Ultrafast dynamics of charge carrier photogeneration and geminate recombination in conjugated polymer:fullerene solar cells

We investigate the nature of ultrafast exciton dissociation and carrier generation in acceptor-doped conjugated polymers. Using a combination of two-pulse femtosecond spectroscopy with photocurrent detection, we compare the exciton dissociation and geminate charge recombination dynamics in blends of two conjugated polymers, MeLPPP [methyl-substituted ladder-type poly($p$-phenylene)] and MDMO-PPV [poly(2-methoxy,5-(3,7-dimethyloctyloxy)-1,4-phenylenevinylene], with the electron accepting fullerene derivative PCBM [1-(3-methoxycarbonyl)-propyl-1-phenyl-$(6,6){\mathrm{C}}_{61}$]. This technique allows us to distinguish between free charge carriers and Coulombically bound polaron pairs. Our results highlight the importance of geminate pair recombination in photovoltaic devices, which limits the device performance. The comparison of different materials allows us to address the dependence of geminate recombination on the film morphology directly at the polymer:fullerene interface. We find that in the MeLPPP:PCBM blend exciton dissociation generates Coulombically bound geminate polaron pairs with a high probability for recombination, which explains the low photocurrent yield found in these samples. In contrast, in the highly efficient MDMO-PPV:PCBM blend the electron transfer leads to the formation of free carriers. The anisotropy dynamics of electronic transitions from neutral and charged states indicate that polarons in MDMO-PPV relax to delocalized states in ordered domains within $500\mathrm{fs}$. The results suggest that this relaxation enlarges the distance of carrier separation within the geminate pair, lowering its binding energy and favoring full dissociation. The difference in geminate pair recombination concurs with distinct dissociation dynamics. The electron transfer is preceded by exciton migration towards the PCBM sites. In MeLPPP:PCBM the exciton migration time decays smoothly with increasing PCBM concentration, indicating a trap-free exciton hopping. In MDMO-PPV:PCBM, however, the exciton migration time is found to show a threshold-like dependence on the PCBM concentration. This observation is explained by efficient interchain energy transfer in ordered MDMO-PPV domains in conjunction with exciton trapping.

Figure 1

Scheme of the experimental setup. A laser pulse with a photon energy of $h{\nu }_1=3.1\mathrm{eV}$ excites the sample, followed by a second laser pulse with a tunable photon energy $h{\nu }_2$. Time-integrated changes of the photocurrent induced by the second pulse are measured as a function of time delay $\Delta t$. Additionally, the transmission change $\Delta T∕T$ of the second pulse due to the excitation pulse is measured. The chemical structure of the organic materials used is also shown.

Figure 2

Film spectra of absorption (dashed) and photoluminescence (solid line) of the polymers MeLPPP and MDMO-PPV. The spectral positions of the two applied laser pulses are indicated by arrows.

Figure 3

(Color online) Time-resolved data of MeLPPP:PCBM samples with various PCBM concentrations. The second laser pulse has a photon energy of $h{\nu }_2=2.51\mathrm{eV}$, corresponding to the SE-band in MeLPPP. (a) The differential transmission of the second laser pulse. The inset shows data and fit curves according to Eq. (1) on a shorter timescale. (b) The relative photocurrent change induced by the second laser pulse. The inset shows $\Delta PC∕PC$ of the neat MeLPPP sample and a sample doped with 1% PCBM on a longer time scale. The neat sample reveals a signal at negative time delay while this is absent for the doped sample.

Figure 4

(Color online) Time-resolved data of MeLPPP:PCBM samples with various PCBM concentrations. The second laser pulse has a photon energy of $h{\nu }_2=1.9\mathrm{eV}$, corresponding to the polaron absorption band in MeLPPP. (a) The differential transmission of the second laser pulse. (b) The relative photocurrent change induced by the second laser pulse. The inset shows data and fit curves according to Eq. (2) on a shorter timescale.

Figure 5

(Color online) Time-resolved data of MDMO-PPV:PCBM samples with various PCBM concentrations. The second laser pulse has a photon energy of $h{\nu }_2=1.97\mathrm{eV}$. (a) The differential transmission of the second laser pulse. The inset shows the data with fit curves according to Eq. (1) on a shorter time scale. (b) The relative photocurrent change induced by the second laser pulse.

Figure 6

(Color online) Time-resolved data of MDMO-PPV:PCBM samples with various PCBM concentrations. The second laser pulse has a photon energy of $h{\nu }_2=1.78\mathrm{eV}$. (a) The differential transmission of the second laser pulse. (b) The relative photocurrent change induced by the second laser pulse.

Figure 7

(a) Anisotropy of $\Delta T∕T$ in the neat polymer films, measured in the SE band (MeLPPP: $h{\nu }_2=2.51\mathrm{eV}$, MDMO-PPV: $h{\nu }_2=1.97\mathrm{eV}$). (b) Anisotropy of $\Delta PC∕PC$ in neat polymer films, measured in the polaron absorption band of (MeLPPP: $h{\nu }_2=1.9\mathrm{eV}$, MDMO-PPV: $h{\nu }_2=1.78\mathrm{eV}$).