7610
Langmuir 2006, 22, 7610-7616
Water-Soluble Perylene Diimides: Solution Photophysics and Layer-by-Layer Incorporation into Polyelectrolyte Films Tingji Tang,† Jianqiang Qu,‡ Klaus Mu¨llen,‡ and Stephen E. Webber*,† Department of Chemistry and Biochemistry and Center for Nano and Molecular Science, The UniVersity of Texas at Austin, Austin, Texas 78712, and Max-Planck-Institut fu¨r Polymerforschung, Ackermannweg 10, D-55128 Mainz, Germany ReceiVed May 18, 2006. In Final Form: June 23, 2006 Recently the synthesis of water-soluble and fluorescent perylene diimides has been reported (Mu¨llen, K.; et al. Angew. Chem., Int. Ed. 2004, 43, 1528; Chem.-Eur. J. 2004, 10, 5297). We have characterized the photophysics of two of these compounds (anionic n-PDI, CAS Reg. No. 694438-88-5. and cationic p-PDI, CAS Reg. No. 817207-4-7) in pure water, dimethyl sulfoxide (DMSO), and aqueous NaCl. These studies, supported by molecular dynamics simulations, have led to the conclusion that these compounds form weakly interacting aggregated species in pure water. n-PDI and p-PDI have been incorporated in polyelectrolyte films of poly(styrene sulfonate) (PSS) and poly(diallyldimethylammonium chloride) (PDAC) following the layer-by-layer (LBL) methodology. The optical density and fluorescence intensity of the PDI-LBL films grew linearly with the number of layers, and the PDI was not extracted by subsequent polyelectrolyte deposition. The PDI fluorescence quantum yield was substantially diminished in these films, which we interpret as a self-quenching effect, enhanced by inter- and intralayer energy transfer. Energy-transfer studies to the incorporated cationic dye Brilliant Green (BG) has demonstrated that the BG resides in the same PSS-rich region as p-PDI and is largely excluded from the region that contains n-PDI (PDAC-rich).
Introduction The preparation of organic thin films from the spontaneous assembly of functional materials has been dramatically sped up by the introduction of the layer-by-layer (LBL) technique by Decher et al., based on the electrostatic interactions between the polycations and polyanions.1 LBL film fabrication can also be influenced by secondary interactions such as hydrogen bonding, donor and acceptor interactions, and π-π interactions.2,3 Because of its simplicity, reproducibility, and large area film uniformity, the LBL technique has been widely used in fabricating all kinds of heterostructural films, incorporating dendrimers, block copolymeric micelles, quantum dots, proteins, and even carbon nanotubes in support of applications such as chemical sensors, fuel cells, nano- or ultrafiltration, and organic light-emitting diodes, etc.4 In addition to the use of polymeric molecules as the building blocks for functional films, small dye molecules can be incorporated into LBL polyelectrolyte films.5,6 Dye molecules (e.g. 1,3,6,8-pyrenetetrasulfonic acid sodium salt) also can be † ‡
The University of Texas at Austin. Max-Planck-Institute fu¨r Polymerforschung.
(1) (a) Decher, G.; Hong, J. Makromol. Chem. Symp. 1991, 46, 321. (b) Decher, G. Science 1997, 277, 1232. (2) (a) Schmitt, J.; Decher, G.; Dressik, W. J.; Brandow, S. L.; Geer, R. E.; Shashidbar, R.; Calvert, J. M. AdV. Mater. 1997, 9, 61. (b) Gao, M. Y.; Richter, B.; Kirstein, S.; Mo¨hwald, H. J. Phys. Chem. B 1998, 102, 4096. (c) Ariga, K.; Lvov, Y.; Kunitake, T. J. Am. Chem. Soc. 1997, 119, 2224. (d) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117. (e) Caruso, F.; Mo¨hwald, H. J. Am. Chem. Soc. 1999, 121, 6039. (3) Tang, T. J.; Qu, J. Q.; Klaus, M.; Webber, S. E. Langmuir 2006, 22, 26-28. (4) (a) Wang, Y.; Tang, Z. Y.; Correa-Duarte, M. A.; Liz-Marza’n, L. M.; Kotov, N. A. J. Am. Chem. Soc. 2003, 125, 2830. (b) Wu, A.; Yoo, J. K.; Rubner, M. F. J. Am. Chem. Soc. 1999, 121, 4883. (c) Hiller, L.; Mendelsohn, J. D.; Rubner, M. F. Nat. Mater. 2002, 1, 59. (d) Tang, Z.; Wang, Y.; Kotov, N. A. Langmuir 2002, 18, 7035. (e) Park, J.; Hammond, P. T. AdV. Mater. 2004, 6, 520. (f) Caruso, F.; Caruso, R. A.; Mo¨hwald, H. Chem. Mater. 2001, 13, 1076. (g) Harris, J. L.; Stair, J. L.; Bruening, M. L. Chem. Mater. 2000, 12, 1941. (h) Ma, N.; Wang, Y.; Wang, A.; Zhang, X. Langmuir 2006, 22, 3906. (5) (a) Tedeschi, C.; Caruso, F.; Mo¨hwald, H.; Kirstein, S. J. Am. Chem. Soc. 2001, 122, 5841. (b) Caruso, F.; Lichtenfeld, H.; Donath, E.; Mo¨hwald, H. Macromolecules 1999, 32, 2317.
used as a probe to obtain detailed information about molecular level local properties such as polarity or molecular mobility.5 Previous studies that applied the LBL process to small charged organic dye molecules7 have shown that regular LBL film formation depends critically on several factors: polymer concentration,8 ionic strength of the polyelectrolyte solution,7,9 and dye molecular shape and dimension.8b The locally high density of small molecules in the LBL films has inspired researchers to design photoharvesting architectures.10,11 However it is often observed that the dye molecules are extracted by the next polyelectrolyte layer (of opposite charge), which significantly impedes the loading efficiency of small molecules into the films, affects the film integrity and stability,5-10 and also compromises the polyelectrolyte dipping solution. Due to their high fluorescence quantum yield and chemical and photostability, perylene diimides (PDIs) are good n-type material candidates for supramolecular assembly blocks, organic thin film transistors, photovoltaics, and molecular electronics,12 (6) (a) Wang, Y. H.; Hu, C. W. Thin Solid Films 2005, 476, 841. (b) Tokuhisa, H.; Hammond, P. T. AdV. Funct. Mater. 2003, 13, 831. (c) dos Santos, D. S., Jr.; Rodrigues, J. J., Jr.; Misoguti, L.; Oliveirs, O. N.; Mendonca, C. R. Biomacromolecules 2003, 4, 1502. (d) Wood, K. C.; Boedicker, J. Q.; Lynn, D. M.; Hammond, P. T. Langmuir 2005, 21, 1603. (7) (a) Araki, K.; Wegner, M. J.; Wrighton, M. S. Langmuir 1996, 12, 5393. (b) Cooper, T.; Campbell, A.; Crane, R. Lamgmuir 1995, 11, 2713. (8) (a) Ariga, K.; Lvov, Y.; Kunitake, T. J. Am. Chem. Soc. 1997, 119, 2224. (b) Saloma¨ki, M.; Tervasma¨ki, P.; Areva, S.; Kankare, J. Langmuir 2004, 20, 3679. (9) Linford, M. R.; Auch, M.; Mo¨hwald, H. J. Am. Chem. Soc. 1998, 120, 178. Dai, Z.; Da¨hne, L.; Donath, E.; Mo¨hwald, H. J. Phys. Chem. B 2002, 106, 11501. Kaschak, D. M.; Lean, J. T.; Waraksa, C. C.; Saupe, G. B.; Usami, H.; Mallouk, T. E. J. Am. Chem. Soc. 1999, 121, 3435. (10) Dai, Z.; Da¨hne, L.; Donath, E.; Mo¨hwald, H. J. Phys. Chem. B 2002, 106, 11501. (11) Kaschak, D. M.; Lean, J. T.; Waraksa, C. C.; Saupe, G. B.; Usami, H.; Mallouk, T. E. J. Am. Chem. Soc. 1999, 121, 3435. (12) (a) Tauber, M. J.; Kelly, R. F.; Giaimo, J. M.; Rybtchinski, B.; Wasielewski, M. R. J. Am. Chem. Soc. 2006, 128, 1282. (b) Ishi-I, T.; Murakami, K.; Imai, K.; Mataka, S. Org. Lett. 2005, 7, 3175. (c) An, Z.; Yu, J.; Jones, S. C.; Barlow, S.; Yoo, S.; Domercq, B.; Prins, P.; Siebbeles, L. D. A.; Kippelen, B.; Marder, S. R. AdV. Mater. 2005, 17, 2580. (d) Weiss, E. A.; Ahrens, M. J.; Sinks, L. E.; Gusev, A. V.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 5577.
10.1021/la061409q CCC: $33.50 © 2006 American Chemical Society Published on Web 07/29/2006
Water-Soluble Perylene Diimides Scheme 1
Langmuir, Vol. 22, No. 18, 2006 7611 “spectroscopic ruler” for characterizing macromolecular interactions in biological or polymeric systems.15 In the present paper FRET between the PDI moieties and BG is utilized to study the PDI distribution in the fabricated LBL films. The characteristics of selftransfer between the PDI moieties are useful for understanding the likely importance of energy transfer within the constructed LBL films, as discussed later in this paper. R0 is the characteristic distance at which 50% energy-transfer efficiency is achieved and is given by the following equations: R06 )
although their low solubility and π-π stacking between the PDIs have hindered these applications. The strong π-π interaction between the PDIs, especially in solid-state films, significantly decreases the PDI fluorescence quantum yield and is often accompanied with broadening and shifting of the absorption and fluorescence spectra due to the formation of (H- or J-) aggregates.13 Mu¨llen et al. have recently reported the synthesis of anionic and cationic water-soluble PDIs (referred to herein as n-PDI and p-PDI, respectively), which have high fluorescence quantum yields in aqueous solution and whose absorption and fluorescence spectra are typical of molecularly dissolved PDIs.14 In the present paper we have studied the solution-phase photophysics of these PDI moieties and demonstrated that they can be employed as a component in LBL polyelectrolyte films without loss by subsequent polyelectrolyte extraction. To our knowledge, this is the first report of the use of water-soluble PDIs as one of the building blocks for LBL film fabrication itself and represents a new methodology for the preparation of perylene diimide films.
J)
9000(ln 10)κ2QD J 128π4Nn4
∫
C0 )
∞FD(ν)A(ν)
ν4
0
(
(1)
dν
(2)
)
(3)
7.66 × 10-8 R0
3
(1) Materials. Positively and negatively charged water-soluble p-PDI (CAS Reg. No. 817207-4-7) and n-PDI (CAS Reg. No. 694438-88-5) (Scheme 1, with dimensions estimated from MD simulations of 2.13 × 1.33 × 1.89 nm and 2.13 × 1.15 × 1.49 nm (LWH), respectively3) were synthesized by Mu¨llen et al.14 All PDI aqueous solutions used for LBL film formation steps were 0.1 mg/mL with no NaCl present (the PDI fluorescence is diminished in 0.5 M NaCl solution, as will be discussed later). Poly(diallyldimethylammonium chloride) (PDAC, MW ) 25 000 g/mol), poly(styrenesulfonate) (PSS, MW ) 70 000 g/mol), and poly(ethylene imine) (PEI) were purchased from Sigma-Aldrich and used without further purification. PSS and PDAC at 1 mg/mL with 0.5 M NaCl solution were used as dipping solutions. The concentration for PEI deposition is around 1.5 mg/mL. Sodium dodecyl sulfate (SDS) was purchased from Aldrich and crystallized in ethanol before use. Dodecyltrimethylammonium chloride (DTAC, puriss grade g99%) was purchased from Fluka and used as received. The effects of surfactants such as SDS and DTAC or NaCl on the absorption and fluorescence emission spectra of p-PDI and n-PDI were studied by diluting concentrated p-PDI and n-PDI aqueous solutions with the SDS and DTAC solutions of different concentrations. All samples were stirred overnight before their spectra were taken. Brilliant Green (BG) was used as received from Sigma-Aldrich. Aqueous solutions of variable concentration were used in the spincasting procedure described below. (2) Fo1 rster Resonance Energy Transfer Study. Fo¨rster resonance energy transfer (FRET) has been frequently employed as a useful
QD is the quantum yield of the donor (p-PDI or n-PDI); FD(ν) is the normalized fluorescence spectrum of the donor; A(ν) is the molar extinction coefficient of the acceptor in M-1 cm-1, J is the overlap integral with the units cm4 M-1 cm-1, which is further expressed in eq 2, and ν is the wavenumber in cm -1.16 In eq 1 κ2 reflects the appropriate average value of the angular dependence of the dipoledipole interaction and can vary from 0 to 4. For rapidly rotating molecules the value is 2/3, as we used for the calculations below, but for randomly oriented and static molecules the value of κ2 is 0.476.16b In homogeneous solution the characteristic concentration (C0) required for 76% quenching is given by eq 3, given in mM in the following. The absorption and emission spectra for p-PDI or n-PDI overlap and their R0 values for self-transfer in aqueous solutions were calculated to be 3.58 nm (C0 ) 9.79 mM) and 4.50 nm (C0 ) 4.93 mM), respectively. These R0 values are on the order of the thickness of individual deposited polyelectrolyte layers, and the C0 values are much lower than our estimates of the PDI local concentration in the polyelectrolyte layers (see later). Thus it is plausible that inter- and intralayer energy transfer could occur between the PDI moieties. R0 values for p-PDI/BG and n-PDI/BG are 4.7 and 5.2 nm, corresponding to a C0 of 4.3 and 3.2 mM, respectively. Therefore FRET to BG that resides in the same layer as either PDI species should be very effective. (3) LBL Procedures. The quartz substrates were dipped in fresh Piranha solution for 10 min (H2SO4:H2O2 ) 7:3 (v/v): Caution! extremely corrosive) followed by RCA solution for 10 min (NH4OH:H2O2:H2O ) 6:2:2 (v/v/v)), both at 60 °C, and then baked at 420 °C for 6 h. The advancing contact angle with water was measured to be approximately 5° for all the substrates. Then the quartz was coated with the PEI layer to make it evenly charged with a high density of cations. The LBL dipping time for all solutions was 10 min, with both sides of the quartz substrate exposed to the dipping solution. A gentle stream of argon gas was employed to dry the films between each dipping step. For n-PDI LBL film fabrication a double layer of PSS/PDAC was used as a foundation layer before the n-PDI layer was deposited. Films with 1, 3, and 5 separation layers (SL) between the adjacent n-PDI layers were fabricated, whose film structure can be denoted as PEI/PSS/PDAC/(n-PDI/PDAC/(PSS/ PDAC)m)n. 1, 3, and 5 SL correspond to m ) 0, 1, and 2, respectively. If p-PDI was used as the PDI building block, the films can be represented as PEI/PSS/PDAC/PSS/(p-PDI/PSS/(PDAC/PSS)m)n with a three-layer foundation. It is worth mentioning that the number of PSS layers in p-PDI LBL films is nearly twice that in n-PDI films (for n-PDI films the number of PSS layers is NPSS ) 1 + mn and for p-PDI films NPSS ) 2 + (m + 1)n). “Precursor films” of PEI/
(13) (a) Wu¨rthner, F.; Thalacker, C.; Diele, S.; Tschierske, C. Chem.-Eur. J. 2001, 7, 2245. (b) Syamakumari, A.; Schening, A.; Meijer, E. W. Chem.-Eur. J. 2002, 8, 3353. (c) Bohn, P. W. Annu. ReV. Phys. Chem. 1993, 44, 37. (14) (a) Qu, J.-Q.; Kohl, C.; Pottek, M.; Mu¨llen, K. Angew. Chem., Int. Ed. 2004, 43, 1528. (b) Kohl, C.; Weil, T.; Qu, J.-Q.; Mu¨llen, K. Chem.-Eur. J. 2004, 10, 5297.
(15) (a) Stryer, L. Annu. ReV. Biochem. 1978, 47, 819. (b) Bokinsky, G.; Zhuang, X. W. Acc. Chem. Res. 2005, 38, 566. (c) Jones, C. D.; MaGratch, J. G.; Lyon, L. A. J. Phys. Chem. B 2004, 108, 12652. (16) (a) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Interscience: New York, 1971. (b) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer Academic/Plenum: New York, 1999.
Experimental Section
7612 Langmuir, Vol. 22, No. 18, 2006 (PSS/PDAC)n (n ) 1, 2, and 3) were fabricated in 0.5 M NaCl for the study of the rate of diffusion of n-PDI into the LBL films. For the study of the rate of diffusion of p-PDI into LBL films, precursor films of PEI/(PSS/PDAC)n/PSS (n ) 0, 1, and 2) were fabricated in 0.5 M NaCl. (4) Spin-Assisted Layer-by-Layer (SA-LBL).17 Brilliant Green was used as energy acceptor for the energy-transfer study. Incorporation of BG following the same protocol used for PDI deposition resulted in such a high BG concentration in the entire LBL film that the PDI fluorescence was totally quenched. Therefore, a strategy of spin-casting BG on top of the dried nonswollen solid films was employed to reduce the BG uptake. A few mL of an aqueous solution of BG at the desired concentration was placed on the film, and the film was immediately spun cast at a speed of 2000 rpm for 20 s, followed by washing four times by DI water at the same speed. BG was spun-coated onto both sides of the LBL films. The amount of BG taken up by the films was proportional to the BG concentration in the aqueous solution used and was essentially independent of the number of layers in the film. This deposition protocol will also result in a higher concentration of BG in the outer portion of the film, although we are unable to quantify the BG concentration profile. (5) UV-Vis and Fluorescence Spectroscopy Measurements. Both UV-vis and fluorescence spectroscopy were measured to monitor the LBL deposition process. Absorption spectra were obtained with a HP 8453 diode array spectrometer. The fluorimeter used was a SPEX Fluorolog-τ2 equipped with a 450 W xenon light source, Czerny-Turner double grating excitation and emission monochromators. A photomultiplier voltage of 950 V was typically used, and the excitation and emission slit widths were set at 2/2/4/4 mm for LBL samples. For solution samples, the slit widths of 1/1/2/2 mm were used. The quartz disk was fixed into a homemade sample holder that allowed the fluorescence to be collected in front face mode. To average out positional uncertainties, all the fluorescence spectra were obtained in triplicate for every PDI layer deposition. (6) Atomic Force Microscopy Measurements. For the atomic force microscopy (AFM) measurements a commercial microscope (D-3000, Digital Instruments, Santa Barbara, CA) was used. Images were recorded in tapping mode in air, employing cantilevers with a nominal radius of less than 10 nm, resonance frequency at 300 kHz, and a spring force constant at 40 N/m. The root mean square (rms) surface roughness of n-PDI LBL films were 4.5 nm before and 2.6 nm after n-PDI deposition. Two representative images are presented in Figure S1 (Supporting Information).
Results and Discussion This section is divided into three subsections: (1) an examination of the photophysical properties of n-PDI and p-PDI in aqueous solution, with the objective of elucidating their state of aggregation; (2) photophysical characterization of the LBL films prepared using n-PDI or p-PDI as components; (3) FRET studies using BG as an energy acceptor, demonstrating the strong localization of the PDI moieties into regions of oppositely charged polyelectrolyte. (1) Absorption and Emission Spectra of p-PDI and n-PDI in Solutions. Self-association of dyes in aqueous solution is often encountered and similar behavior would be expected for n- and p-PDI. Since all the LBL deposition steps discussed later occur either in pure water or 0.5 M NaCl, we wished to ascertain if n- and p-PDI were dissolved molecularly in water or if there was some degree of association between these moieties. In Figure 1 is shown the absorption and emission spectra of p-PDI and n-PDI in water and DMSO, and their photophysical parameters are collected in Table 1 (n-PDI in water has been studied in (17) (a) Jiang, C. Y.; Markutsya, S.; Tsukruk, V. V. AdV. Mater. 2004, 16, 157. (b) Jiang, C. Y.; Markutsya, S.; Pikus, Y.; Tsukruk, V. V. Nat. Mater. 2004, 3, 721.
Tang et al.
Figure 1. (a, b) Absorption and emission spectra of 1.5 × 10-6 M n-PDI and p-PDI in water (solid line) and DMSO (dotted line), respectively. Table 1. Photophysical Parameters for n-PDI and p-PDI in Different Solutions n-PDI p-PDI
media
λabsmax (nm)
λemmax (nm)
max (M-1 cm-1)
ΦFa
water DMSO water DMSO
568 577 587 581
618 611 628 612
29700 34400 32800 38500
0.54 0.82 0.14 0.55
a The quantum yield (ΦF) was measured relative to the standard PDI compound N,N′-bis(2,6-dimethylphenyl)-3,4,9,10-perylenetetracarboxylic diimide in dichloroethane, for which ΦF ) 1 is assumed.
detail by Margineanu et al.,18 and our values in Table 1 are in good agreement with theirs). As can be seen for both compounds, the extinction coefficient is slightly larger in DMSO, the absorption spectrum is more structured, and the fluorescence intensity is considerably increased, accompanied by a slight blue shift. These effects could be a simple solvent effect or could imply that the PDI compounds are weakly aggregated in pure water. Hence we explored the effect of NaCl and surfactants on aqueous solution photophysics of the PDIs. From these experiments, and supported by molecular dynamics simulations of nand p-PDI dimers in water, we conclude that both n- and p-PDI moieties form weakly interacting aggregates (not necessarily dimers) in aqueous solution. (a) Effect of NaCl and Surfactant on the n-PDI and p-PDI Solution Photophysics. The addition of salt to aqueous solutions is known to encourage the precipitation of sparingly soluble compounds, and in the case of our PDI compounds will reduce the electrostatic repulsion between their charged groups. While our PDI compounds are always deposited from salt-free solution in the preparation of the LBL films, the polyelectrolyte solutions are 0.5 M in NaCl such that there could be “salt effects” on the PDI aggregation state in the films. The effect on the absorption spectrum of adding NaCl to p-PDI or n-PDI solutions is shown in Figure 2. As can be seen, the optical density (OD) is reduced by about 20% over the full range of NaCl concentration, which we interpret as a hypochromic effect19 arising from aggregation. The change in the absorption spectra is similar for both n-PDI (18) Margineanu, A.; Holfkens, J.; Cotlet, M.; Habuchi, S.; Stefan, A.; Qu, J. Q.; Kohl, C.; Mu¨llen, K.; Vercammen, J.; Engelborghs, Y.; Gensch, T.; De Schryver, F. C. J. Phys. Chem. B 2004, 108, 12242. (19) Rhodes, W.; Chase M. ReV. Mod. Phys. 1967, 39 (2), 348.
Water-Soluble Perylene Diimides
Langmuir, Vol. 22, No. 18, 2006 7613
Figure 3. MD structure for a p-PDI dimer in pure water. Figure 2. (a) Absorption spectra of n-PDI and (b) p-PDI as a function of NaCl concentration, with insets showing the S-V fluorescence quenching curves. The arrow indicates increasing NaCl concentration.
and p-PDI, but the accompanying decrease in fluorescence is stronger for p-PDI. The corresponding Stern-Volmer plot of F0/F yields slopes of 6.23 and 42.8 M-1 for n-PDI and p-PDI, respectively. We propose that the addition of NaCl encourages dimer or larger aggregate formation that perturbs the absorption spectrum and diminishes the fluorescence. We also examined the effect of adding SDS or DTAC surfactants to the n- or p-PDI solutions. Above the surfactant critical micelle concentration (CMC) the absorption spectrum became more structured, the extinction coefficient increased, and there was a dramatic increase of the fluorescence quantum yield (see Figure S2, Supporting Information), which indicates the incorporation of PDI into the surfactant core monomerically. However the PDI/surfactant system is considerably more complex than the addition of NaCl and is the subject of continuing work.20 (b) Molecular Dynamics Simulation of p-PDI Dimer. Computer simulation of the structure of a possible p-PDI dimer was carried out using the IMPAC software (with periodic boundary conditions) from the Schroedinger Corp.21 For these calculations two p-PDI molecules were placed in the vicinity of each other with Cl- counterions (for simplicity) and the remaining volume filled with H2O molecules. After equilibration of this starting structure, the H2O molecules were removed and then reinserted into the available volume and the structure reequilibrated (see Figure 3). The atomic coordinates from this final structure were used to calculate the molecular center-to-center distance (7.76 Å) and the value of κ2 ) 0.2804 (see the Supporting Information). An analogous simulation was carried out for a n-PDI dimer (structure not shown) with a similar center-to-center distance (7.49 Å) but a much smaller value of κ2 (0.0574). The center-to-center distances are larger than one would expect for classical excimer formation, but for this distance and mutual orientation of the transition dipoles it is plausible that there could be a significant interaction between the p-PDIs in the excited state, but a much weaker one for the n-PDI dimer. As can be seen from Figure 3, the two molecular planes are approximately parallel (and the twisting of (20) Tang, T. Work in progress. (21) These calculations were carried out by Jessica Kingsberg in the laboratory of Prof. E. Dormidontova, Department of Macromolecular Science and Engineering, Case Western Reserve University, Cleveland, Ohio.
the perylene out of planarity is also easily seen). The structure of the n-PDI dimer is qualitatively similar. Our findings are analogous to those recently reported by Marcon et al. for a different class of water-soluble PDIs22 and Jung et al. on water-soluble terrylene diimides.23 Evidently the n- and p-PDIs interact sufficiently in these aggregates to slightly perturb the absorption spectra and diminish the fluorescence quantum yield, but these effects are minor compared to those reported by Xie et al. for an asymmetric cationic PDI complex in water.24 Wang et al. have reported extraordinary shifts in the fluorescence spectra of imide-substituted PDI compounds as a function of their concentration in chloroform.25 These compounds do not have substituents in the bay positions, allowing very strong π-stacking with concomitant shifts in the fluorescence and extensive broadening of the absorption spectrum. (2) Formation and Characterization of (PSS/PDAC)n LBL Films Containing PDI. (a) PDI Penetration into PEI/(PSS/ PDAC)n LBL Films. A control experiment was carried out to assess the ability of n- or p-PDI to diffuse into “precursor films” PEI/(PSS/PDAC)n (n ) 1, 2, and 3 were investigated here) (see Experimental Section). The absorbance and fluorescence intensity as a function of the dipping time for n-PDI are illustrated in Figure 4 (see Figure S3, Supporting Information, for the corresponding absorption and fluorescence spectra). The time required for the n-PDI concentration to reach equilibrium in the film is about 10 min, which is the deposition time used for the PDI LBL film fabrication described next. It can be seen that for thicker PEI/(PSS/PDAC)n films more n-PDI is absorbed. This implies that the PDI molecules are capable of extensive penetration and migration within the LBL films. Similar experimental results were obtained for p-PDI and (PSS/PDAC)n/PSS (n ) 0, 1, and 2) LBL films (not shown here). This result contrasts strongly with the situation when PDI is interposed between (PSS/PDAC)n layers (see next). (b) LBL of n-PDI or p-PDI with PSS and PDAC. A typical series of absorption spectra of PEI/PSS/PDAC/(n-PDI/PDAC/ (22) Marcon, R. O.; dos Santos, J. G.; Figueiredo, K. M.; Brochsztain, S. Langmuir 2006, 22, 1680. (23) Jung, C.; Mu¨ller, B. K.; Lamb, D. C.; Nolde, F.; Mu¨llen, K.; Bra¨uchle, C. J. Am. Chem. Soc. 2006, 128, 5283. (24) Xie, A. F.; Liu, B.; Hall, J. E.; Barron, S. L.; Higgins, D. A. Langmuir 2005, 21, 4149. (25) Wang, W.; Han, J. J.; Wang, L. Q.; Li, L. S.; Shaw, W. J.; A. D. Q. Li, Nano Lett. 2003, 3. 455-458.
7614 Langmuir, Vol. 22, No. 18, 2006
Tang et al. Table 2. Characteristics of UV-Vis and Fluorescence Spectra of p-PDI and n-PDI LBL Films with 1, 3, and 5 SLs p-PDI SLa absb (×1000) abs peakc fluor peakc quenching factord
1 9.6 587 621 35
n-PDI 3 10.9 592 620 25
5 9.8 588 621 9
1 11.2 575 612 43
3 11.5 576 614 33
5 11.6 577 613 22
a 0.5 M NaCl was used for the polyelectrolyte solutions. b Slopes of absorbance around 580 nm from either p-PDI or n-PDI as outside layer versus the number of PDI layers; c S0 f S1 UV-vis peak position or S1 f S0 fluorescence emission peak position for either p-PDI or n-PDI as the outmost layers; d Quenching factors obtained by comparing the fluorescence intensity of the p-PDI and n-PDI aqueous solutions in thin cells with p- or n-PDI LBL films.
Figure 4. (a) Absorbance at 580 nm and (b) fluorescence intensity of n-PDI as a function of dipping time for “precursor” LBL films of PEI/(PSS/PDAC)n (n ) 1, 2, and 3).
Figure 6. (a) Fluorescence emission spectra of n-PDI and (b) p-PDI with PSS and PDAC (in 0.5 M NaCl) LBL films. Black spectra indicate the emission spectra when the PDI layer was formed as the outside layer and red when the following double layer of either PSS/PDAC or PDAC/PSS formed. The insets show the fluorescence intensity as a function of the number of PDI layers.
Figure 5. (a) Absorption spectra of n-PDI and (b) p-PDI layerby-layer assembly with PSS and PDAC solutions (in 0.5 M NaCl) and 3 SL. Black spectra are for the PDI layer as the outside layer and red when the following double layer of either PSS/PDAC or PDAC/PSS is added with insets showing the absorbance as a function of the number PDI layers.
PSS/PDAC)n with 3 SL and different numbers of n-PDI deposition layers are shown in Figure 5a, together with the inset showing the absorbance at the maximum (∼580 nm) as a function of the number of n-PDI layers. Figure 5b provides the same information for p-PDI with 3 SL (PEI/PSS/PDAC/PSS/(p-PDI/PSS/PDAC/ PSS)n). The slopes of the absorbance vs layer number for 1, 3, and 5 SL are given in Table 2 and, within experimental error, are independent of the number of SL, in contrast to the penetration study in the previous subsection. Thus we conclude that the
previously deposited PDI blocks the subsequently deposited PDI from penetrating deeper into the polyelectrolyte layers. The absorbance of the PDI is unchanged when the next double layer (PDAC/PSS) is deposited (see the insets of Figure 5), which demonstrates that no PDI was extracted by the subsequent PDAC or PSS depositions. This behavior is noteworthy as extraction of small dye molecules is often observed by the deposition of subsequent polyelectrolyte layers.6-10 For both cases the fluorescence intensity showed a good linearity with the number of PDI layers (Figure 6), although there is a discernible increase in the fluorescence intensity when the next polyelectrolyte double layer is deposited on top of the PDI. The nonzero intercept of the fluorescence intensity vs the number of PDI layers indicates that the fluorescence quantum yield is diminished for n > 1, presumably because of additional self-quenching, perhaps enhanced by interlayer FRET. By carefully comparing the fluorescence intensity of the films with an aqueous solution of PDI contained in a small path length cell with a similar OD (the thin cell was used to eliminate any optical artifacts in the front-face fluorescence mode), we find that the fluorescence intensity for the LBL film much lower than for the
Water-Soluble Perylene Diimides
Langmuir, Vol. 22, No. 18, 2006 7615
Figure 7. Absorption vs the number of n-PDI layers deposited (b) and the normalized fluorescence intensity (9) for each layer. The number of separation layers is indicated between each n-PDI deposition (SL ) 1, 3, 5, 7, 9).
PDI solution species (the quenching factors in Table 2 vary from 9 to 43, depending on the number of SL). In general the fluorescence quenching factor is diminished as the number of SL layers increases, which we believe is a result of the PDI diffusing into the adjacent SL, thereby reducing the local PDI concentration and the degree of self-quenching. This effect is illustrated in Figure 7, in which the number of SL was increased between PDI depositions. Although the absorbance of the PDI increases linearly with the number of PDI layers, the fluorescence intensity increases supralinearly. (c) Consideration of the Local Concentration and Intralayer FRET. Based on the polyelectrolyte estimated thickness of 6 nm,26 the local n-PDI and p-PDI concentration in a single layer can be deduced from Beer’s Law to be 0.64 and 0.56 M, respectively (this estimate ignores diffusion of the PDI into adjacent layers, which, as discussed above, almost certainly occurs). This concentration corresponds to a volume per molecule of ca. 2.3-2.6 nm3, which is smaller than the approximate molecular volume in water (3.65 nm3) estimated from MD simulations (see Experimental Section). Thus there are ample opportunities for molecular interactions or the formation of aggregates that could lead to quenching. These estimated local concentrations are also much higher than C0 (9.79 and 4.93 mM for p-PDI and n-PDI, respectively; see eq 3) for these molecules, suggesting that facile intralayer Fo¨rster energy transfer could occur. PDI diffusion away from such a high local concentration into the adjacent layers could easily provide a concentration higher than C0, thus leading to the possibility of interlayer FRET. (3) FRET from PDI in Polyelectrolyte Films to Brilliant Green. FRET is often used to elucidate the proximity of an energy donor and acceptor.15 Brilliant Green was chosen as an energy acceptor in the present studies because of its good spectral overlap with the PDI emission (see earlier R0 and C0 estimates). The FRET efficiency was measured by comparing the PDI fluorescence intensity of a film before and after deposition of BG (there was negligible absorption by the BG at the excitation wavelength). In a series of preliminary experiments we measured the FRET efficiency versus the concentration of spin-casting BG solution for three-layer PDI films (with 3 SLs) and found that the BG concentration required to significantly quench the n-PDI fluorescence was much larger than for p-PDI (see Supporting Information, Figure S4). We then measured FRET as a function of the number of layers for n- and p-PDI (Figure 8). The shapes of these curves are very similar to each other, but a much larger BG concentration is required to achieve significant quenching for n-PDI (cf. 8.3 × 10-6 and 1.3 × 10-3 for p-PDI and n-PDI, (26) Park. J.; Hammond, P. T. AdV. Mater. 2005, 16, 520.
Figure 8. (a) Energy-transfer efficiency as a function of the number of n-PDI layers at a BG concentration of 1.3 × 10-3 M. (b) Energytransfer efficiency as a function of the number of p-PDI layers at a BG concentration of 8.3 × 10-6 M. Scheme 2. Schematic Representation of BG Distribution in p-PDI and n-PDI LBL Films with 3 SLs
respectively, a factor of approximately 150). Given the similarity in the R0 values for p-PDI/BG and n-PDI/BG (4.7 and 5.2 nm, corresponding to C0 at 4.3 and 3.2 mM, respectively), this large difference is surprising. We suggest that this effect is the result of the different local environments for the two PDI moieties. A layer of p-PDI is surrounded on both sides by PSS, which would be expected to strongly attract the cationic BG. If there is significant intermingling of the PSS layers and the p-PDI, the BG could be brought into relatively close proximity to the p-PDI.27 By the same argument, the BG would be excluded from the immediate vicinity of the n-PDI (these qualitative ideas are represented in Scheme 2). Thus these experiments have demonstrated a kind of microphase separation of the PDI moieties into “BG-philic” and “BG-phobic” regions. We expect that this phenomenon would extend to other charged molecular species that are capable of penetrating a multilayer polyelectrolyte film. We presume that the decrease of the FRET with the number of (27) Lo¨sche, M.; Schmitt. J.; Decher, G.; Bouwman, W.; Kjaer, K. Macromolecules 1998, 31, 8893.
7616 Langmuir, Vol. 22, No. 18, 2006
PDI layers is primarily a result of the penetration of the BG into the film (at least on the order of six to seven layers), although interlayer energy transfer cannot be ruled out as a contributor to the BG quenching. It is also interesting that PDI layers do not block the penetration of BG (which is a smaller molecule with a +1 charge), while they do block subsequent PDI penetration.
Summary In this work we have characterized the photophysics of the water-soluble perylene diimide dyes n- and p-PDI (Scheme 1) in aqueous solution and explored their incorporation into polyelectrolyte multilayers using the layer-by-layer dipping technique. Additionally we have used FRET from the PDI moieties in the polyelectrolyte film to absorbed Brilliant Green to demonstrate the high degree of localization of the two PDI in different regions of the layered polyelectrolyte film. We briefly summarize our findings in the following: (1) Photophysics in Solution. The absorption spectra and fluorescence spectra and quantum yield of n- and p-PDI in pure water are consistent with a molecularly dissolved species, but in both cases the absorption spectrum becomes more structured and the extinction coefficient and fluorescence quantum yield increases in DMSO (see Table 1). Addition of NaCl to aqueous solutions of n- or p-PDI has the opposite effect (reduced extinction coefficient and fluorescence quantum yield). We have interpreted these observations to be the result of breaking up or encouraging the formation of small aggregates or dimers, in which the PDIs must be rather weakly interacting (the fluorescence yield of strongly interacting PDI moieties is expected to be very low). The predicted structure from molecular dynamics for a n- or p-PDI dimer is consistent with this idea, as the PDI aromatic rings are nearly coplanar but offset from each other, with a centerto-center distance of approximately 7.5-8 Å. This is not the favored geometry for excimer formation but would give rise to an H-aggregate type coupling.28 (2) Incorporation of n- and p-PDI into PSS/PDAC Polyelectrolyte Films. Because the PDI molecules studied herein have a (4 charge, it is not surprising that they can be incorporated into anionic or cationic polyelectrolyte films with charge overcompensation, thereby allowing a continuation of the LBL film formation process. Unlike other multiply charged dyes (e.g. 1,3,6,8-pyrenetetrasulfonic acid salt)5 the PDIs are not extracted by the next polyelectrolyte layer deposition, perhaps because of their larger volume or as a consequence of their π-π interactions. Consistent with the idea of the role of molecular volume, Li et al. have reported LBL films prepared from the macrocycle nickel phthalocyanine tetrasulfonate (NiPc) and poly(diallydimethylammonium) chloride (PDDA), and there is no explicit mention of the loss of NiPc from the films upon subsequent PDDA deposition.29 The fluorescence of both n- and p-PDI is strongly quenched in these films, and it is reasonable to think that this is because of self-quenching. Incorporation of a charged PDI into a polyelectrolyte film can encourage a more intimate association of already-formed dimers or higher aggregates, and the high local concentration (on the order of 0.5 M) is much larger than the critical concentration for energy transfer (on the order of 5 mM) such that intralayer (and perhaps interlayer) energy transfer to quenching sites is quite probable. (28) (a) Ford, W. E. J. Photochem. 1987, 37, 189. (b) Hofkens, J.; Vosch, T.; Maus, M,; Ko¨hn, F.; Cotlet, M.; Weil, T.; Herrmann, A.; Grebel-Koether, D.; Qu. J. Q.; Mu¨llen, K.; De Schryver, F. C. Chem. Phys. Lett. 2001, 333, 255-263. (c) Liu, D.; De Feyter, S.; Cotlet, U.-M.; Weil, T.; Herrmann, A.; Grebel-Koether, D.; Qu. J. Q.; Mu¨llen, K.; De Schryver, F. C. Macromolecules 2003, 36, 5918. (29) Li, D.-Q.; Lu¨tt, M.; Fitzsimmons, M. R.; R. Synowicki,; Hawley, M. E.; Brown, G. W. J. Am. Chem. Soc. 1998, 120, 8797-8804.
Tang et al.
In these experiments we also systematically varied the number of PSS/PDAC double layers (SL ) separation layers) that separate successive PDI depositions. While the amount of PDI deposited was essentially independent of the number of SL, there was a systematic increase in the fluorescence intensity (by factors of 2-3) with increased numbers of SL (see Figure 7 and Table 2). We conclude from these data that the PDI is able to diffuse into adjacent SL during the subsequent deposition steps, thereby lowering its local concentration and self-quenching. (3) FRET from PDI in Polyelectrolyte Films to Brilliant Green. BG is a very effective energy acceptor for both n- and p-PDI (R0 ) 5.2 and 4.7 nm, respectively). We find that to achieve the same degree of energy-transfer quenching requires approximately 150 times more BG for n-PDI than p-PDI (see Figure 8). Since BG has a +1 charge, it is expected to prefer to localize in PSS-rich regions, as is the p-PDI (see Scheme 2). Conversely the BG should be preferentially excluded from the PDAC-rich regions. This explanation qualitatively accounts for the very large difference in the FRET quenching efficiency. This result is surprising given the general picture that there is considerable intermingling of the oppositely charged polymers in films of this type.1,27,30
Conclusions Water-soluble perylene diimide dyes have potential applications as biochemical dyes14a,31 and, because of their photostability and good fluorescence quantum yield, can be used in single molecule experiments.18 However they also have considerable potential in molecular electronics given their favorable photophysical, electrochemical, and n-type semiconductor properties.12 We have shown that they can be readily incorporated into polyelectrolyte films by the LBL technique, which opens up a convenient methodology for processing large scale surfaces of arbitrary topologies (for an alternative approach see ref 22). We presume that improvement of the fluorescence quantum yield in PDI arrays of the type described here would require additional steric protection of the PDI core and, most likely, a larger number of charged groups per molecular species. Acknowledgment. S.E.W. thanks the Welch Foundation Program (Grant F-356) for its financial support of this work. We would also like to acknowledge Jessica Kingsberg for running the molecular simulation of the p-PDI dimer and Prof. Elena Dormidontova (both of Case Western Reserve University) for helpful discussions concerning these simulations. K.M. wishes to acknowledge the support of the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 625 and Schwerpunktprogramm Organische Feldeffekttransistoren). E.E.D. would like to acknowledge the support of the National Science Foundation (CAREER Grant CHE-0348302). Supporting Information Available: AFM images of p-PDI LBL films, absorption spectra of n-PDI with DTAC and p-PDI with SDS at different concentrations, absorption and fluorescence emission spectra of n-PDI dipping into the preassembled PEI(PSS/PdAC)6 as a function of immersion time, energy-transfer efficiency as a function of spin-casting BG concentration for n-PDI and p-PDI LBL films with 3 SLs, and vector information from MD simulations of p- and n-PDI dimers. This material is available free of charge via the Internet at http://pubs.acs.org. LA061409Q (30) Kellog, G. J.; Mayes, A. M.; Stockton, W. B.; Ferreira, M.; Rubner, M. F. Langmuir 1996, 12, 5190. (31) Wei, T.; Abdalla, M. A.; Jatzke, C.; Hengstler, J. Biomacromolecules 2005, 6, 68.
