ARTICLE pubs.acs.org/Langmuir
Energy Transfer between Conjugated Polyelectrolytes in Layer-by-Layer Assembled Films Quentin Bricaud, Roxane M. Fabre, Robert N. Brookins, Kirk S. Schanze,* and John R. Reynolds* Department of Chemistry, Center for Macromolecular Science and Engineering, University of Florida, P.O. Box 117200, Gainesville, Florida 32611-7200, United States
bS Supporting Information ABSTRACT: We present a study of F€orster resonance energy transfer (FRET) between two emissive conjugated polyelectrolytes (CPEs) in layer-by-layer (LbL) self-assembled films as a means of examining their organization and architecture. The two CPEs are a carboxylic acid functionalized polyfluorene (PFl-CO2) and thienylene linked poly(phenylene ethynylene) (PPE-Th-CO2). The PFl-CO2 presents a maximum emission at 418 nm, while the PPE-Th-CO2 has an absorption λmax centered at 431 nm, in sufficient proximity for effective FRET. Several LbL films have been constructed using varied concentrations of the deposition solutions and identity of the buffer layers separating the two emissive layers, using a system of either weak polyelectrolytes, poly(allylamine hydrochloride) (PAH)/poly(sodium methacrylate) (PMA), or strong polyelectrolytes, poly(diallylammonium chloride) (PDDA)/poly(styrene sulfonate) sodium (PSS). The efficiency of FRET has been monitored using fluorescence spectroscopy. Initially, the fluorescence of the PFlCO2 (Eg ∼ 3.0 eV), which emits at 420 nm, is quenched by the lower band gap PPETh-CO2 (Eg ∼ 2.5 eV). For films using the PAH/PMA system as buffer bilayers and deposited from 1 mM solutions, the PFl-CO2 fluorescence is progressively recovered as the number of intervening buffer bilayers is increased. Ellipsometry measurements indicate that energy transfer between the two emissive layers is efficient to a distance of ca. 7 nm.
’ INTRODUCTION Due to its versatility and inexpensive character, the layer-bylayer (LbL) technique can be used as a means for fabrication of nanostructured films and assemblies for fine control of spatial, molecular, and supramolecular architecture.15 The LbL technique generally utilizes oppositely charged polyelectrolytes which form multi-ionic interacting simplexes (polyelectrolyte complexes). To form an LbL film, a substrate is first exposed to a charged polyelectrolyte solution for a certain time determined by diffusion, forming the first layer. Subsequent exposure of this treated surface to a solution of an oppositely charged polyelectrolyte solution leads to complete charge reversal at the surface and formation of a second layer. The process is then repeated for film buildup. In this way, it could be expected that the films formed have a well-ordered lamellar structure; however, in many instances the multilayer film architecture is significantly more complex, as the layers can mix and interpenetrate due to ionic and hydrophobic interactions, along with rearranging their supramolecular structures after deposition.68 Models tend to show gradient structures distributed homogeneously across a lateral structure, but complications arise and, in fact, some phenomena including aggregation,911 deconstruction,12 and porosity13,14 of the LbL films can occur (Scheme 1). There have been extensive investigations that have probed the structure and organization of LbL polyelectrolyte films as a r 2011 American Chemical Society
function of solution deposition conditions. These studies have spanned the effect of various parameters such as pH, presence/ absence of added salt, polyelectrolyte nature, concentration, ionic strength, adsorption time, and temperature.1521 In contrast, there have been relatively few studies that have probed the structure and organization in LbL films constructed using conjugated polyelectrolytes (CPEs).2224 CPEs are multifunctional polymers that feature a π-conjugated backbone with a set of pendant ionic sites. These ionic groups induce solubility in polar solvents (such as water or methanol) allowing processing from these environmentally friendly systems. The resulting materials exhibit optoelectronic and redox properties characteristic of the conjugated backbone structure25,26 and have been used to construct a variety of devices, including polymer light-emitting diodes (PLEDs),2729 organic photovoltaic (OPV) cells,30,31 bio- and chemosensors,32,33 electrochromics,3436 and others.37 In previous work, our groups investigated both the fundamental properties and potential applications of CPE-based LbL films and used these as model systems to understand exciton and charge transport.22 In particular, we have developed a family of variable gap CPEs with different charge types, charge densities, Received: December 27, 2010 Revised: March 14, 2011 Published: March 29, 2011 5021
dx.doi.org/10.1021/la105113k | Langmuir 2011, 27, 5021–5028
Langmuir
ARTICLE
Scheme 1. Schematic Model of LbL Film Architecture Showing the Extremes from Distinct Layers to Mixed Interlayers and Finally Evolving into Complicated Aggregates with Surface Reconstruction
Scheme 2. Repeat Unit Structures of the PFl-CO2 and PPE-Th-CO2 CPEs Employed
and distinct optical properties controlled by the π-conjugated backbone.38,39 With the present investigation, we take advantage of these properties to study the F€orster resonance energy transfer (FRET) from a polyfluorene-based CPE (PFl-CO2) to a thienylene-linked poly(phenylene ethynylene) type CPE (PPE-ThCO2) having the structures shown in Scheme 2. We note that the emission spectrum of PFl-CO2 features a maximum wavelength around 420 nm and, as required for effective FRET, matches well with the absorption spectrum of the PPE-Th-CO2 which shows a λmax around 430 nm. Our objective was to use FRET as a probe of the interactions of the two CPEs in the LbL films where they are separated by a series of insulating polyelectrolyte buffer bilayers. Through the course of this investigation, we have found that the efficiency of energy transfer is strongly dependent on both the film deposition conditions and the nature of the insulating polyelectrolytes. Parallel studies using ellipsometry reveal that the modulation of the FRET efficiency reflects the efficacy of buildup of multilayers of the intervening buffer layers as the active polymers are separated in distance. We find that a poly(diallylammonium chloride) (PDDA)/poly(styrene sulfonate) sodium (PSS) combination does not lead to effective LbL film deposition with a few bilayers deposited. Studies of optimized film conditions show that buffer bilayers made of poly(allylamine hydrochloride) (PAH) and poly(sodium methacrylate) (PMA) are especially effective spacers and the F€orster distance studies are consistent with ellipsometric film thicknesses.
’ EXPERIMENTAL SECTION Materials. Poly(diallylammonium chloride) (PDDA), poly(allylamine hydrochloride) (PAH), and poly(sodium methacrylate) (PMA) were obtained from Aldrich and used as received. Poly(styrene sulfonate) sodium (PSS) was purchased from Scientific Polymer Products and purified by dialysis before use. The two CPEs, a carboxylic acid functionalized polyfluorene (PFl-CO2) and thienylene linked poly(phenylene ethynylene) (PPE-Th-CO2), were synthesized according to the published references.40,41 Polyelectrolyte aqueous solutions
were prepared using 18.2 MΩ water (obtained from a Millipore Milli-Q system). Borosilicate glass slides were purchased from Corning. Silicon wafers were purchased from Siltron Inc., Korea. Substrate Preparation and Film Deposition. Prior to use, glass slides were cleaned by sonication in sodium dodecyl sulfate/water, water, acetone, and then isopropanol, for 15 min each, and subsequently dried. Silicon wafers were cleaned by immersion into a Piranha solution (concentrated sulfuric acid and 30% hydrogen peroxide solution 3:1) for 30 min and subsequently rinsed with Milli-Q water. (Caution: Piranha solution is extremely corrosive and must be handled with proper care.) Films were prepared via the LbL self-assembly technique using a programmable robot (Nanostrata StratoSequence IV). Substrates were alternatively dipped in solutions of cationic and anionic polyelectrolytes for 10 min each. Between each deposition step, the substrates were rinsed three times by immersion in Milli-Q water for 3, 1, and 1 min, respectively. After each rinsing cycle, the water was refreshed. All of the experiments were carried out at room temperature under ambient atmosphere. The typical film architecture is the following: substrate // (polycation/PFl-CO2)1 // (polycation/polyanion)x // (polycation/ PPE-Th-CO2)1. After the deposition process, the films were gently dried using a stream of filtered, compressed air. Before any measurements, the coated substrates were allowed to dry overnight under vacuum at room temperature. All measurements were made within two days of film preparation. Measurements. UVvisible absorption measurements were performed on a Perkin-Elmer Lambda 25 spectrophotometer, and fluorescence spectra were recorded using a Jobin-Yvon Fluorolog-3 spectrofluorimeter. For both measurements, the spectra shown are an average of three different spectra recorded from different areas on the films. Film thickness measurements were determined by ellipsometry using a commercial EP3SW imaging system (Nanofilm Surface Analysis, Germany) that employed a frequency-doubled Nd:YAG laser (adjustable power up to 20 mW) at 532 nm. Four-zone measurements were performed in air at a constant angle of incidence of 61.5° on Si/ SiO2 substrate wafers. Samples were analyzed with the variation of the number of buffer layers (x = 0 to 7). Three different LbL samples were prepared and studied by ellipsometry. For each sample, fifteen different regions were used to calculate the average thickness value. The thickness of the total polymer film was determined from a two-layer model (Si substrate/SiO2/polymer/air) using a constant refractive index for the polymer layer of 1.54.42,43 The measured angles were fitted by the optical model based on the Fresnel theory as a function of the optical parameters and the angle of incidence. Samples analyzed in this work were modeled using AnalysR software, which uses a LevenbergMarquardt algorithm that minimizes the error between the simulated and measured data. Atomic force microscopy (AFM) images were obtained with a Veeco diInova atomic microscope using the tapping mode.
’ RESULTS AND DISCUSSION This work focused on a fundamental study of singletsinglet excited state energy transfer between two fluorescent CPEs 5022
dx.doi.org/10.1021/la105113k |Langmuir 2011, 27, 5021–5028
Langmuir
ARTICLE
Scheme 3. Repeat Unit Structures of the Nonconjugated Polyelectrolytes Used to Form the Buffer Multilayers
Figure 1. Schematic representation of the architecture of built up LbL films.
having matched absorption/emission characteristics for FRET in order to attain an understanding of the physical processes at work in the formation and structure of LbL films. As illustrated by their repeat unit structures in Scheme 3, we selected two pairs of nonconjugated and nonvisible light absorbing cationic and anionic polyelectrolytes to serve as intervening buffer layers, specifically, poly(diallylammonium chloride) (PDDA) and poly(styrene sulfonate) sodium (PSS), along with poly(allylamine hydrochloride) (PAH) and poly(sodium methacrylate) (PMA). The PDDA/PSS system constitutes a pair of strong polyelectrolytes that are fully charged in solution, while the PAH/PMA system constitutes a pair of weak polyelectrolytes (partially dissociated at neutral pH). All of the films were constructed using the specific architecture depicted in Figure 1. A first layer of cationic buffer polyelectrolyte, either PDDA or PAH, was deposited directly onto the substrate, followed by a single layer of PFl-CO2 which deposited via charge compensation with the polycation leading to charge reversal. Next, a set of buffer bilayers was incorporated varying the number of bilayers with the idea of creating a buffer interlayer of variable thickness and regular spacing upon which the PPE-Th-CO2 single layer would then be deposited after capping with either PDDA or PAH. Overall, the multilayer structure can be depicted by the following: Substrate==ðPDDA=PFI-CO2 Þ1 ==ðPDDA=PSSÞx == ðPDDA=PPE-Th-CO2 Þ1
or
Substrate==ðPAH=PFI-CO2 Þ1 ==ðPAH=PMAÞx == ðPAH=PPE-Th-CO2 Þ1 Figure 2 shows the absorbance and emission spectra of films of both active conjugated polyelectrolytes deposited as single layers on a PDDA-treated substrate. The PFl-CO2 (Figure 2a) exhibits an absorption λmax at 362 nm (with an onset ca. 410 nm), while the PPE-Th-CO2 exhibits a λmax at 431 nm and an onset wavelength at ca. 500 nm. The PFl-CO2 fluorescence spectrum shows a maximum at 418 nm with a shoulder around 440 nm, while that of PPE-Th-CO2 exhibits a fluorescence maximum at 480 nm and a shoulder around 512 nm.
Figure 2. UVvis absorbance and PL spectra of single bilayer films of (a) PDDA/PFl-CO2 and (b) PDDA/PPE-Th-CO2 constructed on glass.
This figure nicely illustrates the FRET compatibility of these two polymers, as the strong emission from the PFl-CO2 between 400 and 500 nm strongly overlaps the PPE-Th-CO2 absorption. By adjusting the nominal distance of separating these two components via the number of photophysically inert buffer polymer bilayers interjected between active films, FRET studies give us a probe of the internal architecture of the multilayer films. In initial studies, multilayer thin films incorporating the two active CPEs and a set of 58 buffer bilayers consisting of PDDA/ PSS were prepared using 0.1 mM aqueous solutions of the polymers, and the films were characterized by UVvisible absorption and fluorescence spectroscopy as shown in Figure 3. Two absorption bands of constant intensity are observed with λmax ∼ 365 and 435 nm corresponding to the pFl and the PPETh-CO2 as expected, since the photoactive layers within the films remain constant for this series. In each case, the fluorescence spectra are nearly identical, and they feature the PPE-Th-CO2 emission only with a maximum at 480 nm and a shoulder near 512 nm. By increasing the number of intervening buffer bilayers, we expected to progressively recover the PFl-CO2 fluorescence and observe a diminution in the PPE-Th-CO2 fluorescence intensity. As this was not the case, we subsequently prepared multilayer films with 10 to 19 buffer bilayers using the same deposition conditions. Again, the fluorescence spectra (shown in Supporting Information, Figure S1) showed only PPE-Th-CO2 fluorescence with a constant intensity for each case. While this behavior was not expected, it demonstrates that the FRET remains efficient regardless of the number of PDDA/PSS deposition steps, suggesting that the distance between the two photoactive CPEs layers is short and changes little. In control experiments directed to ensure that our active materials remained stable during LbL deposition, we used the 5023
dx.doi.org/10.1021/la105113k |Langmuir 2011, 27, 5021–5028
Langmuir
Figure 3. Absorbance and PL spectra of films incorporating PFl-CO2 and PPE-Th-CO2 layers separated by PDDA/PSS buffer layers (from 5 to 8 buffer bilayers). All polyelectrolytes have been deposited from 0.1 mM solutions. The typical architecture is as follows: glass // (PDDA/PFl-CO2)1 // (PDDA/PSS)x // (PDDA/PPE-Th-CO2)1.
same deposition conditions to prepare three different films: one film incorporating only PFl-CO2 as the CPE capped by a set of four PDDA/PSS buffer bilayers, another film depositing a set of four PDDA/PSS buffer bilayers capped by PPE-Th-CO2, and a final multilayer film depositing the PFl-CO2 first, followed by a set of four PDDA/PSS buffer bilayers which are then capped by PPE-Th-CO2. The emission spectrum of the first film with PFlCO2 as the only active material (see Supporting Information, Figure S2) shows a maximum at 420 nm as expected for the PFlCO2 fluorescence. This result indicates that the PFl-CO2 layer is deposited as expected into the heterostructure assembly and there is no problem with the deposition process. For the film incorporating only PPE-Th-CO2 as an active material, as expected, the fluorescence spectrum λmax is found at 480 nm with a shoulder near 510 nm, which is the signature of the PPE-Th-CO2 emission. Finally, the last film incorporating both PFl-CO2 and PPE-Th-CO2 gave absorption and emission spectra similar to those shown in Figure 3 as only the PPE-Th-CO2 fluorescence is observed. This control experiment clearly shows that the PFlCO2 is properly incorporated in the multilayer films, and as the PFl-CO2 fluorescence is not recovered after almost 20 buffer bilayers, it is necessary to consider other means to increase the distance between the two active layers. In previous work by our group,44 CPE solutions of 0.1 mM had sufficient concentration to build up multilayer structures with a regular growth in thickness. As it is known that the efficacy of the deposition can be a function of the polyelectrolyte concentration and the ionic strength of the deposition solutions, it appeared that, for the PDDA/PSS buffer layers used in this work, 0.1 mM might be too dilute to induce layer buildup, as there is an equilibrium between polyelectrolyte deposition and redissolution. It is likely that CPE deposition is efficient at the lower concentrations due to the significant hydrophobic character of the conjugated backbone, which makes the chains less soluble in solution, thus favoring their deposition. As shown in Figure 4, by using a higher concentration of 3 mM nonconjugated polyelectrolytes, and varying the number of intervening PDDA/PSS buffer bilayers from 2 to 8, the PFlCO2 emission could be recovered. With 2 buffer bilayers, the emission spectrum shows that the fluorescence is mainly due to PPE-Th-CO2, though a small peak near 420 nm indicates
ARTICLE
Figure 4. PL spectra of multilayer films constructed on a glass substrate and incorporating PFl-CO2 and PPE-Th-CO2 layers separated by 2 to 8 PDDA/PSS (3 mM each) buffer bilayers. The typical architecture is as follows: glass // (PDDA/PFl-CO2)1 // (PDDA/PSS)x // (PDDA/ PPE-Th-CO2)1.
recovery of a portion of the PFl-CO2 emission. This 420 nm feature increases in intensity from 4 to 8 buffer bilayers as the FRET interaction between the two CPEs is decreased, even though the PPE-Th-CO2 emission always remains stronger. These results demonstrated that the strong polyelectrolyte PDDA and PSS system was quite sensitive to deposition conditions, making it difficult to vary and control the separation between the different CPEs in the multilayer films. We then decided to switch our inert buffer system to poly(allylamine hydrochloride) and poly(methacrylic acid) PAH/PMA, a system that uses weak polyelectrolytes.22,45 A set of six multilayer films were prepared incorporating from 0 to 5 PAH/PMA buffer bilayers using a concentration of 3 mM for the polyelectrolytes and a pH of 3.5 for both the polyelectrolyte and rinsing solutions. As with the earlier studies, a layer of PAH polycation was first deposited, followed by the single layer of PFl-CO2, and subsequently, the varied number of buffer bilayers was capped by a single layer of PPE-Th-CO2. As depicted in Figure 5a, when there is no buffer bilayer between the two CPE layers, only the fluorescence of the PPE-Th-CO2 is observed due to the effective FRET. Even with just one buffer bilayer separating the two active layers (orange spectrum in Figure 5a), the fluorescence of the PFl-CO2 at 420 nm begins to emerge, while the PPE-Th-CO2 fluorescence remains strong. With two buffer bilayers incorporated, the two peaks belonging to the PFl-CO2 fluorescence can be clearly observed at 420 and 440 nm and the fluorescence intensity corresponding to the PFl-CO2 becomes stronger than that of the PPE-Th-CO2. As the number of bilayers is increased, the PFl-CO2 fluorescence continues to evolve until it essentially fully dominates the spectrum after 3 or 4 bilayers. Considering that the goal of our experiment is to use the number of bilayers as a sensitive means for controlling CPE separation distance, it was felt that the PFl-CO2 fluorescence recovered too rapidly, while the PPE-Th-CO2 fluorescence disappears too quickly, under these conditions (3 mM PE concentration). As such, we decided to decrease the polyelectrolyte concentration to 1 mM for the PAH/PMA buffer layer system. Repeating the experiment of Figure 5a with 1 mM buffer polyelectrolyte provided good control of the relative PFl-CO2 and PPE-Th-CO2 emission intensities as depicted in Figure 5b. 5024
dx.doi.org/10.1021/la105113k |Langmuir 2011, 27, 5021–5028
Langmuir
ARTICLE
Figure 5. PL spectra of multilayer films constructed on glass substrates and incorporating PFl-CO2 and PPE-Th-CO2 layers separated by 0 to 5 PAH/ PMA buffer bilayers. All polyelectrolytes have been deposited from 3 mM (Figure 5a) or from 1 mM (Figure 5b) solutions and the pH adjusted to 3.5. The typical architecture is as follows: glass // (PAH/PFl-CO2)1 // (PAH/PMA)x // (PAH/PPE-Th-CO2)1.
Figure 6. Thickness of heterostructure assemblies as a function of the number of buffer bilayers for (a) (PAH/PMA) and (b) (PDDA/PSS) separating PFl-CO2 and PPE-Th-CO2 (all deposited from 1 mM solutions) measured by ellipsometry on a silicon wafer with a 15 Å silicon oxide surface. The standard deviations for the ellipsometric thickness values are
