Current Applied Physics 4 (2004) 167–170 www.elsevier.com/locate/cap
Nanoporous inorganic materials from mineral templates K.J.D. MacKenzie a
, K. Okada b, J. Temuujin
MacDiarmid Institute for Advanced Materials and Nanotechnology, Victoria University of Wellington, P.O. Box 600, Wellington, New Zealand b Department of Metallurgy and Ceramic Science, Tokyo Institute of Technology, Tokyo 152-8552, Japan c Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences, Ulaanbaatar 51, Mongolia
Abstract The crystal architecture of layer-lattice minerals has been used to prepare inorganic materials containing pores of closely controlled size and shape. The template minerals are activated, either by thermal or mechanical treatment, and selectively leached with acid to remove the octahedrally-coordinated component, or with alkali to remove the tetrahedral (silica) component. Some acidleached products contain hydrophilic slit-shaped pores with narrow size distributions typically in the nanopore range, whereas the alkali-leached products contain unimodal pore distributions of larger size (2–3 nm). The leaching by-products have also been used to þ prepare combined anion and cation exchange material potentially useful for simultaneous removal of PO 4 and NH4 from waste waters. Ó 2003 Elsevier B.V. All rights reserved. PACS: 81.05.Rm; 81.05.Zx; 82.56.Hg; 83.80.Nb Keywords: Selective chemical leaching; Layer-lattice minerals; Nanoporous materials
1. Introduction The ability to form porous materials with closely reproducible pore sizes is of considerable interest for applications as catalysts and adsorption agents. It is an added bonus to be able to control closely the shape and surface properties of the pores (whether they are hydrophilic or hydrophobic); this determines their ultimate applications, which range from gas absorbers through supports for cracking catalysts to humidity regulators. Mineral structures contain a variety of repeating units which can be used as templates for the creation of nanoporous materials. This idea is not new, having been exploited previously in the development of the pillared clay materials (self-assembled materials in which minerals of both layered and column structures were assembled to produce a mix of galleries and pores ). Close control of the shape and size of these galleries presented diﬃculties, however, and their early promise does not seem to have been realised fully. However, * Corresponding author. Current address: Industrial Research Ltd., P.O. Box 31-310, Lower Hutt, New Zealand. E-mail addresses: [email protected]
, [email protected]
vuw.ac.nz (K.J.D. MacKenzie).
1567-1739/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.cap.2003.10.023
other possibilities presented by the crystal architecture of layer-lattice minerals are now being exploited. The lattice dimensions of a typical layer aluminosilicate (kaolinite) contain a repeat distance of 0.72 nm between adjacent octahedral layers (Fig. 1A). In other layer lattice minerals this repeat distance varies with the stacking sequence and mineral composition, but in all cases falls in the nano region, thus presenting the prospect of preparing nanoporous structures from these minerals. Since the crystal architecture of the layer-structured minerals is the key to these nanoporous materials, some of the possible variations in these structures minerals should be considered. Typically they all consist of sheets, composed either of tetrahedral (predominantly silicate) layers coordinated to oxygen or hydroxyl groups, or octahedral layers similarly coordinated. There is a greater variety of elements making up the octahedral sheets; aluminium and magnesium are the most common, but these can be substituted to a greater or lesser extent by other elements such as iron. Substitution of silicon in the tetrahedral layers, for example by aluminium, is also possible. In this case, the charge imbalance caused by replacement of quadrivalent Si by trivalent Al must be neutralised by the presence of, say, monovalent Na or K or divalent Ca, usually located
K.J.D. MacKenzie et al. / Current Applied Physics 4 (2004) 167–170
Fig. 1. Schematic representation of the key dimensions in the alternating layer structure of 1:1 aluminosilicate kaolinite, (A) before thermal activation, (B) after thermal activation to metakaolinite, (C) after acid leaching. Adapted from Ref. .
between the layers. The octahedral/tetrahedral (Oh/Td) stacking sequence can also vary, with Td/Oh/Td/Oh (1:1) and Td/Oh/Td/Td/Oh/Td (2:1) sequences being the most common. Now let us consider the consequence of removing a signiﬁcant portion of the octahedral cations of a 1:1 layer mineral in such a way as to leave the tetrahedral layers intact. In principle, the silicate sheets may not fully collapse on removal of the octahedral layers, but may form a structure with a repeat distance of 0.66 nm (Fig. 1C) . Furthermore, removal of the octahedral cations will also essentially remove all the structural (hydroxyl) water since this is associated predominantly with the octahedral layer. The interlayer distance should therefore collapse to 0.54 nm .
2. Selective acid leaching of layer mineral structures Selective removal of octahedral cations from layerlattice aluminosilicates has been achieved by exploiting
the chemistry of the c-form of aluminium oxide, which, unlike the a-form (corundum) is relatively soluble in strong acids such as H2 SO4 . To facilitate the acid-leaching process, the clay is ﬁrst activated by heating to about 550 °C to remove the coordinated hydroxyl groups. Solid state MAS NMR shows that this process changes the Al coordination from octahedral Al(VI)–O to a mixture of predominantly Al(V)–O and tetrahedral Al(IV)–O (Fig. 2) , both of which states show increased sensitivity to acid treatment. An alternative activation method is high-energy grinding which degrades the crystal structure, reduces the particle size, and, as with thermal activation, lowers the coordination number of the Al (Fig. 2) . Chemical analysis at regular intervals indicates the smooth removal of the aluminium; solid state MAS NMR of the solid product shows that the acid attacks the Al(V) ﬁrst, followed by Al(IV), followed by Al(VI). Our studies of the acid leaching behaviour of a variety of layer-lattice minerals activated by both thermal and mechanical (grinding) methods, indicate 1. The chemical composition of the mineral determines its structure and hence its reactivity to acid . 2. 2:1 layer lattice minerals generally produce silicas of greater surface area, but may be more resistant to acid leaching. 3. In 2:1 layer-lattice minerals, the interlayer cations are removed preferentially, followed by the removal of the octahedral cations . The tetrahedral network is the most stable to acid, even where substitution of Si4þ by Al3þ is present. 4. Minerals with some tetrahedral substitution are less stable to acid attack than minerals with no tetrahedral substitution .
Fig. 2. 27 Al MAS NMR spectra acquitted at 11.7 T of kaolinite, before and after thermal and mechanical activation. Tetrahedral sites resonate at 53–57 ppm, octahedral at 2–4 ppm. The resonance at 28–30 ppm has been attributed to pentacoordinated Al–O, while the sharp peak at 13 ppm is a-alumina (corundum) from the grinding media.
K.J.D. MacKenzie et al. / Current Applied Physics 4 (2004) 167–170
3. The porous properties of acid-leached layer-lattice minerals Gas adsorption measurements have been used to determine the pore size distribution and surface areas of a number of the silica products prepared by acid leaching. Thermally-activated kaolinite (a 1:1 layer aluminosilicate) shows a very sharp distribution proﬁle at about 0.6 nm (Fig. 3A), in agreement with the value predicted for the complete removal of the octahedral component (Fig. 1C) . Analysis of the adsorption hysteresis curve shape indicates that the pores are slitshaped, and their behaviour towards other absorbates indicates that they have hydrophilic surface properties . This is in contrast to zeolites which have cylindrical pores but can be either hydrophilic or hydrophobic in nature, and activated carbons and pillared clays which have slit-shaped pores but are hydrophobic. The slitshaped pores in acid-leached kaolinite arise from the fact that the thermal activation does not completely disrupt the original layer morphology, as can be seen from electron micrographs of the leached products. By contrast, mechanically activated kaolinite shows a sharp pore size distribution of 1.9 nm when leached, with evidence of some additional micropores (0.9 nm radius) and some larger mesopores [5,6]. The latter may result from shearing of the kaolinite structure during grinding, facilitating separation of the plates during subsequent leaching, with the formation of stable mesopores between the primary kaolinite particles [5,6]. By comparison, the corresponding 2:1 layer lattice aluminosilicate pyrophyllite is very resistant to acid leaching even when activated by grinding, which produces a wide range of pore sizes but with more than one
Fig. 3. Pore size distributions determined from gas adsorption measurements of leached kaolinite. (A) Thermally activated at 550 °C and leached in 20 mass% H2 SO4 at 90 °C for the indicated times. (B) Thermally activated at 980 °C and leached in 4 M KOH for 1 h at the indicated temperatures. Adapted from Refs. [2,11].
distribution peak in the nanoporous region (about 1.5 and 3.5 nm in samples ground for 18 h) . The analogous 2:1 magnesium silicate talc is even more resistant to acid leaching, even when thermally activated at 1000 °C, giving a silica product of low surface area and large mesopores (10–20 nm radius) corresponding to the interparticle rather than the intraparticle spaces . Mechanically activated talc is more readily leached but the porous properties of the resulting silica-rich product are not improved, showing a broad particle size distribution of both micropores and macropores . However, acid leaching of the other 2:1 layer minerals phlogopite mica and vermiculite proceeds more readily and produces nanoporous silica with bimodal pore distributions of micropores and macropores (0.7 and 4 nm respectively in phlogopite  and 0.6 and 2 nm in vermiculite ). The micropores are slit-shaped and correspond to the spaces left by removal of the octahedral layer, while the larger mesopores probably result from the formation of a framework structure by rearrangement and condensation of the silica tetrahedral.
4. Selective alkali leaching of layer aluminosilicates Nanoporous materials can also be prepared by leaching aluminosilicate minerals with alkali, exploiting the fact that the strongly polarised Si–O bonds are susceptible to attack only by nucleophillic reagents such as OH (alkali) and ﬂuorine or ﬂuorides. Thus, the silica component will be preferentially dissolved by alkali such as KOH, leaving the alumina component essentially intact. To facilitate this reaction in the 1:1 mineral
Fig. 4. Schemes for the utilisation of the various products and byproducts of selective leaching of kaolinite.
K.J.D. MacKenzie et al. / Current Applied Physics 4 (2004) 167–170
Fig. 5. Simultaneous uptake characteristics of (A) NHþ 4 and (B) PO4 by an alumina/potassium aluminosilicate composite formed from leached kaolinite.
kaolinite, the association between the alumina and silica components must be loosened by thermal treatment at 980 °C . The silica-rich regions can then be dissolved in KOH, leaving behind intact particles of c-alumina which form agglomerates with a narrow unimodal pore radius distribution of 2–3 nm (Fig. 3B)  suitable as a catalyst support for the Claus Process, in petroleum reﬁning or in automobile catalytic converters.
5. Ion exchange compounds as leaching by-products Since the alkaline leaching solution remaining after the preparation of porous alumina contains dissolved Si, K and a small amount of Al, it is a potentially useful material in its own right. When neutralised with acid, amorphous potassium aluminosilicate is precipitated , which we have found to be nanoporous with useful cation exchange capacity for the ammonium ion (a signiﬁcant polluting agent in some waste waters). Possibilities for the utilisation of the acid and alkali leaching products are summarised schematically in Fig. 4. If the porous c-alumina is not separated out before acidiﬁcation, a composite material can be formed consisting of potassium aluminosilicate and c-alumina . This composite has the interesting and unique property of functioning as a simultaneous cation and anion exchanger (Fig. 5) with special aﬃnity for ammonium and phosphate ions , two of the most common pollutants of waste water, especially in rural and agricultural areas.
This may well constitute one of the ﬁrst of a new class of weapons in the ﬁght to protect the environment for future generations.
References  G.W. Brindley, R.E. Sempels, Clay Miner. 12 (1977) 229.  K. Okada, A. Shimai, T. Takei, S. Hayashi, A. Yasumori, K.J.D. MacKenzie, Microporous Mesoporous Mater. 21 (1998) 289.  K.J.D. MacKenzie, M.E. Smith, Multinuclear Solid State NMR of Inorganic Materials, Pergamon Materials Series 6, PergamonElsevier, Oxford, 2002.  J. Temuujin, K. Okada, T.S. Jadambaa, K.J.D. MacKenzie, J. Amarsanaa, J. Eur. Ceram. Soc. 23 (2003) 1277.  J. Temuujin, K. Okada, K.J.D. MacKenzie, Ts. Jadambaa, Powder Technol. 121 (2001) 259.  J. Temuujin, G. Burmaa, J. Amgalan, K. Okada, Ts. Jadambaa, K.J.D. MacKenzie, J. Porous Mater. 8 (2001) 233.  K. Okada, J. Temuujin, Y. Kameshima, K.J.D. MacKenzie, Clay Sci., submitted for publication.  J. Temuujin, K. Okada, Ts. Jadambaa, K.J.D. MacKenzie, J. Amarsaana, J. Mater. Sci. Lett. 21 (2002) 1607.  K. Okada, N. Nakazawa, Y. Kameshima, A. Yasumori, J. Temuujin, K.J.D. MacKenzie, M.E. Smith, Clays Clay Mins. 50 (2002) 624.  J. Temuujin, K. Okada, K.J.D. MacKenzie, Appl. Clay Sci. 22 (2003) 187.  K. Okada, H. Kawashima, Y. Saito, S. Hayashi, A. Yasumori, J. Mater. Chem. 5 (1995) 1241.  J. Temuujin, K. Okada, K.J.D. MacKenzie, Appl. Clay Sci. 21 (2002) 125.  K. Okada, J. Temuujin, Y. Kameshima, K.J.D. MacKenzie, Mater. Res. Bull. 38 (2003) 749.