Energy Transfer and Emission Quantum Yields ... - ACS Publications

Departamento de Física and CICECO and Departamento de Química and CICECO, Universidade de .... A. S. Ferreira , Carole Carcel , Verónica de Zea Bermud...
7 downloads 0 Views 217KB Size
J. Phys. Chem. C 2007, 111, 3275-3284

3275

Energy Transfer and Emission Quantum Yields of Organic-Inorganic Hybrids Lacking Metal Activator Centers So´ nia S. Nobre,† Patrı´cia P. Lima,†,‡ Luı´s Mafra,§ Rute A. Sa´ Ferreira,† Ricardo O. Freire,‡ Lianshe Fu,† Uwe Pischel,| Vero´ nica de Zea Bermudez,⊥ Oscar L. Malta,‡ and Luı´s D. Carlos*,† Departamento de Fı´sica and CICECO and Departamento de Quı´mica and CICECO, UniVersidade de AVeiro, 3810 -193 AVeiro, Portugal, Departamento de Quı´mica Fundamental - CCEN-UFPE, Cidade UniVersita´ ria, Recife-PE, 50670-901, Brasil, Instituto de Tecnologı´a Quı´mica, UPV-CSIC, UniVersidad Polite´ cnica de Valeˆ ncia, AVenida de los Naranjos s/n, E-46022 Valencia, Spain, and Departamento de Quı´mica and CQ-VR, UniVersidade de Tra´ s-os-Montes e Alto Douro, 5000-801 Vila Real, Portugal ReceiVed: October 18, 2006; In Final Form: December 14, 2006

This work discusses quantitatively the energy transfer mechanism that occurs in the white-light emission of sol-gel derived amine- and amide-functionalized hybrids. The white-light photoluminescence (PL) results from a convolution of the emission originated in the NH/CdO groups of the organic/inorganic cross-links with electron-hole recombinations occurring in the siloxane nanoclusters, both emissions typical of donoracceptor pairs. Two model compounds that reproduce separately the two hybrid’s emissions were synthesized and characterized by X-ray diffraction, 29Si/1H/13C magic-angle spinning NMR, diffuse reflectance, Fourier transform-IR, and photoluminescence spectroscopy to support their use as organic and inorganic structural models for the two counterparts of the hybrids. The comparison between the lifetimes of the two emissions of the inorganic and organic model compounds with those of the hybrids, the Arrhenius dependence with temperature of the siliceous-related lifetime in the hybrids, and the nonexponential behavior of the decay curve of the siliceous-related emission under lower excitation wavelengths are experimental evidence supporting the occurrence of energy transfer in the hybrids. This energy transfer rate is quantitatively estimated for d-U(600) (the diureasil host with smaller number of polymer repeat units) generalizing the ideas proposed recently for the intramolecular energy transfer between singlet and triplet ligand levels and ligand-to-metal charge transfer states in lanthanide coordination compounds. The dipole-dipole energy transfer rate between the two emitting centers is 1.3 × 109 s-1, larger than the value estimated for the transfer rate mediated by the exchange mechanism, 3.7 × 108 s-1. The predicted room-temperature emission quantum yield of that diureasil hybrid is comparable to the corresponding experimental value (7 ( 1 %), pointing out a strong dependence of the radiative component values of the two emissions with temperature, induced by the glass-rubber phase transition of the hybrid’s polymer chains.

Introduction hybrids1

Organic-inorganic are an emerging class of multifunctional nanostructured materials with tailored properties seldom seen in other types of materials and unparalleled performances suitable for promising applications in many different areas, ranging from optics and electronics to energy, environment, biology, and medicine. Applications include membranes and separation devices, functional smart coatings, a new generation of photovoltaic and fuel cells, sensors, smart microelectronics, microoptical and photonic components, systems for nanophotonics, innovative cosmetics, intelligent therapeutic vectors combining targeting, imaging, therapy and controlled release of active molecules, nanoceramic-polymer composites for the automobile or packaging industries, etc.2 The concept of “hybrid organic-inorganic” materials emerged very recently with the birth of the “soft” inorganic chemistry * Corresponding author. Tel: +351-234-370946. Fax: +351-234424965. E-mail: [email protected]. † Departamento de Fı´sica and CICECO, Universidade de Aveiro. ‡ Cidade Universita ´ ria. § Departamento de Quı´mica and CICECO, Universidade de Aveiro. | Universidad Polite ´ cnica de Valeˆncia. ⊥ Universidade de Tra ´ s-os-Montes e Alto Douro.

processes, namely the sol-gel route.3 The unique characteristics of this process (such as the low-temperature processing and shaping, high-sample homogeneity and purity, availability of numerous metallo-organic precursors, and the processing versatility of the colloidal state) permit the synthesis of multifunctional organic-inorganic hybrid structures through a molecular nanotechnology bottom-up approach based on a tailored assembly of organic, inorganic, and even biological building blocks.1,4 In the past few years, the hybrid strategy has been increasingly adopted for the development of low cost siloxane-based matrices with attractive photonic features (e.g., optical switching and data storage,5 high laser efficiency and photostability,6 photopattern waveguiding for integrated optics,7,8 electroluminescence,6,9 and nonlinear optics.10) Several stable and efficient white-light photoluminescent amine- and amide-functionalized hybrids lacking metal activator ions have been introduced, such as those obtained from 3-aminopropyltriethoxysilane (APTES), 3-glycidyloxypropyltrimethoxysilane (GPTES),11-15 urea (-NHC(dO)NH-), urethane (-NHC(dO)O-), and amide (-NHC(dO)-) crosslinked precursors, classed as diureasils,16-21 diurethanesils,22 and

10.1021/jp0668448 CCC: $37.00 © 2007 American Chemical Society Published on Web 01/27/2007

3276 J. Phys. Chem. C, Vol. 111, No. 8, 2007 diamidosils,23 respectively, and from layered perovskite-type APTES-PbCl4 hybrids.24 The APTES hybrid including formic acid is one of the most efficient phosphors known among those not containing metal activator ions,with an external photoluminescence (PL) quantum yield of 35 ( 1%.11 The APTESacetic acid analog exhibits two distinct emissions with quantum yields between 12 and 20%12 and similar values were reported for the diureasils, diurethanesils and diamidosils.21,22c,23 The potential technological relevance of these amine- or amidefunctionalized hybrids for the fabrication of several nanostructured systems, such as large-area neutron detectors,25 electrolytes for dye-sensitized photoelectrochemical cells,26 and efficient white-light room-temperature emitters,16-23 was recently demonstrated. The efficient PL (quantum yields higher than those reported for amorphous porous silicon27 and similar to the typical values of the most efficient conjugated polymers28) results from a convolution of the emission originated in the NH/CdO groups of the urea, urethane, or amide bridges with electron-hole recombinations occurring in the siloxane nanoclusters. Experimental evidence, such as the sublinearly dependence of the integrated emission on the excitation intensity, the increase of the lifetimes of the two emissions with the initial time delay, and the red-shift of the time-resolved spectra relatively to the steady state one, indicate that the two emissions are related to radiative recombination mechanisms typical of donor-acceptor (D-A) pairs.21,22c,23 While the chemical species involved in the emission related to the siliceous nanodomains were assigned to •O-O-Si≡(CO2) oxygen-related defects, detected by electron paramagnetic resonance (EPR),21b those responsible for the NH/CdO-related component have not yet been unequivocally addressed. Despite the assumption done previously associating this emission to the photoinduced proton-transfer between defects, NH2+/N- in diureasils, diurethanesils, and diamidosils or NH3+/NH- in the diureasil organic precursors (i.e., the diamines) and in the amine-functionalized hybrids based on APTES and GPTES,21b the identification of the emitting centers subjacent to that emission remains an open question. In an attempt to gain more insight into the origin and mechanisms responsible for the photonic properties of these amine-functionalized hybrids, the present work examines selectively the two emissions associated with the organic and inorganic counterparts. Two model compounds were synthesized and characterized: one reproducing the organic component of the hybrid structure and the other emulating the siliceous inorganic skeleton. Furthermore, time-resolved spectroscopy clearly demonstrates the existence of energy transfer between the two distinct emitting centers. The energy transfer rate between the Si- and NH-related states is quantitatively estimated for a diureasil hybrid, generalizing the ideas proposed recently for the intramolecular energy transfer between singlet and triplet ligand levels and ligand-to-metal charge transfer (LMCT) states in lanthanide coordination compounds.29 The predicted roomtemperature emission quantum yield of that diureasil hybrid is compared with the corresponding experimental value pointing out a strong dependence of the radiative component values of the two emissions with temperature, induced by the glassrubber phase transition of the hybrid’s polymer chains. Despite the continuous worldwide effort on the elucidation of photonic properties of amine- and amide-functionalized organicinorganic hybrids, the energy transfer mechanisms were only qualitatively discussed in a handful of reports (essentially involving energy transfer from the hybrid host and lanthanide ions.30-32) This is the first attempt to explicitly quantify the

Nobre et al. SCHEME 1: Synthesis of the Inorganic Model Compound

SCHEME 2: Synthesis of the Organic Model Compound

energy transfer rate between distinct hybrid emitting centers. The detailed characterization of these energy transfer mechanisms will definitively contribute to the recognition of the paths needed for the development of siloxane-based hybrids characterized by interesting photonic features and high light emission efficiency. Moreover, the procedure reported provides a theoretical scheme that might be useful in guiding the interpretation of experimental data and in the modeling of new organicinorganic hybrids. Experimental Section Sample Preparation. Materials. Tetraethoxysilane (TEOS, Aldrich), propyl isocyanate (PIC, 99%, Aldrich), propyltrimethoxysilane (PTMOS, 97%, Fluka), and O,O’-bis-(2-aminopropyl) polypropylene glycol-block-polyethylene glycol-blockpolypropylene glycol (commercially designated as Jeffamine ED-2001, Fluka, average molecular weight Mw ∼ 2000 g/mol-1) were used as received. Tetrahydrofuran (THF, Merck) and ethanol (Merck) were stored over molecular sieves. Synthesis of the Inorganic Model Compound. A typical synthesis of the inorganic model compound involves the mixture of 2.0 mL (8.78 mmol) of TEOS with 6.32 mL (35.1 mmol) of PTMOS, with a molar ratio TEOS/PTMOS ) 1:4 in a volume of 10.2 mL of CH3CH2OH. Then, 2.53 mL of 0.01 M HCl was added to this solution. The resulting sol was stirred in nitrogen (N2) at room temperature for 24 h. The sol was poured into a mold for gelation. Finally, the TEOS/PTMOS gel was dried at ∼50 °C for around one week. The reaction is shown in Scheme 1. Synthesis of the Organic Model Compound. The first stage of the synthesis of the organic model compound is the dissolution of 2.0 g (1.0 mmol) of Jeffamine ED-2001 in 10 mL of THF using ultrasonic agitation. Then 0.19 mL (2.0 mmol) of PIC was added to this solution with a molar ratio of Jeffamine ED-2001/PIC ) 1:2. The mixture was stirred in N2 at room temperature for 24 h. The resulting product was placed at room temperature to evaporate the solvent. Finally a solid product, Jeffamine ED-2001/PIC, was obtained. The reaction equation is described in Scheme 2. Experimental Techniques. Powder X-ray Diffraction (XRD). Powder XRD patterns were recorded on a Philips X’Pert MPD diffractometer, using Cu KR radiation (λ ) 1.54 Å), between 1 and 80° (2θ). Magic-Angle Spinning (MAS) Nuclear Magnetic Resonance (NMR). 29Si, 1H, and 13C MAS NMR spectra were recorded on a Bruker Avance 400 (9.4 T) spectrometer (DSX model) at 79.49, 400.11, and 100.62 MHz, respectively. The samples were

Organic-Inorganic Hybrids Lacking Metal Activator Centers

Figure 1. XRD patterns of the inorganic (open circles) and organic (open triangles) model compounds.

analyzed employing a 7 mm (for 29Si) and 4 mm (1H and 13C) double bearing probes. 29Si MAS NMR: A rf pulse length of 2 µs (equivalent to 30° flip angle), a recycle delay (RD) of 60 s and a spinning rate (νR) of 5.0 kHz was employed to record the spectrum. 1H MAS NMR: An excitation pulse length of 3 µs (equivalent to a 90° flip angle) was employed. RD ) 4 s and νR ) 800 Hz. 13C MAS NMR: An excitation pulse length of 2.8 µs (equivalent to 45° flip angle) was employed. RD ) 50 s and νR ) 800 Hz. 13C spectra with and without 1H decoupling were recorded (see further details in figure captions). Chemical shifts are quoted in ppm from tetramethylsilane for all nuclei. Diffuse Reflectance. The ultraviolet/visible (UV/vis) diffuse reflectance spectrum was recorded on a JASCO V-560 instrument by using a powdered sample. Fourier Transform Infrared Spectra (FT-IR). FT-IR spectroscopy of powdered samples suspended in KBr pallets was acquired between 1400 and 1900 cm-1 using a Mattson Mod 7000 spectrometer. PL Spectroscopy. The PL spectra were recorded at roomtemperature with a modular double grating excitation spectrofluorimeter with a TRIAX 320 emission monochromator (Fluorolog-3, Jobin Yvon-Spex) coupled to a R928 Hamamatsu photomultiplier, using the front face acquisition mode. The excitation source was a 450 W Xe arc lamp. The emission spectra were corrected for detection and optical spectral response of the spectrofluorimeter and the excitation spectra were weighted for the spectral distribution of the lamp intensity using a photodiode reference detector. The lifetime measurements were acquired at 14 K with the setup described for the PL spectra using a pulsed Xe-Hg lamp (6 µs pulse at half width and 20-30 µs tail). The emission spectra detected under different power excitations were recorded using a Jobin YvonSpex spectrometer (HR 460) coupled to a R928 Hamamatsu photomultiplier. The excitation source was a c.w. He-Cd laser (325 nm) coupled to neutral density filters. The spectra were corrected for the detection spectral sensitivity. Results and Discussion Local Structure. The X-ray diffraction patterns of the organic and inorganic model compounds are depicted in Figure 1. The diffraction pattern of the inorganic model compound shows two broad peaks, the most intense one centered at 24.22°, associated with the presence of amorphous siliceous domains. Using the Bragg law, the structural unit distance, d, was calculated to be 3.67 Å. The coherent length, L, over which the structural unit survives, was estimated fitting the XRD patterns with pseudoVoigt functions and using the modified Scherrer equation,22c

J. Phys. Chem. C, Vol. 111, No. 8, 2007 3277

Figure 2.

29

Si NMR MAS spectrum of the inorganic model compound.

being 11 ( 2 Å. The least intense broad peak, centered at 7.49°, was assigned to an interference effect between siliceous domains spatially correlated. Both peaks are deviated about 4° relative to the hybrid’s diffraction pattern, which can be related with (i) changes in the conformation of the siliceous skeleton upon incorporation into the hybrid, (ii) absence of the urea groups, and (iii) presence of new Si local environments. The diffraction patterns of the organic model compound, very similar to those of poly(ethylene glycol), PEG, and d-U(2000) diureasil,18b is composed of a series of well-defined Bragg peaks (the most intense ones at 19.03° and 23.23°) assigned to the diffraction of crystalline oxyethylene, OCH2CH2, unities. The 29Si MAS NMR spectra of the inorganic model compound, depicted in Figure 2, exhibit broad signals at approximately -90.9, -100.5, and -108.8 ppm assigned to (≡SiO)2Si(OH)2 (Q2, geminal silanols), (≡SiO)3SiOH (Q3, single silanol), and (≡SiO)4Si (Q4, siloxane) local environments, respectively,33 and three peaks at approximately -50.0, -55.7, and -66.3 ppm, ascribed to organosiloxane atoms R′Si(OSi)(OR)2 (T1), R′Si(OSi)2(OR) (T2), and R′Si(OSi)3 (T3), respectively.22c The 13C MAS NMR spectra of the organic model compound, Jeffamine ED-2001/PIC, are shown in Figure 3a,b with and without 1H decoupling during the signal acquisition. Figure 3b has the advantage to evidence the J-couplings of 1H-13C spin pairs since this interaction is accessible because of the high mobility of the Jeffamine ED-2001/PIC solid sample (see the insets of Figure 3c,d for more detail). Then the assignment of magnetically inequivalent 13C resonances falling in the same chemical shift region can be facilitated exploiting the Jmultiplets making the differentiation between the -CH- and -CH2- and -CH3 groups more intuitive. The 13C resonance assignments of the Jeffamine ED-2001/PIC sample may be obtained based on the typical chemical shift range of the different chemical species and on a previous assignment study of the 13C spectrum of a diureasil based cross-linked hybrid dubbed as d-U(600) containing the Jeffamine ED-2001/PIC unit as the organic part.22c The two 13C resonances located at ca. 11.2 and 11.7 ppm are assigned to the terminal -CH3 groups while the peaks at ca. 17.5 and 18.5 ppm are attributed to the different branched -CH3 groups of the molecular chain (see Scheme 2).34 The peaks at ca. 21.2 and 23.8 ppm are typical of -CH2- aliphatic carbons. In the range of 40-48 ppm are found the -NCH2- groups. The peak centered at ca. 46 ppm may be assigned to -NCHgroups because it shows a doublet shape instead of a triplet (Figure 3d). The region between 60 and 80 ppm refers to the different -OCHn (n ) 1, 2) groups. On the basis of the assignment proposed by other authors,35 the strongest peak in

3278 J. Phys. Chem. C, Vol. 111, No. 8, 2007

Nobre et al.

Figure 4. 1H MAS NMR spectrum of the Jeffamine ED-2001/PIC organic model employing. Asterisks depict spinning sidebands.

SCHEME 3: Transformation of a Urea Group into an Enol Group Figure 3. 13C MAS NMR spectra of the Jeffamine ED-2001/PIC organic model employing (a) 1H decoupling and (b) without 1H decoupling. The insets shows the (c) 60-85 and the (d) 5-50 ppm regions of (a) and (b) to better observe the J-multiplets of each carbon resonances. The value of the1JCH coupling constants ranges from ca. 92 Hz (corresponding to the 45.8 ppm peak) to ca. 141 Hz (corresponding to the 70.7 ppm peak).

this region (ca. 70.6 ppm) is attributed to the -(OCH2CH2)groups (Figure 3c). The high-frequency shifted peaks at ca. 158 and 159 ppm are typical of the urea CdO groups22c and unequivocally prove the integrity of the organic model. The two CdO environments (already observed in diureasils) strongly suggest the presence of CdO‚‚‚H-N hydrogen bonds having different strengths, which influence their chemical shifts.36 This fact may be supported by the presence of different amide proton environments centered at ca. 6 ppm (Figure 4). Unexpectedly, an additional resonance appears at ca. 149 ppm, which is absent in the d-U(600) hybrid.22c Figure 3b show that this peak has no splitting and thus it must be attributed to a quaternary carbon. This peak is located in the region of tertiary olefinic carbons (140-150 ppm)37 and the only choice to obtain a single peak in this region, considering the different Jmultiplets, is when an enol group involving the urea group is formed. This transformation can be favored because of the acidic character of the amide protons (see Scheme 3). Thus, there is also a possibility to establish hydrogen bonding, for example, of the type C-OH‚‚‚OdC-, which also may lead to the second CdO peak present in the 13C NMR spectrum. It is worth noting that this peak cannot belong to any of the precursors because neither of them contain 13C resonances in this region. In addition, a further evidence of the enol formation may be visible by the OH protons typically resonating between 6 and 7 ppm, which fall in the same region of the amide NH protons (Figure 4).38 The connection between the enol group formation and the PL features of the NH/CdO component will be further addressed in a subsequent publication. The presence of -NHC(dO)NH- groups in Jeffamine ED2001/PIC was monitored also through the analysis of the infrared

absorption bands in the “amide I” region (1500-1800 cm-1). The amide I mode is a highly complex vibration that involves primarily the contribution of the CdO stretching. The amide I region of the organic model compound (Figure 5) shows several components involving hydrogen bonding between the NH groups of the urea and the oxygen atoms of ether or carbonyl groups, like it was reported for diureasil hybrids.18a,22c To access the frequency of the vibrational modes, a deconvolution fitting procedure using Gaussian functions (Figure 5) was employed. Three components were isolated for the “amide I” envelope at ∼ 1652, 1691 and 1722 cm-1. The former peak is ascribed to the formation of strong self-associated hydrogen-bonded ureaurea associations and the remaining ones are attributed to the vibration of NHC(dO)NH groups belonging to urea-polyether chains. The absence of a band at ∼ 1750 cm-1 indicates that neither CdO nor N-H groups were left free.18a,22c All the structural data of the Jeffamine ED-2001/PIC and TEOS/PTMOS compounds support their use as, respectively, organic and inorganic structural models for the two counterparts of the organic-inorganic diureasil hybrid. In the strict sense, the Jeffamine ED-2001/PIC compound models the d-U(2000) organic chains structurally because of its polymer average molecular weight. However, the situation found in this material (i.e., the presence of larger quantities of urea-urea associations) resembles closely that observed in the diureasil with lower molecular weight d-U(600) and, therefore, this compound will be used in the next section as a model to discuss the emission component originated in the NH/CdO groups of the urea, urethane, or amide bridges. Photoluminescence. Figure 6a,b show the room-temperature emission spectra of the inorganic and organic model compounds, respectively. All the spectra are formed of a large broad band whose peak position deviates to the red as the excitation

Organic-Inorganic Hybrids Lacking Metal Activator Centers

J. Phys. Chem. C, Vol. 111, No. 8, 2007 3279

Figure 5. FT-IR spectrum of the amide I region of the organic model compound. The dashed lines represent the curve-fitting results. Figure 7. Time-resolved emission spectra of the inorganic (open circles), organic (open triangles) model compounds, and d-U(600) (full circles) excited at 300, 350, and 365 nm, respectively, with an integration window of 20.00 ms. The starting delays were set as 0.05 ms (model compounds) and 20.00 ms (hybrid). The arrow indicates the siliceous-related component in the d-U(600) emission spectrum.

Figure 6. Room-temperature emission spectra of the (a) inorganic and (b) organic model compounds excited at different wavelengths: 280 nm (open triangles), 300 nm (full circles), 330 nm (full up triangles), 350 nm (open circles), 365 nm (down full triangles) and 380 nm (open squares).

wavelength increases. This behavior was observed for the hybrids11,15,18b,19,20-23 and nonhydrolyzed precursors,19-21 and it was recently modeled as radiative recombinations involving thermal relaxation within localized states in the framework of the extended multiple trapping approach.39 The emission bands of Jeffamine ED-2001/PIC and TEOS/PTMOS compounds are well reproduced by a single Gaussian function whose energy peak position varies from 3.17 to 2.54 eV and 3.00-2.89 eV for the organic and inorganic model compound, respectively, as the excitation wavelength increases from 300 to 420 nm (4.13-2.95 eV) and 280-365 nm (4.42-3.40 eV). Comparing these values with those observed for the NH/CdO groups and siliceous-related emission in the hybrid materials,18b,c a blue shift of the emission of the Jeffamine ED-2001/PIC compound is observed, whereas the hybrids’ siliceous-related emission energy is approximately that observed in the TEOS/PTMOS model. Also, the emission spectra of the organic compound are blueshifted with respect to that of the diamine Jeffamine ED2001. Figure 7 compares the time-resolved emission spectra of the d-U(2000) diureasil hybrid and those of the model compounds. For starting delays below 5.00 ms, the hybrids’ emission displays two main components at ∼427 and ∼500 nm.18b,21 The former component were ascribed to the siliceous network and the latter one was assigned to emission arising from the NH/CdO groups within the urea cross linkages.18b,21 The emission of the model compounds is formed by a single emission component (similarly to that observed for the steady-state emission, Figure 6a,b), being that the emission of the inorganic model compound is blue-

shifted with respect to that of the organic model compound like observed in the hybrid material. Moreover, the spectral range observed for the emission of each one of the model compounds is in perfect agreement with the emission assignment proposed for the hybrids time-resolved spectra. Additional measurements were carried out to compare the physical nature behind the emission features of the model compounds with those of the hybrids, in particular the presence of radiative emission mediated by donor-acceptor pairs. The existence of donor-acceptor pairs can be inferred from the analysis of the behavior of the emission intensity, I, as the excitation power, L, is varied. The emission intensity depends on the excitation power according to the power-law I ∝ Lk. Whereas k e 1 is characteristic of D-A pairs, 1 < k < 2 is typical of exciton-like transitions.21b,40 As shown in Figure 8 the emission integrated intensity depends linearly with the power excitation, revealing a slope k of 1.00 and 0.98, for the inorganic and organic model compounds, respectively. These values are similar to those reported for different organic-inorganic hybrids, such as the diureasils21b and the diamidosils,23 indicating that the hybrids’ and the model compounds’ emission is mediated by donor-acceptors pairs. Further experimental evidence of donor-acceptor pair transitions can be inferred from timeresolved spectroscopy. For longer delay times, a red shift in the spectrum is expected because distant pairs have a smaller recombination probability.41 The respective energy levels have therefore larger lifetimes, a situation that is favored with increasing delay time. As a consequence, the recombination takes place at lower energies. The inset in Figure 8 shows for the organic model compound the energetic shift (∆E) between the energy of the maximum intensity of the time-resolved spectra, measured for a starting delay between 0.05 and 20.00 ms, and the corresponding value detected under continuous excitation. Increasing the starting delay induces an emission shift toward the red of 430 meV. For diureasil and diamidosil hybrids and for the starting delay interval between 0.05 and 20.00 ms, ∆E is 150 meV.21b,23 For the inorganic compound, ∆E is null for starting delays between 0.00 and 100.00 ms. Although ∆E * 0 was expected, this may not controvert a D-A pair recombination mechanism. In fact, a similar behavior already was reported for the emission of silicon nanocrystals mediated by donor-acceptor pairs,42 and for the siliceous-related component in diureasils and diamidosils.21b,23

3280 J. Phys. Chem. C, Vol. 111, No. 8, 2007

Nobre et al.

Figure 9. Emission decay curves of the d-U(2000) diureasil acquired at 14 K with a starting delay of 0.08 ms and monitored at 488 nm under (open circles) 365 nm and (solid triangles) 420 nm.

Figure 8. Emission integrated intensity of the inorganic (open circles) and organic (open triangles) model compounds at different power excitations. The solid lines correspond to linear fits to the data (R > 0.99). The inset shows the shift (∆E) of the organic model compound emission between the energy of the maximum intensity of the timeresolved spectra and the corresponding value measured under continuous excitation. The spectra were measured at 350 nm for a fixed acquisition window of 10.00 ms and for starting delays between 0.05 and 20.00 ms. The dashed line is a guide for the eyes.

The lifetime values for the two model compounds were monitored at 14 K within the maximum intensity bands in Figure 6 using the same experimental conditions reported for the hybrids in particular starting delays of 40.00 and 0.08 ms for the NH/CdO- and siliceous-related emissions, respectively.21a Such decay curves are well reproduced by singleexponential functions revealing lifetime values of 211.7 ( 11.4 ms and 161.9 ( 10.1 ms for the inorganic and organic model compound, respectively. At room temperature, the lifetimes of the two model compounds are of the same order of magnitude than those of the two emissions in the hybrids, 10-8-10-9 s21. Comparing the values measured at 14 K with those known for d-U(2000), the lifetime of the organic model compound is the same as the one found for the NH/CdO-related emission in the d-U(2000) diureasil, whereas the lifetime value estimated for the inorganic model compound is two orders of magnitudes higher than the value ascribed to the siliceous-related emission in the hybrid. This significant increase strongly suggests that nonradiative channels associated with the siliceous network exist in the hybrids that were not present in the isolated inorganic model compound. The Arrhenius dependence of the lifetime of the siliceous-related component in the hybrids between 14 and ∼220 K (whereas the lifetime values of the NH/CdO-related emission is constant)21a points out that such nonradiative channels may be related with the existence of energy transfer. The nonradiative desexcitation probability of the siliceous levels may be approximately described by the Mott-Seitz model, which expresses the temperature dependence of the experimental lifetimes (τ) as ∆E1

τ-1 ) τ0-1 + k1e- kBT

(1)

where τ0 is the lifetime at T ) 0 K, k1 is the migration energy rate, ∆E1 is the activation energy for desexcitation of the siliceous emitting states, and kB is the Boltzmann’s constant. With the Mott-Seitz model applied to the siliceous lifetime

dependence on the temperature (14-220 K) for the d-U(2000) data,21a the activation energy estimated is 565 cm-1. While the activation energy predicted by the Mott-Seitz model value is known to be underestimated,43 such a small value is compatible with the above suggestion that the thermally activated nonradiative mechanism associated with the siliceous-related component may involve energy transfer from the NH/CdO groups. This point will be further addressed below. Other evidence of energy transfer from the siliceous-related component to the NH/CdO groups are found by monitoring the lifetime of the latter component under selective excitation wavelength (420 nm) and under excitation wavelengths that simultaneously excite both emissions (350-400 nm).18b,c,21 For the hybrids, the NH/CdO-related decay curves depend on the starting delay time and excitation wavelength. While the decay curve acquired under 420 nm can be described by a singleexponential function independently of the selected starting delay, the decay curves excited under 350-400 nm display a nonexponential behavior, for starting delays smaller than 40.00 ms. These observations are illustrated in Figure 9 in which the decay curves monitored within the NH/CdO groups under 365 and 420 nm are plotted in a logarithmic scale, where the linear behavior can only be applied to the 420 nm decay curve. Geometry Optimization and Absorption Spectrum. The optimization of the diureasil geometry was performed considering only one organic-inorganic chain in which the silicon atoms located at the beginning and end of the chain are both bonded to three hydroxyl groups therefore preventing the polycondensation (Figure 10a). For the sake of the simplicity of the calculations, the geometry optimization was performed for the diureasil host with less polymer repeat units (i.e., the d-U(600) hybrid). The ground-state geometry for the d-U(600) was calculated using the AM144 semiempirical model implemented in the Mopac2002 program.45 The MOPAC keywords used were: PRECISE, GNORM ) 0.25, BFGS (BFGS geometry optimizer), XYZ (the geometry optimizations were performed in Cartesian coordinates), and SCFCRT ) 1.D-10 (in order to increase the SCF convergence criterion). Once the ground-state geometry of d-U(600) has been established, we have predicted their singlet and triplet excited states using configuration interaction (only single excitations, CIS) based on the intermediate neglect of differential overlap/ spectroscopic (INDO/S) technique,46,47 implemented in the ZINDO48 program. The singlet excited states frequencies and oscillator strengths for each chain were used to predict its electronic absorption spectra. For these, we have adjusted to a Lorentzian line shape compatible with the bandwidth obtained experimentally (about 25 nm). The profiles of the theoretical

Organic-Inorganic Hybrids Lacking Metal Activator Centers

J. Phys. Chem. C, Vol. 111, No. 8, 2007 3281

Figure 11. Energy level diagram and coordinate systems used to describe the interaction between the electrons involved in the π f π* and φ f φ* transitions. The indexes j and k indicate the electron that undergoes a transition between molecular orbitals of the •O-O-Si≡(CO2) oxygen-related defects and NH-based ones, respectively, in coordinate systems placed at the baricenters of the corresponding excited states.

Figure 10. (a) Calculated ground-state geometry of the d-U(600) diureasil from the AM1 model. (b) Absorption spectra of the d-U(600) diureasil: experimental (s) and theoretical (...).

TABLE 1: Energy Transfer Rates between the Levels of the Oxygen-Related Defects and Those of the NH/CdO-based Ones Due to the Exchange (W) and Dipole-Dipole (WD-D) Mechanisms in d-U(600)a d-U(600) pγSi (cm-1) pγNH (cm-1) σSi (cm-1) σNH (cm-1) ∆ (cm-1) τSi (s) τNH (s) R(cm) W (s-1) WD-D(s-1)

3915 4571 24485 22809 1676 3.5 × 10-3 163.5 × 10-3 5.7 × 10-8 3.7 × 108 1.3 × 109

φ, and φ* molecular orbitals ascribed to the electronic energy levels of the •O-O-Si≡(CO2) oxygen-related defects and the NH/CdO-based ones, respectively (Figure 11). With this approach, the exchange mechanism is taken into account without the need of using the exchange operator, and the relevant matrix element of the interaction Hamiltonian (the isotropic term) between the two electronic clouds is given by29

〈f|H|i〉 ) (

(2)

where e is the electronic charge. In accordance to Fermi’s Golden rule, the transfer rate between the levels of the •OO-Si≡(CO2) oxygen-related defects and NH/CdO-based ones is

W)

2π |〈f|H|i〉|2F p

(3)

where the temperature-dependent factor F contains a sum over Franck-Condon factors and the appropriate energy mismatch conditions. If eq 2 is substituted into eq 3, the energy transfer rate has the following form

a The parameters used for estimate these energy transfer rates are displayed also.

and experimental absorption bands are very similar (Figure 10b). The blue shift in the theoretical curve with respect to the experimental one might be attributed to the fact that only part of the structure is considered in the calculations.49 We should note however that despite this blue shift between the theoretical and experimental absorption spectra, the essential purpose of the AM1 model is the calculus of the average donor-acceptor distance, R, (average distance between the •O-O-Si≡(CO2) oxygen-related defects and the NH/CdO-based ones, respectively) from the optimized ground-state geometry of d-U(600) (Table 1). The value obtained is in very good agreement with the mean distance between the siloxane domains and the urea groups determined by single-crystal X-ray diffraction in a nonhydrolyzed lamellar-bridged silsesquioxane precursor,50 whose molecular structure is analogous to that of the amineand amide-functionalized hybrids discussed in this work. Energy Transfer Rates. Following the theoretical procedure recently proposed for the intramolecular energy transfer between singlet and triplet ligand levels and LMCT states in lanthanide coordination compounds,29 the energy transfer between the donor and acceptor centers can be estimated considering twoelectron determinantal states (|i〉 and |f〉) involving the π, π*,

e2 〈φ*|π*〉〈φ|π〉 R

W)

2π e4 * * 2 〈φ |π 〉 〈φ|π〉2F p R2

(4)

Actually, this transfer rate is highly sensitive to the distance R through the overlap integrals 〈φ*|π*〉 and 〈φ|π〉. The form of eq 3 also reflects the dominance of the exchange mechanism in the isotropic contribution to the transfer rates. As the widths at half-height of the two donor and acceptor bands, pγSi and pγNH, respectively, are of the same order, the factor F is given by29

F)

1 ln 2 2 xπ p γSi γNH

× exp

[ [( 1 4

(

{[( ) ( ) ] } 1 2 1 + p γNH p γSi

)

2∆ ln 2 (p γSi)2 1 2 1 + p γNH p γSi

2

ln 2

- 1/2

2

) ( )] 2

ln 2

( )

∆ 2 ln 2 p γSi

]

(5)

where ∆ is the energy difference between the donor and acceptor excited states of the two centers. For this purpose, the excitedstate energy of each emitting center in d-U(600) has been determined from the crossing point between the excitation and

3282 J. Phys. Chem. C, Vol. 111, No. 8, 2007

Nobre et al.

Figure 12. Room-temperature emission (dash line with open circles, excitation wavelength of 300 nm) and excitation spectra (full line with open triangles, monitored wavelength of 490 nm) of d-U(600). The • O-O-Si≡(CO2) oxygen-related and NH/CdO-based components are represented by the dot and dash lines, respectively. The crossing points between each excitation and emission curves (black crosses) determine the excited-state energy of each emitting center.

emission curves (Figure 12). These transition energies in wavenumbers (σSi and σNH) together with the ∆ energy difference and the widths at half-height of the two bands (pγSi and pγNH) are given in Table 1. The energy transfer rate between the levels of the oxygenrelated defects and those of the NH/CdO-based ones due to the dipole-dipole mechanism (WD-D) can be given by a Fo¨rsterDexter type expression

WD-D )

2π SSi SNH F p G G R6 Si

(6)

NH

where SSi and SNH are the dipole strengths of the π f π* and φ f φ* transitions, respectively, in units of (esu)2 cm2, which can be given in terms of the radiative lifetimes (τr) and transitions energies in wavenumbers by

pc3 3 S) 4 (2πcσ)3 τ

Figure 13. Configurational-coordinate diagram for the d-U(600) diureasil illustrating the radiative and nonradiative processes of the two emitting centers and the energy transfer between them.

appropriate rate equations for the level populations in the steadystate regime are

-(A3 + W′η1)η3 + Wη2η4 ) 0 -(A2 + Wη4)η2 + (Φ + W′η3)η1 ) 0 -(Φ + W′η3)η1 + (A2 + Wη4)η2 ) 0 -Wη4η2 + (A3 + W′η1)η3 ) 0

(8)

where Φ is the absorption rate in the •O-O-Si≡(CO2) oxygenrelated defects, A represents total decay rates (Arad + ANrad, radiative and nonradiative decay rates, respectively), W and W’ are the forward and backward energy transfer rates, respectively, and η stands for the normalized level populations satisfying the condition

η1 + η2 + η3 + η4 ) 1 From the first and second equations in eq 8, we get

Φη1 ) A2η2 + A3η3 (7) r

GSi and GNH in eq 6 are the degeneracies of the levels of the •O-O-Sit(CO ) oxygen-related defects and those of the NH/ 2 CdO-based ones, respectively. Since the excited-state radiative lifetimes of these centers (τSi and τNH) are considerably long (Table 1), it is reasonable to assume that these states have a triplet like character, while the respective ground-states are both singlets. Thus, GSi ) 3 and GNH ) 1. However, the usual electric dipole selections rules may be considerably relaxed through vibronic-spin-orbit perturbations, particularly if promoting modes become operative at higher temperatures. The dipole strengths were initially estimated from the experimental lifetimes of the model compounds (i.e., donor in the absence of acceptor and Vice-Versa), measured at low temperature (14 K) to circumvent the contributions of the thermally activated nonradiative component. Rough estimates of the overlap integrals 〈φ*|π*〉 and 〈φ|π〉 indicate that they may assume values between 0.01 and 0.1.29 For R ) 5.7 × 10-8 cm, we have assumed the rather conservative value of 0.01. Table 1 gathers the estimated energy transfer rates, as well as the values of all the parameters required for the calculations. Rate Equations. To rationalize the total emission quantum yield of the d-U(600) diureasil (experimental value 7 ( 1%21), we considered the four level system in Figure 13. The

(9)

(10)

The total emission quantum yield, q, which is the ratio between the numbers of emitted and absorbed photons, is given by:

q)

rad Arad 1 2 η2 + A3 η3 ) Φη1 1+x

(11)

where x is

x)

A 2η 2 rad A 2 η2

+ A3η3 + Arad 3 η3

(12)

The experimental room-temperature quantum yield of dU(600) (7 ( 1%) and the temperature dependence of the lifetimes of the excited states (ranging from 10-3 s, between 12 and ∼220 K, to 10-8-10-9 s, at room temperature) can be explained only by assuming that the radiative component values at room temperature are much higher than the corresponding ones measured from the lifetimes at low temperature (∼5-10 s-1, Table 1). We can see no other paths to explain these experimentally observed results. This increase in the radiative component values could be induced by the glass-rubber phase transition of the polymer chains of d-U(600) for which the glass transition temperature Tg is ∼220-250 K.51 This coincides with the temperature range in which the lifetimes of the two emitting

Organic-Inorganic Hybrids Lacking Metal Activator Centers

J. Phys. Chem. C, Vol. 111, No. 8, 2007 3283

states decrease abruptly from the millisecond to the nanosecond scale. Therefore, as the polymer chains soften and pass from a stiff-glassy state to a rubbery one, the configurationalcoordinates of the two emitting molecular states change. This possibly induces the appearance of new promoting modes within the parabolas increasing the vibronic-spin-orbit interaction and subsequently the radiative decay rates. An increase from 5-10 s-1 to 106-107 s-1 would lead to a value of x in eq 12 of the order of 10 and, consequently, to a value of the emission quantum yield of the order of the experimental one. In this case, eqs 6 and 7 lead to a dipole-dipole energy transfer rate between the two emitting centers of the order of 109 s-1, which is at least as high as the transfer rate estimated by the exchange mechanism (Table 1).

energy ∼565 cm-1), as well as the nonexponential behavior of the decay curve of the siliceous-related emission under lower excitation wavelengths (both emissions simultaneously) also support the occurrence of this energy transfer process. This energy transfer rate is quantitatively estimated for d-U(600) (the diureasil host used with smallest number of polymer repeat units) generalizing the ideas proposed recently for the intramolecular energy transfer between singlet and triplet ligand levels and ligand-to-metal charge transfer states in lanthanide coordination compounds.29 The dipole-dipole energy transfer rate between the two emitting centers of the d-U(600) diureasil is 1.3 × 109 s-1, larger than the value estimated for the transfer rate mediated by the exchange mechanism, 3.7 × 108 s-1. The predicted roomtemperature emission quantum yield of that diureasil hybrid is comparable to the corresponding experimental value (7 ( 1%), pointing out a strong dependence of the radiative component values of the two emissions with temperature, induced by the glass-rubber phase transition of the hybrid’s polymer chains. To the best of our knowledge, this is the first attempt to explicitly quantify the energy transfer rate between distinct hybrid emitting centers. The detailed characterization of these energy transfer mechanisms will definitively contribute to the recognition of the paths needed for the development of siloxanebased hybrids characterized by interesting photonic features and high-light emission efficiency. Moreover, the procedure reported provides a theoretical scheme that might be useful in guiding the interpretation of experimental data and in the modeling of new organic-inorganic hybrids.

Conclusions In an attempt to shed more light on the origin and mechanisms responsible for the photonic properties of amine- and amidefunctionalized organic-inorganic hybrids, this work examined independently the two emissions associated with their corresponding organic and inorganic counterparts. Two model compounds reproducing, respectively, the organic (Jeffamine ED-2001/PIC) and inorganic (TEOS/PTMOS) parts of a diureasil hybrid were synthesized and characterized in detail by XRD, 29Si/1H/13C MAS NMR, diffuse reflectance, FT-IR, and PL spectroscopy. XRD, 29Si/1H/13C MAS NMR, and FT-IR of Jeffamine ED-2001/PIC and TEOS/PTMOS support their use as, respectively, organic and inorganic structural models for the two counterparts of the diureasil hybrids. The emission bands of Jeffamine ED-2001/PIC and TEOS/ PTMOS compounds are well reproduced by a single Gaussian, whose energy peak changes from 3.17 to 2.54 eV and 3.002.89 eV for the organic and inorganic model compounds, respectively, as the excitation wavelength increases from 300 to 420 nm (4.13-2.95 eV) and 280-365 nm (4.42-3.40 eV). Comparing these values with those observed for the NH/Cd O- and siliceous-related emissions in the amine- and amidefunctionalized hybrids, we observed a blue shift of the emission of the Jeffamine ED-2001/PIC compound, whereas the hybrids’ siliceous-related emission energy is approximately that observed for the TEOS/PTMOS model compound. The linear dependence of the emission integrated intensity of both materials on the power excitation, slopes of ∼1.00 and 0.98, for the inorganic and organic model compounds, respectively, and the energetic shift between the energy of the maximum intensity of each timeresolved spectra, measured for starting delays between 0.05 and 20.00 ms, and the corresponding value detected under continuous excitation, point out that the model compounds emission is mediated also by donor-acceptors pairs, as it has been reported for the diureasils and diamidosils.21b,23 The decays curves of the two materials are well reproduced by single-exponential functions revealing lifetime values of 211.7 ( 11.4 ms and 161.9 ( 10.1 ms, for the inorganic and organic model compound, respectively. Comparison of these values with those for d-U(2000) reveals that the lifetime of Jeffamine ED-2001/PIC is the same as that found for the NH/ CdO-related emission, whereas the value estimated for TEOS/ PTMOS is two orders of magnitudes higher than the value ascribed to the siliceous-related emission in the hybrid. This significant increase suggests the existence of thermally activated nonradiative mechanism associated with the siliceous-related component that may involve energy transfer from the NH/Cd O groups. The Arrhenius dependence with temperature of the siliceous-related lifetime in the d-U(2000) hybrid (activation

Acknowledgment. The support of NoE “Functionalised Advanced Materials Engineering of Hybrids and Ceramics” (FAME) is gratefully acknowledged. This work was partially supported also by FEDER and Fundac¸ a˜o para a Cieˆncia e Tecnologia (Portuguese Agency, program CTM/59075/04), CAPES and CNPq (Brazilian Agencies), and the RENAMI project (Brazilian Molecular and Interfaces Nanotechnology Network). The authors thank N. J. O. Silva (University of Aveiro) for his help with the XRD analysis and IMMC (Instituto do Milnio de Materiais Complexos) for providing the computational facilities. Note Added After ASAP Publication. This article was published ASAP on January 27, 2007. In the text, the term “Sparkle/AM1” was changed to “AM1”. The correct version was published on February 1, 2007. References and Notes (1) Functional Hybrid Materials, Gomez-Romero, P.; Sanchez, C., Eds.; Wiley-Interscience: New York, 2003. (2) Sanchez, C.; Julia´n, B.; Belleville, P.; Popall, M. J. Mater. Chem. 2005, 15, 3559. (3) Brinker, C. J.; Scherer, G. W. Sol-Gel Science, The Physics and Chemistry of Sol-Gel Processing; Academic Press: San Diego, CA, 1990. (4) Sanchez, C.; Soler-Illia, G. J. de A. A.; Ribot, F.; Lalot, T.; Mayer, C. R.; Cabuil, V. Chem. Mater. 2001, 13, 3061. (5) Sanchez, C.; Ribot, F.; Lebeau, B. J. Mater. Chem. 1999, 9, 35. (6) Sanchez, C.; Lebeau, B. MRS Bull. 2001, 26, 377. (7) Buestrich, R.; Kahlenberg, F.; Popall, M.; Dannberg, P.; M.-Fiedler, R.; Ro¨sch, O. J. Sol-Gel Sci. Technol. 2001, 20, 181. (8) Molina, C.; Moreira, P. J.; Gonc¸ alves, R. R.; Sa´ Ferreira, R. A.; Messaddeq, Y.; Ribeiro, S. J. L.; Soppera, O.; Leite, A. P.; Marques, P. V. S.; de Zea Bermudez, V.; Carlos, L. D. J. Mater. Chem. 2005, 15, 3937. (9) Dantas de Morais, T.; Chaput, F.; Boilot, J. P.; Lahlil, K.; Darracq, B.; Levey, Y. AdV. Mater. 1999, 11, 107. (10) Innocenzia, P.; Lebeau, B. J. Mater. Chem. 2005, 15, 3821. (11) Green, W. H.; Le, K. P.; Grey, J.; Au, T. T.; Sailor, M. J. Science 1997, 276, 1826. (12) (a) Bekiari, V.; Lianos, P. Langmuir 1998, 14, 3459. (b) Bekiari, V.; Lianos, P. Chem. Mater. 1998, 10, 3777.

3284 J. Phys. Chem. C, Vol. 111, No. 8, 2007 (13) Uchida, Y.; Nobu, Y.-I.; Momiji, I.; Matsui, K. J. Sol-Gel Sci. Technol. 2000, 19, 705. (14) (a) Han, Y.; Lin, J.; Zhang, H. J. Mater. Lett. 2002, 54, 389. (b) Cordoncillo, E.; Guaita, F. J.; Escribano, P.; Philippe, C.; Viana, B.; Sanchez, C. Opt. Mater. 2001, 18, 309. (15) (a) Brankova, T.; Bekiari, V.; Lianos, P. Chem. Mater. 2003, 15, 1855. (b) Stathatos E.; Lianos, P.; Orel, B.; Vuk, A. S.; Jese, R. Langmuir 2003, 19, 7587. (16) De Zea Bermudez, V.; Carlos, L. D.; Duarte, M. C.; Silva, M. M.; Silva, C. J.; Smith, M. J.; Assunc¸ a˜o, M.; Alca´cer, L. J. Alloys Compd. 1998, 277, 21. (17) Ribeiro, S. J. L; Dahmouche, K.; Ribeiro, C. A.; Santilli, C. V.; Pulcinelli, S. H. J. J. Sol-Gel Sci. Technol. 1998, 13, 427. (18) (a) De Zea Bermudez, V.; Carlos, L. D.; Alca´cer, L. Chem. Mater. 1999, 11, 569. (b) Carlos, L. D.; De Zea Bermudez, V.; Sa´ Ferreira, R. A.; Marques, L.; Assunc¸ a˜o, M. Chem. Mater. 1999, 11, 581. (c) Sa´ Ferreira, R. A.; Carlos, L. D.; de Zea Bermudez, V. Thin Solid Films 1999, 343, 470. (19) Stathatos, E.; Lianos, P.; Stangar, U. L.; Orel, B.; Judeinstein, P. Langmuir 2000, 16, 8672. (20) Bekiari, V.; Stathatos, E.; Lianos, P.; Stangar, U. L.; Orel, B.; Judeinstein, P. Chem. Mater. 2000, 12, 3095. (21) (a) Carlos, L. D.; Sa´ Ferreira, R. A.; de Zea Bermudez, V.; Ribeiro, S. J. L. AdV. Func. Mater. 2001, 2, 111. (b) Carlos, L. D.; Sa´ Ferreira, R. A.; Pereira, R. N.; Assunc¸ a˜o, M.; de Zea Bermudez, V. J. Phys. Chem. B. 2004, 108, 14924. (22) (a) Gonc¸ alves, M. C.; de Zea Bermudez, V.; Sa´ Ferreira, R. A.; Ostrovskii, D.; Rocha, J.; Carlos, L. D. Chem. Mater. 2004, 16, 2530. (b) Gonc¸ alves, M. C.; Silva, N. J. O.; de Zea Bermudez, V.; Sa´ Ferreira, R. A.; Carlos, L. D.; Dahmouche, K.; Santilli, C. V.; Ostrovskii, D.; Correia Vilela, I. C.; Craievich, A. F. J. Phys. Chem. B. 2005, 109, 20093. (c) Fu, L. S.; Sa´ Ferreira, R. A.; Silva, N. J. O.; Carlos, L. D.; de Zea Bermudez, V.; Rocha. J. Chem. Mater. 2004, 16, 1507. (23) Nunes, S. C.; de Zea Bermudez, V.; Cybinska, J.; Sa´ Ferreira, R. A.; Legendziewicz, J.; Carlos, L. D.; Silva, M. M.; Smith, M. J.; Ostrovskii, D.; Rocha, J. J. Mater. Chem. 2005, 15, 3937. (24) Cheng, Z. Y.; Gao, B. X.; Pang, M. L.; Wang, S. Y.; Han, Y. C.; Lin. J. Chem. Mater. 2003, 15, 4705. (25) Im, H. J.; Willis, C.; Stephan, A. C.; Pawel, M. D.; Saengkerdsub, S.; Dai, S. Appl. Phys. Lett. 2004, 84, 2448. (26) Stathatos, E.; Lianos, P.; Lavrencic-Stangar, U.; Orel, B. AdV. Mater. 2002, 14, 354. (27) Pavesi, L.; Dal Negro, L.; Mazzoleni, C.; Franzo´, G.; Priolo, F. Nature 2000, 408, 440. (28) Greenham, N. C.; Samuel, I. D. W.; Hayes, G. R.; Philips, R. T.; Kessenger, Y. A. R. R.; Moratti, S. C.; Holmes, A. B.; Friend, R. H. Chem. Phys. Lett. 1995, 241, 89.

Nobre et al. (29) Faustino, W. M.; Malta, O. L.; de Sa´, G. F. J. Chem. Phys. 2005, 112, 054109. (30) Sa´ Ferreira, R. A.; Carlos, L. D.; Gonc¸ alves, R. R.; Ribeiro, S. J. L.; de Zea Bermudez, V. Chem. Mater. 2001, 13, 2991. (31) Fu, L. S.; Sa´ Ferreira, R. A.; Silva, N. J. O.; Fernandes, J. A.; Ribeiro-Claro, P.; Gonc¸ alves, I. S.; de Zea Bermudez, V.; Carlos, L. D. J. Mater. Chem. 2005, 15, 3117. (32) Lima, P. P.; Sa´ Ferreira, R. A.; Freire, R. O.; Paz, F. A. A.; Fu, L. S.; Alves, S., Jr.; Carlos, L. D.; Malta, O. L. ChemPhysChem 2006, 7, 735746. (33) Kolodziejsli, W.; Corma, A.; Navarro, M. T.; Perez-Parientte, J. Solid State Nucl. Magn. Reson. 1993, 2, 253. (34) Stothers, J. B. Carbon-13 NMR Spectroscopy; Academic: New York, 1972. (35) Ribot, F.; Lafuma, A.; Eychenne-Baron, C.; Sanchez, C. AdV. Mater. 2002, 14, 1496. (36) Asakawa, N.; Kuroki, S.; Kurosu, H.; Ando, I.; Shoji, A.; Ozaki, T. J. Am. Chem. Soc. 1992, 114, 3261. (37) Kalinowsky, H. O.; Berger, S.; Braun, S. 13C NMR Spektroskopie; Georg Thieme Verlag: Stuttgart, Germany, 1984. (38) Bergens, S. H.; Bosnich, B. J. Am. Chem. Soc. 1991, 113, 958. (39) (a) Sa´ Ferreira, R. A.; Ferreira, A. L.; Carlos, L. D. Eur. Phys. J. B. 2006, 50, 371. (b) Sa´ Ferreira, R. A.; Ferreira, A. L.; Carlos, L. D. J. Non-Cryst. Solids 2006, 352, 1225. (40) Schmidt, T.; Lischka, K.; Zulehner, W. Phys. ReV. B. 1992, 45, 8989. (41) Koyama, H.; Ozaki, T.; Koshida, N. Phys. ReV. B. 1995, 52, 11561. (42) Zhuravlev, K. S.; Gilinsky, A. M.; Kobitsky, A. Y. Appl. Phys. Lett. 1998, 73, 2962. (43) Canny B., Curie D. In AdVances in NonradiatiVe Processes in Solids; DiBartolo, B., Ed.; Plenum Press: New York, 1991; p 1. (44) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 3902. (45) Stewart J. J. P. MOPAC 2002; Fujitsu Limited: Tokyo, Japan, 1999. (46) Ridley, J. E.; Zerner, M. C. Theor. Chim. Acta 1976, 42, 223. (47) Zerner, M. C.; Loew, G. H.; Kirchner, R. F.; Mueller-Westerhoff, U. T. J. Am. Chem. Soc. 1980, 102, 589. (48) Zerner, M. C. ZINDO Manual, QTP; University of Florida: Gainesville, FL, 1990. (49) Batista, H. J.; Longo, R. L. Int. J. Quantum Chem. 2001, 90, 924. (50) Moreau, J. J. E.; Pichon, B. P.; Wong Chi Man, M.; Bied, C.; Pritzkow, H.; Bantignies, J.-L.; Dieudonn, P.; Sauvajol, J.-L. Angew. Chem., Int. Ed. 2004, 43, 203. (51) (a) Armand, M.; Poinsignon, C.; Sanchez, J. Y.; de Zea Bermudez, V.; U. S. Patent 5,283310, 1994. (b) de Zea Bermudez. Ph.D. Thesis, University of Grenoble, France, 1992.