three species. Two independently decaying species represent the iPPs and EXs and a rising component represents the states resulted from iPP-mediated hole transfer.