Step-wise dealumination of natural clinoptilolite: Structural and physicochemical characterization

Microporous and Mesoporous Materials 135 (2010) 187–196

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Step-wise dealumination of natural clinoptilolite: Structural and physicochemical characterization Yunier Garcia-Basabe a,b, Inocente Rodriguez-Iznaga c, Louis-Charles de Menorval d,**, Philip Llewellyn e, Guillaume Maurin d, Dewi W. Lewis f,***, Russell Binions f, Miguel Autie g, A. Rabdel Ruiz-Salvador a,* a

Zeolite Engineering Laboratory, Institute of Materials Research and Engineering (IMRE), University of Havana, Havana 10400, Cuba Department of Physics, Faculty of Mechanical and Chemical Engineering, University of Matanzas, Matanzas 40100, Cuba Materials Technology Laboratory, Institute of Materials Research and Engineering (IMRE), University of Havana, Havana 10400, Cuba d Institut Charles Gerhardt Montpellier, CNRS, Université Montpellier 2, Place E. Bataillon, 34095 Montpellier cedex 05, France e Laboratoire Madirel UMR CNRS 6121, Université de Provence, Centre St Jérôme, 13397 Marseille cedex 20, France f Department of Chemistry, University College London, 20 Gordon St., London WC1H 0AJ, UK g Molecular Engineering Laboratory, Institute of Materials Research and Engineering (IMRE), University of Havana, Havana 10400, Cuba b c

a r t i c l e

i n f o

Article history: Received 20 March 2010 Received in revised form 16 July 2010 Accepted 17 July 2010 Available online 22 July 2010 Keywords: Clinoptilolite Heulandite Dealumination Al vacancy Silanol nest

a b s t r a c t The step-wise dealumination of a natural clinoptilolite has been achieved through a new milder treatment, comprising cycles of aqueous solutions of hydrochloric acid with washing steps. The course of the dealumination was monitored by XRD, FTIR, TGA, UV–vis-DRS and NMR. The XRD patterns show a contraction of cell volume during progressive dealumination steps and a decrease in crystallinity after the third dealumination cycle. The framework and OH vibrations in FTIR spectra show progressive extraction of aluminum atoms from zeolite framework and consequently the formation of silanol nests. 27Al MAS NMR indicates that low levels of octahedral aluminum species are created during the treatments. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction By modifying the chemical composition of the framework, the properties of a zeolite can be fine tuned [1–3]. Such modifications can be brought about during synthesis – for example by varying the synthesis gel, particularly Si/Al, or by inclusion of other species which will isomorphously substitute into the framework, such as iron or titanium. Alternatively, changes can be made post-synthesis and the main route here is to remove Al from the framework, with replacement by Si possible. The most important example is that of dealumination of synthetic faujasite structured materials, where the hydrophilic zeolite X (Si/Al 1.5) can be converted into hydrophobic zeolite Y (with far greater Si/Al). These modifications to the framework, which do not modify the topology, can have significant impact on both the chemical and physical properties of zeolites. From a chemical point view, most of the changes in catalytic and sorption properties can be considered as being the result

* Corresponding author. Tel.: +53 7 8780903; fax: +53 7 8794651. ** Corresponding author. Tel.: +33 467143341; fax: +33 467143304. *** Corresponding author. Tel.: +44 20 76794779; fax: +44 20 76797463. E-mail addresses: [email protected] (L.-C. de Menorval), [email protected] (D.W. Lewis), [email protected] (A.R. Ruiz-Salvador). 1387-1811/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2010.07.008

of the variation in the overall electronegativity of the solid, as discussed by Mortier and co-workers [4–6]. But another important result of such treatments are physical changes, such as to the thermal stability and pore volume, which result from both the changes to the framework and also to concomitant formation of extra-framework species [7,8]. Of course, where natural zeolites are concerned, only post-synthesis treatments are applicable! Dealumination is most commonly achieved by the hydrolysis of Al–O bonds by acid leaching or through hydrothermal treatment in the presence of water vapors [9–11]. Dealumination leads not only to the breaking of Al–O bonds – making the Al available for extraction by very diluted acid leaching – but can also result to a loss of crystallinity, particularly where low Si/Al materials are treated. Thus, milder methods for dealumination are sought. Clinoptilolite is the most common natural zeolite, which whilst isotypic with heulandite has a higher Si/Al: clinoptilolites having Si/Al > 4 and and heulandites Si/Al < 4 [12]. The abundance, low extraction cost and high chemical stability of clinoptilolite make this zeolite a very attractive material for adsorption and environmental applications. In this type of applications it is known that usually the performance of the zeolites are very dependent of Si/ Al ratio. The framework of clinoptilolite (IZA code HEU), constructed of 4-4-1 secondary building units, consists of three chan-

188

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

nels connected in two dimensions: two of them formed of 8-membered rings and the third of 10-membered rings. The 10-membered channel (A) and one of the 8-membered channels (B) runs parallel to the crystallographic axis c, with the second 8-membered (C) parallel to the a axis (Supplementary Fig. S1 in Electronic Supporting Information). The HEU topology also feature layers joined by reticularly rigid oxygen bridges. It is well known that a significant proportion of the Al resides in the tetrahedral sites (T2 following the notation of Alberti [13]) that form the interlayer links (Fig. 1a); the T2 site is about 40% occupied by Al, depending of zeolite sample [13–15]. Thus, it is clear that extensive Al removal can cause interlayer disconnection and consequent breakdown of the zeolite structure (Fig. 1b). Furthermore, it would be expected that only 50% of the Al can be extracted without destroying the zeolite. However, appropriate hydrothermal treatments can result in migration of Si to fill the Al vacancies and thus stabilize the otherwise depillared structure. This appears to be the mechanism that explains the pioneering results of Barrer and Makki who treated natural clinoptilolite with hydrochloric acid of different strengths, obtaining a series of highly crystalline solids with progressive changes in adsorption properties [9]. More recently, clinoptilolite and heulandite dealumination has been considered by a number of authors, following similar or related treatments [16–20]. In these studies the acid treatments are performed on natural cationic forms of the zeolites, where decationization (ion exchange between compensation cations in the zeolites and the H3O+ ions in the acid solution) and the dealumination process can occur simultaneously in a competitive way. Moreover, these studies also showed that in order to remove more than 30% of the aluminum from the clinoptilolite framework it is necessary to use a high acid concentration (P1 M), which of course will also result in a significant reduction in the crystallinity. In this work, using a natural Cuban clinoptilolite, we have developed a step-wise dealumination methodology, using milder conditions than previously considered, whose aim is to minimize

loss of crystallinity and porosity upon Al removal. The progress of the dealumination is followed through different spectroscopic techniques and X-ray diffraction. 2. Experimental 2.1. Starting materials The clinoptilolite used in this work is the main phase present in the zeolitic rock (the raw mineral) from the Caimanes deposit, Moa (Cuba). In addition, Mordenite, Quartz, Feldspar and Montmorillonite are present. The raw mineral was ground and sieved to 32–90 lm grains and then purified by washing with distilled water in a fluidized bed process [21] to remove the non-zeolitic mineral phases. In addition, the denser and lighter fractions were discarded as they often contain feldspar and clays to obtain a material more rich in clinoptilolite phase [22]. After vacuum filtration and drying, this resulted in a mixture of about 87% clinoptilolite, 8.5% quartz and 4.5% Mordenite [23]. Herein this sample is referred to as purified zeolite or natural clinoptilolite (NZ). The chemical composition of the NZ sample is showed in Table 1. The ammonia form of the natural clinoptilolite zeolite (denoted NH4Z) was obtained by ion exchange with aqueous 1 M NH4Cl solution at 100 °C for 40 h. The solid:liquid ratio was 1 g of solid: 10 ml of solution following a methodology described elsewhere [24]. The solution was changed every 8 h. The NH4Z was washed with distilled water until no Cl was detected in the wash water by a AgNO3 test, and then dried at 60 °C overnight. The acid form of this material, denoted HZ, was obtained by air calcining NH4Z by heating at a rate of 5 °C min1 to 400 °C and maintaining this temperature for 16 h. 2.2. Dealumination procedures In order to minimize framework damage we use a lower concentration of HCl than that in many previous procedures [9,17,19,20,25]. Since one of the main effects of hydronium ions in zeolites is to also exchange with the extra-framework cations, this is avoided by the use of the acid form of the zeolite. Thus, HZ was treated with a 0.6 M HCl solution. The temperature and time of reaction was 100 °C and 2 h, respectively. The solid:liquid ratio was 1 g of solid: 10 ml of solution. After this treatment the zeolite was washed 3 times in an ultrasonic bath for 15 min each time with a dilute solution of 0.05 M of HCl at 70 °C. The solid:li-

Table 1 Chemical compositions and distance between tetrahedral sites-oxygen of the zeolite framework of the natural and modified zeolites.

Fig. 1. The clinoptilolite framework view perpendicular to c axis. (a) Aluminum atoms in tetrahedral site T2. (b) Aluminum vacancies in T2 site.

% Atomic

NZ

HZ

HZD1

HZD3

HZD5

O Al Si Fe K Ca Mg Si/Al Si/Fe Al per unit cell Vacancy per unit cell /Å

65.2(1) 5.41(3) 24.2(5) 0.98(9) 0.74(9) 2.58(5) 0.79(7) 4.48 24.69 6.57 0 1.647

66.1(2) 6.15(4) 26.7(4) 0.72(4) 0.34(8) – – 4.34 37.08 6.37 0.11 1.647

65.3(1) 4.18(2) 29.3(3) 0.89(6) 0.33(9) – – 7.00 32.92 4.21 2.36 1.645

65.6(4) 3.94(1) 29.8(4) 0.40(5) 0.25(4) – – 7.55 74.50 3.9 2.67 1.644

65.1(2) 3.09(2) 30.9(2) 0.23(3) 0.19(6) – – 10.00 134.35 3.57 3.62 1.643

The values in parenthesis are the standard deviations for atomic percentage in each element. Vacancy per unit cell is the number of vacancies in tetrahedral framework and is calculated from: Vacancyuc ¼ 36  Sinat  Al Si  Sinat , where Sinat is the Silicon atoms by unit cell in natural zeolite and Al/Si is the aluminum/silicon ratio of the sample. Mean tetrahedral atom to oxygen distance < TO >¼ nSi 1:605þnAl361:72þnv 1:68.

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

quid ratio was 1 g:25 ml. Finally, the sample was washed with distilled water until the wash water was free of Cl and then dried at 60 °C overnight. This first dealuminated zeolite is now denoted HZD1. Two further dealuminated samples were obtained by repeating the procedure described three and five times on fresh samples of HZ: these samples are denoted HZD3 and HZD5, respectively. 2.3. Characterization Powder X-ray diffraction patterns (PXRD) were collected in the 2h range 4–60° at a scan rate of 2°min1 on a XPERT PRO diffractometer with Cu Ka radiation (k = 1.5406 Å). Elemental analysis was performed using EDAX Spectroscopy on a S8OB Cambridge scanning microscope with a microprobe analyzer EDAX DX4. The TG analyses were preformed on TGA Q500 V6.5 Build 196 instrument in flowing He using a heating rate of 5 K min1. N2 gas adsorption isotherms at 77 K were measured by means of a Micromeritics ASAP-2000 device. Prior to adsorption measurements, the samples were evacuated in vacuum for 8 h at 300 °C. Infrared spectra were recorded on an FTIR PU9600 Pye Unicam spectrometer at room temperature. Spectra were collected by summing 250 scans at 2 cm1 resolution in the region 4000–350 cm1 of samples prepared as standard KBr pellets. Prior to the FTIR experiments, the samples were first placed in an oven at 100 °C for 2 h and kept in a desiccator during cooling down in order to remove much of the adsorbed water. UV–vis diffuse reflectance spectra were obtained with a Varian Cary 300 spectrometer, equipped with a standard diffuse reflectance unit using BaSO4 as reference. The samples were finely ground and the spectra were recorded in the 190–850 nm range. The trituration was done gently to avoid solid state chemical reactions. The MAS NMR spectra of the solid samples were collected at room temperature using an ASX-300 Brucker spectrometer for the 27Al and 29Si nuclei. The 29Si MAS NMR spectra were recorded 59.62 MHz using 4.70 ls pulse with a range 2 s recycle time and 4– 16  103 scans. The magic angle spinning rate for all 29Si spectra was 10 kHz and chemical shifts were referenced to tetramethylsilane (TMS). 27Al MAS NMR spectra were recorded at 78.2 MHz using a 3.6 ls pulse with a 1 s recycle delay and 2–16  103 scans. A 1% aqueous solution of Al(NO3)3 was used as a reference for the chemical shifts, and the samples were spun at 10 kHz.

189

amount of iron. However, we are not able to state, as yet, whether this is from the framework or from a secondary phase.

3.1. X-ray powder diffraction Normalized XRD patterns of the samples are show in Fig. 2. The intensity of the clinoptilolite (0 2 0), (2 0 0), (2 0 1), (3 1 1), (1 1 1), (4 0 0), (4 2 2) and (4 4 0) peaks are the most affected during the dealumination. The ratio of intensity of these peaks to that of the (0 1 1) peak of contaminant quartz phase (which should be unaffected by the treatments) is shown in Table 2. From these XRD patterns and the values in Table 2 we see little change between HZ and HZD1, whilst in the further dealumination steps, forming HZD3 and HZD5, the intensities of main peaks in the pattern decrease significantly in comparison to those resulting from the quartz phase. In order to develop a quantitative evaluation of the crystallinity of the samples, we determined the integral width, Hw of the diffraction peaks using a the LeBail profile method, as implemented in the FULLPROF suite programs [30]. The profiles were modeled using a pseudo-Voigt peak-shape function in the C2/m space group (the space group of the parent clinoptilolite) [14,31,32]. The plot of the difference in Hw values between consecutive dealumination processes versus 2h (Fig. 3) reveals that during the initial Al extraction (from HZ to HZD1) changes in the crystallinity of the sample, as measured by the broadening of the peaks, is negligible, in agreement with the result inferred from the relative peak intensities (Table 2). This would indicate that the stress and local distortions caused by forming (on average) two Al vacancies per unit cell (Table 1) is well dissipated and do not produce significant degradation of the structure. However, the subsequent Al extraction steps both result in significant changes to the crystallinity, even though the decrease in average Al content per unit cell is not as dramatic as in the first dealumination step. Beside this quantitative change in Hw, the loss of crystallinity is also visible by the appearance of

3. Results and discussion Elemental analysis of the natural (NZ), H-form (HZ) and dealuminated zeolites HZD1, HZD3 and HZD5 are showed in Table 1. As reported in Ref. [24] a complete elimination of Na, Ca and Mg is achieved during the preparation of the HZ. However the potassium concentration is only reduced 50%, which suggests that in this zeolite (Caimanes deposit) there exists either a particular Si–Al distribution that strongly traps the K+ (see Ruiz-Salvador et al. for an analysis of cation location and Si–Al distribution [26]) or alternatively a potassium rich secondary phase remains. The first dealumination step (forming HZD1) has resulted in a significant decrease in Al content whilst the subsequent two steps (forming HZD3) do not result in any significant decrease. However, a large decrease is again found when the final two steps are undertaken (forming HZD5). The state of Fe in clinoptilolite samples has been the source of some discussion [27–29]. Here, when HZ is formed from NZ, the portion of iron lost can be attributed to the exchange of extraframework Fe2+ or loss from a minor secondary phase. Whilst there appear no changes in iron content (relative to silicon) when HZ is first dealuminated, subsequent acid treatments do reduce the

Fig. 2. XRD patterns of natural and modified zeolites. Miller indices of the main peaks of clinoptilolite and quartz are indicated.

190

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

Table 2 Intensity ratio of the most affect peaks in XRD patterns of the clinoptilolite phase and the peak (0 1 1) of the quartz phase for natural and modified clinoptilolites. Sample

I020Cl/I011Q

I200Cl/I011Q

I201Cl/I011Q

I311Cl/I011Q

I111Cl/I011Q

I400Cl/I011Q

I422Cl/I011Q

I440Cl/I011Q

NZ HZ HZD1 HZD3 HZD5

170 176 122 97 55

59 92 61 55 32

40 44 45 38 28

48 34 26 24 23

55 58 40 39 24

179 184 121 101 66

57 55 41 32 25

73 54 45 39 18

Fig. 3. Differences between the XRD peaks in integral width (Hw) in consecutive dealumination steps.

the broad low baseline in the XRD pattern of HZD3 and more significantly in HZD5, likely to due to the partial amorphization of the zeolite phase. As can be seen from Table 1, about 40% of Al is removed in HZD3 and, according to the XRD, data no significant loss of crystallinity results. However, further dealumination results in noticeable collapse caused for the removal of Al at T2 sites (see Fig. 1b), as discussed above. Table 3 shows the unit cell parameters and volume for the NZ, HZ, HZD1, HZD3 and HZD5 samples, determined from the Le Bail fits, with the position of peak (0 1 1) of the quartz phase being used as an internal standard. The cell volume and the mean distance (Table 1) all decrease from HZ to HZD5 (Fig. 4), as might be expected due to the difference between typical Al–O (1.72 Å) and Si–O (1.605 Å) distances in zeolite frameworks: thus when Al is removed, the framework is likely to contract. However, as expressed in the equation in Table 1 used to determine the , the presence of Al vacancies leads to a small increase in the Si–O distances due to a weakening of the bond, due to softening of the interaction due the presence of the capping silanol groups (of the hydroxyl nest defect). Note that the cell volume of NZ, despite having similar distances is much lower than that of HZ due to the presence

Fig. 4. Variations of the cell volume of the zeolites as function of the mean distances as defined in Table 1.

of extra-framework cations which have strong interactions with the oxygen atoms of the framework [33].

3.2. N2 adsorption analysis The results of nitrogen adsorption measurement on the different zeolitic materials (Table 4) reveal a considerable increase in the specific surface area, micropore volume and external surface area Aext of the acid form HZ, compared to the parent NZ. These results are in agreement with those reported by other authors for clinoptilolite modified by ion-exchanged ammonium followed by calcination [24,34]. The abrupt increment in specific surface area SBET, micropore volume Vmic and external surface area Aext in HZ compared to NZ can be due to the almost complete replacement of the metal cations by H+, together with the elimination of some impurities, also resulting in more available space within the zeolite. Similar to previous reports on acid treated clinoptilolite [18,24,35,36] the specific surface area and micropore volume show a distinct increment with each dealumination step, from HZ to

Table 4 Structural parameters calculated from adsorption–desorption isotherms of nitrogen at 77 K. The absolute error is ±3%. Table 3 The unit cell parameters and cell volume of NZ, HZ, and acid modified samples. Sample

a (Å)

b (Å)

c (Å)

b (°)

NZ HZ HZD1 HZD3 HZD5

17.715(2) 17.726(2) 17.676(3) 17.666(4) 17.621(4)

18.019(6) 18.092(2) 18.020(5) 18.035(6) 18.064(5)

7.421(2) 7.436(1) 7.428(1) 7.434(2) 7.372(1)

116.33(1) 115.972(8) 115.92(2) 115.76(2) 115.82(2)

0

3

Volume (Å A ) 2123.2(8) 2144.2(4) 2130.1(8) 2121.1(10) 2112.5(9)

Zeolite

Vmic/cm3 g1

SBET/m2 g1

Aext/m2 g1

AL/m2 g1

NZ HZ HZD1 HZD3 HZD5

0.0032 0.0984 0.1028 0.1180 0.1141

20.037 248.327 259.198 291.925 305.267

13.467 53.617 55.714 58.227 79.137

29.769 356.975 372.507 419.177 439.782

Vmic Micropore volume; Aext external surface area calculated using the t-plot method. SBET is the specific surface area calculated using the standard Brunauer– Emmet–Teller (BET) method. AL Langmuir surface area from Langmuir method.

191

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

HZD3. These results are associated to creation of aluminum vacancies in the clinoptilolite framework and elimination of extraframework aluminum species during our dealumination procedure. A slight decrease in the micropore volume of HZD5 is evidence of the effect of the amorphization of the zeolite porosity. In contrast, the specific surface area continues to increase in HZD5, which can be attributed to an increment in external surface area and not of micropore surface area (SBET–Aext). The Langmuir surface area AL follow the same trend of the BET specific surface area SBET. 3.3. TG/DTG analysis Fig. 5 shows the differential thermogravimetric curves (DTG) of NZ, HZ and the dealuminated zeolites: the thermogravimetric (TG) curves are also given for completeness as Electronic Supporting Information (Supplementary Fig. S2). From these data we determine the number of water molecules lost (Table 5) in the temperature intervals 25–150 °C, 150–300 °C, 300–600 °C and 600– 800 °C, respectively. In NZ, the water molecules (ca. 24) are hydrating the extra-framework cations, whilst in HZ we estimate that one third are present as H3O+ – in order to charge balance the framework – with the remainder as H2O hydrating these hydronium cations. A similar proportion is found in the dealuminated samples, but with a number of the water molecules now also forming hydrogen bond to the silanol groups surrounding the Al vacancies. These differences explain in general terms the differences in the number of water molecules in the samples with the NZ and HZ

Fig. 5. DTG curves of the natural and modified zeolites. The temperature of the large water loss peak is shown.

samples having the most and the least water molecules, respectively. The relative weight loss (see Table 5) in the first temperature interval (T < 150 °C) increases with the level of dealumination, whilst that in the second stage (150 < T < 300 °C) decreases, suggesting that in dealuminated zeolites the major portion of water molecules is only weakly coordinated to the zeolite structure: whilst when there are more hydroxyl nests, coordination of water molecules to these defects is restricted as they can form intra-nest hydrogen bonds. The fact that all the dealuminated species have essentially dehydrated below 150 °C is also associated with the lack of extra-framework cations, which have the effect of strongly binding the water molecules within the pores. As can be seen in Table 5, the total number of water molecules increases from HZ to HZD3; which may be correlated to the increase in microporosity resulting from dealumination. However, a quantitative inspection reveals that the relative increase in the number of the water molecules is over twice that of the change in micropore and surface areas (Supplementary Table S1). We therefore conclude that the increase in water is due to the availability of sites associated with silanol nests that can form strong hydrogen fond interactions to water. In the case of HZD5, the partially amorphization (Table 5 and Fig. 2) is likely to be the cause of the decrease in the number of water molecules present.

3.4. Infrared analysis IR spectra of the NZ, HZ, HZD1, HZD3 and HZD5 are shown in Fig. 6. The main changes, on successive dealuminations are in the peak around 1062 cm1 associated with the asymmetric internal T–O stretching vibrations of the tetrahedra, whilst the peaks at 1211 cm1 (also internal asymmetric stretching in tetrahedra), 608 cm1 (external tetrahedral double ring) and 459 cm1 (internal tetrahedral bending) show only very small shifts. Similar result has been reported by Cakicioglu-Ozkan et al. [18] during a study of the effect of HCl treatments on a natural clinoptilolite from the Bigadic deposit (Turkey). These authors founded that the T–O asymmetric stretching band at 1056 cm1 shifted to 1080 cm1 as the degree of dealumination increased from 20 to 57%. Christidis et al. [16] also found that in a study of the treatment with 6 N HCl of a number of different heulandites that this band shifts from 1028 to 1092 cm1, while the internal asymmetric stretching (1202 cm1) does not change. As we convert NZ to HZ, these changes are due to the relative strengthening of the T–O bonds, where the loss of the extra-framework cation–oxygen interactions accounts for the shift of the frequencies. Upon dealumination the average charge on a tetrahedral atom increases, resulting in an average increase in the T–O bond strength (indicated by the changes to the peak at 1062 cm1), a conclusion supported by the fact that the other three bands that can be affected by dealumination do not changes during the formation of HZ from NZ. The doublet at 795 and 777 cm1 indicates the presence of the quartz impurity in the sample [37].

Table 5 Number of water molecules lost and their relative lost percent at each dehydration stage. Sample

NZ HZ HZD1 HZD3 HZD5

(25–150 °C)

(150–300 °C)

(300–600 °C)

(600–800 °C)

Total

Water molecules lost

Weight lost %

Water molecules lost

Weight lost %

Water molecules lost

Weight lost %

Water molecules lost

Weight lost %

Water molecules lost

7.96 7.43 10.06 13.67 12.45

36.41 42.23 51.10 57.85 59.46

6.53 5.11 4.32 4.66 3.67

29.42 29.05 21.93 19.69 17.54

5.73 2.36 3.26 2.96 2.86

24.86 13.43 16.59 12.56 13.68

2.23 2.67 2.04 2.33 1.93

9.31 15.23 10.38 9.88 9.24

24.19 17.60 19.69 23.64 20.94

The weight lost reported in each interval of temperature is calculated by ratio between (weight lost % in each temperature range and the total weight loss %)  100.

192

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

elimination of aluminum from the framework. Simultaneously, a clear increase in the peak intensity at 3444 cm1 which is indicative of the formation of silanol nest defects due to the formation of Al vacancies [48]. We also find an increase in the width of this peak as dealumination proceeds, which is in agreement with the TG analysis, where the amount of the loosely bound water increases. The shoulder at 3745 cm1 remains constant during dealumination which means that the acid treatments result principally in the formation of silanol nests and not isolated silanol groups. 3.5. UV–vis, diffuse reflectance analysis (DRS) The use of DRS (UV–vis) spectra for evaluating the presence of extra-framework aluminum in zeolite structure after dealumination was initially proposed by Garbowski [49] and described in more detail by Zanjanchi et al. during studies of the dealumination of various zeolites [50,51]. Their results are based on the fact that tetrahedral framework aluminum bonded to oxygen gives rise to a single band centered around 260 nm, while octahedral extraframework aluminum species usually generate a broad band at 280–370 nm. The broadening of this band further depends on the distortion of the local environment of the extra-framework aluminum species. In the region 200–300 nm, centered on 260 nm, there is also an absorption band due to dp–pp charge-transfer transition between any framework iron and oxygen, whilst the band around 360 nm will also contain a similar contribution from extra-framework octahedrally coordinated Fe3+ species. A weak broad band about 500 nm is associated with F2O3 species [52–55]. Fig. 7 compares the DRS (UV–vis) spectra of NZ and the modified samples (HZ, HZD1, HZD3 and HZD5) in region from 190 to 600 nm. We can see the increase in the intensity and broadening of the band around 360 nm in HZ spectra compared to NZ. This result confirms that during the preparation of HZ sample, a portion of the aluminum and some iron is extracted from zeolite framework forming distortional extra-framework species. However a decrease in the intensity of this 360 nm band together with a narrowing of its profile during the acid treatments is indicative of a decrease in the distortion of the environment of the extra-framework Al and iron species. The weak broad band around 500 nm associated with Fe2O3 remains unchanged during the dealumination processes. 3.6. Fig. 6. FTIR spectra of natural and modified zeolites in the (a) 2000–380 cm1 range and (b) 4000–3000 cm1 range.

The bands at 1400 cm1 present in HZ and HZD1, but absent in HZD3 and HZD5, are associated with a vibrational mode of NHþ 4 [38,39] which remains after the calcination processes. The bending 1 frequency of water molecules (at 1648 cm ) changes only marginally with the various treatments, meaning that the state of the water molecules in all these zeolites is very similar. A new absorption band appears at approximately 933 cm1 in the IR spectra of HZD1, HZD3 and (more intensively in) HZD5, which has been assigned to uncoupled (Si–O) vibrations belonging to Si–OH groups present at internal hydroxyl nest defects [40–43]. Fig. 6b shows the effect of dealumination on the O–H stretching region (3900–3000 cm1). Three characteristic bands appear in the spectra of all the samples, at 3745 cm1, 3625 cm1 and 3444 cm1. The small shoulder at 3745 cm1 is due to isolated silanol groups (Si–OH), the peak at 3625 cm1 is due to Si–O(H)–Al and the broad peak at 3444 cm1 is attributed to hydrogen bonded Si–OH groups in nest defects and hydrogen bonding of loosely held water molecules [44–47]. The intensity of the peak at 3620 cm1 (Si–O(H)–Al) decreases during the dealumination, confirming

27

Al and

29

Si MAS NMR spectra analysis

Only one signal associated to tetrahedral aluminium species at 55 ppm is observed in the 27Al NMR spectrum of natural zeolite NZ (see Fig. 8a). However, the 27Al NMR of the protonic and dealumi-

Fig. 7. UV–vis-DRS spectra of the natural and modified zeolites.

193

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

Fig. 8.

Table 6 Quantitative analysis of the

27

Al MAS NMR spectra of NZ (a), HZ (b), HZD1 (c), HZD3 (d) and HZD5 (e) samples.

27

Al NMR spectra of the natural and modified clinoptilolite.

Sample

Tetrahedral-Al (%)

Octahedral-Al (%)

(Si/Al)afr

(Si/Al)total

NZ HZ HZD1 HZD3 HZD5

100 93.3 94.4 92.5 92.2

0 6.7 5.6 7.5 7.8

4.48 4.81 7.42 8.16 10.84

4.48 4.34 7.00 7.55 10.00

a Determined from respectively.

atomic

Tetrahedral Al by unit cell NMR

Octahedral Al by unit cell NMR

6.57 5.94 3.97 3.61 3.20

0 0.43 0.36 0.29 0.26

27

Al MAS NMR used the expression: (Si/Al)fr = (Si/Al)total(I55 + I0)/I55, where I55 and I0 are the integrated area intensities of signals at 55 and 0 ppm,

nated samples has an additional line at 0 ppm, associated with extra-framework octahedral aluminium species. The width of this

peak decreases during the acid treatments which, similar to the above discussion of the UV–vis data, are indicative of a decrease

194

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

of the distortion in the environment of octahedral Al species. In Table 6, the percentage of tetrahedral framework and octahedral extra-framework aluminium in each sample is shown. It can be seen how the number of aluminium in tetrahedral sites T Al by unit cell, determined by NMR, decreases from HZD1 to HZD5 samples. Similarly the amount of aluminium in octahedral sites decreases with the number of treatments, which indicates the effectiveness of the ultrasonic washing steps applied after each dealumination procedure in removing the octahedral Al from the zeolite channels.

Fig. 9.

29

The residual extra-framework Al in HZD5 is less than 5% of the total Al content of the parent natural sample. Table 6 also confirms that the Si/Al ratio determined by chemical analysis and 27Al NMR are comparable. No evidence is found of any other aluminium species during the acid treatments performed in this work (for example penta-coordinated Al species which are visible in the NMR spectra at around 30 ppm [56,57]). The 29Si MAS spectra of natural and dealuminated samples are shown in Fig. 9. The majority of 29Si NMR studies of clinoptilolite

Si MAS NMR spectra of NZ (a), HZ (b), HZD1 (c), HZD3 (d) and HZD5 (e) samples.

195

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196 Table 7 Peaks position and relative area percentage extracted from he de-convolution of the Sample

NZ HZ HZD1 HZD3 HZD5

Peak1 Si(2Al)

Peak2 Si(1Al)

29

Si MAS NMR spectra.

Peak3 Si(1Al)

Peak5 Si(Fe3+)

Peak4 Si(0Al)

Position (ppm)

Area %

Position (ppm)

Area %

Position (ppm)

Area %

Position (ppm)

Area %

Position (ppm)

Area %

94.9 93.9 96.6 95.1 96.3

5.7 6.1 4.0 3.1 2.7

100.9 100.8 101.9 102.5 102.5

34.0 30.3 25.5 23.2 19.8

106.9 106.7 107.4 107.6 107.6

45.3 42.1 42.0 39.9 39.1

112.7 113.7 112.7 112.7 112.8

13.7 20.3 28.3 33.8 38.4

97.8 96.7 – – –

2.3 1.2 – – –

report four resonance lines, corresponding to Si with between one and four aluminum neighbours. Here, however, we observe five peaks associated with different Si environment in the framework. The first four signals are related to the following configurations: Si(2Al) 95 ppm, Si(1Al) 100 ppm, Si(1Al) 106 ppm and Si(0Al) 111 ppm, respectively [58–60]. The 5th peak at 97 ppm is, we suggest, related to the presence of Fe3+ framework species to NMR spectra, i.e. Si centers which have iron in the second coordination sphere, rather than aluminum [61–63]. This peak has a very small intensity in the NZ and HZ samples but then disappears in the spectra of the HZD1, HZD3 and HZD5 samples. Our assignment also concurs with that of Rodriguez-Fuentes et al. [27] who found that contribution of the Fe in zeolite framework is localized around 95 ppm, overlapping with the Si(2Al) signal, and who further observed that this band disappeared after the zeolite was treated with orthophosphoric acid. Table 7 shows the positions and area of each peak determined by a de-convolution process of the NMR spectra. We can see that the intensity of the Si(0Al) peak increases as the result of dealumination of zeolite framework, while the intensities of peaks associated with Si (1Al) and Si(2Al) decreases. 4. Conclusion In the present work a new mild condition strategy for the stepwise dealumination of zeolites is introduced and applied to clinoptilolite. A significant amount of aluminum (46%) is extracted from the clinoptilolite framework – far more than previous studies [17,18,35,64–66] – with no significant effect on the zeolite crystallinity and a practically negligible amount of extra-framework aluminum species being retained in the pores. The formation of vacancies in the zeolite framework is confirmed by IR and NMR spectra. In particular, the sample obtained by three cycles of the acid treatment (HZD3) shows little sign of amorphisation and is a good candidate for evaluating the effect of increasing Si/Al in clinoptilolite on gas adsorption and separation processes: current work is evaluating the performance of transition metal exchanged dealumination samples in such applications. Acknowledgments This work was partially supported by European Union-sponsored Alfa NANOGASTOR Network (II-0493-FA-FI). Partial support from University of Havana is also acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.micromeso.2010.07.008. References [1] J. Weitkamp, M. Hunger, in: J. Cejka, H. van Beekkum, A. Corma, F. Schuth (Eds.), Introduction to zeolite science and pratice, third rev. ed., Elsevier, Amsterdam, 2007, p. 787

[2] P. Sazama, J. Dedecek, V. Gabova, B. Wichterlova, J. Catal. 254 (2008) 180. [3] J. O’Brien-Abraham, Y.S. Lin, Ind. Eng. Chem. Res. 49 (2010) 809. [4] W.J. Mortier, K.U. Leuven, Electronegativity, Springer Berlin/Heidelberg, Berlin, 1987. p. 125. [5] R. Heidler, W.J. Mortier, G.O.A. Janssens, R.A. Schoonheydt, J. Phys. Chem. 100 (1996) 19728. [6] G.O.A. Janssens, H.T. Baekelandt, W.J. Mortier, R.A. Schoonheydt, J. Phys. Chem. 99 (1995) 3251. [7] M. Johnson, D. O’Connor, P. Barnes, C. Richard, A. Catlow, S.L. Owens, G. Sankar, R. Bell, S.J. Teat, R. Stephenson, J. Phys. Chem. B 107 (2003) 942. [8] N. Dobelin, T. Armbruster, Micropor. Mesopor. Mater. 61 (2003) 85. [9] R.M. Barrer, M.B. Makki, Canad. J. Chem. 42 (1964) 1481. [10] C.V. McDaniel, P.K. Maher (Eds.), Molecular Sieves, London, 1968. [11] R. Giudici, H.W. Kouwenhoven, R. Prins, Appl. Catal. A 203 (2000) 101. [12] D.S. Coombs, A. Alberti, T. Armbruster, G. Artioli, C. Colella, E. Calli, J.D. Grice, F. Liebau, J.A. Mandarino, H. Minato, E.H. Nickel, E. Passaglia, D.R. Peacor, S. Quartieri, R. Rinaldi, M. Ross, R.A. Sheppard, E. Tillmanns, G. Vezzalini, Can. Mineral 35 (1997) 1571. [13] A. Alberti, Mineral Petrogr. Mitt. 18 (1972) 129. [14] K. Koyama, Y. Takeuchi, Z. Kristallogr. 145 (1977) 216. [15] T. Armbruster, M.E. Gunter, Am. Miner. 76 (1991) 1872. [16] G.E. Christidis, D. Moraetis, E. Keheyan, L. Akhalbedashvili, N. Kekelidze, R. Gevorkyan, H. Yeritsyan, H. Sargsyan, Appl. Clay Sci. 24 (2003) 79. [17] A. Radosavljevic-Mihajlovic, V. Dondur, A. Dradakovic, J. Lemic, M. TomaseviCanovic, J. Serb. Chem. Soc. 69 (2004) 273. [18] F. Cakicioglu-Ozkan, S. Ulku, Micropor. Mesopor. Mater. 77 (2005) 47. [19] P. Misaelides, A. Godelitsas, F. Link, H. Baumann, Microp. Mater. 6 (1996) 37. [20] A. Arcoya, J.G. González, N. Travieso, X.L. Seoane, Clay Miner. 29 (1994) 123. [21] G. Rodriguez-Fuentes, M.A. Barrios, A. Iraizoz, I. Perdomo, B. Cedre, Zeolites 19 (1997) 441. [22] I. Rodriguez-Iznaga, A. Gomez, G. Rodriguez-Fuentes, A. Benitez, J. Serrano, Micropor. Mesopor. Mater. 53 (2002) 71. [23] Y. Garcia-Basabe, A. Gomez, I. Rodriguez-Iznaga, A. Montero, G. Vlaic, A. Lausi, A.R. Ruiz-Salvador, J. Phys. Chem. C 114 (2010) 5964. [24] T. Farías, A.R. Ruiz-Salvador, L. Velazco, L.C. de Ménorval, A. Rivera, Mater. Chem. Phys. 118 (2009) 322. [25] M. Rozic, S. Cerjan-Stefanovic, S. Kurajica, M. Rozmaric Maeefat, K. Margeta, A. Farkas, J. Colloid Interface Sci. 284 (2005) 48. [26] A.R. Ruiz-Salvador, A. Gomez, D.W. Lewis, C.R.A. Catlow, L.M. RodríguezAlbelo, L. Montero, G. Rodríguez-Fuentes, Phys. Chem. Chem. Phys. 2 (2000) 1803. [27] G. Rodriguez-Fuentes, L.C. de Ménorval, E. Reguera, F. Chávez Rivas, Micropor. Mesopor. Mater. 111 (2008) 577. [28] A. Arcoya, X.L. Seoane, J. Soria, J. Chem. Tech. Biotechnol. 68 (1997) 171. [29] J.F. Marco, M. Gracia, J.R. Gancedo, T. González-Carreño, A. Arcoya, X.L. Seoane, Hyperfine Interact. 95 (1995) 53. [30] J. Rodriguez-Carvajal, FULLPROF-Suite 2005, Institute Leon Brillouin, Scaly, 2005. [31] T. Armbruster, Am. Miner. 78 (1993) 260. [32] A. Alberti, Miner. Petr. Mitt. 22 (1975) 25. [33] A.R. Ruiz-Salvador, N. Almora-Barrios, A. Gomez, D.W. Lewis, Phys. Chem. Chem. Phys. 9 (2007) 521. [34] H. Kurama, A. Zimmer, W. Reschetilowski, Chem. Eng. Technol. 25 (2002) 301. [35] M.A. Hernandez, F. Rojas, V.H. Lara, J. Porous Mater. 7 (2000) 443. [36] O. Korkuna, R. Leboda, J. Skubiszewska-Zieba, T. Vrublevska, V.M. Gunko, J. Ryczkowski, Micropor. Mesopor. Mater. 87 (2006) 243. [37] W. Mozgawa, T. Bajda, J. Mol. Struct. Theochem. 170 (2006) 792–793. [38] B. Tomazovic, T. Ceranic, G. Sijaric, Zeolites 16 (1996) 301. [39] B. Tomazovic, T. Ceranic´, G. Sijaric, Zeolites 16 (1996) 309. [40] M. Decottigues, J. Phalippou, Z. Zarzycki, J. Mater. Sci. 13 (1978) 2605. [41] M.A. Camblor, A. Corma, A. Martinez, J. Perez-Pariente, J. Chem. Soc. Chem. Commun. 589 (1995). [42] S. Dzwigaj, M.J. Peltre, P. Massiani, A. Davidson, M. Che, T. Sen, S. Sivasanker, Chem. Commun. (1998) 87. [43] R. Soda, Bull. Chem. Soc. 34 (1961) 1491. [44] J. Datka, M. Boczar, B. Gil, Colloid Surf. A 105 (1995) 1. [45] M. Maache, A. Janin, J.C. Lavalley, E. Benazzi, Zeolites 15 (1995) 507. [46] D.M. Roberge, H. Hausmann, W.F. Holderich, Phys. Chem. Chem. Phys. 4 (2002) 3128. [47] M.S. Joshi, V.V. Joshi, A.L. Choudhari, M.W. Kasture, Mater. Chem. Phys. 48 (1997).

196

Y. Garcia-Basabe et al. / Microporous and Mesoporous Materials 135 (2010) 187–196

[48] A.A. Sokol, C.R.A. Catlow, J.M. Garcés, A. Kuperman, J. Phys. Condens. Matter 16 (2004) 2781. [49] E.D. Garbowski, C. Mirodatos, J. Phys. Chem. 86 (1982) 97. [50] M.A. Zanjanchi, A. Razavi, Spectrochim. Acta A 57 (2001) 119. [51] M.A. Zanjanchi, M. Hemmati, Mater. Chem. Phys. 85 (2004) 334. [52] S. Bordiga, R. Buzzoni, F. Geobaldo, C. Lamberti, E. Giamello, A. Zecchina, G. Leofanti, G. Petrini, G. Tozzola, G. Vlaic, J. Catal. 158 (1996) 486. [53] J. Pérez-Ramírez, M. Santhosh Kumar, A. Brückner, J. Catal. 223 (2004) 13. [54] G. Berlier, G. Spoto, S. Bordiga, G. Ricchiardi, P. Fisicaro, A. Zecchina, I. Rossetti, E. Selli, L. Forni, E. Giamello, C. Lamberti, J. Catal. 208 (2002) 64. [55] Y. Jia, W. Han, G. Xiong, W. Yang, Sci. Technol. Adv. Mater. 8 (2007) 106. [56] J.A. van Bokhoven, D.C. Koningsberger, P. Kunkeler, H. van Bekkum, A.P.M. Kentgens, J. Am. Chem. Soc. 122 (2000) 12842. [57] F. Deng, Y. Yue, C. Ye, J. Phys. Chem. B 102 (1998) 5252. [58] A. Rivera, T. Farías, A.R. Ruiz-Salvador, L.C. de Ménorval, Micropor. Mesopor. Mater. 61 (2003) 249.

[59] G. Engelhardt, D. Michel, High Resolution Solid State NMR of Silicates and Zeolites, Wiley, New York, 1987. [60] S. Ramdas, J. Klinowski, Nature 308 (1984) 521. [61] C.A. Fyfe, Y. Feng, H. Grondey, G.T. Kokotailo, H. Gies, Chem. Rev. 91 (1991) 1525. [62] E. Lippmaa, M. Magi, A. Samoson, M. Tarmak, G. Engelhardt, J. Am. Chem. Soc. 103 (1981) 4992. [63] C.A. Fyfe, J.M. Thomas, J. Klinowski, C.G. Gobbi, Angew. Chem. 22 (1983) 259. [64] M.P. Elizalde-González, M.A. Pérez-Cruz, J. Colloid Interf. Sci. 312 (2007) 317. [65] E. Kouvelos, K. Kesore, T. Steriotis, H. Grigoropoulou, D. Bouloubasi, N. Theophilou, S. Tzintzos, N. Kanelopoulos, Micropor. Mesopor. Mater. 99 (2007) 106. [66] V.O. Vasylechko, G.V. Gryshchouk, Y.B. Kuzma, V.P. Zakordonskiy, L.O. Vasylechko, L.O. Lebedynets, M.B. Kalytovska, Micropor. Mesopor. Mater. 60 (2003) 183–196.