Samples containing allophane with molar Al/Si ratios from 1.0 to 2.0 have been examined by infrared spectroscopy, X-ray fluorescence, and phosphate adsorption methods. The infrared spectra of allophane with Al/Si ratios close to 2.0 showed that the wall of the allophane spherules is made up of imogolite structural units similar to “proto-imogolite.” X-ray fluorescence gave no clear evidence of Al in tetrahedral sites (AlIV), while pyridine adsorption results suggested that a small number of Bronsted acid sites (AlIVOH) are present in silica-rich allophanes. Lewis acid sites (AlH2O) are present in both silica-rich and alumina-rich allophanes. The results suggest that the framework for the allophane structure is an Al octahedral sheet.

Allophanes with Al/Si ratios close to 1.0 contain condensed silicate units either on the outside surface of the Al octahedral sheet, giving rise to a halloysite-like structure, or bonded on the inside surface of an imogolite-like structure. Allophanes with Al/Si ratios between 1.0 and 2.0 appear to be mixtures of the “proto-imogolite” structure and the allophane (Al/Si = 1.0) structure.

Imogolite occurs as tiny bumps less than 0.05 µm in diameter and about 0.3 µm long on amorphous thin layers on the surface of weathered plagioclase. The bumps grow outward from the surface to form projections, which then develop into a fibrous imogolite. The fibers branch out and grow into widespread networks or thin films which finally cover the entire surface of the plagioclase grain. The fibers are about 0.05 µm in diameter as seen by transmission electron microscopy.

Allophane samples from soils, pumice, and stream beds have been studied by electron optical, infrared absorption, X-ray fluorescence, gas chromatography, and phosphate adsorption methods. The allophane particles were hollow spherules or polyhedra 35 and 50 Å in diameter with molar Al/Si ratios close to 2.0. The thickness of the wall of the spherules was estimated to be 7 Å. For the pumice allophanes, the wall was largely composed of imogolite structural units (OH)3Al2O3SiOH. Defects or pores were present in the wall and probably were the sites where phosphate was adsorbed. It is suggested that these allophanes with molar Al/Si ratios close to 2.0 should be called “proto-imogolite” allophane.

Two soil allophanes had a similar structure to the allophane from Stratford pumice, but small amounts of layer silicates, including halloysite, were also present in the soil samples, as indicated by infrared bands at 470, 1030, and 1100 cm−1. The allophane from the stream bed at Silica Springs had an infrared spectrum similar to feldspathoids, and it did not have the imogolite structure.

Allophanes with SiO2/Al2O3 molar ratios from 1.38 to 1.92 were heated at temperatures up to 500°C, and the changes induced were investigated by means of infrared spectroscopy (IR) and X-ray powder diffraction (XRD). Heat treatment caused the IR absorption band near 1000 cm−1 due to Si-O stretching to shift towards higher frequencies, and the band near 450 cm−1 due to O-Si-O bending to increase markedly in intensity. These results are probably due to condensation of SiO tetrahedra, following breakdown of Si-O-Al linkages and dehydroxylation of SiOH groups. Concurrent intensity increase of the IR absorption at about 700 cm−1 suggested an increase in the amount of 4-coordinated Al in the heated materials. ‘Imogolite structures’ were decomposed progressively at higher temperatures and were almost absent at about 400°C, as indicated by the weakening and disappearance of the IR band at 348 cm−1 and of the XRD reflections at 2.25 and 1.40 Å. The XRD reflection at 3.3–3.45 Å shifted to about 3.6 Å, probably as a result of thermal condensation of the silica component. The observed thermal changes were much greater for samples with low SiO2Al2O3 ratios (1.38–1.51) than for those with high ratios (1.81–1.92), indicating a lower thermal stability for the former materials. Thus, the thermal stability of allophanes appears to be related to the content of ‘imogolite structures’ and to the polymerization status of the silica component.

The adsorption of sodium, chloride, and phosphate ions by allophane, imogolite, and halloysite has been studied in relation to the surface structure of the mineral samples. The high adsorption of phosphate (>200 μmole/g) and chloride (10–30 meq/100 g at pH 4) by allophane is ascribed to the small particle size of allophane, its high surface area (∼800 m2/g), and the presence at the surface of Al-OH-Al groups and defect sites. In contrast, halloysite has a relatively large particle size and a Si-O-Si surface. Accordingly, the adsorption of phosphate (5–10 μmole/g) and chloride (1 meq/100 g) by halloysite is very much lower as compared with allophane. Phosphate adsorption by halloysite is also related to particle morphology and the number of edge sites. Thus, a sample consisting entirely of spheroidal particles adsorbed only 5 μmole/g at a solution concentration of 1 × 10−4 M, whereas the tubular types of comparable surface area adsorbed 7–10 μmole/g at the same concentration. This is because spheroidal halloysite particles have few, if any, edge sites at which phosphate can adsorb. The relative degree of order and hydration of halloysite, as indicated by infrared spectroscopy, also affects phosphate adsorption. However, this factor is apparently less important than particle morphology and surface structure. Although imogolite also has an Al-OH-Al surface, it contains relatively few defect sites where phosphate can adsorb. Consequently, much less phosphate (120 μmole/g) was adsorbed as compared with allophane.

Imogolite was synthesized at 25°C by aging partially neutralized solutions containing monomeric silicic acid and polymeric hydroxy-aluminum ions for 7 years. Solutions having an initial Si/Al molar ratio of about 0.5 and pHs of 4.0–4.5 produced the largest yields of imogolite, followed by those having an initial Si/Al ratio of about 1, although imogolite was not the principal product. Electron microscopic examination showed a small amount of imogolite fibers embedded in a noncrystalline gel-like substance. Traces of imogolite were detected in solutions having an initial Si/Al ratio of about 2, but no imogolite was found by electron microscopy in products from solutions having an initial Si/Al ratio of about 4. Only gibbsite formed from solutions having initial Si/Al ratios of <0.27. The diameter of the tubular structural unit of the imogolite produced in these experiments was 23 ± 2Å, close to that of natural imogolite.

Small-angle X-ray powder diffraction analyses and high-resolution electron microscopy of allophane samples (SiO2/Al2O3 ratio, 1.12 to 1.68) showed that allophanes consist of nearly identical spherical particles with diameters of about 40 Å and retain their characteristic “hollow” spherical morphology at different ambient moisture and even after dehydroxylation by heating at 500° to 600°C. Unheated allophane samples gave another X-ray powder diffraction band whose maximum position varied from 12.3 to 14.5 A depending on their SiO2/Al2O3 ratio. The appearance of this band may denote some long-range ordering in the structure of allophane. Unlike the spherical particles of allophane, the tube unit of imogolite collapsed on dehydroxylation. This observation suggests that imogolite and allophane are different in their framework structures and that a Si- or Si(Al)-tetrahedral sheet rather than an Al-octahedral sheet constitutes the framework structure of allophane, irrespective of its SiO2/Al2O3 ratio.

Reactions between hydroxy-Al ions and orthosilicic acid as influenced by citric acid were studied at an initial Si concentration of 1.6 × 10−3 M, Si/Al molar ratios of 0.5 and 1.0, OH/Al molar ratios of 1.0-3.0, and citric acid/Al molar ratios of 0-0.3. In the absence of citric acid and at OH/A1 ratios of 1.0-2.8, imogolite (>0.01 μm) was dominant in the precipitates. At citric acid/Al ratios of 0.01-0.1, imogolite and/or pseudoboehmite were dominant in the precipitates at OH/Al ratios of 1.0 and 2.0, and imogolite and/or ill-defined aluminosilicate complexes at OH/Al ratio of 2.8. Instead of allophane or “proto-imogolite” allophane being the predominant species in the precipitates, the formation of ill-defined aluminosilicate complexes at OH/Al ratio of 3.0 was steadily promoted by increasing the solution citric acid/Al ratios from 0 to 0.3. The freeze-dried soluble products (<0.01 μ) ranged from silica gel to “proto-imogolite,” depending upon the basicity of Al and the level of citric acid of the parent solution. The amount of “proto-imogolite” increased with increasing citric acid/Al ratios from 0 to 0.1 in solution. Complexing low molecular weight organic acids, such as citric acid, impeded the formation of the short-range ordered aluminosilicates, allophanes and imogolite.

Adsorption of Cu2+ and Co2+ by synthetic imogolite, synthetic allophanes with a range of SiO2/ Al2O3 ratios, and allophanic clay fractions from volcanic ash soils was measured in an ionic medium of 0.05 M Ca(NO3)2. The effect of pH (and metal concentration) on adsorption was qualitatively similar for the synthetic and natural allophanes with relatively minor changes in behavior caused by variable SiO2/Al2O3 ratios. Cu and Co were chemisorbed by allophane at pH 5.0–5.5 and 6.9–7.2 (pH values for 50% adsorption level), respectively, with concomitant release of 1.6–1.9 protons/metal ion adsorbed. Quantitatively, adsorption by imogolite was less than that by the allophanes, presumably because of fewer sites available for chemisorption on the tubular structure of imogolite. Electron spin resonance studies of the imogolite and allophanes revealed that Cu2+ was adsorbed as a monomer on two types of surface sites. The preferred sites were likely adjacent AlOH groups binding Cu2+ by a binuclear mechanism; weaker bonding occurred at isolated AlOH or SiOH groups. These chemisorbed forms of Cu2+ were readily extracted by EDTA, CH3COOH, and metals capable of specific adsorption, but were not exchangeable. In addition, the H2O and/or OH− ligands of chemisorbed Cu2+ were readily displaced by NH3, with the formation of ternary Cu-ammonia-surface complexes.

Hollow spherical particles of aluminosilicate having diameters of 80 to 100 Å and high electric charge were synthesized by heating solutions containing Si and Al concentrations of about 2 mM (Si/Al molar ratio = 1.0–1.2; pH = 9.0) at 95°–100°C for 7 days. Electron microscopy, X-ray powder diffraction, and infrared spectroscopy suggested that the walls of the spherical particles consisted of monolayer aluminosilicate, most probably having a 1:1 layer structure. Chemical analysis supported this supposition but indicated modifications as shown in the formula (Si1.85Alo0.62)IV(Al1.38)VI(OH)4.10O4.96Na0.62, compared with (Si2)lv(Al2)VI(OH)4O5 for an ideal 1:1 layer silicate. High-resolution 29Si and 27A1 nuclear magnetic resonance spectroscopy indicated the substitution of Al for Si in the tetrahedral sheet, the presence of A1IV and A1VI, and the presence of Si bonded through oxygen to three Al atoms and one H. A natural counterpart of the synthesized aluminosilicate may be a precursor of halloysites formed by weathering of volcanic ash and pumice and have a close structural relation with allophane and imogolite

Imogolites in which a part or all of Si was substituted with Ge were synthesized from solutions containing aluminum ion and silicic and/or germanic acid. The products were similar to natural imogolite in their tubular morphology, electron diffraction patterns, and differential thermal analysis curves. However, the external diameter of the tube increased with increasing Ge substitution up to about 33 Å, and associated changes appeared in the X-ray powder diffraction patterns, suggesting that the substitution of Ge for Si caused a decrease in the curvature of the gibbsite sheet with which SiO- or GeO-tetrahedra are associated. Infrared absorption bands at 995 and 930 cm−1 in imogolite disappeared and new bands appeared at 910 and 810 cm−1 on substitution of Ge for Si, whereas those in the region between 700 and 300 cm−1 remained unchanged or changed little. The former two bands were assigned to Si-O vibrations and the latter bands to Al-O vibrations.

Solid-state nuclear magnetic resonance (NMR) spectroscopy, thermal analysis, and X-ray powder diffraction data on the tubular, hydrous aluminosilicate imogolite were found to be fully consistent with a previously proposed crystal structure consisting of a rolled-up, 6-coordinate Al-O(OH) sheet, bonded to isolated orthosilicate groups. The calculated 29Si chemical shift of this structure agreed with the observed shift within 3 ppm. Thermal dehydroxylation of the Al-O(OH) sheet produced predominantly NMR-transparent 5-coordinate Al, but a few 4- and 6-coordinate sites and some residual hydroxyl groups may also have formed, as shown by NMR spectroscopy. Changes in the 29Si NMR spectrum on dehydroxylation suggest a condensation of the orthosilicate groups, but steric considerations rule out bonding between adjacent silicons. To account for these observations, an alternative mechanism to orthosilicate condensation has been proposed, involving the fracture and unrolling of the tubes, followed by the condensation of fragments to form a layer structure. The layer structure has a calculated 29Si chemical shift of -95.6 ppm, in good agreement with the observed value of -93 ppm.

The negative surface charge of synthetic allophanes with a range of Si/Al ratios decreased and positive charge increased with increasing alumina content at a given pH. The phosphate adsorption capacity also increased with increasing Al content. That this relationship between composition and chemical reactivity was not found for the soil allophanes is attributed to the presence of specifically adsorbed organic or inorganic anions on the natural material. Both synthetic and natural imogolites had a much lower capacity to adsorb phosphate than the allophanes and adsorbed anomalously high amounts of Cl− and ClO4− at high pH. It is proposed that intercalation of salt occurs in imogolite, although electron spin resonance studies using spin probes failed to reveal the trapping of small organic molecules in imogolite tubes. These spin probes in the carboxylated form did, however, suggest an electrostatic retention of carboxylate by imogolite and a more specific adsorption by allophane involving ligand exchange of surface hydroxyl. The results illustrate the inherent differences in charge and surface properties of allophane and imogolite despite the common structural unit which the two minerals incorporate.

Examination of two volcanic and two precipitated allophanes by solid-state NMR, thermal analysis and X-ray powder diffraction shows three of the samples to contain structural features similar to both tubular imogolite and defect layer-lattice aluminosilicates such as kaolinite. The fourth allophane, a precipitated sample from New Zealand, had no imogolite-like features and contained tetrahedral as well as octahedral aluminum. The imogolite-like units in allophane are less stable thermally than tubular imogolite. The NMR spectra and their changes on heating can be accounted for by a structural model in which a two-sheet, kaolinite-like structure containing defects (holes in the tetrahedral sheet) is curved into a sphere in which imogolite-like orthosilicate units are anchored into the octahedral sheet and fit into the tetrahedral defects. Computer simulation shows that the model is crystallographically sound, and accounts for all the known facts, including the spherical morphology, the solid-state NMR spectra and the thermal dehydroxylation behavior of all except the New Zealand allophane, which is of a different structural type.

The dielectric behavior of kaolinite, montmorillonite, allophane, and imogolite samples adjusted to a water potential of 33 kPa was examined using a time-domain reflectometry method over a wide frequency range of 103-1010 Hz. A dielectric relaxation peak owing to bound H2O was observed. The observation of this peak required the precise determination of the contributions of dc conductivity. The peak is located at 10 MHz, indicating that the relaxation time of the bound H2O is approximately ten times longer than the relaxation time of bound H2O with organic polymers, such as an aqueous globular-protein solution. The structure of bound H2O differs between phyllosilicates and amorphous phases, based on differences in relaxation strength and the pattern of distribution of the relaxation times. The dielectric process involving rotation of bulk H2O molecules was also observed at 20 GHz. The relaxation strength of bulk H2O increased with an increase in the water content. The interfacial polarization in the diffuse double layer occurred only in montmorillonite and kaolinite, indicating that mechanisms involving the Maxwell-Wagner and surface-polarization effects cannot be extended to include allophane and imogolite. Although these results suggest that additional work is required, a tentative conclusion is that a tangential migration of counter-ions along clay surfaces may be important.

Electrophoretic mobility of imogolite has been reported as positive (migration toward the negative electrode) below pH 9, and zero above pH 9. However, when mobility of dilute imogolite suspensions (5 × 10−3 kg/m3) was measured, it was found to be negative above pH 9. The reason that imogolite does not behave as a negative colloid when the clay concentration is not very dilute is because the imogolite forms floccules large enough to prevent migration. Imogolite has a PZNC at about pH 6, and has a PZC at pH 8.5–9.0 showing a relatively low absolute mobility under alkaline conditions compared to that under acid conditions. The fact that imogolite behaves like this is understandable given the location of negative charge appearing on the inside surface of the thin fibrous tube, according to the structural model of imogolite.

Water retention curves of colloidal allophane and imogolite with different charges and different pretreatments were measured using a tension plate and a pressure plate apparatus. An irreversible effect of air-drying was found for the water retention of colloidal allophane even for particles of less than 50 nm in Stokes’ diameter collected from air-dried soil. This was attributed to the irreversible submicroscopic aggregation of allophane. Fresh allophane colloids with a low absolute net charge retained more water above −100 J Kg−1 than did highly charged ones due to the formation of a microporous structure. Allophane did not swell under ordinary conditions, but the relatively highly charged allophane recovered water retention above −100 J kg−1 during the wetting process. Imogolite retained 1.5 times more water than Na-montmorillonite at about −650 J kg−1 due to micropores formed by intertwining fibrous particles.

Imogolite is a tubular aluminosilicate which is common in Andosols and Spodosols. The high pH at point-of-zero charge at the outer parts of the tube and the anomalously high chloride adsorption of imogolite suggested that there may be structural charge associated with this mineral. The structural charge may arise because of changes in bond valence imposed by the incorporation of orthosilicate anions in a gibbsite-type sheet. By using a Basic Stem Model approach, it is shown that the surface charge properties of imogolite are explained if the mean Al-O bond valence of the outer -Al2OH groups is higher than the inner -Al2OHSiO3 groups. Hence, a weak positive charge is developed on the outer tube walls whereas a negative charge develops in the tubular pores. The best model fits were obtained where either one or two units of structural charge per unit cell of tube were assumed. The model may also explain why imogolite tubes are normally aggregated in large bundles in close hexagonal packing, because bound counterions may hold the tubes together. However, to arrive at good model descriptions, the deprotonation of -Al2OH groups must occur at a higher pH than that expected when assuming that all surface oxygens form two hydrogen bridges with H2O. A more precise structure of imogolite is required to test fully this hypothesis.