Tuning Intermolecular Interactions in Dioctyl-Substituted Polyfluorene via HydrostaticPressure

K. Paudel, H. Knoll, M. Chandrasekhar, and S. Guha*Department of Physics and Astronomy, UniVersity of Missouri, Columbia, Missouri 65211

ReceiVed: December 13, 2009; ReVised Manuscript ReceiVed: February 20, 2010

Polyfluorenes (PFs) represent a unique class of poly-para-phenylene-based blue-emitting polymers with intriguingstructure-property relationships. Slight variations in the choice of functionalizing side chains result in dramaticdifferences in the inter- and intrachain structures in PFs. Dioctyl-substituted PF (PF8) is characterized by differentbackbone conformations that depend upon the torsion angle between the monomers. We present photoluminescence(PL) and Raman scattering studies of bulk samples and thin films of dioctyl-substituted PF (PF8) under hydrostaticpressure. The bulk sample was further thermally annealed and studied as a function of pressure. The PL energiesof the as-is and thermally annealed samples both red shift but at very different rates, and the difference betweentheir pressure coefficients elucidates the role of the backbone torsional angle. This is further corroborated by densityfunctional theoretical calculations of a fluorene oligomer, where the energy gap is calculated as a function of boththe torsion angle as well as compression. The Raman peaks harden with increasing pressures; the intraring C-Cstretch frequency at 1600 cm-1 has a pressure coefficient of 7.2 cm-1/GPa and exhibits asymmetric line shapes athigher pressures, characteristic of a strong electron-phonon interaction.

I. Introduction

Blue-emitting polymers continue to attract a lot of attentionfor display applications. Among them are polyfluorenes (PFs),which belong to a class of the poly-para-phenylene (PPP)system. They are extremely attractive from the technologicalperspective due to their high photoluminescence quantum yieldand because their rich phase morphology makes them modelsystems for understanding fundamental photophysical phenom-ena. Molecular-level attributes such as local chain structure andside-chain conformations in these systems strongly impact theirtransport and device characteristics. Slight variations in thechoice of functionalizing side chains result in dramatic differ-ences in the inter- and intrachain structures in PFs. Highlightingthese differences are two prototypical PFs, poly-(9,9-(dio-ctyl)fluorene) (PF8) and poly-(9,9-(di ethylhexyl)fluorene) (PF2/6). Side-chain substitution gives rise to a rich array ofmesomorphic behavior, with the appearance of a nematic liquidcrystalline (n-LC) phase at higher temperatures.1 In the pastdecade, many research groups have focused on structural studiesof PFs using X-ray scattering,2-4 small-angle neutron scattering,5,6

and optical spectroscopy.7-10

PFs are characterized by 3-D crystalline phases as well asconformational phases or chromophores, where the latterdepends on the structure of the individual chains. PF2/6 has alimited number of conformational isomers that form a five-foldhelix (5/2 or 5/1).11 These helices in turn self-organize into three-chain unit cells, resulting in a well-ordered semicrystallinehexagonal phase with coherence lengths exceeding 50 nm.4

Thus, the optical properties of PF2/6 are relatively insensitiveto the exact crystallographic state, thermal history, or molecularweight. PF8, on the other hand, has at least three conformationalisomers that depend explicitly upon the torsional angle betweenthe adjacent monomers.12 In addition to these isomers, structuralstudies of PF8 have identified many crystalline phases.13

Recent studies have identified three different conformationalisomers in terms of the local PF backbone torsional angle: theCR conformer with a torsional angle of 135°, the Cγ with anaverage torsional angle of 155°, and the almost planar con-former, C�, with an average torsional angle of 165°.12 The C�

conformer, conventionally known as the � phase, has a separatelong-wavelength feature in both optical absorption and emissionspectra.7 Although this phase appears as a minority constituent,it dominates the optical emission due to this lower emissionenergy. Our recent work has shown that the morphology adoptedby the side chains of PF8 is closely linked to the structure ofthe molecular backbone.14,15 For example, the � phase backboneis stabilized when the side chains in PF8 adopt an anti-gauche-gauche conformation, as shown in the inset of Figure1. The rich phase morphology of PF8 is also seen in thin filmsof its model oligomers.16

Hydrostatic pressure allows a study of materials in a regionof phase space not accessible by chemical techniques. Applica-

* To whom correspondence should be addressed. E-mail: guhas@missouri.edu.

Figure 1. PL and absorption spectra of as-is and thermally cycledPF8 films under ambient conditions. The inset shows a PF monomerwith dioctyl side chains in an anti-gauche-gauche conformation.

J. Phys. Chem. A 2010, 114, 4680–46884680

10.1021/jp911778r 2010 American Chemical SocietyPublished on Web 03/17/2010

tion of hydrostatic pressure allows tuning of both intermolecularand intramolecular interactions in polymer chains withoutchanges in the chemical makeup.17 The photophysical studiesof conjugated materials under high pressure indicate thatenhanced intermolecular interaction produces an increaseddegree of conjugation, increased exciton trapping in organicpolycrystalline molecules,18 red shift and broadening of thephotoluminescence,19-22 enhancement of excited-state dynamicsat polymer-polymer heterojunctions,23 destabilization of local-ized states as in methylated ladder-type PPP (MeLPPP),24 andchanges in the ring torsion motion as in nonplanar para-hexaphenyl (PHP) and other oligophenyls.25,26 The pressure-dependent optical studies in PHP and other oligophenyls furtherreveal the individual contributions of the intermolecular andintramolecular interactions.25

Optical studies in PFs under pressure allow tuning of the richphase space and its impact on electron-phonon interactions.Our previous work on PF2/6 under pressure shows dramaticchanges in the PL spectrum that mainly originate from defectand aggregate states.27 As-is bulk PF8 was studied underpressure recently;28 however, this study could not disentanglethe pressure shifts due to planarization of the backbone andchanges in the optical energy gap owing to enhanced intrachain/interchain interactions. The as-is samples have some fractionof the planar C� conformation; hence, the PL is alwaysdominated by this phase. Thermal annealing of PF8 eliminatesthe low-energy C� conformation, making the polymer morenonplanar. A comparison of the as-is and thermally annealedsamples as a function of pressure (where pressure has beenknown to planarize oligophenyls) therefore provides an idealplatform for separating the two effects that contribute to theshift in the optical energy gap, changes in the torsional angleand a higher overlap of the electronic wave function due toenhanced intra/interchain interactions. Furthermore, PL studiesas a function of pressure can isolate contributions due tointrachain and interchain interactions. Typically the PL energyshift as a function of pressure is a direct consequence of theintrachain interactions, whereas PL line widths as a function ofpressure reflect interchain interactions.17

The Raman spectra of PF8 as a function of pressure for bothas-is and thermally annealed samples show a similar behaviorand follow the same trends as those in PF2/6. The intraringstretch mode at 1600 cm-1 exhibits unexpected antiresonanceline shapes at higher pressures, indicating anharmonic interactions.

II. Methodology

A. Experimental Details. The PF8 sample was obtainedfrom American Dye Source (BE-129) and loaded as-is in thepressure cell. For thin-film studies under pressure, a solutionof PF8 (in toluene) was drop-casted on the bottom surface ofthe diamond. The sticking coefficient of PF films on diamondis quite poor, and therefore, they experience a hydrostaticenvironment. The pressure studies were conducted in aMerrill-Bassett-type diamond anvil cell (DAC) with cryogeni-cally loaded argon as the pressure medium. The pressure wasmeasured using the luminescence of a ruby chip located in thepressure chamber. The photoluminescence (PL) spectra wereexcited using the 325 nm line of a HeCd laser or the 351 nmline of an Ar ion laser. The luminescence excitation wasanalyzed with an Ocean Optics 2000 spectrometer with 25 µmslits. The Raman spectra were collected using an Invia Renishawspectrometer attached to a confocal microscope with a ×50 longworking distance objective and the 785 nm line of a diode laseras the excitation wavelength. The typical laser power was a

few mW on the sample. The bulk sample under pressure wasalso thermally cycled at room temperature (RT) from its n-LCphase at 160 °C, using similar thermal steps as in previousambient pressure studies.14

B. Computation. Theoretical computation of fluorene oli-gomers was performed with hybrid density functional theoryin the GAUSSIAN03 program.29 We employed the B3LYP(Becke’s three parameter hybrid) functional combined with the3-21G* basis set to perform geometry optimization.30 Other basissets such as 6-311G* gave comparable results for fluoreneoligomers. The energy gap was obtained by taking the differencebetween the highest energy of the single-particle highestoccupied molecular orbital (HOMO) and the lowest energy ofthe single-particle lowest unoccupied molecular orbital (LUMO).One must bear in mind that this is not the optical gap observedexperimentally, which takes into account the many-body exci-tonic levels and thus has a smaller energy than the HOMO-LUMO gap. Also, hybrid functionals within DFT that have asmall fraction of exact Hartree-Fock exchange incorporatedinto it may overestimate the gap energy. Furthermore, thecalculations in this work are for a dimer which has a higherenergy gap compared to an infinite chain. The band gap of afluorene dimer was computed as a function of the torsional anglefrom 135° to 180° to simulate the various conformations. TheC� conformation was modeled as being fully planar with atorsion angle of 180°. The important results of these calculationsare the trends and not the absolute energies of the HOMO-LUMO gap.

Pressure effects were simulated by compressing the C-Cbond between the two monomers in a single chain. By usingthe Murnaghan Equation of State, which expresses the unit cellvolume (V(P)) as a function of pressure by31

where B0 is the bulk modulus, B0′ is its derivative, and V0 is theunit cell volume at P ) 0 (optimized geometry), we estimatethe pressure. Since the monomer units are connected by C-Csingle bonds, one can use the bulk modulus of PPP moleculessuch as para-hexaphenyl (PHP). Here, we use B0 and B0′ valuesfor PHP.32 Since only one of the axes is changed in our unitcell, the ratio of the volume of the compressed chain at pressureP to the P ) 0 volume is known; using eq 1, one can determinethe corresponding pressure.

III. Steady-State Photoluminescence

The absorption spectrum of as-is PF8 (slightly aged in atoluene solution) shows the signature of the C� conformer, whichis characterized by a distinct peak at 2.8 eV. Since thisconformer has a lower energy compared to other conformations,the PL is dominated by this phase. The � phase can undergo anirreversible transformation, in terms of the local intrachainstructure, to a mixture of CR- and Cγ-type conformers afterthermal cycling (TC) to RT by heating the polymer in excessof 150 °C. Figure 1 shows the PL and absorption spectra froma thin film of PF8 (cast from toluene) under ambient pressureconditions at RT. The as-is film refers to a film cast from atoluene solution. The “after TC” spectra are from the same filmthat was slowly thermally cycled to RT from 160 °C. A clearshoulder at 2.8 eV is seen in the absorption spectrum of theas-is film, indicating the presence of the � phase.

V(P) ) V0(1 +B0′PB0

)-1/B0′(1)

Tuning Intermolecular Interactions in PF8 J. Phys. Chem. A, Vol. 114, No. 13, 2010 4681

Vibronic progressions in the PL are clearly seen in thin filmsof PF8, indicating a coupling of the backbone carbon-carbonstretch vibration to the electronic transitions. The transitionhighest in energy is the 0-0 transition, which takes placebetween the zeroth vibronic level in the excited state and thezeroth vibronic level in the ground state. The 0-1 energyinvolves the creation of one phonon. The 0-0 transition of the� conformer in PL may be up to ∼100 meV red-shiftedcompared with the CR conformer and is sensitive to the natureof the solvent.33 This is indicative also of an overall crystal-lization upon thermal cycling, which is seen even under pressure.The 2.83 eV 0-0 emission for the ambient pressure data of theas-is film (� phase) is red-shifted by almost 40 meV comparedto the TC film in our data. The slightly lower difference of the0-0 peak energies for the as-is and TC films is most likelydue to the presence of both CR and Cγ conformers after TC.

The PL studies under pressure were carried out from twosamples of PF8 loaded in DACs, one from a thin film cast fromtoluene and the other from the bulk sample. Both the film andbulk samples showed a signature of the � phase before loadingin the DAC. The bulk sample was thermally cycled at 1.9 GPaby heating it to 160 °C and bringing it slowly back to RT. Thetemperature steps used here were similar to the annealing processunder ambient pressure, where the C� conformer is known todisappear after TC.14 This is attributed to a change in thecrystalline phase,15 discussed in greater detail in the next section.The pressure was cycled up and down several times, and therewas almost no hysteresis in the PL data. In principle, TC underpressure could have been carried out at any pressure point.

Figure 2 shows the PL spectra from the thin film sample atselected values of pressure. This sample was not annealed, andpresumably under pressure, the emission is from the dominantC� conformer. The PL vibronics are seen to red shift uponenhanced pressures. The vibronic peaks smear out beyond 5.0GPa. The PL spectra in Figure 2 were measured while thepressure was increased. The hysteresis is small; upon decreasingthe pressure, the PL spectra were almost identical to the onesshown.

Figure 3 compares the PL spectra of the as-is bulk samplebefore and after it was thermally cycled for a few pressurevalues. The PL spectra of the TC sample were first measuredupon increasing the pressure until 6.5 GPa and then decreasingit down to 0.3 GPa. The PL spectra further broaden after TC.Due to the higher self-absorption in the bulk sample, the high-energy vibronic peaks are not very well-resolved. To obtain theindividual vibronic energies and peak widths, the spectra werefitted with four Gaussian peaks, as shown for the 4.0 GPa data;three of the peaks are the 0-0, 0-1, and 0-2 vibronics, and

the low-energy peak is attributed to defect or aggregate emission.Such a feature was also seen in PF2/6 under pressure.27 Beyond5 GPa, the PL spectra need to be fit with an additional low-energy peak at 2.0 eV, as shown for the 5.6 GPa data in Figure2. Most probably, this peak arises from an excimer-typeemission.

A. Discussion of PL Results. The complexity in decipheringthe optical spectra of PF8 arises from the fact that the polymermust be simultaneously assessed in terms of its crystallographicphase and the distribution of conformational isomers. PF8s castfrom various solvents appear in a metastable structure at RT.14

Cooling from the n-LC phase yields the R and R′ crystallinephases. X-ray diffraction (XRD) studies show the R′ crystallinepolymorph to be similar to the R phase, and in addition, it isexceptionally well oriented with respect to the surface normalin both ultrathin and moderately thick films.2 The progressionto the R-crystalline phase is extremely variable; PF8 films castfrom various solvents are marked by an intermediate M phaseat RT when cooled quickly. This phase is intermediate to Rand the R′ and presumably corresponds to the Cγ-type family.A detailed Raman scattering study of PF8 as a function ofthermal cycling showed that heating the polymer to 130 °Creduces the C� conformer and enhances the fraction of CRconformers.14 Upon cooling the polymer back to RT from then-LC phase, a higher fraction of the Cγ and CR conformers isseen.

Both the film and the bulk samples of PF8 in this work wereinitially dominated by the C� conformer. The red-shift of thePL vibronics and the spectra as a whole under pressure, as seenin Figures 2 and 3, are common to almost all conjugatedpolymers and molecules indicating a higher degree of effectiveconjugation. This arises due to the higher overlap between theπ -electron wave functions.

When PF8 is thermally cycled by cooling the sample slowlyfrom its n-LC phase, it changes to an overall three-dimensionalM crystalline phase that precludes the C� conformer.15 Theresulting disordered CR conformer, which has a nonplanarbackbone conformation, shows a blue-shifted 0-0 peak (Figure1). Upon thermal cycling, the PF8 sample at 1.9 GPa (indicatedby the arrow in Figure 3), the overall spectrum, and theindividual vibronics broaden. At this pressure, the differencebetween the PL vibronic energies between the thermally cycledand the as-is samples is almost 40 meV, similar to the trendseen at ambient pressure. This implies that thermal annealingof PF8 at higher pressure results in similar conformationalchanges as those seen under ambient pressure.

Figure 2. PL spectra of the as-is PF8 film at selected pressuresmeasured at room temperature. Figure 3. The right panel shows the PL spectra of the bulk sample at

selected values of pressure measured at room temperature. The leftpanel shows the PL spectra of the same sample after undergoing thermalcycling.

4682 J. Phys. Chem. A, Vol. 114, No. 13, 2010 Paudel et al.

The peak positions of the 0-0 and the 0-1 PL vibronicsfor the bulk sample before and after TC are plotted in Figure 4.The open symbols depict the 0-0 PL peak positions, and thefilled symbols represent the 0-1 peak positions. To highlightthe key features, the film data are not shown in this figure;however, the pressure coefficients of the PL vibronics from allsamples are listed in Table 1. Both as-is and annealed samplesshow a nonlinear red shift in PL energies with pressure. AfterTC, the overall PL energies are higher, confirming the conver-sion from the C� to CR/Cγ conformer. Figure 4 clearly showsthat the PL energies before and after TC do not shift in a parallelfashion with increasing pressures. This trend is clearer for the0-1 peak position, which is shown by the lightly shaded regionin the figure.

Planarization of oligophenyls has been previously observedunder pressure. A sharp decrease in the pressure coefficient isobserved in the PL vibronics of oligophenyls around ∼1.5GPa.25 Repulsion between the ortho-hydrogen atoms leads to atorsion of neighboring phenyl rings with respect to the singlebond connecting them. Application of hydrostatic pressuredecreases the torsion angle, resulting in a planarization of themolecule. Such effects are typically reversible under pressure.First-principles calculation of a PPP chain predicts a decreasein the band gap energy by almost 1 eV when the torsional anglebetween the phenyl rings changes from 50 to 0°.36 We notethere is a nomenclature difference; in the oligophenyls, 0° isreferred to being fully planarized, and in PFs, 180° is the fullyplanar conformation.

In the sample after TC, the PF8 backbone conformation isless planar (CR and Cγ). Therefore, the TC sample is expectedto exhibit a red shift due to planarization and due to an enhancedintrachain interaction leading to a higher overlap of the π wavefunctions, the former being more dominant at lower pressures,

that is, larger pressure coefficients at lower pressures. In contrast,the before TC sample should just exhibit red shifts due toenhanced intrachain effects. In other words, a CR conformerplanarizes and becomes more like C� under pressure. This isshown in the lightly shaded region in Figure 4. The energydifference between the 0-1 PL energies before and after TC islarger at lower pressures, clearly indicating the planarizationof the CR backbone torsional angle, whereby the system is beingdriven toward the C� conformation. To further understand thisbehavior, we performed model calculations of the energy gapsof a fluorene dimer by changing the torsion angles. Additionally,the Raman scattering results presented in section IV also reveala planarization of the annealed PF8 sample as a function ofpressure.

Figure 5 plots the energy difference of the 0-1 PL energiesof the annealed and as-is samples as a function of pressure.The difference decreases by almost 60 meV between 1 and 6GPa. If the impact of pressure were similar for the as-is andannealed samples, one would expect the difference in the PLenergies between the two samples to be a constant value as afunction of pressure. The uncertainties in peak positions arehigher at higher pressures, as seen in Figure 4; hence, the slightincrease of the difference energy beyond 5 GPa may be anartifact. One finds almost a one-to-one correspondence betweenthe trends of the difference in PL energy to the change in thecalculated gap energy as a function of the dihedral angle. Byfitting both curves to a second-order polynomial, the experi-mental energy difference shows a linear shift of -0.030 ( 0.003eV/GPa, and the calculated energy gap shows a linear shift of-0.055 ( 0.004 eV/°. From the resemblance of the two shifts,one may conclude a ∼10° change per GPa in the torsion angleof the annealed sample.

The inset of Figure 5 shows the calculated gap energies of afluorene dimer as a function of pressure for the three differentconformations, CR, Cγ, and C�. The torsion angle for each ofthe conformers was held constant for the various pressure points.Since the simulated pressure effects did not consider any 3Dinteractions and simply took into account the compression ofthe C-C bond, the pressure coefficients are an order ofmagnitude higher than the experimental results. We note thatthe theoretical gap energies also show a nonlinear red shift withincreasing pressures for all conformations. Experimentally, theannealed sample does not follow the theoretical phase space of

Figure 4. Peak positions of the 0-0 and 0-1 PL vibronics beforeand after TC of bulk PF8 under pressure. The open and filled symbolsrepresent the 0-0 and 0-1 peak positions, respectively.

TABLE 1: Pressure Coefficients for Backbone PL EmissionPeaks in PF8 Bulk (before and after thermal cycling) andfor the Film Samplea

sample C0-0 (meV/GPa) C0-1 (meV/GPa)

PF8 bulk (before TC) -31 ( 7 -45 ( 2PF8 film (before TC) -38 ( 4 -37 ( 8PF8 bulk (after TC) -51 ( 5 -76 ( 3PF2/6 filmb -40 ( 1 -28 ( 1

a The pressure coefficients in PF8 are determined by a nonlinearfit of the second order, given by E(P) ) E(0) + CP + C′P2. Onlythe linear components are shown here. The last row shows thepressure coefficients of the PL vibronic energies in PF2/6.b Reference 27.

Figure 5. The 0-1 PL energy difference between the as-is andannealed samples as a function of pressure. The open triangles representthe calculated energy gap of a fluorene dimer as a function of the torsion(dihedral) angle. The inset shows the calculated energy gap of fluorenedimers with CR, Cγ, and C� conformations as a function of pressure.The dash-dot line is a schematic of a plausible transition for theexperimental annealed PF8 sample.

Tuning Intermolecular Interactions in PF8 J. Phys. Chem. A, Vol. 114, No. 13, 2010 4683

the CR or Cγ conformation as a function of pressure, but rather,due to a planarization of the torsion angle, the annealed samplefollows some intermediate path schematically shown by thedash-dot line in the inset of Figure 5.

Currently, the only XRD data under pressure that exist forthe PF family (polymer/oligomer) are for crystalline fluorene,where a reversible phase transition is seen at 3.6 GPa.34 Thistransition corresponds to a π-stacking molecular arrangementwith an increase in the bulk modulus. The crystal structure ofPF8 is quite different from that of fluorene itself, but at higherpressures beyond 5 GPa, it may be similar to that of fluorenewith an enhanced π-stacking geometry, which would induce astronger interchain interaction. Additionally, theoretical calcula-tions involving 3-D interactions of PPP chains show that anenhanced interchain interaction results in a broadening of theabsorption/PL with a very small red shift of the band gap.24

Our experimental data show a large increase in the PL line widthof the TC sample beyond 5 GPa (Figure 6), indicating that atthese pressures, the predominant effect of hydrostatic pressureis an enhanced interchain interaction. PL broadening as afunction of pressure sheds light on the nature of intermolecularinteractions and originates from the differences in compress-ibility of the ground and excited electronic states in configurationcoordinate space.35

The differences in the PL line widths between the as-is andTC samples also reflect upon the crystallinity in the TC sampleversus the amorphous phase in the as-is sample. The full widthat half-maximum (fwhm) of the 0-1 PL vibronic is plotted inFigure 6; the 0-0 peak also shows a similar behavior. The as-is sample clearly shows a steady increase of PL broadening withincreasing pressures. In amorphous polymers such as MeLPPP,a similar broadening of the PL vibronics under pressure hasbeen observed,17 reflecting a strong interaction perpendicularto the chain axis. The PF8 sample that was thermally cycled,on the other hand, shows a very different dependence; onaverage, the PL vibronics do not show any broadening until 6GPa, beyond which a large increase in the fwhm is observed.No changes in the PL line widths have been observed inpolycrystalline molecules such as PHP.17 Since TC induces anoverall crystalline phase in PF8, no changes in the line widthsuntil 6 GPa indicate a lower compressibility perpendicular thechain axis. The sudden increase in the PL line widths beyond6 GPa might imply a change in the 3-D crystalline phase. It ishard to predict exactly how the crystallinity changes beyond 6GPa solely from optical measurements.

B. Pressure and Temperature Dependence of the Huang-Rhys Factor. The fine features in the absorption and emissionspectra of conjugated molecules/polymers are described byFranck-Condon (FC) coupling. The π-π* electronic transitionis accompanied by a well-resolved FC-type progression ofvibronic sub-bands. These vibronic bands are dominated bymodes representing local C-C stretching in the vicinity of1200-1600 cm-1. In the emission process, the Huang-Rhysfactor (S) corresponds to an average number of phonons thatare involved when an excited molecule relaxes to its ground-state configuration from its new equilibrium configuration inthe excited state. Assuming that the vibrational frequency isthe same for the ground and excited states and that the potentialsare perfectly parabolic, S may be experimentally determinedfrom the fractional intensity of the vibronic peaks. The relativeintensities of the features coupled by a single-phonon frequency(ω) are described by

where Itotal is the total intensity of individual transitions. I0fn isthe intensity of the transition from the 0th vibronic excited stateto the nth vibronic state of the electronic ground state. TheHuang-Rhys factor therefore corresponds to the averagenumber of phonons that are involved when the excited moleculerelaxes to its ground-state configuration from the excited state,and Spω is the relaxation energy. S may be determined fromthe fractional intensity of the vibronic peaks

I0f1, I0f2, and I0f3 refer to the intensity of the emission fromthe zeroth vibrational level excited state to the first, second,and third vibrational levels of the ground state, respectively.

The relative strengths of the vibronic transitions change withboth temperature and pressure. Since the individual vibronicsare discernible for a range of pressure values for the film sample,we determine the Huang-Rhys factor only for the PF8 filmsample as a function of both pressure and temperature. Theresults for PF8 as a function of temperature and pressure areshown in Figure 7, where the S factor was determined by usingeq 2 and the vibronic intensities beyond 0-3 were neglected.In a prior work, we have shown that the Huang-Rhys factorin small molecules and long-chain polymers decreases with

Figure 6. The fwhm (in nm) of the 0-1 PL vibronic as a function ofpressure. The open ∆ represents the as-is bulk sample, and b representsthe bulk sample after TC.

Figure 7. The Huang-Rhys factor versus temperature (left and bottomaxes) and pressure (right and top axes) for film samples of PF8.

I0fn

Itotal) e-SSn

n!(2)

S ) (I0f1 + 2I0f2 + 3I0f3 + ...)/Itotal (3)

4684 J. Phys. Chem. A, Vol. 114, No. 13, 2010 Paudel et al.

decreasing temperatures.10 Smaller conjugated molecules typi-cally show a larger value for S, which arises due to their largenormal coordinate displacement between the ground and excitedelectronic states.

At ambient pressure and RT, the S factor in PF8 is determinedas 0.4, similar to values obtained for the � phase by Khan etal.37 At 40 K (ambient pressure), the S factor decreases to ∼0.25.Upon increasing pressures, the S factor steadily increases to 1.3.Lower values of the Huang-Rhys factor at ambient conditionsimply a delocalized excited state along with a smaller geometricrelaxation that follows a transition from the excited state to theground state.

A comparison of the Huang-Rhys factors in PF8 as afunction of temperature and pressure show that they increase,albeit at different rates. Upon increasing the temperature, thesinglet excitons typically become more localized in smallerconjugated chain segments.10 This localization results in a highervalue for S with increasing temperatures. The red shift of thePL spectra upon increasing pressures is a clear sign of anenhanced effective conjugation and thus represents a higherdelocalization of the excited state. Under enhanced pressuresthe geometric relaxation of the electronic states is increased,increasing the S factor. A higher value of the S factor withenhanced pressures therefore has a different origin as comparedwith that at enhanced temperatures.

IV. Raman Scattering Studies

The Raman spectrum of PF is characterized by numerousintramolecular C-C/C-H stretch and bend modes spanningfrom 100 to 1600 cm-1. Both the vibrational frequencies andintensities determined by Raman spectroscopy are stronglyinfluenced by variations in the backbone as well as side-chainconformations. The Raman peaks in the 1600 cm-1 region arisefrom an intraring C-C stretch frequency and is best fit withtwo peaks, an overwhelmingly dominant peak at 1605 cm-1

and at least one or two weak peaks in the range of 1570-1600cm-1. These weaker peaks correspond to a breathing motion ofthe pentagon within the monomer. The Raman frequencies inthe 1250-1350 cm-1 region are associated with the backboneC-C stretch motion. Due to the strong Raman peak of diamond,the 1200 cm-1 region is not observed in a DAC (Figure 8 inset).

The low-frequency Raman peaks in the 200-700 cm-1 range,shown in Figure 8, originate from the alkyl side chains. Thisregion is thus an extremely sensitive indicator of both side-

chain composition and ordering. In perfect transplanar alkanechains, the low-frequency Raman spectrum shows intense bandsthat arise from a longitudinal accordion motion (LAM) of thechain skeleton and its overtone.38 In nonadecane, for example,LAM1 and LAM3 modes appear at 125 and 342 cm-1,respectively.39 Upon thermal cycling of PF8, the backboneconformation is nonplanar (CR/Cγ), which is a direct conse-quence of the dioctyl side chains adopting a linear conformationas an anti-anti-anti form. The LAM modes are then clearlyseen in the Raman spectrum.14,15

The ring torsion mode at 480 cm-1 is seen in all PFs withvarious side-group substitutions and is almost independent oftemperature. Since the Raman peaks around 600 cm-1 and inthe 200-400 cm-1 range are mainly from alkyl side chains,they have very different signatures for various side-chainsubstituted PFs. There are predominantly two Raman peaks inthe 600 cm-1 region; the low frequency at 620 cm-1 correspondsto a stretching motion of the bridging C atom connected to thefirst CH2 group of the alkyl chain, and the higher-frequencypeak at 633 cm-1 involves a torsional motion of the phenylrings. Shorter alkyl chains such as ethylhexyl or an octyl groupwith gauche defects in PFs are characterized only by the higher-frequency Raman peak in this region. The relative ratio of the620 and 633 cm-1 Raman peaks in PF8 further track thepresence of the C� conformation. The intensity of the 633 cm-1

peak is stronger in the presence of the � phase. It is clear fromour ambient pressure data in Figure 8 that the PF8 film as wellas the as-is bulk PF8 (not shown in the figure) had a significantfraction of C� conformation before loading the DAC.

The 735 cm-1 peak in PF corresponds to the 747 cm-1

fundamental of a biphenyl with A1 symmetry. The strongerRaman peak in the 850 cm-1 region is a combination of ringdistortion of the phenyl rings and a C-C stretch of the bridgingcarbon. The weaker Raman peaks around 900 cm -1 relate toC-C stretch-type motion of the alkyl side chains.15 Frequenciesin the 1100-1200 cm-1 region (as shown in Figure 10) arealso sensitive to side-group substitution; the 1120 and 1135 cm-1

peaks mainly arise from the C-H bending modes (either localor between phenyl units).

A. 200-1200 cm-1 Region. Figure 9 shows the ambientpressure data of a PF8 film sample before and after TC alongwith the spectra of the bulk sample after TC under pressure inthe low-frequency region. Due to the red shift of the PLspectrum under pressure, the background of the Raman spectrumis quite high in this region. We select two pressure points wherewe could clearly identify the Raman spectra in this low-

Figure 8. The Raman spectrum of an as-is PF8 film at ambient pressureand RT at low frequencies. The doublet at 600 cm-1 marked by thearrows tracks the conformational isomers. The inset shows the high-frequency region of the Raman spectrum of the PF8 film; the sharppeak at 1320 cm-1 is the diamond peak from DAC.

Figure 9. Low-frequency Raman spectra of PF8. The bottom twospectra are at ambient pressure before and after TC. The top two spectraare at elevated pressures after TC.

Tuning Intermolecular Interactions in PF8 J. Phys. Chem. A, Vol. 114, No. 13, 2010 4685

frequency range. The ambient pressure result of PF8 after TCclearly shows the appearance of a new peak at 370 cm-1, whichis a LAM mode with all anti conformation of the alkyl sidechain.14 Such a conformation of the side chains precludes the �phase, suggesting that the individual backbone conformation isCγ- or CR-type. We note that in the presence of the �conformation, LAM modes are not observed. Under pressure(2.3 GPa), there might be a shoulder at around 370 cm-1 fromsuch a LAM mode. Unfortunately, this low-frequency regionis difficult to systematically track in Raman scattering for allvalues of pressure due to the rising background.

The Raman peaks in the 1140 cm-1 region originate from aC-H in-plane bend motion along with a ring distortion. In theplanar (�) conformation of the polymer or in a monomer, thereis mainly one Raman band at 1135 cm-1 that originates fromthe terminal phenyl rings. For nonplanar conformations, thismotion splits into two or more vibrations originating frommonomer units about the center of symmetry of the moleculeand from the end rings. Upon thermal cycling the polymer fromits n-LC phase, which concomitantly induces an overall crystal-line phase2 and a reduction of the � conformer, the 1135 andthe 1172 cm-1 peaks broaden, as seen in the ambient pressuredata in Figure 10a (two lower spectra). At enhanced pressures,there is a broadening of the 1135 cm -1 peak;28 however, unlikeambient pressure, TC does not induce a further broadening ofthis peak at high pressures. The pressure coefficients of the 1135cm-1 Raman peak are 1.6 and 2.9 cm-1/GPa for the as-is andthermally cycled samples, respectively (Figure 10b). Thesepressure coefficients are lower than those of the high-frequency1600 cm-1 Raman peak, as discussed in the next section.

B. 1600 cm-1 Region. A signature of the various confor-mational isomers is better deciphered by the backbone C-Cstretch modes in the 1300 cm-1 region. Unfortunately, underpressure in a DAC, this region is swamped by the Raman peakof the diamond. However, the intraring C-C stretch peak at1605 cm-1 is clearly seen at all pressures for both as-is and TCsamples. We present a detailed analysis of this peak as a functionof pressure in this section.

Since the origin of the 1600 cm-1 Raman peak is fromintraring C-C stretch motion, it is not sensitive to the variouscrystalline phases or conformations. Figure 11 shows a fewselected spectra of the 1600 cm-1 region of the as-is bulk sampleunder pressure; for a comparison, the 3.1 GPa spectrum of theTC sample is also shown. The Raman background increaseswith pressure with a distinct asymmetry of the 1605 cm-1 peak,as shown in Figure 11. Such an asymmetry was observed inPF2/6 under pressure and was attributed to a strong electron-

phonon interaction between the Raman peaks and the electroniccontinuum.27 This is characteristic of a Breit Wigner Fano(BWF) resonance; in PF2/6, many of the vibrational peaks alsoshowed an antiresonance behavior. To determine the peakposition, asymmetry parameter, and line width as a function ofpressure, we fit the 1605 cm-1 peak with a BWF line shapegiven by

where ω0 is the discrete phonon frequency and Γ is the widthof the resonant interference between the continuum and discretescattering channels. The asymmetry parameter (1/q) dependson the average electron-phonon matrix element, M, and theRaman matrix elements between the ground and excited statesof the phonon and electron. The broadening parameter is givenby Γ ) πM2D(ω), where D(ω) is the combined density of statesfor the electronic transitions.40

Figure 12a shows the peak position of the 1605 cm-1 Ramanpeak as a function of pressure. The pressure coefficient of thebulk PF8 sample before and after TC is almost the same at 7.2cm-1/GPa, which is higher than that observed for the 1135 cm-1

Raman peak. Pressure-induced changes in the vibrationalfrequencies is a measure of the Gruneisen parameter in thepotential energy surface of the particular vibrational coordinate.For bulk inorganic solids, the mode Gruneisen parameter isusually volume-independent. In molecular solids, due to the

Figure 10. (a) The Raman spectra in the 1100 cm-1 region at ambientpressure and higher pressures before and after TC. (b) Peak positionof the 1135 cm-1 Raman peak before and after TC as a function ofpressure.

Figure 11. Raman spectra of the 1600 cm-1 region (before TC) inPF8 at selected value of pressures. The spectrum at 3.1 GPa is fromthe PF8 sample after TC.

Figure 12. (a) Position of the 1605 cm-1 Raman peak, obtained usinga BWF fit, as a function of pressure for both the as-is and TC PF8bulk sample. The inset shows the line width of the same peak as afunction of pressure. (b) Asymmetry parameter (1/q) of the 1605 cm-1

Raman peak versus pressure; 1/q is obtained by fits to the Raman peakwith a BWF line shape (eq 4).

I(ω) ) I0

[(ω - ω0)/q + Γ]2

(ω - ω0)2 + Γ2

(4)

4686 J. Phys. Chem. A, Vol. 114, No. 13, 2010 Paudel et al.

differences between the local volume compression relative tothat of the bulk, a volume-dependent Grüneisen parameter isoften observed.41 It is therefore not surprising that the pressurecoefficients of the 1135 and 1600 cm-1 Raman frequencies aredifferent. We note that the mode Gruneisen parameter, whichis given by (1/ω)dω/dp, is higher for the 1600 cm-1 peakcompared to that of the 1135 cm-1 peak.

The inset of Figure 12a shows the line width to increasealmost linearly with pressure for samples before and after TC.The asymmetry parameter (1/q) from a BWF fit also increaseslinearly with pressure, as seen in Figure 12b. Since the 1600cm-1 peaks originate from an intraring C-C stretch motion,thermal cycling and thus the changes in the induced phases havealmost no impact on this Raman peak. The sign of theasymmetry parameter in a BWF resonance, which arises froman interaction of the electronic continuum, is an indicator ofthe energy of the continuum. The negative value of q in the fitsshow that the center of the electronic continuum lies below thediscrete mode frequency of 1605 cm-1 (0.2 eV).

Since the vibrational frequencies of a harmonic solid areindependent of compression, pressure-induced changes in theRaman spectrum provide insight into the anharmonicity of thesolid-state potential.42 The linear shift of the 1605 cm-1 Ramanpeak is a result of such an anharmonicity of the potential.Although one expects an additional shift of the phononfrequencies due to the BWF interaction, it is not possible toextract this information from our data due to the absence of asimilar defect-free polymer. The electronic continuum may arisefrom defect states such as the fluorenone defects or excimericstates. We see subtle differences between PF8 and PF2/6; thelatter showed a higher asymmetry parameter and a square lawdependence of the line width as a function of pressure.27

V. Summary and Prospect

A comparison of the photoluminescence spectra of as-is andthermally cycled PF8 under pressure provides an ideal tool forisolating the different ingredients that impact the shift of theoptical energy gap. Furthermore, such studies from crystallineand amorphous components of the polymer shed light on inter-and intrachain interactions. Both samples show a red shift ofthe PL energies upon increasing pressures, but the annealedsample shows a higher pressure coefficient. This is interpretedas planarization of the torsional angle along with an enhancedintrachain interaction upon increasing the pressure.

By comparing the PL energy difference of as-is and annealedsamples with pressure to the changes in the theoretical energygap of a fluorene dimer by varying the backbone torsion angle,we infer that the CR and Cγ conformations in the annealedsamples are driven to the more planar C� form. Beyond 5 GPa,both the as-is and annealed samples behave in a similar fashion,indicating that the PL energies are dominated by the C�

conformations. It is hard to decipher the exact fraction of C�

conformers in these experiments. We further point out that thesetransitions are reversible; decreasing the pressure of the annealedsample shows a PL spectrum that may be attributed to CR orCγ conformers.

The line width of the PL vibronics as a function of pressureshows differences in the as-is and annealed PF8 samples. Theas-is sample shows a continuous broadening of the PL vibronicsas a function of pressure. Such a dependence is seen in otheramorphous polymers and signals an enhanced interchain interac-tion. The annealed sample, which is in one of the crystallinephases, shows almost no change in the PL line widths until ∼6GPa. These results are similar to crystalline oligophenyls where

there is hardly any change of PL line widths with increasingpressures. The main impact of pressure at this range is a changein the backbone conformation of the polymer.

The Raman frequencies of PF8 harden with pressure. Due tolimitations in a DAC, only the 1135 and 1600 cm-1 Ramanpeaks could be tracked as a function of pressure. Theseintramolecular vibrations are not the best candidates for unravel-ing changes in the conformations or crystalline phases in PF8.The 1600 cm-1 shows asymmetric line shapes characteristic ofa BWF resonance, signaling a strong electron-phonon interac-tion between the Raman phonons and the electronic continuum.The origin of the enhanced Raman line widths is due to suchelectron-phonon interactions.

In light of our experimental results on the optical propertiesof PF8 under pressure, where a rich change in the phasemorphology is observed, future XRD measurements underpressure would be invaluable to map the conformational andcrystalline changes in both PF8 and PF2/6.

Acknowledgment. This work was partly supported throughthe National Science Foundation under Grant No. ECCS-0823563. We thank M. Arif for the ambient pressure Ramandata.

References and Notes

(1) Grell, M.; Bradley, D. D. C.; Ungar, G.; Hill, J.; Whitehead, K. S.Macromolecules 1999, 32, 5810.

(2) Chen, S. H.; Su, A. C.; Su, C. H.; Chen, S. A. Macromolecules2005, 38, 379.

(3) Knaapila, M.; Lyons, B. P.; Hase, T. P. A.; Pearson, C.; Petty,M. C.; Bouchenoire, L.; Thomspon, P.; Serimaa, R.; Torkkeli, M.;Monkman, A. P. AdV. Funct. Mater. 2005, 15, 1517.

(4) Tanto, B.; Guha, S.; Martin, C. C.; Scherf, U.; Winokur, M. J.Macromolecules 2004, 37, 9438.

(5) Rahman, M. H.; Chen, C.-Y.; Liao, S.-C.; Chen, H.-L.; Tsao, C.-S.; Chen, J.-H.; Liao, J-L; Ivanov, V. A.; Chen, S.-A. Macromolecules 2007,40, 6572.

(6) Knaapila, M.; Garamus, V. M.; Dias, F. B.; Almasy, L.; Galbrecht,F.; Charas, A.; Morgado, J.; Burrows, H. D.; Scherf, U.; Monkman, A. P.Macromolecules 2006, 39, 6505.

(7) Cadby, A. J.; Lane, P. A.; Mellor, H.; Martin, S. J.; Grell, M.; C.Giebeler, C.; Bradley, D. D. C.; Wohlgenannt, M.; An, C.; Vardeny, Z. V.Phys. ReV. B 2000, 62, 15604.

(8) Winokur, M. J.; Slinker, J.; Huber, D. L. Phys. ReV. B 2003, 67,184106.

(9) Liem, H.; Etchegoin, P.; Whitehead, K. S.; Bradley, D. D. C.J. Appl. Phys. 2002, 92, 1154.

(10) Guha, S.; Rice, J. D.; Yau, Y. T.; Martin, C. M.; Chandrasekhar,M.; Chandrasekhar, H. R.; Guentner, R.; de Freitas, P. S.; Scherf, U. Phys.ReV. B 2003, 67, 125204.

(11) Lieser, G.; Oda, M.; Miteva, T.; Meisel, A.; Nothofer, H. G.; Scherf,U. Macromolecules 2003, 33, 4490.

(12) Chunwaschirasiri, W.; Tanto, B.; Huber, D. L.; Winokur, M. J.Phys. ReV. Lett. 2005, 94, 107402.

(13) Chen, S. H.; Chou, H. L.; Su, A. C.; Chen, S. Macromolecules2004, 37, 6833.

(14) Arif, M.; Volz, C.; Guha, S. Phys. ReV. Lett. 2006, 96, 025503.(15) Volz, C.; Arif, M.; Guha, S. J. Chem. Phys. 2007, 126, 064905.(16) Tsoi, W. C.; Lidzey, D. G J. Phys.: Condens. Matter 2008, 20,

125213.(17) Guha, S.; Graupner, W.; Yang, S.; Chandrasekhar, M. ; Chan-

drasekhar, H. R. Approaches to Polar Order; Glaser, R., Kaszynski, P.,Eds.; ACS Symposium Series; American Chemical Society: Washington,DC, 2001; Vol. 798, p 127.

(18) Gangilenka, V. R.; DeSilva, A.; Wagner, H. P.; Tallman, R. E.;Weinstein, B. A.; Scholz, R. Phys. ReV. B 2008, 77, 115206.

(19) Chandrasekhar, M.; Guha, S.; Graupner, W. AdV. Mater. 2001, 13,613.

(20) Webster, S.; Batchelder, D. N. Polymer 1996, 37, 4961.(21) Yang, G; Li, Y.; White, J. O.; Drickamer, H. G. J. Phys. Chem. B

1999, 103, 7853.(22) Morandi, V.; Galli, M.; Marabelli, F.; Comoretto, D. Phys. ReV. B

2009, 79, 045202.(23) Schmidtke, J. P.; Friend, R. H.; Silva, C. Phys. ReV. Lett. 2008,

100, 157401.

Tuning Intermolecular Interactions in PF8 J. Phys. Chem. A, Vol. 114, No. 13, 2010 4687

(24) Yang, S. C.; Graupner, W.; Guha, S.; Puschnig, P.; Martin, C.;Chandrasekhar, H. R.; Chandrasekhar, M.; Leising, G.; Ambrosch-Draxl,C. Phys. ReV. Lett. 2000, 85, 2388.

(25) Guha, S.; Graupner, W.; Resel, R.; Chandrasekhar, M; Chan-drasekhar, H. R.; Glaser, R.; Leising, G. Phys. ReV. Lett. 1999, 82, 3625.

(26) Martin, C. M.; Cai, Q.; Guha, S.; Graupner, W.; Chandrasekhar,M.; Chandrasekhar, H. R. Phys. Status Solidi B 2004, 241, 3339.

(27) Martin, C. M.; Guha, S.; Chandrasekhar, M.; Chandrasekhar, H. R.;Guentner, R.; de Freitas, P. S.; Scherf, U. Phys. ReV. B 2003, 68, 115203.

(28) Paudel, K.; Arif, M.; Chandrasekhar, M.; Guha, S. Phys. StatusSolidi B 2009, 246, 563.

(29) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03;Gaussian, Inc.: Pittsburgh, PA, 2003.

(30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.(31) Murnaghan, F. D. Am. J. Math. 1937, 59, 235.(32) Heimel, G.; Puschnig, P.; Oehzelt, M.; Hummer, K.; Koppelhuber-

Bitschnau, B.; Porsch, F.; Ambrosch-Draxl, C.; Resel, R. Mater. Res. Soc.Symp. Proc. 2003, 771, L7.22.1.

(33) Cheun, H.; Tanto, B.; Chunwaschirasiri, W.; Larson, B.; Winokur,M. J. Appl. Phys. Lett. 2004, 84, 22.

(34) Heimel, G.; Hummer, K.; Ambrosch-Draxl, C.; Chunwachirasiri,W.; Winokur, M. J.; Hanfland, M.; Oehzelt, M.; Aichholzer, A.; Resel, R.Phys. ReV. B 2006, 73, 024109.

(35) Drickamer, H. G.; Frank, C. W. Electronic Transitions and the HighPressure Chemistry and Physics of Solids; Chapman and Hall, Ltd.: NewYork, 1973.

(36) Ambrosch-Draxl, C.; Majewski, J. A.; Vogl, P.; Leising, G. Phys.ReV. B 1995, 51, 9668.

(37) Khan, A. L. T.; Sreearunothai, P.; Herz, L. M.; Banach, M. J.;Kohler, A. Phys. ReV. B 2004, 69, 085201.

(38) Schaufele, R.; Shimanouchi, T. J. Chem. Phys. 1967, 47, 3605.(39) Zerbi, G.; Magni, R.; Gussoni, M.; Mortiz, K. H.; Bigotto, A.;

Dirlikov, S. J. Chem. Phys. 1981, 75, 3175.(40) Klein, M. V. Light Scattering in Solids; Cardona, M., Ed.; Springer-

Verlag: New York, 1975.(41) Zallen, R.; Slade, M. L. Phys. ReV. B 1978, 18, 5775.(42) Ferraro, J. R. Vibrational Spectroscopy at High External Pressures;

Academic Press: New York, 1984.

JP911778R

4688 J. Phys. Chem. A, Vol. 114, No. 13, 2010 Paudel et al.