Controlling Optical Properties and Interchain Interactions in


Sep 2, 2005 - The photophysics of MEH−PPV incorporated into the pores of periodic silica hosts has been investigated in an effort to understand the ...
0 downloads 11 Views 250KB Size


J. Phys. Chem. B 2005, 109, 17879-17886

17879

Controlling Optical Properties and Interchain Interactions in Semiconducting Polymers by Encapsulation in Periodic Nanoporous Silicas with Different Pore Sizes Ashley J. Cadby and Sarah H. Tolbert* Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles, Los Angeles, California 90095-1569 ReceiVed: July 5, 2005

The photophysics of MEH-PPV incorporated into the pores of periodic silica hosts has been investigated in an effort to understand the role played by interchain aggregation and chain morphology in polaron production. In this work, guest/host interactions were used to incorporate MEH-PPV into the straight, homogeneous pores of hexagonal surfactant- or polymer-templated mesoporous silicas of varying pore diameters. Polarized photoluminescence and photoluminescence excitation spectroscopy were then used to investigate the polymers’ environment within the silica pores. Experiments exploiting luminescence peak shifts and depolarization indicate that depending on the pore size and preparation conditions, the alignment and packing of the polymer chains within the pores could be controlled. Samples could be produced with isolated chains, interacting straight chains, and coiled interacting chains. The sub-bandgap absorption by polarons was then measured with photoinduced absorption as a function of pore size. Small-diameter pores that allowed single polymer chains to reside within the pore showed little evidence of interchain contact and had a low polaron yield. Increasing the number of polymer chains within the pore increased the polaron yield. Finally, when the pores were large enough that the chains could coil, strong polaron absorption was observed, indicative of a further increase in polaron yield or an increase in polaron lifetime. The polaron absorption spectra also sharpen and red shift with increasing pore diameter, suggesting that excitons may migrate to lower energy polymer segments in samples where polymer chains are both coiled and interacting.

Introduction Polarons are believed to play an important role in the photophysics of semiconducting polymers and polymer devices. For example, they are considered to be the dominant free carriers in MEH-PPV.1,2 After photoexcitation, primary excitations can relax via photoluminescence and/or internal conversion3 or can produce nonemissive species such as triplets or polarons. Triplet excitons can be produced by intersystem crossing from the singlet manifold to the triplet manifold.4 Alternatively the exciton can transfer an electron to a neighboring chain or the next conjugated segment, thus forming a charge-transfer interchain or intrachain exciton.5 Separation of the exciton further than the Onsager radius leads to the formation of free charge carriers (polarons).6 Strong polaron absorption bands are seen in many conjugated polymers, an indication that polarons are intrinsic to polymer semiconductors.7 Controlling the spatial and energetic location of polarons as well as their production method is very important for the development of polymer devices, including polymer photovoltaics, polymer light emmiting diodes (PLED’s), and lasing materials.8 Polarons have been shown to cause photoluminescence (PL) quenching, thus reducing the PL efficiency,9 and as such their interaction with singlet excitons should be limited for high-efficiency luminescent devices. However, for photovoltaic devices and PLEDs the manipulation of polarons and handling of currents within the devices are important10 so that the charged species are used efficiently. The production mechanism for polaron pairs suggested above requires the separation of an exciton onto separate conjugated sections, either on the same polymer chain or on separate chains. * To whom correspondence should be addressed.

Two important questions to be answered are thus: Is the polaron production method strongly influenced by chain aggregation? And, can polaron production be controlled? Another way to ask this question is perhapssis there a difference between interand intrachain polaron formation efficiencies? We note, however, that some form of chain aggregation is generally required in polymer devices for transport. For example, nanoscale charge transport has been measured in MEH-PPV devices and shows that an increase in the local density of aggregation in a polymer device provides better local conductivity.11,12 Given this requirement, it is desirable to understand how the conformation of the polymer chains, in addition to the degree of aggregation, controls polaron production. The conformation of conjugated polymers has been extensively studied and much insight has been gained into the effects of chain morphology on polymer photophysics.13 Although a great deal is now understood concerning structure/electronic property relationships in these materials, it is still difficult to understand the effects of morphology on polaron production because controlling the degree of aggregation in polymer films is not straightforward. Chain aggregation can be controlled to a limited extent by the choice of solvent used in the spin coating process12,14 or by thermally annealing the entire sample,15,16 but these processes cannot be used to precisely control the local chain conformation. In this study we thus use guest/host chemistry in highly regular, nanoporous materials to control the number of polymer chains that aggregate and their conformation. This allows us to investigate the effect aggregation has on various aspects of the polymer photophysics as we increase the number of interacting polymers and change their conformation. We use solution-phase self-assembly to form regular periodic porous silica with pore

10.1021/jp0536753 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/02/2005

17880 J. Phys. Chem. B, Vol. 109, No. 38, 2005 diameters in the nanoscale size regime.17-19 These pores are then diffusively filled with polymer to either separate the polymer chains or selectively aggregate them in a well-defined manner.20-22 Because of the small size of the pores used in this study, the pores can also be used to control the possible conformations of polymer chains. Beyond allowing us to study the effects of aggregation and polymer chain conformation on polaron production, these experiments also allow us to develop methods for combining the properties of semiconducting polymers with the advantages of “bottom up design” of optical materials based on nanoscale molecular self-assembly.23-25 Because the mobility of organic semiconductors is lower than that of their inorganic counterparts, methods for controlling charge transfer in polymer semiconductors are vitally important. The mobility of a polymer depends on chemical structure and microscopic conformation of the chains,26 and these host/guest composite materials provide a unique avenue to control that polymer conformation. Experimental Section Three mesoporous silica samples were synthesized with differing pore diameters. The three different pore sizes used were created using the two different methods outlined below. The smallest pore samples were produced by surfactant templating of silica,17 while the medium and large pore samples were produced using amphiphilic triblock copolymers as templates.27,28 Method for the Synthesis of Small Pore Silicas. The smallest diameter mesoporous samples were prepared following the method given in ref 29. In the synthesis of these materials, 0.8 g of cetyltrimethylammonium bromide (CTAB) was dissolved in 39.2 g of deionized water and 5.0 g of 2.0 M NaOH. The solution was stirred at 50 °C for several hours until the CTAB was fully dissolved. The solution was allowed to cool to room temperature and 3.85 g of tetraethyl orthosilicate (TEOS) was added. The solution was then stirred at room temperature for 3 h. The precipitate was recovered by filtration and the product was washed with acetone and ethanol to remove residual surfactant before being dried at 100 °C for 1 h. Method for the Synthesis of Medium and Large Pore Silicas. The medium and large pore samples were prepared following the method set out in refs 27 and 28. In the synthesis of these materials, 4.0 g of the triblock copolyner poly(ethyeleneoxide20-block-propyleneoxide70-block-ethyleneoxide20) (P123) was dissolved in 30 g of deionized water and 120 g of 2.0 M hydrochloric acid (HCl). The solution was stirred until the polymer fully dissolved. Next, 8.5 g of TEOS was added and the solution was stirred at 35 °C for 8 h. The temperature was then raised to 60 °C and the solution stirred for 48 h. The solution was placed in a hydrothermal bomb at 80 °C for a further 24 h for the medium pore or at 100 °C for 24 h for the large pore. The treated samples were allowed to cool to room temperature and were filtered to recover the solid product. The precipitate was washed and dried as in the small pore sample preparation. All three samples were then calcined to produce nanoporous solids. The samples were heated to 500 °C over 8 h under flowing nitrogen. After a 6 h hold at 500 °C, the nitrogen was replaced with flowing oxygen for an additional 6 h, followed by slow cooling over 6 h under nitrogen. The inner pore surface was then made hydrophobic by treating the materials with phenyldimethylchlorosilane [PDMCS] to aid in incorporation of the polymer into the pores.21 The samples were treated by stirring the silica powder under inert conditions for 1 h with 1

Cadby and Tolbert ml of triethylamine, which serves as an activating base, and enough PDMCS to cover the sample. After reaction, the samples were washed with ethanol and water, filtered, and cured at 100 °C for an hour to bind the silane to the silica surface. A noticeable increase in the weight of the sample of around 50% was observed after this reaction. Low-angle X-ray diffraction patterns were collected in two ways. Patterns with broader peaks (small and medium pore materials) were obtained with Mo KR radiation from a Rigaku UltraX 18 rotating anode X-ray generator and detected with a Roper Scientific X-ray CCD camera. Patterns on large pore materials (with narrower peaks) were collected at the Stanford Synchrotron Radiation Laboratory on beamline 1-4, using 0.1488 nm X-rays. N2 absorption/desorption data were collected to determine pore size and surface area using a Micromeritics ASAP 2000 porosimeter. The semiconducting polymer MEH-PPV was synthesized in-house via the Gilch route.30,31 The average polymer molecular weight was determined to be between 50 000 and 60 000 Da based on GPC analysis relative to polystyrene. To incorporate the polymer into the pores, the silica samples were first stirred in a 1% MEH-PPV in chlorobenzene solution until well mixed. The mixed silica/polymer solutions were then left in a dark oven at 80 °C for 48 h. After 48 h the samples were filtered, washed with fresh solvent, and dried in air. The resulting powder samples were stored in a dark nitrogen atmosphere. For some samples, additional solvent washing was used in an effort to preferentially remove polymer from areas external to the pores, leaving only polymer chains within the pores. To wash the small and medium pore samples, the samples were stirred in excess chloroform at room temperature for varying amounts of time. The washed powders were then recovered by filtration. Because of increased solvent access to the pores, the large pore samples required different washing conditions. Soaking in chloroform was found to completely remove the polymers, and thus to wash the large pore materials, the samples were simply placed on filter paper in a vacuum filter flask and chloroform was dropped onto the sample with a pipet. The extent of washing is thus measured in terms of the number of times chloroform was applied to the sample. For optical measurements, silica/polymer samples were suspended in glycerol to match the refractive index of the samples environment to that of the sample to reduce scattering. The glycerol samples were stirred for several hours at 40 °C until a homogeneous distribution of sample was achieved. For polarization measurements, samples were transferred to 1-cm quartz cuvettes and illuminated with vertically polarized light from the 532-nm line of a solid-state doubled Nd:Yag laser. The photoluminescence spectra of the samples were collected in both the vertical and horizontal polarizations, using a Kodak sheet polarizer and an Ocean optics CCD model USB 2000. The spectrometer was checked and found to have no intrinsic polarization bias, so measured spectra could be used without correction for an instrument function. The reported polarization ratios are the ratio of the emission spectra collected parallel to the excitation polarization divided by the emission spectra taken at 90° to the excitation. Absorption spectra were collected with a Perkin-Elmer UV-vis spectrometer and were corrected for the scattering background. Photoinduced absorption measurements were carried out on a custom-built spectrometer. The silica/polymer samples were suspended in glycerol in a 1-mm cuvette and placed in an APD variable-temperature cryostat. The samples were excited with the 488-nm line of an Ar+ laser with a power of 200 mW,

Incorporation of MEH-PPV into Mesoporous Silicas

Figure 1. Normalized low-angle X-ray diffraction patterns collected on the three nanoporous silica samples used in this work. Patterns for the medium and large pore samples are offset vertically for clarity. All patterns index to a p6mm hexagonal honeycomb structure, although the repeat distance of the honeycomb is different for each material.

Figure 2. The pore size distribution for the three nanoporous silica samples used in this work calculated using the BJH method applied to nitrogen desorption isotherms. All samples were calcined and coated with phenyldimethylchlorosilane prior to the measurement, so these values represent the average pore diameter available to the incorporated polymer. All pore size distributions are narrow, and the values are well spaced between 2.5 and 8 nm.

chopped at 170 Hz with a Stanford Research Systems optical chopper. The output from a tungsten lamp was focused on the sample and the transmitted light was collected with a CVI 1/4 m monochromator. The resulting change in the transition caused by absorption of the laser line was recorded on a Stanford Research Systems lock-in amplifier and were normalized to the dc transmitted light. The resulting ∆T/T spectrum is the photoinduced absorption spectra. Details of this setup can be found in ref 32. Photoluminescence and photoluminescence excitation (PLE) spectra were collected on a Yvon-Jobin FluoroLog-3 spectrometer. Samples suspended in glycerol were placed in a 1-cm curvette and front face collection was used to record PL and PLE spectra. Results and Discussion Structural Characterization of the Nanoporous Silica Host. X-ray diffraction and nitrogen gas absorption spectra were measured for each of the three silica pore sizes and are shown in Figures 1 and 2. Figure 1 shows the typical 1:x3:2 peak ratio of the (10), (11), and (20) hexagonal peaks for all materials, indicating that despite differences in size scale, all three materials have the same basic p6mm hexagonal honeycomb structure. Differences in peak widths are due to the manner in which the data were collected (synchrotron radiation versus a conventional

J. Phys. Chem. B, Vol. 109, No. 38, 2005 17881 X-ray source) rather than to differences in the degree of periodicity of the silica frameworks. As the size scale of the material increases, the diffraction patterns are observed to shift to lower q or lower angle. While this shift provides a precise measure of the change in nanoscale repeat distance in the material, it does not indicate exactly how the pore size changes because the thickness of the silica wall is also subject to change and the organic surface coating occupies some of the pore volume. Pore sizes after phenyldimethylchlorosilane coating were thus directly measured using nitrogen adsorption/desorption data fit using the BJH method. The resulting desorption pore size distributions are shown in Figure 2. The results indicate that materials can be created with narrow pore size distribution and diameters cleanly spanning the range from 2 to 8 nm. We note that the pore size difference between the medium and large pore samples is greater that the difference in the diffraction peak position. This is because the synthesis conditions used to produce the medium pore material generate a sample with significantly thicker silica walls than in the large pore case.27 For the remainder of this paper, these 2, 5, and 8 nm pore diameter samples will be referred to as small, medium, and large pore materials. Polymer Incorporation and Spectroscopic Characterization. Increased interaction between polymer chains in general decreases the quantum efficiency of the polymer system and red shifts the PL spectra. These phenomena can be explained by energy transfer between polymer chains allowing excited states to access longer chain segments and thus increase the possibility of excitons encountering a nonradiative site. Energy transfer between polymer chromophores, either from the same chain or from neighboring chains, is expected to be dominated by Fo¨rster transfer.33 Fo¨rster transfer is a nonradiative energy transfer that occurs between a donor and acceptor site separated by some distance r. The rate of transfer between donar and acceptor [KDfA] is proportional to the inverse sixth power of the distance between the two sites. Thus the transfer rate falls off rapidly with increased distance. Fo¨rster’s original theory was used to describe luminescence depolarization of molecular solutions with increasing concentration.34 As the concentration of luminescent molecules in a solution is increased, the Fo¨rster interaction increases due to a reduction in the average distance between chromophore, allowing more energy transfer between chromophores to occur. This scrambles the polarization as the emitting molecule is no longer required to have the same spatial orientation as the absorbing molecule. A reduction in the polarization ratio is thus an indication of increased intermolecular interaction. As such, the photoluminescence polarization ratio allows us to investigate interactions and energy transfer between polymer chains. For an ensemble of isolated and randomly oriented absorbing chromophores excited with a given polarization of light, the photoluminescence will be stronger in the direction parallel to the excited polarization than at 90° to the excitation. For a randomly homogeneous solution of isolated chromophores with parallel absorption and emission dipoles, the maximum ratio for parallel emission/perpendicular emission is 3,35 which corresponds to a polarization anisotropy of 0.4.36 If the absorption and emission dipoles are not parallel, this maximum value can be somewhat reduced. More importantly, however, if all of the neighboring emitting dipoles in a material are parallel to each other and to the absorbing dipole, Fo¨rster energy transfer can occur without resulting in a reduction of the polarization ratio from the maximum possible value stated above.

17882 J. Phys. Chem. B, Vol. 109, No. 38, 2005 Beyond polarization changes, the optical properties of these molecular systems are strongly affected by the environment of the molecule.15 In addition to the Fo¨rster transfer discussed above, which allows for red-shifted emission because of energy transfer to lower energy polymer segments, stabilization of either the ground state or the emitting excited state of a polymer by its environment can also lead to shifts in both the absorption and emission spectra. If the surrounding environment stabilizes the ground state more than the excited state then a blue shift in the absorption and emission will be seen. Conversely, if the environment preferentially stabilizes the excited state, red-shifted absorption and emission is observed. This solvatochromism is thus a function of the difference in the dipole moments of the ground and excited states and the refractive index of the medium.37 The high refractive index of polymer chains surrounded by other polymers is expected to result in a red shift in both absorption and emission, compared to a polymer chain that is surrounded only by silica or sliane. To understand both the alignment and the degree of interactions between polymer chains in pores of various sizes, we thus examined both the polarization of the emitted light and any red or blue shifts in the polymer absorption or emission. For comparison to data taken on samples of polymers in silica pores, photoluminescence spectra were also collected on a drop cast film and a dilute solution of MEH-PPV in chlorobenzene. The photoluminescence spectra of the film sample and the solution sample are given in Figure 3. The PL spectra of the film is, as expected, significantly red shifted compared to that from the solution sample. This red shift is likely due to a combination of increased Fo¨rster transfer, which allows for luminescence from lower energy, longer polymer segments, and increases in the dielectric constant of the environment surrounding each chromophore.38 The film sample also shows a red tail in PL and a reduction in the vibronic structure, which are signs of aggregation within the sample.39 Figure 4 shows wavelength-dependent polarization ratios for a series of samples. The lowest ratio is found in film samples, which show a value of approximately 1.0 at all wavelengths, indicating that energy transfer between polymer chains with different spatial orientations is rapid and thus scrambles the polarization effectively. For the solution samples polarization ratios between 1.8 and 1.9 were achieved. In this case, there is small decrease in this value at redder wavelengths, although the decrease is minimal. This reduction of the polarization ratio from the maximum value of 3 has been seen previously in solutions of MEH-PPV and other related semiconducting polymers.21,20,40 The reduced polarization ratio has been attributed to Fo¨rster energy hopping along different segments of the same conjugated polymer chains. In addition, if the absorption and emission dipoles are not parallel, this can further reduce the maximum achievable value, and some recent experiments on polymers in silica nanopores suggest that this may well be the case.41 Also shown in Figure 3 are spectra taken for typical silica pore/polymer samples. The spectra of all five samples fall into two distinct groups: the small pore and solution samples have emission peaks at about 560 nm, whereas the medium pore, large pore, and film samples have emission peaks at 590 nm. The difference between the two groups is likely due to the change in the polymer’s environment: In the redder samples, the polymer chains are surrounded by other polymers. This both produces a higher dielectric environment and allows for Fo¨rster energy transfer between polymer chains. In the bluer samples, the polymer chains are isolated and are surrounded either with

Cadby and Tolbert

Figure 3. Photoluminescence spectra for a dilute solution of MEHPPV (cyan), a drop cast polymer film (red), and the three washed polymer-in-silica nanopore samples. The smallest pore material is shown in blue, the medium pore material in brown, and the large pore material in orange. The data clearly show that PL from polymer incorporated into small silica pores is qualitatively similar to data collected on a dilute solution, while PL from polymer incorporated into medium or large pores looks much more like data collected on a polymer film sample.

Figure 4. Photoluminescence polarization ratios, defined as I|/I⊥, for polymer incorporated into small pore silica as a function of washing time. Data for a dilute polymer solution and a drop cast polymer film are included for comparison. As the sample is washed, the polarization ratio increases from a low value similar to a polymer film to a value almost as high as that achieved in solution, suggesting that washing selectively removes unaligned polymer chains adhering to the external silica surface. All samples show minimal wavelength dependence to the polarization ratio.

silica or solvent. The results thus provide strong spectroscopic evidence that small pore samples contain only one polymer chain per pore, while medium and large pores samples contain multiple chains per pore. These spectroscopic results are in good agreement with geometric results obtained from nitrogen adsorption/desorption measurements. From the data given in Figure 2, the small pore samples have a 2 nm pore diameter. From X-ray diffraction work performed on films of MEH-PPV, the side-to-side packing distance between polymer chains was measured to be 1.6 nm; the stacking distance between two chains was found to be 0.71 nm.42 From these distances it would be very difficult for more than one polymer chain to exist within a single small pore of 2 nm diameter. Therefore, we can assume that the majority of polymer chains in the small pores are isolated. By contrast, for the large and medium pore samples with diameters of 5 and 8 nm, there should be ample space for multiple polymer chains per pore. One complication that needs to be addressed in all samples is the possibility of polymer adhered to the outside of the silica grains, or polymer chains that are partly in and partly outside

Incorporation of MEH-PPV into Mesoporous Silicas

Figure 5. Photoluminescence polarization ratios, defined as I|/I⊥, for polymer incorporated into medium pore silica as a function of washing time. By contrast to the small pore material, for this material washing reduces the polarization ratio. This result indicates that washing removes interacting, parallel polymer chains from the pores, allowing more space for polymer chains to kink or coil. The wavelength dependence to the polarization ratio also indicates that polymer chains are in contact.

of the silica pores. Such “outside” chains would be expected to agglomerate in a manner similar to that observed in polymer films.20,21 Directly after polymer incorporation, an unwashed sample is thus expected to be covered in MEH-PPV on the outside and also have polymer within the pores. The polymer on the outside is expected to have a polarization ratio near 1, thus lowering the overall polarization ratio for the entire sample. With time soaking in a good solvent for the polymer, however, chains adhered to the outside of the pores and chains that are partly inside the pores and partly outside should be selectively removed relative to polymer chains that are fully embedded in the pores. Figure 4 thus shows wavelength-dependent polarization ratios for the small pore silica sample after soaking in a good solvent for the polymer for various periods of time. In agreement with previous data, as the washing time is increased, the polarization ratio also increases.21 The increased polarization ratio is due to the selective removal of polymer chains that are not isolated and aligned within the pores. In previous studies of macroscopically aligned nanoporous silica hosts with a similar pore size, washing produced polarization ratios well above 3, indicating that the polymer was incorporated in the silica nanopores.20,21 Moreover, in Figure 4, the polarization spectra show no wavelength dependence, which is a further indication that the aligned chains are isolated and as such, energy transfer to longer chains is limited. The highest polarization ratios achieved for small pore materials (1.7-1.8) are very close to those obtained from dilute solutions of isolated polymer chains. This value is probably lower than the theoretical value of 3 because of some combination of off-axis transition dipoles, as discussed above, and a small amount of unincorporated polymer that is not removed by solvent washing.20 The medium pore samples have a pore diameter of 5 nm as measured by nitrogen absorption/desorption measurements. This should allow multiple polymer chains to exist within each pore. Despite this fact, very high polarization ratios are observed for this material, as shown in Figure 5. Moreover, the highest polarization ratio is now seen in a sample that has not been extensively washed. This is very different from that of the small pore sample, and suggests that because of increased access to the larger pores, this sample does not contain a large amount of polymer that is partly in the pores and partly outside. One way to explain both the high polarization ratio and the redshifted PL data (Figure 3) is to postulate that these pores contain multiple, straight parallel polymer chains. As a result, a high

J. Phys. Chem. B, Vol. 109, No. 38, 2005 17883 dielectric environment is present and Fo¨rster transfer is allowed, but the energy transfer does not result in a scrambling of the polarization. In agreement with this idea, washing the sample reduces the polarization ratio, a phenomenon that could result when some of the tightly packed, aligned chains in the pores are removed, allowing the remaining chains to coil within the pores. The wavelength dependence of the polarization ratio for the medium pore system is also different from that of the small pore system (Figure 5). The peak of the polarization ratio is at about 600 nm, and the ratio decreases with increasing wavelength. This is again an indication that energy transfer can occur between polymer chains in the same pore. The reduction in polarization ratio at red wavelengths may be due to energy transfer out of the pore, or to energy transfer to some less ordered polymer chromophores which are not fully aligned with the pore axis. As the sample is washed and the chains become isolated, this wavelength dependence to the polarization ratio disappears (Figure 5). The pore size of the large pore sample is found to be 8.3 nm. This pore size should allow many polymer chains to coexist with a single pore and should potentially allow space for polymer chains to coil within a single pore. Extrapolation of light scattering data from much higher molecular weight MEHPPV, indicates that the size of these large pores is on the same order as the hydrodynamic radius of a single polymer chain in solution,39 thus opening up the potential for polymer to incorporate as a coiled entity. In agreement with this idea, the PL polarization ratios for the large pore samples are always low and show no significant increase or decrease with washing. The polarization ratio is one for an unwashed sample and increases to a maximum value of 1.2 with washing; we were unable to achieve a higher ratio than this and further washing of the sample reduces the polarization ratio. Apparently solvent simply and rapidly removes coiled polymer from the silica with no real gain in PL ratio. The slight increase in the polarization ratio after brief washing may be due to removal of bulk polymer from the outside of the grains. However, even after this washing, the sample still resembles a polymer film in PL and has a low polarization ratio indicative of randomly coiled, aggregated polymer chains. We thus postulate that this sample contains mostly coiled polymer chains contained within the pores forming small film-like sections of polymer with limited domain size. The environment of the polymer chains can also be studied by photoluminescence excitation spectroscopy (PLE). In a PLE experiment the intensity of a single emission band within the sample’s PL spectra is recorded as the excitation wavelength is scanned through the sample’s absorption band. For an isolated chromophore one would expect the PLE spectra to be identical with the absorption spectra. If aggregation occurs within the sample leading to chromophores with differing emissive properties, however, PLE can be used to selectively examine absorptions from these less emissive species. Figure 6 shows the PLE spectra taken at 660 nm for the small pore sample. The choice of 660 nm is based on the fact that this fairly red wavelength has been associated with an aggregation feature in MEH-PPV.12 As the washing time is increased, the tail of the PLE spectra blue shifts. The blue shift is associated with a change in the dielectric environment of the polymer as well as a reduction in polymer aggregation,43 both of which occur when film-like polymer chains are removed from the outside of the silica domains. Data from samples washed for 2 h are very similar those obtained after 60 min. Combined with the polarization data presented in Figure 4, these results indicate

17884 J. Phys. Chem. B, Vol. 109, No. 38, 2005

Figure 6. The normalized PLE spectra collected at 660 nm for polymers contained in small silica pores. As the washing time is increased, the PLE spectra shift to the blue. The result indicates that washing selectively removes film-like aggregated polymer chains adhering to the outer surface of the silica material. Data collected on samples washed for 2 h are very similar to the 60 min data shown here.

Figure 7. Normalized PLE spectra for the three sizes of silica/polymer samples and a film sample, all collected at 660 nm. In each case, samples with the highest polarization ratios were used. A steady red shift is observed in going from small to medium to larges pores and on to films. Emission at 660 nm is associated with absorption from aggregated polymer species,12 indicating a steady increase in chain aggregation across these samples.

that after 60 min or more of washing, the polymer chains in the sample are well isolated. The PLE spectra for the medium and large pore samples show similar results: as the samples are washed the PLE spectra blue shift. This result indicates that polymer can be washed out the pores of larger pore materials, leaving more isolated chains. The difference is that the polarization ratio decreases as the PLE shifts to the blue with the medium pore sample because the isolated polymer chains now have space to coil. As discussed above, the large pore materials show no real polarization change as the PLE shifts. Figure 7 compares the PLE spectra for the three silica/polymer sample with different pore sizes and a film sample. In each case, the PLE was collected on the samples that had the highest polarization ratio for that pore sizesthe small sample has been washed for 2 h, while the medium pore and large pore samples had only been washed briefly to remove excess polymer. It can be seen that although the small pore and medium pore materials have the same polarization ratio, the PLE for the medium pore sample is significantly red shifted. In addition, the PLE spectra for the film is redder than any of the polymer in nanopore materials, indicating that even for the largest pores, the polymer environment differs somewhat from a bulk film.

Cadby and Tolbert

Figure 8. The maximum polarization ratio achieved for each of the samples used in this work. Small pore, medium pore, and solution samples all show ratios in the range between 1.8 and 1.9, while large pore and film samples show significantly lower ratios. These data, combined with the PL and PLE data presented in Figures 3, 6, and 7, are used to make conclusions about the polymer conformation and degree of aggregation in each sample. Cartoons showing the likely polymer configuration are presented for each material.

Together, the spectroscopic experiments presented above show three very different systems. The small pore silica samples, after heavy washing, show optical properties very similar to those of the dilute solution. The PLE and PL spectra indicate that polymer chains are isolated and the polarization ratio is as high as the solution. The medium pore sample with minimal washing has a large polarization value, yet the PL spectrum resembles a film and the PLE spectrum indicates some polymer chain aggregation. In this case it is expected that the polymer chains are aggregated but aligned within the pore. In the large pore sample the polarization is low, and both the PL and PLE spectra indicate film-like behavior. We thus postulate that these large pore samples contain coiled, aggregated polymer chains contained within the pores. Cartoons of these different conformations, along with a summary of the polarization data, are presented in Figure 8. Photoinduced Absorption. The sub-gap excited-state species of MEH-PPV have been extensively studied with a wide range of techniques.44,45 Work using photoinduced absorption (PA) and absorption detected magnetic resonance (ADMR) have shown that the PA spectrum from 1 to 2 eV is dominated by a triplet absorption centered at 1.35 eV,46 with broad underlying polaron absorption between 1.2 and 1.8 eV. This is beautifully shown by the spin 1/2 ADMR work performed by Wei et al.47 In this work, only the spin 1/2 polaron absorption is detected and it can be seen spanning the 1-2 eV energy range with a maximum at 1.4 eV. The temperature dependence of the PA signal from triplets in PPV and MEH-PPV is markedly different from that of polarons, as shown in refs 47 and 48. The magnitude of both the triplet and the polaron absorption in PA shows a strong dependence on temperature. The triplet absorption spectrum has a maximum at low temperatures, i.e., lower than 50 K. The change in the triplet signal with temperature is significantly greater than that of the polaron, however, and while the polaron absorption is still strong above 200 K the triplet signal is completely removed above 170 K.48,49 Measurements on the pore samples were thus carried out at 220 K to remove triplet absorption from the PA spectra while retaining the polaron signal. Figure 9 shows PA spectra for the three samples. Each sample used had the highest achievable polarization ratio for that sample type: 1.8, 1.8, and 1.2 for small, medium, and large pore materials, respectively. For the data in Figure 9, spectra

Incorporation of MEH-PPV into Mesoporous Silicas

Figure 9. Photoinduced absorption spectra from the three silica/ polymer samples used in this work. The spectra have been normalized to the optical density of the samples at 488 nm, the wavelength used to excite the samples. In all cases, the peak is associated with absorption from polarons. As the size of the pore increases, this polaron absorption intensity dramatically increases. Increases in going from small to medium pore materials are associated with an increase in the probability of polaron production. Increases in going from medium to large pore materials are associated with either increased polaron production rates or an increased polaron lifetime.

were normalized with respect to the maximum optical density of the sample at 488 nm, the excitation wavelength used in the experiment. All three samples show a broad absorption feature ranging from 1.2 to 1.8 eV. The most dramatic difference between the PA spectra is the fact that the magnitude of the feature increases with pore size. This result indicates that increased polymer interaction increases polaron production and may also change polaron stability and lifetime. The low polaron absorption in the small pore sample surely stems from the isolated nature of the chains and indicates that chain-chain interactions are required for exciton break-up. This indicates that intrachain polaron production is not effective. A large increase in the polaron signal is seen in the medium pore sample in which several chains are aligned within the pore. The larger signal in the medium pore sample is a clear indication that polymerpolymer interactions are beneficial for polaron production and that interchain polaron production is more effective than intrachain. There is again an increase in the polaron signal from the larger pore sample. In this sample it is believed that both aggregation and coiling of the polymer chains occurs. It is difficult to say whether the large increase in polaron signal, a factor of 2, is due to extended aggregation or the effects of kinks in the polymer chains. Ideally, polaron production involving aggregation should only require nearest neighbor interactions 50,51 so the addition of more chains may not have a dramatic effect. Increasing numbers of chains should change the average dielectric, however, which could affect the probability of polaron formation. In addition, the PA experiment is a relatively slow measurement, and so lifetime effects can also affect the total polaron absorption. Kinks or breaks in conjugation of the polymer chains may provide long-lived trap sites for polarons.52 In agreement with this idea, recent results by the authors on these same nanopore samples indicate that the polaron lifetimes in the small and medium pore materials are similar and fairly short, while polaron lifetimes for the large pore materials are much longer.53 Thus, the increase in signal for the large pore material is attributed to efficient polaron production because of polymer chain aggregation, combined with increased polaron lifetimes because of the presence of kink sites.

J. Phys. Chem. B, Vol. 109, No. 38, 2005 17885

Figure 10. Normalized photoinduced absorption spectra for small pore (gray), medium pore (dashed), large pore (solid), and film (dotted) samples. As the size of the pore increases, the polaron absorption is seen to progressively shift to the red. Large pore materials show spectra nearly identical with those of polymer films. These results suggest that polymer-polymer interactions allow polarons to delocalize and find lower energy sites, but that the kink sites present only in the large pore and film samples may form the ultimate low energy trap sites for polarons.

In addition to changes in PA intensity, the position of the PA peaks is also observed to systematically shift to the red as the pore size is increased. To make this shift clearer, Figure 10 shows normalized PA spectra for all the nanopore samples and a film sample taken under the same experimental conditions. The peak of the absorption shifts from 1.65 eV in the small pore sample, to 1.5 eV in the medium pore sample, and on to 1.45 in the large pore material. The film sample shows a spectrum that is nearly identical to that of the large pore material. Like the spectral shifts observed in luminescence, the red shift with increasing pore size is likely due to some combination of solvatochromic (dielectric) effects and increased connectivity between polymer chains, which allows excited states to access lower energy chain sections before relaxing.54 In the small pore sample any excited state formed on the isolated chains cannot migrate to lower energy sites and therefore any polarons formed will be restricted to the chain where the initial excitation occurred. For larger pore sizes, in which several chains interact, excitations formed on a polymer chain will be able to migrate to a wider distribution of chain lengths. It appears, however, that kink sites form the ultimate low energy polaron traps because once the chains have space to coil in the large pore material, a limiting spectrum identical to a film spectrum is observed, despite the fact that the pore’s size is only 8 nm and the average dielectric constant differs significantly from that of a bulk polymer film. This result has important consequences for devices based on nanostructured semiconducting polymers. A red-shifted polaron absorption could be associated with a trapped, less mobile charge carrier. By straightening polymer chains within the silica nanopores, it may to be possible to remove these low energy kink trap sites and thus potentially to produce higher mobility charge carriers. We note that the formation of liquid crystalline domains can also be used to straighten and align polymer chains, but in this case, grain boundaries form between domains and these grain boundaries have a highly detrimental effect on bulk mobilities.55 For small pore materials with single polymers per pore, chain ends and defects would surely have a larger negative impact on bulk mobility than any domain boundaries in liquid crystalline materials. For the medium pore samples, however, the ability to combine multiple chains in a single pore means that continuous, wire-like structures can potentially be formed.

17886 J. Phys. Chem. B, Vol. 109, No. 38, 2005 By aligning chains in a continuous manner within silica nanopores, it may someday be possible to achieve high-mobility continuous domains of aligned polymer chains.22 Conclusions We have shown that the manipulation of conjugated polymers by guest/host chemistry in conjunction with nanoscale selfassembly leads to the manipulation of semiconducting polymers on the molecular level. We can achieve highly aligned and isolated polymer chains in the case of the small pore samples. In the medium pore samples we have closely packed, yet parallel polymer chains. Finally, the large pore materials have a more film-like environment with aggregated and coiled polymer chains. The interconnections between polymer chains in the medium pore material, coupled with the directionality of the silica pores, could be a promising route for the development of molecular wires. In small pore samples there is only a small band associated with absorption from polarons. The chains are expected to be isolated and thus any polarons must be formed on single, straight chains. Medium pore samples show an increased polaron yield that is attributed to close packing of the polymer within the pores. This is an indication that the dominant method for production of polarons in MEH-PPV is by interchain polaron production rather than intrachain production. In large pore materials, even more polaron signal is observed, although this may be due to lifetime changes, rather than differences in polaron production rate. These results indicate that polaron production and stability can be simply controlled by modifying polymer chain aggregation and conformation. Acknowledgment. This paper includes data collected at the Stanford Synchrotron Radiation Laboratory (SSRL), which is operated by the Department of Energy, Office of Basic Energy Sciences. This work was supported by the Office of Naval Research (N00014-04-1-0410 and N00014-99-1-0568) and by the Beckman Foundation. S.H.T. is an Alfred P. Sloan Foundation Research Fellow. References and Notes (1) Motamedi, F.; Ihn, K. J.; Ni Z.; Srdanov, G.; Wudl, F. Polymer 1992, 33, 1102. (2) Frankevich, E. L.; Lymarev, A. A.; Sokolik, I.; Karasz, F. E.; Blumstengel, S.; Baughman, R. H.; Ho¨rhold, H. H. Phys. ReV. B 1992, 46, 9320. (3) Heeger, A. J. ReV. Mod. Phys. 2001, 73, 681-700. (4) McGlynn, S. P.; Azumi, T.; Kinoshita, M. Molecular Spectroscopy of the Triplet State; Prentice-Hall: Englewood Cliffs, NJ, 1969. (5) Dyakonov, V.; Frankevich, E. Chem. Phys. 1998, 227, 203-217. (6) Frankevich, E.; Muller, J. G.; Lemmer, U. Chem. Phys. 2002, 285, 13. (7) Lane, P. A.; Wei, X.; Vardeny, Z. V. Phys. ReV. Lett. 1996, 77, 1544-1547. (8) Hide, F.; Schwartz, B. J.; Diaz-Garcia, M. A.; Heeger, A. J. Synth. Met. 1997, 91, 35. (9) List, E. J. W.; Kim, C. H.; Graupner, W.; Leising, G.; Shinar, J. Synth. Met. 2001, 119, 511. (10) Wilkinson, C. I.; Lidzey, D. G.; Palilis, L. C.; Fletcher, R. B.; Martin, S. J.; Wang, X. H.; Bradley, D. D. C. Appl. Phys. Lett. 2001, 79, 171. (11) Lin, H. N.; Lin, H. L.; Wang, S. S.; Yu, L. S.; Perng, G. Y.; Chen, S. A.; Chen, S. H. Appl. Phys. Lett. 2002, 81, 2572-2574. (12) Nguyen, T. Q.; Martini, I. B.; Liu, J.; Schwartz, B. J. J. Phys. Chem. B 2000, 104, 237-55. (13) Schwartz, B. J. Annu. ReV. Phys. Chem. 2003, 54, 141-172. (14) Nguyen, T. Q.; Schwartz, B. J. J. Chem. Phys. 2002, 116, 8198208. (15) Schaller, R. D.; Lee, L. F.; Johnson, J. C.; Haber, L. H.; Saykally, R. J.; Vieceli, J.; Benjamin, I.; Nguyen, T.-Q.; Schwartz, B. J. J. Phys. Chem. B 2002, 106, 9496-506.

Cadby and Tolbert (16) Nguyen, T.-Q.; Kwong, R. C.; Thompson, M. E.; Schwartz, B. J. Appl. Phys. Lett. 2000, 76, 2545. (17) Kresge, C. T.; Leonowitz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (18) Hou, Q.; Margolese, D. I.; Stucky, G. D. Chem. Mater. 1996, 8, 1147. (19) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. T.; et al. J. Am. Chem. Soc. 1992, 114, 10834. (20) Nguyen, T. Q.; Wu, J. J.; Doan, V.; Schwartz, B. J.; Tolbert, S. H. Science 2000, 288, 652-656. (21) Wu, J. J.; Gross, A. F.; Tolbert, S. H. J. Phys. Chem. B. 1999, 103, 2374-2384. (22) Molenkamp, W. C.; Watanabe, M.; Miyata, H.; Tolbert, S. H. J. Am. Chem. Soc. 2004, 126, 4476-4477. (23) Wu, C.-G.; Bein, T. Science 1994, 264, 1757. (24) Marlow, F.; McGehee, M. D.; Zhao, D. Z.; Chmelka, B. F.; Stucky, G. D. AdV. Mater. 1999, 11, 632. (25) Lu, Y.; Yang, Y.; Sellinger, A.; Lu, M.; Huang, J.; Fan, H.; Haddad, R.; Lopez, G.; Burns, A. R.; Sasaki, D. Y.; Shelnutt, J.; Brinker, C. J. Nature 2001, 410, 913. (26) Lupton, J. M.; Samuel, I. D. W.; Beavington, R.; Frampton, M. J.; Burn, P. L.; Bassler, H. Phys. ReV. B 2001, 63, 155206. (27) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548. (28) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc. 1998, 120, 6024. (29) Gallis, K. W.; Landry, C. C. Chem. Mater. 1997, 9, 2035. (30) Wudl, F.; Alleman, P. M.; Srdanov, G.; Ni, Z.; McBranch, D. ACS Symp. Ser. 1991, 455. (31) Wudl, F.; Hoger, S. PCT Patent Application WO 94/20589, 1991. (32) Lane, P. A.; Lipson, S. M.; Cadby, A. J.; O’Brien, D. F.; Mellor, H.; Martin, S. J.; Blau, W. J. Synth. Met. 2001, 119, 661-662. (33) Fo¨rster, T. H. Ann. Phys. 1948, 2, 55. (34) Herz, L. M.; Silva, C.; Friend, R. H.; Phillips, R. T.; Setayesh, S.; Becker, S.; Marsitsky, D.; Mu¨llen, K. Phys. ReV. B 2001, 64, 195203 and references therein. (35) Lakowicz J. R. Principles of Fluorescence Spectroscopy; Klumwer Academic/Plenum Publishers: New York, 1999. (36) Cantor C. R.; Schimel P. R. Biophysical Chemistry, Part II; Freeman, San Francisco, CA, 1980. (37) Greenfield, S. R.; Svec, W. A.; Gosztola, D.; Wasielewski, M. R. J. Am. Chem. Soc. 1996, 118, 6767-6777. (38) See, e.g.: Hess, B. C.; Kanner, G. S.; Vardeny, Z. Phys. ReV. B 1994, 47, 1407. (39) Nguyen, T.-Q.; Doan, V.; Schwartz, B. J. J. Chem. Phys. 1999, 110, 4068. (40) Watanabe, A.; Kodaira, P.; Ito, O. Chem. Phys. Lett. 1997, 273, 227. (41) Martini, I. B.; Molenkamp, W.; Tolbert, S. H.; Schwartz, B. J. Unpublished data. (42) Yang, C. Y.; Hide, F.; Diaz-Garcia, M. A.; Heeger, A. J.; Cao, Y. Polymer 1998, 39, 2299. (43) Collison, C. J.; Rothberg, L. J.; Treemaneekarn, V.; Li, Y. Macromolecules 2001, 34, 2346. (44) Smilowitz, L.; Heeger, A. J. Synth. Met. 1992, 48, 193-202. (45) Martin, S. J.; Bradley, D. D. C.; Lane, P. A.; Mellor, H.; Burn, P. L. Phys. ReV. B 1999, 59, 15133. (46) Christiaans, M. P. T.; Van Hal, P. A.; Janssen, R. A. J.; Wienk, M. M.; Kroon, J. M. Synth. Met. 1999, 101, 265. (47) Wei, X.; Vardeny, Z. V.; Sariciftci, N. S.; Heeger, A. J. Phys. ReV. B 1996, 53, 2187. (48) Colaneri, N. F.; Bradley, D. D. C.; Friend, R. H.; Burn, P. L.; Holmes, A. B.; Spangler, C. W. Phys. ReV. B 1990, 42, 11670. (49) Wei, X.; Hess, B. C.; Vardeny, Z. V.; Wudl, F. Phys. ReV. Lett. 1992, 68, 666. (50) (a) Frankevich, E. L.; Sokolik, I. A.; Lymarev, A. A. Synth. Met. 1991, 41, 213. (b) Stevens, M. A.; Silva, C.; Russell, D. M: Friend, R. H. Phys. ReV. B 2001, 63, 165213. (51) (a) Collison, C. J.; Rothberg, L. J.; Treemaneekarn, V.; Li, Y. Macromolecules 2001, 34, 2346. (b) Wang, P.; Cuppoletti, C. M.; Rothberg, L. J. Synth. Met. 2003, 137, 1461. (52) (a) Koslowski, T.; Jurjiu, A.; Blumen, A. J. Phys. Chem. B 2004, 108, 3283-3288. (b) Lipson, S. M.; Coleman, J. N.; Drury, A.; O’Brien, D. F.; Blau, W. J.; Cadby, A. J.; Lane, P. A.; Bradley, D. D. C. J. Appl. Phys. 2004, 95, 6138-6144. (53) Cadby, A.; Schwartz, R. N.; Tolbert, S. H. Manuscript in preparation. (54) Gadermaier, C.; List, E. J. W.; Markart, P.; Graupner, W.; Partee, J.; Shinar, J.; Smith, R.; Gin, D.; Leising, G. Synth. Met. 2000, 111, 523. (55) Kline, J.; McGehee, M. D.; Kadnikova, E. N.; Liu, J.; Frechet, J. M. J. AdV. Mater. (Weinheim, Ger.) 2003, 15, 1519.