Spectroscopy: selective excitation and probing

The proposed method to study hole transfer is based on two main pillars that allow distinguishing the hole transfer process from other photoinduced processes. First, PC71BM

29 was selectively photoexcited at an energy lower than the polymer bandgap (fig. 2.1 and S2.1) to ensure as weak polymer excitation as possible. Second, the polymer was frequency-selectively probed for an appearance of holes. The appearance of holes is delayed by the hole transfer time³¹(excitation at the interface) or fullerene exciton diffusion. Exciton diffusion occurs at significantly longer time scales (e.g. 10-100 ps^{42, 43}) than the hole transfer, which is the focus of this Chapter (for the fullerene exciton diffusion, see Chapter 3). Effectively, fullerene is the light-absorbing and hole-donating material, while the polymer is the material used to detect charges via polaron (charge-induced) absorption⁴⁴. The blend composition was varied by changing the volume fractions of polymer and PC71BM in order to address the influence of morphology on the hole transfer time.

0 PC₇₁BM/polymer absorption ratio (right axis, dashed-dotted lines).

For the investigation of P3HT polymers, the excitation wavelength at 680 nm was chosen, where PC71BM absorption contrast is the highest (see the small bump at 680 nm in fig. S2.1). For MDMO-PPV blends the excitation at 630 nm was used, which is close to the

a )

b )

30

maximum contrast in these blends. Excitation at 630 nm also allows comparing our results with earlier studies of MDMO-PPV:PC61BM³¹. Selectivity of PC71BM excitation is an important parameter, which determines the fraction of photons being absorbed by PC₇₁BM.

In order to characterize the selectivity of PC₇₁BM excitation, absorption coefficients of all blends were measured (fig. 2.1). Absorption coefficients were calculated as:

(2.1)

where OD – the optical density of a sample (fig. S2.1) and L is the length of a sample in centimeters (fig. S2.2).

Absorption coefficients increase linearly with an increase of PC₇₁BM fraction for all polymer:PC₇₁BM blends as confirmed by the linear fits. This linear increase was expected because PC₇₁BM has significantly higher absorption coefficient than polymer at the energy just below polymer bandgap and, therefore, the film absorption is mostly determined by PC71BM.

Determination of absorption coefficients has limited degree of accuracy defined mainly by two factors: i) light reflection and scattering effects, which are present in measured absorption, especially for pristine polymers at energy lower than the bandgap (due to a relatively low absorption coefficient as compared to other effects), and ii) accuracy of measured film thickness. Note that the absorption coefficient obtained with a spectrometer for pristine RRe-P3HT is higher than for blends with <2% PC71BM. More careful measurements with a laser at 660 nm wavelength have revealed a deviating OD dependence on PC71BM content in the range of 0-2% PC71BM (fig. S2.1f). This can be attributed to the lower laser light scattering (because of better focusing) as compared to the spectrometer.

Therefore, OD at 680 nm for PC71BM loads of 0-2% was replaced with the measurements captured using the laser light. To compensate the excessive OD obtained by the measurement with the laser light of 660 nm the 0.05 was subtracted from the original measured value to match the OD of the measurements with a standard spectrometer at higher PC71BM loads (> 5%).

The linear fit of absorption coefficients was performed to make a more accurate estimate of the absorption coefficients, which is important for further data analysis.

Absorption coefficients of pristine materials as deduced from the linear fits (fig. 2.1) are listed in Table 2.1. The fitted absorption coefficients of polymers are: RRa-P3HT

31 2.1·10⁴ cm^-1 (at 680 nm), MDMO-PPV 1.17·10⁵ cm^-1(at 630 nm), RRe-P3HT 8.7·10⁴ cm^-1 (at 680 nm) and PC71BM 2.9·10⁶ cm^-1 (at 630 nm). The absorption coefficient of PC71BM at the wavelength of 680 nm is estimated to be equal to 1.55·10⁶ cm^-1 from averaging two fits: a) RRa-P3HT (1.5·10⁶ cm^-1), and b) RRe-P3HT (1.6·10⁶ cm^-1) blend sets.

Table 2.1. Results of linear fits for absorption coefficients.

630 nm 680 nm

MDMO-PPV PC₇₁BM RRa-P3HT RRe-P3HT PC₇₁BM Absorption coeff.

(·10⁵ cm^-1) 1.18±0.2 29±1 0.2±0.1 0.9±0.2 15.5±1

k (PC71BM to polymer

absorption ratio)

25 N/A 74 18 N/A

The ratio of absorption coefficients of PC71BM and polymer will be called excitation contrast k further on (Table 2.1). The absorption contrast of PC71BM to polymer depends on the wavelength because the two materials have different absorption spectra. The ratio of ODs between PC71BM and polymer is presented in the insets of fig. 2.1. The OD ratio gives an idea, which wavelength (or photon energy) is optimal for selective photoexcitation of PC71BM. Note that the PC71BM OD to polymer OD ratio does not represent the actual PC71BM excitation contrast because the ODs of PC71BM and polymer are not normalized to the film thickness.

The pristine polymer has some small absorption even below the bandgap due to the absorption of polymer chains with an extended π-conjugation length, nanocrystals that are significantly larger than average size, the presence of chemical impurities, defect states and charge transfer (CT) states^{45, 46} between energy disordered regions⁴⁷. The presence of extremely long π-conjugated polymer chains, which could absorb the excitation pulses used in this work, can hardly account for the overall observed absorption amplitude (at excitation energy) in the pristine polymers of RRa-P3HT and MDMO-PPV because the energy of excitation pulse is significantly lower than the bandgap. It is much more likely that the presence of chemical impurities, the defect states and CT transitions between energy disordered regions are responsible for the below bandgap absorption of a polymer.

Detection of the charges on the polymer was chosen by probing at ~0.4 eV energy, which is related to charge induced absorption^{44, 48-50} (see discussion in Chapter 4 and fig. 4.8).

32

This energy is spectrally separated from the response of most other electronic transitions, e.g.

excitons^{20, 51}, ground state bleaching and stimulated emission^52-54.

0

Fig. 2.2 The PIA transients of polymer:PC₇₁BM blends with different PC₇₁BM weight fractions (shown at the right-hand side) normalized to a number of absorbed photons: a) RRa-P3HT:PC₇₁BM (green circles) b) MDMO-PPV:PC₇₁BM (red circles) c) P3HT:PC₇₁BM (blue circles). Circles represent data points;

black lines represent global fit explained in section 2.6. All transients are corrected for PC₇₁BM background response (section S2.2.3).

2.3. Dynamics