Infrared Spectra of Oriented and Nonoriented Layered Double

We have extended the experimental examination of layered double hydroxides of the types Mg2Al(OH)6X and Mg3Al(OH)8X, both as isotropic pellets and as ...
0 downloads 0 Views 250KB Size
Download PDF
J. Phys. Chem. C 2007, 111, 4209-4215

4209

Infrared Spectra of Oriented and Nonoriented Layered Double Hydroxides in the Range from 4000 to 250 cm-1, with Evidence for Regular Short-Range Order in a Synthetic Magnesium-Aluminum LDH with Mg:Al ) 2:1 but Not with Mg:Al ) 3:1 Mickey C. Richardson and Paul S. Braterman* Departments of Chemistry and Materials Science, UniVersity of North Texas, Denton, Texas 76203 ReceiVed: July 25, 2006; In Final Form: December 16, 2006

We have extended the experimental examination of layered double hydroxides of the types Mg2Al(OH)6X and Mg3Al(OH)8X, both as isotropic pellets and as oriented supported specimens, to cover the entire range from 4000 to 250 cm-1 by the use of CsI pellets and CdTe and polyethylene supports and compared our results with theoretical expectations. Close inspection of the group theoretically predicted vibrational modes for these materials shows that formal allowedness, while necessary, is not sufficient for spectroscopic intensity. The OH z-translational mode occurs around 700 cm-1, while the x,y-translations and the OH rotation (MOH bend) cover a broad region from around 1000 cm-1 down. The hydroxide O-H stretching mode, which is suppressed by orientation, is concentrated at the higher end of the total O-H stretching region. Aged material of the type Mg2Al(OH)6X shows reproducible structure around 400 cm-1 both when X is chloride and when X is 1/4ferrocyanide. This structure is weak or absent in freshly prepared Mg2Al(OH)6Cl and in both fresh and aged Mg3Al(OH)8X, strongly suggesting that local superlattice formation, required for charge avoidance when M(II):M(III) ) 2:1 but not when M(II):M(III) ) 3:1, is specific to the 2:1 material. We regard the inferred structural differences between fresh and aged Mg2Al(OH)6Cl as evidence for ripening by a solutionreprecipitation mechanism.

Introduction Layered double hydroxides (LDHs) are a class of natural and synthetic mixed metal hydroxides with the general formula [MII1-xMIIIx(OH)2]x+[Y]x/nn-‚zH2O,1,2 although LDHs of the type [LiAl2(OH)6Y1/nn- are also readily prepared.3 The range of divalent to trivalent metal ions varies, but the most commonly studied types have a 2:1 or 3:1 ratio. The structure of LDHs closely resembles brucite, but the incorporation of trivalent metal ions causes a net positive charge, which has to be countered by anions to attain electroneutrality. The metal hydroxide sheets are held together through hydrogen bonding by both interlayer water and anions. These materials have found wide applications as catalysts, polymer additives, and catalyst and ceramic precursors and, more recently, in novel composites4 and as pharmaceutical controlled release agents.5 Given their widespread occurrence in nature, they have also been suggested as possible information stores6 and as catalysts and concentrating agents7 in the processes leading up to the origins of life on earth. The most common LDHs, synthetic or natural, have a rhombohedral stacking sequence (3R1), although LDHs can also occur in nature with hexagonal (2H1) arrangements,8,9 and a 3R2 synthetic hydrotalcite has been described.10 One question of general interest for each of these materials is the extent of order versus disorder at several levels. EXAFS11 has been used to obtain information about the nearest and next nearest neighbors of metal ions in LDHs, with implications about the existence or otherwise of short-range order. Simple orderly stacking of layers one above the other gives rise to a range of possible polytypes, distinguishable from each other by X-ray * To whom correspondence should be addressed. Phone: (940) 5652357. Fax: (940) 565-4318. E-mail: [email protected].

diffraction, as discussed by Drits and Bookin.9 In a few cases, X-ray diffraction also reveals orderly arrangements of cations, and sometimes of anions, in mineral specimens, giving rise to superlattice spacings.12 It is also possible for such superlattice spacings to arise in synthetic material from orderly arrangements of the interlayer anions,13 but it is not possible to say whether such order is also present in the cation distribution. Unfortunately, Mg(II) and Al(III), which give rise to the most common LDHs, both natural and synthetic, are virtually indistinguishable by X-ray techniques. While X-ray diffraction will certainly remain the principal method for characterization of LDHs, there is growing interest in the application of vibrational spectroscopy to LDHs14 and to layer minerals in general.15 In this paper, we argue that infrared spectroscopy can give insight, not obtainable by other means, into structural details of LDHs and in particular into the question of cation order. We demonstrate that useful information is obtainable in the generally neglected region between 400 and 250 cm-1, now accessible using moderately priced commercial instruments with cesium iodide optics. Finally, we report on ways of extending our technique16 of oriented infrared spectroscopy, previously limited to the range above 800 cm-1 by the properties of our support material, to the full wavelength region and demonstrate its use as an aid in vibrational assignment. As sample materials, we use magnesium aluminum LDHs, with Mg:Al ratios of 2:1 and 3:1, both fresh (as initially prepared) and aged (annealed17 to enhance crystallinity). As anions, we use chloride (because of its absence of internal vibrations, which could complicate the analysis) and ferrocyanide (because its internal vibrations,18 and the effect of the orientation on them when incorporated into LDHs,16 are already well understood).

10.1021/jp064744w CCC: $37.00 © 2007 American Chemical Society Published on Web 02/28/2007

4210 J. Phys. Chem. C, Vol. 111, No. 11, 2007

Richardson and Braterman

TABLE 1: Metals Analysis LDH sample

[Mg], %

[Al], %

fresh 21 Mg-Al LDH-Cl aged 21 Mg-Al LDH-Cl fresh 31 Mg-Al LDH-Cl aged 31 Mg-Al LDH-Cl fresh 21 Mg-Al LDH-ferro aged 21 Mg-Al LDH-ferro fresh 31 Mg-Al LDH-ferro aged 31 Mg-Al LDH-ferro

15.85 15.34 23.50 23.82 15.41 14.63 23.06 23.29

9.95 8.52 6.21 6.18 8.32 8.29 6.55 6.37

Experimental Section General Procedures. Solutions were prepared and materials washed using fresh deionized water (∼18 MΩ/cm) from a Millipore Milli-Q Plus water purification system. K4Fe(CN)6‚ 3H2O, MgCl2‚6H2O, and AlCl3‚6H2O (Sigma-Aldrich) and 50% (w/w) NaOH (Alfa Aesar) were used as supplied. The metals were analyzed by atomic absorption spectroscopy using a Perkin-Elmer AAnalyst 300 spectrometer with standards provided by Perkin-Elmer. An air-acetylene flame was used for Mg and Fe and N2O-acetylene for Al. Synthesis. The parent LDH-Cl materials were prepared, under a steady stream of nitrogen gas to prevent adventitious CO2 uptake, by the standard technique of addition of stoichiometric amounts of 50% NaOH (Alfa Aesar) (6 mol of OH- for every mole of Al3+ for a 2:1 LDH and 8 mol of OH- for every mole of Al3+ for a 3:1 LDH) to a solution 0.1 M in AlCl3 and either 0.2 or 0.3 M in MgCl2. The fresh LDH was collected after 1 h of stirring, while the aged materials were washed once by centrifuge and gently refluxed in water for 24 h. Both fresh and aged materials were thoroughly washed by centrifuge. Materials for exchange with ferrocyanide were used as prepared; materials for spectroscopic examination were dried in a vacuum desiccator over a mixture of Drierite and molecular sieves. For the LDH-ferrocyanide samples, 1 g of fresh or aged LDH chloride (as calculated from the conditions of preparation) was suspended in 25 mL of water and the calculated amount of potassium ferrocyanide (i.e., ferrocyanide:Al ) 1:4) added in 25 mL of water. Use of excess ferrocyanide led to the formation of the cubic material that we have described previously.19 The materials were stirred under nitrogen for 1 h, thoroughly washed by centrifuge, and dried in a vacuum desiccator over a mixture of Drierite and molecular sieves. Spectroscopic Methods. Infrared spectra were obtained using a Perkin-Elmer Spectrum One spectrometer, with an internal CsI beam splitter and automatic water and CO2 absorption correction. Supporting and reference materials were of infrared grade and supplied through (KBr) Alfa Aesar, (CsI) Wilmad, (CdTe disks) New Era Enterprises, and (polyethylene ST-IR cards) Thermo Electron Corp. All spectra were averaged over 40 scans at a resolution of 4 cm-1 and are shown normalized in absorbance mode to an ordinate maximum of 1.0 for consistency of interpretation. KBr/CsI (Pellet) FT-IR Samples. The conventional (pellet) IR spectra were obtained using KBr and CsI against disks consisting of either KBr or CsI alone (0.2000 g of KBr or 0.2500 g of CsI, respectively) as a reference. The samples were prepared by measuring out approximately 1% (by mass) sample (with respect to the background mass) and then adding either KBr or CsI to come as close to the original background mass as possible, since we have found that differences as small as 0.5 mg between the sample disk mass and background disk mass are enough to introduce spurious background features into the sample spectrum. The spectral scans were performed from 4000 to 400 cm-1 for KBr and from 4000 to 250 cm-1 for CsI.

[Fe], %

Mg/Al

Al/Fe

Mg/Fe

4.02 4.39 3.63 3.11

1.8 2.0 2.7 2.8 2.1 2.0 3.1 3.2

4.3 3.9 3.1 3.5

8.80 7.67 9.46 11.16

CdTe/Polyethylene (Oriented) FT-IR Samples. Oriented IR spectra were obtained using CdTe and polyethylene supports, chosen for their insolubility in water and transparency over a useful range of wavelengths. For each CdTe sample, approximately three to five drops of a dilute LDH suspension were placed on top of each disk and allowed to dry, at room temperature, in a sealed evacuated desiccator. If the infrared spectrum was too weak, more material was added. The polyethylene samples were prepared in a similar fashion, but more sample (seven to ten drops) seemed to be required. Infrared absorbances will appear in the pellet samples regardless of polarization, while, to the extent that the material is aligned as horizontal platelets on the support, only in-planepolarized absorption bands will appear in the oriented spectra. This selection effect is well-established for the ferrocyanide derivatives,16 and the similarity reported here between orientation effects in the regions of the spectra common to the chloride and ferrocyanide materials implies a comparable degree of alignment. We have in addition verified by SEM that the ferrocyanide-exchanged materials share the same familiar hexagonal platelet habit. Results and Discussion Preparation of Materials. All the materials discussed in this work are thoroughly familiar, and the elemental analyses (Table 1) are within the range normally found. The only feature of interest is the relatively high uptake of ferrocyanide, especially on the fresh materials. This finding, if real, could indicate surface adsorption beyond the ideal stoichiometry. Comparison of Experimental Conditions. We were pleased with the performance of polyethylene films around 2000 cm-1 and below 600 cm-1. In the 3000-4000 cm-1 region, especially with fresh material (with which it was more difficult to achieve total coverage), and also in comparison with air, they gave rise to a weak modulation which we attribute to imperfectly compensated interference effects, strongest near an absorption band (Figure 1). We therefore confine our use of polyethylene to the regions around 2000 cm-1 and below 600 cm-1.

Figure 1. Spectrum of a polyethylene film against a polyethylene film reference. Note incompletely compensated peaks and interference modulation.

IR Spectra of Oriented and Nonoriented LDHs

Figure 2. Infrared spectrum (2000-2200 cm-1) of Mg2Al(OH)6‚1/4Fe(CN)6‚xH2O from an aged chloride precursor: (A) CdTe support, (B) polyethylene support, (C) KBr pellet, (D) CsI pellet.

For ordinary pellet transmission work, potassium bromide would seem to be preferred. Cesium iodide must be used below 400 cm-1, but gives rise to major distortions in the observed spectra in the neighborhood of strong, sharp peaks (Figure 2). We would also draw attention to its higher cost, greater tendency to adsorb water, and greater chemical reactivity under the conditions of localized pressure and temperature that take place during pellet formation. Cadmium telluride pellets were suitable substrates down to 400 cm-1, but are expensive, toxic, and easily damaged. Attempts to prepare our own cadmium telluride pellets from powder in an ordinary pellet press were unsuccessful. Group Theoretical Analysis and Its Limitations. Published treatments20,21 of the hydroxide layer vibrations of the systems Mg2Al(OH)6 and Mg3Al(OH)8 assume an idealized orderly arrangement, with factor group D3d. The factor group analysis proceeds in much the same way as the familiar point group analysis of individual molecules, except that translationally equivalent atoms are treated as formally identical. Those motions that correspond to an overall translation or rotation of the unit cell are discarded and the number of allowed modes (and, implicitly, their orientation, although previous authors had no reason to comment on this) compared with the observed spectra. This procedure can be regarded as an extension of that applied by Mitra in a classical review paper22 to magnesium hydroxide, and the modifications that follow are closely related to those that we have offered, for magnesium hydroxide,23 on the basis of more extensive vibrational mode analysis and molecular dynamics simulations.

J. Phys. Chem. C, Vol. 111, No. 11, 2007 4211 The use of the factor group presupposes an orderly arrangement, forming a superlattice, of the two different kinds of metal, although this need not be sufficiently extensive to give rise to superlattice spacings in the X-ray diffraction pattern. Such order is bound to arise locally in systems of the type MII2MIII(OH)6, if M(III)-M(III) nearest metal neighbor approaches are to be avoided, but the same is not true for systems of the type MII3MIII(OH)8. As we shall see below, there is good evidence from our spectra for local symmetry in annealed samples of the former type, but not in very freshly prepared samples, nor in any samples of the latter type. The factor group treatment, by its nature, finds only those vibrations in which all unit cells are moving in phase, what is sometimes referred to as the k ) 0 member of the Brillouin zone associated with each mode. As we have shown elsewhere,23 this neglects motions in which groups in neighboring unit cells are moving more or less out of phase with each other, giving rise to motions of higher frequency than the k ) 0 member. In magnesium hydroxide, these modes are inactive in infrared and Raman spectroscopy, but active in neutron scattering. In LDHs, there is the added complication24 that the interlayer anions and water molecules form a dynamic system of no particular symmetry, hydrogen bonded to each other and to the lattice hydroxides, and molecular dynamics simulations and normalmode analysis using the CLAYFF force field25 show that this has the effect26 of localizing the hydroxide modes and of undermining the k ) 0 selection rule and its corollary, the predicted absence from the spectrum of modes corresponding to the translation or rotation of the entire chosen unit cell. For completeness, we give the results of formal factor group analysis, classified by the group or atom motions principally involved, in Tables 2 and 3. Of course, different modes of the same symmetry will mix, especially if they involve the same kind of motion. It is worth noting that several formally symmetry-allowed modes involving the hydroxide groups are expected to be, in reality, vanishingly weak. This is because these groups reside on sites of less than 3-fold symmetry, so that some combinations of z-direction motions are formally allowed to have components in the x,y-plane and vice versa. Infrared Spectra. In all cases, as expected, the H2O bending mode (1620-1640 cm-1)27 did not show any major orientation effects. This peak remains consistent in all samples and will not be further discussed. CN Stretching Region of LDH Ferrocyanides. We consider first the aged materials Mg2Al(OH)6‚1/4Fe(CN)6‚xH2O as a test of our methodology. Here our earlier work using barium fluoride as a support had established16 that the T1u infrared-active mode is split into two components, a slightly sharper band at 2034 cm-1, absent in the oriented spectrum, and a broader band at 2045 cm-1 that is still present on orientation. We interpreted

TABLE 2: Formal Analysis of the Motions of the Mg2Al(OH)6 Unit, Factor Group D3d group or atom

type of motion

effect on bonding

irreducible representations spanned and predicted activitya

OH OH OH OH Al Al Mg Mg

stretch z-axis motion x,y-axis motion rotation z-axis motion x,y-axis motion z-axis motion x,y-axis motion

O-H stretch M-O-M stretch/bend M-O-M bend/stretch M-O-H bend M-O-M stretch/bend M-O-M bend/stretch M-O-M stretch/bend M-O-M bend/stretch

A1g(R,pol) + A2u(IR;z) + Eg(R,depol) + Eu(IR;x,y;f) A1g(R,pol) + A2u(IR;z)b + Eg(R,depol)b + Eu(IR;x,y;f) A1g(R,pol) + A1u + A2gb + A2u(IR;z,f) + 2Eg(R,depol) + 2Eu(IR;x,y)b A1g(R,pol) + A1u+ A2g + A2u(IR;z,f) + 2Eg(R,depol) + 2Eu(IR;x,y) A2u(IR;z)b Eu(IR;x,y)b A1g(R,pol) + A2u(IR;z)b Eg(R,depol) + Eu(IR;x,y)b

a Key: R, Raman-allowed; pol, polarized; depol, depolarized; z, z-axis (c-direction) polarized; x,y, x,y-axis (a,b-plane) polarized; f, formally (group-theoretically) allowed, but expected to be vanishingly weak (see the text). b Also contributes significantly to unit cell translation or rotation (total A2u + Eu + A2g + Eg), spectroscopically inactive.

4212 J. Phys. Chem. C, Vol. 111, No. 11, 2007

Richardson and Braterman

TABLE 3: Formal Analysis of the Motions of the Mg3Al(OH)8 Unit, Factor Group D3d group or atom

type of motion

effect on bonding

irreducible representations spanned and predicted activitya

OH(1)b OH(2)b OH(1) OH(2) OH(1)

stretch stretch z-axis motion z-axis motion x,y-axis motion

O-H stretch O-H stretch M-O-M stretch/bend M-O-M stretch/bend M-O-M bend/stretch

OH(2) OH(1) OH(2) Al Al Mg Mg

x,y-axis motion rotation rotation z-axis motion x,y-axis motion z-axis motion x,y-axis motion

M-O-M bend/stretch M-O-H bend M-O-H bend M-O-M stretch/bend M-O-M bend/stretch M-O-M stretch/bend M-O-M bend/stretch

A1g(R,pol) + A2u(IR;z) + Eg(R,depol) + Eu(IR;x,y;f) A2g + A2u(IR;z) A1g(R,pol) + A2u(IR;z)c + Eg(R,depol)c + Eu(IR;x,y;f) A2g + A2u(IR;z)c A1g(R,pol) + A1u + A2gc + A2u(IR;z,f) + 2Eg(R,depol) + 2Eu(IR;x,y)c Eu(IR;x,y)c + Eg(R,depol) A1g(R,pol) + A1u + A2g + A2u(IR;z,f) + 2Eg(R,depol) + 2Eu(IR;x,y) Eu(IR;x,y) + Eg(R,depol) A2u(IR;z)c Eu(IR;x,y)c A2u(IR,z)c + Eu(IR;x,y;f) A1u + A2u(IR;z;f) + 2Eu(IR,x,y)c

a Key: R, Raman-allowed; pol, polarized; depol, depolarized; z, z-axis (c-direction) polarized; x,y, x,y-axis (a,b-plane) polarized; f, formally (group-theoretically) allowed, but expected to be vanishingly weak (see the text). b OH(1) bridges two Mg atoms and one Al atom. OH(2) bridges three Mg atoms. OH(1) and OH(2) modes of the same symmetry will be strongly mixed, especially when derived from the same type of motion. c Also contributes significantly to unit cell translation or rotation (total A2u + Eu + A2g + Eg), spectroscopically inactive.

Figure 3. Infrared spectrum (2000-2200 cm-1) of Mg2Al(OH)6‚1/4Fe(CN)6‚xH2O from a fresh chloride precursor: (A) CdTe support, (B) polyethylene support, (C) KBr pellet, (D) CsI pellet.

this finding as consistent with D3d site symmetry for the ferrocyanide, which will split T1u (Oh) into A2u (z-oriented) and Eu (x,y-oriented) components. Figure 2 compares the spectra obtained with a CdTe support, polyethylene support, KBr disk, and CsI disk in this region. The suppression of the 2034 cm-1 band in the supported materials shows that orientation has indeed taken place and establishes the suitability of both these supports for our work. The spectrum in CsI shows a new broad peak around 2094 cm-1, clearly indicative of reaction of some kind, together with an apparent shift of the authentic ferrocyanide feature to slightly higher frequency and a minimum at its low-wavenumber edge. We regard the latter effects as optical artifacts, perhaps connected with the high refractive index of CsI, and draw the lesson that sharp peaks in CsI should be evaluated with care. Figure 3 shows the spectrum in this region of fresh 2:1 LDH chloride, exchanged with ferrocyanide. In this case, the separate components are not resolved, and, somewhat remarkably, the ordering of A2u and Eu components appears to be reversed. We infer (a) that the interaction of the CN stretching motion with its environment (presumably by hydrogen bonding to layer OH) is mildly direction-dependent and (b) that the distinction between fresh and aged materials survives anion exchange at room temperature.

Figure 4. Infrared spectrum (1100-250 cm-1) of aged Mg2Al(OH)6Cl‚ xH2O: (A) CdTe support, (B) polyethylene support (asterisked peaks are artifacts; compare with Figure 1), (C) KBr pellet, (D) CsI pellet.

OH Bend-MOH Lattice Region, 1100-250 cm-1. There are striking differences between fresh and aged 2:1 Mg-Al LDHCl in this region and significant orientation effects. The most obvious difference between these two materials is within the 1100-250 cm-1 range (Figures 4 and 5). The aged material in CsI shows a sharp peak at 447 cm-1, which is weak or on occasion absent in the fresh material, together with a sharp peak at 390 cm-1; the 447 cm-1 peak is also present in the KBr pellet spectrum, as expected. The fresh material in CsI or on polyethylene shows a broad peak with a maximum around 395 cm-1. In addition, all the spectra show broad features extending from low frequencies to 1000 cm-1 or beyond, with defined peaks at around 560 and 1025 cm-1 in the aged materials only, and all nonoriented specimens show a broad absorption around 700 cm-1, absent in the oriented specimens. We interpret these facts as follows. The 447 cm-1 band is diagnostic of lattice ordering, which is required if the system is to avoid having tervalent cations as each other’s nearest neighbors (compare Lowenstein’s rule28 for aluminosilicates). This supports the suggestion that a disorderly as-formed material matures by a solution-reprecipitation process, since it is not clear how Mg and Al could change places within an intact lattice. However, even this initially formed material must show sufficient crystallinity for orientation effects to be significant.

IR Spectra of Oriented and Nonoriented LDHs

Figure 5. Infrared spectrum (1100-250 cm-1) of fresh Mg2Al(OH)6Cl‚ xH2O: (A) CdTe support, (B) polyethylene support (the question mark indicates a possible artifact; compare with Figure 1), (C) KBr pellet, (D) CsI pellet.

Although OH rotation (MOH bending) and OH x,y-translation (mainly MO bending/stretching) modes are allowed to mix, we would suggest that the broad, orientation-independent absorption corresponds mainly to the former and that this is cofirmed by the existence of inelastic neutron scattering modes found in this region and regarded21 as diagnostic of motions that correspond to large displacements of hydrogen atoms. They are closely related to similar modes in Mg(OH)223,29 except that (as confirmed by molecular dynamics models24,26) hydrogen bonding to interlayer water causes loss of microscopic symmetry and allows spectroscopic activity. The minor differences between nominally similar spectra may be caused by small changes in water content. Absorptions due to A2u modes are selectively suppressed by orientation, and three such modes are formally predicted by the group-theoretical treatment (Table 1); however, only one of these derives from a motion that will in reality convey z-oriented intensity. We therefore assign the z-oriented absorption around 700 cm-1 to this mode, the OH z-direction translation with its assocated M-O stretching and MOM bending. There remain the sharp peaks at 447 and 397 cm-1 in the aged materials and at 390 cm-1 in the fresh LDH. We assign these to the two allowed Eu modes derived from OH x,ytranslational (M-O-M bend stretching) motions. The existence of two such modes within the same unit cell is a result of coupling between groups of the same type, i.e., OH groups each with one Al and two Mg neighbors. We then attribute the fact that the band is broader, and unsplit, in fresh LDH to the absence of a superlattice, which will in this case imply the existence of OH groups in a near continuum of slightly different environments. In the region below 1000 cm-1, the ferrocyanides echo the behavior of their parent chlorides, with the expected addition in both film and pellet spectra of the known18 M-C-N bending mode at 590 cm-1 (literature value 583 cm-1 in water) and a band at 560 cm-1 of unknown origin, perhaps related to the effect of the M-C-N bend on the near continuum of x,ypolarized motions. The MC stretching mode expected at 416 cm-1 is not detectable against the strong layer absorption in this region, but the suppression of the above-mentioned 700 cm-1 (OH z-motion) peak in the oriented spectra is particularly clear.

J. Phys. Chem. C, Vol. 111, No. 11, 2007 4213

Figure 6. Infrared spectrum (1100-250 cm-1) of aged Mg3Al(OH)8Cl‚ xH2O: (A) CdTe support, (B) polyethylene support (asterisked peaks are artifacts; compare with Figure 1), (C) KBr pellet, (D) CsI pellet.

Figure 7. Infrared spectrum (1100-250 cm-1) of fresh Mg3Al(OH)8Cl‚ xH2O: (A) CdTe support, (B) polyethylene support (asterisked peaks are artifacts; compare with Figure 1), (C) KBr pellet, (D) CsI pellet.

In stark contrast to the 2:1 materials, the fresh and aged 3:1 materials (Figures 6 and 7) show little difference in any frequency range. The features around 450-380 cm-1 in the 2:1 materials are in all cases replaced by a single broad peak close to 400 cm-1. This could conceivably be the result of overlap between the absorption bands of the two distinct types of OH present (Mg3-coordinated vs Mg2Al-coordinated) in a wellordered structure, which seems to us unlikely, especially in fresh LDH. On the contrary, we note that, in 3:1 material, it is possible to avoid nearest metal neighbor Al-Al approaches without imposing any regularity of structure and infer that the aged 3:1 material, as well as the fresh, lacks superlattice ordering. In this case, the range of different OH environments will be large, and the symmetry restrictions of Table 2 will not apply; thus, we expect a wide range of slightly different frequencies, much as in the fresh 2:1 material. Slightly more structure is apparent in the oriented than in the nonoriented spectra, a fact that we very tentatively attribute to suppression of the OH z-axis motion in the former. OH Stretching Region, 3000-4000 cm-1. For the 2:1 materials, all the spectra show strong, broad absorption in the range 3600-3200 cm-1 (Figures 8 and 9), with subtle but consistent differences between the oriented and the pellet spectra. The

4214 J. Phys. Chem. C, Vol. 111, No. 11, 2007

Figure 8. Infrared spectrum (3000-4000 cm-1) of aged Mg2Al(OH)6Cl‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

Figure 9. Infrared spectrum (3000-4000 cm-1) of fresh Mg2Al(OH)6Cl‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

pellet spectra have their intensity at somewhat higher frequency, and in the aged chloride material, there is evidence for a resolvable band around 3500 cm-1, especially in the oriented specimens. We therefore relate the 3500 cm-1 band and the low-end intensity to the antisymmetric and symmetric OH stretching modes of interlayer water, which in liquid water occur at 3490 and 3280 cm-1. These vibrations are expected to be largely confined to the x,y-plane,24 leading us to locate the OH stretch (which is, of course, predominantly z-polarized) toward the higher frequency end of the feature, perhaps with intensity concentrated around 3450 cm-1, although it will presumably be broadened by local fluctuations in H-bonding. We also note an overall reduction in the width of the absorption on aging, suggesting that the cation ordering also, indirectly, affects the interlayer water. For the 2:1 ferrocyanides (Figures 10 and 11), the differences between fresh and aged materials, and between oriented and pellet spectra, are qualitatively similar to those found for the chloride, and we invoke similar explanations. We draw attention to the small sharp resolved in-plane absorption at 3620 cm-1. We ascribe this to interstitial water in an environment with very little hydrogen bonding, a situation more readily achieved with ferrocyanide than with chloride. The frequency is too low for brucite or gibbsite contamination, and in any case the in-plane orientation excludes assignment to layer hydroxide. Note also that aging leads, as it did with 2:1 chloride, to a general reduction in the width of the OH stretching absorption In the 3:1 materials as well (Figures 12 and 13), the oriented spectra in this region carry more intensity at lower frequencies. We offer the same explanation as we did in the 2:1 case, that the layer OH stretch is toward the higher frequency end of the range spanned by water OH. Again, we find less structure, and

Richardson and Braterman

Figure 10. Infrared spectrum (3000-4000 cm-1) of aged Mg2Al(OH)6‚ 1 /4Fe(CN)6‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

Figure 11. Infrared spectrum (3000-4000 cm-1) of fresh Mg2Al(OH)6‚ 1 /4Fe(CN)6‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

Figure 12. Infrared spectrum (3000-4000 cm-1) of aged Mg3Al(OH)8Cl‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

less difference between fresh and aged materials, than in the case of 2:1 LDH, consistent with our suggestion that the development of spectroscopically significant structure on aging is limited to the 2:1 materials. Structural and Methodological Implications. It is clear that all samples, including the fresh material, show sufficient anisotropy for orientation effects to be observable. It is also clear that, at least in the case of the 2:1 materials, an initially less orderly structure is converted to one that imposes clearer constraints on the infrared spectrum. For the reasons given above, we interpret this as the formation of a x3×x3 superlattice, required at this charge ratio for M(III)-M(III) nearest neighbor avoidance. We would suggest that such a change in cation distribution is unlikely to take place by migration within existing layers and is therefore more likely to arise by a solution-reprecipitation process. In any case, our work should be compared with that of Xu et al., who have

IR Spectra of Oriented and Nonoriented LDHs

J. Phys. Chem. C, Vol. 111, No. 11, 2007 4215 (OH)6Cl‚xH2O and its ferrocyanide derivatives, aging leads to a reduction in bandwidth in the OH stretching region, indicating that the cation order in the aged material also influences the interlayer water. Acknowledgment. We thank the Welch Foundation (Grant B-1445) for support and Dr. R. T. Cygan for helpful discussions. References and Notes

Figure 13. Infrared spectrum (3000-4000 cm-1) of fresh Mg3Al(OH)8Cl‚xH2O: (A) (offset) CdTe support, (B) KBr pellet.

shown30 that the formation of mature LDH is not a simple single-step process. It is also of interest to compare our results with those of Wang, Kirkpatrick, and colleagues,24,31 who have investigated the low-frequency region of a wide range of layered materials, including 3:1 Mg-Al LDH. We agree with their reporting for this stoichiometry a single absorption band close to 400 cm-1, across a range of anions, and their assignment of this band to a motion involving layer hydroxide oxygen. Our finding that ordered and disordered phases show significant differences in this region may help explain why their experimental spectra (obtained for material that is presumably disordered) show less structure than their calculated power spectra, which were determined for model material showing cation order. We would further suggest that since it is now possible to distinguish experimentally between z-polarized and x,y-polarized motions, it may be of value to consider separately the zz and xx (or yy) autocorrelation power spectra and also to compare the spectra for hydroxide oxygen in this region with those for hydroxide hydrogen to distinguish more effectively between OH translational and librational motions. Conclusions Extension of the region scanned down to 250 cm-1 and the use of oriented specimens both give novel information about band assignments and about the degree of order in the materials studied. Thus, it is possible to distinguish among OH x,y-motions around 400 cm-1, OH rotations (MOH bending modes) across the entire region below 1000 cm-1, and the OH z-oriented motion around 700 cm-1. In addition, it is apparent that aged Mg2Al(OH)6Cl‚xH2O shows a degree of ordering absent in the fresh material, which we attribute to the achievement of a local superlattice structure, as required for this ratio by an Al(III)Al(III) neighbor avoidance rule. Both fresh and aged Mg3Al(OH)8Cl‚xH2O show spectra very similar to those of fresh Mg2Al(OH)6Cl‚xH2O, and we infer that in this case the materials achieve Al(III)-Al(III) neighbor avoidance without developing a regular repeat structure. In the OH stretching region, intensity is in all cases concentrated toward lower frequency in the oriented spectra, indicating that the lattice hydroxide absorption is concentrated more toward higher frequencies. In Mg2Al-

(1) Layered Double Hydroxides: Present and Future; Rives, V., Ed.; Nova Science Publishers: Huntington, NY, 2001. (2) Braterman, P. S.; Xu, Z. P.; Yarberry, F. In Handbook of Layered Materials; Auerbach, S. M., Carrado, K. A., Dutta, P. K., Eds.; Marcel Dekker: New York, 2004; p 373. (3) Leroux, F; Taviot-Gueho, C. J. Mater. Chem. 2005, 15, 3628. (4) Sissoko, I.; Iyagba, E. T.; Sahai, R.; Biloen, P. J. Solid State Chem. 1985, 60, 283. (5) See, e.g.: Kriven, W. M.; Kwak, S. Y.; Wallig, M. A.; Choy, J. H. MRS Bull. 2004, 29, 33. (6) Greenwell, H. C.; Coveney, P. V. Origin Life EVol. Biosphere 2006, 36, 13. (7) Boclair, J. W.; Braterman. P. S.; Brister, B. D.; Jiang, J., Lou, S., Wang, Z., Yarberry, F. Origin Life EVol. Biosphere 2001, 23, 53. (8) Allmann, R. Chimia 1970, 24, 99. (9) Drits, V. A.; Bookin, A. S. In Layered Double Hydroxides: Present and Future; Rives, V., Ed.; Nova Science Publishers: Huntington, NY, 2001; p 39. (10) Newman, S. P.; Jones, W.; O’Conner, P.; Stamires, E. N. J. Mater. Chem. 2002, 12, 153. (11) See, e.g.: Roussel, H.; Briois, V.; Elkaim, E.; de Roy, A.; Besse, J. P.; Jolivet, J. P. Chem. Mater. 2001, 13, 329. (12) Solin, S. A.; Hines, D. R.; Seidler, G. T.; Treacy, M. M. J. J. Phys. Chem. Solids 1996, 57, 1043-1048. (13) Vucelic, M.; Moggridge, G. D.; Jones, W. J. Phys. Chem. 2005, 99, 8328. (14) Kloprogge, J. T.; Frost, R. L. In Layered Double Hydroxides: Present and Future; Rives, V., Ed.; Nova Science Publishers: Huntington, NY, 2001; p 139. (15) The Application of Vibrational Spectroscopy to Clay Minerals and Layered Double Hydroxides; Kloprogge, J. T., Ed.; CMS Workshop Lectures, Vol. 13; The Clay Minerals Society: Chantilly, VA, 2005. (16) Braterman, P. S.; Tan, C.; Zhao, J. Mater. Res. Bull. 1994, 29, 1217. (17) Yarberry, F. Layered Double Hydroxides via Aluminate. Ph.D. Dissertation, University of North Texas, Denton, TX, 2002. (18) Jones, L. H. Inorg. Chem. 1963, 2, 777. (19) Boclair, J. W.; Braterman. P. S.; Brister, B. D.; Wang, Z.; Yarberry, F. J. Solid State Chem. 2001, 161, 249. (20) Hernandez-Moreno, M. J.; Ulibarri, M. A.; Rendon, J. L.; Serna, C. J. Phys. Chem. Miner. 1985, 12, 34. (21) Kagunya, W; Baddour-Hadjean, R.; Kooli, F.; Jones, W. Chem. Phys. 1998, 236, 225. (22) Mitra, S. S. Solid State Phys. 1962, 13, 1. (23) Braterman, P. S.; Cygan, R. T. Am. Mineral. 2006, 91, 1188. (24) Wang, J.; Kalinichev, G.; A., Amonette, J. E.; Kirkpatrick, R. J. Am. Mineral. 2003, 88, 398. (25) Cygan, R. T.; Liang, J. J.; Kalinichev, A. G. J. Phys. Chem. B 2004, 108, 1255. (26) Braterman, P. S. Cygan, R. T. Manuscript in preparation. (27) Eisenberg, D.; Kauzmann, W. The structure and properties of water; Oxford University Press: London, 1969. (28) Lowenstein, W. Am. Mineral. 1954, 39, 92. (29) Chakoumakos, B. C.; Loong, C. K.; Schultz, A. J. J. Phys. Chem. B 1997, 101, 9458. (30) Xu, Z. P.; Lu, G. Q. Chem. Mater. 2005, 17, 1055. (31) Kirkpatrick, R. J.; Kalinichev, A. G.; Wang, J.; Hou, X; Amonette, J. E. In The Application of Vibrational Spectroscopy to Clay Minerals and Layered Double Hydroxides; Kloprogge, J. T., Ed.; CMS Workshop Lectures, Vol. 13; The Clay Minerals Society: Chantilly, VA, 2005; p 240 ff.