J. Phys. Chem. B 2008, 112, 16415–16421
16415
Fractal Aggregates of Polyfluorene-Polyaniline Triblock Copolymer in Solution State Matti Knaapila,*,† Vasil M. Garamus,‡ La´szlo´ Alma´sy,§,| Jing S. Pang,⊥ Michael Forster,# Andrea Gutacker,# Ullrich Scherf,# and Andrew P. Monkman⊥ Department of Physics, Institute for Energy Technology, NO-2027 Kjeller, Norway; GKSS Research Centre, DE-21502 Geesthacht, Germany; Research Institute for Solid State Physics and Optics, Budapest-1525, Hungary; Laboratory for Neutron Scattering, ETHZ & PSI, CH-5232 Villigen, Switzerland; Department of Physics, UniVersity of Durham, DH1 3LE Durham, England; and Fachbereich Chemie, Bergische UniVersita¨t Wuppertal, DE-42097 Wuppertal, Germany ReceiVed: July 30, 2008; ReVised Manuscript ReceiVed: October 2, 2008
We report on the aggregate structure of a symmetrical A-B-A 9,9-dialkylfluorene/2-alkylaniline “coil-rod-coil” triblock copolymer (or PF/PANI11112-b-PANI11)sconsisting of 2-dodecylanilines as A blocks and 9,9di(3,7,11-trimethyldodecyl)fluorene)s as B blockssmixed in deuterated toluene, chloroform, and methylcyclohexane. These mixtures contain micrometer scale aggregates. Small-angle neutron scattering data indicate that the interface between the aggregates and solvents manifests surface fractal-like structure. The upper and lower limit length scales which show fractal character are of the order of >100 and 15 nm. The surface fractal dimension Ds varies from 2.2 to 2.8. The data show that stiff polyfluorene (PF) segments (persistence length g9.5 nm, diameter ∼ 2 nm) of PF/PANI11112-b-PANI11 are dissolved down to the molecular level. Photoluminescence data indicate that most PF units are isolated from both polyaniline (PANi) units and each others. Together with the scattering data, this implies that the disordered interface consists of stiff isolated PF blocks linked together via domains of associated PANi blocks. Introduction self-organization1
Phase behavior and represent central facets in the physics of polyfluorenes (PFs).2 Besides intramolecular details, the intermolecular assemblies and nano- and microstructures of PFs have become increasingly important. The preparation strategies of polyfluorene nanostructures include topdown methods such as electrospinning3 and microporous templates4 and bottom-up methods such as synthesis of blends5 and block copolymers.6 Recently, Scherf and co-workers introduced a series of PF block copolymers with polyaniline (PANi)7 and polythiophene8 blocks. Essential to this class of polymers are not only competing interactions in bulk but the segregation controlled by solvent, leading to important photophysical implications such as selective tuning of the photoluminescence quenching. It is expected that this emerging field would benefit from further structural studies. Elsewhere, solution structures of PFs have been studied in diverse conditions, which have given us two apparent perspectives. First, PFs manifest hierarchy from single molecules9 to intermolecular assemblies10 and all the way to micrometer scale networks.11 Second, solution structures are conveniently designed by the parameters such as the nature of backbone,12 side chain,13 or solvent.14 Despite these efforts, PF block copolymer solutions have not been studied from a structural perspective. The loose structures adopted by the swelled polymers in polymer-solvent systems15 may be interpreted as random * Corresponding author: Tel +47-6380-6081; Fax +47-6381-0920; e-mail
[email protected]. † Institute for Energy Technology. ‡ GKSS Research Centre. § Research Institute for Solid State Physics and Optics. | ETHZ & PSI. ⊥ University of Durham. # Bergische Universita ¨ t Wuppertal.
fractals with a lower limit value (inner cutoff) and an upper limit value (outer cutoff). Direct imaging has often been used to track fractal-like structures in diverse polymer systems in the micron-submicron scale.16 As understood in the scaling of small-angle scattering,17 fractals are reported for submicronnanometer scale dispersions of polymers,18 block copolymers,19 and conductive organic materials such as carbon nanotubes20 or carbon black.21 When limiting our discussion to π-conjugated polymers, the landmark report is the one of Chen et al.22 By using small-angle neutron scattering (SANS), these authors demonstrated that poly(2,3-diphenyl-5-hexyl-1,4-phenylenevinylene) (DP6-PPV) adopts network aggregates with mass fractal dimensions of 2.2-2.7. In solutions, PFs can form sheetlike aggregates10 or large network-like aggregates.11 However, whether PF aggregates can be interpreted as fractals remains unclear. In this brief paper we report on the dispersions of symmetric A-B-A triblock copolymer with poly(2-dodecylaniline) as A blocks and poly(9,9-di(3,7,11-trimethyldodecyl)fluorene)s as B blocks (PF/PANI11112-b-PANI11). This study aims at providing first details on the dispersion structure of PF/PANI11112b-PANI11. Moreover, we aim at illustrating whether such a system can be interpreted in terms of fractals using the scaling behavior of small-angle scattering.17 We find that PF/PANI11112b-PANI11 forms micrometer scale aggregates when dispersed in deuterated toluene (toluene-d8), chloroform (CDCl3), or methylcyclohexane (MCH-d14). The aggregates have a compact core, and the interfaces between the aggregates and matrices can be interpreted as surface fractals with the outer cutoff more than 100 nm and the inner cutoff of 15 nm. The surface fractal dimensions vary from 2.2 to 2.8. Below this length scale the polymer forms a network where PF blocks are fully dissolved and where links are formed between associated PANi blocks. This study implies that PFs can form fractal aggregates. In
10.1021/jp806763d CCC: $40.75 2008 American Chemical Society Published on Web 11/18/2008
16416 J. Phys. Chem. B, Vol. 112, No. 51, 2008
Knaapila et al.
CHART 1: Chemical Structure of Symmetric PANi-PF-PANi Triblock Copolymer (PF/ PANI11112-b-PANI11) (x ) 23, y ) 15)
Figure 1. Optical micrograph of PF/PANI11112-b-PANI11 in 10 mg/ mL toluene-d8. T ) 25 °C.
particular, while Chen et al.22 demonstrated mass fractals for a π-conjugated polymer, our results illustrate surface fractals for π-conjugated block copolymer. Also, due to the triblock structure, the reported aggregates differ clearly from molecular micelles of diblock copolymers. Experimental Section Materials. The synthesis of PF/PANI11112-b-PANI11 triblock copolymer (Chart 1) has been described elsewhere.7 PF block has branched alkyl side chains, whereas PANi blocks have linear alkyl side chains. In short, the synthetic route involves the aryl-aryl coupling of a 2,7-dibromo-9,9-dialkylfluorene in the presence of unprotected 4-bromoaniline according to Yamamoto followed by the oxidative coupling of the NH2-terminated prepolymer with the AB-type bifunctional monomer 2-undecylaniline. The molecular weight of the PF prepolymer is controlled by the feed ratio of bifunctional dibromofluorene/4bromoaniline. For the employed triblock copolymer the numberaveraged and weight-averaged molecular weights were respectively Mn ) 31.9 kg/mol and Mw ) 39.7 kg/mol. The degree of polymerization was ∼23 for the PF block and ∼15 for the two terminal PANi blocks. For neutron scattering and photoluminescence the polymer was mixed with deuterated toluene-d8 (99.5% D, Cambridge Isotope Laboratories Inc.), CDCl3 (99.8% D, Apollo Scientific Ltd.), or MCH-d14 (99.5% D, Apollo Scientific Ltd.) and stirred for 24 h. The employed concentrations ranged from 1 to 10 mg/mL (from ∼0.1 to 1 wt %). Instrumentation and Methods. Optical microscopy was performed using a Nikon 115 optical microscope. SANS measurements were performed using both the Yellow Submarine instrument at the BNC in Budapest (Hungary)23 and the SANS-1 instrument at the GKSS Research Centre in Geesthacht (Germany).24 The overall q range was from 0.004 to 0.4 Å-1. The samples were filled in Hellma quartz cells of 2 mm path length and placed in a thermostatic holder. The raw scattering patterns were corrected for sample transmission, room background, and sample cell scattering. The isotropic 2-dimensional scattering patterns were azimuthally averaged, converted to an absolute scale, and corrected for detector efficiency dividing by the incoherent scattering spectra of 1 mm thick pure water. The scattering from deuterated solvents used for the sample preparation was subtracted as a background; the small incoherent scattering due to the nondeuterated polymer was taken into account by fitting procedure. The scattering functions were interpreted using scaling concepts. The simple interpretation was enhanced by numerical modeling to certain geometric shapes. The data were analyzed further by indirect Fourier transforms (IFT) by Glatter25 in approximation of rodlike objects and trace scattering from large
aggregates. The large q part q > 0.02 Å-1 was additionally fitted by model of stiff cylinder,26 and so the length of cylinder and its radius of cross section were obtained. Optical absorption measurements were performed using a Perkin-Elmer Lambda 19 spectrometer. Photoluminescence (PL) measurements of dispersions were performed using a JobinYvon Fluoromax fluorimeter and quartz cell at 20 ( 1 °C. The excitation wavelength was 380 nm, representing the maximum of the excitation profile. Results and Discussion Major examples of fractal π-conjugated polymer, DP6-PPV, are reported for toluene and chloroform mixtures.22 Nanostructured PF/PANI11112-b-PANI11 films have been processed also from toluene and chloroform.7 Elsewhere, it has been shown in turn that phase behavior of PF solutions can be altered when moving from toluene to MCH.14 In order to make straightforward comparison to the earlier reports, these three solvent matrices were selected for the present study. When ∼0.1-1% PF/PANI11112-b-PANI11 is mixed with toluene-d8, CDCl3, or MCH-d14, a visually black dispersion is formed. Material is easily dispersed at room temperature, but it is not dissolved down to the microscopic level and at least part of material remains as needlelike particles whose thickness and lengths are respectively of the order of 1 and 10 µm. As an example, Figure 1 shows an optical micrograph of PF/PANI11112-b-PANI11 in 10 mg/mL toluene-d8. The fact that dispersion, not dissolution, is formed arises two aspects. First, as all the solvents used here are good solvents for PF block but too nonpolar for PANi block (see discussion in ref 27), aggregate formation is expected to stem from the association of PANi blocks. Second, as it is not clear whether a part of the material is dissolved down to the microscopic level, the dispersions were studied both after mixing and after filtration (vide infra). Attention was next placed on the potential fractal nature of PF/PANI11112-b-PANI11. The fractal structure17 is manifested by a self-similarity so that the smallest fractal units (sometimes called as monomers) builds up larger structures. In real material the fractal pattern repeats itself over a limited length scale L, and the lowest and highest limits of the fractal structure are characterized by the radius of monomer (r0) and the size of fractal aggregate (Lm). These two parameters can be estimated using the position of the upper (qmax) and the lower limit (qmin) of the power-law region of the scattering curve.17 The estimated limit of the length scales, L, for which the structure is fractal can be defined as 1/qmax < L < π/qmin. One should take into account that the observed q interval for fractal behavior may be limited by the lowest q value of the instrument and the
Fractals of Polyfluorene Block Copolymer
J. Phys. Chem. B, Vol. 112, No. 51, 2008 16417
influence of nonfractal particles. Therefore, the actual cutoffs for the fractal behavior can be wider than experimentally observed. An evidence for fractals in the submicron and nanometer scale can be conveniently derived from small-angle scattering based on well-known dimensional analysis.17 The power law of the scattering intensity I(q) can be described as
I(q) ∼ q-R
(1)
where the exponent R indicates whether the microscopic structure of the scatterer can be understood as mass fractals or surface fractals. When the angular coefficient R of the log I(q) vs log q plot is determined, it correlates to the dimensions of mass and surface fractals, Dv and Ds, as
R ) 2Dv - Ds
(2)
Figure 2. SANS data of PF/PANI11112-b-PANI11 mixed in toluened8. The concentrations were 10 mg/mL (open circles) and 1 mg/mL (solid circles). Solid red lines show the fractal (q < 0.02 Å-1) and rigidrod (q > 0.02 Å-1) models to the data. Dotted lines show -1, -3, and -4 decays for comparison. T ) 25 °C.
If 1 < R < 3, the scattering curve can be associated with mass fractal material. In these systems, which are often referred as volume or bulky fractals, the surface and volume have the same fractal dimension, i.e., Dv ) Ds. The volume (V) and the surface area (S) of the particle are related to the dimensions Dv and Ds as
V(r) ∼ rDv
(3)
S(r) ∼ rDs
(4)
and
where r is the radius of the scatterer. If 3 < R < 4, the scattering curve can be associated with surface fractal material. Such particles have a dense core and rough surface. The core has an Euclidean dimension, Dv ) 3, whereas the surface obey a relation Ds ) 6 - R. If R ) 4, the material is not fractal but has a uniform, dense core with perfectly smooth surface. This relates to the dimensions as Dv ) 3 and Ds ) 2. As neutron scattering is sensitive to the contrast between matrix and polymer, essentially no scattering would arise from the domains in bulk PF/PANI11112-b-PANI11. Black dispersions would be impractical to study by light scattering, and thus SANS was selected for our study. Yet SANS experiments have certain limitations. The data may indicate that in the considered q range the material is dominated by a certain fractal type. However, as scattering gives an average picture of the material, the coexistence of different kinds of fractals cannot be excluded. Furthermore, polydispersity of aggregates can affect the obtained values of fractal dimension. In present case, the number distribution of aggregates was unknown and the polydispersity corrections could not be done. Figure 2 plots SANS data of PF/PANI11112-b-PANI11 in toluene-d8 in 10 mg/mL (∼1%) and 1 mg/mL (∼0.1%) mixtures as well as the fits to the IFT models with assumption of the trace contributions of large fractal objects and rodlike particles. Note that there are two models for each data set. The data for both concentrations superimpose for q < 0.1 Å-1. At low q (0.02 Å-1) all the data follows ∼q-1 decay. A slight downturn is seen for higher concentration for >0.1 Å-1. Figure 3 plots SANS data of PF/PANI11112-b-PANI11 in 10 mg/mL CDCl3 with
Figure 3. SANS data of PF/PANI11112-b-PANI11 mixed in CDCl3 (open diamonds). The concentration was 10 mg/mL. Solid red line show the rigid-rod (q > 0.02 Å-1) model to the data. Dotted lines show -1, -3, and -4 decays for comparison. T ) 25 °C.
Figure 4. SANS data of PF/PANI11112-b-PANI11 mixed in MCHd14. The concentrations were 10 mg/mL (open squares) and 1 mg/mL (solid squares). Solid red lines show the fractal (q < 0.02 Å-1) and rigid-rod (q > 0.02 Å-1) models to the data. Dotted lines show -1, -3, and -4 decays for comparison. T ) 25 °C.
appropriate IFT fits. The data are similar to those shown in Figure 2. Figure 4 shows SANS data of PF/PANI11112-bPANI11 in 10 and 1 mg/mL MCH-d14. When the concentration is 1 mg/mL, MCH-d14 dispersion PF/PANI11112-b-PANI11 follows the characteristics of toluene-d8 and CDCl3 dispersions (cf. Figures 2 and 3). However, when the concentration is raised up to 10 mg/mL, the low-q regime (
