Revealing Structure Formation in PCPDTBT by Optical ......Revealing Structure Formation in PCPDTBT by Optical Spectroscopy Christina Scharsich,1,2 Florian S. U. Fischer,3 Kevin Wilma,1,4

Revealing Structure Formation in PCPDTBT by Optical Spectroscopy

Christina Scharsich,1,2 Florian S. U. Fischer,3 Kevin Wilma,1,4 Richard Hildner,1,4

Sabine Ludwigs,3 Anna K€ohler1,2

1Bayreuth Institute of Macromolecular Research (BIMF), University of Bayreuth, Bayreuth 95440, Germany2Experimental Physics II, University of Bayreuth, Bayreuth 95440, Germany3IPOC-Functional Polymers, University of Stuttgart, Stuttgart 70569, Germany4Experimental Physics IV, University of Bayreuth, Bayreuth 95440, Germany

Correspondence to: S. Ludwigs (E-mail: [email protected]) or A. K€ohler (E-mail: [email protected])

Received 5 March 2015; accepted 15 June 2015; published online 29 July 2015

DOI: 10.1002/polb.23780

ABSTRACT: The low-bandgap polymer poly{[4,4-bis(2-ethylhexyl)-

cyclopenta-(2,1-b;3,4-b0)dithiophen]-2,6-diyl-alt-(2,1,3-benzo-

thiadiazole)24,7-diyl} (PCPDTBT) is widely used for organic

solar cell applications. Here, we present a comprehensive

study of the optical properties as a function of temperature

for PCPDTBT in solution and in thin films with two distinct

morphologies. Using absorption and photoluminescence

spectroscopy as well as Franck-Condon analyses, we show

that PCPDTBT in solution undergoes a phase transformation

at 300 K from a disordered to a truly aggregated state on cool-

ing. The saturation value of aggregates in solution is reached

in PCPDTBT thin films at any temperature. In addition, we

demonstrate that the photophysical properties of the aggre-

gates in films are similar to those in solution and that a

low percentage of thermally activated excimer states is pres-

ent in the films at temperatures above 200 K. VC 2015 Wiley

Periodicals, Inc. J. Polym. Sci., Part B: Polym. Phys. 2015, 53,

1416–1430

KEYWORDS: aggregates; low-bandgap polymer; phase transition

INTRODUCTION The tendency of a conjugated polymer mate-rial to aggregate is a key factor in controlling both the chargecarrier mobility as well as the ability of excited states to disso-ciate when combined with an electron-accepting material.1–3

In this way, the aggregation process controls the performanceof a polymer in solar cell applications as well as in thin-fieldtransistors.3–5 The interdependence of device performanceand thin film morphology has been extensively studied in par-ticular for the polymer poly(3-hexylthiophene) (P3HT).2,4,6–8

The conformation of P3HT chains in thin films of neat materi-als or when blended with an electron acceptor such as phenyl-C61-butyric acid methyl ester (PCBM) can be controlled by arange of methods such as thermal annealing, solvent vaporannealing, and spin-coating from solutions containing high-boiling-point solvent additives. The reason why P3HT is soamenable to these processing techniques lies in the semicrys-talline nature of P3HT. Although it adopts a randomly disor-dered chain conformation in a good solvent, planarization andconcomitant crystallization occurs when the solvent quality isdeteriorated.9,10 Similar effects have been reported for otherhomopolymers such as the poly(p-phenylene vinylene) deriva-tive MEH-PPV and for poly(fluorene).11–14

Although solar cells with high efficiencies have beenobtained by optimizing the morphology of P3HT/PCBMblends, the lower energy part of the sun’s irradiation spec-trum is not used by this polymer/acceptor combination. Analternative that is more matched to the incident photon fluxconsists of the use of low-bandgap polymers.15–17 Low-optical gaps can be obtained by combining electron-rich andelectron-deficient moieties to donor–acceptor copolymers.18

Examples for such copolymers that have been used for solarcell applications are cyclopentadithiophene-based polymerssuch as poly{[4,4-bis(2-ethylhexyl)-cyclopenta-(2,1-b;3,4-b’)dithiophen]-2,6-diyl-alt-(2,1,3-benzo-thiadiazole)24,7-diyl}(PCPDTBT), carbazole-based polymers such as PCDTBT,diketopyrrole-based polymers such as PDPP-TPT, thienothio-phene-benzodithiophene-based polymers such as PTB7, andso forth (see Supporting Information for full chemicalnames).19 Like P3HT, the performance of bulk heterojunctionsolar cells made with these polymers depends critically onthe processing conditions and the resulting thin film mor-phology. Consequently, great effort is made to understandand control the bulk heterojunction morphologies, forexample, by thermal annealing, solvent vapor annealing, and

Additional Supporting Information may be found in the online version of this article.

VC 2015 Wiley Periodicals, Inc.

1416 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2015, 53, 1416–1430

FULL PAPER WWW.POLYMERPHYSICS.ORGJOURNAL OF

POLYMER SCIENCE

spin-coating from solutions containing solvent addi-tives.2,20–25 Although the blends used for solar cell applica-tions are studied intensively, less attention is paid to theaggregation properties of the individual blend components.Knowledge of what controls the conformation and assemblyof an individual low-bandgap polymer, however, is a prereq-uisite to understand its behavior in blends.

Here, we have therefore studied the formation of aggregatesin solution and neat films of the widely used low-bandgappolymer PCPDTBT. PCPDTBT reaches power conversion effi-ciencies up to 5.5% in bulk heterojunction cells withPCBM,24,26 yet the properties of structure formation inPCPDTBT themselves are not well understood, and few stud-ies address the properties of the pure polymer.15,16,27–29 Ini-tial structural studies indicated that PCPDTBT does notbecome higher crystalline upon temperature annealing. Ithas been considered as “marginally crystallizable,” quite incontrast to derivatives with the same backbone yet differentside chains that readily form ordered structures.30–32 Subse-quent work, however, demonstrated that PCPDTBT formsrandomly oriented aggregates or even fibrils of about 40 nmsize when spin-coated from high-boiling-point solvents suchas chlorobenzene (CB) or 1-CN.27 When spin-coated fromsolvents containing additives such as 1,8-diiodooctane (DIO)or 1,8-octanedithiol (ODT), the amount of aggregates in thefilm increases as evidenced by changes in the absorptionspectra, yet the structure of the aggregates remains.27,28 Thestructures of the spin-coated films were investigated in theliterature mainly by GIWAXS measurements. A semicrystal-line edge-on morphology with long-range p-stacking (3.8 Å)in-plane and lamellar stacking (11–12 Å) of the alkyl chainsout-of-plane, similar to P3HT, was suggested.29,33,71 Recently,a different crystal structure containing polymer dimers withno evidence for long-range p-stacking was reported forPCPDTBT films prepared by solvent vapor annealing.34,73

Here, we present a detailed study on the process of aggrega-tion in PCPDTBT using steady-state absorption and photolu-minescence spectroscopy as well as time-correlated single-photon counting (TCSPC). First, we investigate PCPDTBT insolution where high degree of conformational freedomallows for facile structure formation. Next, we compare theoptical properties of the polymer in solution with the prop-erties in thin films of two different morphologies wherestructure formation is more constraint. One type of film wasspun using the high-boiling-point additive DIO, and the sec-ond type was solvent-annealed with CB after spin-coatingaccording to Fischer et al.34

The temperature-dependent measurements of PCPDTBT insolution reveal a conformational transformation to occurfrom the disordered state to an ordered, aggregated statewith extended, planarized chains. The critical temperatureof the phase transformation is at about 300 K. Thus,PCPDTBT in solution tends to form preaggregates thataffect the aggregation process of solution-processed thinfilms and devices.

In the photoluminescence of PCPDTBT thin films, a similarplanarized structure as in solution is found at low tempera-tures. At room temperature and at temperatures with effi-ciently extended diffusion lengths, however, relaxation totrap sites takes place. We show that spectral diffusion withinthe aggregates evolves as expected for individual chromo-phores and that the populated trap sites can be understoodto be excimers as typically formed at grain boundaries andinterfaces in the solid state.35 In addition, we show thatthere is a nearly temperature-independent fraction of about40% of aggregates within the films in accordance with thesaturation value of aggregates in solution.

EXPERIMENTAL

Sample PreparationPCPDTBT was purchased from 1-Material and has a molecularweight (Mw) of 23 kDa with a polydispersity of 1.7 (HT-SEC, 1608C, trichlorobenzene, against PS standards). The solvents (p.a.grade) chloroform (CHCl3), CB, and 2-methyltetrahydrofuran(MTHF), as well as the processing additive DIO, were boughtfrom Sigma Aldrich. For temperature-dependent solutionmeasurements, the PCPDTBT was dissolved in the low-melting-point solvent MTHF with a concentration of 0.25 mg/mL.Concentration-dependent solution measurements were donewith PCPDTBT in MTHF. The solutions were stirred at about 508C for at least 15 min and exposed shortly to ultrasonic soundbefore further dilution and measurements.

Film preparation was done under nitrogen atmosphere. Spec-trosil B quartz substrates were cleaned with a CO2 snow jetfollowed by exposure to oxygen plasma (Diener Femto 100W) for at least 300 s. Two different techniques were used. (i)Films were spin-coated onto the quartz substrate at1000 rpm for 240 s from 3 mg/mL solution of CB containing2 wt % of DIO. We refer these films as CB/DIO films. (ii)Films were also prepared by first continuously stirring a3 mg/mL CHCl3 solution for 1–2 h at elevated temperatures(50–60 8C). From this solution, spin-coating took place at1000 rpm for 30 s within 24 h after preparation of the solu-tions. Subsequently, these films were exposed to controlled CBvapor by placing the spin-coated substrates, held at 42 8C, ina chamber with the vapor atmosphere set to 50 8C. The filmswere first swollen to a solution-like state and then recrystal-lized by slowly decreasing the vapor pressure as previouslydescribed.34,36 We refer these films as CB-annealed films. Inaddition, films for Raman measurements were prepared byspin-coating from the 3 mg/mL CHCl3 solution onto a goldsurface that had been evaporated on top of a silicon wafer.

MicroscopyThe atomic force microscopy (AFM) was performed on aDimension Icon AFM from Bruker operated in tapping modeunder ambient air. The polarized microscopy images (POM)were captured with an Axio Imager.A1 from Zeiss usingcrossed polarizers as indicated by arrows in the POM images.

JOURNAL OFPOLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPER

WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2015, 53, 1416–1430 1417

Optical SpectroscopyThe home-built setup for temperature-dependent absorptionand photoluminescence measurements of solution and thinfilms includes a monochromator (CVI Instruments Digikrom240) with Si-photodiode (by Thorlabs) and lock-in technique(Stanford Research Systems, Model SR830 DSP Lock-InAmplifier) for signal detection, a Xe lamp (LAX 1530 byM€uller GmbH Elektronik–Optik) with monochromator (LTIby Amko Light Technology Instruments) for illumination(absorption), and a Compass continuous-wave 405-nm diodelaser by Coherent for excitation (photoluminescence). Duringthe measurement, the sample (the film on the quartz sub-strate or a 1-mm quartz cuvette with the MTHF solution) isheld in a continuous-flow helium cryostate, with the temper-ature controlled by an ITC503 Oxford Instruments tempera-ture controller. To measure the absorption for differentconcentrations, we used a Varian Carry 5000 UV/vis spec-trometer in a standard double-beam transmission alignment.Photoluminescence of the concentration series was measuredin front geometry using a home-built setup that includes aCompass continuous-wave 405-nm diode laser by Coherentfor excitation and an iDus CCD camera from Andor Technol-ogy combined with a MS125 spectrograph by Newport forsignal detection. Spectral correction was determined bymeans of an SESSL-1-5 calibrated spectral flux standard bySphereOptics.

Temperature-dependent absorption of PCPDTBT thin filmswas measured using a Zeiss spectrometer with MCS621VISII detector combined with a HotStage THMS600 by Linkamfor heating. Photoluminescence lifetime was measured ona home-built confocal microscope using TCSPC. The excita-tion light (532 nm, 20 MHz, <80 ps pulse duration;LDH-P-FA-530L; PicoQuant) was directed via a dichroicbeam splitter (z532 RDC, AHF Analysentechnik AG) to aninfinity-corrected oil-immersion objective (603, PlanApo,NA5 1.45; Olympus) and focused onto PCPDTBT films.The photoluminescence was collected by the same objec-tive, passed suitable dielectric long-pass filter, and wasdetected by single-photon sensitive photodiode (MicroPhoton Devices). The Micro Photon Devices’s electricalsignal was used to start a TCSPC module (Timeharp 200;PicoQuant). The quantum efficiency of PCPDTBT in MTHF(c5 0.25 mg/mL) was measured with a FP-8600 fluores-cence spectrometer by JASCO Analytical Instruments. Theexcitation wavelength was 405 nm. For Raman measure-ments, a Jobin Yvon HR800 confocal Raman microscopefrom Horiba with a He–Ne Laser (500:1 polarization,632.82 nm, and 20 mW) was used. Films were measuredin air for 60 s (laser power: 2 lW). The Raman spectraare shown in Supporting Information Figure S1.

Franck-Condon AnalysisFor an appropriate analysis of the spectra, Franck-Condon(FC) fits were carried out by modeling the absorption andphotoluminescence spectrum as a sum of FC transitionsbased on several intramolecular vibrational modes, i. In theharmonic approximation, vibrational frequencies x remain

constant for both ground and excited states. Thus, the inten-sity of individual emission and absorption lines is j deter-mined by the Huang-Rhys parameter S,

S ¼ Mx2�h

DQð Þ2; (1)

where M is the reduced mass of a single oscillator, and DQ isthe configurational displacement between ground andexcited states. The intensities I02m of the vibronic transitions02mi of mode i are related to the Huang-Rhys parameter viaI02mi ¼ e2Si Simi=mi!. We modeled the photoluminescencespectra, PL �hxð Þ, and the absorption spectra, A �hxð Þ, asfollows:37

PL �hxð Þ=n3 � �hxð Þ3 ¼X

mi

Y

i

I02mi Cd½�hx2ð�hx02X

i

mi�hxiÞ�;

(2)

and

A �hxð Þ=n � �hxð Þ ¼X

mi

Y

i

I02mi Cd½�hx2ð�hx01X

i

mi�hxiÞ�;

(3)

where n is the refractive index of the surrounding medium,mi5 0, 1, 2, 3, . . . being the vibration quantum number of theith vibrational mode, �x0 is the energetic position of the 0–0line, and C denotes the Gaussian line shape function with con-stant standard deviation r. The normalization terms n3(�x)3

and n(�x) arise from accounting for the photon density ofstates (DOS) in the surrounding medium effecting the emit-ter’s emission and absorption coefficient, respectively.37–39

RESULTS

PCPDTBT in SolutionEarlier work has shown that PCPDTBT can adopt differentmorphologies.27,34 To investigate the self-assembly processin PCPDTBT, we first concentrate on solutions of the poly-mer. In a solution, the chain experiences less external con-straints than during the process of film preparation, whereinterfaces (substrate–solution and solution–air), shear forces(during spin-coating or swelling), and kinetic aspects influ-ence the self-aggregation process. By cooling the cuvette, thequality of the solvent can be modified gradually from a goodsolvent to a poor solvent without changing the chemicalstructure. Thus, optical spectroscopy in solution is a conven-ient tool to study the self-assembly process in a controlledfashion.9,11,40,41 First, we investigated PCPDTBT in an MTHFsolution with concentration c5 0.25 mg/mL. For bothabsorption and emission spectra, we observe a clear changeof the spectra with temperature [Fig. 1(a)]. At elevated tem-peratures (340 K), the absorption spectrum features a broadand unstructured band peaking at about 1.8 eV. At higherenergies, the second absorption band has a center peak atabout 3.05 eV. On cooling, the first absorption band shiftscontinuously to red. At 300 K, an additional low-energyshoulder appears at about 1.55 eV. This shoulder grows in


POLYMER SCIENCE


intensity with decreasing temperature, forming a sharp maxi-mum at 1.50 eV at 180 K. The spectral range from 1.60 to2.00 eV is examined in more detail in Figure 1(b). It canclearly be seen how the absorption shifts to lower energiesfrom 340 K to 280 K. From 280 K onward, the spectra forman isosbestic point at 1.705 eV. The integrated intensity ofthe first absorption band increases linearly by 23% on cool-ing the solution from 340 K to 180 K. The absorption of thesecond absorption band (2.5–3.2 eV) shifts similarly to thered and increases with decreasing temperature.

These changes in the absorption spectrum—a bathochromicshift, an isosbestic point, and an increase in intensity—aresimilar to those observed for the polymers P3HT and MEH-PPV.9,11,41 Accordingly, we attribute the initial red-shift oncooling to a swelling of the polymer chain that leads toincreased conjugation length and thus lower energy absorp-tion. The isosbestic point is clear evidence for the transfor-mation between two phases. The lower energy of this band,its dominant 0–0 peak, and the associated clear vibrationalstructure indicate a planarized conformation of the lower-energy phase with concomitant longer conjugation length.

In the photoluminescence spectra of the PCPDTBT solution[Fig. 1(c)], we observe a similar evolution. At high tempera-tures, the emission maximum is centered at 1.6 eV, and withdecreasing temperatures, the intensity of this emissiondecreases and a second emission emerges at 1.42 eV domi-nating the photoluminescence spectrum below 260 K. Thesecond emission at lower energies shows higher 0–0 peaksand is more structured than the high energy emission. Weassign the high-energy emission to the PCPDTBT-disorderedphase in which more geometric disorder such as torsions ispresent after excitation than in the planarized phase. In con-trast, the more structured low-energy emission indicatesmore ordered and planarized chains.

To explore whether these spectral changes are associatedmerely with planarized conformations of individual chains orwhether they actually indicate the formation of aggregatedplanar chains, we measured the absorption and photolumi-nescence at room temperature as a function of concentration(Fig. 2). On raising the concentration by factors of

ffiffiffiffiffiffi10p

from0.025 to 0.79 mg/mL, we observe an increase of theshoulder at 1.6 eV in the absorption spectrum, indicatingthat increased concentration assists the formation of thelow-energy band. The same trend is observed in the room-temperature photoluminescence. For the lowest concentra-tion, 0.025 mg/mL, we still observe two features, oneshoulder centered at 1.6 eV and a peak centered at about1.45 eV. With increasing concentration, the high-energyshoulder disappears, and the low-energy peak shifts to evenlower energies. This evolution is similar to that observedwhen lowering the temperature at a fixed concentration.This concentration dependence clarifies that the low-energytransition is associated not only with planarized chains butrather with aggregated chains. Figure 2(c) demonstrates howthe fraction of aggregates formed increases with

FIGURE 1 PCPDTBT in solution (c 5 0.25 mg/mL in MTHF) for

different temperatures between 340 K and 180 K: (a) optical

density as measured, (b) optical density illustrating the iso-

sbestic point, and (c) intensity of the photoluminescence as

measured. [Color figure can be viewed in the online issue,

which is available at wileyonlinelibrary.com.]



http://wileyonlinelibrary.com

concentration. This fraction has been determined from theabsorption spectra as described in detail further below inthe context of Figure 6.

It is possible to separate the measured spectra into the spec-tra arising from the aggregated phase and the disorderedphase by using a thorough FC analysis according to eqs 2and 3 for absorption and photoluminescence, respectively.The multimode FC analysis was based on five modes withthe following energies: 62, 106, 136, 167, and 190 meV. The62-meV mode is assigned to torsional relaxation, whereasother modes are effective vibrational modes taken fromRaman measurements (see Supporting Information). Therefractive index of the surrounding medium was assumed tobe constant over the considered spectral range. Figure 3(a)shows the normalized photoluminescence and absorption ofPCPDTBT in solution measured at 240 K. Figure 3(b,c)shows the normalized underlying spectra of the disorderedand the aggregated phase, respectively, that were obtainedfrom the FC analysis. The FC fits and their comparisonagainst the measured spectra are detailed in the Supporting

FIGURE 2 (a) Absorption of PCPDTBT in MTHF with concentra-

tions of 0.025, 0.079, 0.25, and 0.79 mg/mL, respectively. Arrows

indicate increasing concentration. (b) Photoluminescence of

PCPDTBT in MTHF with concentrations of 0.025, 0.079, 0.25, 0.79,

and 2.5 mg/mL, respectively. (c) Fraction of aggregates in solution

as a function of concentration (PCPDTBT in MTHF) as determined

from absorption spectra by means of Franck-Condon analysis.

[Color figure can be viewed in the online issue, which is available

at wileyonlinelibrary.com.]

FIGURE 3 Normalized photoluminescence and optical density

of PCPDTBT in solution at 240 K: (a) spectra as measured, (b)

spectra of disordered phase, and (c) spectra of aggregated

phase. [Color figure can be viewed in the online issue, which is

available at wileyonlinelibrary.com.]


POLYMER SCIENCE




Information. This decomposition of the spectra into the twophases shows that photoluminescence and absorption of thedisordered phase intersect at 1.66 eV, whereas for the aggre-gated phase, this intersection point is at 1.48 eV. Thus, thespectra associated with two distinct phases are 180 meVapart.

In this manner, we decomposed the absorption spectra inthe temperature range from 320 K to 180 K. The obtainedabsorption spectra for the aggregated PCPDTBT phase areshown in Figure 4(a). The associated photoluminescencespectra [Fig. 4(b)] were found by subtracting the disorderedphase photoluminescence (as measured at 340 K yet normal-ized to match the high-energy shoulder) from the photolumi-nescence spectrum measured at an arbitrary temperature.We see that in absorption, the aggregate spectrum shifts con-tinuously by about 150 meV to lower energies (up to 1.495

eV) and narrows with decreasing temperature. Moreover, theratio of the 0–0 to the 0–1 absorption peak increases. Inphotoluminescence, the spectrum of the aggregates shiftsonly by about 50 meV in a continuous way to the red (1.420eV) and becomes narrower.

The parameters of the FC analysis are summarized in Table1, including the position of the 0–0 transition E0, the stand-ard deviation of the Gaussian linewidth r, the peak ratio ofthe 0–0 and 0–1 transition I0–0/I0–1, and the relative contri-bution of the aggregates to the total spectra. To determinethe peak ratio, the 0–0 intensity was set to unity, and the0–1 intensity was calculated as the sum of all Huang-Rhysparameters of the above-mentioned modes. In Table 1, theFC fits are listed for spectra dominated by the aggregatedphase (180 K), dominated by the disordered phase (340 K),and one spectrum exemplarily containing both aggregatedand disordered phase (240 K) with roughly equal weights inabsorption. The actual spectra and fits are presented in Sup-porting Information Figure S2.

Although for the aggregated phase, the I0–0/I0–1 ratio isabout 1.0, this value is about 0.7 for the disordered phase,indicating a stronger geometric distortion of the excitedstates. The disorder parameter r is around 356 5 meV forthe aggregated phase, yet 556 9 meV for the disorderedphase, with the lower value pertaining to the photolumines-cence spectra. These data support the notion of a higherenergy phase with more conformational disorder and alower energy phase where the chains adopt a more planargeometry so that the excited states are more delocalized andconformationally induced energetic disorder is reduced.From Table 1 and Supporting Information Figure S2, we alsosee that the contribution of the aggregates to the steady-statespectra is much stronger in emission than in absorption forthe spectra taken below 340 K. Obviously, there is energytransfer from the disordered phase to the aggregated phase,implying that at the concentration used (c5 0.25 mg/mL), thetwo phases are reasonably adjacent.

PCPDTBT Thin FilmsKnowing the characteristics of PCPDTBT in solution, weexpanded our studies to PCPDTBT thin films, as films aremost relevant to device applications.24,26,34,42 We used twodifferent preparation conditions. Our analysis focuses onfilms prepared via the device preparation protocol fororganic solar cells, that is, spin-coating from CB (3 mg/mL)with 2 wt % DIO added as solvent additive (CB/DIO film).To explore to which extent the aggregation is sensible to theprocessing conditions, we analyzed a second type of filmwhich was spun from CHCl3 and afterward annealed in CBvapor atmospheres (CB-annealed). The detailed preparationprotocol together with a tentative crystal structure is accord-ing to an earlier publication.34,73

The comparison of absorption spectra of the two PCPDTBTfilms for temperatures between room temperature and about500 K to PCPDTBT in solution is given in Figure 5. Althoughin solution, the low-energy peak appears suddenly from

FIGURE 4 PCPDTBT in solution (c 5 0.25 mg/mL in MTHF): (a)

optical density of aggregated phase as a result of Franck-

Condon analysis of the first absorption band, and (b) photolu-

minescence intensity of the aggregated phase. [Color figure

can be viewed in the online issue, which is available at

wileyonlinelibrary.com.]




300 K onward and there is an isosbestic point resulting from280 K onward, the film spectra show no isosbestic point,and the low-energy peak evolves gradually, yielding peakpositions of 1.56 and 1.67 eV at room temperature for theCB/DIO film and the CB-annealed film, respectively. Further-more, the 0–0 peak is less pronounced in the thin film spec-tra than in the solution spectra.

The absorption spectra of the films at room temperature arecompared with the spectra of the solution at low tempera-ture (240 K) in Figure 5. We see that the solution spectrumand the spectrum belonging to the CB/DIO film coincidewith regard to the energetic position as well as the 0–0/0–1peak ratio. In contrast, the spectrum of the CB-annealed filmis shifted to higher energy (140 meV to the blue), is broader,and shows a high-energy shoulder at about 2.05 eV. Fromthis, we infer that the aggregates in the CB/DIO film may bestructurally similar to the aggregates found in solution oncooling, whereas the aggregates formed in the CB-annealedfilm differ in their structure.

Analogous to the absorption spectra measured in solution, itis possible to separate the thin film absorption spectra into asuperposition of two spectra using FC analysis on the basisof the Raman spectra. This spectral decomposition is shownin Supporting Information Figure S3. Note that FC analysis ofthe experimental spectra under the assumption of a single-emitting excited state is not possible, consistent with ourinterpretation of there being two phases, namely, a disor-dered phase and an aggregated phase. From the spectraldecomposition, we can determine the fraction of aggregatespresent in the films. This requires to correct the fraction ofaggregate absorption observed in the absorption spectra toaccount for the change in oscillator strength that takes placeon aggregation. This change in oscillator strength can beobtained, following the procedure by Clark et al.,43 from theabsorption spectra of PCPDTBT in solution as, in solution,there is an isosbestic point. Thus, we compared the addi-tional contribution of absorption of the aggregated phasewith the reduction in absorption of the disordered phasewhen going from 280 K to 240 K in solution. From this, we

TABLE 1 Fitting Parameters of the Franck-Condon Analyses for the Photoluminescence and Absorption Spectra of the Aggregated

Phase (agg.) and the Disordered Phase (dis.) of PCPDTBT in Solution

Photoluminescence Absorption

FC Parameter Unit f (%) E0 (eV) r (meV) I0–0/I0–1 f (%) E0 (eV) r (meV) I0–0/I0–1

180 K agg. 100 1.420 30 0.94 67 1.495 33 1.06

180 K dis. 0 – – – 33 1.720 60 0.70

240 K agg. 86 1.434 37 0.98 57 1.520 40 0.94

240 K dis. 14 1.598 46 0.76 43 1.720 63 0.66

340 K agg. 0 – – – 0 – – –

340 K dis. 100 1.614 46 0.76 100 1.720 63 0.76

Note: f is the relative contribution of the phase to the total spectrum; E0 is the position of 0–0 transition; r is the Gaussian standard deviation; and

I0–0/I0–1 is the peak ratio of the 0–0 line and the 0–1 line, the latter is calculated as the sum of all Huang-Rhys parameters of the present modes with

the following energies: 62, 106, 136, 167, and 190 meV.

FIGURE 5 Optical density of PCPDTBT for different tempera-

tures: (from top to bottom) solution (c 5 0.25 mg/mL), CB/DIO

film, and CB-annealed film; and (bottom) normalized absorp-

tion of PCPDTBT in solution at 240 K (black line) and absorp-

tion of CB/DIO film (pink line) and CB-annealed film (green

line) at room temperature. [Color figure can be viewed in the

online issue, which is available at wileyonlinelibrary.com.]


POLYMER SCIENCE



find the oscillator strength of the aggregated chains to be1.456 0.10 higher than the oscillator strength of the disor-dered chains. This factor is in good agreement with relativeoscillator strengths found for transformation of disorderedto aggregated chains in P3HT solution and thin films.9,43 Thefraction of aggregates in the film can then be obtained bydividing the fraction of aggregate absorption observed in thespectra by the relative change in oscillator strength. Theresulting fraction of aggregates for PCPDTBT thin films andsolution is shown in Figure 6. In solution, the fraction ofaggregates increases with decreasing temperature until a sat-uration value of about 45% is reached at temperaturesbelow 200 K. In PCPDTBT thin films, the fraction of aggre-gates does not significantly depend on temperature. For theCB-annealed film, the aggregate fraction stays constantly at38%. In the CB/DIO film, the aggregate fraction slightly risesfrom 40 to 45% with decreasing temperature. Thus, bothfilm morphologies show similar fractions of aggregates whencompared with the saturation value in solution.

The photoluminescence spectra of the two films are com-pared with the photoluminescence in solution in Figure 7.Although in solution, emission is observed from both thedisordered phase (at 1.60 eV) and the aggregated phase (at1.42 eV), the thin film spectra show only one emissionwith a 0–0 peak at 5 K centered at 1.40 eV for the CB/DIOfilm and at 1.43 eV for the CB-annealed film. The photolu-minescence spectra of the films shift continuously to thered from 400 K to 5 K, with the shift being 60 meV (CB/DIO film) and 47 meV (CB-annealed film), respectively, asdetermined from the 0–0 peak positions obtained by FCanalysis. Concomitantly, the shape of the spectra changeswith temperature. Figure 8(a) shows the temperature-

dependent photoluminescence normalized such that the 0–0 transition as determined by the FC analysis (shown inthe Supporting Information) is at unity. The solution spec-tra are also displayed for ease of comparison. Evidently, inall spectra, the Gaussian linewidth broadens with increas-ing temperature from values around 30 meV to about 55meV. In addition, for the films, roughly 100 meV below the0–0 peak, a Gaussian peak increases strongly in intensityon heating above 150 K. The evolution with temperaturefor the PL spectra in solution can be modeled well by con-voluting the 180 K solution spectrum with a Gaussian line-width function of increasing linewidth to simulatetemperature-dependent inhomogeneous broadening, as dis-played in Figure 8(b). This is not possible for the film spec-tra. The change in the thin film spectra with temperaturecan only be reproduced if one of the two followingapproaches is used. One option is to presume a single-FIGURE 6 Fraction of aggregates as function of temperature for

PCPDTBT solution (squares), CB/DIO film (circles), and CB-

annealed film (triangles). This fraction was determined by calcu-

lating the fraction of aggregate absorption and the relative

change in oscillator strength for disordered and planarized

chains via Franck-Condon analysis. [Color figure can be viewed

in the online issue, which is available at wileyonlinelibrary.com.]

FIGURE 7 Intensity of photoluminescence of PCPDTBT for dif-

ferent temperatures: (from top to bottom) solution

(c 5 0.25 mg/mL), CB/DIO film, and CB-annealed film. [Color fig-

ure can be viewed in the online issue, which is available at

wileyonlinelibrary.com.]





emitting state and to model the spectra using FC analysisaccording to eq 2. To account for the spectral changes withtemperature, one needs to presume a modified, reducedintensity of the 0–0 peak and a strong increase (fromS5 0.16 to S5 0.63) of the vibrational mode at 136 meV(1096 cm21) from 100 K onward. This is shown in theSupporting Information. The second possibility is to postu-late two emitting states. One can then reproduce the spec-tra by superimposing a usual FC progression (eq 2) with abroad Gaussian peak at 70–130 meV below the 0–0 peak.The standard deviation r of this additional Gaussian peakis comparable with r of the corresponding FC progression.The intensity of the additional Gaussian increases withtemperature, whereas the Huang-Rhys parameters of theFC progression remain essentially unaffected. Figure 8(b)shows this superposition exemplarily for the film spectraat 250 K. In addition, we measured the photoluminescencedecay of the PCPDTBT thin films at room temperature andfound an exponential decay with lifetimes of s 5 250 ps forboth films (see Supporting Information). The quantumyield U of both films is less than 0.1% at room tempera-ture (below measuring limit). The radiative decay rate kr

and the nonradiative decay rate knr can be calculatedaccording to the following equation:

U ¼ kr � s ¼kr

kr1knr; (4)

resulting in kr 5 0.004 ns21 and knr 5 4.0 ns21.

DISCUSSION

Two Phases in PCPDTBT SolutionSeveral semiconducting polymers are known to show a con-formational transformation from a disordered phase consist-ing of randomly disordered chains to a mixed phase whereordered aggregates are embedded in a disordered, amor-phous matrix of disordered chains. This has been shownextensively for P3HT, yet also for MEH-PPV and PFO. In solu-tion, the transformation can be induced by reducing the sol-vent quality or temperature.9,11,44–48 As shown in exemplaryfashion for low-polydispersity P3HT, the transformation fromthe fully disordered form to the phase containing aggregatesproceeds in a sequence of three steps on lowering the tem-perature of the solution.41 First, the absorption and emission

FIGURE 8 (a) Photoluminescence spectra of PCPDTBT thin films and in solution normalized to 0–0 transition as determined by

Franck-Condon analysis. The temperature ranges from 180 to 320 K for solution and from 5 to 480 K for films; arrows indicate the

direction of increasing temperature. (b; Top) 180 K PL spectrum convoluted with Gaussian line of standard deviation r 5 10, 20, 30,

40, and 50 meV; arrow indicates the direction of increasing r. (b; Center/Bottom) Photon flux, that is, PL intensity/energy3, of 250 K

PL film spectra (blue squares) and Franck-Condon fits (red line) comprising FC progression of the aggregated phase (green line

with filled circles) and a Gaussian-shaped emission (wine line with triangles). [Color figure can be viewed in the online issue,

which is available at wileyonlinelibrary.com.]


POLYMER SCIENCE



spectra of the disordered form shift to slightly lower ener-gies and the oscillator strength increases slightly. This hasbeen interpreted to indicate an increase in conjugationlength due to planarization of the polymer backbone. Next,an isosbestic point can be observed that has been taken toindicate the transformation from disordered chains to aggre-gated chains. Finally, a continuous bathochromic shift of the0–0 peak of the aggregate absorption indicates an ongoingfurther planarization that leads to a longer conjugationlength. These observations are also made for PCPDTBT oncooling the solution (Fig. 1). In addition, the measurementsof absorption and emission spectra as a function of concen-tration (Fig. 2) demonstrate that the lower energy absorp-tion and emission results from the interaction of polymerchains with each other. It is thus clear that PCPDTBT canexist in two phases, namely, a fully disordered phase and aphase containing aggregated chains, consistent with earlierobservations by Peet et al.28 As revealed by Figure 3, the S1excited state of the disordered form and the aggregated formdiffers by 180 meV. This shift is consistent with the energydifferences between emission from disordered and aggre-gated chains in P3HT (200 meV; ref. 41), PPV (180 meV; ref.37), MEH-PPV (120 meV; ref. 11), and PFO (110 meV; refs.12 and 14; 100 meV; ref. 49).

The phenomenon of a disorder-order transition is oftenregarded as specific to P3HT. It is well known that it plays amajor role in controlling the efficiency of bulk heterojunctionsolar cells of P3HT with a fullerene. Evidently, this disorder-order transition can be found in a range of chemically differ-ent polymers, and it proceeds in an analogous manner forP3HT, MEH-PPV, and PCPDTBT. Existing differences relate tothe temperature at which the transformation proceeds, thesharpness of the temperature-dependent transformation, themaximum fraction of aggregation obtained, and the predomi-nance of H- or J-type character in the spectra.

Although MEH-PPV and P3HT show transformation tempera-tures at about 200 K and 250 K, respectively, PCPDTBT showsthe conformational transformation already at room tempera-ture. Clearly, this has a strong impact on films prepared atroom temperature for applications of PCPDTBT in organicsemiconductor devices. Considering the sharpness of the con-formational transformation, Figure 6 shows that for ourPCPDTBT with a molecular weight of 23 kDa (Mw) and a PDIof 1.7, the phase transformation extends from 320 K to260 K, that is, over a range of about 60 K. This is the sametemperature range as observed for polydisperse, commercialMEH-PPV and for P3HT with a molecular weight of 19 kDa(Mn) and a PDI of 2.0.11,41 Note that when using a P3HT withsame molecular weight yet lower PDI of 1.1, the transitionrange reduces to only 25 K. With respect to the saturationfraction of aggregates, the saturation value of 40% of aggre-gates in solution that we observe for PCPDTBT is consistentwith the similar values found for P3HT and MEH-PPV.9,11,41

P3HT and MEH-PPV are known to form weakly interactingaggregates with either H-type or J-type character, depending

on the detailed conditions in the solutions (with respect tosolvent mixture, solvation process, molecular weight of thepolymer, polydispersity, etc.).11,40,41,50–55 Whether H-type or J-type character prevails depends on the relative orientation ofadjacent chains to each other as well as the amount of ener-getic disorder that is present in the sample and that may arisefrom variations in structure or in polarization of the environ-ment.56,57 On cooling solutions with THF or MTHF, low- tomedium-molecular-weight P3HT usually forms H-type aggre-gates, whereas MEH-PPV adopts a mixed H/J character.39,58 Asevidenced by Figure 2, PCPDTBT is clearly an aggregate. Itsvibrational structure, however, can be reproduced withoutneed to suppress or enhance the 0–0 peak relative to theremaining vibrational progression. Thus, PCPDTBT has neithera predominant H-type character nor an eminent J-type nature.Evidently, the inherent energetic disorder is large when com-pared with the amount of coupling between chromophores sothat the signatures of a prevailing H- or J-type character aremasked.39 The PCPDTBT aggregates are perhaps bestdescribed as strongly disordered H/J aggregates.

Aggregate Formation in PCPDTBT Thin Films(CB/DIO Films)When studying the formation of the low-energy b-phase inPFO thin films, Khan et al.13,49 showed that small aggregatespresent in the solution act as nucleation sites during thespin-coating process. Here, we have evidence for PCPDTBTaggregates in an MTHF solution at room temperature, andwe have shown that the formation of aggregates depends onthe concentration of the solution. Thus, we consider that pre-existing aggregates act as nucleation sites when the film for-mation takes place in the presence of a slowly evaporatingsolvent as is the case when spin-coating a PCPDTBT filmfrom a CB/DIO solution. The aggregates formed in solutionand those formed in a CB/DIO spin-coated film seem to beidentical. This can be inferred from the nearly identicalabsorption spectra of aggregates in solution at 240 K and aCB/DIO film at room temperature (Fig. 5), as well as theclose similarity of the photoluminescence spectra at 180 Kin solution and at 5 K in CB/DIO film (Fig. 7).

The fraction of aggregates found in the CB/DIO film is com-parable with the saturation value obtained in solution, sug-gesting that aggregate formation has been optimized by filmpreparation. The lack of an isosbestic point when cooling afilm, in contrast to its presence when cooling a solution, isremarkable. We attribute this to the change in environmentalpolarization in the film on cooling. In solution, the polymerchains are surrounded by solvent molecules, whose polariza-tion does not change significantly on cooling. In contrast, infilm, the chains are embedded in neighboring chains, whoseconjugation length, overall density, and thus polarizationchanges with temperature. The associated shift in polariza-tion energy thus masks any isosbestic point.

Energy Transfer in PCPDTBT Thin FilmsIf we decompose the absorption spectra into absorption fromnoninteracting chains and absorption from aggregated chains,



we find similar fraction of aggregates for solution below thetransformation temperature regime as for thin films. Theamount of aggregates present in the films corresponds to thesaturation value in solution (Fig. 6). In contrast, in PL, weobserve differences. The PL spectra of PCPDTBT in solutionshow emission from both disordered chains and aggregatedchains, whereas the film spectra show only aggregate emis-sion. In thin films, energy transfer can take place efficiently byexciton diffusion, that is, a multistep hopping process witheach hopping occurring by F€orster transfer.49,59–61 We attrib-ute the fact that we see only aggregate emission in thin filmscontaining 60% of amorphous material to exciton diffusionfrom amorphous parts of the film to aggregated parts. Thisalso implies that the approximate diameter of any amorphousphase is less than twice the exciton diffusion length at therespective temperature and that there is close intermixing ofamorphous phase with aggregates. It seems that the aggre-gated parts are present as short-range ordered regionsembedded in the amorphous matrix rather than as long-rangeordered assemblies as found in polymer crystallites.

The analysis of the spectra has already shown that the chro-mophores constituting the aggregate are only weakly inter-acting so that the spectra retain the vibrational structurethat is characteristic for noninteracting chromophores. Inthis context, one may question whether the exciton dynam-ics, that is, the diffusion of excitations, in weakly interactingaggregates differs from that typically found for amorphousfilms. In Figure 9, we plotted the spectral shift De between theposition of the 0–0 transition in the photoluminescence andthe center of the DOS normalized to the Gaussian linewidth rversus kT/r as determined via FC analysis of absorption andphotoluminescence spectra (see Supporting Information forabsorption spectra at low temperatures). When comparedwith the theoretical dependence De(T)/r(T)5 2r(T)/kT,62–64

we see good agreement between theory and experimentaldata for kT/r> 0.3 (above 150 K). At lower temperatures, sat-uration of De/r occurs since the life time of the excitation lim-its the relaxation progress within the DOS, as is wellestablished for this process of spectral diffusion.62 Thus, wefound from the analysis of the steady-state spectra that thetemperature-dependent dynamics in these weakly interactingaggregates does not differ from that of noninteracting chromo-phores. This further supports the notion of a short-rangeordered character of the aggregates.

Temperature-Induced Structural ChangesThe changes induced by temperature in the thin filmabsorption spectra (Fig. 5) are weak. Overall, the spectrashift slightly and continuously to the red spectral range andthey become more structured on cooling. As already men-tioned when discussing the lack of an isosbestic point inthe films, we attribute the bathochromic shift to anincreased polarization when the density of the polymerchains decreases on cooling, whereas the reduced inhomo-geneous broadening is commonly attributed to a reductionof molecular vibrations.

In contrast to the absorption spectra, the photoluminescencespectra show significant differences in their evolution withtemperature between the spectra taken in solution and thosetaken in thin films (Fig. 8). In solution, the spectral shapereflects temperature-induced linewidth broadening. This islikely to arise from increased torsional motion as torsionalmotion has also been shown to be the source oftemperature-induced linewidth broadening for MEH-PPV inMTHF.65 In contrast, in the films, an increased intensity inthe red tail of the photoluminescence adds to the broaden-ing. As mentioned above, this may be accounted for by atemperature-induced increase of the 136 meV vibration orby the temperature-induced population of a lower-energyemissive state. We consider that the latter is the most likelyexplanation for the following reasons:

� Modeling the spectra as resulting from only one emit-ting state requires the use of a modified FC analysiswith a suppressed 0–0 intensity (see Supporting Infor-mation) from about 150 K onward. This would suggesta different orientation and order of the chain at highertemperatures. For the solution spectra, this is not neces-sary. It is not plausible why, in a film, where chainmovement is constrained when compared with solution,such a reorientation should occur while it is absent insolution. In addition, the spectra considered are emis-sion spectra. The vibrational peaks observed relate tothe vibrations induced in the molecule in the ground statewhen the nuclei want to return to their ground-state equilib-rium position after the vertical electronic transition from theexcited state. It is not obvious why raising the temperatureshould change the intensity of a particular ground-state

FIGURE 9 (Solid symbols) Energetic difference De between the

position of the 0–0 transition in the photoluminescence and the

center of the density of states (DOS) normalized to the stand-

ard deviation r versus kT/r for the CB/DIO film. De was deter-

mined by Franck-Condon analysis of the absorption spectra.

(Dashed line) The theoretical dependence: De(T)/r(T) 5 2r(T)/

kT. [Color figure can be viewed in the online issue, which is

available at wileyonlinelibrary.com.]


POLYMER SCIENCE



vibrational mode. Thus, we discard this approach as physi-cally unlikely.

� The model of two emitting states, in contrast, allows repro-ducing the spectra with the same FC analysis as used forsolution, without need for suppressing the 0–0 peak. Atemperature-induced population of a lower energy emissivestate is possible. As demonstrated by Mikhnenko et al.,66

exciton diffusion becomes temperature-activated fromabout 150 K onward. If a very low amount of red-emittingsites were incorporated in a film, the amount of emissionfrom those sites would thus increase with temperaturefrom 150 K onward, consistent with the observation in ourdata. Note that if low-energy sites are formed in solution,the lack of efficient diffusion across the solution preventsthem to be sufficiently populated for detection.

Although we can assign the spectral changes with temperatureto red-emissive trap sites, we are not able to unambiguouslyidentify the nature of these sites merely on the basis of thespectroscopic data. The broad emission shape would be con-sistent with excimer-like sites, forming to a low percentage inthe film, as a possible origin.67,68 In polycrystalline samples,excimers have been demonstrated to form in particular atgrain boundaries as a result of structural dislocations. Thus,excimer formation comprising structural trapping is promotedby small domain sizes, a large amount of interfaces, and fastexciton diffusion.35 To assess whether the optical data are con-sistent with an excimer-like origin, we compare the radiativedecay rate kr and the nonradiative decay rate knr obtained forthe PCPDTBT films at room temperature with typical valuesfor molecular assemblies and solids collected by Gierschnerand Park.69 For excimer emission of distyrylbenzene deriva-tives in the solid state, Gierschner and Park find a quantum

yield of 0.01, a lifetime of 8.3 ns, and a radiative decay rate of0.001 ns21. This results in a nonradiative decay rate of 0.12ns21. The radiative decay rate is comparable with the valuewe found for PCPDTBT thin films, whereas our values for thenonradiative decay rates is 30 times higher. However, the dis-tyrylbenzene derivative investigated by Gierschner and Parkemits at 3.3 eV, whereas our material emits at 1.4 eV. As thenonradiative decay rate in polymers increases drastically withdecreasing energy gap between ground and excited states, afactor of 30 is fully consistent with the energy gap law.70 Thus,we tentatively attribute the low-energy emissive state inPCPDTBT thin films to a low percentage of excimer statespresent whose population is thermally activated.

Correlation of Optical Properties in CB/DIO Films toStructural AnalysisInformation on the structure of PCPDTBT has only becomeavailable in the recent years. On the basis of grazing incidencewide-angle X-ray experiments,29,33,71 Russell et al., Nelsonet al. and Bazan et al. could show that PCPDTBT forms anedge-on morphology with long-range p-stacking with a dis-tance between the plane of the backbone at around 3.8Å.29,33,71 This applies to PCPDTBT films prepared by spin-coating from neat CB and from CB solution containing solventadditives such as DIO or ODT (2 wt %). These aggregates typi-cally extend over about 40 nm and are randomly oriented inthe film. Figure 10(a) shows an image taken by AFM.34 This isthe structure that prevails in our CB/DIO films.

The absorption and emission spectra in the CB/DIO filmclosely resemble the spectra and the energetic positionsobtained in solution (Figs. 5 and 7). Thus, it appears likelythat PCPDTBT may also prevail in p-stacked assemblies of

FIGURE 10 AFM height images of a CB/DIO film (a) and a CB-annealed film (c). (b) Polarized microscopy image of a CB-annealed

film. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]




comparable size (�40 nm) in solution. This would alsoaccount for the similar percentage of 40% aggregatesobserved in both solution and CB/DIO films.

Comparison of CB/DIO Films to Solvent-Annealed FilmsIt was demonstrated earlier that conjugated homopolymerssuch as P3HT can be recrystallized into highly orderedstructures using solvent vapor annealing protocols.36,72 In arecent publication, it has been shown that also fromPCPDTBT highly crystalline films can be obtained by solventvapor annealing using CB as solvent.34 The films prepared inthis manner (CB-annealed films) show clear evidence foraggregates, yet their spectral signature differs in someaspects indicating an altered structure when compared withthe CB/DIO films. For example, the absorption and photolu-minescence spectra (Fig. 7) in the CB-annealed film areshifted hypsochromically (Fig. 5) when compared with theCB/CIO film. Furthermore, the photoluminescence spectrashow more inhomogeneous broadening at room temperature,and they evolve differently with temperature due to a higheramount of low-energy emissive states when compared withthe CB/DIO film (Fig. 8). We consider that this difference inthe optical spectra reflects an underlying difference instructure.

Indeed, for the CB-annealed films, AFM and polarizingmicroscopy indicate the formation of extended aggregates[Fig. 10(b,c)]. Recent structural investigations by Ludwigset al. using DSC, TEM/ED, and GIWAXS have shown that suchsamples are highly crystalline and give rise to a melting pointaround 280 8C.34 The PCPDTBT crystallizes in polymorph FormI which involves a high symmetry packing of four chains in anorthorhombic cell (Pccn space group) with pseudohexagonalsymmetry. The chains are paired into dimers with a � 3.7 Åinterchain p-stacking distance and a segregated mode of stack-ing of cyclopentadithiophene and benzothiadiazole (BT). Thisstructure does not allow for long-range pi-interaction.73

Although the edge-on, p-stacked structure found for the CB/DIO films is akin to the structure of P3HT aggregates, the dimerstructure observed for the CB-annealed films is more similar tostructures reported for poly(fluorene)s.74

Remarkably, the hypsochromic shift seen in Figures 5 and 7for the CB-annealed films applies to both the aggregatedphase and the disordered phase. In principle, a hypsochro-mic shift can arise from two effects, that is, first, a shorteraverage conjugation length of the chains, and second, areduced polarization of the medium surrounding the poly-mer chains. Our experiments do not allow us to distinguishbetween these two options. However, we point out that theabsence of long-range p-stacking that is found in the CB-annealed films when compared with the CB/DIO films wouldseem consistent with a reduced polarization for the aggre-gated phase and could also impact on disordered chainsembedded between the crystalline structures. In this context,the larger inhomogeneous broadening in the CB-annealedfilm would indicate either a larger variation of conjugationlengths or a stronger spatial variation of the polarizability.

The slightly lower fraction of aggregates present in the CB-annealed film when compared with the CB/DIO film, in par-ticular at room temperature (Fig. 6), suggests that aggregateformation is somewhat more optimized in the CB/DIO film.This is also consistent with the reduced amount of low-energy trap sites in the CB/DIO film when compared withthe CB-annealed film.

CONCLUSIONS

Like P3HT and MEH-PPV, the low-bandgap copolymerPCPDTBT that is frequently used for solar cell applicationsundergoes a conformational transformation on cooling insolution. Above the temperature Tc for the transformation,the polymer prevails in a nonaggregated manner in a ran-domly disordered conformation. Below Tc, up to about 40–45% of the polymer chains aggregate and adopt a more pla-narized conformation. For PCPDTBT, the temperature Tc atwhich the phase transformation takes place in MTHF isaround 300 K. Comparison with structural investigationssuggest that p-stacked aggregates prevailing in solution atroom temperature serve as nucleation points during the pro-cess of spin-coating, thus resulting in films containing ran-domly oriented aggregates. In contrast, solvent vaporannealing a CHCl3-spun film using CB vapor induces a dimerstructure. We show that using suitable processing, the maxi-mum fraction of about 42% aggregates can also be obtainedin thin films of PCPDTBT. In these films, a small percentageof low-energy trap states prevails that are populated by exci-ton diffusion for temperatures above 150 K. The emissioncharacteristics of these states are consistent with an originfrom excimers resulting from structural defects such as grainboundaries.

ACKNOWLEDGMENTS

This work received funding from the Graduiertenkolleg1640 of the Deutsche Forschungsgemeinschaft, the FreeState of Bavaria in the context of the program “Solar Tech-nologies go Hybrid,” and the Bayerisches Programm zurF€orderung der Chancengleichheit f€ur Frauen in Forschungund Lehre. The authors thank Heinz B€assler for fruitful dis-cussions and Fabian Panzer for assistance with the spectro-scopic setup. We thank M. Brinkmann for discussion andcollaboration on the structural model of PCPDTBT and theDFG for funding within the priority program “ElementaryProcesses in Organic Solar Cells” (SPP-1355) and the EmmyNoether Program.

REFERENCES AND NOTES

1 L. G. Li, G. H. Lu, X. N. Yang, J. Mater. Chem. 2008, 18, 1984.

2 Y. Zhao, Z. Y. Xie, Y. Qu, Y. H. Geng, L. X. Wang, Appl. Phys.

Lett. 2007, 90.

3 M. Schubert, D. Dolfen, J. Frisch, S. Roland, R. Steyrleuthner,

B. Stiller, Z. H. Chen, U. Scherf, N. Koch, A. Facchetti, D.

Neher, Adv. Energy Mater. 2012, 2, 369.


POLYMER SCIENCE


4 A. Zen, J. Pflaum, S. Hirschmann, W. Zhuang, F. Jaiser, U.

Asawapirom, J. P. Rabe, U. Scherf, D. Neher, Adv. Funct.

Mater. 2004, 14, 757.

5 A. A. Virkar, S. Mannsfeld, Z. A. Bao, N. Stingelin, Adv.

Mater. 2010, 22, 3857.

6 X. N. Yang, J. Loos, S. C. Veenstra, W. J. H. Verhees, M. M.

Wienk, J. M. Kroon, M. A. J. Michels, R. A. J. Janssen, Nano

Lett. 2005, 5, 579.

7 D. Chirvase, J. Parisi, J. C. Hummelen, V. Dyakonov, Nano-

technology 2004, 15, 1317.

8 J. S. Liu, T. Tanaka, K. Sivula, A. P. Alivisatos, J. M. J.

Frechet, J. Am. Chem. Soc. 2004, 126, 6550.

9 C. Scharsich, R. H. Lohwasser, M. Sommer, U. Asawapirom,

U. Scherf, M. Thelakkat, D. Neher, A. K€ohler, J. Polym. Sci. Part

B: Polym. Phys. 2012, 50, 442.

10 C. E. Johnson, D. S. Boucher, J. Polym. Sci. Part B: Polym.

Phys. 2014, 52, 526.

11 A. K€ohler, S. T. Hoffmann, H. B€assler, J. Am. Chem. Soc.

2012, 134, 11594.

12 J. Peet, E. Brocker, Y. H. Xu, G. C. Bazan, Adv. Mater. 2008,

20, 1882.

13 A. L. T. Khan, M. J. Banach, A. K€ohler, Synth. Met. 2003,

139, 905.

14 M. Ariu, D. G. Lidzey, M. Sims, A. J. Cadby, P. A. Lane, D.

D. C. Bradley, J. Phys.: Condens. Matter 2002, 14, 9975.

15 D. Fazzi, G. Grancini, M. Maiuri, D. Brida, G. Cerullo, G.

Lanzani, Phys. Chem. Chem. Phys. 2012, 14, 6367.

16 L. Marin, H. Penxten, S. Van Mierloo, R. Carleer, L. Lutsen,

D. Vanderzande, W. Maes, J. Polym. Sci. Part A: Polym. Chem.

2013, 51, 4912.

17 L. T. Dou, J. B. You, J. Yang, C. C. Chen, Y. J. He, S. Murase, T.

Moriarty, K. Emery, G. Li, Y. Yang, Nat. Photonics 2012, 6, 180.

18 E. E. Havinga, W. Tenhoeve, H. Wynberg, Polym. Bull. 1992,

29, 119.

19 F. Liu, Y. Gu, J. W. Jung, W. H. Jo, T. P. Russell, J. Polym.

Sci. Part B: Polym. Phys. 2012, 50, 1018.

20 W. L. Ma, C. Y. Yang, X. Gong, K. Lee, A. J. Heeger, Adv.

Funct. Mater. 2005, 15, 1617.

21 Y. Kim, S. A. Choulis, J. Nelson, D. D. C. Bradley, S. Cook,

J. R. Durrant, Appl. Phys. Lett. 2005, 86.

22 G. Li, V. Shrotriya, Y. Yao, J. S. Huang, Y. Yang, J. Mater.

Chem. 2007, 17, 3126.

23 C. V. Hoven, X. D. Dang, R. C. Coffin, J. Peet, T. Q. Nguyen,

G. C. Bazan, Adv. Mater. 2010, 22, E63.

24 J. K. Lee, W. L. Ma, C. J. Brabec, J. Yuen, J. S. Moon, J. Y.

Kim, K. Lee, G. C. Bazan, A. J. Heeger, J. Am. Chem. Soc.

2008, 130, 3619.

25 F. Etzold, I. A. Howard, N. Forler, D. M. Cho, M. Meister, H.

Mangold, J. Shu, M. R. Hansen, K. Mullen, F. Laquai, J. Am.

Chem. Soc. 2012, 134, 10569.

26 J. Peet, J. Y. Kim, N. E. Coates, W. L. Ma, D. Moses, A. J.

Heeger, G. C. Bazan, Nat. Mater. 2007, 6, 497.

27 F. S. U. Fischer, K. Tremel, A. K. Saur, S. Link, N. Kayunkid,

M. Brinkmann, D. Herrero-Carvajal, J. T. L. Navarrete, M. C. R.

Delgado, S. Ludwigs, Macromolecules 2013, 46, 4924.

28 J. Peet, N. S. Cho, S. K. Lee, G. C. Bazan, Macromolecules

2008, 41, 8655.

29 Y. Gu, C. Wang, T. P. Russell, Adv. Energy Mater. 2012, 2, 683.

30 D. Niedzialek, V. Lemaur, D. Dudenko, J. Shu, M. R. Hansen,

J. W. Andreasen, W. Pisula, K. Mullen, J. Cornil, D. Beljonne,

Adv. Mater. 2013, 25, 1939.

31 S. H. Wang, M. Kappl, I. Liebewirth, M. Muller, K. Kirchhoff,

W. Pisula, K. Mullen, Adv. Mater. 2012, 24, 417.

32 H. N. Tsao, D. M. Cho, I. Park, M. R. Hansen, A. Mavrinskiy,

D. Y. Yoon, R. Graf, W. Pisula, H. W. Spiess, K. Mullen, J. Am.

Chem. Soc. 2011, 133, 2605.

33 T. Agostinelli, T. A. M. Ferenczi, E. Pires, S. Foster, A.

Maurano, C. Muller, A. Ballantyne, M. Hampton, S. Lilliu, M.

Campoy-Quiles, H. Azimi, M. Morana, D. D. C. Bradley, J.

Durrant, J. E. Macdonald, N. Stingelin, J. Nelson, J. Polym.

Sci. Part B: Polym. Phys. 2011, 49, 717.

34 F. S. U. Fischer, D. Trefz, J. Back, N. Kayunkid, B. Tornow,

S. Albrecht, K. Yager, G. Singh, A. Karim, D. Neher, M.

Brinkmann, S. Ludwigs, Adv. Mater. 2015, 27, 1223.

35 J. Gierschner, L. Luer, B. Milian-Medina, D. Oelkrug, H. J.

Egelhaaf, J. Phys. Chem. Lett. 2013, 4, 2686.

36 E. J. W. Crossland, K. Rahimi, G. Reiter, U. Steiner, S.

Ludwigs, Adv. Funct. Mater. 2011, 21, 518.

37 P. K. H. Ho, J. S. Kim, N. Tessler, R. H. Friend, J. Chem.

Phys. 2001, 115, 2709.

38 S. J. Strickler, R. A. Berg, J. Chem. Phys. 1962, 37, 814.

39 H. Yamagata, N. J. Hestand, F. C. Spano, A. K€ohler, C.

Scharsich, S. T. Hoffmann, H. B€assler, J. Chem. Phys. 2013, 139.

40 A. P. Roque, L. A. Mercante, V. P. Scagion, J. E. Oliveira, L.

H. C. Mattoso, L. De Boni, C. R. Mendonca, D. S. Correa, J.

Polym. Sci. Part B: Polym. Phys. 2014, 52, 1388.

41 F. Panzer, H. B€assler, R. Lohwasser, M. Thelakkat, A. K€ohler,

J. Phys. Chem. Lett. 2014, 5, 2742.

42 H. P. Chen, Y. C. Hsiao, B. Hu, M. Dadmun, Adv. Funct.

Mater. 2014, 24, 5129.

43 J. Clark, J. F. Chang, F. C. Spano, R. H. Friend, C. Silva,

Appl. Phys. Lett. 2009, 94.

44 F. B. Dias, J. Morgado, A. L. Macanita, F. P. da Costa, H. D.

Burrows, A. P. Monkman, Macromolecules 2006, 39, 5854.

45 R. F. Cossiello, M. D. Susman, P. F. Aramendia, T. D. Z.

Atvars, J. Lumin. 2010, 130, 415.

46 G. A. Sherwood, R. Cheng, T. M. Smith, J. H. Werner, A. P.

Shreve, L. A. Peteanu, J. Wildeman, J. Phys. Chem. C 2009,

113, 18851.

47 C. C. Kitts, D. A. Vanden Bout, Polymer 2007, 48, 2322.

48 O. Ingan€as, W. R. Salaneck, J. E. Osterholm, J. Laakso,

Synth. Met. 1988, 22, 395.

49 A. L. T. Khan, P. Sreearunothai, L. M. Herz, M. J. Banach, A.

K€ohler, Phys. Rev. B 2004, 69.

50 E. T. Niles, J. D. Roehling, H. Yamagata, A. J. Wise, F. C.

Spano, A. J. Moule, J. K. Grey, J. Phys. Chem. Lett. 2012, 3, 259.

51 W. Y. So, J. Y. Hong, J. J. Kim, G. A. Sherwood, K. Chacon-

Madrid, J. H. Werner, A. P. Shreve, L. A. Peteanu, J. Phys.

Chem. B 2012, 116, 10504.

52 R. T. Zhou, W. Chen, X. Y. Jiang, S. F. Wang, Q. H. Gong,

Appl. Phys. Lett. 2010, 96.

53 K. P. Chao, S. L. Biswal, Langmuir 2014, 30, 4236.

54 A. M. Crotty, A. N. Gizzi, H. J. Rivera-Jacquez, A. E.

Masunov, Z. J. Hu, J. A. Geldmeier, A. J. Gesquiere, J. Phys.

Chem. C 2014, 118, 19975.

55 T. P. Martin, A. J. Wise, E. Busby, J. Gao, J. D. Roehling, M.

J. Ford, D. S. Larsen, A. J. Moule, J. K. Grey, J. Phys. Chem. B

2013, 117, 4478.

56 M. Kasha, Radiat. Res. 1963, 20, 55.

57 F. C. Spano, C. Silva, Annu. Rev. Phys. Chem. 2014, 65, 477.

58 J. Clark, C. Silva, R. H. Friend, F. C. Spano, Phys. Rev. Lett.

2007, 98.



59 R. Q. Albuquerque, C. C. Hofmann, J. K€ohler, A. K€ohler, J.

Phys. Chem. B 2011, 115, 8063.

60 S. A. Bagnich, C. Im, H. B€assler, D. Neher, U. Scherf, Chem.

Phys. 2004, 299, 11.

61 L. M. Herz, C. Silva, A. C. Grimsdale, K. M€ullen, R. T.

Phillips, Phys. Rev. B 2004, 70.

62 S. T. Hoffmann, H. B€assler, J. M. Koenen, M. Forster, U.

Scherf, E. Scheler, P. Strohriegl, A. K€ohler, Phys. Rev. B 2010, 81.

63 H. B€assler, Phys. Status Solidi B: Basic Res. 1993, 175, 15.

64 S. Athanasopoulos, S. T. Hoffmann, H. B€assler, A. K€ohler, D.

Beljonne, J. Phys. Chem. Lett. 2013, 4, 1694.

65 S. T. Hoffmann, H. B€assler, A. K€ohler, J. Phys. Chem. B

2010, 114, 17037.

66 O. V. Mikhnenko, M. Kuik, J. Lin, N. van der Kaap, T. Q.

Nguyen, P. W. M. Blom, Adv. Mater. 2014, 26, 1912.

67 H. J. Egelhaaf, J. Gierschner, D. Oelkrug, Synth. Met. 2002,

127, 221.

68 J. Gierschner, H. J. Egelhaaf, D. Oelkrug, K. M€ullen, J. Fluo-

resc. 1998, 8, 37.

69 J. Gierschner, S. Y. Park, J. Mater. Chem. C 2013, 1, 5818.

70 J. S. Wilson, N. Chawdhury, M. R. A. Al-Mandhary, M.

Younus, M. S. Khan, P. R. Raithby, A. K€ohler, R. H. Friend, J.

Am. Chem. Soc. 2001, 123, 9412.

71 J. T. Rogers, K. Schmidt, M. F. Toney, E. J. Kramer, G. C.

Bazan, Adv. Mater. 2011, 23, 2284.

72 E. J. W. Crossland, K. Tremel, F. Fischer, K. Rahimi, G.

Reiter, U. Steiner, S. Ludwigs, Adv. Mater. 2012, 24, 839.

73 F.S.U. Fischer, N. Kayunkid, D. Trefz, S. Ludwigs, M.

Brinkmann, Macromolecules. 2015, 48, 3974.

74 M. Brinkmann, Macromolecules 2007, 40, 7532.


POLYMER SCIENCE


Documents

Revealing Structure Formation in PCPDTBT by Optical ......Revealing Structure Formation in PCPDTBT by Optical Spectroscopy Christina Scharsich,1,2 Florian S. U. Fischer,3 Kevin Wilma,1,4