Preparation of Layered Double Hydroxides

PREPARATION OF L A Y E R E D D O U B L E H Y D R O X I D E S

EIJI KANEZAKI Department of Chemical Science and Technology, Faculty of Engineering, The University of Tokushima, 2-1 Minamijosanjima, Tokushima 770-8506- JAPAN E-mail: [email protected]

Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors)

9 2004 Elsevier Ltd. All rights reserved.

346

E. Kanezaki

I - Preparation of layered double hydroxide with interlayer carbonate.

Layered double hydroxides (LDH) have a general formula of [MaM'b(OH)2(a§ abbreviated as MM'/A-LDH where M and M' are a divalent and a trivalent metal cation in the layer of the hydroxide, respectively, and the excess positive charge of the layer is compensated by the negative one on the interlayer monoanion A, which can be replaced by a di-, tri- or other multivalent anion or in some cases a mixture of them, thus LDH is called an anionic clay mineral. LDH is a mimic of naturally occurring hydrotalcite (MgA1/CO3-LDH), has a similar layered structure, which is called hereafter the hydrotalcite-like layered structure, and is also called the hydrotalcite-like compound. Over the last two decades, interest has been growing in the availability for the intercalation of various organic anions having flexible or rigid molecular frameworks into LDH not only from the scientific but also from industrial viewpoints [ 1]. LDH is usually prepared in ordinary conditions of temperature, pressure and so on as a precipitate from solutions by means of environmentally benign methods. Therefore, it is easy to synthesize the LDH which has the desired anion(s) at the interlayer region when we carefully select the combination of the metals and the organic compound. Both the facility in the synthesis and the potential variety in the combination of the components promise the usefulness of LDHs for developing new-type of materials. A lot of studies have been done on the synthesis and characterization of LDH with the variety of divalent (Zn 2§ Ni 2§ Fe 2§ Co 2§ etc.) or trivalent metal cations (Fe 3+, Cr 3+, Sc 3+, etc.) in the layers of LDH. It is outstanding that many kinds of anion, indifferent to mono- or multivalent, are intercalated at the interlayer of LDH because the volume of the interlayer gallery is variable and the interlayer distance between the adjacent layers is enhanced or reduced in proportion to the molecular size of the anion. The interlayer anion of LDH is organic, inorganic or coordination compounds which could be introduced into LDH by means of anion-exchange, coprecipitation or rehydration the last of which is very characteristic of the intercalation of LDH and is described below in detail. The rehydration method originates in a unique nature of LDH; the hydrotalcite-like layered structure is reproduced when the precursor LDH is calcined at high temperature to collapse the layered structure followed by swelling in aqueous solution containing anions to be intercalated. Collapsing the layered structure in the precursor LDH results in only amorphous phases of the mixed metal oxides of MO and M'203 in the calcined solid, all of which are absent in the X-ray powder diffraction patterns (XRD). When the intercalation of anions occurs at the interlayer gallery region of LDH, the XRD pattern usually changes drastically and the explicit layered structure appears again; the magnitude of the basal spacing, calculated from the lowest 20 angle of diffraction lines, indicates the molecular size of the intercalated anion perpendicular to the normal axis of the stacking layers. Therefore, in the study of the solid-state chemistry of LDH, the XRD measurement before and after the intercalation of a particular anion is essential with rare exceptions. Furthermore, the inspection of the thermal change of the XRD pattern together with that of differential thermal analysis/thermal gravimetry (DTA/TG) allows us the fruitful discussion since the collapse of the layered structure, which results from the dehydroxylation of the double hydroxide, takes place at a moderate temperature. However, thermally metastable solid phases are sometimes missed in conventional measurements in which samples are heated and cooled outside the sample chambers. Therefore, high temperature in situ measurements are necessary in order to investigate the thermal change of the solids

Preparation of Layered Double Hydroxides

347

precisely. In particular, the in situ high temperature XRD measurement (in situ HTXRD) is favorable for the study of LDH since the layered structure recovers very soon in an ordinary atmosphere owing to the adsorption of water and CO2 in air. Results of these measurements are illustrated below for two LDH compounds, which have the interlayer carbonate and are frequently used as the precursor for the intercalation of the particular anion. I.I - Mg and AI layered double hydroxide with interlayer carbonate Figure 1 illustrates a general view of the in situ HTXRD patterns of MgA1/CO3-LDH in the temperature range from 30~ to 1000~ There are three regions of temperature (T), which have the common HTXRD pattern; 30_
E. Kanezaki

348

thickness of one layer ofthe double hydroxides (4.8 A) and the interlayer distance (3.0 A) [ 1]. The interlayer distance of this layered structure is determined by the dimension of the interlayer carbonate dianions which are located with their molecular plane parallel to the internal surface of the layer [ 1,2,4]. On the other hand, the interlayer distance of Phase II is calculated 1.8 A with the same assumption for the thickness of one double hydroxides layer as that of the hydrotalcite-like layered structure. The smaller value of the interlayer distance of Phase II than that of the hydrotalcite-like layered structure suggests strongly that the interlayer species are different from each other between these two solid phases.

,

,

0 20

40

,

~ 20 / deg.

' 60

Figure 1 - In situ HTXRD diffraction patterns of MgAI/CO3-LDH with the atomic ratio Mg/AI=2 between 30~176 temperatures are 1000~ (top), 900~ 800~ 700~ 600~ 500~ 400~ 380~ 360~ 340~ 320~ 300~ 280~ 260~ 240~ 220~ 200~ 180~ 160~ 140~ 120~ 100~ and 30~ (bottom). Reprinted from [40] with permissionfrom Elsevier. Table 1 - Indexing of the in situ HTXRD pattern of Phase II produced by calcination of MgAI/CO3-LDH (Mg/AI=3) at 300~ hexagonal lattice with ao=3.070 A Co=6.592 A is assumed. Reprinted from [40] with permission by Elsevier.

20/degree 13.40 26.96 33.86 36.40 43.74 60.24 61.90

observed I/Io 100 8 10 20 10 22 10

d/A 6.592 3.304 2.645 2.466 2.068 1.535 1.498

Calculated indexing d/A (001) 6.592 (002) 3.296 (100) 2.659 (101) 2.466 (102) 2.069 (110) 1.535 (111) 1.495

,

.

Figure 2 gives the time course of the XRD pattern of Phase II, which is observed after the calcination of Mg/A1/CO3-LDH at 300~ in an electric furnace. Though the pattern does not change at one day after the calcination, the intensity of the Phase II diffraction decreases after 8 days and it becomes weak significantly at 56 days

349

Preparation of Layered Double Hydroxides

after the preparation when the calcined sample is placed in an air-tight silica gel desiccator after preparation; (see lines indicated by arrows in Figure 2). In contrast to the decrease of the Phase II diffraction, Phase I appears weakly after 8 days and becomes distinct at 56 days after the preparation. The same XRD pattern in Figure 2 (d) is also observed at three months after the calcination when the sample is still placed in the same desiccator. These results indicate clearly that Phase II is metastable and goes to an XRD silent amorphous phase slowly and that a small part of the amorphous phase goes Phase I in a dry condition. When the Phase II sample powder is placed in a humid atmosphere, the change is accelerated and Phase I is predominant in the XRD pattem after a few days. These results indicate that two processes are present in the change from (a) to (d) in Figure 2; first, the metastable Phase II degrades to an amorphous phase in a dry condition. Second, a small quantity of Phase I is formed from this amorphous phase owing to the contact to trace quantity of H20 (and probably of CO2) in a desiccator. Keeping the unchanged XRD pattern in Figure 2 (d) for three months after the sample preparation above is elucidated by the lack of enough H20 (and probably of CO2), which are necessary to return to Phase I, in the desiccator. A direct phase transition from Phase II to Phase I is doubtful. Remembering that the total structure of the layer is determined by the chemical status of the magnesium ions in the layer, it is concluded that the brucite-like layered structure having the octahedral unit [Mg(OH)6] is still maintained in Phase II since MgO is not observed in the temperature range where Phase II is alive (180<_T_<380~ in Figure 1). Therefore, the thermal shrinkage of the basal spacing in Phase II is not due to the change in the layer but due to the change in the interlayer species, which is further discussed below in the result of FT-IR.

Phas~L I yl~ Phase II = ID

•

20

40

20 / deg.

60

Figure 2- X-ray diffraction pattern of Phase II; (a) immediately, (b) I day, (c) 8 days and (d) 56 days after sample preparation, positions of the most intense line of Phase I and that of Phase H in Figure 1 are indicated by arrows. Reprinted from [39] with permission by Kluwer Academic~Plenum Publishers. Differential thermal MgA1/CO3-LDH are illustrated in observed at 196~ (endotherm A) throughout the temperature range

analysis/thermal gravimetry (DTA/TG) of Figure 3. In the DTA curve, two endotherms are and at 379~ (endotherm B). The total mass loss agrees well with the summation of all volatile

350

E. Kanezaki

components (H20 and CO2) in the chemical formula. It is noted that temperatures of these two endotherms correspond to those of formation and degradation of Phase II; Phase II is formed at the temperature of endotherm A and is degraded at the temperature of endotherm B. Since the indexing of the pattern of Phase II in Table 1 suggests strongly that the layered structure is still maintained in this solid, it is reasonable to as-sume that the mass loss at endotherm A in the TG curve is due to the elimination of interlayer carbonate and water in the solid sample. Interlayer carbonate ((CO32-)inter.) in MgA1/CO3-LDH is thermally oxidized by a nearby water molecule and produces both volatile CO/and interlayer hydroxyl anion ((OH)inter.) as in eq. (1) [5].

(CO32")inter.+ H 2 0

~

2(OH')inter.+ CO2 T

(Eq. 1)

The calculated mass loss 16.4% at endotherm A by using the chemical formula is close to the observed one 17% at the endotherm A in the TG curve of Figure 3 thus supporting the elimination of species in eq. (1). Namely the hydroxyl anions are produced at the interlayer region of LDH when Phase II is formed at the temperature of endotherm A. Less bulky nature of this anion than a carbonate explains the reduction of the interlayer distance in XRD patterns from 3.0 A (hydrotalcite-like layered structure) to 1.8 _A(Phase II). This reduction may be related to the migration of AI 3§ into the interlayer, which has been proposed recently [6]. In the rehydration method, the LDH with interlayer carbonate is used as the precursor and is calcined at the temperature which is very important because it is fatal to the successful reconstruction of the hydrotalcite-like layered structure in the rehydration method. In the calcination of the precursor LDH, the temperature should be higher than that of the layer collapse but at the same time lower than that of the the spinel phase formation because this solid phase is stable and is hard to be converted to LDH in water. Namely, for MgA1/CO3-LDH above, the temperature is usually set between 400~ and 700~ since the former temperature corresponds to the collapse of the layer of the double hydroxides and the latter to the beginning of the spinel formation. In X-ray photoelectron (XP) spectra of Mg/AI/CO3-LDH (not shown), several photoelectron peaks appear, which originate in core level electrons emitted from Mg, A1, C and O in the solid compound. Since A13§ ions occupy the octahedral Mg 2§ sites isomorphously, the coordination number of A13§ in the layer is six (Oh A1). This is verified in the 27A1 MAS NMR spectra obtained by a Brucker Avance300 solid state NMR spectrometer of the University of Tokushima with the reference of Al(NO3)3 aqueous solution (0.3 moldm-3; 5 = -0.1 ppm) after the data accumulation of 64-128 times. Sample powder was put in a capped ZrO-rotor (7 mm or 4 mm in diameter) spinning at the frequency in the range 4.0-13 kHz in order to avoid spinning side bands. Figure 4 illustrates 27A1 MAS NMR spectra of Mg/A1/CO3-LDH at room temperature for two spinning frequencies in order to eliminate the spinning side bands. Indifferent to the spinning frequency, only one absorption maximum corresponding to the change in the angular momentum of the nuclear spin in 27A1 Am/=l is observed at ~5=8.39 ppm with good S/N ratio. It has been reported that the chemical shift of this nucleus is 8.1 ppm for Mg/A1/(CO3+NO3)-LDH [7], 9.2 ppm for Mg/A1/B4Os(OH)4-LDH [8], 14.5 ppm for Zn/A1/CO3-LDH [9] and the same value for Cd/AgNO3-LDH [ 10]. All of these values including the chemical shift in this study fall in the range of the aluminum nucleus having the octahedral hexa-coordination. It has been also reported that the absorption due to the

351

Preparation of Layered Double Hydroxides

nucleus having the tetrahedral four-coordination (Td A1) is weakly observed at about 5=80 ppm on elevating the sample temperature in situ at 100~ in 27A1MAS NMR and that the intensity of the absorption increases (but does not predominate) with the increase of the temperature up to 400~ [7]. Caution should be paid on eliminating spinning side bands by means of changing the spinning frequency. TGA

DTA exo.

endo. 20

O ra~ t~

40

200

600

Temp./I~C

1000

Figure 3 - DTA/TG curves of MgAI/COs-LDH (Mg/AI-3). Reprinted from [39] with permission by Kluwer Academic~Plenum Publishers.

i

I

400

I

I

i

200

i

0

i

I

-200

(B)

4;0

.

. 200 .

.

.

0

-2;0

'

Chemical Shitt/ ppm

Figure 4 - 2ZAlMAS NMR spectra of MgAI/COs-LDH (Mg/Al=2. 6) at spinning frequency 4kHz (A) and 7kHz (B), maxima with an asterisk are spinning sidebands. [49]

~ ~-

~o~

t~

0

o

-I~ 0 0 0

_

0 0

;>

J

t~

Transmittance

S

r

Preparation of Layered Double Hydroxides

353

Figure 5 (A) shows the FT-IR spectrum of MgA1/CO3-LDH at room temperature, which has been reported by many authors [11-15]. A broad band which is located at 3500-3600 cm 1 is the O-H stretching vibration both of the hydroxide group in the layer and of the interlayer water molecules. The band located at 1652 cm 1 is the O-H bending of the interlayer water. Figures 5 (B)-(D) are FT-IR spectra of this compound at 210~ at 400~ and at 500~ respectively. On elevating the temperature, the transmittance of the sample disk becomes low at the high energy side in the spectra because the probe light is scattered increasingly owing to the proceeding of the microcrystallization of KBr. At 210~ (Figure 5(B)), two major differences are observed in comparison with the spectrum at room temperature; first, the band due to the O-H bending of the interlayer water disappears, which band is originally observed at 1652 cm -1 at room temperature in Figure 5(A). Since this band is not observed above this temperature, it is concluded that the interlayer water is lost up to 210~ In other words, the interlayer water eliminates from the sample exhaustively at endotherm A in Figure 3. Comparing the water content in this solid sample with the mass loss in Figure 3, the mass loss at endotherm A has the other origin(s) than the interlayer water. Secondly, the intensity at around 1500 cm -~ increases because the v3 vibration of CO32 is doubly split; one component is located at 1414 cm 1 and the other one at 1505 cm ~ the latter of which components is overlapped by the weak band due to the water in air observed in the spectrum at room temperature. The splitting v3 vibration of the carbonate has been reported as the result of the coordination of oxygen atoms in the carbonate to a metal cation [ 16]. This is the case which is observed in Figure 5(B); the oxygen atom of carbonate makes chemical bond directly to a metal atom in the layer. The carbonate still maintains the trigonal symmetry at 210~ since the totally symmetric v~ vibration, which is forbidden in the D3h symmetry [17], is not observed in Figure 5(B). Since the metastable Phase II is alive between 180~ and 380~ in this solid (Mg/AI=2.6), as is described in the result of XRD, it is concluded that the spectrum in Figure 5(B) is of Phase II. Constantino and Pinnavaia have reported the thermal shrinkage (1.85 A) of the basal spacing at 250~ concluding that the interlayer sulfonate changes thermally to the "grafting sulfonate" at the interlayer region on elevating the temperature [ 18]. Since the shrinkage (1.29 A) in this study from Phase I to Phase II is in the same order of magnitude it is inevitable to assume that the interlayer carbonate is grafting in the non-aqueous circumstance of the interlayer gallery region at 210~ in this study. This conclusion seems reasonable because it is difficult to suppose that the anion would be located solely at the interlayer gallery region at 210~ without cations. The grafting carbonate would be unstable since it has three equivalent C-O bonds directed to metal cations in the adjacent layers and at the same time it holds the trigonal symmetry. This unstable nature may explains the metastable profile of Phase II in the HTXRD pattern in Figure 1. The FT-IR spectrum at 400~ in Figure 5(C) shows four major differences; first, the intensity of the O-H stretching at 3500-3600 cm "1 decreases, secondly, the sharp band at 1645 cm 1 appears, thirdly, the intensity of the band at around 1500 cm 1 decreases, and in fourth, the band at 1083 cm "1 newly appears. The decrease in the intensity of the O-H stretching is elucidated by the thermal event in which the layered structure of Phase II collapses at 380~ owing to the thermal dehydroxylation of the double hydroxide. The grafting carbonate, whichis observed at210~ decomposes on elevating the temperature up to 400~ and a part of the carbonate is oxidized to COz. The band at 1645 cm ~ is

354

E. Kanezaki

assigned to the C-O stretching vibration (V3) of the coordinating CO 2 molecule to the metal cation [ 16]. It is received that this vibration is observed at 2349 cm "1 for the free CO2 molecule and shifts drastically to the lower energy side as 1600-1800 cm ~ on the coordination to metal. Though the Vl vibration of the coordinating CO2 locates in the region 1100-1300 cm "1 [ 16], it is hidden behind the steep increase of the absorption at 1414 cm "1 in Figure 5(C). It has been reported that the brisk CO2 evolution is observed at 300-400~ [10], which CO2 molecule presumably participates in the coordinating CO2. The formation of the coordinating CO2 is important because this species is expected as the useful C 1 source in the catalytic reaction and the formation of the coordination bond to metals could open the possibility to go further reaction. Both of the latter two in the four differences described before in the spectrum at 400~ (Figure 5(C)) are due to another carbon species which is also produced after the grafting carbonate decomposes thermally. The band at 1083 cm "1 has been observed in the IR spectrum of CaCO3 (aragonite) as the Vl vibration of the carbonate [ 16]. It has been reported that the Vl vibration of the carbonate is allowed in the trans-Cs symmetry of the O-O-C=O structure in CaCO3 (aragonite) although it is for-bidden in the planar trigonal structure of the dianion as mentioned before [17]. The band at 1410 cm -1 is the v3 vibration of the carbonate, which has the aragonite-like symmetry. Although it has been reported that the v3 band of the carbonate is doubly split in the trans-Cs symmetry [ 16,17], the counterpart may be hidden behind the band cluster around 1500 cm -1 in Figure 5(C). The sharp band located at 856 cm ~ and a shoulder at 878 cm ~ are of the doubly split v2 vibration of the carbonate and the weak band at 707 cm ~ and the band at 616 cm ~ are of the doubly split v4 vibration of the dianion. Thus, there are two carbon species at 400~ the CO2 complex and the aragonite-like carbonate. In the IR spectrum at 500~ (Figure 5(D)), the intensity of all bands decreases. A broad band is still observed at 3500-3600 cm "1 which is probably due to the OH stretching of amorphous aluminum hydroxide in the form of Al(OH)3 or A10(OH). Three bands which are located at 1470 cm 1 (broad), 852 cm -1 and at 671 cm "1 are the v3, v2 and the v4 vibration bands of the carbonate, respectively. The CO2 complex and the aragonite-like carbonate both of which are observed at 400~ decompose on elevating the temperature to 500~ another carbonate having the CaCO3 (calcite)-like structure with the forbidden v~ vibration of the carbonate appears in the IR spectrum in Figure 5(D) [ 16]. Since aragonite is metastable in an ordinary atmosphere and thermally transforms to calcite, the assignment above is reasonable. A sharp band located at 882 cm ~ is assigned [ 16] to the Mg-O vibration band in MgO which is produced after the collapse of Phase II. Since this solid is also observed in HTXRD pattems above 400~ in Figure 1, it is likely that this Mg-O vibration overlaps the band at 878 cm -~ in the spectrum at 400~ (Figure 5(C)). The location of the v3 vibration of the carbonate at 500~ (1470 cm "1) shifts to the higher energy side than that at room temperature (1394 cml). In spite of the uncertainty of the band location due to the broadness in Figure 5(D), this shift is significant. In other words, the interlayer carbonate has smaller force constant in this vibration mode at room temperature than the calcite-like carbonate does at 500~ probably owing to the presence of the hydrogen bonds in the interlayer gallery region at the room temperature as previously reported [13]. Both the carbonate and the hydroxylate which are left in the solid sample at 500~ would eliminate from the sample as CO2 and H20, respectively, on further increase of the temperature although we cannot follow the spectral change above 500~ from the instrumental reason. The elimination of these species would elucidate the

Preparation of Layered Double Hydroxides

355

gentle mass loss observed in the TG curve above this temperature in Figure 3. In IR spectra of the other two solids with different atomic ratios Mg/A1 in the layer, the parallel discussion is done. In conclusion, thermal change in the layered structure of MgA1/CO3-LDH associated with elevating temperature below 1000~ is understood by means of in situ HTXRD, DTA/TG analysis and the high temperature in situ FT-IR measurements in which the thermal change in the chemical status of interlayer species is studied in detail. These results indicate that the thermal shrinkage of the basal spacing in Phase II is due to the thermal oxidation of the anion mainly and the residual carbonate makes direct bonding to metal atoms in the layers. 1.2 - Zn and AI layered double hydroxide with interlayer carbonate

When the element of divalent metal cation is changed, the different results are observed in these measurements. Figure 6 shows a general view of in situ HTXRD of ZnA1/CO3-LDH on increasing the temperature from 30~ and 1000~ The hydrotalcite-like layered structure is observed up to the temperature of 160~ with the intense diffraction line at 20 = 11.3 A (d=7.8 A) associated with some discrete lines, which is very similar to that of MgA1/CO3-LDH in the temperature between 30~ and 180~ of Figure 1. A metastable phase appears at 180~ with diffraction lines slightly shifted to the higher 20 angle thus giving a smaller basal spacing d=7.4 A (20 = 11.9 A). The reduced basal spacing has been also reported at 120~ in the HTXRD measurement of ZnA1/C1-LDH whereas the reduced value recovers on further increasing the sample temperature followed by the collapse of the layer structure [ 18]. The reason for appearing a metastable phase having smaller basal spacing is assumed to be the same as that of MgA1/CO3-LDH described before. It is common to the three LDHs above that the basal spacing is reduced at lower temperature than that of the layer collapse. This metastable phase disappears promptly at 200~ which is lower than the temperature that Phase II disappears in Figure 1, suggesting the thermal instability of the metastable phase of ZnA1/CO3-LDH. The layered structure collapses and the dehydroxylation occurs at this temperature producing metal oxides, finally a spinel phase appears above 800~ The temperature at which the layered structure collapses differs depending on the divalent metal species in the layers; 200~ for Zn and 379~ for Mg. DTA/TG thermal analysis in Figure 7 also illustrates the thermal change of the hydrotalcite-like layered structure of ZnA1/CO3-LDH. A sharp endotherm at 180~ appears with a large mass loss. In the DTA curve of ZnA1/C1-LDH, however, it has been reported that the intense endotherm splits in two; one is assigned to the elimination of interlayer water and the other to that of lattice one due to the layer collapse [19]. The endotherm in Figure 7, however, includes the elimination of the interlayer water/carbonate and dehydroxylation of the layers of the double hydroxide which induces the collapse of the layered structure. The endotherm at 180~ in Figure 7 corresponds to the metastable phase also observed at 180~ in Figure 6 and this phase promptly disappears on increasing the temperature. In Figure 8, three absorption bands are observed in 27A1 MAS NMR spectrum of the calcined ZnA1/CO3-LDH at 500~ A sharp band located at 6=8.890 ppm, a shoulder at 5=50.015 ppm and a broad one at 5=72.289 ppm in this figure agrees with those bands which have been assigned to the 27A1nuclei in OhA1, in the penta-coordinated A1 [21,23] and in the tetra-coordinated one (TdA1) [20-24], respectively.

356

E. Kanezaki

or9 +.a

@ o,.u ,k.a

o,.~

20

z~O

20/deg.

60

Figure 6 - In situ HTXRRD patterns of ZnA/CO3-LDH on increasing the temperature; at room temperature, IO0~ 120~ 140~ 160~ 180~ 200~ 220~ 240~ 260~ 280~ 300~ 320~ 340~ 360~ 380~ 400~ 500~ 600~ 700~ 800~ 900~ and 1000~ (from bottom to top). [50] DTA

TGA

exo.

0 f~ g3

t"

O ur f/3

endo.

20 '~

n

n

200

t

t

I

I

n

n

600 ~ t,,~',-,'rem".mc'

t

I

1000

Figure 7- DTA/TG thermal analysis of ZnAI/CO3-LDH. [49] Both bands of the non-octahedrally coordinated A13+in this figure disappear in the NMR spectrum of the rehydrated LDH; all of the AI3+ ions uniformly locate in the octahedral coordination sphere after the intercalation of anions in the rehydration process. Although this uniformity has been also reported previously for some intercalated anions [20,21 ], no definite explanation has been made; abundant resources of donating atoms in

Preparation of Layered Double Hydroxides

357

aqueous solution containing anions likely cover the deficiency in the donating atom of non-octahedral A13§ upon the rehydration.

50.015 ppm 9

4;0

'

11

2;0

' () ' -200 ' 5/ppm Figure 8 - eZAlMAS NMR spectrum of ZnAI/CO3-LDH calcined at 500~ measured with the spinning frequency of 7 kHz; chemical shifts of three maxima are indicated, peaks with asterisks are spinning side bands. Reprinted from [47] with permission by Kluwer Academic/Plenum Publishers. 2 - Intercalation of organic anions to ldh 2.1 - Intercalation by rehydration method

It has been reported that aliphatic [4] and aromatic [2, 25-30] molecular dianions are intercalated in hydrotalcite-like layered compounds. When these large organic ions are intercalated, an increase of the basal spacing is usually observed in powder XRD patterns as described before. Furthermore, the magnitude of the interlayer distance depends on the orientation of the interlayer organic molecules since organic molecules generally have spatial anisotropy in the dimension of molecules and since the dimension, which is measured along the direction perpendicular to the layer surface, reflects on the magnitude of the interlayer distance of the intercalated products. A careful examination of XRD patterns of these LDH gives possible orientation of the interlayer molecules and thereby insight into the interaction(s) between the molecules and other component(s) in the products as well as understanding the properties of the products for potential use. In this section, intercalation of naphthalenedisulfonates by the rehydration method is described, in which the calcined ZnA1/CO3-LDH is used as the precursor. Three isomers of naphthalenedisulfonates (naphthalene-2,6-, -1,5- and-2,7-disulfonates; abbreviated as N26DS, N 15DS and N27DS, respectively, and collectively abbreviated as NijDS) are intercalated in the interlayer region of Zn and A1 layered double-hydroxides. Figure 9 exhibits two XRD patterns of ZnA1/N26DS-LDH; the basal diffraction appears at 20=5.74 A (d=15,38 A) and is associated with some prominent (00/)-type diffractions in Figure 9A (the 15 A phase). It has been reported that the unit cell of ZnA1/CO3-LDH has hexagonal symmetry with a lattice constant a0=3.06 A and with the c axis perpendicular to the layers [2,13]. The XRD patterns of ZnA1/NijDS-LDH were thus analyzed on the assumption that the unit cell symmetry and the a0 dimension was the same as those of ZnA1/CO3-LDH but that the dimension of the other lattice constant Co

E. Kanezaki

358

differs among the products. Indexing of the pattem in Figure 9 indicates that only one XRD-active phase exists and that the predominant crystal growth occurs in the direction parallel to the c axis in this solid sample, which points out well developed stacking of the layers of Zn and A1 double hydroxide along this axis [1,2]. On the other hand, the basal diffraction shifts to a slightly smaller diffraction angle of 20-5.18 A (d=17.05 A) in Figure 9B (the 17A phase).

002

A

i~

004 003

006

I]0

001

B 002

=

l*

@

I,I 003 ill i

004 [

20.0

005

:007 I

, ~104

110

112

40.0

20/deg.

60.0

Figure 9 - XRD patterns of ZnAI/N26DSLDHs; (A) the 15.,{phase with diffraction lines at 20=5.740/[ (001), 11.58/~ (002), 17.44 ~ (003) 23.36/~ (004), 29.38 ~ (005), 34.54 (006), 41.50 ~{ (007), 46. 74/[ (008), 52.94 ~{ (009), 60.20/~ (00 10) and 61.50 ~ (110): (B) the17A phase with lines at 20=5.18 A (001), 10.50/[ (002), 15.72/[ (003), 21.12 (004), 26.42 A (005), 37.32 ~ (007), 53.12/[ (00 10) and 60.22 A (11o); Zines with asterisks are those of the host ZnAI/CO3-LDH without interlayer aromatic molecules. Reprinted from [48] with permission by Kluwer Academic~Plenum Publishers. The 17A phase of ZnAI/N26DS-LDH was obtained when the amount of the organic salt was increased relatively to that of aluminum in the calcined powder; the ratio of (organic salt)/A1 =5 and 1 for the 17A and the 15A phases, respectively. In addition to the intense (001) diffraction of the 17A phase associated with some (00/)-type lines, another sequence of diffraction lines (marked with asterisks in Figure 9B) is present. Since the additional sequence agrees with the XRD pattern of the precursor, it is concluded that the synthetic condition for the 17A phase of the ZnAI/N26DS-LDH also favors the formation of the carbonate-intercalated product in which interlayer carbonate

Preparation of Layered Double Hydroxides

359

are originally solvated species in the reaction mixture. The XRD patterns in Figure 10A and B are of ZnA1/N15DS-LDH; the 15A phase with the basal diffraction at 20=5.82 A (d=15.17 A) and the 17A phase with the diffraction at 20=5.24 A (d= 16.85 A), respectively. Both phases in this figure exhibit well developed stacking of the layers, though the 17A phase in Figure 10B appears together with the carbonate-intercalated product which is also observed in Figure 9B. Only the 15A phase in Figure 10A is formed after leaving the precipitate in the reaction mixture for 40 days, therefore, it is suggested that this phase is thermodynamically more stable than the 17A phase. The 15A phase has been observed within several kinds of LDH by some authors [25,30]. In contrast, only the 17A phase is observed in the XRD pattern (not shown) of ZnA1/N27DS-LDH, in which the basal diffraction appears at 20=5.34 A (d= 16.54 A). It should be noted that the magnitude of the interlayer distances of the 17A phases of all the NijDS-intercalated products in Table 2 is in good agreement with the estimated molecular size of the individual guest organic anions using MO calculations. This result is significant and will be discussed later.

t

002

00l [ 004 [ [ oo [

r,r

c (D

c c o

. ,,,,~ r

. ,,,,~

A

~

l

006 [103008

112

B

001 rae)

c

002

004

=ll i 0:

c

.o

~

_

2o.o

4o.o

_~A~,..~ _

20Meg.

6o.o

Figure 1 0 - X R D p a t t e r n s o f ZnAI/N15DS-LDH: (A) the 15,4 phase with diffraction lines at 20=5.820 ~ (001), 11.68 ~ (002), 17.58 ~ (003) 23.48 ~ (004), 29.24 ~ (005), 35.60 ~ (006), 60.20 ~ (110) and 61.72 ~ (0010); (B) the 17,4 phase with lines at 20=5.24 ~ (001), 10.54 ~ (002), 15.80 ~ (003), 21.18 ~ (004), 26.54 ~ (005), 37.36 ~ (007), 56.48 ~ (00 10) and 60.26 ~ (110); lines with asterisks are the same as those in Figure 9. Reprinted from [48] with permission by Kluwer Academic~Plenum Publishers.

E. Kanezaki

360

The 17A phase of ZnA1/N26DS-LDH has two maxima located at kmax = 277 nm and at 7~max-314 nm in the diffuse reflectance (DR) spectrum (not shown), which are assigned as a 1B2,(=, zt*) and a 1B3u(rc, n*) transitions, respectively in the naphthalene moiety [ 14]. This profile in the DR spectra is commonly observed in all the spectra of ZnA1/NijDS-LDH in this study. Therefore, it strongly suggests that the interlayer organic anions retain the planar molecular skeleton of the naphthalene moiety with the stable conjugated p-electron framework. A wide scan XPS spectrum of the 15A phase of ZnA1/N26DS-LDH shows several peaks which are assigned to core level electrons emitted from Zn, A1, C, S and O atoms of the solid samples [32]. Table 2 - Summary of results and synthetic conditions of ZnAI/NijDS-LDH. Reprinted from [48] with permission by Kluwer Academic/Plenum Publishers.

Ij

IM~)/A

lLZ)lA NijDS/CO33) Zn:AI:NijDS 4)

26

12.66

15

12.05

12.3 10.6 12.1 10.4 10.3 10.3 10.4 10.8 11.7

27

12.08

0.67 0.56 0.56 0.39

0.26

4:2:5 4:2:1 4:2:1 4:2:1

4:2:1

MM ''5)

notes

ZnA1

17A phase[43] 15A phase[26] 17A phase[26] 15A phase[43] 15A phase[Z5] 15A phase[30] 15A phase[30] 15A phase[30] 17A phase[26]

ZnA1 MgA1 ZnA1 ZnCr CaA1 ZnA1

1) Molecular size: twice the anionic radius of oxygen (2.8 ~) plus the interatomic distance between two anionic groups of an aromatic dianion whose geometry is optimized in MO calculation. 2) Interlayer distance: the interplanar spacing d(O01) obtained in XRD patterns minus the thickness of a layer (4. 8~). 3) Molar ratio of intercalating NijDS and C032- based on the chemical analysis of intercalated products with the composition of the layer [ZnAlo.5o(OH)3.oo] and with compositions of the interlayer are (C03)o.ss~26DS)o.so, (C03)o.26(N26DS)o.o9, (C03)o.s6(N15DS)o.og, (C09o.sg(N15DS)o.oTfor the 17,,{phase and for the 15~ phase of ZnAl/N2 6DS-LDH and of ZnAI/N15DS-LDH, respectively, and (C0 3)o.2o(N27DS)o.o5for the 17~ phase of ZnAI/N27DS-LDH; water and chloride are omitted. 4) Molar~atomic ratio in preparation. 5) Divalent (M) and trivalent (M') metal ions within layers. In the precursor LDH with interlayer carbonate, the binding energy of the A12p electron, Eb(A1 2p), is 74.2+ 0.6 eV (1 eV=l.602xl0 19 J) which agrees with that of the octahedrally coordinated A13+ions previously reported [23]. No significant shift of Eb(A1 2p) is observed after intercalation of NijDS within experimental error; Eb(A1 2p) of the 15A and of the 17A phases are 74.9+ 0.6 and 75.0+ 0.6 eV for ZnAIfN26DS-LDH, 74.6 -----0.7 and 74.9-+-0.6 eV for ZnA1/N 15DS-LDH, respectively, and 74.4--- 0.6 eV for the ZnAl~27DS-LDH. Therefore, no dependence of Eb(A12p) on the magnitude of the basal spacing results in this study. Since the binding energy is a good measure of the chemical status of metals [32], it is concluded that the aluminum ions in the layers still exist as

361

Preparation of Layered Double Hydroxides

trivalent cations and are coordinated octahedrally in all phases of ZnA1/NijDS-LDH. In other words, the two values for the basal spacings of ZnA1/N26DS-LDH and ZnA1/N15DS-LDH do not originate from any change in the chemical status of the A13+ ion in the layer. A typical DTA/TG thermal analysis curve of ZnA1/%I27DS-LDH (Figure 11) shows two prominent endotherms (at 195~ and at 283~ and two exotherms (at 520~ and at 565~ in the DTA curve9 This profile agrees well with that of the 15A phase of ZnA1/N26DS-LDH [26]; two endotherms at 185~ and at 308~ were both assigned to the elimination of water and CO2, and exotherms at 487~ and at 565~ to thermal decomposition of N26DS at the outer surface and in the interlayer region, respectively. The result of the previous study leads us to the parallel conclusion that the two endotherms are assigned to the elimination of water and CO2 molecules and that the two exotherms are due to the decomposition of N27DS anions at the outer surface (520~ and in the interlayer region (565~ It is important that two anion-sites at which NijDS ions locate, the outer surface and the interlayer region, are distinguished thermally [26]. The total mass loss in the TG curve in the temperature range up to 800~ is 35.8 % which corresponds to the sum of water, carbonate and N27DS contents (36.0 %) being estimated from the result of the chemical analysis in Table 2. All of the ZnAI/NijDS-LDHs give the similar DTA/TG profile, which indicates that the peak locations are insensitive both to the difference in the magnitude of the basal spacing and to the kind of guest isomer. TG

DTA 520

i,

100

80

Endo. 195 60

2()0

400

600 Temperature / I,C

800

Figure 11 - DTA/TG analysis of ZnAl/N27DS-LDH in the range of (room temperature)_
362

E. Kanezaki

in his early crystallographic work [ 13 ]. He concluded that the interlayer CO32 anion has the conformation in which the molecular plane of the ion lies parallel to the two-dimensional layers thus the magnitude of the interlayer distance is essentially equal to the diameter of the O2 anion. Although an orientation of interlayer organic anions, which is perpendicular to the layer surface, was concluded in earlier studies [25,27,28], a tilted orientation of the anions has been also suggested in recent XRD studies when basal spacings of the intercalated products are smaller than molecular sizes of the anions [2,26,29,30]. Since the interlayer distance of the 17A phase in Table 2 agrees with the estimated molecular size of the anion in all of the ZnA1/NijDS-LDHs, this result suggests strongly that the interlayer NijDS anion in this phase takes the bridging orientation which links a pair of A13§ ions in the neighboring layers like a fastener zipping on these two layers. The line that connects two anionic oxygen atoms (the O-O line), each of which belongs to different sulfonate groups in one bridging NijDS, is parallel to the c axis in this orientation. In the 15A phase, however, the interlayer distance is smaller than the molecular size of the guest in Table 2. A plausible explanation is displayed in Figure 12 for the case of interlayer N 15DS; the guest molecule has the orientation in which the O-O line is not parallel to the c axis. It is the author's opinion that the variation in the magnitude of the basal spacing is caused by the variation in the magnitude of the enhanced interlayer distance owing to an alternate orientation by the interlayer molecular anions in the cases of N26DS and N15DS. The variation in the basal spacing does not result fi'om the difference in the thickness of the layer because this variation is so small in both cases (1.7A) that it cannot be explained by the change of the stacking number of the octahedron unit. A pair of A13+ ions in the 15A phase, which are linked by one N15DS molecule, do not locate directly opposite to each other but are separated in length (R) by R-6.0• A measured along the inner surface of the layers (Figure 12) although R=0 in the 17A phase. The separation in this figure agrees with that value of R-6.8_0.5 A previously reported for the case of the interlayer N26DS anion in the 15A phase [26] and both values are approximately equal to twice the interatomic distance between two adjacent metal ions (Ro=3.12 A [34] ) in the layers of Mg and A1 double hydroxide. If it is assumed that interatomic distances in the layers of Mg and A1 double hydroxide and in the layers of Zn and A1 double hydroxide are the same, it is interesting that only two values of the ratio R/R0 are observed; (R/Ro)=2 for the 15A phases and (R/Ro)=0 for the 17A phases. Although the author cannot propose the reason why a value of 1 for the ratio is missing, it is assumed that the reason may be steric hindrance due to the bulky naphthalene moiety. The chemical compositions of the interlayers of the intercalated products in Table 1 shows both of carbonate and NijDS anions coexist in all solid samples of ZnA1/NijDS-LDH. It has been reported that some kinds of LDH with interlayer carbonate are synthesized [4]; the interlayer distance (ca. 3 A) of them, which is equal to the molecular size of carbonate, is much smaller than the estimated molecular size of NijDS in Table 2. Therefore, even when carbonate and NijDS ions coexist in the interlayer, we cannot recognize the coexistence of the two ions in XRD patterns because NijDS ions would act as pillars between two layers of double hydroxide, which is the case which we observe in the 15A phases. In contrast, when two kinds of intercalated product are formed as a mixture, one is ZnA1/NijDS-LDH and the other is ZnA1/CO3- LDH, we can recognize this situation in the XRD measurement, which is the case which we observe in the 17A phases. No evidence to support the presence of superlattice is observed so far.

Preparation of Layered Double Hydroxides

363

The molecular ratio NijDS/CO3 is always larger in the 17A phase than in the 15A phase for ZnA1/N26DS-LDH and ZnA1/N 15DS-LDH. Considering that the 17A phases of these two intercalated products are formed together with the ZnA1/CO3-LDH and that the compositions of the 17A phases in Table 1 are thus of a mixture of them, the real ratio in the pure phase is more larger than the value listed in Table 1 for both 17A phases.

Figure 12 - A plausible orientation of the interlayer N15DS between two layers of Zn and AI double hydroxide in the 15/[ phase; other anions and water molecules are omitted. Reprinted from [48] with permission by Kluwer Academic~Plenum Publishers. We therefore conclude that the guest molecules in the 17A phase are packed in the interlayer region more closely than in the 15A phase, which is realized by taking the orientation of the guest molecule in the 17A phase as described above. The reason why more than one basal spacing is not observed in ZnA1/N27DS-LDH is not known. In the previous work, we reported that only the 15A phase was observed when N26DS is intercalated and that only the 17A phase was observed when N15DS or N27DS was intercalated [26] as listed in Table 2. Together with the results of Drezdzon [25] and of Meyn et al [30], this study reveals that the duality in the magnitude of the interlayer distance due to the orientation change of the interlayer molecules is commonly observed when N26DS or N 15DS is intercalated between layers of Zn/and A1 double oxide. This duality has been also reported by our laboratory when AQ26 is intercalated between layers of Mg and A1 double hydroxide [2] which is described in the next section. It is

364

E. Kanezaki

important that the magnitude of the basal spacing is changeable by means of controlled preparation conditions. With this changeable magnitude of the basal spacing, we would be able to design materials having desirable pore-sizes in the mesopore region, which is of value in reaction catalysis and in adsorption technology.

2.2 - Intercalation of organic anions by the coprecipitation method 2.2.1 - Intercalation of Naphthalenedisulfonate at the interlayer region of Mg and AI double hydroxide.

In the high temperature in situ XRD patterns of MgA1/N26DS-LDH (Figure 13), there are three temperature regions which are classified by the common XRD pattem; room temperature_
Preparation of Layered Double Hydroxides

365

interlayer water and Vco vibration of the interlayer carbonate are observed in 3500-3400 cm ~, at 1634 cm 1 and at 1364 cm ~, respectively, in this spectrum. The other sharp absorption bands located in the region 1500-600 cm "l have been assigned to molecular vibrations of the interlayer N26DS [35,36]. Three S-O bonds in the sulfo group of the interlayer N26DS are classified in two categories; one bond has the oxygen atom coordinating to the metal cation in the layer and the other two bonds have not. The Vso vibration in the former bond is observed in the FT-IR spectrum of the free sodium salt of N26DS (not shown) at 973 cm 1 whereas this band shifts to the lower energy side in all of the FT-IR spectra of the interlayer N26DS up to 400~ (Figures 15(A)-(D)) [ 16].

~5

2'0

40

20/deg.

60

Figure 13 - In situ HTXRD patterns of MgAI/N26DS-LDH at different temperature from bottom to top, room temperature, 100~ 120~ 140~ 160~ 180~ 200~ 220~ 240~ 260~ 280~ 300~ 340~ 380~ 400~ 500~ 600~ 700~ 800~ 900~ and 1000~ indexing o f the diffraction at room temperature is 20 (hkl) = 5.26 ~ (001), 10.56 ~ (002), 15.86 ~ (003), 21.24 ~ (004), 26.44 ~ (005), 34.68 ~ (102) and 60.92 ~ (110) under the assumption of a hexagonal unit cell with lattice constants ao=3.04 ~ and Co=16. 77/[ [49] Figure 15(A) shows the FT-IR spectrum of MgA1/N26DS-LDH at room temperature. The broad VoH vibration of the hydroxyl group, the 5HOHvibration of the interlayer water and Vco vibration of the interlayer carbonate are observed in 3500-3400 cm 1, at 1634 cm l and at 1364 cm -1, respectively, in this spectrum. The other sharp absorption bands located in the region 1500-600 cm 1 have been assigned to molecular vibrations of the interlayer N26DS [35,36]. Three S-O bonds in the sulfo group of the interlayer N26DS are classified in two categories; one bond has the oxygen atom coordinating to the metal cation in the layer and the other two bonds have not. The Vso vibration in the former bond is observed in the FT-IR spectrum of the free sodium salt of N26DS (not shown) at 973 cm -~ whereas this band shifts to the lower energy side in all of the FT-IR spectra ofthe interlayer N26DS up to 400~ (Figures 15(A)-(D)) [ 16]. Two Vso vibrations in the latter two S-O bonds are observed at 1037 cm l and at 1234 cm ~ both in the spectra of the free salt and that of the interlayer N26DS (Figures 15(A)-(D)) although the latter band is not distinct accidentally in Figure 15(A). The coordination of the sulfo

366

E. Kanezaki

group(s) in the interlayerN26DS would have so large influence on the results of XRD and of FT-IR of MgAI/N26DS-LDH that its thermal profiles in both measurements are different from those of MgA1/CO3-LDH.

8.68 ppm

L

4()0

'

200

'

0

'

-200

Chemical Shift/ppm Figure 14 - 27.4l MAS NMR spectra of MgAI/N26DS-LDH at the spinning rate of 7 kHz; peaks with an asterisk are spinning side bands. [49]

Figures 15(B)-(E) are the FT-IR spectra of MgA1/N26DS-LDH at 140~ 210~ 400~ and at 500~ respectively. On elevating the temperature to 140~ an absorption appears at 866-862 cm1. Although this band locates near the VMgOvibration in the spectra of MgA1/CO3-LDH, it should not be the VMgOvibration because MgO is not observed in the XRD pattern of Mg/A1/N26DS-LDH below 500~ in Figure 13. Therefore, the band at 866-862 cm1 is tentatively assigned to the VA10vibration of the amorphous aluminum oxide. This assignment is reasonable because the observed energy (866-862 cm"1) agrees with one energy quantum of the VA10vibration (cr(A1-O)=865 cm q) calculated by the equation (1) with the assumption that the force constant in the VA10 vibration K(A1-O) is equal to the constant in the VMgOvibration ~:(Mg-O), o(AI-O)/o(Mg-O)=0.98 {~c(A1-O)/K(Mg-O)}0.5

(Eq. 2)

where o(Mg-O) is one energy quantum of the VMgOvibration and is equal to 882 cm"1. Since A1203 is not observed in the FT-IR spectra of MgA1/COa-LDH it is likely that the presence of the interlayer N26DS is one of the reasons for observing this solid in Figure 15(B). It is supposed that the aluminum ion which is in the Oh coordination sphere of the layer and at the same time is coordinated by the interlayer N26DS at room temperature would be unstable because of the bulky nature of the coordinating sulfo group in the interlayer N26DS and thus the metal cation undergoes the change into the Td structure on elevating the temperature with keeping the coordination by the N26DS. Since the aluminum ion in the Td structure cannot remain in the layer, it moves to the interlayer region thus producing a vacancy in the layer. This thermal change lowers the regularity along the c-axis because the interlayer organic anion no longer supports the width of the

Preparation of Layered Double Hydroxides

367

interlayer distance whereas the regularity along the ab-plane has little influence by the change. The resulting Td A13+would be observed in the high temperature measurement of 27A1 MAS NMR spectra. In the spectrum at 200~ (Figure 15(C)), the absorption by the interlayer water observed at 1634-1627 cm 1 in Figures 15(A) and (B) disappears because of the thermal elimination of water, which elimination is also described in the spectrum of MgA1/COB-LDH at 210~ (Figure 5(B)). The disappearance of the interlayer water corresponds to the first endotherm at 191~ in DTA/TG thermal analysis curve of MgA1/N26DS-LDH. The spectral profile in the molecular vibrations of the interlayer N26DS at 200~ is almost the same as that at 140~ indicating no change in the chemical status of the organic anion. On elevating the temperature to 400~ (Figure 15(D)), the absorption due to the interlayer carbonate located at 1362-1364 cm 1 in the spectra below 200~ still remains. The interlayer carbonate in MgA1/N26DS-LDH does not change its chemical form on elevating the temperature to 400~ which result differs from that of MgAI/CO3-LDH. This difference is elucidated from the fact that the interlayer carbonates in the latter compound are in the close contact with the metal cations in the layers because of the narrow spacing of the interlayer gallery region and hence the anions easily undergo the thermal reaction with the cations whereas the anions in the former compound are not. However, the intensity of the bands for the interlayer carbonate is more weak at 400~ than that at 200~ because a part of this anion eliminate from the sample at the former temperature. The elimination of the interlayer carbonate is indistinguishable in the DTA/TG thermal analysis curve of MgAI/N26DS-LDH as an independent event because it is included in the tail of the broad endotherm centered at 434~ The impossibility of distinguishing the CO2 elimination in the thermal analysis curve has been also described in the result of MgA1/CO3-LDH [42]. Almost all of the molecular vibrations originated from the interlayer N26DS are still observed in the spectrum at 400~ (Figure 15(D)) except for the skeleton vibration of the naphthalene moiety at 1496-1500 cm -~ observed in the spectra below 200~ (Figures 15(A)-(C)) because of the thermal deformation of the skeleton of the moiety. This absence also occurs in the spectrum of the free sodium salt of N26DS at 400~ In the IR spectrum at 500~ (Figure 15 (E)), however, the intensity of all bands decreases in common. The broad VOH vibration at 3500-3600 cm 1 becomes extremely weak because of the dehydroxylation of the layers, which agrees with the results in XRD and DTA/TG, where the layered structure disappears completely at 500~ in the former measurement and the endotherm at 434~ appears in the latter thermal analysis curve. Furthermore, the profile in the region of the molecular vibration of N26DS is changed considerably because of the thermal decomposition of this organic anion. The strong absorption located at 1180-1187 cm -~ in Figures 15(A)-(D) has been assigned to the ~SCHvibration of the naphthalene moiety in the interlayer N26DS [35,36] and disappears in Figure 15 (E) as the result of the thermal decomposition of this moiety. The absorption located at 1032 cm -~ in Figure 15(E), which is assigned to one of the Vso vibrations of the non-coordinating S-O bond in the sulfo group of the interlayer N26DS as described before, also becomes a weak shoulder at 500~ The decomposition of the interlayer naphthalenedisulfonate is also observed both in the XRD pattern and in the DTA/TG thermal analysis curve; MgSO4 appears in the XRD pattern at 600~ as a reaction product between Mg 2+ and the organic anion (Figure 13) and the broad endotherm centered at 749~ is observed in the latter thermal analysis curve (the result is not shown). In conclusion, the thermal change in the layered structure of

368

E. Kanezaki

MgA1/N26DS-LDH on elevating the sample temperature to 1000~ is understood by means of in situ HTXRD with the aid of DTA/TG, 27A1 MAS NMR and the high temperature in situ FT-IR study.

/

'

/

~

(E) at 500I~C

E

I 4000

I 3000

I 2000

o I cm 1

I 1000

I 400

I

4000

I

3000

I

2000

o I crrf I

I

I

1000

400

Figure 15 - High temperature in situ FT-IR spectra o f MgAl/N26DS-LDH at room temperature (A), 140~ (B), 200~ (C), 400~ (D) and at 500~ (E). [49] 2.2.2 - Intercalation of 9,10-Anthraquinonedisulfonates at the interlayer region of Mg and AI double hydroxide

In this section, the intercalation of 9,10-Anthraquinone-disulfonates (AQij; ij=l 5, 18, 26 and 27) between layers of Mg and AI double hydroxide is described. The synthetic method of the coprecipitation has been described previously [2,43]. Aqueous solution of MgCl2 and A1C13,metal sources of the double hydroxide, were simultaneously added to an aqueous solution of a sodium salt of AQij with the molar ratio of Mg:Al:AQij and with vigorous stirring at room temperature in air and with keeping pH followed by aging overnight at 73-74~ The precipitate forms immediately after the addition of solutions altogether and is washed until a negative result in the test for CI with Ag § for the filtrate. Owing to the intercalation of AQ27 between the layers of Mg and AI double hydroxide, the basal spacing in XRD patterns is enhanced to 19 A, 15 A or 12 A and the resulting solids are called hereafter as 19A phase, 15A phase and 12A phase, respectively. The patterns of these solids are indexed and also displayed altogether in Figure 16 with the assumption of the hexagonal unit cell symmetry. It is observed that the magnitude of

Preparation of Layered Double Hydroxides

369

one lattice constant ao (3.06 A) is the same as that of MgAI/CO3-LDH in Table 1 although the other lattice constant Co is large. The symmetric and sharp basal diffraction indexed (001) is associated with a sequence of prominent (00/) lines in all the patterns and shows that each one of three phases are formed by the pure product. This observation indicates that the brucite-like layers (4.769 A in thickness [4]) in the intercalated products have the well developed stacking in the direction parallel to the c-axis. Basal spacings of all the phases are listed together with the other results in Table 3. Two values of the basal spacing are also observed when AQ26 is intercalated instead of AQ27 by means of the same method, although only the 12A phase has been obtained so far under the hydrothermal condition at 127+ 1~ in aging precipitates. Only a single phase has been observed when AQ 15 or AQ 18 is intercalated. The position of two -SO3- groups in the interlayer AQij thus governs whether the multiplicity in the basal spacing appears or not as well as the degree of the multiplicity. Although the mechanism is not known, it is interesting that the multiplicity appears preferably with isomers of 2,6- and 2,7-substitutions both of which have two -SO3 groups at the most distant positions from the carbonyl groups in anthraquinone. An absorption at around ~,=315 nm has been reported due to the 1B2u(n,7~*) transition of anthraquinone [44]. The guest AQij molecule is responsible for all the absorption maxima in the DR spectra of AQij-intercalated products except for a gradual signal rise to the short wavelength side. Considering that no significant change is noticed when the spectra are compared with each other among all the AQij-intercalated products, it is concluded that the interlayer AQij molecules commonly hold the planar framework of the conjugated n-electron system in the anthraquinone moiety. A weak absorption due to the 1B~g(n,n*) transition to the lowest singlet state of anthraquinone [44] is located at around k=485 nm in the DR spectrum of the solid sodium salt of AQij. Instead of this band location, a broad absorption appears at ~,max=501-526 nm. The check does not confirm that this absorption is due to the intercalation of AQij. Judging from the comparable intensity of this band to the allowed (n,n*) transition described above, it is suggested that this absorption originates from electronic transition(s) between the conjugated n-electron orbital(s) and other orbital(s) which results from the interaction(s) between the intercalated AQij and other species in the intercalated product. XP wide scan spectra of the MgA1/AQij-LDHs commonly show several peaks which are assigned to core level electrons emitted fi'om C, S, O, A1 and Mg [32]. The Eb(A12p) is 74.0+0.9 eV and 73.7+ 0.9 eV for the 19A and the 12A phases, respectively; the variation of the energy due to the difference in the basal spacing is insignificant. These values agree with that of the octahedrally coordinated A13§ within experimental errors [23]. Since MgA1/CO3-LDH has Eb(A1 2p)=73.6 eV [2,26], it is concluded that insignificant change takes place in the chemical status of this cation in the layers upon the intercalation of AQij. A very weak signal at around 200 eV is additionally observed in a few samples after they are etched by an Ar § bombardment for 60 s. It corresponds to tiny quantity of CI detected by the potentiometric titration previously described. Typical results of MgA1/AQij-LDHs are listed in Table 3 including basal spacings and chemical compositions. Since each of these phases has never been produced as a mixture, it is concluded that the optimum synthetic condition for each phase is not overlapped. Although the 12A phase in MgA1/AQ26-LDH is preferable in the hydrothermal condition as previously described, both the phases in MgAI/AQ26-LDH and all of three phases in MgA1/AQ27-LDH are thermodynamically stable at room

370

E. Kanezaki

temperature because no mutual transformation is observed in XRD patterns after these solid samples are left in air for a month. The results ofXRD, XPS and DR spectra force us to conclude that the multiplicity in the basal spacing is originated from the distinct dimension of interlayer distance which is govemed by the size of the interlayer guest AQij molecules along the c-axis.

001 I

r~

19A phase

004 2 I 005 [003.1, I

001 002 003 [ I 004 [

|

001

002 I

15A phase

006

,

003

I

i

12A phase

I

2o

20 / deg. 40

60

Figure 1 6 - Indexing in XRD patters of MgAl/AQ27-LDH by means of the Ni-filtered CUKal line 0,=1.5405 ~); thel9 /[ phase (upper) with diffraction lines at 20= 4.780 ~ (001), 9.500 ~ (002), 14.20 ~ (003), 18.82 ~ (004), 24.10 ~ (005) and 61.28 ~ (110); the 15~ phase (middle) with lines at 5.880 ~ (001), 11.94 ~ (002), 17.84 ~ (003), 24.14 ~ (004), 34.82 ~ (006) and 60.70 ~ (110); the 12~ phase (lower) with lines at 20=7.22 ~ (001), 14.34 ~ (002), 22.18 ~ (003), 35.60 ~ (005) and 60.82 ~ (110). Reprinted from [43] with permission by Taylor & Francis.

Preparation of Layered Double Hydroxides

371

The A1/(AI+Mg) ratio in Table 3 remains within the limit for Mg and A1 double hydroxide reported previously [4]. The interlayer distance of the 19A phase of the AQ26-intercalated product is 14.26 A which is just less than the calculated guest molecular size lM=15.285 A in Table 3. This is rationalized if we assume that an orientation of the interlayer AQ26 molecule bridging as a bidentate between adjacent layers of Mg and AI double hydroxide, which has been proposed in the cases of N26DS between layers of Zn and A1 double hydroxide (Figure 12), C104 and 8042"both between layers of Mg and A1 double hydroxide [2,4,26,45,46]. This orientation is held by the attraction between the positive charge on A13+ in the layer and the negative one on -SO3 in the AQ26 anion at the interlayer. A slightly tilted molecular plane of the interlayer AQ26 relative to the inner surface of the layer is suggested for the 19A phase and a more tilted one is supposed for the 12A phase. Similar orientations are proposed for interlayer AQ27 anions in three phases of intercalated products. Table 3 - Summary of synthetic and analytical results of MgAI/AQij-LDHs. Reprinted from [43] with permission by Taylor & Francis.

AQi~ AQ15 AQ 18 AQ26

E1)/eV 74.0+__0.9 73.8+ 0.8 73.9___0.9

IMZ)/A 12.882 10.444 15.285

AQ27

73.8-t-0.9

15.520

d3)/A 11.64 11.46 12.73 19.03 12.23 15.02 18.47

Formula MgAlo.3I(OH)2.93(CO3)0.06(AQ15)0.09 MgAlo.43(OH)z.ss(AQ 18)o.21 MgA10.47(OH)2.94(CO3)0.12(AQ26)0.12 MgAlo.47(OH)z.95(CO3)o.o4(AQ26)o.19 MgAlo.22(OH)2.39(AQ27)0.13 MgA10.33(OH)2.63(AQ27)0.19

1) binding energy of Al 2p electron in XPS. 2) molecular size: the sum of ionic diameter of anionic oxygen (2.56 ~) and the interatomic distance between two anionic oxygens in different -SOl groups in AQij whose geometry is optimized by MO calculation, see text. 3) interplanar spacing obtained d(OOl) in XRD patterns. This bridging model of the guest anions leads to the upper limit of the AQ26/A1 ratio is 0.5. When the ratio in the analysis is lower than this limit, the intercalation of CO32 is not negligible as is indicated in the analytical formulae in Table 2. In the 19A phase, the ratio are larger than that in the 12A phase. It is plausible that the orientation of the interlayer AQ26 anions in the former phase is such that they take the close-packed structure thereby a high ratio is realized whereas the anions in the latter phase do not take the structure. Steric hindrance due to the bulky anthraquinone moiety is supposed for the reason of it although no reason is supposed for the AQ27-intercalated products. 3 - References [ 1] F. Cavini, F. Trifiro and A. Vaccari, Catalysis Today, 11 (1991) 173; A. Roy, C. Forano, K. E. Malki and J. Besse, Expanded Clays and Other Microporous Solids; Anionic Clays: Trends in Pillaring Chemistry, Eds. M. L. Occelli and H. Robson, Van Nostrand Reinhold, New York, 1992, chap. 7; D. O'Hare, ed., Inorganic Materials, 2nd ed., John Wiley & Sons, Chichester, 1997; V. Rives and M. A. Ulibarri, Coordination Chem. Rev.,181 (1999) 61; T. J. Pinnavaia and G. W.

372

E. Kanezaki

Beall, Eds., Polymer-Clay Nanocomposites, John Wiley & Sons, Chichester, 2000. [2] E. Kanezaki, S. Sugiyama and Y. Ishikawa, J. Mater. Chem., 5 (1995) 1969. [3] MgO (periclase) JCPDS card No.45-9846; MgA1204 (spinel) JCPDS card No.21-1152. [4] S. Miyata, Clays Clay Miner., 23 (1975) 369. [5] W. Yang, Y. Kim, P. K. T. Liu, M. Sahimi and T. T. Tsotsis, Chem. Engin. Sci., 57 (2002) 294 [6] M. Belloto, B. Rebours, O. Clause, J. Lynch, D. Bazin and E. Elkaim, J. Phys. Chem., 100(1996) 8535. [7] M. J. Hudson, S. Carlino and D.C. Apperley, J. Mater. Chem., 5 (1995) 323. [8] L. Li, S. Ma, X. Liu, Y. Yue, J. Hui, R. Xu, Y. Bao and J. Rocha, Chem. Mater., 8 (1996) 204. [9] S. Velu, V. Ramkumar, A. Narayanan and C. S. Swamy, J. Mater. Sci., 32 (1997) 957. [ 10] F.M. Vichi and O.L. Alves, J. Mater. Chem., 7 (1997) 1631. [11] S. Kannan and C.S. Swamy, J. Mater. Sci., 32 (1997) 1623. [12] M. Sato, H. Kuwabara and S. Sato, Clays Sci., 8 (1992) 309. [13] T. Hibino, Y. Yamashita, K. Kosuge, A. Tsunashima, Clays Clay Miner., 43 (1995) 427. [14] J. Santhanalakshmi and T. Raja, Appl. Catal., A147 (1996) 69. [ 15] F. Kooli, V. Rives and M. A. Ulibarri, Inorg. Chem., 34 (1995) 5122. [16] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, 5th ed., sers. A and B, John Wiley & Sons, New York, 1997. [ 17] F. A. Cotton, Chemical Application of Group Theory, 2nd ed., John Wiley & Sons, New York, 1971, chap. 10. [ 18] E. M. Moujahid, J. Besse and F. Leroux, J. Mater. Chem., 13 (2003) 258. [ 19] V. Prevor, C. Forano and J. Besse, Appl. Clay Sci., 18 (2001) 3. [20] S. Velu, V. Ramkumar, A. Narayanan and C.S. Swany, J. Mater. Sci., 32 (1997) 957. [21 ] C. Depege, F. E1Metoui, C. Forano, A. de Roy, J. Duouis and J. Besse, Chem. Mater., 8(1996)952. [22] M.J. Hudson, S. Carlino and D.C. Apperley, J. Mater. Chem., 5 (1995) 323 [23] W.T. Reichele, S.Y. Kang and D.S. Everhardt, J. Catal., 101 (1986) 352. [24] S. Kannan and C.S. Swamy, J. Mater. Sci., 32 (1997) 1323. [25] M.A. Drezdzon, Inorg. Chem., 27 (1988) 4628. [26] E. Kanezaki, K. Kinugawa, and Y. Ishikawa, Chem. Phys. Lett., 226 (1994) 325. [27] K. Chibwe and W. Jones, J. Chem. Soc. Chem. Commum., (1989) 926. [28] I.Y. Park, K. Kuroda, and C. Kato, Chem.Lett., (1989) 2057. [29] K.R. Franklin, E. Lee and C.C. Nunn, J. Mater. Chem., 5 (1995) 565. [30] M. Meyn, K.Beneke, and G. Lagaly, Inorg. Chem., 29 (1990) 5201. [31] A. Streitwieser, Jr., Molecular Orbital Theory for Organic Chemists, Wiley, New York, (1961), Chap. 8. [32] D. Briggs and M.P. Seah, Practical Surface Analysis-Auger and X-ray Photoelectron Spectroscopy, J. Wiley & Son, New York (1982). [33] R. Allman, Chimica, 24 (1970) 99; ibid, Acta Crystallogr., B24 (1968) 972. [34] D.D. Elleman and D. Williams, J. Chem. Phys., 25 (1956) 742. [35] E. Kanezaki, J. Incl. Phen., 34 (2000) 447. [36] E. Kanezaki, Mater. Res. Bull., 34 (1999) 1435.

Preparation of Layered Double Hydroxides

373

[37] MgSO4; JCPDS card No. 21-0546. [38] E. Kanezaki, Inorg. Chem., 37 (1998) 2588. [39] E. Kanezaki, J. Mater. Sci. Lea., 17 (1998) 371. [40] E. Kanezaki, Solid State Ionics, 106 (1998) 279. [41] E. Kanezaki, Mater. Res. Bull., 33 (1998) 773. [42] V. Rives, Inorg. Chem., 38 (1999) 406. [43] E. Kanezaki, Mol. Cryst. Liq. Cryst., 286 (1996) 153. [44] E. Kanezaki, N. Nishi and M. Kinoshita, Bull. Chem. Soc. Jpn., 52 (1979) 2836. [45] E. Kanezaki, J. Mater. Sci., 30 (1995) 4926. [46] V.R.L. Constantino and T.J. Pinnavaia, Inorg. Chem., 34 (1995) 883. [47] E. Kanezaki, J. Incl. Phen., 46 (2003) 89. [48] E. Kanezaki, J. Incl. Phen., 24 (1996)41. [49] E. Kanezaki, unpublished result. [50] E. Kanezaki et al, Mater. Res. Bull., submitted.