dispersion–orientation effects of fulleropyrrolidine in zone annealed block-copolymer films toward...

10
Dispersioneorientation effects of fulleropyrrolidine in zone annealed block-copolymer lms toward optimizing OPV interfaces Praveen Pitliya a , Gurpreet Singh b , Jose Chapa b , Alamgir Karim b, ** , Dharmaraj Raghavan a, * a Department of Chemistry, Howard University, 525 College Street, NW, Washington, DC 20059, USA b Department of Polymer Engineering, University of Akron, Akron, OH 44325, USA article info Article history: Received 10 October 2012 Received in revised form 19 December 2012 Accepted 22 December 2012 Available online 5 January 2013 Keywords: Fullerene Block copolymer Orientational effects abstract Introduction of an insulating polystyrene block (BCP) polymethylmethacrylate copolymer (PS-b-PMMA) layer between the poly poly(3,4-ethylenedioxythiophene)/poly(styrene sulfonate) (PEDOT:PSS) and the photoactive layer (fullerene or functionalized fullerene/poly-3-hexylthiophene) has been reported to improve solar cell performance. We explore how the morphological structure of this ordered interfacial BCP layer may be modied with novel synthesized electron accepting fulleropyrrolidine nanoparticles (f- NP), processed via a novel dynamic zone-annealing (ZA) method. N-methyl-2-(4-nitro phenyl) full- eropyrrolidine and N-methyl-2-(4-cyano phenyl) fulleropyrrolidine were synthesized by 1,3-dipolar cycloaddition of azomethineylides to fullerene and characterized by 1 H NMR, 13 C NMR, MALDI-TOFMS, cyclic voltammetry, and thermogravimetry. The newly synthesized f-NPs exhibited higher thermal stability and equivalent electronic properties compared to conventionally used [6,6]-phenyl-C 61-butyric acid methyl ester (PCBM) for photoactive layer. f-NP lled PS-b-PMMA thin lms processed using uni- form oven annealing promoted phase segregation driven aggregation of f-NP located at the defect junction points of the block copolymer lms. In contrast, ZA of f-NP lled BCP lms led to the homo- geneous dispersion of f-NPs within the lms, however the f-NP had a synergistic orientation effect on BCP lms, switching PMMA cylinders from vertical to parallel orientation in the ZA lms. This effect is presumably due to a lowering of thermal conductivity of the BCP lm by nanoparticles that scatter the phonons thereby decreasing the mean free path length of phonon propagation. These results may be important for the self-assembly of thermally stable and interfacially insulating BCP lms for improved solar cell devices. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction There is signicant interest in efcient organic photovoltaics (OPVs), largely due to the ease with which OPVs can be processed at lower costs and can be customized at the molecular level, com- pared to inorganic PVs. OPVs carry a distinct advantage of being mechanically more robust, cheap, lightweight, and potentially easy to dispose off than the conventional Si-based technologies [1,2]. Fundamental advances in the photovoltaic eld require develop- ment of novel design concepts, new nanomaterials and processing strategies, which is the focus of the present paper. Currently, most OPVs are based on a combination of mixture of donor and acceptor materials, which include perylenebisimide de- rivatives, hexabenzocoronenes, phthalocyanines, polythiophenes, alkoxy-poly (2-methoxy-5-(2 0 -ethylhexoxy)-p-phenylene), as well as C60 and higher fullerene derivatives. However, the most widely studied system for OPV application is a blend of regioregular poly(3- hexylthiophene) (rr-P3HT) and [6,6]-phenyl-C61-butyric acid methyl ester (PCBM). Typically, BHJ solar cells based on a P3HT/ PCBM system give efciencies in the range of 3e5%. Despite signif- icant attempts to improve the efciency of the blend system, P3HT/ PCBM based BHJ solar cells still suffer from two major drawbacks: a poorly controlled electron donor (D) P3HT/acceptor (A) PCBM domain size distribution and inherent thermal instability [3]. The phase morphology of donor/acceptor system is also strongly dependent on processing conditions. In this regard, thermal annealing, slow solvent evaporation, co-solvent approach etc. have been investigated to improve the phase separation ability of binary blend in nanometer range (10 nm). Efforts have been made to improve the overall environmental and thermal stability of conjugated polymer and BHJ solar cell so as to address the overall power conversion efciency of the devices. Diffusion of oxygen in PEDOT:PSS oxidizes the photoactive * Corresponding author. Tel.: þ1 202 806 4427; fax: þ1 202 806 5442. ** Corresponding author. E-mail address: [email protected] (D. Raghavan). Contents lists available at SciVerse ScienceDirect Polymer journal homepage: www.elsevier.com/locate/polymer 0032-3861/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.polymer.2012.12.068 Polymer 54 (2013) 1415e1424

Upload: dharmaraj

Post on 08-Dec-2016

214 views

Category:

Documents


2 download

TRANSCRIPT

at SciVerse ScienceDirect

Polymer 54 (2013) 1415e1424

Contents lists available

Polymer

journal homepage: www.elsevier .com/locate/polymer

Dispersioneorientation effects of fulleropyrrolidine in zone annealedblock-copolymer films toward optimizing OPV interfaces

Praveen Pitliya a, Gurpreet Singh b, Jose Chapa b, Alamgir Karim b,**, Dharmaraj Raghavan a,*

aDepartment of Chemistry, Howard University, 525 College Street, NW, Washington, DC 20059, USAbDepartment of Polymer Engineering, University of Akron, Akron, OH 44325, USA

a r t i c l e i n f o

Article history:Received 10 October 2012Received in revised form19 December 2012Accepted 22 December 2012Available online 5 January 2013

Keywords:FullereneBlock copolymerOrientational effects

* Corresponding author. Tel.: þ1 202 806 4427; fax** Corresponding author.

E-mail address: [email protected] (D. Ragha

0032-3861/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.polymer.2012.12.068

a b s t r a c t

Introduction of an insulating polystyrene block (BCP) polymethylmethacrylate copolymer (PS-b-PMMA)layer between the poly poly(3,4-ethylenedioxythiophene)/poly(styrene sulfonate) (PEDOT:PSS) and thephotoactive layer (fullerene or functionalized fullerene/poly-3-hexylthiophene) has been reported toimprove solar cell performance. We explore how the morphological structure of this ordered interfacialBCP layer may be modified with novel synthesized electron accepting fulleropyrrolidine nanoparticles (f-NP), processed via a novel dynamic zone-annealing (ZA) method. N-methyl-2-(4-nitro phenyl) full-eropyrrolidine and N-methyl-2-(4-cyano phenyl) fulleropyrrolidine were synthesized by 1,3-dipolarcycloaddition of azomethineylides to fullerene and characterized by 1H NMR, 13C NMR, MALDI-TOFMS,cyclic voltammetry, and thermogravimetry. The newly synthesized f-NP’s exhibited higher thermalstability and equivalent electronic properties compared to conventionally used [6,6]-phenyl-C 61-butyricacid methyl ester (PCBM) for photoactive layer. f-NP filled PS-b-PMMA thin films processed using uni-form oven annealing promoted phase segregation driven aggregation of f-NP located at the defectjunction points of the block copolymer films. In contrast, ZA of f-NP filled BCP films led to the homo-geneous dispersion of f-NPs within the films, however the f-NP had a synergistic orientation effect onBCP films, switching PMMA cylinders from vertical to parallel orientation in the ZA films. This effect ispresumably due to a lowering of thermal conductivity of the BCP film by nanoparticles that scatter thephonons thereby decreasing the mean free path length of phonon propagation. These results may beimportant for the self-assembly of thermally stable and interfacially insulating BCP films for improvedsolar cell devices.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

There is significant interest in efficient organic photovoltaics(OPVs), largely due to the easewithwhich OPVs can be processed atlower costs and can be customized at the molecular level, com-pared to inorganic PVs. OPVs carry a distinct advantage of beingmechanically more robust, cheap, lightweight, and potentially easyto dispose off than the conventional Si-based technologies [1,2].Fundamental advances in the photovoltaic field require develop-ment of novel design concepts, new nanomaterials and processingstrategies, which is the focus of the present paper.

Currently, most OPVs are based on a combination of mixture ofdonor and acceptor materials, which include perylenebisimide de-rivatives, hexabenzocoronenes, phthalocyanines, polythiophenes,

: þ1 202 806 5442.

van).

All rights reserved.

alkoxy-poly (2-methoxy-5-(20-ethylhexoxy)-p-phenylene), as wellas C60 and higher fullerene derivatives. However, the most widelystudied system for OPV application is a blend of regioregular poly(3-hexylthiophene) (rr-P3HT) and [6,6]-phenyl-C61-butyric acidmethyl ester (PCBM). Typically, BHJ solar cells based on a P3HT/PCBM system give efficiencies in the range of 3e5%. Despite signif-icant attempts to improve the efficiency of the blend system, P3HT/PCBM based BHJ solar cells still suffer from two major drawbacks:a poorly controlled electron donor (D) P3HT/acceptor (A) PCBMdomain size distribution and inherent thermal instability [3]. Thephase morphology of donor/acceptor system is also stronglydependent on processing conditions. In this regard, thermalannealing, slow solvent evaporation, co-solvent approach etc. havebeen investigated to improve the phase separation ability of binaryblend in nanometer range (10 nm).

Efforts have been made to improve the overall environmentaland thermal stability of conjugated polymer and BHJ solar cell so asto address the overall power conversion efficiency of the devices.Diffusion of oxygen in PEDOT:PSS oxidizes the photoactive

P. Pitliya et al. / Polymer 54 (2013) 1415e14241416

polymer, forming poor contact which reduces the fill factor (FF) andthe open circuit voltage (Voc) especially for systems annealed athigh temperature or stored in air for prolonged period [5]. Onemethod often used to improve the mechanical and thermal prop-erties of a specific conducting polymer is to prepare compositeusing selective inorganic oxides such as SiO2, TiO2, and zeolite.Alternatively, the preparation of blends, composites, or copolymersof polythiophene using insulating polymers as a processing aid hasalso shown to improve the cell performance [6]. Ideally, suchcomposite would possess a combination of the outstanding pro-cessability and thermal stability characteristic of the insulatingpolymer and the electrical conductivity and optical properties ofthe conducting polymers. It is also reported that the3-butylpolythiophene (P3BT)/insulating polymer (PS and PMMA)composite bilayer showed much improved conductivity comparedto pure P3BT [7]. It is believed that the P3BT/insulating polymercomposite form continuous network structures of the conductingpolymer in the blend system. This network structure not onlyprovides extremely large interfacial area between conjugated pol-ymer and insulating polymer matrix, but also highly efficientconductive channels throughout the composite [7]. Interestingly,a P3HT/PS composite does not show similar improved conductivitycompared to pure P3HT [7]. Clearly, the role of insulating layer indesign of OPVs deserves further investigation.

Block copolymers (BCP) composed of amorphous (non-conju-gated) polymers have been investigated recently as well for OPVsystems for a variety of functions. In a recent study, Sun et al. [3]used PS-P3HT BCP to serve as an effective interfacial compatibil-izer in PCBM/P3HT blends for achieving nanosize domains, criticalfor high efficiency photovoltaics. The study also highlights theinteraction between PS block and the PCBM molecules as thedriving force to control phase separation in P3HT/PCBM blends [3].It is well known that a compatibilizer can reduce interfacial tensionand suppress coalescence, thereby limiting donor/acceptor domainsizes and improving morphological stability in a blend system.Similarly Yang et al., for instance, synthesized rod-coil block co-polymers, and added them to P3HT/PCBM blends as a surfactant,resulting in a substantial improvement in efficiency [9]. Studies byChoi et al. has shown that adding homopolymer PMMA to P3HTePMMA block copolymer significantly improves the self orderingof P3HT phase in BCP, which is highly desirable for obtaining anefficient percolation pathway [8].

Recently, a thin layer of PSePMMA BCP has been studied as aninsulating layer between the active layer and PEDOT:PSS to improvethe overall efficiency of solar cell [5].The thin PSePMMA layerserves to protect the photoactive layer, block the dark current,ensure photocurrent tunneling and improve overall interfacialcontact. Under high temperature annealing conditions, the inter-face between PEDOT:PSS and BHJ can be highly susceptible tomoisture and oxygen [5], the reasons for which require furtherstudy. The insulating layer can protect the active layer by sup-pressing oxygen andmoisture penetration from PEDOT:PSS into thephotoactive layer. For example, Tsai et al. [5] reported that, intro-duction of insulating (PSePMMA) layer between active layer andPEDOT:PSS results in an increase in half-efficiency lifetime ofstandard (PCBM/P3HT) solar cell from 58 h to 97 h. Additionally, theintroduction of insulating PSePMMA interfacial layer between theactive layer and PEDOT:PSS improves the morphology of P3HT/PCBM in the active layer by forming an efficient percolation con-ducting pathway [5]. At present, little is known about thearrangement of PCBM, and more generally, how fullerene deriva-tive molecules organize into the PSePMMA layer through diffusionfrom the active layer and the self orientation of P3HT phase inresponse to the unknown surface morphology of the PCBM infusedBCP layer. In particular, the BCP orientation and the partitioning of

fullerene derivative molecules within an oriented BCP layer isexpected to control the transport and conductivity of the insulatinglayer. The study of fullerene derivatives additives on structure ofblock copolymer thin films by thermally directed assembly is one ofthe key objectives of our study.

Although a few reports [3,12,13] have tried to address thebehavior of PCBM and P3HT in block copolymers, there is still nogeneral understanding of how P3HT or PCBM, a functionalizedfullerene interact with the BCP phases. Literature on the morpho-logical development of nanoparticle loaded block copolymer ofderivatized fullerene chemistry is scarce. We have previouslydemonstrated that a directed assembly method for thin film BCPordering, termed cold zone annealing or CZA, consists of a movingthermal gradient zone with a maximum temperature (Thot) belowthe orderedisorder transition temperature (TODT) of the BCP canlead to preferential alignment of the microstructure. Thesteepness of the temperature gradient has important orienting ef-fects on the BCP structure. A sharp temperature gradient with7T w 45 �C mm�1, termed CZA-S (S ¼ Sharp) produced verticallyoriented BCP thin films [10], while a shallow temperature gradient,7T < 15 �C mm�1 (CZA) created horizontally aligned BCP films [4].These differences are attributed to a competition between an im-pulse driven vertical thermal expansion of the film aligning the BCPnanodomains compared to the horizontal translational motion thatfavor BCP wetting induced layering at slow speeds. Domination ofone over the other produces the different vertical or horizontal BCPorientation. Our paper shows that how small amount (observenevertheless, w1% by polymer mass is a very large number ofnanoparticles) of the fullerene nanoparticles can effect a changefrom vertical to horizontal transition in BCP structure. While thiseffect appears consistent with a homogenization of the thermalgradient by virtue of high thermal conductivity nanoparticles, lit-erature [14] in fact reports a lowering of thermal conductivity withadded nanoparticles, in which case the effect is due to reduced rateof vertical film expansion compared to the horizontal orientingdynamic thermal gradient. Regardless, this is potentially a usefulstrategy for tuning the insulating layer electronic transportproperties.

In this study, we investigate the effect of well-characterizedfulleropyrrolidine nanoparticle (f-NP) derivatives on the surfacemorphology of dynamic CZA-S processed BCP films compared tostatic thermal annealing ones. PSePMMA BCP and N-methyl-2-(4-nitro phenyl) fulleropyrrolidine (P-NO2-NMFP) and N-methyl-2-(4-cyano phenyl) fulleropyrrolidine (P-CN-NMFP) were chosen asmodel systems for this study. Thermally stable fulleropyrrolidineswere synthesized via 1,3-dipolar cycloaddition of in-situ formedazomethineylides to fullerene. Our study clearly shows an inter-esting synergy of effects between f-NP and the CZA-S processingeffects. While the dynamic thermal annealing (CZA-S) drives a ho-mogeneous dispersion of the f-NP in the BCP film (compared tostatic oven annealing where the f-NP aggregate), the f-NP in turnflips the BCP orientation from perpendicular to parallel. Under-standing the correlation between f-NP dispersion and thermalprocessing factors that influence CZA-S will provide a broadenedunderstanding of how to utilize driven block copolymer self-as-sembly to create organizing templates for organizing fullerenederivatized nanoparticles in confined 3-D geometries for advancedorganic photovoltaic applications.

2. Experimental

2.1. Materials

C60 (99.5%) was purchased from Sigma Aldrich, while otherreagents needed for synthesis and characterizationwere purchased

P. Pitliya et al. / Polymer 54 (2013) 1415e1424 1417

from Sigma Aldrich and Fisher Scientific and used without furtherpurification. For column chromatography, silica gel (particle size70e230 mesh) was purchased from Alfa Aesar. Block copolymers ofpoly(styrene-block-methylmethacrylate) (PS-b-PMMA,57 kg mol�1-b-25 kg mol�1; polydispersity of 1.07) purchased fromPolymer Source, Inc. was used as obtained. Toluene and DMF(anhydrous) solvents were used, and tetra n-butyl ammoniumhexafluorophosphate was used without recrystallization as sup-porting electrolyte for CV measurement. PCBM (99.5% purity) wasreceived from Nano-C Inc., and used without further purification.Amorphous quartz substrates were bought from Chemglass, Inc.

2.2. Instrumentation for characterization of synthesized N-methylfulleropyrrolidine derivatives

TLC was performed to identify the appropriate solvent compo-sition needed for column chromatography for effective separationof synthesized compound from unreacted as well as byproducts byeluting the sample through silica gel 60 F254 plates. Subsequently,column chromatography was performed by employing 70e230mesh silica gel (Fluka) and using appropriate solvent conditions.1H NMR and 13C NMR spectra of pure products were recorded onBruker AVANCE 400 spectrometer for 1H NMR (400 MHz) and for13C NMR (100 MHz). Chemical shifts are given in parts per million(d) relative to tetramethylsilane (TMS). For 1H NMR and 13C NMRspectra, peaks are calibrated to 7.26 and 77.16 ppm, respectively. Aminimum of 16 scans was acquired. MALDI mass spectra wereobtained in positive linear mode at 15 kV acceleration on a massspectrometer Reflex�time of flight (Voyager-DE STR Bio-spectrometry), using 2,5 dihydroxybenzoic acid as matrix. Tripli-cate measurements were recorded to obtain reproducible results.

2.3. Synthesis and characterization of N-methyl fulleropyrrolidinederivatives

Reaction procedure was adopted from previously reported lit-erature [15]. A solution of 72 mg (0.1 mmol) of C60, 17.8 mg(0.2 mmol) of N-methyl glycine (sarcosine) and 75.5 mg (0.5 mmol)of 4-nitro benzaldehyde in 70 ml of toluene was stirred at refluxtemperature for 6 h under nitrogen atmosphere. Reaction pro-gression was monitored by TLC. Then the solvent was removed invacuo and residue was purified by column chromatography. Firsta purple color band of unreacted C60 was collected then a browncolor band of product was collected and solvent was removed inrotary evaporator, and finally brown color product N-methyl-2-(4-nitro phenyl) fulleropyrrolidine (P-NO2-NMFP) was dried in vac-uum oven for 6 h. Similar procedure was adopted for synthesis andpurification of compounds (P-CN-NMFP) by replacing 4-nitrobenzaldehyde with 4-cyano benzaldehyde as the starting mate-rial. 1H NMR, 12C NMR, FTIR and MALDI-TOFMS characterizationresults of synthesized compounds (P-NO2-NMFP and P-CN-NMFP)can be found in supporting information.

2.4. Electrochemistry

Cyclic voltammetric experiments were performed on synthe-sized fulleropyrrolidines using a BAS 100 electrochemical analyzer.The electrochemical set-up consisted of a three electrode systemwith platinum as the working electrode, a platinum wire as thecounter electrode, and Ag/Agþ (0.01 M AgNO3, 0.1 M Bu4NPF6 inacetonitrile) as the reference electrode. Tetra n-butyl ammoniumhexafluorophosphate was used as supporting electrolytes withoutrecrystallization. The electrodes were polished and sonicated priorto performing the electrochemical experiments. All potentials werereferenced to the reference electrode. 0.5 mM concentration of

synthesized compounds was prepared by dissolving it in (Tolue-ne:DMF, 3:2) solution. All measurements were done in nitrogenatmosphere and the scan direction was from �ve to þve potentialof �2500 mV to 0 mV. All the measurements were done underambient conditions, using scan rate of 100 mV s�1 and a sensitivityof 10 mA V�1.

2.5. Thermogravimetric analysis

Thermogravimetric analysis was conducted using SEIKO (SeikoInstrument Inc.) Thermal Analyzer SSC/5200 by heating 8e12mg ofsynthesized compounds from room temperature to 100 �C ata heating rate of 10 �C min�1, holding it at 100 �C for 15 min; thenramping it from 100 �C to 550 �C at a rate of 10 �C min�1 and thencooling back to room temperature at a rate of 10 �C min�1 undera nitrogen atmosphere. For comparison purposes, thermogram ofPCBM was also recorded under similar condition.

2.6. PS-b-PMMA/fulleropyrrolidine thin film blends preparation

Thin film samples of diblock copolymer of PS-b-PMMA wereused in these studies as a model copolymer. The total molar mass ofthe block copolymer, polydispersity and mass fraction of PMMAwas 82 kg mol�1, 1.07, and 0.30 respectively. The mass fraction ofPMMA leads to the formation of hexagonally-packed cylindersupon micro-phase separation with cylinder to cylinder spacing,Lo z 50 nm. The glass transition temperature (Tg) for the PS blockand PMMA blocks is 110 �C and 124 �C, respectively. The orderedisorder transition temperature (TODT) for PS-b-PMMA used inthis study, is not accessible, i.e. TODT [ Tg.

Functionalized fullerene derivatives (PCBM, P-NO2-NMFP, andP-CN-NMFP) was first dispersed in toluene and then added toa separately prepared PS-b-PMMA toluene solution. For example,the weight fractions of BCP:PCBM were adjusted so as to achieve20% PCBM loading with respect to the PMMA block of the BCP withthe PS-b-PMMA solution of 3% wt/wt. Additionally, thin films of f-NP loaded PS and PMMA homopolymer films were also preparedunder similar condition for comparative purposes. The thickness offilmwas determined using a thin film interferometer, (F-20 UV ThinFilm Analyzer, Filmetrics, Inc.) with a resolution of 1 nm. All filmswere approx. 250 nm thick, spun cast onto ultra-violet ozone (UVO)cleaned quartz substrates. The residual solvent was removed fromthe films by drying the samples at 60 �C for 24 h under vacuum.

2.7. Sharp cold zone annealing (CZA-S) processing of thin films

We adopted the CZA-S procedure as previously described bySingh et al. and Lovinger et al. [10,11]. Briefly, BCP/f-NP andhomopolymer/f-NP blend films were swept at a translation rate of1 mm s�1 across a sharp gradient temperature profile with a max-imum of temperature, Tmax z 210 �C and temperature gradient,7Tmax z 45 �C mm�1. A schematic of the CZA-S set-up is shown inFig. 1.

The cold blocks were cooled by circulating PDMS oil at �5 �Cusing a chiller system (Julabo F12-ED Refrigerated/Heating Circu-lator). The hot zone is generated by a low resistance (0.025 U cm�1)nickelechrome wire covered with a 3 mm outer diameter ceramicinsulation. The nickelechrome wire is heated by connecting it toa low voltage high current source (Digital DC Power Supply, Model1692, B&K Precision Corp.). For the present experiments the coldblocks are separated from the nickelechrome wire at a distance of3 mm, and the heated nickelechrome wire is adjusted to be within0.3 mm from the substrate. The height of nickelechrome wire isfinely controlled so as to achieve the desired Tmax on the substrate.The temperature gradients thus produced are characterized using

Fig. 1. Pictorial representation of the cold zone annealing (CZA) apparatus. The thinfilm was casted on a quartz substrate, and placed on a moveable arm. The sampletemperature during the run was measured by thermocouple which also presses thesample onto the stage. The sample is moved across a temperature controlled hot block(nickelechrome wire) which is located between two PDMS oil cooled cold block. Thethree blocks assembly (coldehotecold) thus forms an in-plane temperature gradient.The sample movement across the in plane gradient is controlled by computer pro-grammed arm at a prescribed velocity. Gaussian like annealing history is recorded ateach position on the sample with a peak temperature of Tmax.

P. Pitliya et al. / Polymer 54 (2013) 1415e14241418

a thermal IR imaging camera with an accuracy of 0.08 �C (Testo 875Thermal Imager Kit). Additionally, thin films of BCP/f-NP andhomopolymer/f-NP were prepared and annealed by vacuum ovenprocessing. Thin films were oven annealed at 210 �C for 24 h ina Thermo Scientific vacuum oven.

The film topographies were imaged using a Dimension� Icon�

Atomic Force Microscopy (AFM) (Bruker AXS) in the Peak ForceQuantitative Nanomechanical Property mapping mode, followedby digitization of the images (using “Image J (NIH)”) for quantitativestatistical analysis.

3. Results and discussion

3.1. Synthesis and characterization of fulleropyrrolidinenanoparticles

We synthesized N-methyl fulleropyrrolidine analogs via1,3-dipolar cycloaddition of azomethineylides to fullerene. Scheme1 provides an overall reaction scheme for synthesizing

Scheme 1. Representation of reaction procedure used to synthesize fulleropyrrolidine anathineylides in-situ, which in turn reacts with C60 to form fulleropyrrolidine.

fulleropyrrolidine analogs and their structures. In this study, thesubstituents attached at 2 position of N-methyl 3,4-full-eropyrrolidine (NMFP) in fulleropyrrolidine analogs are 4-nitrophenyl and 4-cyano phenyl. The reaction yields desired com-pound along with the formation of several byproducts over pro-longed reaction period. The crude compound was brown in colorwhile the unreacted fullerene was purple in color.

To minimize the formation of multiadduct as a part of endproduct, the reaction was closely monitored by TLC. The reactionwas discontinued (after 6 h) once slight trace of multiadduct wasnoticed on TLC (Rf ¼ 0.74 and 0.50 for compounds P-NO2-NMFP andP-CN-NMFP, respectively). The reaction mixture was elutedthrough a packed silica gel columnwith toluene as eluent. Wewereable to obtain quantifiable yields (53%) of compounds P-NO2-NMFPand P-CN-NMFP.

Fig. 2 is the 1H NMR of P-CN-NMFP collected in (CDCl3:CS2 1:3).The 1H NMR spectrum of the compounds in CDCl3:CS2 (1:3) solventshows two doublets and one singlet which are characteristic fea-tures of the 2-substituted pyrrolidine ring attached to fullerene.Two doublets and one singlet were noticed between d 4.0 and d 5.5for compounds P-NO2-NMFP and P-CN-NMFP, which indicates thatall pyrrolidine ring protons except N-methyl protons are present indifferent chemical environments. N-methyl protonwas observed atd 2.86 and d 2.82 for compounds P-NO2-NMFP and P-CN-NMFP,respectively. In addition, two doublets and one singlet for com-pounds P-NO2-NMFP and P-CN-NMFP appeared at differentchemical shift values which were attributed to difference in func-tionality attached to pyrrolidine ring. In aromatic region, theappearance of a broad peak at d 8.07 and d 7.97 for compounds P-NO2-NMFP and P-CN-NMFP, respectively, is a result of restrictedrotation of phenyl substituent on pyrrolidine ring which has beenwell documented in the literature for other phenyl substituentfulleropyrrolidine derivatives [16]. Restricted rotation is explainedby the close proximity of substituents on the phenyl ring to full-erene cage.

13C NMR spectra were also recorded for both compounds. Wenoticed a number of peaks in the range of around 135e155 ppm,which shows the lack of symmetry typical of asymmetric full-eropyrrolidine system [17]. The signals for sp3 carbon of the pyr-rolidine ring and those at the 6, 6 junction of the C60 cage wereobserved at dw 68e84 and dw 68e70, respectively. For compoundP-NO2-NMFP, sp2 carbon (ortho and meta to nitro group) of phenylring and sp3 carbon of methyl group were observed at d123.95,

logs. Under toluene reflux, N-methyl glycine reacts with aldehyde and form azome-

P. Pitliya et al. / Polymer 54 (2013) 1415e1424 1419

d130.09 and d40.03, respectively, which is in agreement with pre-viously reported literature [15]. Similarly for compound P-CN-NMFP, sp2 carbon (ortho and meta to cyano group) of phenyl ringand sp3 carbon of methyl group were observed at d132.49, d129.97and d 40.01, respectively. Additionally, sp carbon of cyano groupwas also observed at d118.30. The ortho carbon of cyano substitutedphenyl ring was observed to be more downfield (d132.49) thannitro substituted phenyl ring (d123.95), suggesting that this dif-ference may be attributed to inductive and additional anisotropiceffect of cyano substituent attached to phenyl ring of P-CN-NMFP.

Final confirmation of the assigned compound was obtained byperforming MALDI-TOFMS characterization of the synthesizedcompound that wasmixed with 2,5 dihydroxybenzoic acid (matrix)in carbon disulfide. A minimum of three trials was performed toconfirm the reproducibility of the spectra. Quasi-molecular ion(Mþ H) peak was observed at 898.97 and 879.38 for compounds P-NO2-NMFP and P-CN-NMFP respectively by MALDI-TOFMS, con-firming the structure of functionalized fulleropyrrolidine.

3.2. TGA analysis

The thermal stability of the synthesized compound has beendefined by parameters such as the onset decomposition tempera-ture. The onset decomposition temperature (Td) is defined as thetemperature at which 10% mass loss occurs in the compound. Fig. 3shows the thermograms of fulleropyrrolidine derivatives P-NO2-NMFP and P-CN-NMFP under nitrogen atmosphere. It was observedthat themass of P-NO2-NMFP and P-CN-NMFP increased initially by1% at 260 �C and 170 �C, respectively. Similar trend have beenpreviously reported and the explanation put forth is that nitrogen isincluded in the fullerene cage [18]. P-NO2-NMFP start to lose massat 351 �C and losses nearly 10% of its original mass at 550 �C,whereas, P-CN-NMFP start to lose mass at 342 �C and losses nearly

Fig. 2. 1H NMR spectra of N-methyl-2-(4-cyano phenyl) fulleropyrrolidine in (CDCl3:CS2 1:position. Characteristic peaks between 8.1 and 7.8 ppm represent phenyl ring protons at e,

10% of its original mass at 550 �C. On the other hand, PCBM showssmall gradual mass loss from 100 �C to 340 �C. At 359 �C, PCBMshows significant mass loss and nearly 13% of its mass is lost by550 �C. Based on the onset decomposition temperature and massremaining at 550 �C of the three compounds studied, it appears thatthermal stability of P-NO2-NMFP and P-CN-NMFP is slightly betterthan that of PCBM.

3.3. Electrochemical analysis

Electrochemical studies using CV technique were performed soas to evaluate potentials of the different redox active species pro-duced from synthesized compounds. Fig. 4 shows the cyclic vol-tammograms of the 0.5 mM fulleropyrrolidine derivatives at roomtemperature in (Toluene:DMF, 3:2) solution containing TBAPF6(0.1 M) as the supporting electrolyte. Toluene:DMF mixture servedas suitable solvent for all the derivatives and the supporting elec-trolyte. Table 1 summarizes the first, second, and third reductionpotentials of P-NO2-NMFP, and P-CN-NMFP. For comparison pur-poses, the redox potentials of the two parent compounds fullereneC60 and PCBM were also determined and included in Table 1.

We observed fullerene to undergo reversible three-electronreduction indicating the formation of a monoanion, a dianion,and a trianion through reduction of a fullerene moiety. Exper-imentally, we established that the observed reduction potential offullerene closely matches with the reported reduction pattern andpotential in the literature [19,20]. Similar to C60, P-NO2-NMFP andP-CN-NMFP also exhibits three reversible reduction waves in thenegative region. However, the introduction of 2-substituted pyr-rolidine ring on C60 causes a cathodic shift in the first reductionpotential of nearly 120 mV with respect to C60. The shift in thereduction potential can be attributed to the partial disruption of theconjugation in C60 upon functionalization [21]. The 1st reduction

3). Peaks between 4.0 and 5.5 ppm represents pyrrolidine rings protons at (a) and (b)j, f, and i position.

80

85

90

95

100

105

25 100 175 250 325 400 475 550

% W

eig

ht lo

ss

Temperature

P-NO2-NMFP

PCBM

P-CN-NMFP

Fig. 3. TGA spectra of PCBM, P-NO2-NMFP and P-CN-NMFP under nitrogen atmo-sphere at a heating rate of 10 �C min�1.

P. Pitliya et al. / Polymer 54 (2013) 1415e14241420

potential values for compounds P-NO2-NMFP and P-CN-NMFPweresimilar to that of PCBM. This suggests that the electrochemicalproperty of synthesized functionalized fulleropyrrolidine is com-parable to that of PCBM.

Some studies have used LUMO energy level determined from CVdata to measure the electron acceptor capability of the acceptormaterials (fullerene and functionalized fullerene) [22]. Moreimportantly, open circuit voltage (Voc) of the solar cell has beenshown to be sensitive to the LUMO of acceptor materials [22]. Wecalculated LUMO energy level of C60, PCBM, P-NO2-NMFP, and P-CN-NMFP from Eonred value using the following equation:LUMO ¼ �eðEonred þ 4:71ÞeV where the LUMO level is expressed ineV with respect to Ag/Agþ. The LUMO energy level of P-NO2-NMFPand P-CN-NMFP is �3.87 and �3.90 eV based on the onset reduc-tion potential (Ered (onset)) of �0.84 and �0.81 V, respectively.Form the presented data in Table 1, it appears that the LUMO energylevel of fulleropyrrolidines is comparable to that of PCBM.

3.4. Synergistic process-ordering of f-NP loaded PS-b-PMMAcopolymer

To evaluate the morphology of vacuum oven and CZA annealedf-NP filled BCPs, thin films were coated on quartz substrate. Singh

Fig. 4. Cyclic voltammogram of P-NO2-NMFP, P-CN-NMFP, PCBM, and C60 in (Toluene:D100 mV s�1. CV data was recorded at room temperature (25 �C) under nitrogen atmospher

et al. [10] have demonstrated that the BCP morphology is highlysensitive to the in-plane temperature gradient which is stronglydependent on the thermal conductivity of underlying substrate. Inthis regard, low thermal conductivity substrate (quartz, mica etc.)should preserve the sharp thermal gradient whereas high ther-mally conductive substrate (copper, silicon etc.) should yield shal-low thermal gradient. Therefore, we casted BCP films on quartzsubstrate (thermal conductivity of 1.4 Wm�1 k�1) so as to preservethe sharp thermal gradient conditions and to evaluate the ori-entational effects on the morphology of BCPs with and without f-NPs under sharp thermal gradient annealing conditions. Fur-thermore, the surface energy of UVO treated quartz (69 mJ m�2) isgreater than the surface energy of PS and PMMA (43e45 mJ m�2)which should improve the wettability of PS-b-PMMA on thequartz substrate and favor the self ordering of BCP [10].

First, we compared the influence of processing conditions on themorphology of BCP. Fig. 5a and b shows AFM data of neat PS-b-PMMA thin film, 250 nm film thickness, upon conventional vac-uum oven annealing and CZA annealing under sharp thermal gra-dient ‘CZA-S’ respectively. We noticed significant difference inmorphology of vacuum oven processed BCP sample to that of CZAprocessed sample. Previously, BCP morphology dependence on thethermal processing condition has been observed by Singh et al. [10]and Berry et al. [23]. The difference in morphology of vacuum ovenannealed and cold zone annealed PS-b-PMMA thin film is attrib-uted to a competition between relaxation kinetics and wetting ki-netics of respective block copolymer phases under uniform thermalannealing and dynamic thermal gradient annealing conditions.Fig. 5cef shows the morphology of the PS-b-PMMA film blendedwith functionalized fullerenes (1% wt. of fullerenes/wt. of BCP) thatwere either annealed in a conventional vacuum oven (Fig. 5c and e)or CZA-S annealed (Fig. 5d and f). It can be seen that the nano-particle aggregation (average aggregate size w100 nm) is pre-dominantly restricted to vacuum oven annealed nanoparticle/blockcopolymer thin films as seen in Fig. 5c and e. It is anticipated thatnanoparticles will be non-homogenously distributed in a BCP sys-tem, as reported in the literatures since nanoparticles tend to residein preferred domains. Moreover, the segregated location of thenanoparticles is typically dependent on the size of the domainclusters that is governed by thermodynamic enthalpic and entropic

MF, 3:2) containing tetrabutylammonium hexafluorophosphate at the scan rate ofe.

Table 1Reduction potential of fullerene and functionalized fullerene.

Compounds Reduction potential (V)/Ag/Agþ LUMO (eV)

E1/2(1) E1/2(2) E1/2(3)

C60 (Ref. [19]) �0.83 �1.33 �1.91C60 (This work) �0.84 �1.34 �1.91 �4.04PCBM �0.94 �1.43 �2.03 �3.91P-NO2-NMFP �0.95 �1.43 �2.12 �3.87P-CN-NMFP �0.95 �1.42 �2.07 �3.90

P. Pitliya et al. / Polymer 54 (2013) 1415e1424 1421

factors, as well as non-equilibrium kinetic factors and their initialas-cast dispersion state. According to previously reported litera-tures, large particle clusters gain enthalpic interaction by clusteringthat appear to be localized near the center of the preferred domaindue to entropic effects between the preferred chains and particles.If the clustering occurs in as-cast film state, it is often difficult tobreak up or transport the clusters with annealing for the aboveenthalpic gain reason. On the other hand, mobile nanoparticlesmayform clusters as their more thermodynamically favorable state,

Fig. 5. Atomic force microscope images of 250 nm thick PSePMMA (82 kg mol�1) films with(wt./wt.) P-NO2-NMFP nanoparticles (e, f) coated on quartz substrates which were then thermaximum temperature “Tmax” w 210 �C, temperature gradient “PTmax” z 45 �C mm�1

especially if they are trapped in unfavorable polymer domain thatcan limit their mobility. So the final dispersed state of nanoparticlesin BCP thin films can be governed by quite complex set ofparameters.

In our system, AFM clearly shows that particles form clusters(w100 nm), therefore we assumed that our system is also governedby similar phenomenon and we see aggregates residing in PS phasein uniform oven annealing sample. We also performed TEMmeasurement but due to small amount of C60 content (1%) to thatof BCP, a noteworthy contrast of nanoparticles was not observed.The functionalized fullerene nanoparticles agglomerate at the blockcopolymer defect sites, thus relieving stress in the system via chainrelaxation [24]. When identical films were CZA-S annealed, nano-particles dispersion in block copolymer is favored tremendously aswitnessed by the disappearance of nanoparticle agglomerates inFig. 5d and f which we presume is due to enhanced particle dis-persion driving force via the dynamic thermal gradient, but withthe same preference for PS domain, some at the polymer/polymerinterface and PMMA domain without any aggregation. In otherwords, the state of dispersion of nanoparticles is improved with

out nanoparticles (a, b), with 1% (wt./wt.) P-CN-NMFP nanoparticles (c, d) and with 1%mally annealed in a vacuum oven at 210 �C for 24 h (a, c, e) and cold zone annealed atand annealing velocity of 1 mm s�1 (b, d, f).

P. Pitliya et al. / Polymer 54 (2013) 1415e14241422

a sweeping thermal zone i.e. cold zone annealing. This is becausethe localized heating of CZA does not allow for long-range mobilityof the nanoparticles to agglomerate as is possible in uniformthermal oven annealing. This will be reported separately in moredetail in another manuscript, but illustrates our ability to dispersethese synthesized nanoparticles both homogeneously and hetero-geneously in block copolymer films, the target of which is self-assembled organic solar cells using conjugated block copolymerssystems.

Interestingly, BCP film reorients from vertically oriented cylin-ders (Fig. 5b) to horizontally oriented cylinders (Fig. 5d and f)during CZA-S annealing when loaded with functionalized fullerenenanoparticles. We attribute this effect to a recent report where theaddition of low levels of C60 fullerene to PMMA homopolymermatrix in fact reduces the thermal conductivity (s) of the compositefilm substantially, rather than an expected enhancement of thermalconductivity due to high thermal conductivity of the pristinenanoparticles [14]. A lowered BCP film thermal conductivity lowersthe vertical thermal expansion response time when the film en-counters the CZA sharp front. The BCP cylinders align parallel to thesubstrate since a fast asymmetric thermal expansion is required forobtaining vertically aligned BCP cylinders as discussed in Ref. [14].Thermal conductivity varies as s ¼ 1/3 (Cvul), where Cv is theeffective thermal conductivity per unit volume, u is the speed ofphonons, and l is the phonon mean free path. Reduction of s inRef. [14] is primarily attributed to a decrease in the mean free pathlength of phonons due to their scattering by the nanoparticle dis-persion. They observed a 20% reduction of swith 1% addition of C60in PMMA, down from pure PMMA at 0.19 W m�1 K�1 tow0.14 W m�1 K�1, also to be compared to pristine C60’sw0.4 W m�1 K�1 [25].

We have previously observed that the vertical orientation of theblock copolymer is very sensitive to the sharpness of the thermalgradient profile in CZA. A sharp thermal gradient of w45 �C mm�1

compared to a 14 �C mm�1 produced a perpendicular versus par-allel oriented PSePMMA block copolymer respectively. The presentresults suggest that the likely reason for the re-orientation of thesame block copolymer with addition of f-NPs from perpendicular toparallel under sharp thermal gradient ofw45 �Cmm�1 must be dueto the reduced rate of vertical thermal expansion arising from itslowered thermal conductivity. An instantaneous asymmetric nor-mal expansion of a substrate-confined polymeric thin film totemperature is followed by slower isotropic relaxation over severalhours as reported by Beaucage et al. in homopolymer films [10]. Webelieve the reduced thermal conductivity lowers the rapid asym-metric thermal orientation kinetics, so that the slower process,polymer wetting dynamics competes effectively (i.e. the PS block towet the air surface and PMMA to wet the quartz surface), to driveparallel orientation of the BCP cylindrical structures as in shallowCZA gradients. Interestingly, a potentially enhanced in-plane ther-mal conductivity with high thermal conductivity nanoparticlesmay have reduced the thermal gradient magnitude and also haveproduced parallel cylinders, but we believe this is not the casebased on kinetic alignment mechanisms discussed in Ref. [14].

The detailed nature of nanoparticle dispersion in the film anda precise phonon scattering scenario is beyond the scope of thepresent paper as it requires neutron scattering and phonon mod-eling. Nevertheless it is worthy of some discussion. It is very likelythat the non-homogeneous distribution of the nanoparticle in theBCP thin films make the change of thermal conductivity differentfrom even that in the PMMA homopolymer/C60 blended systemwhere more clustering occurs which is clearly observed by AFM. Ahigh dispersion in the PS block domain and lower in the PMMAblock domain would nevertheless result in an overall decrease inthe thermal conductivity of the BCP-NP systems due to increased

phonon scattering. We note that in the homopolymer/C60 system,the effect of altered thermal conductivity is not observable by AFMas any morphological transition whereas in case of BCP whichmicrophase separate upon annealing distinctly reveals the effect ofthermal conductivity by reorientation of vertical cylinder to par-allel. Orientation of BCP cylinders is highly sensitive to the thermalprofile aspects of CZA. It has been reported that thermal conduc-tivity plays a role in reorientation of cylinder by reducing theresponse dynamics of vertical expansion of the BCP film due tonumerous phonon scattering centers. However, there are otherpossibilities which could also affect the BCP morphology that in-cludes (a) potential segregation of nanoparticles to the polymersubstrate interface. According to previously reported literature,substrate located nanoparticles roughness induces perpendicularordering of BCP thin film; (b) compositional change can potentiallylead to swelling of BCP domains that causes reorientation of BCP. Toinclude more possibilities, plasticization effects of C60 has recentlybeen reported where C60 disrupts the molecular packing of poly-mer by plasticization effect in BCP thin film which results in reor-ientation of BCP morphology [26]. According to Green et al. [27],local polymer chain backbone motion in a polymer/nanoparticlecomposite is suppressed relative to neat polymer which results inan increase in Tg. In the melt, dynamics of the polymer segments inthe vicinity of the particle surface are suppressed relative to neatpolymer. The above mentioned phenomena could be playinga collective role in our system.

To better appreciate the large scattering potential for phonons atlow nanoparticles concentrations, we estimated the average dis-tance between dispersed nanoparticles. Assuming the ideal statewhere every nanoparticle is homogeneously dispersed throughoutthe cast or annealed film, in our 0.6 wt% nanoparticle filled BCP filmof thickness 250 nm, the average distance is estimated at onlyw5.2 nm.We assumed a total film volume (L�W� H)¼ 1 inch� 3inch � 250 nm ¼ 4.84 � 10�4 cc and corresponding filmmass ¼ 5.08 � 10�4 g assuming film density of 1.05 g cc�1. Since, 1%by filmmass is f-NP, its mass in film¼ 5.08� 10�6 g, so that numberof moles of f-NP in the film ¼ 5.66 � 10�9 corresponding to total noof f-NPs ¼ 34.1 � 1014. The volume occupied by one f-NPmolecule ¼ Total volume of film/Total no of f-NPs ¼ 142 nm3.Assuming the f-NP’s are located at the centers of cubes with thisvolume, the f-NP to f-NP distance ¼ 5.22 nm. We also made anattempt to estimate the domain size from AFM images afterincorporation of nanoparticle. We did not observe domain swellingupon close examination as it is very difficult to detect the estimatedw1% change of domain size. It is also difficult to imagine a 1%domain size increase will cause reorientation of BCP thin filmmorphology. Macrophase separation of particle is possible if par-ticle loading is high, but in our case loading was only 1%. We nev-ertheless do see aggregates in oven annealing, but the fact thatthese larger aggregates are absent in the CZA samples leads us tobelieve that it can be prevented. Therefore, observed clusters arenot the result of macrophase separation rather possibly due toa partially miscible system. However, developing a phase diagramat these low % level concentration would not be very meaningfulpractically, so we prefer to think of dispersion and clustering effectsinstead.

Next, we conducted morphological studies on cylinder formingPSePMMA diblock copolymers blended with PCBM (Fig. 6). It mustbe noted that unlike previous systems studied, PCBM loading was3 wt% in these systems. PCBM is highly miscible in BCP and theobserved solubility of PCBM in toluene (12 mg ml�1) was found tobe nearly 6e8 times more than P-NO2-NMFP (1.5 mg ml�1) and P-CN-NMFP (2 mg ml�1) and is in general agreement with reportedsolubility data [28]. The better miscibility of PCBM in BCP and tol-uene translates to reasonably good dispersion of nanoparticles even

Fig. 6. Thermal annealed in a vacuum oven at 210 �C for 24 h and cold zone annealed at maximum temperature “Tmax” w 210 �C, temperature gradient DTmax w 45 �C mm�1 andannealing velocity of 1 mm s�1.

P. Pitliya et al. / Polymer 54 (2013) 1415e1424 1423

for the samples that were annealed in a conventional thermalvacuum oven with occasional nanoparticle aggregation sites. Thesize of the PCBM aggregates (w30 nm) in oven annealed BCPsamples is considerably smaller than that of P-NO2-NMFP and P-CN-NMFP (w100 nm) suggesting the subtle role of chemistry offunctionalized fullerene on nanoparticle dispersion in BCP.

As discussed before, for CZA annealed BCP/PCBM film, due todecreased thermal conductivity, CZA-S is not able to yield verticallyoriented cylinders which was observed in BCP system free of full-erene nanoparticles. Complete dispersion of nanoparticles isessential for highly efficient organic photovoltaic devices. Poordispersion or agglomeration of nanoparticles can serve as chargerecombination sites, and thus lead to low efficiencies charge sep-aration and charge transfer. If the dispersed nanoparticles couldform a continuous acceptor path then the composite film will behighly desirable for design of highly efficient OPVs.

Fig. 7. Atomic force microscope images of PS or PMMA films with 1% (wt./wt.) P-CN-NMFPcoated on quartz substrates which were then thermally annealed in a vacuum oven at 210temperature gradient “PTmax z 45 �C mm�1 and annealing velocity of 1 mm s�1 (eeh).

3.5. Morphology of homopolymer blended f-NPs

In next phase of study, we investigated the phase preference offulleropyrrolidine nanoparticle in PS and PMMA homopolymersystem. In order to establish the phase preference of nanoparticlein PSePMMA block co-polymer system, we studied the dispersionof functionalized nanoparticles in homopolymer (PS and PMMA)system via conventional thermal vacuum oven annealing and CZA-S annealing. Fig. 7aeh shows thin film morphology for PS andPMMA blended with the functionalized nanoparticles that werevacuum oven annealed (Fig. 7aed) and CZA-S annealed (Fig. 7eeh). Irrespective of the annealing procedure, more aggregation wasnoticed with PMMA samples and more uniform dispersion ofnanoparticles in PS samples. From these results, it suggests thatthe nanoparticles appear to have more preference to the PS phasein PSePMMA block copolymer system. This is in accordance with

nanoparticles (a, b, e, f) and with 1% (wt./wt.) P-NO2-NMFP nanoparticles (c, d, g, h)�C for 24 h (aed) and cold zone annealed at maximum temperature “Tmax” w 210 �C,Scan dimension is 5 mm � 5 mm.

P. Pitliya et al. / Polymer 54 (2013) 1415e14241424

previous studies on blends of functionalized fullerene (i.e. PCBM)and polystyrene-block-polydimethylsiloxane, where C60 prefer-entially segregates to the PS block [29]. We attribute the morehomogenous dispersion of P-NO2-NMFP, P-CN-NMFP, and PCBM inPS phase compared to PMMA phase to be a direct result of highersolubility, miscibility and hydrophobic e hydrophobic type ofinteraction between polystyrene and fullerene derivatives. Further,studies that include neutron reflectivity and XPS measurementsare needed to confirm this observation. However, our results are ingeneral agreement with recent observation where it was notedthat the interaction between PS block and the PCBM molecules asthe driving force to control phase separation in P3HT/PCBMblends.

In order to further confirm that those aggregates are indeedC60 nanoparticle, XPS was performed of the homopolymer/C60thin film. Due to very close appearance of carbon signal of PS,PMMA and C60, we made an attempt to look for a nitrogen peakwhich was present in only C60 derivative. However, since theC60 derivative (1%) was very low to that of polymer we could notdetect any significant peak for nitrogen in both PS/f-NP andPMMA/f-NP film. However, we could see increase in peak in-tensity associated with carbon in case of PS/C60 derivatives film.These could be attributed to the presence of C60 but it wasdifficult to observe clearly, C60 c1 carbon appear at 283.5 eVwhile PS c1 carbon appear at 284.0 eV.

4. Conclusions

Toward continuously expanding the playfield of electronaccepting nanoparticles for OPV applications beyond PCBMthrough rational synthesis, fulleropyrrolidine derivatives weresynthesized by exploiting Prato reaction. The synthesis of N-methylfulleropyrrolidine analogs with substituents at 2 position wasdemonstrated based on 1,3-dipolar cycloaddition of azomethiney-lides to fullerene. The functionalized fulleropyrrolidines werecharacterized by 1H NMR, 13C NMR, MALDI-TOFMS. Importantly,their redox properties were comparable to PCBM and their thermalproperties were slightly better than PCBM. With regard to OPVsystems with these nanoparticles, we take cue from recent studydemonstrating the use of a well studied block copolymer, PS-b-PMMA as an ultrathin underlayer, (<7 nm) in P3HT:PCBM systemwith notable improvement in conduction pathway and lifetime ofsolar cell. Our AFM results indicate that the dispersion of fullerenederivatives in PSePMMA film is more significantly influenced bythermal processing conditions and less by the chemical structure ofthe nanoparticle. We also find that a dynamic thermal annealing,termed cold zone annealing (CZA), favors the homogenous dis-persion of nanoparticle in PS-b-PMMA in all cases, whereas gen-erally vacuum oven annealed samples show large aggregates thatfavor recombination of charges. Evidently the high thermal con-ductivity of f-NP in BCP is not noticed instead it appears phononscattering appears to dominate resulting in parallel orientation ofBCP as opposed to vertical orientation of cylinders in BCP films.Future work will both correlate and optimize the dispersion mor-phology and orientation of cylinders with corresponding OPV de-vice properties.

Acknowledgments

The entire f-NP synthesis, characterization and their orderingwithin block copolymer aspects of the research for OPV studies wassupported by the U.S. Department of Energy, Division of BasicEnergy Sciences under contract No. DE-FG02-10ER4779. Weacknowledge Prof. David Bucknall and O. Bakare for discussionsrelated to the observed photovoltaic properties of the synthesizednanoparticles. Validating static annealing studies of f-NP filled BCPfilms was supported by AFOSR No. #FA9550-12-1-0306.

References

[1] Singh MK. Flexible Photovoltaic Textiles for Smart Application. In: Brabec C,Dyakonov V, Parisi J, Sariciftici NS, editors, New York: Springer; 2003.

[2] Kippelen B, Bredes JL. Energy Environmental Science 2009;3(2):251e61.[3] Sun Z, Xiao K, Keum JK, Yu X, Hong K, Browning J, et al. Advanced Materials

2011;46(23):5529e35.[4] Yager KG, Fredin NJ, Zhang X, Berry BC, Karim A, Jones RL. Soft Matter 2010;

1(6):92e9.[5] Tsai HW, Pei Z, Huang TH, Li PW, Chan YJ. Organic Electronics 2010;11(11):

1796e801.[6] Nicho ME, Garcia-Escobar CH, Arenas MC, Altuzar-Coello P, Cruz-Silva R,

Guizado-Rodriguez M. Materials Science and Engineering B 2011;17(176):1393e400.

[7] Lu G, Tang H, Qu Y, Li L, Yang X. Macromolecules 2007;18(40):6579e84.[8] Choi SY, Lee JU, Lee JW, Lee S, Song YJ, Jo WH, et al. Macromolecules 2011;

7(44):1771e4.[9] Yang C, Lee JK, Heeger AJ, Wudl F. Journal of Material Chemistry 2009;30(19):

5416e23.[10] Singh G, Yager KG, Smilgies DM, Kulkarni MM, Bucknall DG, Karim A. Mac-

romolecules 2012;17(45):7107e17. Beaucage G, Banach M, Vaia RA. Journal ofPolymer Science Part B: Polymer Physics 2000; 38:2929e36.

[11] Lovinger AJ, Chua JO, Gryte CC. Journal of Physics E: Scientific Instruments1976;11(9):927e8.

[12] Yu X, Xiao K, Chen J, Lavrik NV, Hong K, Sumpter BG, et al. ACS Nano 2011;5(5):3559e67.

[13] Chen J, Yu X, Hong K, Messman JM, Pickel DL, Xiao K, et al. Journal of MaterialChemistry 2012;26(22):13013e22.

[14] Salikhov TKh, Tabarov SKh, Rashidov D, Tuichiev Sh, Hussain A. TechnicalPhysics Letters 2009;11(35):1010e1.

[15] Zhang J, Morton JJL, Sambrook MR, Porfyrakis K, Ardavan A, Briggs GAD.Chemical Physics Letters 2006;4e6(432):523e7.

[16] Cruz PDI, Hoz ADL, Font LM, Langa F, Rodriguez MCP. Tetrahedron Letters1998;33(39):6053e6.

[17] Attanasi OA, Mele G, Filippone P, Mazzetto SE, Vasapollo G. ARKIVOC 2009;8:69e84.

[18] Hong LX, Yin CY. Wuhan University Journal of Natural Sciences 2000;3(5):348e50.

[19] Wang N, Lee Y, He X, Gan H, Lee Y, Huang C, et al. Tetrahedron 2006;6(62):1216e22.

[20] Eiermann M, Hicks RG, Knight BW, Neugebauer H, Wudl F. ElectrochemicalSociety Proceeding 2004;22:407e14.

[21] Maggini M, Karlsson A, Scorrano G, Sandona G, Farnia G, Prato M. ChemicalCommunications 1994;5:589e90.

[22] Chen CP, Chan SH, Chao TC, Ting C, Ko BT. Journal of American ChemicalSociety 2008;38(130):12828e33.

[23] Berry BC, Bosse AW, Douglas JF, Jones RL, Karim A. Nano Letters 2007;9(7):2789e94.

[24] Listak J, Bockstaller MR. Macromolecules 2006;17(39):5820e5.[25] Tea NH, Yu RC, Salamon MB, Lorents DC, Malhotra R, Ruoff RS. Applied Physics

1993;A(56):219e25.[26] Sanz A, Ruppel M, Douglas JF, Cabral JT. Journal of Physics: Condensed Matter

2008;20:104209.[27] Kropka JM, Sakai VG, Green PF. Nano Letters 2008;8(4):1061e5.[28] Kim ST, Cho SY, Lee C, Baek NS, Lee KS, Kim TD. Thin Solid Films 2010;2(519):

690e3.[29] Waller JH, Bucknall DG, Register RA, Beckham HW, Leisen J, Campbell K.

Polymer 2009;17(50):4199e204.