Temperature-independent carrier formation dynamics in bulk heterojunction

We investigated the effects of temperature on the carrier formation dynamics in a small-molecular blend film, 2,5-di-(2-ethylhexyl)-3,6-bis-(5′′-n-hexy-[2,2′,5′,2′′]terthiophen-5-yl)-pyrrolo[3,4-c]pyrrolo-1,4-dione (SMDPPEH)/[6,6]-phenyl C71-butyric acid methyl ester (PC71BM). We spectroscopically determined the absolute numbers of donor () and acceptor () excitons per absorbed photon as functions of the delay time (t), in addition to the relative number of donor carries (). We found that the carrier formation dynamics is independent of temperature at 300 and 80 K: the carrier formation time (τrise = 0.4 ps) is much faster than the decay time (τdecay ≈ 2.5 ps) of donor excitons. The temperature independence strongly suggests that only excitons created near the donor–acceptor interface contribute to the carrier formation.

O rganic solar cells (OSCs) with bulk heterojunctions (BHJs) 1,2) are promising energy conversion devices with flexibility. They have a low-cost production process, e.g., the roll-to-roll process. The BHJ active layer, which is usually sandwiched between a transparent indium tin oxide (ITO) anode and an Al cathode, consists of phaseseparated nano-size domains of the donor (D) and acceptor (A) materials. In this layer, photo-irradiation creates donor (D + ) and acceptor (A + ) excitons within the respective nano domains. The photo-created excitons are considered to migrate to the D=A interface and separate into electrons and holes. Time-resolved spectroscopy is one of the most powerful tools for clarifying the carrier formation dynamics in BHJ layers. [3][4][5][6][7][8][9][10][11][12][13][14] The analyses of photo-induced absorption (PIA) reveal the relative numbers of excitons (D + and A + ) and donor carriers (D + ) as functions of the delay time (t). For example, the spectroscopy revealed that the carrier formation time (τ rise ≈ 0.2 ps) is comparable to the decay time (τ decay ≈ 0.3 ps) 10) of A + in poly[ [4,8- [15][16][17] The nano-size domain structure of the BHJ layers is advantageous for efficient charge formation and, in turn, high power-conversion efficiency (PCE). The complex domain structure of the BHJ layer, however, impedes the microscopic observation of the charge formation process. For example, scanning transmission X-ray microscopy (STXM) has revealed significant fullerene mixing within donor-rich domains. [18][19][20] In addition, Hedley et al. 21) reported substructures of ∼10 nm within the ∼100 nm domains in PTB7= PC 71 BM blend film. Here, we emphasize that the temperature effect provides significant clues on the charge formation process. For example, Moritomo et al. 22) reported that the carrier formation efficiency (Φ CF ), which is defined as the number of the photo-induced carriers per absorbed photon, is independent of temperature in regioregular poly(3-hexylthiophene) (rr-P3HT)=[6,6]-phenyl C 61 -butyric acid methyl ester (PCBM) and PTB7=PC 71 BM blend films. The independence of temperature strongly suggests that the exciton dissociation should be treated using the quantum-mechanical wave-packet picture, rather than the Marcus theory, 23) in which the charge separation is governed by the displacement of surrounding molecules.
Among the donor materials, the oligothiophene-diketopyrrolopyrrole molecule with ethylhexyl substituents (SMDPPEH) is suitable for a detailed spectroscopic investigation on the charge formation dynamics because it shows intense and sharp PIAs due to D + and D + in the infrared region. 12) In addition, the SMDPPEH=PC 71 BM BHJ solar cell shows a high PCE of 3.0%, reflecting the intense absorption of SMDPPEH for relatively long wavelengths. [24][25][26] In our previous paper, 12) we performed time-resolved spectroscopy in the SMDPPEH=PC 71 BM blend film at 300 K and derived the relative numbers of D + , A + , and D + as functions of t. However, the data points were too scattered to reveal the carrier formation dynamics in detail.
In this paper, we investigated the effects of temperature on the carrier formation dynamics in SMDPPEH=PC 71 BM blend film. By comparing the absolute intensities of the PIAs between the blend and neat films, we determined the absolute numbers of the donor (n D Ã ) and acceptor (n A Ã ) excitons per absorbed photon as functions of t. The improved data acquisition and analysis reveals that the carrier formation time (τ rise = 0.4 ps) is less than the decay times (τ decay ≈ 2.5 ps) of D + , indicating that the late decay component (t ≥ τ rise ) does not contribute to the carrier formation process. The independence of temperature and the low value of τ rise strongly suggest that only the excitons created near the D=A interface contribute to the carrier formation process.
The SMDPPEH=PC 71 BM blend film was spin-coated on quartz substrates using a chlorobenzene solution of SMDPPEH:PC 71 BM of 1 : 1 by weight. Then, the blend film was dried in an inert N 2 atmosphere. SMDPPEH was purchased from Sigma-Aldrich and used as received. For comparison, we prepared spin-coated SMDPPEH (PC 71 BM) films on quartz substrates using chlorobenzene (chloroform) solution. The thicknesses of the SMDPPEH neat, PC 71 BM neat, and SMDPPEH=PC 71 BM blend films were 39, 50, and 96 nm, respectively.
Time-resolved spectroscopy was performed in a pumpprobe configuration at 300 and 80 K, the details of which are described in the literature. 10) The blend film was placed on the cold head of a cryostat, the temperature of which was controlled using liquid nitrogen. As the light source, we employed a regenerative amplified Ti:sapphire laser with a pulse width of 100 fs and repetition rate of 1000 Hz. The 400 nm excitation pulse was generated as second harmonics in a β-BaB 2 O 4 (BBO) crystal. The excitation intensity was 27-36 µJ=cm 2 . The frequency of the pump pulse was decreased to half (500 Hz) to obtain the "pump-on" and "pump-off " conditions. A white probe pulse (800-1600 nm), generated by self-phase modulation in a sapphire plate, was focused on the sample with the pump pulse. The spots of the pump and probe pulses were 5 and 3 mm in diameter, respectively. The transmitted probe spectra were detected using a 256 ch InGaAs photodiode array attached to a 30 cm imaging spectrometer. The spectral data were accumulated for 20000 pulses to improve the signal=noise ratio. The differential absorption spectrum (ΔOD) is expressed as ÁOD À logðI on =I off Þ, where I on and I off are the transmitted light intensity with and without pump excitation, respectively. The time resolution of the system was ∼0.2 ps. Figure 1 shows ΔOD spectra of (a) SMDPPEH and (b) PC 71 BM neat films. In the neat SMDPPEH film [ Fig. 1(a)], the broad absorption band at ≈1100 nm is ascribed to the PIA due to D + . In the neat PC 70 BM film [ Fig. 1(b)], the structureless absorption extending above ∼800 nm is ascribed to the PIA due to A + . We confirmed that the spectral shape remains unchanged even at 10 ps. Figure 2 shows ΔOD spectra of SMDPPEH=PC 71 BM blend films. At 300 K [ Fig. 2(b)], the ΔOD spectra show a broad absorption band, the peak position if which shows a red-shift from ∼1100 nm at 1 ps to ∼1200 nm at 10 ps. The peak position at 1 ps (≈ 1100 nm) suggests that the spectrum contains a considerable D + component. The red-shift disappears above 10 ps, and the spectral profile becomes independent of t. Therefore, we ascribed the absorption band (t ≥ 10 ps) to the PIA due to D + . In fact, the spectral profile of PIA is similar to that of the electrochemical differential absorption spectra of the SMDPPEH neat film. 22) In the early stage (≤10 ps) after photoexcitation, the PIA signal is considered to originate from the weakly bound state of electrons and holes across the D=A boundary. 27) A similar t-dependent behavior of the ΔOD spectra is observed at 80 K [ Fig. 2(c)].
In order to reveal the carrier formation dynamics, we decomposed the PIA (ϕ exp ) of the SMDPPEH=PC 71 BM blend film into the PIA components due to D + ( D þ), D + ( D Ã ), and A + ( A Ã ). We regarded the ΔOD spectra of the SMDPPEH= PC 71 BM blend film (average between 8 to 10 ps), the SMDPPEH neat (at 1 ps) film, and PC 71 BM neat (at 1 ps) films as the basis functions D þ , D Ã , and A Ã , respectively. The spectral weights, i.e., C D þ , C D Ã , and C A Ã , of the respective components were determined so that they minimize the trial function: where λ i denotes the respective wavenumbers. The unit of D þ , D Ã , and A Ã is optical density. F, C D þ , C D Ã , and C A Ã are functions of t. We found that the average process of D þ significantly improves the scattering of C D þ , C D Ã , and C A Ã against t, which enables us to discuss the difference in τ rise of D + and τ decay of D + and A + . Figure 3(a) shows a prototypical example of the spectral decompositions at 300 K. We observed that the 400 nm excitation excites both D + and A + .
To evaluate the absolute numbers of the donor (n D Ã ) and acceptor (n A Ã ) excitons per absorbed photon spectroscopically, we need the absolute intensities of the PIAs per unit density of D + and A + . We assumed that one absorbed photon creates one D + (A + ) in the SMDPPEH (PC 71 BM) neat film. Then, the PIA intensity per unit density of D + (A + ) becomes α exciton = 0.028 (0.002) nm 2 =exciton on considering the ab-  sorption index. Then, n D Ã (n A Ã ) can be calculated by α photon = α exciton , where α photon is the PIA intensity of the D + (A + ) component perunit photon density in the SMDPPEH= PC 71 BM blend film. Note that we should convert the unit of the D + (A + ) component from optical density to nm 2 =photon by considering the excitation pulse energy and absorption index.
In the upper panel of Fig. 4(a), we plotted the obtained n D Ã and n A Ã as functions of t at 300 K. In the lower panel of Fig. 4(a), we plotted the relative number of n D þ . We plotted adjacent averages in n A Ã because n A Ã significantly scatters owing to the small coefficient (α exciton = 0.002 nm 2 =exciton) between the PIA and exciton density. The solid curves are results of least-squares fittings with the exponential function Cð1 À e Àt= rise Þ for n D þ and Ce Àt= decay for n D Ã and n A Ã . In the analysis of n D þ , we use a single exponential function without distinguishing the two process, i.e., D + → D + and A + → D + . The obtained characteristic times (τ rise and τ decay ) and amplitudes (C ) for D + , D + , and A + are listed in Table I. We found that τ rise (= 0.4 ps) of D + is comparable with τ decay (= 0.4 ps) of A + , indicating that the A + → D + conversion process is completed within ≈0.4 ps. We note that τ rise (= 0.4 ps) of D + of the SMDPPEH=PC 71 BM blend film is comparable to that (= 0.2-0.3 ps 10) ) of the PTB7=PC 71 BM blend film. The sub-picosecond τ rise observed in the BHJ layer is ascribed to molecular mixing [18][19][20] as well as the nano-size domain structure. 21) Our careful analysis revealed that τ decay (= 2.6 ps) of D + is much greater than τ rise (= 0.4 ps) of D + . The longer decay time indicates that the late decay component (t ≥ τ rise ) of D + does not contribute to the carrier formation process. In other words, the exciton dissociation efficiency steeply decreases with t. This is probably because the excess energy 28) of excitons, which is indispensable to compensate for the coulombic binding energy between the electron and hole, steeply decreases with exciton migration within the domain. The excitons that reach the D=A interface after the long migration have no excess energy to separate into electrons and holes. Then, only the excitons created near the D=A interface contribute to the carrier formation process. Such a hot exciton picture is theoretically supported. 29,30) Now, let us proceed to the effects of temperature on the carrier formation dynamics. Figure 4(b) shows n D Ã , n A Ã , and n D þ as functions of t at 80 K. The solid curves are results of least-squares fittings with exponential functions. The obtained τ rise , τ decay , and C for D + , D + , and A + are listed in Table I. We emphasize that τ rise (= 0.4 ps) of D + shows no temperature dependence, even though τ decay of A + significantly increased from 0.4 ps at 300 K to 1.0 ps at 80 K. The effect of temperature on τ decay of A + is understood in terms of the thermally activated exciton diffusion. 31) The fast τ decay of A + at 300 K is ascribed to the fast exciton diffusion and resultant additional recombination process at the trap state. The temperature independence of τ rise is well explained if only the excitons created near the interface contribute to the carrier formation process. In this case, τ rise is hardly influenced by temperature because the process is free from activation-type exciton diffusion.
In summary, we spectroscopically clarified the effects of temperature on the carrier formation and exciton decay dynamics in SMDPPEH=PC 71 BM blend film. We found that τ rise (= 0.4 ps) of D + is independent of temperature. The temperature independence suggests that only the excitons created near the D=A boundary contribute to the carrier formation process.  Table I. Characteristic times (τ rise and τ decay ) and amplitudes (C ) of n D þ , n D Ã , and n A Ã . The parameters were obtained through least-squares fittings with the exponential function Cð1 À e Àt=rise Þ for n D þ and Ce Àt=decay for n D Ã and n A Ã . The amplitude of D + has arbitrary units.