When dissolved aluminum reacts with high-surface-area amorphous silica in NaCl solutions at 25–120 °C, the solids that form are not the thermodynamically stable phases (gibbsite or kaolinite) but X-ray-amorphous metastable aluminosilicates: an octahedrally coordinated Al (Al[6]Al^{[6]}) phase with a kaolinite-like hydration environment and, increasingly at higher temperature, a dominant tetrahedrally coordinated Al (Al[4]Al^{[4]}) phase whose 27^{27}Al{1^{1}H} REDOR (rotational-echo double-resonance NMR, an experiment that measures the proximity of aluminum to nearby hydrogen) dephasing resembles that of the zeolite analcime (a sodium aluminosilicate framework mineral with rigidly bound structural water). The identification, however, rests on comparisons with model minerals rather than on direct structure determination. The open problem is to determine the atomic structures of these metastable phases — including whether the Al[4]Al^{[4]} phase is a hydrous sodium aluminosilicate gel with a tectosilicate framework (a metastable precursor to zeolitic phases such as analcime, or even a hydrous albite-like \ceNaAlSi3O8⋅H2O\ce{NaAlSi3O8\cdot H2O}) or Al occluded in amorphous silica; whether sodium is a structural charge-balancing cation or mainly surface-sorbed; the identity of the kaolinite-like Al[6]Al^{[6]} phase and its relation (if any) to allophane-type material (allophane: a natural short-range-ordered hydrous aluminosilicate) — and to establish the mechanism by which these phases nucleate and grow at the silica surface, including whether aluminum-hydroxide precursors are required.

Background

Aluminum is the third most abundant element in the Earth's crust, and in low-temperature waters its transport is tightly coupled to silicon through the formation of aluminosilicates. In natural and engineered systems — aquifers, soils, geothermal and petroleum production, nuclear-waste glass corrosion — dissolved Al at micromolar to millimolar concentrations interacts with amorphous (non-crystalline) silica surfaces, and the identity of the solid products controls Al mobility and silica reactivity. Prior to the work that raised this question, interpretations of Al uptake on silica relied on thermodynamic expectations that stable aluminum hydroxides or clays should dominate; Houston, Herberg, Maxwell and Carroll (2008) had shown by solution chemistry and NMR that Al associates with silica in ways that simple sorption or gibbsite (\ceAl(OH)3\ce{Al(OH)3}) precipitation could not explain.

The source study (Mason, Maxwell and Carroll, 2011, Geochim. Cosmochim. Acta 75, 6080–6093) reacted dissolved Al with high-surface-area amorphous silica (about 306 m2/gm^2/g) in a 0.1 M NaCl background electrolyte. The 25 and 50 °C flow-through experiments ran at low dissolved-Al concentrations (0.7–1.5 µM) at initial pH 4.3 and 6.3; the 120 °C batch experiments (up to 15 days) used 0.5 mM nominal Al solutions adjusted to measured initial pH 4.3 and 8–9 — which the paper designates the low-pH and near-neutral-pH series — with final pH 2.7–3.0 and 5.1–6.7, respectively. The recovered solids were characterized by solution chemistry (ICP-MS, inductively coupled plasma mass spectrometry, which quantifies dissolved elemental concentrations) and multiple solid-state NMR (nuclear magnetic resonance) techniques. In MAS (magic-angle spinning) NMR, the sample is spun at a magic angle to average out anisotropic broadenings, yielding isotropic-line spectra; 27^{27}Al single-pulse MAS distinguishes octahedral Al[6]Al^{[6]} (peaks near 0–10 ppm) from tetrahedral Al[4]Al^{[4]} (43–56 ppm). Multiple-quantum MAS (MQ/MAS) resolves overlapping peaks of quadrupolar nuclei (nuclei such as 27^{27}Al and 23^{23}Na with nuclear spin greater than 1/2, whose broad lines otherwise overlap) to give isotropic chemical shifts and quadrupole coupling constants. REDOR (rotational-echo double-resonance) NMR measures dipolar dephasing between observed 27^{27}Al and nearby protons, so the REDOR fraction reports on rigid Al–H (water/hydroxyl) proximity; comparing dephasing curves of the unknowns with those of model minerals (gibbsite, kaolinite, analcime, beidellite, mordenite) constrains the hydration environment of each Al site. 29^{29}Si{1^{1}H} CP/MAS (cross-polarization) probes the silicon network (QnQ^n notation counts the number n of bridging oxygens around each Si), and 23^{23}Na MAS probes sodium environments.

The study found that, despite thermodynamic driving forces favoring gibbsite, every experiment produced metastable aluminosilicates: at 25–50 °C a kaolinite-like hydrated Al[6]Al^{[6]} phase (REDOR curves tracking kaolinite) coexists with an Al[4]Al^{[4]} phase whose REDOR curves track analcime — a zeolite whose framework Al requires rigidly bound structural water and a charge-balancing cation, here plausibly Na giving a hydrous albite-like composition \ceNaAlSi3O8⋅H2O\ce{NaAlSi3O8\cdot H2O}. At 120 °C the Al[4]Al^{[4]} phase dominates. A weak 29^{29}Si resonance at −79.3 ppm in one sample suggested a minor allophane-like constituent (allophane and imogolite are natural short-range-ordered aluminosilicates with curved, defect-rich imogolite-like local structure). The authors proposed a descriptive metastable-equilibrium relation, \ceAl2Si2O5(OH)4+2H++Na++SiO2(aq)<=>NaAlSi3O8⋅H2O+Al3++2H2O\ce{Al2Si2O5(OH)4 + 2H+ + Na+ + SiO2(aq) <=> NaAlSi3O8\cdot H2O + Al3+ + 2H2O}, linking the kaolinite-like and Na-bearing tectosilicate-like (tectosilicate = 3-D framework silicate) phases. Crucially, however, these structural models were inferred from model-mineral comparisons on sparse datasets — MQ/MAS was run on only two samples and REDOR on a subset — leaving the actual atomic structures, the state of sodium, and the formation pathway at the silica surface undetermined.

Problem Statement

Determine the atomic structures and the formation mechanism of the metastable aluminosilicates that precipitate when dissolved Al reacts with amorphous silica in the Al–Si–H2H_2O system under the conditions of Mason, Maxwell and Carroll (2011) (0.1 M NaCl; 25–50 °C flow-through experiments at 0.7–1.5 µM dissolved Al and initial pH 4.3 and 6.3; 120 °C batch experiments at 0.5 mM nominal Al with measured initial pH 4.3 — the low-pH series, final pH 2.7–3.0 — and 8–9, final pH 5.1–6.7). Specifically: (i) identify the atomic structure of the dominant tetrahedral-Al phase, discriminating between the two possibilities advanced by the source: a hydrous sodium aluminosilicate gel with a tectosilicate framework representing a metastable precursor to zeolitic phases such as analcime, or even a hydrous albite-like (albite is the NaAlSi3O8 feldspar) \ceNaAlSi3O8⋅H2O\ce{NaAlSi3O8\cdot H2O} composition, versus Al occluded within amorphous silica; (ii) establish whether Na is structurally incorporated as the charge-balancing cation of that framework or is mainly present as a surface-sorbed species; (iii) identify the octahedral-Al phase whose hydration environment resembles kaolinite, and determine its relation (if any) to allophane-type short-range-ordered aluminosilicate; and (iv) establish the mechanism by which these phases form at the silica surface, including whether aluminum-hydroxide precursors are required intermediates or form directly from Al–silica surface association. A complete answer must resolve these ambiguities definitively — not merely by analogy to model minerals — and be consistent with the published 27^{27}Al SP/MAS and MQ/MAS shifts and quadrupole couplings, 27^{27}Al{1^{1}H} REDOR dephasing behavior, 29^{29}Si{1^{1}H} CP/MAS spectra, 23^{23}Na MAS spectra, and solution chemistry of the source experiments.

References

  1. Mason, H. E.; Maxwell, R. S.; Carroll, S. A. (2011). The formation of metastable aluminosilicates in the Al–Si–H2O system: Results from solution chemistry and solid-state NMR spectroscopy. Geochimica et Cosmochimica Acta 75(20), 6080–6093. DOI 10.1016/j.gca.2011.07.031. https://doi.org/10.1016/j.gca.2011.07.031 (source paper)
  2. Houston, J. R.; Herberg, J. L.; Maxwell, R. S.; Carroll, S. A. (2008). Association of dissolved aluminum with silica: Connecting molecular structure to surface reactivity using NMR. Geochimica et Cosmochimica Acta 72(14), 3326–3337. DOI 10.1016/j.gca.2008.04.028. https://doi.org/10.1016/j.gca.2008.04.028
  3. Beardmore, J.; Lopez, X.; Mujika, J. I.; Exley, C. (2016). What is the mechanism of formation of hydroxyaluminosilicates? Scientific Reports 6, 30913. DOI 10.1038/srep30913. https://doi.org/10.1038/srep30913
  4. Levard, C.; Doelsch, E.; Basile-Doelsch, I.; Abidin, Z.; Miche, H.; Masion, A.; Rose, J.; Borschneck, D.; Bottero, J.-Y. (2012). Structure and distribution of allophanes, imogolite and proto-imogolite in volcanic soils. Geoderma 183–184, 100–108. DOI 10.1016/j.geoderma.2012.03.015. https://doi.org/10.1016/j.geoderma.2012.03.015
  5. Shamo, E.; Charpentier, T.; Moskura, M.; Chartier, A.; Miserque, F.; Rousselet, A. (2024). Synthesis methodology of pure N–A–S–H gels with wide range of Si/Al ratios at ambient temperature. Journal of the American Ceramic Society 107(11), 7537–7549. DOI 10.1111/jace.20012. https://doi.org/10.1111/jace.20012
  6. Chen, L.; Bornes, C.; Bengtsson, O.; Erlebach, A.; Slater, B.; Grajciar, L.; Heard, C. J. (2025). A machine learning approach for dynamical modelling of Al distributions in zeolites via 23Na/27Al solid-state NMR. Faraday Discussions 255, 46–71. DOI 10.1039/d4fd00100a. https://doi.org/10.1039/d4fd00100a
  7. Chan, Y.-T.; Kuan, W.-H.; Tzou, Y.-M.; Chen, T.-Y.; Liu, Y.-T.; Wang, M.-K.; Teah, H.-Y. (2016). Molecular Structures of Al/Si and Fe/Si Coprecipitates and the Implication for Selenite Removal. Scientific Reports 6, 24716. DOI 10.1038/srep24716. https://doi.org/10.1038/srep24716