J O U R N A L O F C H E M I S T R Y Materials The influence of template extraction on the properties of primary amine templated aluminosilicate mesoporous molecular sieves Robert Mokaya* and William Jones Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW. E-mail: rm140@cus.cam.ac.uk Received 3rd August 1998, Accepted 19th October 1998 Aluminosilicate mesoporous molecular sieves (Al-MMS) prepared at room temperature using hexadecylamine as template have been subjected to template extraction prior to calcination. Extraction in ethanol alone removes only that part of the template (in neutral form) which is not associated with framework Al while the presence of a cation (Na+ or NH4+) ensures total template removal.Template extraction has no eVect on elemental composition and for dry (non-calcined) samples results in an improvement in structural ordering.The eVect of calcination depends on the mode of extraction; samples extracted in ethanol or ethanol/NH4+ are structurally stable to calcination and possess surface area and pore volume similar to directly calcined samples while ethanol/Na+ extracted samples are relatively unstable and undergo considerable structural degradation resulting in lower surface area and pore volume.Dealumination results from calcination of amine or ammonium ion containing samples while Na containing samples do not undergo any dealumination. The acid content of calcined ethanol and ethanol/NH4+ extracted samples is comparable to that of the directly calcined samples but the extracted samples exhibit higher catalytic activity for the cracking of cumene. Calcined ethanol/Na extracted samples possess very low acidity and exhibit no catalytic activity.soluble silicon/aluminium species and also as a source of Introduction charge balancing protons (during calcination of the as- Recent advances in the synthesis of mesoporous molecular synthesised material ).This is advantageous because Brønsted sieves which possess uniform and sharply distributed pores of acid sites are generated by simple calcination. We and others diameter 20–100 A° have increased the range of well ordered have shown that in contrast to the purely siliceous analogue, solid acid catalysts (previously the domain of microporous where all the amine is in neutral form, a part of the templating zeolites) into the mesoporous regime.1 Synthesis of such silica amine in the Al-MMS samples exists in a protonated form based materials involves the use of surfactants to assemble where it is electrostatically bound to the inorganic framework inorganic species from solution into a solid framework in and cannot therefore be removed by simple solvent extracwhich the organic surfactant template is occluded.2–4 The tion.9,12 The proportion of such protonated amine increases removal of the template to generate the molecular sieve with with the amount of aluminium in the solid framework.9 We regular void spaces is therefore an integral part of the synthetic also found that high amounts of aluminium in the Al-MMS process.For M41S materials which are synthesised using a framework increased the attraction between the templating cationic surfactant and anionic inorganic species, template amine micelles and the framework and thus removal of the removal may be achieved either by calcination or by solvent template by direct calcination resulted in greater structural extraction followed by calcination.5,6 It has been shown that (framework) collapse due to the increase in local heating more of the template is removed during the extraction if the eVects.11 It is thus of interest to investigate other methods of solvent system contains a cation donor such as an acid or template removal. Here we report a detailed comparison of salt.5–7 For Al-containing MCM-41, template removal by the properties of Al-MMS catalysts prepared at room temperasolvent extraction results in an increase in the amount of non- ture using hexadecylamine as template and subjected to various framework octahedral Al (due to partial dealumination).methods of template removal. Directly calcined samples are Furthermore solvent extraction on its own does not readily compared with samples in which template removal is achieved remove any template associated with tetrahedrally coordinated by extraction using ethanol alone followed by calcination or framework Al and therefore calcination at elevated tempera- by solvent extraction using ethanol in the presence of an acid ture is required to ensure complete removal of the template.7 generating (NH4+) or a non-acidic (Na+) cation prior to Tanev and Pinnavaia8 have, however, shown that in the calcination.We have investigated the eVects of the extraction absence of strong electrostatic interactions between the surfac- procedure on the structural integrity of the Al-MMS materials tant and framework (e.g. when neutral primary amine micelles and also on the nature of the Al they contain. We also report are used to direct the assembly of neutral silica inorganic on the influence of the template extraction procedure on the species) it is possible to achieve complete template removal by acidity and catalytic performance for cumene cracking. solvent extraction alone.Particular emphasis is given to the eVect of cations during the We have recently reported the synthesis, acidity and catalytic extraction.properties of Al-containing mesoporous molecular sieves (Al- MMS) prepared using primary amines as the templates.9–11 We have shown that when compared to equivalent Al- Experimental containing MCM-41 materials the Al-MMS materials possess significantly higher Brønsted acid content and consequently Synthesis of materials exhibit higher catalytic activity for Brønsted acid catalysed The as-synthesised aluminosilicate materials were prepared as reactions such as the cracking of cumene.10,11 In the synthesis follows: aluminiumisopropoxide [Al(i-C3H7O)3 dissolved in 35 ml of Al-MMS materials the primary amine surfactant micelles act both as structure directing agents during the assembly of isopropanol ] was mixed with 0.2 mol tetraethylorthosilicate J.Mater. Chem., 1998, 8, 2819–2826 2819(TEOS, in 80 ml ethanol ) at Si/Al molar ratios in the range Catalytic testing 40–5/1 and vigorously stirred at room temperature for 15 The conversion of cumene was performed at 300 °C and a minutes. The template solution was separately prepared by WHSV of 4.0 using a tubular stainless steel, continuous flow dissolving 0.05 mol hexadecylamine in a mixture of 80 ml fixed-bed microreactor (of internal diameter 10 mm) with water and 120 ml ethanol.The TEOS–Al(i-C3H7O)3 mixture helium (25 ml min-1) as carrier gas. The catalyst bed (100 mg; was added to the template solution under vigorous stirring at 30–60 mesh) was first activated for 1.5 h at 500 °C under room temperature. The pH of the synthetic mixtures was close helium (25 ml min-1). For the reaction a stream of cumene to 9.5.The resulting gel mixture was allowed to react at room vapour in helium was generated using a saturator at room temperature for 20 hours following which the solid product temperature. The reaction products were separated and ana- was obtained by filtration and air dried at room temperature. lysed using a Carlo Erba HRGC 5300 gas chromatograph on Template removal was achieved either by direct calcination in line with the microreactor.Gas chromatographs were obtained air at 650 °C for 4 hours or by solvent extraction prior to automatically on samples of the product stream which were calcination. Solvent extraction was performed as follows: 1.5 g collected at regular intervals using a Valco 6-port valve.The of the dry (as synthesised) material was added to 150 ml of gas chromatographs were used to calculate the percent overall extraction media and stirred at 65 °C for 1 h followed by cumene conversion. filtration and washing with ethanol. This procedure was repeated once to give the ‘dry extracted’ samples. Three types Results and discussion of extraction media were used: (1) 150 ml ethanol alone; (2) 1.5 g of sodium acetate in 150 ml ethanol; and (3) 1.5 g Chemical composition and thermal analysis ammonium acetate in 150 ml ethanol.All the ‘dry extracted’ samples were then calcined in air at 650 °C for 4 hours. The The bulk Si/Al molar ratios of the parent (directly calcined) Al-MMS samples, shown in Table 1, indicate that Si and Al directly calcined materials are designated Al-MMSX (where X is the Si/Al ratio used in the synthesis gel ).The extracted are incorporated into the solid framework in proportions which are largely in line with the synthesis gel composition. samples are named with a prefix indicating the extraction media (Et for ethanol alone; NH for ammonium acetate and The mode of template extraction does not have any significant or systematic eVect on the Si/Al ratio except that for samples Na for sodium acetate) derived from Al-MMS20 there is some preferential leaching Characterisation of a small amount of silica (see Table 1) reflected in lower Si/Al ratios.However samples derived from Al-MMS10 (data Elemental composition (via EDA analysis) was obtained using not shown) had very similar Si/Al ratios (in the range a Camscan S4 scanning electron microscope at 20 kV.The 12.8±0.3) which implies that silica leaching is not necessarily data were processed through a ZAF4 program running on a a characteristic of solvent mediated template extraction. Link 860 series 2 processor. TGA and DTG curves were Fig. 1 shows the TGA and DTG curves obtained for the obtained using a Polymer Laboratories TG analyser with a as-synthesised purely siliceous sample (prepared according to heating rate of 20 °Cmin-1 under nitrogen flow of ref. 9) and the Al-MMS samples. The first mass loss between 25 ml min-1. Powder X-ray diVraction (XRD) patterns were 25 and 100 °C is attributed to the release of water and/or recorded using a Philips 1710 powder diVractometer with Cuethanol.All samples show a further mass loss centred at ca. Ka radiation (40 kV, 40 mA), 0.02° step size and 1 s step time. 300 °C which is due to amine desorption. In addition the Al- Textural properties (surface area, pore volume and pore size) MMS samples show a third mass loss centred at 450 °C were determined at -196 °C using nitrogen in a conventional which increases as the (synthetic gel ) Si/Al ratio reduces and volumetric technique by a Micromeritics ASAP 2400 sorptodevelops at the expense of the mass loss centred at 300 °C.meter. Before measurement each sample was oven dried at We have previously shown that for primary amine templated 280 °C and evacuated overnight at 200 °C under vacuum. 27Al aluminosilicate samples such as those described here, the magic-angle-spinning (MAS) NMR spectra were recorded at occluded amine exists in two forms, i.e., neutral ‘low 9.4 T using a Chemagnetics CMX-400 spectrometer with zirtemperature’ amine similar to that present in the purely conia rotors 4 mm in diameter spun at 8 kHz.The spectra siliceous material and electrostatically bound ‘high tempera- were measured at 104.3 MHz with 0.3 s recycle delays and ture’ amine.9 During thermogravimetric analysis the former corrected by subtracting the spectrum of the empty MAS less strongly bound template is desorbed and decomposed rotor.External Al(H2O)63+ was used as a reference. To ensure between 100 and 350 °C while the electrostatically bound quantitative reliability all calcined samples were fully hydrated template is removed between 350 and 500 °C.However, and equilibrated with room air prior to the measurements.13 despite the varying amounts of neutral and charged amine, it is clear from Fig. 1 that the total amine occluded remains Acidity measurements more or less the same regardless of the Si/Al ratio. This implies that both the neutral and charged amine are occluded The acid content of the materials was measured by using TPD of cyclohexylamine.The samples were exposed to liquid cyclo- in micellar arrangements which are encased by the inorganic framework. As more Al is incorporated into the framework hexylamine at room temperature after which they were kept overnight (at room temperature) and then in an oven at 80 °C more of the occluded amine is protonated so as to balance the resulting negative charge.for 2 hours so as to allow the base to permeate the samples. The oven temperature was then raised to 250 °C and main- Fig. 2 compares the TGA and DTG curves of the dry parent (as-synthesised) Al-MMS10 sample before and after tained at that temperature for a further 2 hours. The samples were then cooled to room temperature under dry nitrogen being subjected to various modes of template extraction.The mass loss between 150 and 800 °C, which excludes adsorbed following which they were subjected to thermogravimetric analysis using a Polymer Laboratories TG analyser with a water or ethanol, was 36% for the parent as-synthesised Al- MMS10 sample and 26, 11 and 10% for Et-, NH- and Na- heating rate of 20 °Cmin-1 under nitrogen flow of 25 ml min-1. The weight loss associated with desorption of Al-MMS10 samples respectively. Unlike the parent sample, the ethanol-extracted sample (Et-Al-MMS10) shows only one the base from acid sites occurred between 300 and 450 °C, with a maximum at ca. 370 °C. This weight loss was used to mass loss due to the desorption of electrostatically bound amine.Thus extraction in ethanol alone removes the neutral quantify the acid content (in mmol of cyclohexylamine per gram of sample) assuming that each mole of cyclohexylamine amine but has no noticeable eVect on the electrostatically bound amine. On the other hand the TGA and DTG curves corresponds to one mole of protons.11,14 2820 J. Mater. Chem., 1998, 8, 2819–2826Table 1 Elemental composition, d spacing and textural properties of calcined Al-MMS samples subjected to various modes of template removal Surface area/ Pore volume/ Wall Sample Si/Al ratio d100/A° m2 g-1 cm3 g-1 APDa/A° a0 b/A° thickness/A° Directly calcined Al-MMS40 44.1 35.1 1165 0.67 25.1 40.5 15.4 Al-MMS20 26.3 34.2 (35.8)c 1050 0.58 23.7 39.5 15.8 Al-MMS10 12.8 32.2 1104 0.47 21.2 37.2 16.0 Al-MMS5 7.6 31.6 901 0.40 20.1 36.5 16.4 Extracted prior to calcination Et-Al-MMS20 22.7 36.0 (37.0)c 941 0.81 26.4 41.6 15.2 NH-Al-MMS20 24.1 36.2 (38.4)c 894 0.71 25.7 41.8 16.1 Na-Al-MMS20 24.6 32.5 (36.0)c 731 0.53 27.8 37.5 9.7 aAPD=Average pore diameter (determined using BJH analysis). ba0=Lattice parameter, from the XRD data using the formula a0=2d100/Ó3.cValues in parentheses are d spacing before calcination.of NH- and Na-Al-MMS10 indicate that extraction in the presence of a cation removes all the occluded amine. Indeed FTIR spectroscopy confirmed the total removal of amine from samples subjected to template extraction in the presence of NH4+ and Na+ ions, i.e., we did not observe any peaks attributable to hexadecylamine. This means that the events in the TGA curve of the NH- and Na-Al-MMS10 samples are new and are not attributable to amine desorption.For Na- Al-MMS10 we attribute the mass loss centred at ca. 550 °C to dehydroxylation. The dehydroxylation starts at 200 °C and is complete at 700 °C. The DTG of the NH-Al-MMS10 sample shows two mass losses in the temperature range 200–500 °C which we attribute to the decomposition of ammonium ions.The presence of two distinct mass losses suggests two types of ammonium ions, one more stable than the other. If we consider the NH-Al-MMS10 sample as being similar to an ammonia saturated solid Brønsted acid material, with the base adsorbed as ammonium ions on Brønsted acid sites, it is then possible to ascribe the mass loss centred at 350 °C to diamination from weaker acid sites while ammonium ions held on stronger acid sites are desorbed at the higher (420 °C) temperature.This explanation is supported by Fig. 3 which shows TGA and DTG curves of all the Al-MMS samples after template extraction in the presence Fig. 1 Thermogravimetric analysis (TGA) and diVerential thermo- of NH4 ions. As the Si/Al reduces, the mass loss at 420 °C gravimetric (DTG) analysis curves for as-synthesised purely siliceous increases which is agreement with higher population of strong (MMS) and Al-MMS samples.acid sites due to increasing Al incorporation. Fig. 2 (A) Thermogravimetric analysis (TGA) and (B) diVerential thermogravimetric (DTG) analysis curves for an Al-MMS sample prepared at Si/Al=10 before and after template extraction; (a) as-synthesised, (b) extracted in ethanol alone, (c) extracted in ethanol/NH4+, (d) extracted in ethanol/Na+.J. Mater. Chem., 1998, 8, 2819–2826 2821which again emphasises the absence of long range order. As expected for such mesoporous materials,15 the intensity of the basal peak reduces with increase in Al incorporation indicating that Al has a deleterious eVect on structural ordering. Furthermore as the amount of Al incorporated increases the d spacing reduces (see Table 1).Fig. 5 compares the XRD patterns of the parent assynthesised (dry) and calcined Al-MMS20 samples with equivalent samples after being subjected to template extraction (dry samples) and subsequent calcination. For the dry samples template extraction has the eVect of increasing the intensity of the basal (100) peak.This may be related to better ordering occasioned by the benign removal of all or some of the template and formation of cross-linking siloxane or Si–OH–Al bonds. Upon calcination the intensity of the basal peak for Et-Al-MMS20 and NH-Al-MMS20 is maintained suggesting that structural ordering is retained in the calcined forms of these samples.For Na-Al-MMS20 the decrease in intensity of the basal peak is indicative of some structural collapse which in turn suggests that template free Na-containing Al-MMS materials are unstable towards calcination (see interpretation of porosity data). Similar behaviour has been previously reported for aluminosilicate Na-MCM-41 materials from which most of the template had been extracted.7 The instability of template free Na-containing Al-MMS materials is further Fig. 3 Thermogravimetric analysis (TGA) and diVerential thermogravimetric (DTG) analysis curves for Al-MMS samples after template highlighted by the extent of lattice contraction due to calciextraction in ethanol/NH4+. nation; the basal spacing of Na-Al-MMS20 reduces by 9.7% compared to 4.5, 2.7 and 5.7% for the parent (as-synthesised), Et-Al-MMS20 and NH-Al-MMS20 samples respectively (see Physical characterisation Table 1).The higher contraction for Na-Al-MMS20 suggests The XRD patterns of the as-synthesised and (directly) calcined extensive dehydroxylation which supports our interpretation parent Al-MMS materials are shown in Fig. 4. The samples of the TGA results described above.We note that calcined Etmainly exhibit a single basal (100) peak which is characteristic Al-MMS20 and NH-Al-MMS20 samples have similar d100 of materials possessing short range hexagonal ordering.8 As- spacings despite the larger lattice contraction in the latter. synthesised materials prepared at Si/Al>10 show an additional The textural parameters of the parent (directly calcined) Alweak and diVuse peak at ca. 1.8 nm which may be an indication MMS materials are given in Table 1 while Fig. 6 shows the of slightly better long range ordering in these silica-rich corresponding nitrogen sorption isotherms. The isotherms materials. On calcination the scattering intensity of the basal exhibit a mesopore filling step in the relative pressure (p/po) peak increases.The increase in scattering intensity observed range 0.05 to 0.4 which is characteristic of such materials.16 here may be due to better ordering of the inorganic framework The height and steepness of the step which are an indication (as a result template removal and formation of cross-linking of the extent and uniformity of the mesopores reduce at lower siloxane bonds) or increase in scattering domain size.However Si/Al ratios. This is due to increasing amount of framework despite the increase in the intensity of the basal peak (especially for Al-MMS40), the diVuse peak at ca. 1.8 nm disappears Fig. 5 Powder XRD patterns of dry and calcined Al-MMS samples Fig. 4 Powder XRD patterns of as-synthesised and calcined Al-MMS prepared at Si/Al=20; (a) as-synthesised, (b) extracted in ethanol alone, (c) extracted in ethanol/NH4+, (d) extracted in ethanol/Na+.samples prepared at Si/Al of (a) 5, (b) 10, (c) 20 and (d) 40. 2822 J. Mater. Chem., 1998, 8, 2819–2826Fig. 7 Nitrogen sorption isotherms of the directly calcined Al-MMS20 Fig. 6 Nitrogen sorption isotherms of directly calcined Al-MMS sample compared to equivalent samples subjected to template extracsamples.tion prior to calcination. Inset is the pore size distribution of the samples. Al which results in lower structural ordering and therefore less well defined framework confined mesoporosity. Na-Al-MMS sample is indicative of a material with lower Furthermore the mesopore filling range generally shifts to structural ordering which is in agreement with the XRD and lower partial pressures as Si/Al reduces which is an indication TGA results described earlier.The picture that emerges is that of decrease in pore size as shown in Table 1. There is no template free Na-AlMMS materials are unstable to calcination systematic variation of interparticle or ‘textural’ mesoporosity during which they undergo extensive dehydroxylation and (which is indicated by the presence of high pressure hyster- significant structural collapse. esis)16 with Si/Al ratio.The surface area, pore volume and average pore diameter (APD) given in Table 1 are consistent Nature of Al nuclei (solid state 27Al NMR) with those previously reported for similar mesoporous materials such as MCM-41.17,18 In general the pore volume and The 27Al NMR spectra of the parent as-synthesised and directly calcined samples is shown in Fig. 8. The as-synthesised pore sizes decrease as the amount of aluminium incorporated increases while the surface area does not change in any samples show a relatively sharp resonance at 53 ppm due to tetrahedrally coordinated Al. As expected this resonance systematic way except for a significant reduction for the most aluminous (Al-MMS5) sample.The decrease in pore volume increases with increasing incorporation of Al into the inorganic framework. Most if not all the Al is in tetrahedral coordination and pore size is probably due to some collapse of the structure (during calcination to remove the template) caused by local for samples prepared at Si/Al10.The sample prepared at Si/Al ratio of 5 (Al-MMS5) exhibits a broad low intensity heating eVects associated with the presence of increasing amounts of framework aluminium.12,15 The wall thickness peak at ca. 0 ppm indicating that some of the Al is in octahedral coordination, i.e., extra-framework Al (EFAL). values obtained by subtracting the average pore diameter (APD) from the lattice parameter (a0), are consistent with Calcination in air results in some dealumination in all the samples and the amount of EFAL generally increases with those previously reported for similar materials.16 The wall thickness increases slightly with increasing amounts of incor- lowering Si/Al ratio.In Fig. 9 we compare the 27Al NMR spectra of calcined template extracted samples (prepared at porated Al which is consistent with the presence of increasing amounts of Al in the framework; the Al–O bond being longer Si/Al=10) with the directly calcined sample (Al-MMS10).The spectra of the dry extracted samples (not shown) were than the Si–O bond. Fig. 7 compares the sorption isotherms of directly calcined similar to that of as-synthesised Al-MMS10 shown in Fig. 8 implying that extraction on its own had no eVect on the Al-MMS20 with equivalent materials subjected to template extraction prior to calcination. The isotherms of Et-Al-MMS20 environment of Al nuclei. The spectra of the calcined Et- and NH-Al-MMS10 samples indicate the presence of EFAL while and NH-Al-MMS20 are similar to that of the parent material indicating that these three samples have a similar extent of that of Na-Al-MMS10 exhibits only one resonance at 53 ppm due to framework Al.We may therefore infer that on calci- mesopore uniformity. The height and steepness of the mesopore filling step for Na-Al-MMS20 is lower indicating a less nation Et- and NH-Al-MMS samples undergo some dealumination while Na-Al-MMS samples apparently retain all the Al well defined mesopore structure.The inferior mesopore structure of Na-Al-MMS20 compared to the other samples is in tetrahedral coordination. In this respect the Et- and NHAl- MMS samples are similar to the directly calcined sample. confirmed by the pore size distribution curves shown in Fig. 7 (inset). Furthermore as shown in Table 1, Et- and NH-Al- The factor that distinguishes these three samples from Na-Al- MMS samples is that during calcination they undergo diamin- MMS20 samples have surface area and pore volume comparable or higher than those of the parent (directly calcined) Al- ation.In the directly calcined and Et-Al-MMS samples the decomposing species is protonated amine while for the NH- MMS20 while Na-Al-MMS20 on the other hand exhibits lower values. In addition the wall thickness of Na-Al-MMS20 Al-MMS sample the diaminating species are ammonium (NH4+) ions.Since diamination is an exothermic process it is at 9.7 A° , is much lower than those of the other three samples which have walls of thickness in the range 15.2 to 16.1 A° ; the likely that it causes local heating in the vicinity of the (framework) Al sites on which the diaminating species are attached thinner wall for the Na-AlMMS sample is consistent with structural collapse during calcination.The textural data of the resulting in the extraction of some of the Al. No such local J. Mater. Chem., 1998, 8, 2819–2826 2823Fig. 8 27Al MAS NMR spectra of as-synthesised and calcined Al-MMS samples prepared at Si/Al ratio of (from top to bottom) 5, 10, 20, 40.present during the calcination of dry relatively template free samples.7 Acidity and catalytic activity The acid content of the samples was obtained using TPD of cyclohexylamine.14,15 The results are given in Table 2. As expected the acid contents of the directly calcined materials is dependent on the Si/Al ratio and increases with the Al content. The acid contents of Et- and NH-Al-MMS samples prepared at Si/Al=10 or 20 are similar to those of the equivalent directly calcined samples. This is in agreement with the elemental composition (see Table 1) and 27Al NMR results.We may therefore assume that for Et- and NH-Al-MMS materials, template extraction prior to calcination does not have any significant eVect on the population of acid sites (mainly of the Brønsted type)14 which are strong enough to retain adsorbed cyclohexylamine after thermal treatment at 250 °C. Na-Al- MMS samples on the other hand exhibited very low acid content.This is due to the blocking of potential Brønsted acid Fig. 9 27Al MAS NMR spectra of a directly calcined Al-MMS10 sites by Na+ ions which for these samples are the charge sample compared to equivalent samples subjected to template extrac- balancing cations.Na-Al-MMS samples therefore require an tion prior to calcination. extra ion exchange (with ammonium ions) and calcination step to generate Brønsted acid sites. The conversion of cumene over directly calcined Al-MMS heating eVects exist for the Na-sample and therefore no samples as a function of time is shown in Fig. 10. The curves dealumination occurs. Furthermore it seems that amine of some reference materials are included for comparison. The decomposition causes greater dealumination than the conversion achieved over the Al-MMS samples is dependent ammonium ion; from the deconvoluted NMR spectra we on the Si/Al ratio while the rate of deactivation is comparable estimate that ca. 68% of the Al in the directly calcined and for all the samples.From Fig. 10 we may conclude that the Et-Al-MMS10 samples is in four coordination while for the activity of hexadecylamine templated Al-MMS materials is NH-Al-MMS10 sample the value is close to 78% (see Fig. 9). higher than that of aluminosilicate MCM-41 and amorphous Presumably the diamination of the larger amine molecules aluminosilicates but lower than that of ultra-stable Y (USY) causes more local heating than the decomposition of the zeolite.The activity per acid site for Al-MMS samples is given smaller ammonium ion. Our interpretation of the NMR results in Table 2 as the turnover frequency (TOF). The TOF values suggests therefore that despite lower structural stability were obtained by dividing the rate of cumene conversion towards calcination, dry template free Na-Al-MMS materials (mmol g-1 h-1) after 10 minutes time on stream with the acid are able to retain most if not all Al in tetrahedral coordination.content (mmol g-1). The TOF increases as the Al content This in turn implies that structural disintegration of Al-MMS reduces implying that the catalytic eYciency of the acid sites materials is not always linked to dealumination and for on Al-MMS materials reduces as their density increases.calcined Na-Al-MMS materials the tetrahedral Al may exist Similar trends have been observed for zeolites.19 Fig. 11 com- in a disordered (or amorphous) phase as in the case of pares the activity of the directly calcined (Al-MMS20) sample amorphous aluminosilicates.Indeed it has previously been with equivalent calcined extracted samples. The Na-Al-MMS reported that the formation of amorphous (or even crystalline samples had no observable catalytic activity under the reaction alkali metal–silicate phases) can be favoured over the formation of the sodium form of AlMCM-41 if alkali cations are conditions used. The initial conversion achieved by the Et- 2824 J.Mater. Chem., 1998, 8, 2819–2826Table 2 Acidity and catalytic activity of the study materials Cumene conversion Sample Acidity/mmol H+ g-1 Initial ratea TOFb Directly calcined Al-MMS5 650 1830 2.82 Al-MMS10 535 1780 3.31 Al-MMS20 330 1493 4.52 Al-MMS40 210 1210 5.76 Extracted prior to calcination Et-Al-MMS10 520 1936 3.72 Et-Al-MMS20 320 1760 5.50 NH-Al-MMS10 550 2002 3.64 NH-Al-MMS20 330 1905 5.77 Na-Al-MMS10 80 No catalytic activity Na-Al-MMS20 50 No catalytic activity aObtained after 10 min time on stream in mmol (g cat h)-1. bObtained by dividing initial rate by acid content.Fig. 11 EVect of template extraction on the catalytic activity and Fig. 10 Catalytic activity and deactivation behaviour of Al-MMS samples and reference materials compared at 300 °C and WHSV of deactivation behaviour of Al-MMS samples prepared at Si/Al=20 compared at 300 °C and WHSV of 4.0; Al-MMS20 (+); Et-Al- 4.0.Al-MMS5 (#), Al-MMS10 (%), Al-MMS20 (6), Al-MMS40 ((), USY (CBV740) Si/Al=21 ($), H+-MCM-41-20, Si/Al=20 (+) MMS20 (&) and NH-Al-MMS20 ($). and amorphous silica–alumina (ASA12) Si/Al=12 (&). and NH-Al-MMS extracted samples is higher than that of be achieved by direct calcination, extraction in ethanol followed by calcination or extraction in a cation-containing their directly calcined analogue; the order of initial conversion is Et-Al-MMS20>NH-Al-MMS20>Al-MMS20.This order ethanol solution. Extraction in ethanol alone removes only that part of the template which is not associated with frame- of activity is reflected by the TOF values for the samples given in Table 2.Fig. 11 also shows that the rate of deactivation is work Al while extraction in the presence of a cation results in complete template removal. Template extraction has no eVect higher for the extracted samples; this may be an indication that extracted samples possess stronger acid sites (which on elemental composition and for dry (non-calcined) samples results in an improvement in structural ordering.The eVect of therefore deactivate more rapidly) than those on the directly calcined sample. The lack of catalytic activity for the Na-Al- calcination depends on the mode of extraction. Samples subjected to template extraction in ethanol or ethanol/NH4+ are MMS20 sample is due to the absence of Brønsted acid sites which as mentioned earlier are blocked by the charge balancing structurally stable to calcination and possess textural parameters (surface area and pore volume) similar or higher than Na+ ions.for directly calcined samples. Samples subjected to template extraction in ethanol/Na+ are, on the other hand, relatively Conclusions unstable to calcination and undergo considerable structural degradation resulting in lower surface area and pore volume.Aluminosilicate mesoporous molecular sieves (Al-MMS) exhibiting short range hexagonal order have been prepared at Dealumination results from calcination of samples containing amine or ammonium ions (i.e., directly calcined, ethanol and room temperature using the primary amine hexadecylamine as structure director.The templating amine is occluded in the ethanol/NH4+ extracted samples) while Na containing (i.e., ethanol/Na+ extracted) samples do not appear to undergo materials either in neutral or charged (protonated) form. The relative amounts of neutral and protonated amine depend on any dealumination. We conclude therefore that removal of aluminium from framework positions is due to local heating the Si/Al ratio.The proportion of protonated amine increases with increasing framework Al content. Template removal may eVects arising from diamination of protonated amine or J. Mater. Chem., 1998, 8, 2819–2826 2825P. Sieger, R. Leon, P. M. PetroV, F. Schuth and G. D. Stucky, ammonium ions. The acid content of calcined ethanol and Nature, 1994, 368, 317.ethanol/NH4+ extracted samples is comparable to that of 4 J. S. Beck and J. C. Vartuli, Curr. Opin. Solid State Mater. Sci., directly calcined samples but the former exhibit higher catalytic 1996, 1, 76. activity for the cracking of cumene. Calcined ethanol/Na+ 5 C.-Y. Chen, H.-X. Li and M. E. Davis, Microporous Mater., 1993, extracted samples possess very low acidity and exhibit no 2, 17. 6 R. Schmidt, D. Akporiaye, M. Stocker and O. H. Ellestad, Stud. catalytic activity. Surf. Sci. Catal., 1994, 84, 677. 7 S. Hitz and R. Prins, J. Catal., 1997, 168, 194. 8 P. T. Tanev and T. J. Pinnavaia, Science, 1995, 267, 865. 9 R. Mokaya and W. Jones, Chem. Commun., 1996, 981. Aknowledgements 10 R. Mokaya and W. Jones, Chem. Commun., 1996, 983. 11 R. Mokaya and W. Jones, J. Catal., 1997, 172, 211. R.M. is grateful to the EPSRC for an Advanced Fellowship 12 A. Tuel and R. Gontier, Chem. Mater., 1996, 8, 114. and Trinity College, Cambridge for a Research Fellowship. 13 R. J. Ray, B. L. Meyes and C. L. Marshall, Zeolites, 1987, 7, 307. The assistance of Dr H.Y. He with the NMR measurements 14 C. Breen, Clay Miner., 1991, 26, 487. and Laporte Adsorbents with surface area determinations is 15 R. Mokaya, W. Jones, Z. Luan, M. D. Alba and J. Klinowski, Catal. Lett., 1996, 37, 113. appreciated. 16 P. T. Tanev and T. J. Pinnavaia, Chem.Mater., 1996, 8, 2068. 17 P. J. Branton, P. G. Hall, K. S. W. Sing, H. Reichert, F. Scuth and K. K. Unger, J. Chem. Soc., Faraday Trans., 1994, 90, 2965. 18 J. Rathousky. A. Zukal, O. Franke and G. J. ShulzekloV, J. Chem. References Soc., Faraday Trans., 1994, 90, 2821. 19 M. J. Remy, D. Stanica, G. Poncelet, E. J. P. Feijen, P. J. Grobet, 1 A. Corma, Chem. Rev., 1997, 97, 2373. J. A. Martens and P. A. Jacobs, J. Phys. Chem., 1996, 100, 12440. 2 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992, 359, 710. 3 Q. S. Hue, D. I. Margolese, U. Ciesla, P. Y. Feng, T. E. Gier, Paper 8/06049E 2826 J. Mater. Chem., 1998, 8, 2819–2826