Fluorescence Decay Study of Anisotropic Rotations of Substituted

10308

J. Phys. Chem. B 2001, 105, 10308-10315

Fluorescence Decay Study of Anisotropic Rotations of Substituted Pyrenes Physisorbed and Chemically Attached to a Fumed Silica Surface Ilia N. Ivanov, Reza Dabestani,* A. C. Buchanan, III, and Michael E. Sigman Chemical & Analytical Sciences DiVision, Oak Ridge National Laboratory, P.O. Box 2008, MS-6100, Oak Ridge, Tennessee 37831-6100 ReceiVed: March 6, 2001; In Final Form: June 6, 2001

Optical polarization spectroscopy has been used to investigate molecular dynamics of the fluorescent probes 1-pyrenebutanol and 1-pyrenebutyric acid at the solid/air interface of cab-o-sil (fumed silica nanoparticles). Pyrenebutanol was chemically attached to the surface of cab-o-sil through a silyl ether bond, while pyrenebutyric acid was physisorbed on the surface. Dynamics of fluorescence depolarization for both molecules was studied under steady-state and time-resolved conditions. Low and high loadings of the probe molecules were used in our studies to examine the dynamics of the probes motion and excimer formation. Our data indicate that excimer formation is mostly static in nature for the adsorbed probe, but shows more dynamic character for the chemically attached probe with spacer molecules present. Fluorescence lifetimes were dependent on the concentration of the probe molecule and became shorter at higher surface loadings for both the chemically attached and physisorbed probes. For the chemically attached probe, the presence of excess co-attached biphenyl molecules was found to provide a 2-D solvent-like environment.

Introduction An understanding of the fundamental interactions between adsorbed molecules and heterogeneous surfaces has been the subject of much research in recent years.1-4 These fundamental interactions may play an important role in many chemical reactions and processes including environmental chemistry,5-13 surface catalysis, and solid-state synthesis.14-22 In solution chemistry, of central importance is the understanding of the nature and dynamics of local structures induced by solute and solute-solvent interaction.23-28 For heterogeneous systems (e.g., silica), the nature of the network structure and the changes that it will impose on physisorbed molecules depends on the surface termination groups (e.g., silanols) or modifier groups.29-34 One of the most powerful methods for probing these interactions in solution involves studying the reorientation dynamics of a fluorophore molecule (probe) to see how the relaxation dynamics of the probe responds to various properties of the solvent (e.g., viscosity, dielectric constant, and dipole moment).35-39 Due to photoselection, fluorescence emission is usually depolarized, which results from rotational diffusion of the probe in the absence of energy transfer.37 The degree of this depolarization can be related to the dynamics of interaction between the solvent and solute. There are several reports where emission anisotropy has been used for heterogeneous systems to probe sol-gel, gel-glass conversion,40 microviscosity in the cavities of titania gel,41 rotational diffusion of tracer spheres in packing and dispersion of colloidal spheres,38 dynamics of pyrenyl labeled polyallyl carbosilane dendrimer,42 pyrene dimer emission dynamics on silica,43 and photophysical properties of ruthenium polypyridyl SiO2 gels.44 Cab-o-sil silica (12 nm particle size) was chosen as a model system because of its flat surface (no defects or pores) to avoid different depolarization of light scattered from the surface which * To whom correspondence should be addressed. Phone: (423)576-7325. Fax: (423)574-7596. E-mail: [email protected].

can complicate the analysis if defects or pores were present.45,46 The cab-o-sil surface contains a known density of silanol groups that can be chemically derivatized at a range of surface densities. Two-component surfaces can be prepared in which a fluorescent probe molecule is anchored to the surface and surrounded by spacer molecules of variable chemical structure.47,48 The spacer molecules will serve as modifiers to the fluorescent probesilica surface interaction. Because the average distance between silanol groups in cab-o-sil is 5 nm,1 the modifier and probe molecules will be separated by the same distance creating a pseudo “two-dimensional” model49 where the probe molecule is surrounded by a “ring” of modifier molecules. Such a system can be envisioned as the intermediate between the solution and gas/solid phase where the probe can experience both environments (the probe experiences the gas/solid environment and, in the presence of modifier molecules, it also experiences a “pseudosolution” environment). In this paper we report our findings on the steady-state and time-resolved fluorescence anisotropy of 1-substituted pyrenes physisorbed and chemically attached to the cab-o-sil surface. In-plane and out-of-plane rotational rate constants for the motion of the chemically attached probe molecule, calculated from the anisotropy data, appear to be affected by the presence of spacer molecules. Our study shows that in heterogeneous systems molecular dynamics is influenced by the environment that molecules experience. Materials and Methods Cyclohexane HPLC grade (J. P. Baker) was dried over 8-12 mesh molecular sieve (J. P. Baker). 1-Pyrenebutyric acid (Molecular probe), 1-pyrenebutanol (Aldrich), and glycerol (Kodak) were used as received. 4-Phenylphenol (hydroxybiphenyl) was purified as previously described.47,48 Sample Preparation. The derivatized silica surfaces containing the fluorescent probe (butylpyrene) and modifier (biphenyl) molecules (Figure 1) were prepared by the two-step method described previously for other two-component surfaces.47,48 In

10.1021/jp010858q CCC: $20.00 © 2001 American Chemical Society Published on Web 09/27/2001

Anisotropic Rotations of Substituted Pyrenes

J. Phys. Chem. B, Vol. 105, No. 42, 2001 10309

Figure 2. Schematic diagram of the apparatus employed for timeresolved photoluminescence measurements. Complete details are given in the Experimental Section.

Figure 1. Molecular model for (A) chemically attached 1-pyrenebutanol with biphenyl as the spacer and (B) physically adsorbed 1-pyrenebutyric acid on silica surface.

the first step, an initial surface coverage of biphenyl (ca. 600 µmol/g) was prepared by condensation of 4-phenylphenol with the surface silanols of the cab-o-sil. The biphenyl molecules are attached to the silica surface through a Si-O-Caryl linkage. In a subsequent step, 1-pyrenebutanol molecules were chemically exchanged onto the surface to replace a small portion of the hydroxybiphenyl molecules. Two surfaces were generated that contained 4.0 and 47 µmol/g of the fluorescent probe in the presence of a large excess (ca. 570 µmol/g) of the attached biphenyl modifier. This exchange process was accomplished by adding 3.20 or 19.2 mg of 1-pyrenebutanol to 2.02 or 1.07 g, respectively, of the biphenyl-modified silica in a dry methylene chloride (40 mL) slurry. After stirring and removal of the solvent on a rotovap, the solids were heated in a sand bath (200 °C for 1 h) under vacuum (1 × 10-5 Torr) in a sealed tube. The recovered solids were extracted with dry methylene chloride (100 mL, stirred for 1 h, filtered, and washed with additional 20 mL) and dried in a vacuum desiccator. Analysis of surface coverages involved digestion of the silica in aqueous base, recovery of the 1-pyrenebutanol and 4-phenylphenol, and analysis by GC with internal standards as previously described.47,48 Samples of the probe (pyrenebutyric acid) physisorbed on the cab-o-sil surface (Figure 1) were prepared by a previously reported method.50-56 Two loadings of 1.04 and 4.17 µmol/g 1-pyrene-butyric acid on cab-o-sil were prepared, and a portion of each sample was placed in a rectangular 1 cm quartz cuvette. Both samples were degassed under vacuum (10-5 Torr) for 2 days to ensure complete removal of residual solvent and oxygen. The degassed samples were then flame-sealed. Samples with chemically attached pyrenebutanol were also degassed for 2 days and flame-sealed. The probe concentration on cab-o-sil was chosen such that at low loadings mainly monomer emission would be observed while at high loadings both monomer and excimer emission would be observed. Fluorescence Anisotropy. Steady-State Measurements. Steadystate fluorescence polarization measurements were conducted on a Spex Fluorolog-2 spectrophotometer equipped with a 450

W Xe lamp as the excitation light source and monochromators on both excitation and emission sides. Two Glen-Thompson polarizers were used to adjust the excitation and emission polarization. Emission and excitation spectra were monitored at the front face of the sample. Time-ResolVed Measurements. The geometry of the timeresolved setup for measuring dynamics of anisotropy is shown in Figure 2. Circularly polarized 337 nm light from a nitrogen laser (VSL-337LRF, Laser Science), which provides 4-6 ns laser pulses with a pulse energy of 120 µJ, was used as the excitation source. A small portion of the excitation beam reflected off a quartz plate beam splitter onto a photodiode (201/ 579-7227, Thornlabs, Inc.) was used to trigger data acquisition. Polarization of the excitation beam was done using a GlenThompson polarizer prism. The emission from the sample was collected with two lenses and passed through a polarizer into the monochromator. The exit port of monochromator was attached to the detector equipped with a side on R928 photomultiplier tube (Hamamatsu, PMT spectral response 185-900 nm, and rise time 2.2 ns). The PMT signal was amplified using a DC-300 MHz amplifier (model SR445, Stanford Research System) and the amplified signal was fed into a 500 MHz LeCroy transient digitizer (model 9354A). For each kinetic decay profile 1000 signals were sum-averaged by the transient digitizer. The averaged signals were exported as ASCII files to a personal computer. Data collection and transfer were done from the personal computer using data acquisition software written using LabView 5.0 (National Instruments). Timeresolved anisotropy values were calculated from the experimental kinetic decays using Origin 6.0 (Microcal Software Inc.) software. The experiments showed that signal correction for background light scattering due to silica neither changed the absolute value of the signal nor improved the signal-to-noise ratio of anisotropy signals. This is probably due to the fact that scattered light from the silica surface has a lifetime within the response time of our detector. On the basis of these findings, no correction for the background scattering was applied in determination of the anisotropy values. The fluorescence decay profile observed for various experiments was converted to fluorescence anisotropy, r(t), using eq 1

r(t) )

IVV(t) - G(t) IVH(t) IVV(t) + 2G(t) IVH(t)

(1)

where IVV(t) and IVH(t) represent the fluorescence decay profiles

10310 J. Phys. Chem. B, Vol. 105, No. 42, 2001

Ivanov et al.

Figure 3. Molecular model of a 1-substituted pyrene superimposed on the in-plane projection of an oblate ellipsoid. The z-axis is perpendicular to the plane of the molecule.

at time t observed with emission polarization parallel and perpendicular to the direction of excitation polarization (vertical) and G is a correction factor accounting for variation in sensitivity of the PMT to differently polarized light expressed as

G(t) )

IHV(t) IHH(t)

(2)

where IHV(t) and IHH(t) are vertical and horizontal polarized emission intensities detected with horizontal polarized excitation. The expression in the denominator of eq 1 represents total emission expressed as F(t) ) IVV(t) + 2GIVH(t). A similar approach for evaluation of fluorescence anisotropy was used for the steady-state experiments. The emission intensity in this case is an integral value and is time independent. Theoretical Treatment of the Probe’s Motion on Cab-oSil. To describe the motion of chemically attached 1-pyrenebutanol on cab-o-sil, we used a theoretical approach to anisotropy decay of ellipsoids.37,39,57 The probe was treated as an oblate revolving ellipsoid (the unique axis is shorter than the other two having the shape of a flattened sphere), a simplified version of a general revolving ellipsoid (three unequal semiaxes), Figure 3. The relatively long butyl substituent allows free in-plane and out-of-plane rotational modes.37 Because of the disklike shape of the pyrene probe, the transition dipoles are within the plane of the ring. This is possible provided the angle between the absorption transition moment A (or emission transition moment, E) and the unique axis of the ellipsoid is 90° (βA ) βE ) 90°, Figure 3). In this particular case, the angle between absorption and emission moments is equal to the angle between their projection onto the plane of the molecule, β ) ξ (Figure 3). Assuming the ellipsoid is rigid, a simplified expression for anisotropy decay as a function of time will have the form37

r(t) ) 0.3 cos2ξ exp{- (4DVV + 2DVH)t} + 0.1 exp{- (6DVH)t} (3) where DVV and DVH are the rotational rate constants for inplane and out-of-plane rotation. We have utilized eq 3 in our time-resolved anisotropy experiments to calculate the rotational rate constants and ξ. Results and Discussion A. 1-Pyrenebutyric Acid in Glycerol. This study was carried out to observe the probe’s behavior in a viscous solvent. At low temperature, the vibrational bands are better resolved in glycerol than in MeOH with sharper peaks at 377, 387, 397, 416, 444, and 472 nm, respectively (not shown). The excitation spectrum shows two similar vibrational progressions when

Figure 4. Emission and excitation spectrum of 1.04 µmol of 1-pyrenebutyric acid physisorbed on cab-o-sil. The emission spectrum (B) was measured when the sample was excited with 337 nm light of a 450 W Xe lamp. The excitation spectrum (A) was recorded monitoring at 396 nm. Anisotropy of the emission and excitation was estimated using the following equation: r ) IVV - GIVH/IVV + 2GIVH.

monitored at 397 and 416 nm (not shown). The excitation spectrum monitored at 444 and 472 nm also exhibits characteristic pyrene bands, which are shifted to longer wavelengths (390, 413, and 439 nm). This is likely due to the aggregate form of 1-pyrenebutyric acid.1,31,33 The emission anisotropy has a relatively constant value of 0.14 throughout the spectrum (not shown) that is lower than the theoretically allowed maximum value of 0.4.37 The excitation anisotropy (not shown) has a similar shape for 397 and 416 nm bands and is constant throughout the first vibrational series with a value of 0.15 decreasing stepwise in the second vibrational series to 0.05. A relatively higher anisotropy value (0.35) was observed for 439 nm vibrational progression when the emission was monitored at 444 and 472 nm (not shown). A first-order fit of the 377 nm emission kinetics for 1-pyrenebutyric acid in glycerol (at ambient temperature) gave an emission lifetime of 165.5 ( 1.1 ns. Anisotropy decay was positive and could be fitted to a double exponential equation giving a lifetime of 33.7 ns for the short-lived component and 809 ns for the long-lived component. B. 1-Pyrenebutyric acId Adsorbed on Cab-o-Sil. Ba. Low Surface CoVerage. Steady-state emission spectrum of 1.04 µmol/g of 1-pyrenebutyric acid adsorbed on cab-o-sil shows two maxima at 376 and 396 nm, a shoulder at 418 nm, and a broad band in the 450-500 nm wavelength region that could be attributed to pyrene excimer, Figure 4B. The excitation spectrum monitored at 396 nm exhibits two vibrational progression bands with maxima at 262 and 273 nm (first progression) and 308, 320, and 339 nm (second progression), respectively, Figure 4A. The excitation spectrum monitored at 470 nm (not shown) shows vibrational bands whose maxima coincide with the minima of the excitation spectrum monitored at 396 nm. This anticoincident behavior in the excitation spectra when monitored under the monomer (396 nm) and excimer (470 nm) emission has been attributed to the static nature of excimer formation.31,33,35,45,58 Emission anisotropy (excitation at 337 nm) is in the range 0.004-0.02 throughout the spectrum, Figure 4B. Excitation anisotropy (monitored at 396 nm) varies from -0.03 to -0.01 in the 250-360 nm wavelength region, Figure 4A.

Anisotropic Rotations of Substituted Pyrenes

J. Phys. Chem. B, Vol. 105, No. 42, 2001 10311

Figure 5. Decay profile and the fit (fit residuals) for total emission, F(t) ) IVV(t) + 2GIVH(t), of 1.04 µmol of 1-pyrenebutyric acid adsorbed on cab-o-sil monitored at 376 nm. Excitation was with 337 nm light of a nitrogen laser. Two exponential fit parameters are A1 ) 0.31 ( 0.01, t1 ) 49.1 ( 3.30 ns and A2 ) 0.67 ( 0.02, t2 ) 189.95 ( 3.02 ns.

TABLE 1: Anisotropy Lifetimes Obtained from Double Exponential Fit of Anisotropy Decay Profiles for Low and High Surface Coverage of Physisorbed and Chemically Attached Probes physisorbed probe

chemically attached probe

loading

τ1, ns

τ2, ns

τ1, ns

τ2, ns

low high

530 444

131 25

299 115

114 171

Negative anisotropy values throughout the excitation spectrum of the monomer (compared to a value of 0.01 in glycerol) suggest that the dipole moment for the first transition, S0 f S1, is affected by the silanol groups present on the silica surface. Molecules are aligned (laid down) on the surface to such a degree that they mainly emit horizontally polarized light. Excitation anisotropy monitored at 470 nm (not shown) is 0.15 and is invariable throughout the spectrum, suggesting 1-pyrenebutyric aggregates (dimers) interaction with the negatively charged silica surface is not very pronounced as in the case of monomer. The emission lifetimes estimated from the two exponential fit of the total emission decay at 376 nm were found to be 49.1 ( 3.3 and 190.0 ( 3.0 ns, respectively (Figure 5). A doubleexponential fit of the total emission at 470 nm gave a lifetime of 21.7 ( 1.8 ns for the short-lived component and 131.9 ( 7.7 ns for the long-lived component of the decay. The observed decrease in lifetimes may be attributable to the contribution of aggregates at this wavelength. Time-dependent anisotropy decay, which exhibits a negative rise followed by its recovery to zero baseline, is shown in Figure 6. A negative rise in the signal followed by its recovery back to the baseline is consistent with eq 3 that predicts a negative term if ξ is greater than 45° (contribution from cos 2ξ).

Figure 6. Anisotropy decay and the fit (fit residuals) for 1-pyrenebutyric acid on cab-o-sil (1.04 µmol/g) at 298 K. Excitation was with 337 nm light of a nitrogen laser. The data were fit to equation: r(t) ) 0.3 cos 2ξ exp(-t/0.530 µs) + 0.1 exp(-t/0.131 µs) where ξ ) 56.4°.

Furthermore, it suggests that the emission moment is oriented in the x-y plane and that emission depolarization takes place mostly through rotational relaxation within the same plane. Fitting the experimental time-dependent anisotropy profiles to eq 3 allows estimation of the rate constant for in-plane and outof-plane rotation along with the value for ξ. A doubleexponential fit of the anisotropy decay profile gives lifetimes of 131 and 530 ns, respectively (Table 1). Estimation of the rotational rates, DVH and DVV, and ξ using eq 3 and the above lifetimes gives the following values: DVH ) 4.59 × 107 s-1, DVV ) 8.22 × 106 s-1, and ξ ) 56.4° (Table 2). The value of 56.4° obtained for ξ using eq 3 further supports our treatment of the probe as an oblate ellipsoid of rotation with absorption and emission transition moments positioned in the plane of the molecule. For the oblate revolving ellipsoid one would anticipate the shortest rotational time to correspond to the in-plane rotation and the longest one to the out-of-plane rotation. Opposite results can be explained in terms of restriction imposed on the probe by the solid surface, provided the out of plane rotation is not a complete, but rather a wagging, motion between the energetically similar conformations of the molecule with respect to the surface. A similar segmental motion was observed in the studies of immunoglobulin G(lgG) complexed with plasma membrane receptor.59 Bb. High Surface CoVerage. Besides monomer bands observed at 376, 397, and 420 nm, the emission spectrum of 4.17 µmol of 1-pyrenebutyric acid adsorbed on cab-o-sil shows a significant excimer emission around 466 nm (not shown). The excitation spectrum monitored at 487 nm (not shown) shows a set of absorption bands similar to the monomer but shifted to longer wavelength and anticoincident to the excitation spectra monitored at 376 (monomer). Emission anisotropy (λex ) 337 nm) was found to be positive (0.05) throughout the emission

TABLE 2: Rate Constants for In-plane and out-of-plane Rotation of the Probe Physisorbed or Chemically Attached to the Surface of Cab-o-sil, and the Estimated Values for the Angle between Absorption and Emission Transition Moments physisorbed probe loading low high

DVV,

s-1

8.22 × 5.65 × 107 106

DVH,

s-1

4.59 × 3.88 × 108 107

chemically attached probe ξ, deg 56.4 ( 1.32 45.5 ( 2.0

DVV,

s-1

1.54 × 6.94 × 107 107

DVH, s-1

ξ, deg

5.29 × 106 3.5 × 107

57.7 ( 4.4 64.0 ( 1.5

10312 J. Phys. Chem. B, Vol. 105, No. 42, 2001

Ivanov et al.

Figure 7. Emission (λex ) 337 nm), excitation and anisotropy spectra of 1-pyrenebutanol attached to cab-o-sil (4 µmol/g) with 0.57 mmol/g of biphenyl as the spacer. Emission spectrum (B) was recorded when the sample was excited with 337 nm light of a 450-W Xe lamp. Excitation spectrum (A) was recorded monitoring at 397 nm. Anisotropy of the emission and excitation was estimated using the following equation: r ) IVV - GIVH/IVV + 2GIVH.

spectrum (not shown). The excitation anisotropy (λm ) 376 nm) was found to be 0.01 in the 300-360 nm wavelength region, increasing to 0.03 for the 240-280 nm wavelength region (not shown). Positive excitation and emission anisotropy suggest lesser interaction between the silica surface and physisorbed molecules for this sample compared to a low loading of 1-pyrenebuturic acid. Higher loadings of the probe on the cabo-sil surface most likely create a situation where the adsorbed molecules have to rearrange on silica active sites in order to accommodate the incoming molecules (to form aggregates). The 376 nm emission decay (not shown) was best fitted to a single-exponential, giving a lifetime of 147.4 ( 1.9 ns. The decay of total emission at 470 nm was best fitted to a double exponential, giving lifetimes of 26.17 ( 3.7 and 122.6 ( 9.8 ns, respectively. The shorter lifetime of the 376 nm emission at higher surface coverage compared to lower surface loadings may be attributable to a self-quenching process.58,60,61 We could not resolve the rise of the 450 nm emission on our time-resolved setup. This is probably due to fast,