- Email: [email protected]
Textural properties of layered double hydroxides: effect of magnesium substitution by copper or iron
Microporous and Mesoporous Materials 47 (2001) 275±284
www.elsevier.com/locate/micromeso
Textural properties of layered double hydroxides: eect of magnesium substitution by copper or iron Gabriela Carja *,1, Ryuichi Nakamura, Takashi Aida, Hiroo Niiyama Department of Chemical Engineering, Tokyo Institute of Technology 2-12-1, O-okayama, Meguro-ku, Tokyo 152-8550, Japan Received 6 June 2000; received in revised form 16 April 2001; accepted 9 June 2001
Abstract Layered double hydroxides (LDHs), in which magnesium was partially substituted by iron or copper, were synthesized by a coprecipitation method. The materials were characterized by X-ray diraction, X-ray ¯uorescence, FT-IR spectroscopy, N2 adsorption, scanning electron microscopy and transmission electron microscopy. The treatment of the adsorption±desorption isotherms with dierent computation models permitted the determination of several parameters, useful for thorough characterization of the microstructures of the samples. The substituted samples preserved the hydrotalcite-type layered structure though their textural properties underwent important modi®cations. For copper substituted LDH, a relatively uniform porous structure with emphasized mesoporous characteristics emerged, the mesopores were enlarged and the speci®c surface area was decreased. On the contrary, for iron substituted LDH, the microporosity features developed, nonuniformity and constrictions in the porous structure was accentuated, the pore size was decreased and the speci®c surface area was increased. Microscopic morphology characteristics contributed to establish the textural properties of the samples. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Hydrotalcites; Textural properties; Mesoporosity; Microporosity; Microscopic morphology
1. Introduction Hydrotalcite is a naturally occurring anionic clay with the formula Mg6 Al2 (OH)16 CO3 4H2 O. It presents a positively charged brucite-like layers (Mg(OH)2 ) in which some of Mg2 are replaced by Al3 in the octahedral sites of hydroxide sheets. * Corresponding author. Address: Mathematical Seminary ``Al. Myller'', University Al. I. Cuza, Bd. Copou no. 11, Iasi 6600, Romania. E-mail address: [email protected] (G. Carja). 1 On leave from: Department of Physical Chemistry, Faculty of Industrial Chemistry, Technical University ``Ghe. Asachi'' Bd. D. Mangeron, Iasi 6600, Romania.
Interstitial layers formed by CO23 anions and water molecules compensate the positive charge resulting from this substitution [1]. Both magnesium and aluminum can be isomorphously substituted by other divalent or trivalent cations, and a wide range of compositions containing various combinations of M(II), M(III) and dierent anions An can be synthesized [2,3]. The resulting materials, known as layered double hydroxides (LDHs) or hydrotalcite-like anionic clays, have received considerable interests in recent years owing to their applications as catalysts, catalyst precursors, adsorbents and ion exchangers [4]. In de®ning or tailoring the catalytic and adsorption properties the crystallinity and textural
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 3 8 7 - 0
276
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
characteristics are of particular importance. For example, the shape and the size of pores, associated with the uniformity of the porous structure, can promote shape selectivity; a higher surface area will facilitate guest±host interactions; even mass transfer process depends on the porosity type. Despite this, many of the published studies concerning LDHs deal only brie¯y or not at all with their textural characteristics and with the possibility to tailor them for required applications [5± 10]. In order to get new information about the possibility to tailor textural properties of LDHs, the aim of this study is to investigate how these properties can be altered through direct synthesis when magnesium is partially substituted by other metals. In the present study, we synthesized MgCuAl and MgFeAl ternary LDHs. The eects of magnesium substitution by copper or iron on the microscopic morphology and the textural properties of the materials are reported. 2. Experimental 2.1. Synthesis of layered double hydroxides All the samples were synthesized by a coprecipitation method in low supersaturation conditions [11], at 313 K, under a bubbling constant ¯ow of nitrogen in the reaction medium and vigorous stirring. After an aging step, the precipitates were separated by centrifugation, washed extensively with warm deionized water until sodium free and dried under vacuum at 338 K. 2.1.1. MgAlLDH An aqueous solution (100 ml) of Mg(NO3 )2 6H2 O (0.03 mol)/Al(NO3 )3 9H2 O (0.01 mol) and an aqueous solution of NaOH/Na2 CO3 (CO23 = Al3 Mg2 0:67, HO =Al3 Mg2 2:25), were added dropwise together, in such a way that the pH remained at a constant value of 9.5. The resulting white precipitates were aged at 338 K for 24 h under stirring. 2.1.2. CuLDHs An aqueous solution (85 ml) of Mg(NO3 )2 6H2 O ((0.03-q) mol)/Al(NO3 )3 9H2 O (0.01 mol)/
Cu(NO3 )2 3H2 O (q mol, 0:005 6 q 6 0:01) and an aqueous solution of Na2 CO3 (1 M, 30 ml) were added dropwise together over a period of 2 h at a constant pH value of 8.2. The resulting light blue precipitates were aged at 305 K for 24 h under stirring. 2.1.3. FeLDHs An aqueous solution (100 ml) of Mg(NO3 )2 6H2 O ((0.03-r) mol)/Al(NO3 )3 9H2 O (0.01 mol)/ FeSO4 7H2 O (r mol, 0.005 6 r 6 0.01) and an aqueous solution of Na2 CO3 (1 M, 35 ml) were added together over a 2 h period at a constant pH of 8.5. The resulting orange precipitates were aged at 305 K for 24 h under stirring. 2.2. Characterization The chemical compositions of synthesized samples were determined by X-ray ¯uorescence (XRF) spectroscopy (Shimadzu XRF-1700 sequential XRF spectrometer). X-ray powder diraction (XRD) patterns were recorded on a Philips PW 1840 diffractometer using monochromatic CuKa radiation (k 0:154 nm), operating at 40 kV and 30 mA over a 2h range from 4° to 70°. IR spectra were recorded on a Shimadzu FTIR spectrometer under the following experimental conditions: 200 scans in the mid-IR range (500±4000 cm 1 ) using KBr (ratio 5/95 wt.%) pellets, and a resolution of 4.0 cm 1 . N2 adsorption±desorption isotherms were measured on a Coulter SA 3100 automated gas adsorption system. Prior to the measurements the samples were heated for 5 h under vacuum at 383 K in order to expel the interlayer water molecules. The BET speci®c surface area (SBET ) was calculated by using the standard Brunauer, Emmett and Teller method [12] on the basis of the adsorption data. Pore volume, micropore area and mesopore area were determined by using the t-plot method of De Boer [14]. Pore size distributions were calculated from the desorption branches of the isotherms using the Barret, Joyner and Halenda method and the corrected Kelvin equation [13,14]. All these models are considered to be fully applicable to mesoporous materials [15]. The IUPAC classi®cation of pores and isotherms were used in this study [16]. A Hitachi S-800 scanning
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
277
electron microscope (SEM) was utilized for the observation of microscopic morphologies. The accelerating voltage was 15 kV. Transmission electron microscopy (TEM) analysis was performed on a Hitachi instrument operating at 200 kV. The samples were prepared by dispersing powders in ethanol.
3. Results and discussion 3.1. Structural characteristics The composition of the synthesized ternary hydrotalcites can be fully de®ned by the formula: Mex Mg3 x Al, where Me is Cu or Fe (atomic ratio: Mg=Me=Al 3 x=x=1). Therefore we will use the notation xMeLDH to denote the composition of the samples and to de®ne the ternary sample name. The chemical compositions and some structural properties of the synthesized samples are presented in Table 1. The Mg=Me=Al atomic ratios determined by XRF are coincident with those of the starting mixed aqueous solutions within the experimental errors. The XRF results reveal that only magnesium (no aluminum) was substituted by iron during the synthesis procedure. The XRD patterns (Fig. 1) exhibit sharp and symmetric re¯ections for the (0 0 3), (0 0 6), (1 1 0) and (1 1 3) planes and broad symmetric peaks for the (1 0 2), (1 0 5) and (1 0 8) planes, which are characteristics of these materials [17]. The lattice parameters are calculated by indexing the peaks under a hexagonal crystal system, using a least squared method. The parameter a corresponds to the cation±cation distance within the brucite like layer while the parameter c is related to the total
Fig. 1. XRD patterns of: (a) MgAlLDH, (b) 0.5CuLDH, (c) CuLDH, (d) 0.5FeLDH and (e) FeLDH.
thickness of the brucite-like layer and the interlayer distance [1,18]. For the xCuLDH samples the parameter a has almost the same value as for the MgAlLDH. The result is due to the correspondence between the ionic radius of Cu2 hexacoor Shannon ionic radii [19]) and the dinate (0.73 A ionic radius of Mg2 hexacoordinate (0.72 A, Shannon ionic radii). The decrease of the parameter a for the xFeLDHs samples clearly indicates the incorporation of iron into the brucite-like
Table 1 Chemical composition and lattice parameters of MgAlLDH and MeLDHs Sample
Mg:Me:Al from XRF
Me (% mass) from XRF
XRD phase
Lattice parameters (A)
IFS (A)
a
c
d0 0 3
MgAlLDH 0.5CuLDH CuLDH 0.5FeLDH FeLDH
2.94:0:1 2.4:0.6:1 2.1:1:1.1 2.4:0.57:1 2:1.07:1
0 27 46 24 50
LDH LDH LDH LDH LDH
3.059 3.059 3.058 3.044 3.040
23.38 23.21 23.17 22.80 22.76
7.79 7.73 7.72 7.60 7.58
2.99 2.93 2.92 2.80 2.78
278
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
The layer, as the ionic radius of Fe2 is 0.61 A. decrease of the parameter c and the d0 0 3 (basal spacing calculated from the 0 0 3 re¯ection position in the XRD pattern [1,3]) can be attributed to the modi®ed electrostatic interactions between the layer and the interlayer network when another metal is introduced in the LDH layer [18]. The interlayer free spacing (IFS) values are calculated by subtracting the thickness of the LDH layer (4.8 [1]) from the calculated d0 0 3 spacing. The values A, for xCuLDHs and to decrease from 2.99 to 2.92 A 2.78 A for xFeLDHs, respectively. The decrease in IFS value suggests a distortion of the LDH network induced in the process of isomorphous substitution of magnesium by copper or iron. The distortion of LDH network is accentuated for the iron substituted samples. XRF results, together with the structural changes that are revealed by XRD, demonstrate the incorporation of iron and copper in the LDHs network. Fig. 2 presents the FTIR spectra of MgAlLDH, CuLDH and FeLDH. In the case of LDHs IR analysis is used to get information about the anions present in the interlayer between the bru-
cite-like sheets [1]. For all the samples IR spectra are quite similar though some dierences can be noticed in the intensity and the broadness of the bands. All the spectra exhibit an intense broad band between 3800 and 2700 cm 1 which comprises the vibrations of physisorbed water [11,20], vibrations of structural OH groups [20], characteristic valency vibrations of OH OH and/or characteristic stretching vibrations of CO23 OH in hydrotalcite [21]. The band corresponding to the vibration mode dOH appears between 1644± 1648 cm 1 and may be assigned to the adsorbed interlayer water [22]. An intense absorption band between 1370±1383 cm 1 is attributed to the m3 vibration mode of CO23 [23]. In the same range the m3 vibration mode of nitrate anion could appear if this still exists in the interlayer. The presence of the band between 1504±1537 cm 1 can be ascribed to a lowered carbonate symmetry that activated the vibration mode m1 of CO23 [24]. For FeLDH the band that appear at 1122 cm 1 is due to the m3 vibration mode of SO24 anion that still exists in the interlayer; a very weak band corresponding to the m4 vibration mode of this anion also appears at 650 cm 1 [25]. The bands due to m4 and respective m2 vibration modes of CO23 anion appear in all the spectra between 667±680 and 870±880 cm 1 , respectively [24]. We can remark some differences in the intensities of these bands between MgAlLDH and the substituted samples; this fact can be ascribed to a dierent orientation of CO23 in the interlayer as a consequence of dierent electrostatic forces that act in the network of substituted samples. For all the spectra weak bands appear at 635, 770 and 939 cm 1 corresponding to the hydrotalcite lattice vibrations [26,27]. 3.2. Textural characteristics
Fig. 2. FTIR spectra of: (a) MgAlLDH, (b) CuLDH and (c) FeLDH.
The nitrogen adsorption isotherms of MgAlLDH, CuLDH and FeLDH are shown in Fig. 3(a)±(c). Their dierent appearances suggest the modi®cations of the porosity characteristics when copper or iron is substituted for magnesium in the LDH network. For MgAlLDH (Fig. 3(a)) we observe a type IV isotherm with a broad type H3 hysteresis loop, in the middle range of relative
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
279
Fig. 3. N2 adsorption±desorption isotherms of (a) MgAlLDH, (b) CuLDH and (c) FeLDH.
pressure. At low relative pressure the prevailing process is the formation of a monolayer while a multilayer adsorption takes place at a high relative pressure. For the values of relative pressure higher than 0.8 condensation takes place giving a sharp adsorption volume increase. This behavior indicates that this sample has a mesoporous character. Hysteresis loop type shows that aggregates of plate-like particles forming enough nonuniform slit shaped pores [16] are present in this sample. For CuLDH (Fig. 3(b)) the shape of the curves is typical for a type IV adsorption isotherm and a type H1 hysteresis. The isotherm form and the
narrow hysteresis loop, with almost parallel adsorption and desorption branches, indicate that cylindrical pores with mesopore size range and almost regular geometry, more uniform in size and/or shape than those present in MgAlLDH, exist in the copper substituted sample. For FeLDH (Fig. 3(c)) the adsorption isotherm is very similar to that found for porous zircon [28] and presents particular characteristics. That is: · for ps =p0 < 0:2, multilayer adsorption is present (see also Table 2) revealing microporous characteristics for this sample;
280
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
Table 2 Characteristics of N2 adsorption/desorption MgAlLDH, CuLDH and FeLDH Sample
nav ps =p0 < 0:25
ps =p0 1
MgAlLDH CuLDH FeLDH
1 1.30 2.4
15 21.7 6
a
data
VN2 (ml/g)
dp a (nm)
460 432 189
12.8 20 4.1
for
dp 4VN2 =SBET .
· for 0:2 < ps =p0 < 0:8, the isotherm shape reveals the presence of mesopores with a wide size range; · for ps =p0 > 0:8, it can be assimilated with a type II isotherm, characteristic for adsorption in pores with sizes larger than 50 nm. The irregular form of type H3 hysteresis loop also indicates the presence of nonuniformity in pores size and/or shape. These features allow us to conclude that, for FeLDH, both microporous and mesoporous properties contribute to establish the characteristics of the microstructure. Pore size distribution (PSD) is further used to check and to complete the previous results. The PSD curves, Fig. 4, point out important modi®cations in the pore size range of the samples. For FeLDH a sharp and high peak with a maximum
Fig. 4. Pore distribution in MgAlLDH, CuLDH and FeLDH.
around 3.5 nm appears. This shows that small pores (micropores and very small mesopores) establish the porosity characteristics of this material. Large mesopores and even some macropores also developed and the porous structure of FeLDH has a pronounced nonuniformity. The PSD curve of MgAlLDH is broader and is shifted to a larger size with two maximums at 9.2 and 12.3 nm. Pores with sizes belonging to an entire range characteristic for mesoporosity appear in this case. The PSD curve of the copper substituted sample with two maximums at 27 and 37 nm shows clearly the enlargement of the pore size. All the pores are larger than 5 nm and we observe no mesopores with sizes larger than 45 nm. Consequently neither micropores nor macropores are detected, revealing that the uniformity of the porous structure is accentuated. Therefore, the PSD results complete the analysis of the isotherms and demonstrate that for the copper containing hydrotalcite-like sample the pore size increased by nearly 15 nm though for the iron substituted sample the mesopores size decreased signi®cantly and some very large mesopores are also detected. The contribution of pores, as a function of their size range, to the total pore volume was calculated using the PSD results. It can be seen (Fig. 5) that the pores belonging to the range of 5±20 nm contribute by nearly 60% to the total pore volume of MgAlLDH, while the contribution due to the pores with sizes between 20 and 40 nm is only 22%. For CuLDH, almost 80% of the pore volume is due to the pores belonging to the range of 20±40 nm. For FeLDH it is not possible to see a distinct major contribution as pores belonging to the range of 0±5 nm contribute by around 22% to the pore volume and pores with sizes larger than 40 nm are present with a contribution equal by nearly 27%. The result also emphasizes that CuLDH possesses a more uniform porous structure while nonuniformity of the porous structure is accentuated for FeLDH. Other aspects of the alteration of the textural characteristics of LDHs are revealed by the average number of nitrogen layers (nav ) formed during N2 adsorption. The nav values are calculated by using the adsorbed monolayer volume and the
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
Fig. 5. Contribution of pores, as a function of their size range, to the total pore volume in MgAlLDH, CuLDH and FeLDH.
values of total adsorbed volume (VN2 ). The results are presented in Table 2. At low relative pressure, (ps =p0 < 0:25), a monolayer is formed for both MgAlLDH and CuLDH, though a multilayer adsorption takes place in the case of FeLDH. The number of nitrogen layers is signi®cantly higher for ps =p0 values close to unity, being 6 for FeLDH, 15 for MgAlLDH and nearly 22 for CuLDH. The volume of nitrogen adsorbed at the highest value of relative pressure, (VN2 ), decreases for FeLDH. These results also con®rm the accentuation of the mesoporous character of CuLDH (as multilayer adsorption is stressed at higher values of relative pressure) and the development of the microporosity features in FeLDH (as multilayer adsorption take place at low values of relative pressure; nav is lower for values of relative pressure close to unity; volume of nitrogen adsorbed at the highest value ps =p0 decreases). The estimation of the average pore diameter (dp ), by the classical Gurvitsch rule [13], is given in the last column of Table 2. According to the Gurvitsch rule, average pore diameter (which in fact does not mean the eective pore diameter because the contribution of the external surface in
281
the total surface is also included [29]) decreases from 20 nm for CuLDH to 12.8 nm for MgLDH and becomes equal to 4.1 nm for FeLDH. The values are enough close to those obtained from PSD. To reveal another estimation for the modi®cations of the textural properties we calculated the values of the total pore volume (Vp ), and the extent of micropore area (%lpA) and mesopore area (%mesopA) in the total t-plot area (see Table 3). The values of BET surface area are also given in the second column of the table. It can be seen that the FeLDHs develop higher BET surface areas in comparison to MgAlLDH while a decline of BET surface areas is characteristic for the copper substituted samples. The extent of area formed by mesopores in the total t-plot area is nearly 99% for CuLDH but decreases to 86% for MgAlLDH while the value is only 65% for FeLDH. The contribution of mesopores and micropores to the total surface area indicates that the increase in the surface area of iron containing LDH is due to the development of microporosity characteristics; the decrease of the surface area of copper containing LDH is due to the accentuation of mesoporosity characteristics. The indirect proportionality between the pore volume and the BET surface area suggests that the porosity properties are in¯uenced by the microscopic morphology of the samples. All the computation models applied to the adsorption isotherm data gave similar results that point out important transformations of the textural properties for copper and iron substituted LDHs. With copper substitution the mesopore size increases and a more uniform porous structure appears. For iron containing LDH the pore size
Table 3 BET surface area and porosity characteristics obtained from N2 adsorption data using the t-plot method of De Boer Sample
BET area (m2 /g)
Vp (ml/g)
%lpA
%mesopA
MgAlLDH 0.5CuLDH CuLDH 0.5FeLDH FeLDH
132 117 86 162 181
0.640 0.647 0.661 0.411 0.277
14 11 1 22 35
86 89 99 79 65
282
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
decreases (85% of the pores are smaller than 5 nm) and the nonuniformity of the porous structure was accentuated. Hence, micropores, very small mesopores and also some large pores coexist. A large decrease of the BET surface area is characteristic for the copper substituted samples though the pore volume is enhanced. Contrary for iron substituted LDHs the BET surface area values increase while the pore volume and the adsorption capacity show lower values. These transformations are the result of the enhancement of mesoporosity characteristics for the copper substituted samples and the
development of microporosity features for the iron substituted samples. It is quite dicult to explain the transformations that appear in the porosity features of LDHs when a part of magnesium is substituted by copper and iron. The LDH layered structure with an inter cannot be relayer free space smaller than 3 A sponsible for any mesoporosity characteristics. The mesoporous structure of LDHs may arise from interparticle space hence the characteristics of mesopores could be de®ned by the particle size and shape and also by the particles interconnection
Fig. 6. (a) SEM micrograph of MgAlLDH sample. (b) SEM micrograph of CuLDH sample. (c) SEM micrograph of FeLDH sample.
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
patterns. In this view the changes in the porosity features are the result of altering the microscopic morphology characteristics. The SEM images (Fig. 6(a)±(c)) show that MgAlLDH and the substituted samples are highly crystalline. Intersected, nearly hexagonal-shaped particles, interconnected with each other, emerged for all the samples. Notwithstanding this, a different aspect emerged in the appearances. The differences are mainly due to the dierent mode of interconnection and agglomeration of the particles in ternary LDHs and also to the modi®cation of the particle size. Interparticle space is larger for CuLDH (Fig. 6(b)) than for MgAlLDH (Fig. 6(a)) or FeLDH (Fig. 6(c)). Wider interparticle cavities developed in the iron-substituted sample; these can be considered as the origins of the large pores.
283
The average value of particle size is 150 nm for CuLDH, 130 nm for MgAlLDH and 100 nm for FeLDH. The TEM image of the Cu substituted sample (Fig. 7) shows the presence of aggregates of well-de®ned particles, characteristics for hydrotalcite-like materials [30,31]. A careful inspection of the image reveals that the particles themselves do not present any mesoporosity. The SEM and TEM results further support the idea of the connection between the textural properties and the microscopic morphology of the samples. The results reported previously [10,30] claimed that the textural properties of hydrotalcite-like materials are strongly in¯uenced by the synthesis conditions. We assume that the dierent pH values, the distinct nature and electronegativities of the ions present in the synthesis medium, and the dierent concentrations of the starting solutions could be important factors to control the microscopic morphology and to determine the changes of the textural properties.
4. Conclusions The current study demonstrated that the microstructures of ternary layered double hydroxides are signi®cantly transformed when copper or iron substituted magnesium in the LDH layer. As a general trend copper substituted LDH presented a mesoporous structure with a higher uniformity than MgAlLDH, while microporous characteristics developed in iron substituted LDH and a microporous±mesoporous structure with a more accentuated nonuniformity emerged. The possibility of altering the textural properties of LDHs when a new metal is introduced in the layer may oer new perspectives in tailoring the textural characteristics of these materials for potential applications.
Acknowledgements
Fig. 7. Characteristic TEM image of Cu substituted sample.
G.C. acknowledges a research grant from Ministry of Education, Science, Sports and Culture of Japan.
284
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
References [1] F. Cavani, F. Tri®ro, A. Vaccari, Catal. Today 11 (1991) 173 and references therein. [2] W. Jones, M. Chibwe, in: I.V. Mitchel (Ed.), Pillared Layer Structures Current Trends and Applications, Elsevier, London, 1991, p. 67. [3] F. Tri®ro, A. Vaccari, in: J.L. Atwood, J.E. Davies, D.D. Mac Nicol, F.Vogtle (Eds.), Comprehensive Supramolecular Chemistry, vol. 7, Pergamon Press, Oxford, 1996, p. 251. [4] A. Vaccari, Catal. Today 41 (1998) 53. [5] H. Nijs, M. De Bock, E.F. Vansant, Micropor. Mesopor. Mater. 30 (1999) 245. [6] K. Chibwe, W. Jones, Chem. Mater. 1 (1989) 489. [7] J. Don Wang, G. Serrette, Y. Tian, A. Clear®eld, Appl. Clay Sci. 10 (1995) 105. [8] T. Kwon, T. Pinnavaia, Chem. Mater. 1 (1989) 381. [9] A. Corma, V. Fornes, F. Rey, J. Catal. 151 (1995) 50. [10] F. Malherbe, C. Forano, J.P. Besse, Micropor. Mater. 10 (1997) 67. [11] S. Miyata, Clays Clay Miner. 31 (1983) 305. [12] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309. [13] M. Jaroniec, M. Kruk, A. Sayari, in: L. Bonneviot, F. Beland, C. Danumah, S. Giasson, S. Kaliaguine (Eds.), Studies in Surface Science and Catalysis, vol. 117, Elsevier, Amsterdam, 1994, p. 325. [14] S.J. Gregg, K.S.W. Sing, Adsorption Surface Area and Porosity, second ed., Academic Press, London, 1982. [15] M. Kruk, M. Jaroniek, A. Sayari, Micropor. Mesopor. Mater. 27 (1999) 217.
[16] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603. [17] A. Vaccari, M. Gazzano, in: G. Poncelet, J. Martens, B. Delmon, P.A. Jacobs, P. Grange (Eds.), Preparation of Catalysts VI, Studies in Surface Science and Catalysis, vol. 91, Elsevier, Amsterdam, 1995, p. 309. [18] S. Velu, N. Shah, T.M. Jyothi, S. Sivasanker, Micropor. Mesopor. Mater. 33 (1999) 61. [19] R.D. Shanon, Acta Crystallogr. A 32 (1976) 751. [20] M.J.H. Hernandez, M.A. Ulibarri, J.L. Rendon, C.J. Serna, Phys. Chem. Miner. 12 (1985) 34. [21] D. Roy, R. Roy, E. Osborn, Am. J. Sci. 251 (1973) 337. [22] R. Allmann, Chimia 24 (1970) 99. [23] D.L. Bish, G.W. Brindley, Am. Miner. 62 (1977) 458. [24] S. Miyata, A. Okada, Clays Clay Miner. 44 (1977) 14. [25] S. Kannan, C.S. Swamy in: G. Poncelet, J. Martens, B. Delmon, P.A. Jacobs, P. Grange (Eds.), Preparation of Catalysts VI, Studies in Surface Science and Catalysis, vol. 91, Elsevier, Amsterdam, 1995, p. 903. [26] S. Miyata, Clays Clay Miner. 23 (1975) 369. [27] M.A. Aramendia, Y. Aviles, J.A. Benitez, V. Borau, C. Jimenez, J.M. Marinas, J.R. Ruis, F.J. Urbano, Micropor. Mesopor. Mater. 29 (1999) 319. [28] R. Valero, B. Durand, J.L. Guth, T. Chopin, Micropor. Mesopor. Mater. 29 (1999) 311. [29] A. Gervasini, Appl. Catal. A General 180 (1999) 71. [30] F. Prinetto, G. Ghiotti, P. Gran, D. Tichit, Micropor. Mesopor. Mater. 39 (2000) 229. [31] O. Clause, B. Rebours, E. Merlen, F. Tri®ro, A. Vaccari, J. Catal. 133 (1992) 231.
www.elsevier.com/locate/micromeso
Textural properties of layered double hydroxides: eect of magnesium substitution by copper or iron Gabriela Carja *,1, Ryuichi Nakamura, Takashi Aida, Hiroo Niiyama Department of Chemical Engineering, Tokyo Institute of Technology 2-12-1, O-okayama, Meguro-ku, Tokyo 152-8550, Japan Received 6 June 2000; received in revised form 16 April 2001; accepted 9 June 2001
Abstract Layered double hydroxides (LDHs), in which magnesium was partially substituted by iron or copper, were synthesized by a coprecipitation method. The materials were characterized by X-ray diraction, X-ray ¯uorescence, FT-IR spectroscopy, N2 adsorption, scanning electron microscopy and transmission electron microscopy. The treatment of the adsorption±desorption isotherms with dierent computation models permitted the determination of several parameters, useful for thorough characterization of the microstructures of the samples. The substituted samples preserved the hydrotalcite-type layered structure though their textural properties underwent important modi®cations. For copper substituted LDH, a relatively uniform porous structure with emphasized mesoporous characteristics emerged, the mesopores were enlarged and the speci®c surface area was decreased. On the contrary, for iron substituted LDH, the microporosity features developed, nonuniformity and constrictions in the porous structure was accentuated, the pore size was decreased and the speci®c surface area was increased. Microscopic morphology characteristics contributed to establish the textural properties of the samples. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Hydrotalcites; Textural properties; Mesoporosity; Microporosity; Microscopic morphology
1. Introduction Hydrotalcite is a naturally occurring anionic clay with the formula Mg6 Al2 (OH)16 CO3 4H2 O. It presents a positively charged brucite-like layers (Mg(OH)2 ) in which some of Mg2 are replaced by Al3 in the octahedral sites of hydroxide sheets. * Corresponding author. Address: Mathematical Seminary ``Al. Myller'', University Al. I. Cuza, Bd. Copou no. 11, Iasi 6600, Romania. E-mail address: [email protected] (G. Carja). 1 On leave from: Department of Physical Chemistry, Faculty of Industrial Chemistry, Technical University ``Ghe. Asachi'' Bd. D. Mangeron, Iasi 6600, Romania.
Interstitial layers formed by CO23 anions and water molecules compensate the positive charge resulting from this substitution [1]. Both magnesium and aluminum can be isomorphously substituted by other divalent or trivalent cations, and a wide range of compositions containing various combinations of M(II), M(III) and dierent anions An can be synthesized [2,3]. The resulting materials, known as layered double hydroxides (LDHs) or hydrotalcite-like anionic clays, have received considerable interests in recent years owing to their applications as catalysts, catalyst precursors, adsorbents and ion exchangers [4]. In de®ning or tailoring the catalytic and adsorption properties the crystallinity and textural
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 3 8 7 - 0
276
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
characteristics are of particular importance. For example, the shape and the size of pores, associated with the uniformity of the porous structure, can promote shape selectivity; a higher surface area will facilitate guest±host interactions; even mass transfer process depends on the porosity type. Despite this, many of the published studies concerning LDHs deal only brie¯y or not at all with their textural characteristics and with the possibility to tailor them for required applications [5± 10]. In order to get new information about the possibility to tailor textural properties of LDHs, the aim of this study is to investigate how these properties can be altered through direct synthesis when magnesium is partially substituted by other metals. In the present study, we synthesized MgCuAl and MgFeAl ternary LDHs. The eects of magnesium substitution by copper or iron on the microscopic morphology and the textural properties of the materials are reported. 2. Experimental 2.1. Synthesis of layered double hydroxides All the samples were synthesized by a coprecipitation method in low supersaturation conditions [11], at 313 K, under a bubbling constant ¯ow of nitrogen in the reaction medium and vigorous stirring. After an aging step, the precipitates were separated by centrifugation, washed extensively with warm deionized water until sodium free and dried under vacuum at 338 K. 2.1.1. MgAlLDH An aqueous solution (100 ml) of Mg(NO3 )2 6H2 O (0.03 mol)/Al(NO3 )3 9H2 O (0.01 mol) and an aqueous solution of NaOH/Na2 CO3 (CO23 = Al3 Mg2 0:67, HO =Al3 Mg2 2:25), were added dropwise together, in such a way that the pH remained at a constant value of 9.5. The resulting white precipitates were aged at 338 K for 24 h under stirring. 2.1.2. CuLDHs An aqueous solution (85 ml) of Mg(NO3 )2 6H2 O ((0.03-q) mol)/Al(NO3 )3 9H2 O (0.01 mol)/
Cu(NO3 )2 3H2 O (q mol, 0:005 6 q 6 0:01) and an aqueous solution of Na2 CO3 (1 M, 30 ml) were added dropwise together over a period of 2 h at a constant pH value of 8.2. The resulting light blue precipitates were aged at 305 K for 24 h under stirring. 2.1.3. FeLDHs An aqueous solution (100 ml) of Mg(NO3 )2 6H2 O ((0.03-r) mol)/Al(NO3 )3 9H2 O (0.01 mol)/ FeSO4 7H2 O (r mol, 0.005 6 r 6 0.01) and an aqueous solution of Na2 CO3 (1 M, 35 ml) were added together over a 2 h period at a constant pH of 8.5. The resulting orange precipitates were aged at 305 K for 24 h under stirring. 2.2. Characterization The chemical compositions of synthesized samples were determined by X-ray ¯uorescence (XRF) spectroscopy (Shimadzu XRF-1700 sequential XRF spectrometer). X-ray powder diraction (XRD) patterns were recorded on a Philips PW 1840 diffractometer using monochromatic CuKa radiation (k 0:154 nm), operating at 40 kV and 30 mA over a 2h range from 4° to 70°. IR spectra were recorded on a Shimadzu FTIR spectrometer under the following experimental conditions: 200 scans in the mid-IR range (500±4000 cm 1 ) using KBr (ratio 5/95 wt.%) pellets, and a resolution of 4.0 cm 1 . N2 adsorption±desorption isotherms were measured on a Coulter SA 3100 automated gas adsorption system. Prior to the measurements the samples were heated for 5 h under vacuum at 383 K in order to expel the interlayer water molecules. The BET speci®c surface area (SBET ) was calculated by using the standard Brunauer, Emmett and Teller method [12] on the basis of the adsorption data. Pore volume, micropore area and mesopore area were determined by using the t-plot method of De Boer [14]. Pore size distributions were calculated from the desorption branches of the isotherms using the Barret, Joyner and Halenda method and the corrected Kelvin equation [13,14]. All these models are considered to be fully applicable to mesoporous materials [15]. The IUPAC classi®cation of pores and isotherms were used in this study [16]. A Hitachi S-800 scanning
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
277
electron microscope (SEM) was utilized for the observation of microscopic morphologies. The accelerating voltage was 15 kV. Transmission electron microscopy (TEM) analysis was performed on a Hitachi instrument operating at 200 kV. The samples were prepared by dispersing powders in ethanol.
3. Results and discussion 3.1. Structural characteristics The composition of the synthesized ternary hydrotalcites can be fully de®ned by the formula: Mex Mg3 x Al, where Me is Cu or Fe (atomic ratio: Mg=Me=Al 3 x=x=1). Therefore we will use the notation xMeLDH to denote the composition of the samples and to de®ne the ternary sample name. The chemical compositions and some structural properties of the synthesized samples are presented in Table 1. The Mg=Me=Al atomic ratios determined by XRF are coincident with those of the starting mixed aqueous solutions within the experimental errors. The XRF results reveal that only magnesium (no aluminum) was substituted by iron during the synthesis procedure. The XRD patterns (Fig. 1) exhibit sharp and symmetric re¯ections for the (0 0 3), (0 0 6), (1 1 0) and (1 1 3) planes and broad symmetric peaks for the (1 0 2), (1 0 5) and (1 0 8) planes, which are characteristics of these materials [17]. The lattice parameters are calculated by indexing the peaks under a hexagonal crystal system, using a least squared method. The parameter a corresponds to the cation±cation distance within the brucite like layer while the parameter c is related to the total
Fig. 1. XRD patterns of: (a) MgAlLDH, (b) 0.5CuLDH, (c) CuLDH, (d) 0.5FeLDH and (e) FeLDH.
thickness of the brucite-like layer and the interlayer distance [1,18]. For the xCuLDH samples the parameter a has almost the same value as for the MgAlLDH. The result is due to the correspondence between the ionic radius of Cu2 hexacoor Shannon ionic radii [19]) and the dinate (0.73 A ionic radius of Mg2 hexacoordinate (0.72 A, Shannon ionic radii). The decrease of the parameter a for the xFeLDHs samples clearly indicates the incorporation of iron into the brucite-like
Table 1 Chemical composition and lattice parameters of MgAlLDH and MeLDHs Sample
Mg:Me:Al from XRF
Me (% mass) from XRF
XRD phase
Lattice parameters (A)
IFS (A)
a
c
d0 0 3
MgAlLDH 0.5CuLDH CuLDH 0.5FeLDH FeLDH
2.94:0:1 2.4:0.6:1 2.1:1:1.1 2.4:0.57:1 2:1.07:1
0 27 46 24 50
LDH LDH LDH LDH LDH
3.059 3.059 3.058 3.044 3.040
23.38 23.21 23.17 22.80 22.76
7.79 7.73 7.72 7.60 7.58
2.99 2.93 2.92 2.80 2.78
278
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
The layer, as the ionic radius of Fe2 is 0.61 A. decrease of the parameter c and the d0 0 3 (basal spacing calculated from the 0 0 3 re¯ection position in the XRD pattern [1,3]) can be attributed to the modi®ed electrostatic interactions between the layer and the interlayer network when another metal is introduced in the LDH layer [18]. The interlayer free spacing (IFS) values are calculated by subtracting the thickness of the LDH layer (4.8 [1]) from the calculated d0 0 3 spacing. The values A, for xCuLDHs and to decrease from 2.99 to 2.92 A 2.78 A for xFeLDHs, respectively. The decrease in IFS value suggests a distortion of the LDH network induced in the process of isomorphous substitution of magnesium by copper or iron. The distortion of LDH network is accentuated for the iron substituted samples. XRF results, together with the structural changes that are revealed by XRD, demonstrate the incorporation of iron and copper in the LDHs network. Fig. 2 presents the FTIR spectra of MgAlLDH, CuLDH and FeLDH. In the case of LDHs IR analysis is used to get information about the anions present in the interlayer between the bru-
cite-like sheets [1]. For all the samples IR spectra are quite similar though some dierences can be noticed in the intensity and the broadness of the bands. All the spectra exhibit an intense broad band between 3800 and 2700 cm 1 which comprises the vibrations of physisorbed water [11,20], vibrations of structural OH groups [20], characteristic valency vibrations of OH OH and/or characteristic stretching vibrations of CO23 OH in hydrotalcite [21]. The band corresponding to the vibration mode dOH appears between 1644± 1648 cm 1 and may be assigned to the adsorbed interlayer water [22]. An intense absorption band between 1370±1383 cm 1 is attributed to the m3 vibration mode of CO23 [23]. In the same range the m3 vibration mode of nitrate anion could appear if this still exists in the interlayer. The presence of the band between 1504±1537 cm 1 can be ascribed to a lowered carbonate symmetry that activated the vibration mode m1 of CO23 [24]. For FeLDH the band that appear at 1122 cm 1 is due to the m3 vibration mode of SO24 anion that still exists in the interlayer; a very weak band corresponding to the m4 vibration mode of this anion also appears at 650 cm 1 [25]. The bands due to m4 and respective m2 vibration modes of CO23 anion appear in all the spectra between 667±680 and 870±880 cm 1 , respectively [24]. We can remark some differences in the intensities of these bands between MgAlLDH and the substituted samples; this fact can be ascribed to a dierent orientation of CO23 in the interlayer as a consequence of dierent electrostatic forces that act in the network of substituted samples. For all the spectra weak bands appear at 635, 770 and 939 cm 1 corresponding to the hydrotalcite lattice vibrations [26,27]. 3.2. Textural characteristics
Fig. 2. FTIR spectra of: (a) MgAlLDH, (b) CuLDH and (c) FeLDH.
The nitrogen adsorption isotherms of MgAlLDH, CuLDH and FeLDH are shown in Fig. 3(a)±(c). Their dierent appearances suggest the modi®cations of the porosity characteristics when copper or iron is substituted for magnesium in the LDH network. For MgAlLDH (Fig. 3(a)) we observe a type IV isotherm with a broad type H3 hysteresis loop, in the middle range of relative
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
279
Fig. 3. N2 adsorption±desorption isotherms of (a) MgAlLDH, (b) CuLDH and (c) FeLDH.
pressure. At low relative pressure the prevailing process is the formation of a monolayer while a multilayer adsorption takes place at a high relative pressure. For the values of relative pressure higher than 0.8 condensation takes place giving a sharp adsorption volume increase. This behavior indicates that this sample has a mesoporous character. Hysteresis loop type shows that aggregates of plate-like particles forming enough nonuniform slit shaped pores [16] are present in this sample. For CuLDH (Fig. 3(b)) the shape of the curves is typical for a type IV adsorption isotherm and a type H1 hysteresis. The isotherm form and the
narrow hysteresis loop, with almost parallel adsorption and desorption branches, indicate that cylindrical pores with mesopore size range and almost regular geometry, more uniform in size and/or shape than those present in MgAlLDH, exist in the copper substituted sample. For FeLDH (Fig. 3(c)) the adsorption isotherm is very similar to that found for porous zircon [28] and presents particular characteristics. That is: · for ps =p0 < 0:2, multilayer adsorption is present (see also Table 2) revealing microporous characteristics for this sample;
280
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
Table 2 Characteristics of N2 adsorption/desorption MgAlLDH, CuLDH and FeLDH Sample
nav ps =p0 < 0:25
ps =p0 1
MgAlLDH CuLDH FeLDH
1 1.30 2.4
15 21.7 6
a
data
VN2 (ml/g)
dp a (nm)
460 432 189
12.8 20 4.1
for
dp 4VN2 =SBET .
· for 0:2 < ps =p0 < 0:8, the isotherm shape reveals the presence of mesopores with a wide size range; · for ps =p0 > 0:8, it can be assimilated with a type II isotherm, characteristic for adsorption in pores with sizes larger than 50 nm. The irregular form of type H3 hysteresis loop also indicates the presence of nonuniformity in pores size and/or shape. These features allow us to conclude that, for FeLDH, both microporous and mesoporous properties contribute to establish the characteristics of the microstructure. Pore size distribution (PSD) is further used to check and to complete the previous results. The PSD curves, Fig. 4, point out important modi®cations in the pore size range of the samples. For FeLDH a sharp and high peak with a maximum
Fig. 4. Pore distribution in MgAlLDH, CuLDH and FeLDH.
around 3.5 nm appears. This shows that small pores (micropores and very small mesopores) establish the porosity characteristics of this material. Large mesopores and even some macropores also developed and the porous structure of FeLDH has a pronounced nonuniformity. The PSD curve of MgAlLDH is broader and is shifted to a larger size with two maximums at 9.2 and 12.3 nm. Pores with sizes belonging to an entire range characteristic for mesoporosity appear in this case. The PSD curve of the copper substituted sample with two maximums at 27 and 37 nm shows clearly the enlargement of the pore size. All the pores are larger than 5 nm and we observe no mesopores with sizes larger than 45 nm. Consequently neither micropores nor macropores are detected, revealing that the uniformity of the porous structure is accentuated. Therefore, the PSD results complete the analysis of the isotherms and demonstrate that for the copper containing hydrotalcite-like sample the pore size increased by nearly 15 nm though for the iron substituted sample the mesopores size decreased signi®cantly and some very large mesopores are also detected. The contribution of pores, as a function of their size range, to the total pore volume was calculated using the PSD results. It can be seen (Fig. 5) that the pores belonging to the range of 5±20 nm contribute by nearly 60% to the total pore volume of MgAlLDH, while the contribution due to the pores with sizes between 20 and 40 nm is only 22%. For CuLDH, almost 80% of the pore volume is due to the pores belonging to the range of 20±40 nm. For FeLDH it is not possible to see a distinct major contribution as pores belonging to the range of 0±5 nm contribute by around 22% to the pore volume and pores with sizes larger than 40 nm are present with a contribution equal by nearly 27%. The result also emphasizes that CuLDH possesses a more uniform porous structure while nonuniformity of the porous structure is accentuated for FeLDH. Other aspects of the alteration of the textural characteristics of LDHs are revealed by the average number of nitrogen layers (nav ) formed during N2 adsorption. The nav values are calculated by using the adsorbed monolayer volume and the
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
Fig. 5. Contribution of pores, as a function of their size range, to the total pore volume in MgAlLDH, CuLDH and FeLDH.
values of total adsorbed volume (VN2 ). The results are presented in Table 2. At low relative pressure, (ps =p0 < 0:25), a monolayer is formed for both MgAlLDH and CuLDH, though a multilayer adsorption takes place in the case of FeLDH. The number of nitrogen layers is signi®cantly higher for ps =p0 values close to unity, being 6 for FeLDH, 15 for MgAlLDH and nearly 22 for CuLDH. The volume of nitrogen adsorbed at the highest value of relative pressure, (VN2 ), decreases for FeLDH. These results also con®rm the accentuation of the mesoporous character of CuLDH (as multilayer adsorption is stressed at higher values of relative pressure) and the development of the microporosity features in FeLDH (as multilayer adsorption take place at low values of relative pressure; nav is lower for values of relative pressure close to unity; volume of nitrogen adsorbed at the highest value ps =p0 decreases). The estimation of the average pore diameter (dp ), by the classical Gurvitsch rule [13], is given in the last column of Table 2. According to the Gurvitsch rule, average pore diameter (which in fact does not mean the eective pore diameter because the contribution of the external surface in
281
the total surface is also included [29]) decreases from 20 nm for CuLDH to 12.8 nm for MgLDH and becomes equal to 4.1 nm for FeLDH. The values are enough close to those obtained from PSD. To reveal another estimation for the modi®cations of the textural properties we calculated the values of the total pore volume (Vp ), and the extent of micropore area (%lpA) and mesopore area (%mesopA) in the total t-plot area (see Table 3). The values of BET surface area are also given in the second column of the table. It can be seen that the FeLDHs develop higher BET surface areas in comparison to MgAlLDH while a decline of BET surface areas is characteristic for the copper substituted samples. The extent of area formed by mesopores in the total t-plot area is nearly 99% for CuLDH but decreases to 86% for MgAlLDH while the value is only 65% for FeLDH. The contribution of mesopores and micropores to the total surface area indicates that the increase in the surface area of iron containing LDH is due to the development of microporosity characteristics; the decrease of the surface area of copper containing LDH is due to the accentuation of mesoporosity characteristics. The indirect proportionality between the pore volume and the BET surface area suggests that the porosity properties are in¯uenced by the microscopic morphology of the samples. All the computation models applied to the adsorption isotherm data gave similar results that point out important transformations of the textural properties for copper and iron substituted LDHs. With copper substitution the mesopore size increases and a more uniform porous structure appears. For iron containing LDH the pore size
Table 3 BET surface area and porosity characteristics obtained from N2 adsorption data using the t-plot method of De Boer Sample
BET area (m2 /g)
Vp (ml/g)
%lpA
%mesopA
MgAlLDH 0.5CuLDH CuLDH 0.5FeLDH FeLDH
132 117 86 162 181
0.640 0.647 0.661 0.411 0.277
14 11 1 22 35
86 89 99 79 65
282
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
decreases (85% of the pores are smaller than 5 nm) and the nonuniformity of the porous structure was accentuated. Hence, micropores, very small mesopores and also some large pores coexist. A large decrease of the BET surface area is characteristic for the copper substituted samples though the pore volume is enhanced. Contrary for iron substituted LDHs the BET surface area values increase while the pore volume and the adsorption capacity show lower values. These transformations are the result of the enhancement of mesoporosity characteristics for the copper substituted samples and the
development of microporosity features for the iron substituted samples. It is quite dicult to explain the transformations that appear in the porosity features of LDHs when a part of magnesium is substituted by copper and iron. The LDH layered structure with an inter cannot be relayer free space smaller than 3 A sponsible for any mesoporosity characteristics. The mesoporous structure of LDHs may arise from interparticle space hence the characteristics of mesopores could be de®ned by the particle size and shape and also by the particles interconnection
Fig. 6. (a) SEM micrograph of MgAlLDH sample. (b) SEM micrograph of CuLDH sample. (c) SEM micrograph of FeLDH sample.
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
patterns. In this view the changes in the porosity features are the result of altering the microscopic morphology characteristics. The SEM images (Fig. 6(a)±(c)) show that MgAlLDH and the substituted samples are highly crystalline. Intersected, nearly hexagonal-shaped particles, interconnected with each other, emerged for all the samples. Notwithstanding this, a different aspect emerged in the appearances. The differences are mainly due to the dierent mode of interconnection and agglomeration of the particles in ternary LDHs and also to the modi®cation of the particle size. Interparticle space is larger for CuLDH (Fig. 6(b)) than for MgAlLDH (Fig. 6(a)) or FeLDH (Fig. 6(c)). Wider interparticle cavities developed in the iron-substituted sample; these can be considered as the origins of the large pores.
283
The average value of particle size is 150 nm for CuLDH, 130 nm for MgAlLDH and 100 nm for FeLDH. The TEM image of the Cu substituted sample (Fig. 7) shows the presence of aggregates of well-de®ned particles, characteristics for hydrotalcite-like materials [30,31]. A careful inspection of the image reveals that the particles themselves do not present any mesoporosity. The SEM and TEM results further support the idea of the connection between the textural properties and the microscopic morphology of the samples. The results reported previously [10,30] claimed that the textural properties of hydrotalcite-like materials are strongly in¯uenced by the synthesis conditions. We assume that the dierent pH values, the distinct nature and electronegativities of the ions present in the synthesis medium, and the dierent concentrations of the starting solutions could be important factors to control the microscopic morphology and to determine the changes of the textural properties.
4. Conclusions The current study demonstrated that the microstructures of ternary layered double hydroxides are signi®cantly transformed when copper or iron substituted magnesium in the LDH layer. As a general trend copper substituted LDH presented a mesoporous structure with a higher uniformity than MgAlLDH, while microporous characteristics developed in iron substituted LDH and a microporous±mesoporous structure with a more accentuated nonuniformity emerged. The possibility of altering the textural properties of LDHs when a new metal is introduced in the layer may oer new perspectives in tailoring the textural characteristics of these materials for potential applications.
Acknowledgements
Fig. 7. Characteristic TEM image of Cu substituted sample.
G.C. acknowledges a research grant from Ministry of Education, Science, Sports and Culture of Japan.
284
G. Carja et al. / Microporous and Mesoporous Materials 47 (2001) 275±284
References [1] F. Cavani, F. Tri®ro, A. Vaccari, Catal. Today 11 (1991) 173 and references therein. [2] W. Jones, M. Chibwe, in: I.V. Mitchel (Ed.), Pillared Layer Structures Current Trends and Applications, Elsevier, London, 1991, p. 67. [3] F. Tri®ro, A. Vaccari, in: J.L. Atwood, J.E. Davies, D.D. Mac Nicol, F.Vogtle (Eds.), Comprehensive Supramolecular Chemistry, vol. 7, Pergamon Press, Oxford, 1996, p. 251. [4] A. Vaccari, Catal. Today 41 (1998) 53. [5] H. Nijs, M. De Bock, E.F. Vansant, Micropor. Mesopor. Mater. 30 (1999) 245. [6] K. Chibwe, W. Jones, Chem. Mater. 1 (1989) 489. [7] J. Don Wang, G. Serrette, Y. Tian, A. Clear®eld, Appl. Clay Sci. 10 (1995) 105. [8] T. Kwon, T. Pinnavaia, Chem. Mater. 1 (1989) 381. [9] A. Corma, V. Fornes, F. Rey, J. Catal. 151 (1995) 50. [10] F. Malherbe, C. Forano, J.P. Besse, Micropor. Mater. 10 (1997) 67. [11] S. Miyata, Clays Clay Miner. 31 (1983) 305. [12] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309. [13] M. Jaroniec, M. Kruk, A. Sayari, in: L. Bonneviot, F. Beland, C. Danumah, S. Giasson, S. Kaliaguine (Eds.), Studies in Surface Science and Catalysis, vol. 117, Elsevier, Amsterdam, 1994, p. 325. [14] S.J. Gregg, K.S.W. Sing, Adsorption Surface Area and Porosity, second ed., Academic Press, London, 1982. [15] M. Kruk, M. Jaroniek, A. Sayari, Micropor. Mesopor. Mater. 27 (1999) 217.
[16] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603. [17] A. Vaccari, M. Gazzano, in: G. Poncelet, J. Martens, B. Delmon, P.A. Jacobs, P. Grange (Eds.), Preparation of Catalysts VI, Studies in Surface Science and Catalysis, vol. 91, Elsevier, Amsterdam, 1995, p. 309. [18] S. Velu, N. Shah, T.M. Jyothi, S. Sivasanker, Micropor. Mesopor. Mater. 33 (1999) 61. [19] R.D. Shanon, Acta Crystallogr. A 32 (1976) 751. [20] M.J.H. Hernandez, M.A. Ulibarri, J.L. Rendon, C.J. Serna, Phys. Chem. Miner. 12 (1985) 34. [21] D. Roy, R. Roy, E. Osborn, Am. J. Sci. 251 (1973) 337. [22] R. Allmann, Chimia 24 (1970) 99. [23] D.L. Bish, G.W. Brindley, Am. Miner. 62 (1977) 458. [24] S. Miyata, A. Okada, Clays Clay Miner. 44 (1977) 14. [25] S. Kannan, C.S. Swamy in: G. Poncelet, J. Martens, B. Delmon, P.A. Jacobs, P. Grange (Eds.), Preparation of Catalysts VI, Studies in Surface Science and Catalysis, vol. 91, Elsevier, Amsterdam, 1995, p. 903. [26] S. Miyata, Clays Clay Miner. 23 (1975) 369. [27] M.A. Aramendia, Y. Aviles, J.A. Benitez, V. Borau, C. Jimenez, J.M. Marinas, J.R. Ruis, F.J. Urbano, Micropor. Mesopor. Mater. 29 (1999) 319. [28] R. Valero, B. Durand, J.L. Guth, T. Chopin, Micropor. Mesopor. Mater. 29 (1999) 311. [29] A. Gervasini, Appl. Catal. A General 180 (1999) 71. [30] F. Prinetto, G. Ghiotti, P. Gran, D. Tichit, Micropor. Mesopor. Mater. 39 (2000) 229. [31] O. Clause, B. Rebours, E. Merlen, F. Tri®ro, A. Vaccari, J. Catal. 133 (1992) 231.