Article pubs.acs.org/Langmuir

Contribution of Aggregate States and Energetic Disorder to a Squaraine System Targeted for Organic Photovoltaic Devices Chenyu Zheng,†,‡,§ Anirudh Raju Penmetcha,† Brandon Cona,† Susan D. Spencer,§ Bi Zhu,† Patrick Heaphy,† Jeremy A. Cody,† and Christopher J. Collison*,†,‡,§,∥ †

School of Chemistry and Materials Science, Rochester Institute of Technology, Rochester, New York 14623, United States Nanopower Research Laboratory, Rochester Institute of Technology, Rochester, New York 14623, United States § Microsystems Engineering, Rochester Institute of Technology, Rochester, New York 14623, United States ‡

S Supporting Information *

ABSTRACT: Squaraine dyes have significant potential for use in organic photovoltaic devices because their chemical and packing structure tunability leads to a broad solid state panchromaticity. Nevertheless, broadening of the spectrum does not always give rise to increasing power conversion efficiencies. Furthermore, the same processing strategy used to make devices from different squaraines does not lead to the same optimized performance. In this work, by varying the environmental conditions of a set of anilinic squaraines, we demonstrate that spincast thin films are made up of a complex set of states, with each state contributing differently to the overall device efficiency. We demonstrate crystallochromy in that small changes in the packing structure give rise to dramatically different absorption spectra. Through a remarkable comparison between squaraines in poly(methyl methacrylate) solid solution and squaraine:PC60BM blends, we also show long-range and orientational disorder broadening, which distorts the ability to correlate qualitative spectroscopic assessment with an understanding of the device mechanism. We conclude that a full quantitative assessment of the populations of each excited state must be carried out in order to make progress toward an improved understanding of each state’s contribution to charge transfer at the bulk heterojunction interface.

■

mass production.10 The golden synergy is to leverage molecules such as squaraines where small chemical modifications can lead to changes in efficiency,11 while structural properties are retained, allowing synthetic reproducibility, quality control, and solvation in nontoxic ink vehicles for processing success.12 Efficiency optimization first requires a comprehensive understanding of the excited state properties of squaraines, which will change based on molecular structure and solid state packing in pure and blended form, and subsequently a testing of associated OPV devices. Indeed, once the theory for the excited states is completely and precisely determined, squaraines can be used as robust and predictable mechanistic probes for a better, consistent, and fruitful understanding of the universal OPV device.13 Furthermore, once the structure−function connection between excited states and device efficiency has been made, then squaraines can be prescriptively functionalized, or tuned, for processing in nontoxic solvents, ready for mass production.10 To recognize predictable trends in squaraine tunability, we can describe the behavior of each member of a small series,

INTRODUCTION Humanity relies on solar energy for a more sustainable future. Organic photovoltaic (OPV) solar cells offer a technology with the promise of low-cost manufacture when spin-coating can be used for prototypes, leading to roll-to-roll printing or spray coating deposition for commercialization.1−3 OPV devices can be mechanically flexible, allowing application to curved structures and, significantly, through chemical tailoring, devices can be modified to increase the spectral range of overlap with the solar spectrum, for example using near-infra-red (NIR) absorbing polymers.4,5 Nevertheless, polymeric OPV cells currently suffer from lower efficiencies and high initial manufacturing barriers. These are both rooted in the difficulty associated with controlling polymer morphology in the solid state,6−8 a function of polymer molecular weight.9 Thus, there remains a substantial need for new easily purified, consistently processable, low-band-gap molecular materials. Moreover, when such materials are fully characterized (both in morphology and for associated excited states), improved efficiencies will subsequently result from an improved understanding of the mechanism of operation. This will lead to prescriptions for optimal fabrication of OPVs. Our goal is to make more efficient OPV devices from optimized molecular compounds, which will facilitate effective © 2015 American Chemical Society

Received: March 20, 2015 Revised: June 22, 2015 Published: July 1, 2015 7717

DOI: 10.1021/acs.langmuir.5b01045 Langmuir 2015, 31, 7717−7726

Article

Langmuir

Such annealing studies provide insight regarding the phase separation that occurs during annealing of SQ:PC60BM blend films. With such a wide range of predetermined environmental conditions more accurate spectroscopic assignments can be made, important because significant confusion abounds regarding the true assignment of thin film absorbance features for squaraines when using typical approaches17−22 based solely on the aggregation-induced spectral shifts23,24 (red for J, blue for H) described by Kasha.25 Assignment can be misleading considering that spectral shifts can arise from sources other than excitonic coupling.24,26 Our work leads to a more accurate assignment of the peaks, necessary such that theoretical models of all squaraine excited states can be effectively developed and validated.

assigning the spectra as the squaraines interact with each other in a range of different environments spanning dilute liquid solution (completely isolated molecules), concentrated solid solutions (through work with poly(methyl methacrylate) films), nanoparticle dispersions in mixed solvents, and in neat and blended films (analogous to active layers for OPV devices). In this work we assign the spectral features of two squaraines (Figure 1), a dihydroxy squaraine, 2,4-bis(4-dibutylamino-2,6-

■

Figure 1. Squaraine molecules studied in this work.

EXPERIMENTAL SECTION

Basic Characterization. Chloroform purchased from SigmaAldrich was used (as received) to make SQ solutions. Squaraine stock solutions were bath-sonicated before being diluted to desired concentrations. The absorption measurements were taken using a Shimadzu UV-2100PC spectrophotometer. Steady-state fluorescence emission and excitation of SQs in chloroform solutions were measured on a HORIBA FluoroMax fluorometer. The photoluminescence quantum yield (PLQY) was measured in the HORIBA FluoroMax fluorometer by using a Quanta-φ integrating sphere. To prevent saturation of the PMT detector, a 0.25% ND filter was used to scan the very intense scattered excitation peaks. For optical absorption spectroscopy measurements, the solution was made by diluting stock solution in chloroform to a concentration of 5 × 10−6 mol L−1 for both squaraines. For fluorescence spectroscopy measurements, the same solution was used. Two correction files obtained from HORIBA were used: one to correct for the spectral response of a reference detector used to account for fluctuation of the incident light, and the second to account for spectral responsivity of the emission detector. For fluorescence quantum yield measurements, chloroform solutions were further diluted to ∼3 × 10−7 mol L−1 yielding an optical density at the absorption peak of ∼0.1). PLQY errors are based on manufacturer’s specifications of ±10% with our own experimental precision measured as being less than 1% for solutions made with similar concentrations and similar instrument parameters (excitation, emission wavelengths, and slit widths). Spectral characterization was completed on solid-state films, analogous to those used in devices, made with identical processes as follows. ITO-coated glass substrates were cleaned, respectively, in ultrasonic baths of acetone and isopropyl alcohol for 30 min. A PEDOT:PSS solution was spin-casted onto the precleaned substrates at a spin speed of 5000 rpm. Chloroform solutions were made and bath-sonicated, and then were spin-coated at a spin speed of 800 rpm. The concentration of SQ in chloroform was 8 mg/mL. When measuring the absorption of films, a blank substrate (PEDOT:PSS coated onto ITO as described above) was used to baseline the spectrum. The absorbance spectrum of a blank substrate, measured independently, was never shown to have an optical density (OD) above 0.1. The saturation limit for our Shimazu UV-2100PC Spectrophotometer is OD of 5 optical density and the instrument has been demonstrated to keep linear responsivity below 2.5 OD. After the absorption measurements, a thermal annealing approach was applied to each film at different temperatures: 80, 120, and 150 °C. For each temperature, the annealing time was 60 s. The annealing was performed on a hot plate at the measured temperature under nitrogen atmosphere to prevent materials degradation caused by ambient water and O2. The cooling process occurred by allowing the film to re-equilibrate back to room temperature by simply removing it from the hot plate. Solid solutions were made by spin-casting chloroform solutions of poly(methyl methacrylate) and squaraine in appropriate proportions denoted as weight percents, with the assumption that the chloroform

dihydroxyphenyl)cyclobutane-1,3-dione (DBSQ(OH)2) and a corresponding deshydroxy squaraine, 2,4-bis(4-dibutylaminophenyl)cyclobutane-1,3-dione (DBSQ). We show the origins of broadening in films used for OPV devices and we explore the influence of the hydroxyl side group on squaraine packing. Spectral features are assigned for neat and blended films, mixed solvent nanodispersions, and solid solutions, with each of these sample environments introduced below. As-cast blended films are approximated as being homogeneously mixed but absorption data indicate that phase separation is enabled through annealing. Phase separation should be controllable so that bulk heterojunction interface area can be maximized while maintaining effective charge transport through short-path ordered domains. Moreover, a cascade of morphology-specific energy levels over a range of extents of phase separation may aid in reducing geminate recombination of separated charge pairs.7,8 To aid in assigning the spectra of pure and blended films we probe molecular packing through mixed solvent work, noting the formation of kinetically stable and thermodynamically stable aggregates in mixed solvent nanodispersions.14−16 Absorption spectra of such nanodispersions are critical to improved excited state assignments because they will be less broadened relative to film spectra. The reduced broadening occurs because (a) the particle size distribution is expected to be relatively small, (b) the solvent environment is quite stable, (c) interactions between nanoparticles, and hence between aggregates, are insignificant, and (d) the nanoparticles, when formed reversibly, represent the thermodynamic minimum in terms of crystal packing. Finally, studies of squaraines dissolved in solid PMMA films lead us to consider longer-range intermolecular interactions. Even with significant intermolecular separation, an assumed random orientation of squaraine molecules leads to substantial broadening, which we explore. With higher squaraine concentrations, increased absorbance broadening also correlates well with a rapid decrease in fluorescence quantum yield, associated with concentration quenching and the formation of intermolecular states. Significantly, the broadening seen in PMMA solid solutions is apparent in the spectra of bulk heterojunction squaraine films and it is considered that there is a cascade of energies through which the (i) exciton and (ii) free charges may move in a working device. Annealing of PMMA films again leads to phase separation that shows the formation of aggregates with smaller intermolecular separation distances. 7718

DOI: 10.1021/acs.langmuir.5b01045 Langmuir 2015, 31, 7717−7726

Langmuir

■

was fully evaporated during the spin-coating process. The concentration of PMMA in chloroform was fixed at 40 mg/mL, while the concentration of SQ was varied for different weight ratios. Once chloroform was added, the mixture was heated at 40 °C for 20 min to fully dissolve the polymer. The spin-coating was conducted at a spin speed of 3000 RMP for 45 s to obtain films of 1.5 μm. With an assumed density of PMMA of 1.18 g/mL, the average intermolecular distance can be calculated from the reciprocal of the squaraine concentration. Assuming random homogeneous mixing of squaraine molecules in (a) chlroform solution at 1 × 10−5 mol/L concentration and (b) PMMA solid solutions with squaraine composition of 5, 15, and 22.5 wt %, estimated intermolecular distances for DBSQ(OH)2 are (a) 55.0 nm and (b) 2.45, 1.64, and 1.39 nm, respectively. Mixed Solvent. For mixed solvent work, solutions were prepared to a desired volume percent by adding “bad” solvent−deionized water obtained from a Barnstead E-Pure Ultrapure Water Purification System to the respective good solvent−dimethyl sulfoxide (DMSO) purchased from Fisher Chemical. These solvents were used without further purification steps. A series of blank solvent blends was prepared with varying DMSO:H2O ratios. A 0.300-mL aliquot of squaraine stock solution was slowly added over a 10-s period to the appropriate solvent mixture while it was bath-sonicated. Once all squaraine was injected, the solution was allowed a further 10 s in the ultrasonic bath to finish mixing. The sample was then immediately measured for absorption using a Shimadzu UV-2100PC spectrophotometer against a blank solvent mixture of equivalent volume percent. Thus, all absorption measurements were initiated within 60 s of making the solution. No squaraine degradation has been observed due to exposure to water, as tested within 30 min of the making of solutions. Device Making. ITO-coated glass substrates were consecutively cleaned in acetone and iso-propyl alcohol (IPA) ultrasonic baths. PEDOT:PSS solution (Cleavios PH 750, σ = 10−100 S cm−1, diluted 1:1 with deionized water) was spin-cast onto ITO at a spin speed of 5000 rpm. SQ and PC60BM were dissolved together into chloroform to achieve the weight ratio of 1:1 SQ to PC60BM with a concentration of SQ of 8 mg/mL. The active layer solution was then allowed to sonicate for 10 min. The active layer was obtained by spin-casting chloroform solution onto the PEDOT:PSS films at a spin speed of 800 rpm in a N2-filling glovebox. After active layer deposition, the films were placed in a dark vacuum chamber to allow the further evaporation of solvent. The films were then annealed at different temperatures (80, 120, and 150 °C) for 60 s before aluminum deposition. At the final step, aluminum was evaporated under a low pressure (