You are on page 1of 5

NANO LETTERS

Coherent Coupling of Molecular Excitons to Electronic Polarizations of Noble Metal Nanoparticles


Gary P. Wiederrecht,* Gregory A. Wurtz, and Jasmina Hranisavljevic
Chemistry DiVision and Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439
Received July 23, 2004; Revised Manuscript Received September 2, 2004

2004 Vol. 4, No. 11 2121-2125

ABSTRACT
We report the observation of coherent polarization coupling between the exciton of a molecular J-aggregate and the electronic polarization of noble metal nanoparticles. For Ag nanoparticles, coherent exciton coupling to the conduction band electrons leads to constructive interference and a long-lived exciton. For Au nanoparticles, the excitonic state interferes destructively with bound electron transition dipoles and the exciton lifetime is reduced by two orders of magnitude. Coherent polarization coupling in heterostructured nanomaterials may strongly impact the functionality of future nanoelectronic and nanooptical materials.

Heterostructured hybrid materials that utilize inorganic templates functionalized with organic or biological constituents can produce optical, electronic, or selective binding properties that are otherwise not possible with a single component, impacting fields such as templated biostructures,1,2 selfassembly,3 molecular sensors,4 and photonics.5,6 Molecular J-aggregates are a target for the organic component of heterostructures because they are well-known for collective excitons with large oscillator strength, while their ease of oxidation makes them useful as sensitizers in the photographic process of electron injection into AgBr semiconductor particles.7,8 They are also of interest for modeling the transition from bulk to molecular electronic behavior, because the exciton coherence length typically extends over only a handful of molecules.9,10 This same property results in interesting ultrafast exciton dynamics and nonlinear optical properties due to a propensity for exciton-exciton annihilation.11-13 Ultimately, their large polarizability and optical extinction, as well as their electrochemical activity and robustness, should make coupling through linear or nonlinear processes to an inorganic substrate unusually strong. In particular, the narrow line width of the exciton band makes coherent interactions with low damping to other polarizable media a possibility. Some of the most polarizable inorganic templates in existence are metal nanoparticles that produce an enhanced optical field near their surfaces when interacting with visible electromagnetic radiation. Thus, metal nanoparticle/molecular J-aggregate heterostructures can produce an unusually large
* Corresponding author. E-mail: wiederrecht@anl.gov. Present address: Department of Pure and Applied Physics, The Queens University of Belfast, Belfast BT7 1NN, United Kingdom. 10.1021/nl0488228 CCC: $27.50 Published on Web 10/02/2004 2004 American Chemical Society

and confined optical field for exciting molecular excitons. We demonstrate here that the narrow exciton band of molecular J-aggregates can coherently and selectively couple to the bound electronic transition dipoles in Au or to the plasmon oscillation of free electrons in Ag nanoparticles. We also show that the exciton spectral and temporal properties vary significantly when interacting with the metallic core. Optics and electronics should benefit from tunable coherence phenomena in hybrid materials at the nanoscale.14 The exciton of molecular J-aggregates manifests itself optically as a narrow and red-shifted band with large extinction as compared to the monomers of the aggregate.9,11 An exciting advance was recently reported that allows for uniformly coating 10 nm diameter colloidal metal nanoparticles in an aqueous environment with molecular J-aggregates that are resonant with electronic transitions of the nanoparticles, while avoiding aggregation and precipitation.15 The dye used is a cyanine derivative, 5,5-dichloro-3,3-disulfopropylthiacyanine sodium salt (TC, Hayashibara Biochemical Laboratories, Japan), shown in Figure 1a.16 As illustrated in Figure 1b, the morphology of the TC aggregate is such that the thiacyanine chromophore is adsorbed on the particle and the SO3- groups extend away from the nanoparticle to provide solubility.17 Previous STM studies indicate that homogeneous coverage is obtained on nanoparticle surfaces.17 This report deals specifically with the spectroscopic signature and dynamics of excitons when their resonances overlap the plasmon or interband transition dipoles of metal nanoparticles. This leads to a very different variation in the nanoparticles optical response as compared to the cases

I |Pexciton + Bparticle|2 ) |Pexciton + Bplasmon + B P B P

Pn Bd-sp|2 n

(1)

n where Pd-sp represents the polarization of the bound electron transition dipoles. To quantify the polarization coupling, we calculate the frequency dependent effective dielectric constant ( eff) of the hybrid system and calculate the corresponding polarizability (Reff) obtained by conserving the dipolar term only from the Mie scattering development.20 These quantities can be expressed as follows:

Reff ) 4emr23
2 m

+1 -2 eff
eff 1 1

(2)

Figure 1. (a) Thiacyanine (TC) molecule, (b) orientation of the J-aggregate on the nanoparticle, (c) ground-state spectra of bare silver or gold nanoparticles, along with the fits to the data using the model described in the text, (d) ground-state spectra of TC J-aggregate coated silver or gold nanoparticles and associated fits with the core-shell model.

eff )

( (

1 1

+ 2 2) + 2( + 2 2) - (

3 2)(r1/r2) 3 2)(r1/r2)

(3)

where the excitonic absorption frequency is out of resonance with the metal nanoparticles plasmon or interband transitions.18,19 In particular, we will show that for such coherent interaction new absorption bands appear in the absorption spectrum that can be assigned to hybridized states. This is the first such study to our knowledge to show coherent polarization effects in J-aggregate/metal nanoparticle heterostructures. Figure 1 also illustrates the ground-state absorption spectra of bare (c) and TC-coated (d) Ag and Au nanoparticles. The ground-state spectra agree well with the results of Kometani et al., illustrating significant modulation of the nanoparticle extinction near the exciton resonance.15 Furthermore, the nanoparticles coated with TC show a sharp peak at 481 nm on Ag nanoparticles, but a pronounced decrease in the Au hybrid nanoparticles extinction at 474 nm. The modified extinction spectra point to constructive and destructive coherence effects between the nanoparticle electronic polarization and that of the exciton transition dipole. To model the striking differences in the ground-state spectra of the TC aggregate coated nanoparticles, we find that theoretical calculations using a core-shell structure of spherical symmetry within the dipolar limit satisfactorily reproduce the data. A schematic of this model system is illustrated in Figure 2a and b. In the frequency range of interest, the calculations simply describe the interaction between two transition dipoles that are homogeneously broadened. In our case, we found this simplification quite justified by the small impact the evaluated 10% size distribution has on the line width. The first transition dipole is the exciton absorption band of Lorentzian shape, and the second is the metallic core polarization, further developed as a sum of intraband and interband contributions. The polarization (P ) RE) of the sample then produces a farfield scattering intensity written as
2122

where 1 is the dielectric constant of the metal core, 2 is the dielectric constant of the J-aggregate shell, and m is the dielectric constant of the surrounding medium taken to be water. The dielectric behavior of the metal ( 1) is described by a sum of the intraband transitions (electronic response of the free electrons in the conduction band) and interband transitions (electronic transitions between the d valence band and the sp conduction band). We utilized models for each of these transitions that have successfully reproduced optical properties of metallic nanoparticles.21,22 The intraband dielectric response is described by a Drude term (eq 4) while the interband electronic response is modeled within the random phase approximation (eq 5): ) p2 ( + iintra())

intra

(4)

inter

)K

E dx
g

px - Eg [1 - F x 2 - x2 - 2 - 2iinter inter (5) (x,T)] 2 2 2 ( - x - inter)2 + 422 inter

where p is the bulk plasmon resonance frequency of the metal, intra and inter respectively characterize the plasmon and interband damping rates, Eg is the gap between a dispersionless d band and a parabolic sp band, and K reflects the oscillator strength of the interband transitions.21 Finally, the exciton resonance has been modeled assuming a Lorentzian absorption line shape, i.e., a homogeneously broadened absorption,23 with transition frequency o, reduced oscillator strength f, and line width agg:24 ) ) + 02f 02 - 2 - iagg

TC

(6)

The best fits to the ground-state spectra are shown along with the experimental data in Figure 1c,d.25 To allow for an
Nano Lett., Vol. 4, No. 11, 2004

Figure 2. (a) Metal-aggregate core-shell model and associated optical dielectric constants, (b) schematic of the electronic polarization in the metallic core with the excitonic polarization of the aggregate, (c) phase and amplitude of the core and shell contributions for the Ag hybrid derived from the fits shown in Figure 1, (d) phase and amplitude of the core and shell contributions for the Au hybrid using the fits shown in Figure 1.

understanding of the different coherent processes in the two (Ag/TC, Au/TC) hybrids, it is insightful to plot the optical response (amplitude and phase) of the metal core and molecular aggregate shell separately. These plots are illustrated in Figures 2c(Ag/TC) and 2d(Au/TC). For the amplitude and phase of the metallic core, we plotted Im(R) and Arg(R) using eq 2 with the dielectric constant relative to Au and Ag as obtained from the fits. For the J-aggregate shell optical response we directly plotted Im( 2) and Arg( 2). The coherent coupling within the core-shell nanosystems is apparent from the phase plots associated with the respective contributors. In particular, one finds that for the Ag hybrid the excitonic transition dipole is coherently coupled to the free electron plasmon dipolar resonance of the metal core. This coupling manifests itself by a constructive interference of the respective transition dipole moments leading mainly to a net increase in the nanoparticles extinction at 485 nm in Figure 2c. Since the interband band edge for Ag lies at approximately 3.9 eV (350 nm), far from the exciton resonance situated at 2.6 eV (474 nm as obtained from the fits), the phase sensitive coupling results solely from the interaction of the excitonic state with the free electron plasmon resonance of the metal. The hybrid peaks are then determined by the respective positions and oscillator strengths of the two bands (dipolar plasmon and exciton). For Au nanoparticle hybrids, the exciton resonance is isoenergetic with interband transition dipoles from the metal core (the interband band edge lies at 2.4 eV below the Fermi level). As shown in Figure 2d, this interband contribution at the exciton resonance produces an out-of-phase response between the excitonic and interband transition dipoles that results in a ground state bleach at 2.6 eV (474 nm), showing
Nano Lett., Vol. 4, No. 11, 2004

very little sensitivity toward the position of the plasmon absorption band and which position is therefore exclusively determined by the underlying exciton absorption frequency. The transition dipole moment frequency of the adsorbed J-aggregate is estimated from the fits to be centered at 2.6 eV (J-agg ) 474 nm). The calculated exciton transition frequency is found to be identical for both hybrids, which indicates that similar intermolecular resonance energies ensure the aggregates cohesion on Ag and Au. The intermolecular coupling strength can be estimated by the shift from o, the monomer transition dipole frequency. The larger red shift indicates that a more robust aggregate is formed when on the surface of the nanoparticles than in solution. In solution without nanoparticles, the J-aggregate exciton absorption is shifted by a value of 220 meV, while a shift by an additional 28% to 280 meV is observed for aggregates on nanoparticles. In a lower limit estimate (nearest-neighbor approximation without taking into account the effects of disorder), a 10% increase in the resonant coupling energy of aggregates on metal nanoparticles results.26 This increased stabilization is due to the interaction of the monomer with the particles surface, as it is not observed in solution.26 Another significant quantitative difference for the two hybrids lies in the excitons bandwidth: agg ) 86 meV on Ag and agg ) 118 meV on Au. These compare to agg in solution of 47 meV and the monomer line width of 250 meV. The bandwidth reflects the degree of localization of the excitonic state in the aggregate and is usually determined by disorder. Disorder includes a static contribution, i.e., siteto-site variations in the molecular transition dipole moments and variations in the intermolecular coupling energies, and a dynamic contribution from the thermal fluctuation of the molecules nuclear coordinates.27 Compared to the aggregate in solution, the exciton line broadening on the metal nanoparticles is likely due to the presence of the particle surface that introduces local field inhomogeneities. Given the large differences in the exciton bandwidth for aggregates formed on Ag or Au nanoparticles, it is reasonable to expect that exciton dynamics following photoexcitation of the metal nanoparticles be strongly modified. The results from ultrafast transient differential absorption spectra (DAS) on the bare and hybrid nanoparticles are shown in Figure 3. Along with the experimental data, we also plotted the fitted DAS. The theoretical DAS were calculated by considering the transient spectra as the sequence of linear absorption experiments where the excited-state absorption spectrum is induced from the ground-state spectrum prepared by the pump beam. This approach is valid for delays between the pump and the probe exceeding the width of the pump beam, i.e., in the weak excitation limit. In these experiments, an ultrafast laser with a 100 fs pulse width is used to optically excite the metallic core by using the 417 nm second harmonic of an amplified Ti:sapphire laser system described previously.28 The excitation is not resonant with the J-aggregate exciton band. For Ag nanoparticles, this photon energy lies in the nanoparticle plasmon absorption band. For gold, 417 nm lies in the 5d to 6sp interband transition region, but prior studies show that electron-electron scattering in Au nano2123

Figure 3. Transient absorption spectra for (a) bare and (b) coated Au nanoparticles. (c) Transient absorption kinetics at the exciton resonance for coated Au nanoparticles. Transient absorption spectra for (d) bare and (e) coated Ag nanoparticles. (f) Transient absorption kinetics at the exciton resonance for coated Ag nanoparticles.

particles produces an equilibration of conduction band electrons that occurs on a 100 fs time scale.29 This produces a transient bleach (TB) of the plasmon band that is faster than the time resolution of the apparatus. As a result, for both Ag and Au nanoparticles, excitation of the J-aggregate exciton is achieved via the evanescent field produced by the nanoparticle plasmon excitation.30 DAS spectra were taken with variable optical delays at probe wavelengths of 450 to 750 nm. A prominent aspect of the data is the large spectral differences for bare and coated nanoparticles, with the Au data shown in the left column (a-c) and Ag data shown in the right column (d-f) of Figure 3. The bare Au nanoparticles (Figure 3a) show spectra similar to that previously reported for uncoated nanoparticles, where an induced absorption on the low and high-energy side of the plasmon bleach is observed in the DAS. This is due largely to a simultaneous increase in the metallic core intraband damping constant (intra) relative to the ground state, from 0.645 to 0.69 eV (at 3 ps), and an increase in the interband damping (inter) from 0.24 to 0.2856 eV. In contrast, the Au hybrid shows two distinctive transient signatures compared to the uncoated nanoparticles. First, the hybrids DAS spectrum shows a strong red-shift in the plasmon bleach (20 nm) and the absence of any induced absorption on its low energy wing (Figure 3b). Second, the blue side of the DAS is now composed of two peaks. The broad peak originates from increased interband scattering whereas the most energetic positive feature is an excited-state absorption from the one2124

exciton to the two-exciton state.31 For the Au cores contribution to the DAS taken at 2.7 ps in Figure 3b, the data can be fit by an increase in inter to 0.31 eV from the ground-state value of 0.245 eV, and an increase of intra from 0.65 eV in the ground state to 0.667 eV in the transient excited state. The fits also show that the red-shifted plasmon bleach is actually due to a blue shift of the plasmon absorption upon excitation to 9.25 eV from 9.20 eV in the ground state, so that the lower energy components of the plasmon band appear as a bleach in the DAS. The short lifetime of the excitons as shown in Figure 3c indicates resonant Forster energy transfer coupling of the excitonic states in the aggregates selectively to the interband transition dipoles in the Au core. This conclusion is supported by the dynamics of the bleach at 474 nm on Au hybrids, which recovers within 3 ps and on the same time scale as the nanoparticles electron-phonon coupling events monitored at 540 nm. We note that the blue shift of the plasmon resonance in the heterostructured Au/TC nanoparticles has not been observed before to our knowledge and is likely due to changes in the environment through interactions with the adsorbed species. However, a change in the conduction electron density through charge transfer character cannot be entirely eliminated. Future study on this effect is warranted. The DAS data from the Ag nanoparticle systems are shown in Figures 3d,e. Note that our laser system prevents us to follow dynamics below 450 nm due to a lack of white light probe intensity. For bare Ag, the shape of the DAS spectrum in Figure 3d is determined by a red shifted excited-state absorption spectrum compared to the ground-state spectrum. This is due to the dispersion of the real part of the effective dielectric constant of the particle in the investigated spectral region.32 Compared to the bare nanoparticle, the Ag hybrids spectrum of Figure 3e shows an additional bleach that has a spectral dispersion of about 5 nm toward longer wavelengths within the first 10 ps following excitation by the pump light. Both the induced absorption and bleach can be fitted by considering a red-shifted excited-state plasmon absorption band (from 9.14 to 8.61 eV) for the metallic core, concurrent with a red-shifted (from 2.6 to 2.58 eV at 0.7 ps) and strongly damped (from 86 to 125 meV) exciton transition dipole moment, the latter being responsible for the transient bleach. The absence of the two-exciton state induced absorption band in the Ag hybrid DAS represents a fundamental divergence from the Au hybrids transient response. To describe this mechanism, we followed the temporal dynamic of the transient bleach in Figure 3e. These data are shown in Figure 3f at a wavelength of 475 nm and provide a value of about 233 ps for the recovery of the one-exciton state, or far longer than the 3 ps electron-phonon coupling decay time of the nanoparticle. Referring to Figure 4, we propose the following interaction mechanism within the Ag/TC hybrid: the oneexciton state (path 1) is created by evanescent field excitation of the J-aggregates on the metal nanoparticles. This is followed by an efficient electron transfer from the oneexciton band to the conduction band of the metallic core (path 2). This scheme prevents any buildup in the exciton concentration that would lead to the often-observed blueNano Lett., Vol. 4, No. 11, 2004

Figure 4. Schematic energy diagram for the Ag hybrid as obtained from electrochemical measurements performed on both TC Jaggregates and on Ag nanoparticles. Please see text for details.

shifted induced absorption.31,33-35 Following the electrons fast relaxation to the Fermi level through electron-electron scattering (path 3), the long recovery time for this state is governed by the non-adiabatic back-transferred charge from the Fermi level of the particle to the ground state of the aggregate (path 4). This reversible oxidation-reduction mechanism accounts for both the absence of induced absorption from the one-exciton state to a higher-lying state as well as the long recovery time of the excitonic band shown in Figure 3f. We have shown that the excitons of J-aggregate coated metal nanoparticles couple coherently to the dipolar responses of metal nanoparticles. The exciton band is shown to selectively couple to the free electron dipolar resonance for Ag nanoparticles, but coherent coupling of the exciton to bound electrons when on Au nanoparticles strongly modifies the exciton dynamics. These results should prove significant for understanding and improving the performance of molecular excitons near nanoscale metal surfaces. Acknowledgment. This work was partially supported by the U.S. Department of Energy, under contract number W-31-109-ENG-38. This work also benefited from the University of Chicago-Argonne National Laboratory Consortium for Nanoscience Research. References
(1) Mao, C.; Flynn, C. E.; Hayhurst, A.; Sweeney, R.; Qi, J.; Georgiou, G.; Iverson, B.; Belcher, A. M., Proc. Nat. Acad. Sci. U.S.A. 2003, 100, 6946. (2) Zhang, H.; Chung, S.-W.; Mirkin, C. A. Nano Lett. 2003, 3, 43. (3) Mirkin, C. A.; Letsinger, R. L.; Music, R. C.; Storhoff, J. J. Nature 1996, 382, 607. (4) Riboh, J. C.; Heas, A. J.; McFarland, A. D.; Youson, C. R.; Duyne, R. P. V. J. Phys. Chem. B 2003, 107, 1772.

(5) Liau, Y.-H.; Egusa, S.; Scherer, N. F. Ultrafast interferometric measurements of plasmonic transport in photonic crystals. Opt. Lett. 2002, 27(10), 857. (6) Nanophotonics; Prasad, P. N., Ed.; John Wiley & Sons: Hoboken, NJ, 2004. (7) Jelley, E. E. Nature 1936, 138, 1009. (8) Scheibe, G. Angew. Chem. 1936, 49, 563. (9) Higgins, D. A.; Reid, P. J.; Barbara, P. F. J. Phys. Chem. 1994, 100, 1174. (10) Sanchez, E. J.; Novotny, L.; Xie, X. S. Near-field fluorescence microscopy based on two-photon excitation with metal tips. Phys. ReV. Lett. 1999, 82(20), 4014-17. (11) J-Aggregates; Kobayashi, T., Ed.; World Scientific: Singapore, 1996. (12) Durrant, J. R.; Knoester, J.; Wiersma, D. A. Chem. Phys. Lett. 1994, 222, 450. (13) Hranisavljevic, J.; Dimitrijevic, N. M.; Wurtz, G. A.; Wiederrecht, G. P. J. Am Chem. Soc. 2002, 124(17), 4536. (14) Saito, K. J. Phys. Chem. B 1999, 103, 6579. (15) Kometani, N.; Tsubonishi, M.; Fujita, T.; Asami, K.; Yonezawa, Y. Langmuir 2001, 17, 578. (16) The J-aggregate coated metal nanoparticles are created by using a TC concentration of 5 10-6 M and a silver nanoparticle concentration of 4 10-9 M, with the exciton absorption band appearing over the course of a few minutes. The J-aggregates do not form in the absence of metal nanoparticles at this TC concentration. (17) Sato, T.; Tsugawa, F.; Tomita, T.; Kawasaki, M. Chem. Lett. 2001, 30, 402. (18) Ozcelik, S.; Ozcelik, I.; Akins, D. L. Appl. Phys. Lett. 1998, 73, 1949-51. (19) Tian, Y.; Guerin, F.; Guldi, D.; Fendler, J. H. In Nanoparticles in Solids and Solutions; Fendler, J. H., Dekany, I., Eds.; Kluwer Academic: Dordrecht, 1996; pp 569-574. (20) Bohren, C. F.; Huffmann, D. R. Absorption and scattering of light by small particles. John Wiley & Sons: New York, 1998. (21) Bigot, J.-Y.; Merle, J.-Y.; Cregut, O.; Daunois, A. Phys. ReV. Lett. 1995, 75(25), 4702. (22) Inouye, H.; Tanaka, K.; Tanahashi, I.; Hirao, K. Phys. ReV. B 1998, 57, 11334. (23) Gaizauskas, E.; Feller, K.-H. Opt. Commun. 2003, 216(1-3), 217. (24) Pockrand, I.; Brillante, A.; Mobius, D. Chem. Phys. Lett. 1980, 69, 499. (25) For the dielectric constant of the metals, the fitting parameters for the ground-state spectra are as follows. For bare Au: K ) -1.7 1037 SI, Eg ) 1.7 eV, pp ) 9.02 eV, ) 6.8. For Au/TC: K ) -3.41 1037 SI, Eg ) 1.7 eV, pp ) 9.2 eV, ) 4.21. For bare Ag: K ) -1.9 1037 SI, Eg ) 3.5 eV, pp ) 9.27 eV, ) 4.925. For Ag/TC: K ) -1.9 1037 SI, Eg ) 3.5 eV, pp ) 9.14 eV, ) 4.925. The same parameters were used for the ground-state spectra to calculate the DAS. For the organic aggregate the parameters are as follows. In solution, f ) 0.0175, p0 ) 2.675 eV, pagg ) 47 meV. On Au, f ) 2.612, p0 ) 2.617 eV, pagg ) 118 meV. On Ag, f ) 2.84, p0 ) 2.62 eV, pagg ) 86 meV. ) 1 in each case. (26) Kirstein, S.; Mohwald, H. J. Chem. Phys. 1995, 103, 826. (27) Nabetani, A.; Tomioka, A.; Tamaru, H.; Miyano, K. J. Chem. Phys. 1995, 102, 5109. (28) Greenfield, S. R.; Wasielewski, M. R. Appl. Opt. 1995, 34, 2688. (29) Hodak, J. H.; Henglein, A.; Hartland, G. V. J. Phys. Chem. B 2000, 104(43), 9954. (30) Wurtz, G. A.; Hranisavljevic, J.; Wiederrecht, G. P. J. Microsc. 2003, 210, 340. (31) Hamanaka, Y.; Kurasawa, H.; Nakamura, A.; Uchiyama, Y.; Marumoto, K.; Kuroda, S. J. Lumin. 2001, 94-95, 451. (32) Halte, V.; Guille, J.; Merle, J.-C.; Perakis, I.; Bigot, J.-Y. Phys. ReV. B 1999, 60, 11738. (33) Sundstrom, V.; Gillbro, T.; Gadonas, R. A.; Piskarskas, A. J. Chem. Phys. 1988, 89, 2754. (34) Minoshima, K.; Taiji, M.; Misawa, K.; Kobayashi, T. Chem. Phys. Lett. 1994, 218, 67. (35) Burgel, M. v.; Wiersma, D. A.; Duppen, K. J. Chem. Phys. 1995, 102, 20.

NL0488228

Nano Lett., Vol. 4, No. 11, 2004

2125