David Richard Walker’s Metrology of the Roman Silver Coinage analyzed the silver content of over 5000 Roman denarii, antoniniani, and drachmae using x-ray fluorescence (XRF) spectrometry. His results have been widely cited and used by scholars in the fields of Roman economic theory and numismatics. This thesis seeks to prove that Walker’s XRF results were not only inaccurate, but inconsistently so. Corrosion and surface enrichment on silver-copper coins have caused surface-level elemental examinations, like XRF, to produce incorrect results. The results from Walker’s XRF analysis have been compared against results from four individual wet chemical studies. The comparisons display striking, and significant, differences. I am forced to conclude that Walker’s data does not in any way align with the true silver content of the coins he analyzed. As a result, this thesis will re-examine several theories and hypotheses posed by scholars who used Walker’s data and propose new, more appropriate, uses for Walker and XRF analysis outside of the examination of corroded silver-copper coins.

Natural nanominerals are abundant in Earth's critical zone and important in innumerable environmental processes that affect water quality. The chemical behavior of many natural nanominerals is related to their extreme small size (<10 nm) and high surface area. Atomic structural and chemical heterogeneity are also important factors affecting nanoparticle reactivity, and are a consequence of the mechanisms and complex (natural) conditions by which they form. The relationships between these factors remain poorly understood and limit our ability to predict the formation, transformation, and chemical behavior of natural nanominerals in the environment. We are using a poorly crystalline ferric hydroxysulfate nanomineral, schwertmannite, as a model system to understand the effect of formation conditions, specifically solution chemistry, on its physico-chemical characteristics. Previous studies indicate schwertmannite has highly variable bulk sulfate (Fe/S molar from 3-15) and water contents (Caraballo et al., 2013). In addition, both natural and synthetic schwertmannites have recently been described as "polyphasic" (i.e., consisting of sulfate-poor, goethite-like ordered domains embedded in a sulfate-rich, amorphous material) from observations using transmission electron microscopy (French et al., 2012). We hypothesize that solution chemistry at the time of schwertmannite formation directly affect its composition and structure. Using a factorial experiment design, we investigated the effects of increasing solution sulfate concentration ([SO4]/[Fe] at 1, 2, 3 and 5) and pH (2.4-5.6) on the crystallinity and composition of the products. Ferric hydroxide and hydroxysulfate solids were precipitated in batches by the rapid oxidation of Fe(II) by hydrogen peroxide, similar to what is seen in natural environmental systems. Sulfate and hydroxide concentrations were varied by addition of NaSO4 and NaOH, respectively. Solids were characterized using synchrotron X-ray diffraction (XRD), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), inductively coupled plasma-mass spectrometry (ICP-MS), scanning electron microscopy (SEM), and high resolution- transmission electron microscopy (HR-TEM). Our results show that schwertmannite is the only precipitate formed at low pH and that goethite rapidly becomes dominant at pH > 3.5. High-resolution TEM showed our synthetic schwertmannite samples consist of poorly crystalline goethite-like nanodomains within an amorphous solid, similarly seen in previous results. ICP-MS results reveal a narrow Fe/S molar ratio of 4.5 ±0.1 for our synthetic schwertmannite, which suggests that schwertmannite chemical composition does not depend strongly on pH or initial solution sulfate concentration. Increasing pH from 2.4 to 3.2 also has little effect on the crystallinity, bulk Fe/S ratio and water contents of schwertmannite. Increasing solution [SO4]/[Fe] also has little to no impact on crystallinity, water content or the amount of sulfate incorporated in schwertmannite. Thus, schwertmannite crystallinity and composition is not affected by initial solution sulfate and concentration under our experimental conditions. Thermal analysis allows us to independently measure OH and SO4 content in synthetic schwertmannite. In doing so, we propose a more accurate chemical formula (Fe8Oz(OH)24-2z-2x(SO4)x). The average stoichiometry based on thermal analysis of schwertmannite precipitated at [SO4]/[Fe] = 1 and pH ranging from ~2.4 2.9 is Fe8O6.51(OH)8.4(SO4)1.28. Interestingly, the calculated number of moles of oxygen is less than 8, which suggests that the standard formula Fe8O8(OH)8-2x(SO4)x is incorrect. These results for synthetic samples provide important constraints for future studies aimed at better understanding the formation, compositional variability and chemical behavior of natural schwertmannite. / Master of Science