Perspective pubs.acs.org/JPCL
High-Resolution Photocurrent Imaging of Bulk Heterojunction Solar Cells Sabyasachi Mukhopadhyay, Anshuman J. Das, and K. S. Narayan* Chemistry and Physics of Materials Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore-560064, India ABSTRACT: Images obtained from photocurrent scanning of organic bulk heterojunction solar cell devices provide a direct measure of correlation of the morphology to the performance parameters. The peripheral photocurrent induced from light coupled to probe tips in the near-field regime of bulk heterojunction layers permits in situ scanning of active solar cells with asymmetric electrodes. We present a methodology involving a combination of atomic force microscopy, near-field optical microscopy, and near-field photocurrent microscopy to decipher the carrier generation and transport regions in the bulk heterojunction layer. The angular Fourier transformation technique is implemented on these images to rationalize the optimum blend concentration in crystalline and amorphous donor systems and provide insights into the role of the bulk heterojunction morphology.
B
like PCDTBT and PCPDTBT with bridging atoms like silicon (Si-PCPDTBT) have been explored over a large spectral range, and in the case of PBDTTT-based BHJs, efficiencies as high as 9.2% have been reported.15−18 These systems typically require larger amounts of PCBM from 50 to 80% in order to achieve higher efficiencies. The effects of thermal annealing and solvent mixtures are also significantly different and system-specific. Morphological studies on these polymers, some of them not as crystalline as P3HT, have indicated that there is only a small change under annealing.19 BHJ morphology in these systems is better tuned via processing conditions by adding solvent additives like alkane thiols or diiodooctane.15,18,20,21 Since morphology is dictated by processing conditions, it has been suggested that the mixed phase represents polymer networks with entrapped acceptor molecules,19 and the photophysical properties as well as the transport processes cannot be understood in terms of a binary phase morphology.22,23 The mixed phase would predominantly be present at the D−A interface, and its extent would depend on the miscibility of the two components in the film.11 Mixed phases have been the theme of numerous recent reports that have employed scattering techniques to determine its existence and role in the functioning of a solar cell. A brief description of popular techniques to study structural information of BHJ-based films is mentioned in the next section. As recent excellent reviews on this topic indicate, these techniques are well-developed and provide rich quantitative assessment of the microstructure.10,24 Morphological Information through Scattering Methods: X-ray and Neutron Scattering. Scattering techniques like grazing incidence wide-angle X-ray scattering25 (GIWAXS) have been successful
ulk heterojunction (BHJ) solar cells rely on a sequence of molecular processes like photogeneration and diffusion of excitons, charge separation, transfer, and collection in order to achieve efficient device performance.1−4 Efficacy of each of the photophysical and transport process appears to be correlated to the morphology of the active BHJ layer.5−7 Each combination of BHJ system that exhibits high-efficiency performance parameters is associated with a characteristic nanomorphology.8−10 Loss processes of geminate and nongeminate recombination events have been shown to be clearly affected by the interface structure and the bulk morphology. The traditional view of assuming the BHJ components to be immiscible with spinodal decomposition driving phase separation resulting in a bicontinuous network has been modified for a large class of donor systems. A prototypical system that has been extensively studied in this category is poly(3-hexylthiophene)/phenyl-C61 butyric acid methyl ester (P3HT/PCBM) blends. More appropriate models have been proposed recently where the role of crystallization and nucleation kinetics in the resulting morphology is included.10−12 These models have taken into account the structural information from X-ray and neutron scattering studies on these systems, which indicate that the P3HT-rich phase forms pure nanocrystals.11 In these models, it is argued that as P3HT crystallizes, PCBM molecules are pushed out of the crystalline regions and form regions that are PCBM-enriched.12 This crystalline phase coexists with the phase consisting of a mixture of amorphous P3HT and PCBM along with a PCBM-rich phase. There have been reports that have questioned the purity of the donor polymer phase as well as the acceptor aggregate phase. The large EQE of P3HT/PCBM has been attributed to the purity of the nanodomains of optimum dimensions.13,14 Though P3HT/PCBM is a widely studied model system, it has limitations in performance that partly arise from the limited range of absorption. Donor−acceptor (D−A) polymers © 2012 American Chemical Society
Received: November 10, 2012 Accepted: December 14, 2012 Published: December 14, 2012 161
dx.doi.org/10.1021/jz3018336 | J. Phys. Chem. Lett. 2013, 4, 161−169
The Journal of Physical Chemistry Letters
Perspective
information about the morphology, interfaces, and domain connectivity. A combination of PISAe, scattering techniques, and electron microscopy should be able to provide a reconstruction of a three-dimensional morphology of the bulk heterojunction film. The prerequisite condition for PISAe is the existence of substantial photocurrent in the electrode perimeter vicinity. We introduce the concepts involved in these measurements initially and then proceed to the imaging methods. Photocurrent in the Electrode Vicinity. Earlier reports from our laboratory have shown a measurable intensity-modulated photocurrent (PC) arising from the peripheral region of the electrode.38 PC decreases progressively as a light spot is scanned away from the overlap region of the electrodes. The origin of PC outside of the overlapping electrode can be attributed to three possible reasons (i) optical effects: scattering at rough interfaces and wave guiding in thin films; (ii) electric field fringe effects at electrode edges; and (iii) carrier diffusion from the generation point that is at a distance from the electrode. The contribution of optical scattering for surface roughness can be estimated by studying the haze parameter, defined as the ratio of the diffused light to the total incident light.39 AFM studies on coverslips and BHJ films reveal an average rms roughness of 10 and 3 nm, respectively. At 532 nm illumination, the computed haze parameter entails that only ∼10% of incident light has diffusive nature. Therefore, incident light mostly gets absorbed or transmitted through the film. Even at normal incidence, light guiding at the polymer−ITO interface is quite negligible. The monotonic nature of PC decay near the electrode edge rules out the optical scattering effects as the dominant factor for the peripheral PC. Peripheral PC is sustained over much larger length scales as compared to the extent of the fringing field (the field strength decays to 10−100 V cm−1 within a distance of 100 nm) is observed for blends with a >80% PC71BM concentration, thereby leading to decreased interconnectivity for the holes in the donor phase and low device efficiency. This trend is utilized to obtain an optimum value for the product of carrier generation, charge separation, and charge transport, which is accessed for a critical acceptor concentration. Beyond the critical acceptor concentration, an enhanced electron mobility results due to a well-connected acceptor network, which creates an unbalanced transport with increased bimolecular recombination in the film.54 Observations from AFM can be rationalized from decay length measurements in amorphous and crystalline blends with varying acceptor concentrations. Large anisotropy in decay length (φ = LhDLeD/(|LhD − LeD|)) is observed for an amorphous system as compared to a crystalline system at the optimum acceptor concentrations. The LeD magnitude varies gradually in the amorphous system, whereas sharp changes are observed in the crystalline system with respect to the acceptor concentration (Table 1). The Si-PCPDTBT/PC71BM film exhibits a
Figure 2. (a,b) Schematic representations of the experimental setup. (c) PC decay profile outside of the Al electrode. (d) Schematic representation of the transmission and near-field PC line profile. Reprinted with permission from ref 56. Copyright (2011) American Chemical Society.
The interpretation from decay profile studies is further substantiated in the T-NSOM and NPC measurements (Figure 2). PC71BM domains in the blend can be identified by comparing the transmission images obtained at −532 and 633 nm illumination.56 Films with low (10−30%) PC71BM concentration showed a homogeneous morphology with PCDTBT as the donor. Phase-separated PC71BM domains corresponding to 20−40 nm length scales in the amorphous polymer matrix spanning over >70 nm were resolved in T-NSOM images obtained from 1:1 D−A ratio films (Figure 3). However, beyond
Table 1. Electron and Hole Decay Lengths from a SingleExponential Fit to the Decay Profiles for PCDTBT/PC71BM and Si-PCPDTBT/PC71BM Blends with Varying PC71BM Concentrationsa PCDTBT/PC71BM PC71BM conc. (%)
LeD (μm)
LhD (μm)
80 75 66 50
60 71 55 44
86 114 125 75
Si-PCPDTBT/PC71BM ϕ
PC71BM conc. (%)
LeD (μm)
LhD (μm)
φ
198 188 98 106
30 40 50 70
67 75 31 37
110 95 45 44
171 356 99 232
a φ is the anisotropy parameter, LhD and LeD are the hole and electron decay lengths, respectively.
strong aggregation feature in the absorption spectra range of 600−850 nm, indicating that polymer chains tends to stack even in the presence of PC71BM.55 This highly crystalline nature of the Si-PCPDTBT improves the charge transport and reduces the probability of formation of a charge-transfer complex and bimolecular recombination rate (kr). The presence of longlived charge-transfer states in the PCDTBT/PC71BM blend is evident from transient measurements, but they are absent in Si-PCPDTBT/PC71BM blends.15 Ratio of the mobility (μh/μe) is found to be a factor of 3−5 lower in the PCDTBT blend as compared to that for the blend with Si-PCPDTBT.16 Si-PCPDTBT devices exhibit a non-Langevin-type recombination with a coefficient (∼kr) of ∼1.2 × 10−12 cm3 s−1, whereas PCDTBT solar cells exhibit Langevin-type recombination with rate of ∼10−10 cm3 s−1.16 The differences in carrier mobility (μh/μe) and recombination rates predominantly direct the anisotropy in the first-order decay lengths for amorphous and crystalline donor polymer blends. The PC decay profiles are analyzed by spreading impedance analysis incorporating a carrier percolation transport parameters, where a morphology-dependent spreading resistance function was utilized to quantify the blend morphology and electrical connectivity of networks.43,53
Figure 3. T-NSOM images of BHJ films with different blend ratios. (a) 1:1 and (b) 1:4 PCDTBT/PC71BM. (c) 3:2 (d) 1:1 SiPCPDTBT/PC71BM films. The scale bar indicates the change in the absorption at the PMT in counts per second (kHz) units.
an 80% PC71BM concentration, phase separation between the components is clearly observable in PCDTBT/PC71BM films. The interconnected PC71BM-rich domains have larger lateral dimensions ∼100−200 nm, which render the blend film more transparent to the probe light (Figure 3a,b). The critical D−A concentration drives PC71BM to aggregate and reduces the fraction of the mixed amorphous region, as seen in the T-NSOM images. Measurements on the Si-PCPDTBT blend depict domains of lateral dimensions of 30−60 nm for
