J. MATER. CHEM., 1994, 4( ll), 1737-1744 Ion-exchange Properties of Lithium Aluminium Layered Double Hydroxides Ian C. Chisem and William Jones* Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, UK CB2 I EW The synthesis of layered lithium aluminium hydrotalcite-like materials is described along with different anion exchange procedures for the preparation of materials intercalated with chloride, nitrate and vanadate. The products have been characterised using elemental chemical analysis, powder X-ray diffraction, Fourier-transform infrared spectroscopy and thermogravimetric analysis. The matrices are found to be reasonably stable to acid treatment at pH 4.5 for periods of up to 72 h, with anion exchange taking place. Results indicate that total exchange of interlayer carbonate for chloride, nitrate and vanadate may be accomplished.The thermal properties of the materials have been studied: they demonstrate interesting differences in thermal behaviour compared with hydrotalcite. Layered double hydroxides (LDHs) are mixed-metal hydroxides with the general formula [MZ+1-xM3+x(OH)2]A+Xm-Alm*nH20,forwhere A=x z=2 and A=2x-1 for z=l.l A recent review describes their synthesis, properties and applications.2 They are re-lated to the naturally occurring mineral hydrotalcite [Mg6Al,(OH),6(CO,). 4H20] whose structure consists of a positively charged framework in which the cations occupy the octahedral sites between the sheets of close-packed anions. The layers are thus similar to those of brucite [Mg(OH),], but with a proportion of the divalent metal cations replaced by trivalent cations.The resulting excess positive charge is balanced by anions incorporated in the interlayer region. Such solids are comparable to the well known and widely studied cationic clays,, and thus hydrotalcite-like materials are some- times referred to as 'anionic clays'. Here, we examine the particular layered lithium alum- inate hydroxide system represented by the formula [LiAl,(OH),] +X-.nH,O, which consists of sheets of alu- minium octahedra with vacancies filled by lithium The symmetry of the cationic sheets is hexagonal, with AB-BA-AB-BA stacking. The sheets are neutralised by exchangeable anions, X-, which are located between the layers. A major difference in the structure of this material in comparison with M"/M"' LDHs, however, is that the latter Experimental Synthesis Details Analytical-grade reagents were used for all preparations described in this work.The various synthetic routes used are outlined in Scheme 1. LiAlCO, (Parent) [(LiA12(0H)6+)2]C032- -nH,O was prepared by a method similar to that used by Serna et al.' and based on a synthesis first described by Feitknecht.', AIC1,.6H20 (250ml of a 0.4mol 1-' solution) was added dropwise to 600ml of a mixture of 1.5 moll-' LiOH*H,O and 0.08 moll-' Na2C03 with vigorous stirring. The addition took 1 h. The pH varied from 12.8 (initially) to 10.2 (finally). The gel-type precipitate was then crystallised at 65°C in a thermal bath with gentle stirring for 18 h.When the product had been coolcd it was filtered off and washed several times with hot distilled water. The material was dried for 18 h at 100°C in air. Acid Exchange ofLiAlC0, LiAlCO, (1.0 g) was added to 150 ml distilled, deioniced water and the pH adjusted to 4.5 by the addition of 0.1 mol 1-' HNO, or HC1. The addition was accompanied by rapid expulsion of CO,. The mixture was stirred whilst maintaining The lithium aluminium system has, in general, been much less widely studied than its MI', M"' analogues. In particular, comparatively little is known about the exchange properties of these materials. Meyn et al., have examined the anion- exchange properties of a variety of LDHs including LiAl materials by exchanging the nitrate-intercalated LDH with aqueous solutions of organic acid salts, and Borja and Dutta7 have exchanged lithiurn aluminium chloride LDH with long- chain fatty acids.Exchange of chloride for 4-nitrohippuric acid was performed by Cooper and Dutta.8 Here, we demon- strate for the first time the anion exchange of carbonate for chloride and nitrate by exchange with acid by a method similar to that used by Bi~h','~ for the M" MI1'systems. A recent paper'' examining the properties of vanadate- exchanged MgAl LDHs has concluded that the decavanadate anion (V100,,6~) may be readily incorporated by a variety of synthetic routes. However, it has previously been reported that V,00286-may not be ion-exchanged into lithium aluminium LDHs.', Here we examine routes to vanadate-exchanged LiAl LDHs and attempt to compare the properties of the LiAl system with the MgAl system.show little evidence for cation ordering (i.e. M" and MIr1 the pH at 4.5 by further addition of acid. The reaction was occupy the same set of octahedral sites), whereas X-ray studies carried out over varying periods of time, of between 30min of the lithium compound show evidence of cation ~rdering.~*~ and 72 h. The resulting product was isolated by fillration or centrifugation and washed several times with hot distilled water. LiA1N03N(72) was prepared by reaction of ,.i mixture of 1.0 g LiA1CO3 in 150 ml distilled, deionised water contain- ing 0.1 mol 1-' NaNO, with the pH adjusted to pH 4.5 by addition of 0.1 mol I-' HNO,.The mixture was stirred whilst maintaining the pH at 4.5 by further addition of acid for a period of 72 h. The resulting product was isolated bj filtration or centrifugation and washed several times with hot distilled water. Direct Synthesis of LiAlCl (D) The preparation was similar to that used by Twu and Dutta', and Serna et a!.' Aluminium (0.05mol) was dissolved in 100ml of 2.0 mol 1-' NaOH solution. LiCl(0.25 moll-') was dissolved in this solution, and the mixture was heated for 48 h at 90°C under a nitrogen atmosphere. The products were washed with hot distilled water followed by 0.1 mol 1-' NaCl. Vanadate Intercalates A 150 ml solution of 0.1 mol 1-' NaVO, was adjusted to pH 4.5 by addition of 0.5 mol 1-' HCl (for LiAlCO, and LiAlCl J.MATER. CHEM., 1994, VOL. 4 heal to 150 "C for 18 hLiAlC03 direct synthesis b 1 treat with HCI LiAICl(0.5) I4h LiAlCl(4)a 124th -LiAICl(24) LiAIV(CI) treat with NaZC03 LiAIC03(CI) LiAK=O,(C) & # Itreatwith inpresence of glycerol LiAIV(CO3)G treat with HN03 LiAIN03(0.5) I4h LiAINO3(4)QI24h LiAIN03(24)1 LiAIN03(72)Qtreat with NaVOT I18hr-lLiAIV(N03) treat with HN03-NaN03I ,ltrrwilhiJdQ4Cl LiAIV(NO3N) treat with NaVOdCI I18h LiAIV(CID)U Scheme 1 The various routes followed for the synthesis of the carbonate, nitrate, chloride and vanadate materials precursors) or 0.5 mol 1-' HNO, (for LiAlNO, precursors).LiAlCO,, LiAlC1(24), LiAlN0,(72) or LiA1N03N(72) (1.0 g) was added and the pH was adjusted to 4.5 by addition of 0.1 mol 1-' acid. The mixture was stirred and the pH was maintained at 4.5 throughout the course of the reaction by further addition of 0.1 mol 1-' acid. The reaction was carried out for periods of between 4 and 24 h. The resulting product was isolated by filtration or centrifugation and washed with hot distilled water. Vanadate Exchange of LiAlCO, in the Presence of Glycerol To a suspension of 0.5 g of LiAlCO, in 50 ml distilled water and 100ml glycerol was added, with stirring, a solution of 0.45 g of NaV0, in 25 ml water acidified with 2.0 mol 1-1 HC1 to a pH of 4.5. The pH was maintained at this value for a further 5 h.During the reaction the system was purged with N, to remove evolved CO,. All processes were carried out at room temperature. The resulting product was isolated by centrifugation and washed with hot distilled water. Techniques Elemental chemical analysis for Li, A1 and V was performed using a Perkin-Elmer 3100 atomic absorption spectrometer. C and H were determined using a Perkin-Elmer 2400C instrument. Powder X-ray diffraction (PXRD) patterns were recorded using a Philips APD 1710 instrument with Ni- filtered Cu-Kn radiation. A step scan of 0.05" (28) at a rate of 0.05" s-' was used. The FTIR spectra of the solids were recorded at room temperature using a Nicolet 205 FTIR spectrometer using the KBr pellet technique, with a resolution of 2 cm-' and 60 scans averaged.Thermogravimetric analysis was carried out using a Polymer Laboratories TGA 1500 with a heating rate of 10°C min-' in all cases. Results and Discussion Elemental Analysis Results for chemical analysis of the samples are given in Table 1. The A1:Li ratio is found to be close to 2 for the parent LiAlCO, sample. Despite the fact that the A1 :Li ratio is rather high (2.5) for LiAlC1(24), there appear to be no systematic trends which might suggest preferential leaching of lithium or aluminium in these samples. For the vanadate- exchanged samples with chloride or carbonate precursors the ratio is significantly less than 2, being around 1.7&0.2. Where LiAlNO, is used as the precursor, the A1 :Li ratio is found to be close to 2.The vanadium content is found to remain fairly constant for samples with chloride or carbonate precursors (ca. 22 2%). In the case of vanadate-exchanged LiAlNO, mate- rials the vanadium content is significantly higher (ca. 27%). The carbon and hydrogen content were determined in selected samples (Table 2). The values are consistent with the presence of carbonate in the interlayer of LiAlCO, but its absence in LiAlNO,, LiAlCl and LiAlV (since a C content of 0.2-0.3O/0 is close to the sensitivity of the instrument). On the basis of the above data it can be assumed that the oxovanadate species V,O,"- are the only anions present in the interlayer of the vanadate-intercalated samples. The excess positive charge on the layers must therefore be balanced by J.MATER. CHEM., 1994, VOL. 4 1739 Table 1 Elemental analysis data for the samples synthesized sample Li(%) Al(Y0) V(%) (1-x) x A1:Li z/ua LiAlCO, 2.9 20.9 -0.35 0.65 1.87 -LiAlCO,(Cl) 2.5 20.7 -0.32 0.68 2.15 -LiAlCl(0.5) 2.8 21.3 -0.34 0.66 1.97 -LiAlCl(4) 2.6 20.8 -0.32 0.68 2.07 -LiAlCl(24) 2.4 23.1 -0.29 0.71 2.50 -LiAlCl(D) 3.1 23.2 -0.35 0.65 1.94 -LiAlNO,( 0.5) 2.8 21.3 -0.34 0.66 1.97 -~LiAlNO,( 4) 2.7 22.4 -0.32 0.68 2.15 LiAlNO, (24) 2.8 21.2 -0.34 0.66 1.96 -LiAlNO,( 72) 2.6 21.5 -0.32 0.68 2.14 -LiAlNO,N( 72) 2.7 22.2 -0.32 0.68 2.13 -LiAlV(C0,) 2.0 13.0 20.3 0.37 0.63 1.69 0.50 LiAlV( C0,)G 1.8 12.9 20.2 0.35 0.65 1.86 0.56 LiAlV( C1) 2.1 12.6 22.5 0.39 0.61 1.56 0.42 60 80 LiAlV (ClD) 2.1 14.3 23.2 0.36 0.64 1.77 0.51 2Bldegrees LiAlV( NO,) 2.0 15.3 26.3 0.33 0.67 1.98 0.56 LiAlV( NO,N) 2.1 15.8 27.2 0.34 0.66 1.95 0.53 Fig.1 Powder X-ray diffraction patterns for: (a) LiAlCO,, (b) " Assuming a formula [Li,~,A~,(OH),~(V,Ob)"~nH2OV-LiAlCl(O.5), (c)LiAlCl(4). (d) LiAlCl(24) and (e) LiAlCl(D Ifor the containing samples. higher values of 28. The basal spacing of 7.!7 A is slightly Table 2 Carbon and hydrogen content for a selection of the samples lowe! than those reported by Masco10'~ (7.60 A), Serna et al.l synthesized (7.6 A) and Sissoko et aL5 (7.56 A). It is well known that the basal spacing for hydrotalcite is influenced strongly by the sample c(Yo) H(%) drying treatment of the material, and hence small differences LiAICO, 2.5 4.0 in basal spacings are not significaat.The value oblained is LiAlCI(24) 0.2 3.7 indicative of a gallery height of 2.7 A, which !pproximates to LiAlNO,( 72) 0.2 3.3 the thickness of .a carbonate anion (2.84A). assuming a LiAlV(C0,) 0.3 2.7 thickness of 4.77 A for the cationic sheets. The material is relatively crystalline and has a well ordered sheet structure. as observed elsewhere. Indexing of the powder pattern based the negative charge on the interlayer anion, with the ratio of on the basis of an or4ered supercell' is given in Table 4.Thez/a balancing the aluminium content. The ratio is 0.53k0.03 reflection at d=4.38 A cannot be indexed on the basis of a for all samples except LiAlV(C1) which has a z/a ratio of 0.42.random cell, providing evidence for cation ordering (i.e. onlyThis sample also exhibits an unusually low Al: Li ratio of on the basis of an ordered cell can the pattern be indexed). 1.56. We can thus describe the anion as a single species with An alternative indexing based on a monoclinic cell5 is also the formula [V10027.5]s-,which may, within the grounds of given.experimental error, be reasonably assumed to be [Vlo02,]6-. (Assuming the vanadium to be present as V5+,an assumption LiAlCl Samplesgiven the inherent instability of V3+ in air.) The validity of Irrespective of the duration of acid treatment therc is little this analysis is somewhat open to question; it is highly likely variation in the observed c parameter when compared with that no single vanadium species is present in the interlayer, LiAlCO, [Fig.ol(a)-(d)].The observed values of c lie between but in fact a number of species are present. 14.9 and 15.2 A, whic! is slightly smaller than that reported by Mascolo14 (15.40 A) and Twu and Dutta12 (15.6 A) for Powder X-Ray Diffraction chloride-treated lithium aluminate. The values are similar to that obtained for LiA1C03, reflecting the similar size of theThe values for the c dimension of the various materials chloride and carbonate anions. As a result, PXRD gives little synthesized are given in Table 3. LiAlCO, Samples Table 4 Indexing of powder pattern for LiAlCO, The PXRD pattern of LiAlCO, [Fig. l(a)] is typical of that for an LDH, with sharp, symmetric basal (001) reflections at 20/ hexagonal cell" monoclinic cell' -~ lower values of 28 and relatively broad, weak reflections at degrees I/Io(%) dobs/A dcalc/A hkl dcalc/A hkl Table 3 c parameters for samples obtained from PXRD data 11.91 100.0 7.43 7.48 002 7.55 002 20.27 10.9 4.38 4.38 101 4.38 110 sample c/A sample CIA 23.72 67.9 3.75 3.74 004 3.78 004 ~ 35.85 42.4 2.50 2.49 006 2.52 006 LiAlCO, 15.0" LiAlNO, (72) 17.7" 35.85 42.4 2.50 2.49 112 LiAICO,(Cl) 15.1" LiAlNO,N( 72) 17.8" 40.02 7.2 2.25 2.19 202 2.25 016 LiAlCl(0.5) 15.1" LiAlV(C03) 22.9b 40.02 7.2 2.25 2.19 106 LiAlClI4) 15.2" LiAlV(C03)G 23.7b 46.92 5.1 1.94 1.98 115 1.99 017 LiAlCl(24) 15.0" LiAlV( C1) 23.5' 63.26 16.8 1.47 1.46 303 1.46 330 LiAICl(D) 14.9" LiAlV( C1D) 23.5b 64.70 13.1 1.44 1.42 10,lO 1.44 600 LiAlNO,(O.5) 15.1" LiAlV(NO3) 23.1 68.51 5.2 1.37 1.36 305 LiAlNO,( 4) 15.1" LiAlV( N0,N) 23.4' LiAlNO,( 24) 15.2" "AssignmFnt based on a hexagonal unit cell with a=5.29 A, C= 14.95 A." c =( 1/3) [2d( 002) +4d( 004) +6d(006)]. 'c =( 113) [44004) + bAssignment bayd on a monoclinic unit cell with a=8.68 A,b= 6d( 006 I]. 5.07 A, C= 15.12 A, /?=92"56'. 1740 useful information about the intercalation of chloride, though it can be concluded that the LDH structure is stable at this pH since the PXRD confirms that the layered structure remains intact even after 24 h. A gradual weakening in the intensity of both the basal and non-basal reflections is observed upon increasing exposure to acid. The PXRD pattern of the directly synthesized chloride [LiAlCl(D), Fig.1 (e)] is similar to that ,Of the other materials and exhibits a basal spacing of 14.9 A. Treatment of LiAlCI( 24) with 0.1 mol 1-Na2C03 solution yields a material [LiAlCO,(Cl)] with a PXRD pattern, IR spectrum and TG curve identical to that of LiAlCO,. We thus conclude that the ion-exchange process is essentially reversible. LiAINO, Samples There is little change in the observed c parameter for treatment times of up to 24 h, but after 72 h of acid treatment a significant increase in the c parameter occurs (Fig. 2). The values of c for samples LiAlNO,(O.S), LiAlN0,(4) and LiAIN0,(24) are similar to those observed for LiAlCO,, suggesting that little nitrate has been intercalated.After 72 treatments, materials with c values of 17.7 and 17.8 A [LiAlNO,( 72) and LiAlNO,N( 72), respectively] are prod-duced. These values are consistent with that of 17.60A reported by Masc01o.l~ It can be thus deduced that nitrate is incorporated, but at a rate which is significantly slower than that of chloride. Again, the structure is found to be stable to the acid treatment, although some weakening in the intensity of the basal reflections is observed. Results indicate that a treatment time of 48 h gives rise to a material wjth two phases corresponding to values of c of 14.9 and 17.7 A, respectively, suggesting co-intercalation of nitrate and carbonate [Fig. 2(e)]. LiAl Vunadates The PXRD patterns for the vanadate intercalated materials are similar in appearance (Fig.3). The first peak is veryobroad and its position is variable between ca. 8.8 and 10.4A. The second and third peaks are relatively sharp, and more similar to those recorded for the carbonate-, chloride- and nitrate- intercalated derivatives. In a similar way to the MgAlV materials" the position of the first reflection is incorrect for (002) if the second and third reflections are indexed as (004) and (006), respectively. It is possible that the material is in fact biphasic, with the broad reflection hiding the (002) of the 1 1 (a4 20 40 60 80 2Bldeg rees Fig. 2 Powder X-ray diffraction patterns for: (a) LiAlCO,, (b) LiAlNO,(O.S), (c) LiAlN0,(4), (d) LiAIN0,(24), (e) LiAIN0,(48),(f)LiAlN0,(72) and (g) LiAlNO,N(72) J.h4ATER. CHEM., 1994, VOL. 4 0 20 40 60 80 2eldegrees Fig. 3 Powder X-ray diffraction patterns for: (a) LiAIV(CO,), (b) LiAlV(CO,)G, (c) LiAlV(Cl), (d) LiAlV(ClD), (e) L,iAIV( NO,) and (f) LiAIV(N0,N) intercalated phase. The reason why the (002) peak should be so weak as to be hidden by this broad reflection is not known, however. Similar broad peaks have been reported for MgAl vanadates'l and ZnAl hydrotalcites intercalated with [a-SiW11039]8- and [a-1,2,3-SiV,W,0,0]7- .I5 ,4s a result of this, the interlayer spacings for the vanadate-intercalated samples have been calculated from the positions of the (004) and (006) reflections only. A gallery height of between 6.7 and 7.0 A is obtained for vanadate intercalatedoLiAl LDHs.Twu and Dutta observed gallery heights of 6.0 A for LiAl vanadate LDHs prepared by a route similar to that used to prepare LiA1V(CID).l2 They proposed that V2074-was the vanadate species at pH 8-11, and V40124- at pH between 3 and 8. They did not report the presence of a broad peak in the X-ray powder pattern. For MgAl vanadate LDHs, Twu and Dutta,I6 Pinnavaia and co- worker~:~and Ulibarri et d." all observed gallery heights of 7.0-7.1 A. In this case, V,002,6- was proposed as the interca- lated species. Pinnavaia and co-workers found that ion exchange of V4OlZ4- into transition-metal hydrotalcite~l~ (at pH 5.5-10.0) led to a gallery height of 4.7 A. 'Twu and Dutta account for the difference in spacing for the V,0,24-in LiAl and MgAl systems by suggesting that there is a stronger electrostatic interaction in the transition-metal LDH due to the higher charge density.It is clear that the value we have obtained for the gallery height of our vanadate samples from PXRD analysis is close to that of MgAl decavanadates and it thus appears that we have Vl,02,6-intercalated LiAl LDHs. Assignment of Monoclinic or Hexagonal Cell Two alternative cells may be used to index the powder patterns of these materials; a monoclinic cell or a heFagonal cell (with cation ordering). If the reflection at d =4.38 A moves upon intercalation, then it cannot be the (1 10) reflection of a monoclinic cell; conversely, if its position does not change upon intercalation then it cannot be the (101) reflection of the hexagonal cell.Fig. 4 shows the change in the position of the reflection in the various samples synthesked, and com- pares the observed value of d for this reflection with the calculated values for a monoclinic and hexagonal cell. Experimentally the data are not a good fit to either line, but it appears that there is no evidence for an increase in dobsas the c parameter increases. Thus it is reasonable to suggest that the reflection does indeed correspond to (110) and a monoclinic cell is more appropriate. J. MATER. CHEM., 1994, VOL. 4 014.3 ! I I I 20 25 30 35 c value Fig. 4 Variation of d spacing for (110), monoclinic, and (101), hexagonal. dabs, 0;dale (hex), broken line; and d,,,,(mono), solid line. Infrared Spectroscopy LiA1C03Samples The infrared spectrum of LiA1CO3 is shown in Fig.5(a), and assignments are given in Table 5. The main bands to note are those due to carbonate at 1370, 1050, 875 and 675 cm-' and the H-bonding between water and carbonate in the interlayer at 3000 cm-'. These bands should therefore be lost on inter- calation with chloride, nitrate or polyvanadate. The presence of the band at 1050cm-' provides some evidence for a lowering of symmetry of the carbonate species, witnessed also by the shoulder at 1400cm-'.18 LiAlCl Samples A gradual loss in the intensity of the carbonate bands is observed with increasing acid treatment time [Fig. 5(b)-(d)]. The H-bonding vibration between H20 and C032- is also lost with increasing treatment time.After 24 h treatment there is little remaining residual carbonate visible. The presence of bands attributable to hydroxy groups at ca. 940 and 600 cm-' but the lack of vibrations due to interlayer anions would seem to indicate the presence of chloride in the interlayer. LiAlCl(D) [Fig. 5(e)] shows a weak carbonate band, suggesting that I I4000 3200 2000 1600 1200 800 400 wavenum berkm-' Fig. 5 Infrared spectra of: (a)LiAlCO,, (b)LiAlCl(OS), (c) LiAlC1(4), (d) LiAlCl(24) and (e)LiAlCl(D) 1741 Table 5 IR data for LiAlCO, wavenumber /cm -assignment 3450 H-bonding stretching vibrations of OH group in layer 3000 H-bonding between the H20 and C0,2-in the interlayer 1650 H20 bending vibration 1400 lowering of symmetry of C0,2- (C2") 1370 CO,'-absorption band (v,) 1050 lowering of symmetry of C0,2-causes v1 mode to become active (usually Raman-active only for D,, co32-1 1030 OH bending vibration in layer 875 CO,'-absorption band (v2) 725 A1-0 (A2u)675 C0,2-absorption band (v4) 550 A1-0 (E,) 455 A1-0 (E,) some carbonate was incorporated by the direct synthesis method, despite attempts to exclude air from the system.LiAlNO, Samples For acid treatment times of up to 24 h there is little change in the appearance of the carbonate bands in the spectrum [Fig. 6(b)-(d)]. Note, however, the appearance of a nitrate band at 1385 cm-' in all samples treated with HN03.It may be concluded that only a small quantity of nitrate is incorpor-ated even after a 24 h treatment.The spectrum of LiA1N03( 72) [Fig. 601 is quite different, however. The complete loss of n 4 30 3200 2000 1600 1200 800 400 wavenumberkm-' Fig. 6 Infrared spectra of (a) LiA1CO3, (b) LiAlN0,t OS), (c) LiA1N03(4), (d) LiA1N03(24), (e) LiAlN0,(48), (f) LiAIN0,(72) and (g) LiAlNO,N(72) 1742 the carbonate bands is observed and there is a strong v3 NO,-absorption at 1385cm-' and a weak v2 band at 825 cm-', as observed by Hernandez-Moreno et a!.'' The v1 band is just visible at around 1050cm-' in some samples. There is evidence for loss of symmetry of the nitrate anion upon intercalation, witnessed by the splitting of the v, band (1425, 1385 cm-'). LiA1N03N(72) has a similar IR spectrum [Fig.6(g)], but the v3 band appears to be more symmetric. The IR evidence thus confirms the observations made by PXRD that little nitrate intercalation occurs for treatment times of up to 24 h, but major conversion is achieved after 72 h. LiAl Vanadates The IR spectra of the LiAl vanadates are shown in Fig. 7. There are several points to note. The shoulder at ca. 3000 cm-' is absent, as are the other bands due to the presence of carbonate anions. The bands associated with hydroxy groups and molecular interlayer water are present, however (ca. 3520 and 1630 cm-'). The main diagnostic bands in these spectra are the medium-intensity bands at around 967 and 951 cm-'. It has been shown that crystalline polyvanadates give rise to absorption bands in the 950-1000 cm-' regionlg attributable to the V=O terminal stretching mode.The nature of the vanadate anion influences the number and position of the bands, with decavanadates giving rise to a single band and hexavanadates giving two bands in this region. Here, the presence of a double band suggests that decavanadate is not the only interlayer species. Other bands at ca. 827, 729 and 668 cm-' are in similar positions to those reported for vana- 1OOr I 40 20 g{" Lc 20 0 !.. :. 4 1100 900 700 500 wavenum berkm-' Fig. 7 Infrared spectra over the regon 1100-400 cm-' for: (a) LiA1V(C03), (b) LiAlV(CO,)G, (c) LiAlV(Cl), (d) LiAlV(ClD), (e) LiA1V(N03),(f) LiA1V(NO3N)and (8)a typical infrared spectrum over the region 4000-400cm-' for an LiAl vanadate LDH [LiAlV(NO,) shownJ J.MATER. CHEM., 1994, VOL. 4 date-pillared MgAl LDHs.~' A comparison of the observed bands in the region 500-1000cm-' is given in Table 6. Metavanadates (i.e. MVO,) give rise to IR bands at 850-863 cm-', 920-935 cm-' for terminal V=O stretching and 475-495 and 685-693cm-' for V-0 stretching in bridging V-0-V bonds. In view of the fact that no bands are seen in these regions, the presence of these species would seem unlikely. Thermal Treatment All thermogravimetric analysis was performed in a nitrogen atmosphere. LiAlCO, Samples The TG curve [Fig. 8(a)] shows an overall weight loss of 43.3%, which corresponds to the loss of interlamellar water (nH,O), dehydroxylation and loss of C02.The weight loss occurs over a wide range of temperatures between 150 and 350°C and is centred around 225-245 "C. It is not possible therefore to distinguish the three transitions, unlike the case of hydrotalcite,21 where a weight loss at 216°C had been identified as due to the loss of interlamellar water and weight losses at 330 and 370°C attributed to simultaneous loss of CO, and dehydroxylation. On the basis of the observed weight loss and chemical Table 6 Bands observed in the 1000-500 cm-' region for MgAl vanadates and LiAl vanadates v( MgAl )/cm- v( LiAl)/cm-' - 967 970 951 825 827 748 729 675 668 605 602 555 i79 10%I I I I I I I I 1 4 100 200 300 400 500 TI'C Fig.8 TG curves for: (a) LiAICO,, (b) LiAlCl(O.S),(c) LiAIC1(4), (d) LiAlCI(24) and (e) LiAICl(D) J. MATER. CHEM., 1994, VOL. 4 analysis, the number of moles of interlamellar water, n=0.26, giving the formula [Lio.3sAl,~,s(OH)2](C03)o~ls-0.26H20. Heating a sample of LiAlCO, to 150°C for 18 h leadsoto a material (LiAlC0,C) with a gallery height of 2.02 A as compared with a height of 2.7 A for the precursor. There is also a significant loss in the intensity of the basal reflections, to such an extent that the (004) reflection is no longer visible. It would appear that prolonged treatment at this temperature is enough to result in significant loss of carbonate and water from the galleries. The TG curve of this sample cooled to room temperature exhibits a weight loss of 24.8% at tempera- tures up to 500°C. This compares with a theoretical weight loss of 31.9% if the sample contained no interlayer carbonate or water but there was no dehydroxylation.This suggests that partial dehydroxylation also occurs upon prolonged treatment at 150"C. Heating the sample beyond 500 and up to 800°C results in virtually no further weight loss. It has been proposed that at temperatures >450 "C the framework is destroyed, leading to the formation of Li20 and y-A120,.'2 LiAlCl Samples When this material is treated with HCl at pH 4.5 significant changes are seen in the TG curve [Fig. 8(b)-(d)]. In all samples, the first transition corresponds to the loss of physi-sorbed surface water at relatively low temperature.After 30min treatment the broad weight loss which is seen in the TG curve of LiAlCO, is resolved into three weight losses centred at 200, 250 and 320°C. After 4 h treatment the transitions may be more clearly resolved and the positions have shifted to 190, 250 and 320°C. After 24 h only two transitions are then visible at 250 and 315°C. This sequence suggests that the weight loss below 200°C must be due to the loss of residual carbonate from the interlayer, since it is not present in the sample treated for 24 h. Indeed, the IR spectrum of this sample heated to 210°C shows no bands in the 1350-1450 cm-' region, proving that carbonate has been lost. We may thus conclude that carbonate becomes increasingly unstable in the interlayer as the amount of intercalated chloride increases.The second and third transitions are due to the loss of interlamellar water and dehydroxylation, respectively. For LiAlCl( 24) calculations show that the second weight loss, centred at 250"C, corresponds to the loss of 0.31 mol of interlamellar water. On the basis of this analysis, the molecular formula for LiAlCl(24) is [Lio,,,Al,,,,(OH)2]C1,~42 .0.31H20. The total observed weight loss is 31.6Y0, which compares with a theoretical value of 31.2% based on the above formula, and is therefore within experimental error. It can be seen that dehydroxylation occurs at a higher temperature in this com- pound than in LiAlCO,, which possibly reflects differences in the H-bonding stability of the hydroxy groups.The TG curve for LiAlCl prepared by direct synthesis [LiAlCl( D), Fig. 8(e)] is similar to that of LiAlCl(24). LiAlNO, Samples With HNO, and for treatment times up to 24 h the TG curves appear to be similar to that of the carbonate material, but with a shift in the weight loss towards lower temperature [Fig. 9(b)-(d)]. PXRD and elemental analysis show these materials to contain predominantly carbonate in the inter- layer. It appears that, as for the LiAlCl samples the carbonate is destabilised as the acid-treatment time increases. The TG curve of LiA1N03(72) is quite different [Fig. 9(f)] and two transitions are seen. The first is associated with the loss of physisorbed water (at low temperature), the second transition, centred at 310 "C, corresponding to dehydroxylation of the layers.There is no weight loss close to 200"C, indicating the I I 1 I I I I I 1-1 100 200 300 400 500 TI'C Fig. 9 TG curves for: (a) LiAlCO,, (b)LiAlNO,(O.S), (c) LiAlY03(4), (d) LiA1N03(24), (e) LiA1N03(48), (f) LiAlN0,(72) and (g) LiAlNO,N( 72) absence of carbonate in the interlayer, consistent ~ith the data obtained by other techniques. There is in addition no observed weight loss associated with the loss of interlayer water in this compound. It is possible that the galleries are so tightly packed with nitrate anions that there rs little remaining space for interlayer water. This would suggest that the H20 bending vibrations seen in the IR spectrum of this compound are in fact due to physisorbed surface water and not interlamellar water.The IR spectrum of this sample heated to a temperature of 250°C shows little change with respect to the precursor, confirming there is no dehydroxyl- ation or nitrate decomposition up to this temperature. The theoretical weight loss based on the formula Li0.32A10.68(OH)2(N03)0.36is 23.4%, and the observed weight loss of 23.7% up to 350°C is thus in agreement. At higher temperatures a further gradual weight loss is observed which may correspond to the loss of the nitrate anion as NO2. A weight loss of 45.2% is observed up to a temperature of 610 "C, whereupon there is no further decomposition. The theoretical weight loss based on the assumption that NO2 is lost is 45.2%, which appears to be within experimental error.A similar TG curve is seen for LiA1N03N(72) [Fig. 9(g)], with a 23.0% weight loss observed at temperatures up to 350 "C. LiAl Vanadates Two transitions are seen in the TG scans of these materials (Fig. 10). The first weight loss occurs at temperatures of <145"C and has previously been ascribed to the loss of interlayer water.12 The second weight loss at around 325 "C may be attributed to the dehydroxylation of the layers. This weight loss is around 13.5% of the initial weight. On the basis of the formula [Lio.36Alo,64(oH)2(V100286-)0.04-] the expected weight loss due to dehydroxylation would be 18.2%. The reason for this discrepancy is as yet unknown. It is possible that this sample contains some co-inter calated glycerol which decomposes upon thermal treatment.f) 1 I 1 I I If I I I 100 200 300 400 500 TI'C Fig. 10 TG curves for: (a) LiAlV(CO,), (b) LiAlV(CO,)G, (c) LiAlV(Cl), (d) LiAlV(ClD), (e)LiAlV(N0,) and (f) LiAlV(N0,N) Conclusions It is clear from this study that LiAlCO, sheets are stable to acid treatment at pH 4.5 with prolonged treatment, resulting in complete anion exchange of the carbonate for the anion of the acid. Exposure of these acid-treated samples to sodium carbonate solution results in the regeneration of the precursor carbonate-containing material. It is also clear that vanadate-intercalated LiAl LDHs may be readily prepared via carbonate, chloride or nitrate precur- sors and the structure and properties of the products appear to be largely independent of the precursor.Thermal treatment of LiAlCO, leads to the simultaneous loss of interlayer water, dehydroxylation of the layers and decomposition of the carbonate anions over an extended temperature range between 150 and 375°C. Treatment of J. MATER. CHEM., 1994, VOL. 4 LiAlCl leads to loss of interlayer water between 235 and 260 "C and dehydroxylation at temperatures between 260 and 365 "C. For LiAlNO,, dehydroxylation of the layers occurs between 285 and 350°C and, at temperatures of >450"C, decomposition of the nitrate anions. There is no evidence for the existence of interlayer water in this compound. For LiAlV there is a loss of interlayer water at temperatures below 145"C.Dehydroxylation of the layers occurs at temperatures above 250°C and is complete by 350°C. We are grateful to the EPSRC for support (studentship to I.C.C.).Discussions with V. Rives are appreciated. This work was performed under the Concerted European Action on Pillared Layered Solids (CEA-PLS). Support via Accion Integrada is appreciated. References 1 C. J. Serna, J. L. Rendon and J. E. Iglesias, Ckiys Clay Mineral., * 1982,30, 180. 2 F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 1991,11, 173. 3 Pillared Clays, in Catal. Today, ed. R. Burch, 19S8, vol. 2. 4 W. T. Reichle, CHEMTECH 1986, 86. 5 I. Sissoko, T. Iyagaba, R. Sahai and P. Bilnen, J. Solid State Chem., 1985,60,283. 6 M. Meyn, K. Beneke and G. Lagaly, Inorg. Chcm., 1990,29,5201. 7 M. Borja and P. K. Dutta, J. Phys. Chem., 1992,96,5434. 8 S. Cooper and P. K. Dutta, J. Phys. Chem., 1990,94, 114. 9 D. L. Bish and G. W. Brindley, Am. Mineral., 1977,62,458. 10 D. L. Bish, Bull. Mineral., 1980, 103, 170. 11 M. A. Ulibarri, F. M. Labajos, V. Rives, R. Trujillano, W. Kagunya and W. Jones, Inorg. Chem., 1994,33,2592. 12 J. Twu and P. K. Dutta, J. Phys. Chem., 1989,93,7863. 13 W. Feitknecht and M. Gerber, Helv. Chim. Actu. 1942,25, 131. 14 G.Mascolo, Thermochim. Acta, 1986,102,67. 15 E. Narita, P. Kaviratna and T. J. Pinnavaia, Chem. Lett., 1991, 805. 16 J. Twu and P. K. Dutta, J. Catal., 1990, 124, 503. 17 T. Kwon, G. A. Tsigdinos and T. J. Pinnavaia, J. Am. Chem. Soc., 1988,110,3653. 18 M. J. Hernandez-Moreno, M. ,4. Ulibarri, J. L. Rendon and C. J. Serna, Phys. Chem. Mineral.. 1985, 12, 34. 19 L. D. Frederickson Jr. and D. M. Hausen, Anal. Chem., 1978, 23,93. 20 E. Lopez Salinas and Y. Ono, Bull. Chem. Soc. Jpn., 1992,65,2465. 21 L. Pesic, S. Salipurovic, V. Markovic, D. Vucelic, W. Kagunya and W. Jones, J. Muter. Chem., 1992,2, 1069. Paper 4/02607A; Receivtd 3rd May, 1994