J. Phys. Chem. B 2006, 110, 25707-25715
25707
Photoluminescence Response of Terbium-Exchanged MFI-Type Materials to Si/Al Ratio, Texture, and Hydration State Carmen Tiseanu,*,† Michael U. Kumke,‡ Vasile I. Parvulescu,§ Andre´ Gessner,‡ Bogdan C. Gagea,| and Johan A. Martens| National Institute for Laser, Plasma, and Radiation Physics, P.O. Box MG-36, RO 76900, Bucharest-Magurele, Romania, Institute of Chemistry, Physical Chemistry, UniVersity of Potsdam, Karl-Liebknecht-Strasse 24-25, 14476 Potsdam-Golm, Germany, UniVersity of Bucharest, Department of Chemical Technology and Catalysis, 4-12 Regina Elisabeta BVd., Bucharest 030016, Romania, and Centre for Surface Chemistry and Catalysis, KU LeuVen, 23 Kasteelpark Arenberg, B-3001 LeuVen (HeVerlee), Belgium ReceiVed: June 13, 2006; In Final Form: August 15, 2006
Terbium-exchanged MFI zeolite type materials, i.e., microporous-mesoporous Zeotile-1 with the Si/Al ratio in the range 33-200, Zeogrid with the Si/Al ratio of 75, and nanocrystalline MFI with the Si/Al ratio of 75, were prepared via an ion exchange procedure. All of these zeolites were investigated by means of timeresolved photoluminescence techniques in various hydration states: as-synthesized (hydrated), calcined (heated at 450 °C in air), and rehydrated (after a six-month exposure to the atmospheric moisture). The photoluminescence decays and spectra were analyzed by discrete exponential fitting, distribution lifetimes analysis, and area-normalized time-resolved photoluminescence spectra. The results sustained a single average terbium species coordinated to both water molecules and framework oxygens in the hydrated zeolites. The framework contribution increased with the Si/Al ratio in Zeotile-1 and was greatest for the nanocrystalline MFI zeolite. For the calcined Zeotile-1 and Zeogrid, two main terbium species of different environments were found. For the nanocrystalline Tb3+-MFI, a distinct number of species could not be inferred, indicating a more heterogeneous distribution. Rehydration further differentiated among the Tb3+-exchanged zeolites. Photoluminescence line shape and decay of Tb3+-Zeotile-1 were between those of the hydrated and calcined states indicating a slow rehydration rate in contrast with the photoluminescence properties of Tb3+-MFI, which fully recovered the values of the hydrated state. Tb3+-Zeogrid presented an intermediate case: while the PL line shape was fully restored to that measured for the hydrated sample, the decay was still longer than that measured with the hydrated sample. Terbium photoluminescence response related to zeolite texture, Si/ Al ratio, and hydration state suggest different sitting and location of terbium in Zeotile-1, Zeogrid, and nanocrystalline MFI materials. In mesoporous Zeotile-1 and Zeogrid, the results sustained two types of terbium sites: one on the internal surface of mesopores, the other inside the pores, while for the nanocrystalline MFI, terbium sites inside the pores predominate.
1. Introduction Zeolites are microporous materials for which many and very interesting applications have been found.1 Although they exhibit large surface areas, these applications are limited to small molecules because molecules with a kinetic diameter higher than 8 Å can hardly penetrate inside the micropores. Under these conditions, a very attractive alternative was represented by the mesoporous materials.2 However, even if mesoporous materials offer great advantages in terms of diffusion properties, many of the chemical and physical applications are limited by the lack of atomic order in the pore walls. Therefore, it is of great interest to synthesize a mesoporous material with crystalline walls that can combine the advantages of both mesoporous and crystalline structures. The first attempts proved that such materials can present a better thermal and hydrothermal stability but also a higher acidity when compared with the amorphous * Corresponding author. E-mail:
[email protected]. † National Institute for Laser, Plasma and Radiation Physics. ‡ Institute of Chemistry, Physical Chemistry. § Department of Chemical Technology and Catalysis. | Centre for Surface Chemistry and Catalysis.
analogue.3-6 Another approach in preparation of mesoporous materials with crystalline walls was proposed by Martens. According to this approach, a family of such materials had been synthesized by stapling together, with the help of secondary templates, MFI precursor units (nanoslabs).7-8 Zeolite socony Mobil-FIve (MFI) structure consists of five-membered ring building units that are linked to form larger 10-ring channels. The interconnection of these chains leads to the formation of the structures of MFI-type9,10 Besides acidity, the hydrophobicity is also an important property because it can differentiate the catalytic behavior in superficial selectively controlled reactions. Ion exchange with cations may provide information about the nature of the location of exchangeable sites, their strength, and the ratio of exchangeable sites inside the pores and on the external surface.11 Because both mesoporous hybrid Zeotile-1 (ZT1) and Zeogrid (ZG) exhibit a very large external surface, the existence of pore exchangeable sites represents good evidence for the hierarchical structure. Lanthanide (Ln3+) porous materials have attracted a great interest in recent years due to their potential application in
10.1021/jp0636827 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/29/2006
25708 J. Phys. Chem. B, Vol. 110, No. 51, 2006 luminescent materials, phosphors, and lasers.12-15 Through the adjustment of zeolite type, Si/Al ratio, and Ln3+ loadings, the fine-tuning of the optical properties was obtained.14 On the other side, lanthanide ions were employed for more than 20 years as photoluminescence (PL) probes for zeolites structures.16-19 Compared to faujasites, the distribution and coordination of lanthanides in MFI or mesoporous-type materials such as MCM41 are less understood. In a first time-resolved photoluminescence (TRES) characterization of Eu3+-doped MFI, it was found that the coordination number in MFI is smaller than in water, with framework oxygen atoms displacing water from the Eu3+ coordination sites.20 More recent studies on the intrazeolitic redox process of Ce3+-Ce4+ in MFI evidenced that most of the Ce3+ resides at the intersecting 10-ring channels.21 FTIR investigations on the La-exchanged MFI pointed to the fivemembered ring as cation sites for the lanthanum ions.22 Density functional calculations determined that lanthanum ions are favorably accommodated in the two 6-T rings of the straight channels of MFI,23 where they coordinate to six oxygen atoms; four of the oxygen atoms come from the zeolite framework, while the other two come from the hydrolysis of lanthanide hydrate. The few data on the lanthanide distribution in MCM41 type mesoporous materials point to a single europium species located in the low-symmetry sites.24 In a nanosized Tb3+MCM-41 synthesized by means of sol-gel-assisted selfassembly, it is claimed that the lanthanide ion was incorporated within a mesoporous structure experiencing up to four different chemical environments.25 The traditional lanthanide ions used as luminescent structural probes for zeolites are Eu3+ and Tb3+ ions. Compared to Eu3+, terbium transitions do not offer the advantage of an internal standard for the environment symmetry due to its large J multiplicities.26 However, terbium energy levels structure determines that the metastable level located in green spectral region (5D4-7F6) is generally free of nonradiative processes via cross relaxation. Therefore, the discrimination of a multisite effect against energy transfer processes from a nonexponential pattern of the luminescence dynamics is simpler.27 On the contrary, the ion blue emission (5D3-7F6) is generally subject to a strong quenching via cross relaxation. This is a concentration-dependent process and therefore may provide information on the clustering effects.28 Further, the energy gap between the metastable state and the ground-state manifold is larger than that of Eu3+ (at ca. 15 000 cm-1). Therefore, the coupling of the terbium-excited states to the proximate OH oscillators (with energy ∼3300-3500 cm-1) inherently present in the sol-gel materials, occurs to the fourth vibrational overtone consistent with a less efficient PL quenching for Tb3+ compared to Eu3+.29 The present study is a continuation of our first work on terbium-exchanged Zeotile-1 materials.27 Here we present a comparative investigation of PL properties of terbium exchanged in Zeotile-1 (ZT1), Zeogrid (ZG), and nanocrystalline MFI (nMFI) materials.To discriminate between sitting and distribution of terbium ions in each zeolite, terbium photoluminescence response is investigated in depth with relation to zeolite texture, Si/Al ratio, and hydration state. II. Experimental Section II.1. Materials Synthesis. Nanoslabs with MFI structure type are prepared through hydrolysis of tetraethyl orthosilicate (TEOS) in a concentrated solution of tetrapropylammonium hydroxide (TPAOH).30 Aluminum-containing nanoslabs were prepared by first dissolving the appropriate amount of Al metal powder in TPAOH 40 wt % followed by addition of TEOS
Tiseanu et al. and water using the procedure for all silica nanoslabs. The final composition of the nanoslab clear solution was Si(9-x)/Al(x)/ TPAOH(25)/H2O(400). The aluminum-containing nanoslabs were further used to prepare hybrid (ZT1) and (ZG) zeolites. Small crystals of MFI zeolite were prepared by heating the nanoslab suspension at 120 °C for three days in a stainless steel autoclave.31 The nanocrystals were separated and washed with water three times by centrifugation at 12 000 rpm, dried at 60 °C, and calcined at 550 °C for 5 h in an air oven. ZT1 materials were prepared following the synthesis method reported previously.7 The nanoslab suspension was mixed with 10 wt % aqueous cetyltrimethylammonium bromide (CTABr) solution at 80 °C under stirring. The precipitate was then separated by filtration, washed, dried at 60 °C for 12 h, and calcined at 550 °C for 5 h using a 1 °C/min. Zeogrid materials were prepared by dropwise addition of CTABr solution (12 wt % in ethanol) under vigorous stirring at room temperature.31-33 The precipitate was left under quiescent conditions for 24 h and then recuperated by filtration and washed with ethanol. The sample was dried overnight at 60 °C and then calcined at 400 °C, one day in N2, and three days in O2. The calcined materials were subjected to an ionic exchange treatment following a previously reported procedure.19 Then 1 g of freshly calcined material was added in 25 mL of 0.004 M Tb3+ aqueous solution and stirred for 30 min at 80 °C. The samples were then recuperated and dried in an air oven at 120 °C for 24 h. Calcination was pursued in air at 450 °C at a heating rate of 5 °C/min. II.2. Structural Characterization. Characterization of these materials was carried out using adsorption-desorption isotherms of nitrogen at 77 K and XRD. Sorption isotherms of N2 were obtained at 77 K with a Micromeritics ASAP 2000 apparatus after outgassing the samples at 423 K for 12 h under vacuum. The XRD patterns were obtained with a SIEMENS D-5000 diffractometer at 40 kV-50 mA, equipped with a variable-slit diffracted-beam monochromator and scintillation counter. The diffraction patterns were recorded in the range 0-80° 2θ using Cu KR radiation (λ ) 1.54183 Å). The N2 adsorption isotherm of Tb3+-ZT1 samples showed a hysteresis loop characteristic to mesoporous materials with a mean pore size of 2.2-2.5 nm (BJH method) and a BET surface area of around 1000 m2/g. Zeolites notation, Si/Al ratio, and Tb3+ concentration (determined by ICP-AES) together with the textural results are gathered in Table 1. As terbium is found only as trivalent ion, “Tb” notation was also used throughout the paper. Tb3+-Zeogrid samples present a type I isotherm due to the presence of only supermicropores of around 1.3 nm and a BET surface area of 1627 m2/g. The Tb3+-nMFI sample exhibited a type I isotherm with an extra adsorption step at very high p/po values from N2 condensation in the interparticle volume created by the small zeolite crystals. The BET surface area for nMFI was of 450m2/g. Figure 1 shows the XRD patterns of the investigated samples. The XRD pattern of Tb3+-nMFI confirms the MFI structure of high crystallinity, while for the mesoporous materials, only low-angle peaks were observed. Tb3+-ZT1 samples present a hexagonal symmetry, while the Tb3+-ZG sample has a laminar structure. The absence of diffraction lines at higher 2θ angles proves that the ZT1 and ZG materials are built from zeolite precursor units that are smaller than the XRD detection limit. II.3. Photoluminescence Measurements. The stationary luminescence measurements were carried out at 20 °C using a Fluoromax 3 spectrofluorometer (Jobin Yvon), operated in the
Terbium-Exchanged MFI-Type Materials
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25709
TABLE 1: Characteristics of Tb3+-exchanged Zeolites
no.
notation
system
1 2 3 4 5 6 7 8
Tb3+-ZT1(33) Tb3+-ZT1(50) Tb3+-ZT1(66) Tb3+-ZT1(75) Tb3+-ZT1(100) Tb3+-ZT1(150) Tb3+-ZT1(200) Tb3+-ZG
9
Tb3+-nMFI
Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI-hexagonal structure Tb3+-zeolite MFI layered structure, distance between planes of ca. 1.5 nm Tb3+-zeolite MFI with small crystals (200 nm)
phosphorescence mode. The repetition rate of the flash lamp was 25 Hz, the integration window 0.5 ms, the delay after flash was set to 0.1 ms, and 50 flashes were accumulated per data point. The slits were set to 2 nm spectral bandwidth in excitation as well as emission. In the time-resolved luminescence measurements, a nitrogen laser (LTB) operated at 10 Hz was employed. The luminescence was detected at a right-angle geometry in a spectral range of 450 nm < λem < 650 nm using an intensified CCD camera (Andor DH720-18H-13, Andor) equipped with a spectrograph (MS257, Oriel Instruments). The detection window was set to 10 µs, which was shifted in steps of 20 µs over a total time of 3.0 ms for the hydrated Tb3+-zeolites and in steps of 60-80 µs over a total time of 12-16 ms for the calcined Tb3+-zeolites. A typical acquisition time of a complete TRES experiment was about 20 min.
Si/Al ratio
Tb3+ content (wt %)
BET surface (m2 g-1)
pore size (nm)
33 50 66 75 100 150 200 75
1.05 0.86 0.71 0.62 0.49 0.41 0.34 0.61
962 986 990 1045 1176 1123 992 1627
2.2 2. 2 2.2 2.2 2.5 2.5 2.5 1.4
75
0.60
450
0.54
decay time. The average decay times were calculated using the n n Aiτi2/∑i)1 Aiτi. The PL decays following formula: 〈τ〉 ) ∑i)1 were also analyzed according to a lifetime distribution analysis with decay times equally spaced in ln(τ), allowing for symmetric distributions as well as Voigt profiles (Gaussian and Lorentzian mixture). The decay curve is given by: I(t) ) ∑j R(ln(kj)) exp(-kjt) or, in terms of normalized intensity I(t)/I(0) ) ∑j R(ln(kj)) exp(-kjt)/∑j R(ln(kj)), where k ) 1/τ is the decay rate and R(ln(kj)) is the relative weight of kj in the total distribution.34 The time-resolved photoluminescence spectra (TRES) were analyzed as both peak-normalized and area-normalized (TRANES) forms. TRANES were constructed by normalizing the area of each spectrum in TRES such that the area of the spectrum at time t is equal to the area of the spectrum at t ) 4 µs.35-36 IV. Results and Discussion
III. Photoluminescence Data Analysis Methodology The PL transients were deconvoluted using a multiexponential n function f(t) ) ∑i)1 Ai exp(- t/τi) + B where Ai is the decay amplitude, B is a constant (the baseline offset), and τi is the
Figure 1. XRD patterns of Tb3+-ZT1(33) (A), Tb3+-ZG(75) (B), and Tb3+-nMFI(C) recorded in the range 0-80° 2θ using Cu KR radiation (λ ) 1.54183 Å).
IV.1. Zeolite-Type Supports. Nanoslabs are zeolite precursor units of 4 × 4 × 1.3 nm that can aggregate and form zeolite crystals when heated at elevated temperatures.30 Zeogrid type was obtained through precipitation of a suspension of MFI nanoslabs upon addition of an ethanolic solution of CTABr. During the precipitation, the nanoslabs are organized in layers with intercalated surfactant molecules (Figure 2). Removal of the surfactant through calcination causes facial and lateral fusion of the nanoslabs.32 Empty spaces left laterally between nanoslabs are responsible for the formation of well-defined slit-shaped supermicropores with a characteristic height of ca. 1.4 nm, which is in accordance with the nanoslabs size. In addition, this material also contains typical MFI micropores. In the case of Zeotile-1 materials, nanoslabs were precipitated using an aqueous solution of CTABr at 80 °C. Because of the synthesis conditions, the nanoslabs aggregated in a hexagonal structure around the CTABr micelles (Figure 2).7 In addition to micropores, this material contains mesopores of 2.2-2.5 nm (Table 1). As mentioned, both ZT-1 and ZG materials were synthesized using MFI building blocks (nanoslabs). Therefore, the walls of
Figure 2. Assembling modes of nanoslabs in Zeogrid and Zeotile-1.
25710 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Tiseanu et al. 545 nm-centered emission line could easily be noticed for all samples investigated. In the time-resolved luminescence spectra, only Tb3+ PL originating from the 5D4 state was observed and no PL of the 5D3 state was found for all samples investigated. The weak emission detected between 630 and 700 nm is attributed to the 5D4-7F0,1,2 transitions (indicated with arrows in Figure 4). The PL decays measured at 545 nm are nonexponential in all samples, irrespective of the hydration state (hydrated, calcined, or rehydrated). The measured terbium lifetime τ related to the emission from the metastable 5D4 state is generally described as:
τ ) 1/(krad + knonrad + knorad*)
Figure 3. Representative steady-state excitation (left) and emission (right) spectra of hydrated (solid lines) and calcined (dotted lines) Tb3+ZT1. Emission and excitation wavelengths are fixed at 545 and 290 nm, respectively. Also included is the excitation spectrum of the aqueous terbium complex (dashed line).
Figure 4. PL decays of hydrated Tb3+-zeolites. Inset: representative time-resolved photoluminescence spectrum of hydrated Tb3+-zeolites (delay after the laser pulse of 4 µs).
these materials have some degree of internal order as opposed to MCM-41 materials. However, because of the nanosize of the building units, the resulting materials present a mixture of physicochemical properties from both MFI and mesoporous materials.37 IV.2. Hydrated Tb3+-Zeolites. Illustrated in Figure 3 is the representative steady-state luminescence spectrum of hydrated Tb3+-ZT1 samples. In the steady-state spectra, the PL of Tb3+ with its prominent emission peaks in the spectral range of 450 nm < λem < 650 corresponding to the 5D4-7F3, 7F4, 7F5, 7F6 transitions was observed. In addition, a broad featureless luminescence with an emission maximum at λem ) 410 nm that was attributed to the host matrix was seen as well. The excitation spectrum of a terbium-doped ZT1 sample follows that of the aqueous terbium complex (Figure 3). After calcination, both emission and excitation spectra of terbium PL become broadened, supporting a stronger interaction between terbium and zeolite framework. A spectral shift of ca. 10 nm of the excitation spectra of the annealed terbium-zeolites relative to those of the hydrated zeolites was also detected. Upon laser excitation at 337 nm, all hydrated terbiumzeolites display the terbium 5D4-related PL (Figure 4). A delay of 4 µs was sufficient to record a background-free Tb3+ PL spectra. Light differences between the spectral shapes of the
(1)
where krad is the natural radiative rate constant, knonrad is the rate constant for the nonradiative relaxation processes due to the -OH coupling, and knorad* is the rate constant for the radiationless relaxation processes due to other processes than the -OH coupling. The vibrational coupling of terbium ions with the -OH oscillators result from both water molecules and silanol groups. The corresponding knonrad is proportional to the number of OH groups contained in the inner coordination shell of the terbium ion.38 knorad* may include resonant ion-ion interactions located in close proximity or/and nonradiative energy transfer to host defects/impurities. Zeolite framework exhibits only low-energy phonons,39 which are not expected to significantly contribute to the nonradiative deactivation of Tb3+ PL. Further, quenching of 5D4 related emission due to ion-ion interactions can be excluded given the low Tb3+ concentration and lack of energy resonances involving this metastable level. The PL decays of hydrated Tb3+-zeolites were best fitted to two exponentials, and the obtained decay parameters are listed in Table 2. The biexponential analysis yielded Tb3+ PL decay times of approximately τ1 ∼ 180 µs and 470 µs < τ2 < 570 µs. The amplitude of τ2 accounted for more than 60% of the total amplitude. Generally, the decays became slower with increasing Si/Al ratio in the Tb3+-ZT1 materials. This is reflected in the calculated average decay time 〈τ〉, which increased by 22% across the whole Si/Al range. Unless a physical model is available, the amplitudes and decay times resulting from the biexponential fit are not readily interpretable in terms of luminescent species. Therefore, a twospecies distribution of terbium in zeolites was further checked by means of site-selective excitation spectra and/or time-resolved PL spectra. As the broad character of the excitation spectra of Tb3+-ZT1 (Figure 3) rendered the site-selective site-excitation measurements useless, TRES spectra at different time delays after the laser pulse were analyzed. The changes in spectral shape, spectral shift, or appearance of the new peaks with time are generally used as an indicative of the system heterogeneity.40 No differences in peak position, spectral line width, and spectral shape were found in the peaknormalized TRES of terbium-zeolites in the time range of 4 µs < t < 3000 µs after laser excitation. Therefore, it was reasonable to conclude that, in the hydrated samples, terbium is found as a single average species with lifetimes varying between 450 and 570 µs, which represent the dominant decay component (Table 2). The observed short decay component at 180 µs is attributed to impurities or defects of the host matrix. Further studies are needed to elucidate the nature and extent of this effect. According to ref 19, the PL lifetime of Tb3+-exchanged MFI measured at 77 K was determined at 630 µs. A single
Terbium-Exchanged MFI-Type Materials
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25711
TABLE 2: PL Decay Parameters of Tb3+-Exchanged Zeolites sample
hydration statea
Tb3+-ZT1(33)
h c rh h c rh h c rh h c rh h c rh
Tb3+-ZT1(150) Tb3+-ZT1(200) -ZG
Tb3+
Tb3+-nMFI
a
decay times τi (µs) and amplitudes Ai (%) in brackets
average decay time 〈τ〉 (µs)
χ2
180(37) 1043(64) 482(66) 183(37) 556(58) 476(69) 184(18) 976(57) 509(68) 178(38) 713(58) 408(74) 159(34) 520(62) 159(34)
421 1285 835 450 1056 772 511 1315 750 420 1200 581 520 780 520
0.9998 0.9995 0.9992 0.9996 0.9996 0.9993 0.9995 0.9992 0.9992 0.9997 0.9998 0.9996 0.9992 0.9996 0.9992
355(24) 67(19) 133(15) 65(16) 325(24) 160(18) 267(30) 107(16) 136(23)
474(63) 2195(12) 1392(15) 506(63) 1423(27) 1304(15) 535(82) 1996(19) 1305(14) 480(62) 1780(12) 1123(10) 570(66) 1310(15) 570(66)
h, hydrated; rh, rehydrated; c, calcined.
exponential decay with a rather short lifetime of only 34 µs was measured for Tb3+ doped MCM-41 synthesized by the solgel-assisted self-assembly method, while up to four chemical environments at the ions site were determined by means of NMR.25 Density functional calculations determined that Ln3+ ions were located in the straight channels, where they are coordinated to six oxygen atoms, four of them being related to the zeolite framework and the other two resulting from the hydrolysis of the lanthanide hydrate.23 In water, Tb3+ shows monoexponential luminescence decay kinetics with a luminescence decay time of 395 ( 10 µs. Therefore, our results (Table 2) also sustain a mixture of framework contribution and water molecules to terbium coordination in MFI-type materials. The effect of the Si/Al ratio on the PL properties of terbium in ZT1 samples was readily observed in the PL decays as well as the line shapes (Figure 4 and ref 27). These findings evidence that the effect of framework on terbium PL in the hydrated ZT1 is sizable, although water molecules are expected to shield terbium ions against matrix effects. The changes with Si/Al ratio might be related to the following effects: a mixed coordination to water and framework oxygen atoms (with framework contribution increasing with Si/ Al ratio), varying local charge (that reflects the Al distribution in the framework), as well as framework geometry at the terbium sites (different bonds and angles in the Al-OH-Si groups41), which also may shape the first coordination sphere of lanthanide’s ions. The mechanisms above contribute to both radiative and nonradiative relaxation rates of terbium and hence they should not be accounted for separately when discussing terbium lifetime changes with the Si/Al ratio (Table 2). However, it is expected that, in the hydrated states of ZT1, the partial bonding to the framework oxygens is mainly responsible for the Tb3+ lifetime increase with increasing the Si/Al ratio (Table 2). Preliminary PL measurements at 80 K show a fine structure with maxima at ca. 542, 545, and 550 nm with the relative intensities depending on the Si/Al ratio. These maxima represent the convolution of the various inter-Stark transitions between the 5D4 and 7F5 multiplets whose relative intensities depend on the local crystal field. The modification of the PL line shapes of terbium-ZT1 with the Si/Al ratio is therefore a clear indication of the crystal-field modification at the terbium sites caused by the compositional tuning. Irrespective of texture, Tb3+ is expected to occupy a position in the close proximity of aluminum with the involvement of a single interaction with the network, although an additional interaction with a second aluminum network or deposition
nearby a silicon are not excluded.42 The longest lifetime shown by Tb3+ in nMFI (570 µs, Table 2) provide evidence that Tb3+ exists predominantly inside zeolite channels, where the interaction with the network is stronger. For ZT1 and ZG materials, the total surface area is composed of the micropores surface area, located in the mesopore walls due to the partial zeolite crystallinity, and the mesopores surface area, with the latter being predominant. Under these conditions, Tb3+ would prefer the sites located on the internal surface of mesopores (the mesopore wall for ZT1 or between the layers of Zeogrid) to those in channels, which require additional mass transfer limitations to be reached. IV.3. Calcined Tb3+-Zeolites. Upon thermal treatment in air at 450 °C, the terbium PL display stronger intensity as well as longer decays compared to those measured with the hydrated samples (Figure 5). Moreover, the Tb3+ PL decays display an increased heterogeneous character, and a minimum of three exponentials were needed for a reasonable fitting (based on the reduced χ2 and randomness of the residuals). The decay times vary between 133 µs < τ1 < 355 µs, 520 µs < τ2 < 1043 µs, and 1310 µs < τ3 < 2195 µs, with τ2 accounting for more than 50% of the total decay amplitude. Upon thermal treatment, the coordination of terbium ions changes as water is removed from the first coordination sphere. The framework modifies to accommodate the coordination requirements of the terbium ions or they migrate to a different position where they can obtain a higher coordination environment.43-46 Although migration of terbium ions inside the
Figure 5. PL decays of calcined Tb3+-zeolites. Inset: comparison between PL decays of hydrated (curve 1), rehydrated (curve 2), and calcined Tb3+-ZG (curve 3).
25712 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Tiseanu et al.
Figure 7. TRANES profiles illustrated for the Tb3+-ZT1(200) and a zoom picture around the isoemissive point at 543.3 nm. TRANES profiles were smoothed by using a log-normal distribution.
Figure 6. Comparison between spectral line shapes of 545 nm-based emission of Tb3+-ZT1(200), Tb3+-ZG, and Tb3+-nMFI in the hydrated (thin lines), rehydrated (dotted lines), and calcined (bold lines) states.
channels of ZT1 at low Si/Al ratio cannot be completely ruled out, with increasing the Si/Al ratio, it is expected to become more difficult even in the presence of water. Once entrapped in the position where the channels are interpenetrated, terbium coordination with water will be limited by the interactions with the network. On the framework adjustment side, both a conformation change of the Al-O tetrahedra and redistribution of the charge in their close neighbor are expected to occur,47-49 which in turn modify the crystal field at the terbium sites. Considering the various factors that affect the spectral profile of an optical transition at a given temperature,50 the stronger emission and PL line shape alteration of the calcined Tb3+zeolites are explained by the modification of crystal field, increased heterogeneity, and the -OH removal at terbium sites. The relative contribution of the above effects to the overall terbium PL properties changes is determined by zeolite type. For example, the emission broadening expressed as the fullwidth at half maximum (fwhm) of the PL line at λem ) 545 nm increases in the order: 10.4 nm (nMFI), 11.2 nm (ZG), and 13.3 nm (ZT1(200)), Figure 6. For comparison, the same values for the hydrated samples were around 8-9 nm. The small dependence of PL lines half widths (fwhm) with temperature (down to 80 K) confirm that the primary broadening mechanism of the terbium ion in the investigated zeolites is inhomogeneous, arising from the increased heterogeneity of the ion environment upon calcination. The PL decay lengthening from ca. 500 µs to more than 2000 µs is determined mainly by the decreased nonradiative relaxation rate, knonrad, induced by the -OH removal, albeit a greater contribution of the radiative relaxation, krad, to terbium lifetime modification is expected compared to the hydrated case due to a stronger interaction between Tb3+ and zeolite matrix (eq 1). We found that this contribution depended on zeolite texture. For the Tb3+-nMFI, only a small variation of the PL shape
was detected. Therefore, the removal of residual hydroxyl groups is considered as the main reason for the stronger emission and longer decays upon calcining. For Tb3+-ZG and Tb3+-ZT1 samples, besides broadening, a more noticeable structuring of the emission was also observed. A combination of the crystalfield modification at the terbium sites as well as the OH removal is most appropriate to explain these PL changes. The inhomogeneous broadening is generally determined by a short-range disorder in the host lattice and similar to glasses is expected to be comparable or greater than the overall crystal-field splitting of the 2S+1LJ multiplets of the lanthanide’s ion. The resolved features in the luminescence spectra of Tb3+-zeolites sustained an increased local order at the terbium sites, which according to Figure 6, follows the sequence: Tb3+-ZT1 > Tb3+-ZG > Tb3+-nMFI. Similar to the hydrated samples, the TRES analysis was used to evaluate a three-terbium species distribution in the calcined zeolites. We found that, in contrast with the hydrated samples, the time-resolved PL spectra display a slight but continuous variation with time with the variation extent depending on zeolite texture.51 A novel method based on the area normalization of TRES (time-resolved area normalized emission spectra, TRANES) proved very useful in the identification of number of species, even in systems that display slight changes of TRES.27,35,36 Compared to TRES, TRANES has a physical significance: an isoemissive point in TRANES spectra indicates the presence of two emissive species in the system. Here, we used TRANES analysis to investigate the heterogeneity of calcined terbiumdoped ZT1 materials with various Si/Al ratios and compare the results with those for the calcined Tb3+-ZG and Tb3+-nMFI. Figure 7 shows the TRANES spectra of Tb3+-ZT1(200) sample. An isoemissive point at 543.3 nm is observed whose intensity is constant and does not change with time. For further demonstration, intensities at three positions (at 541.94 nm, 548.245 nm, and at isoemissive point) have been chosen from TRANES, and their time dependency is plotted in Figure 8 as well. It is obvious that intensity at the isoemissive point is independent of time, whereas intensities on either side (left or right to the isoemissive point) were changing in opposite ways. This further supports the presence of the isoemissive point at 543.3. nm. For Tb3+-ZT1(33) and Tb3+-ZT1(150) samples, TRANES analysis provides similar results. However, some differences between the positions of the isoemissive points were evidenced (Figure 8). For Tb3+-ZT1(33), its value was found to be at 543 nm, which was similar to within our spectral resolution to
Terbium-Exchanged MFI-Type Materials that of Tb3+-ZT1(200). For the Tb3+-ZT1(150), the isoemissive point was shifted 2 nm toward shorter wavelengths at 541.3 nm. Note that, for this sample, a different spectral shape as well luminescence dynamics compared to those of Tb3+-ZT1(33) and Tb3+-ZT1(200) were obtained (Figure 5). A special case is represented by Tb3+-ZT1(33) and Tb3+-ZT1(200). For these samples, TRANES displayed an identical isoemissive point (at 543 nm). However, its relative location toward the wavelengths at which the PL intensity showed a maximum variation with time was different advocating for different terbium environments in the two samples. For Tb3+-ZG, a similar isoemissive point at ca. 543 nm, although not so well-defined as in the case of Zeotile-1 samples, was obtained, while in the case of Tb3+nMFI, TRANES analysis could not provide an isoemissive point. This is most probably due to a more heterogeneous terbium distribution in this material. The exponential fitting of the PL decay curves assumed a number of two or three in the cases of the hydrated and calcined Tb3+-zeolites, respectively. Unless a priori knowledge suggests multiple discrete exponential decay components, it is empirically more effective to assume a decay model corresponding to a continuous lifetime distributions. This could be a more accurate representation of the terbium distribution given the highly heterogeneous nature of these structures. While, to the best of our knowledge, there are no reports on the lifetimes distribution of lanthanides in mesoporous or zeolites, broad lifetime distributions of excited states of guest molecules in zeolites are frequently reported.52 A lifetime distribution analysis based on a recently proposed method34 is applied to terbium-ZT1 samples. The method allows for the symmetric distributions as well as Voigt profiles (Gaussian and Lorentzian mixture). According to the calculated distribution, a multimodal lifetime distribution was obtained for Tb3+-ZT1(150), and except for the shortest component at 70 µs, the distribution is composed by two broad and strongly overlapped distributions centered at ca. 650and 1400 µs (Figure 9). Similar patterns were obtained for all ZT1 samples. The shape of PL lifetimes distribution might be related to two terbium species distributed on a range of environments. Accordingly, the longer lifetime is attributed to terbium located more deeply in zeolites, whereas the shorter lifetime describe terbium ions located on the internal surface of the mesopores. IV.4. Rehydrated Tb3+-Zeolites. Studies on recalcining/ rehydration of metal-exchanged zeolites are generally carried out by using EXAFS, XRD, and synchrotron X-ray powder diffraction.53-54 In particular, PL studies on lanthanideexchanged zeolites were generally pursued on (fully) hydrated and/or calcined states. To further discriminate between the three types of zeolites in terms of their PL response to the hydration states, we have repeated the measurements after a 6-month
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25713
Figure 9. Lifetimes distribution for Tb3+-ZT1(150) recovered from the PL decays observed at 545 nm based on the method described in ref 34. Pure Gaussian distributions were used in the data fitting. Vertical lines represent lifetimes values obtained with a three discrete exponential fit (Table 2).
exposure of samples to the atmospheric moisture. The samples are called rehydrated, and the results regarding the PL decay and spectra were included in Table 2 and Figures 5-6. It is readily noticed that the behavior toward rehydration strongly depend on zeolite texture and Si/Al ratio for Zeotile-1 samples. For Tb3+-ZT1 both line shapes and decays lifetimes values were found between those of calcined and hydrated samples suggesting a rather slow rehydration experienced by the Zeotile-1 materials. The relative decrease of the average lifetime, 〈τ〉, with rehydration is ca. 27% for terbium-ZT1(150) and almost 43% for terbium-ZT1(200). On the contrary, the PL properties of Tb3+-nMFI were restored to those of the hydrated sample (Table 2 and Figure 6). Tb3+-ZG presented an intermediate case: the PL line shape recovered that of the hydrated state, while the average lifetime, 〈τ〉, for the rehydrated sample was 580 µs compared to only 420 µs for the hydrated state (see also Figure 5, inset). The slow rehydration rate of the mesoporous materials might be due to a deeper location of Tb3+ inside the zeolite channels where water diffusion is limited compared to that in nMFI zeolite. In the latter case, terbium is expected to locate close to the pore opening, which supports for the fast recovery of the PL properties after exposure to the atmospheric moisture. V. Conclusions The work presents a first comparison between PL properties of terbium-exchanged Zeotile-1, Zeogrid, and nanocrystalline MFI materials. Because of the lack of structural data on the cation-exchanged MFI materials, the determination of terbium
Figure 8. Time dependency of the intensity measured at the isoemissive point (I) and at two emission wavelengths located to the left (L) and right (R) of the isoemissive point, which display a maximum increase and decrease of the intensity with time. All curves represent experimental data.
25714 J. Phys. Chem. B, Vol. 110, No. 51, 2006 sitting and distribution only by means of photoluminescence spectroscopy remains a difficult task. However, our detailed time-resolved measurements and analysis based on discrete exponential analysis, area-normalized time-resolved photoluminescence spectra, and lifetimes distribution, allow us to draw the following conclusions: in the hydrated Tb3+-exchanged zeolites, Tb3+ was found as a single species with mixed coordination to water molecules (most part) and framework oxygens with lifetime values varying from 470 to 570 µs. Terbium PL could successfully probe for the structural modifications induced by varying the Si/Al ratio, even in the hydrated state of ZT1 where the shielding of water molecules was expected to soften the framework effect. For the calcined ZT1 and ZG, two main terbium species were found with lifetimes varying between ca. 550-1040 µs and 1420-2195 µs, respectively. For terbium in the nanocrystalline MFI, a similar distribution could not be inferred. This is most probably related to the possibility of a more heterogeneous terbium distribution, with lanthanides located in the various inner channels. The resolved features of the luminescence spectra of Tb3+-zeolites sustained an increasing local order at the terbium sites according to the sequence: Tb3+-Zeotile-1 > Tb3+Zeogrid > Tb3+-nMFI. The behavior toward rehydration resulted upon 6 months of exposure to moist atmosphere was found to strongly differ with zeolite texture. Photoluminescence line shape and decay of Tb3+-Zeotile-1 were between those of the hydrated and calcined states, indicating a slow rehydration rate in contrast with photoluminescence properties of Tb3+nMFI, which fully recovered the values of the hydrated state. For Tb3+-Zeogrid, although the PL line shape was fully restored to that measured for the hydrated sample, the decay was still longer than that of the hydrated sample. The overall data suggest that, in Zeotile-1 and Zeogrid materials, two distinct locations of Tb3+ exist, which might be correlated with two types of exchange sites: one on the internal surface of mesopores, the other inside the pores, while for the nanocrystalline MFI, the sites predominate inside the pores. References and Notes (1) Jacobs, P. A.; Jansen, J. C.; Flanigen, E. M.; van Bekkum, H. Introduction to Zeolite Science and Practice; Elsevier Science Ltd.: Amsterdam, 2001. (2) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonovicz, M. E.; Kresge, C. T.; Schmidt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834. (3) Liu, Y.; Zhang, W.; Pinnavaia, T. J. J. Am. Chem. Soc. 2000, 122, 8791. (4) Liu, Y.; Zhang, W.; Pinnavaia, T. J. Angew. Chem., Int. Ed. 2001, 40, 1255. (5) Zhang, Z.; Han, Y.; Zhu, L.; Wang, R.; Yu, Y.; Qiu, S.; Zhao, D.; Xiao, F. S. Angew. Chem., Int. Ed. 2001, 40, 1258. (6) On, D. T.; Kaliaguine, S. Angew. Chem., Int. Ed. 2001, 40, 3248. (7) Kremer, S. P. B.; Kirschhock, C. E. A.; Aerts, A.; Villani, K.; Martens, J. A.; Lebedev, O. I.; van Tendeloo, G. AdV. Mater. 2003, 15, 1705. (8) Kremer, S. P. B.; Kirschhock, C. E. A.; Tielen, M.; Collignon, F.; Grobet, P. J.; Jacobs, P. A.; Martens, J. A. AdV. Funct. Mater. 2002, 12, 286. (9) Kokotailo, G. T.; Lawton, S. L.; Olson, D. H.; Meier, W. M. Nature 1978, 272, 437. (10) Meier, W. M.; Olson, D. H. Atlas of Zeolite Structure Types; Butterworth-Heinemann: London, 1992. (11) Sobalik, Z.; Dedecek, J.; Ikonnikov, I.; Wichterlova, B. Microporous Mesoporous Mater. 1998, 21, 525. (12) Bredol, M.; Kynast, U.; Ronda, C. AdV. Mater. 1991, 3, 361. (13) Justel, T.; Wiechert, D. U.; Lau, C.; Sendor, D.; Kynast, U. AdV. Funct. Mater. 2001, 11, 105.
Tiseanu et al. (14) Rocha, J.; Carlos, L. D. Curr. Opin. Solid State Mater. Sci. 2003, 7, 199. (15) Wada, Y.; Okubo, T.; Ryo, M.; Nakazawa, T.; Hasegawa, Y.; Yanagida, S. J. Am. Chem. Soc. 2000, 122, 8583. (16) Arakawa, T.; Takata, K.; Adachi, G. Y.; Shiokawa, J. J. Lumin. 1979, 20, 325. (17) Suib, S. L.; Zerger, R. P.; Stucky, G. D.; Morrison, T. I.; Shenoy, G. K. J. Phys. Chem. 1984, 80, 2203. (18) Baker, M. D.; Olken, M. M.; Ozin, G. A. J. Am. Chem. Soc. 1988, 110, 5709. (19) Serra, O.; Nassar, E. J.; Zapparolli, G.; Rosa, I. L. V. J. Alloys Compd. 1995, 63, 225. (20) Stucky, G. D.; Iton, L.; Morrison, T.; Shenoy, G.; Suib, S. L.; Zerger, R. P. J. Mol. Catal. 1984, 27, 71. (21) Hong, S. B. J. Phys. Chem. B 2001, 105, 11961. (22) Othman, I.; Mohamed, R. M.; Ibrahim, I. A.; Mohamed, M. M. Appl. Catal., A 2006, 299, 95. (23) Yang, G.; Wang, Y.; Zhou, D.; Zhuang, J.; Liu, X.; Han, X.; Bao, X. J. Chem. Phys. 2003, 119, 9765. (24) Aquino, J. M. F. B.; Araujo, A. S.; Melo, D. M. A.; Silva, J. E. C.; Souza, M. J. B.; Silva, A. O. S. J. Alloys Compd. 2004, 374, 101. (25) Yin, W.; Zhang, M. J. J. Alloys Compd. 2003, 360, 231. (26) Reisfeld, R.; Zigansky, E.; Gaft, M. Mol. Phys. 2004, 102, 1319. (27) Tiseanu, C.; Kumke, M. U.; Parvulescu, V. I.; Koti, A. S. R.; Gagea, B. C.; Martens, J. A. J. Photochem. Photobiol. A: Chem. To be published. http://dx.doi.org/10.1016/j.jphotochem.2006.10.026. (28) Tiseanu, C.; Frunza, L.; Kumke, M. U. Phys. B 2004, 352, 358. (29) Beeby, A.; Clarkson, I. M.; Dickins, R. S.; Faulkner, S.; Parker, D.; Royle, L.; de Sousa, A. S.; Williams, J. A. G.; Woods, M. J. Chem. Soc., Perkin Trans. 2 1999, 493. (30) Kirschhock, C. E. A.; Buschmann, V.; Kremer, S.; Ravishankar, R.; Houssin, C. J. Y.; Mojet, B. L.; van Santen, R. A.; Grobet, P. J.; Jacobs, P. A.; Martens, J. A. Angew. Chem., Int. Ed. 2001, 40, 2637. (31) Aerts, A.; Huybrechts, W.; Kremer, S. P. B.; Kirschhock, C. E. A.; Theunissen, E.; van Isacker, A,; Denayer, J. F. M.; Baron, G. V.; Thybaut, J.; Marin, G. B.; Jacobs, P. A.; Martens, J. A. Chem. Commun. 2003, 15, 1888. (32) Kremer, S. P. B.; Kirschhock, C. E. A.; Jacobs, P. A.; Martens, J. A. Comptes. Rend. Chim. 2005, 8, 379. (33) Aerts, A.; van Isacker, A.; Huybrechts, W.; Kremer, S. P. B.; Kirschhock, C. E. A.; Collignon, F.; Houthoofd, K.; Denayer, J. F. M.; Baron, G. V.; Marin, G. D.; Jacobs, P. A.; Martens, J. A. Appl. Catal., A 2004, 257, 7. (34) Branco, T. J. F.; Botelho do Rego, A. M. H.; Machado, I. F.; Ferreira, L. F. V. J. Phys. Chem. B 2005, 109, 15958. (35) Koti, A. S. R.; Krishna, M. M. G.; Periasamy, N. J. Phys. Chem. A 2001, 105, 1767. (36) Koti, A. S. R.; Periasamy, N. J. Chem. Phys. 2001, 115, 7094. (37) Kirschhock, C. E. A.; Kremer, S. P. B.; Vermant, J.; van Tendeloo, G.; Jacobs, P. A.; Martens, J. A. Chem. Eur. J. 2005, 11, 4306. (38) Horrocks, W. D., Jr.; Sudnick, D. R. J. Am. Chem. Soc. 1979, 101, 334. (39) Maas, H.; Currao, A.; Calzaferri, G. Angew. Chem., Int. Ed. 2002, 41, 2495. (40) Lakowicz, J. R. Principles of Fluorescence Spectrometry; Plenum Press: New York, 1983. (41) Corma, A. Chem. ReV. 1995, 95, 559. (42) Parvulescu, V. I.; Centeno, M. A.; Grange, P. J. Catal. 2000, 191, 445. (43) Norby, P.; Poshni, F. I.; Grey, C. P.; Gualtieri, A. F.; Hanson, J. C. J. Phys. Chem. B 1998, 102, 839. (44) Shy, D. S.; Chen, S. H.; Lievens, J.; Liu, S. B.; Chao, K. J. J. Chem. Soc., Faraday Trans. 1991, 87, 2855. (45) Lee, E. F. T.; Rees, L. V. C. Zeolites 1987, 7, 143. (46) Klein, H.; Fuess, H.; Hunger, M. J. Chem. Soc., Faraday Trans. 1995, 91, 1813. (47) Alberti, A.; Vezallini, G. In Proceedings of the Sixth International Zeolite Conference; Olson, D., Bisio A., Eds.; Butterworths: Markham, ON, Canada, 1984. (48) Bish, D. L. In International Committee on Natural Zeolites; Ming, D. W., Mumpton, F. A., Eds.; Brockport: New York, 1995.
Terbium-Exchanged MFI-Type Materials (49) Kentgens, A. P. M.; Scolle, K. F. M. G. J.; Veeman, W. S. J. Phys. Chem. 1983, 87, 4357. (50) Hehlen, M. P.; Cockroft, N. J.; Bruce, A. J.; Gosnell, T. R. Phys. ReV. B 1997, 56, 9302. (51) Tiseanu, C.; Kumke, M. U.; Parvulescu, V. I.; Gagea, B. C.; Martens, J. A. Stud. Surf. Sci. Catal., to be published.
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25715 (52) Scaiano, J. C.; Kaila, M.; Corrent, S. J. Phys. Chem. B 1997, 101, 8564. (53) Dooryhee, E.; Catlow, C. R. A.; Couves, J. W.; Maddox, P. J.; Thomas, J. M.; Greaves, G. N.; Steel, A. T.; Townsend, R. P. J. Phys. Chem. 1991, 95, 4514. (54) Sankar, G.; Thomas, J. M. Top. Catal. 1999, 8, 1.
