Compositional Characteristics of 18th Century Derby Porcelains: Recipe Changes, Phase Transformations, and Melt Fertility

Abstract

Bulk and phase compositional characteristics of 18th century glassy (frit) and phosphatic Derby porcelains constrain aspects of production methods before and after the acquisition of the Chelsea works in 1770. Based on sherds excavated adjacent to the Derby manufactory site, glassy wares were derived from a recipe containing quartz (30–55wt.%), flint glass frit (34–56%), kaolinite (6–10%), calcite (≤16%), and subordinate bone ash (0–6%) and gypsum (0–1%). The recipe for phosphatic wares contained bone ash (31–43wt.%), kaolinite (21–24%), quartz (31–39%), and lesser amounts of calcite (1–6%). Although both recipes contained calcite and quartz, wollastonite is restricted to glassy sherds because in phosphatic samples, all calcite was consumed by a bytownite-forming reaction (calcite+kaolinite=plagioclase+2H2O +CO2) initiated at a relatively low temperature (≥200°C versus ≥350°C) dictated by the composition of the pore fluid. Due to resorption by a melt phase at peak firing conditions, however, most if not all of the wollastonite in the glassy wares is interpreted as a liquidus phase.The potential for sagging was evaluated using phase diagrams. Phosphatic wares contain a former melt phase (∼37vol.%) formed above the thermal minimum (∼1250°C) in the system Ca3(PO4)2–Al2O3–SiO2. The matrix in the glassy sherds, including rare, sagged samples, is depleted in PbO (<10 mol.% PbO) compared to the minimum melt (∼50 mol.% PbO) in the CaO–PbO–SiO2system. Along with its abundance (>50vol.%), this suggests that melting occurred at elevated temperatures, well above the thermal minimum (∼700°C) in this system. Melt compositions, however, were largely controlled by the incomplete separation of silica polymorphs and other phases during cooling. The composition of the melt phase in both types of ware is nonetheless consistent with kiln temperatures exceeding 1470°C, as indicated by the presence of cristobalite. Both types of ware would have undergone similar amounts of partial melting at the tridymite–cristobalite inversion temperature. It is therefore unlikely that the change to phosphatic recipes after the acquisition of the Chelsea works was due to concerns over kiln losses related to sagging. Sagging was due to overfiring rather than the periodic use of non-refractory glassy recipes.