The enormous advances in methods to synthesize nanometer-sized Fe(III)- hydroxide clusters can answer questions about the Fe(III)-solid surfaces that have puzzled geochemists for decades. These questions include: (i) Which bridging hydroxyls exchange oxygen isotopes with bulk solution via what rate laws? (ii) Where do electronexchanges proceed? (iii) Can our computer models predict this chemistry? (iv) What level of accuracy, that is, expense, is needed? We propose a set of projects involving molecular clusters that have many Fe(III) or Al(III) atoms linked together with sets of hydroxyl bridges. The family of homologous molecules (Fe17, Fe19, Al13 and Al15) have a similar structural core but varying numbers of Al(III) or Fe(III) metal centers. Clusters can be chosen that have, or do not have, bound water molecules. The fact that Fe(III) and Al(III) clusters can be synthesized that are structurally similar is an enormous experimental advantage. The Al13 and Al15 clusters are diamagnetic and allow NMR experiments, including 17O, 27Al, 13C and 1H. Even the Fe(III) clusters allow probing via the methods of isotope-exchange and paramagnetic NMR. The projects that we propose are to: (i) Determine the reactivity of different oxygens in the clusters via isotope-exchange experiments; (ii) Evaluate the dynamics in the 'adsorbed' organic ligand that makes the clusters stable in water, including the residence times in the bound state; and particularly: (iii) Evaluate the tools needed to predict reactivities in Fe(III)-hydroxide solids and molecules. This last point can be restated: 'What level of quantum sophistication is required to obtain a structure accurate to better than 0.05 A? Better than 0.01 A'? In previous studies, we found that a change in bond length of 0.03 A changes the adjacent hydroxide oxygen-isotope exchange rates by a factor of i-103. 'How sophisticated must our methods of treating electron correlation be?' 'What constitutes an adequate basis for wave-function representation?' and: 'What are the most promising types of computationally inexpensive methods that can yield hidden reaction pathways in systems that are sufficiently large to be geochemically relevant?' How does one construct a useful reduced representation of electronic structure? The molecules allow a fairly dense progression from small to large (monomers, dimers, tetramers, Al13, Al15, Fe17, and Fe19 clusters) and this series provides a way to systematically extrapolate accurate calculations that are possible on small molecules to more approximate methods needed for extended solids surfaces and geochemistry. The implications of this research extend well beyond Earth science. This research will be used by many disciplines, including colloid chemistry, nanoscience, and medicine (metalloproteins such as ferritin resemble these clusters). Also, enrollment at UCD draws disproportionately from the high population of new immigrants to the United States, so these funds bring nontraditional but highly talented students into the Earth Sciences. We train a diverse group of students in the latest techniques of quantitative geochemistry and we blur the distinction between Earth science and inorganic chemistry, which is important for the training of fearless young geochemists.