Boron diffusion, related isotope fractionation and the structural role of B in pegmatite forming melts

Abstract

In this study, we experimentally investigate the diffusion of B in flux-rich (1.7 % Li₂O, 2.5 % B₂O₃, 2.7 % P₂O₅ and 3.5 % F) pegmatite forming melts in order to assess the transport mechanisms of B in rare-element pegmatite deposits. Glasses were synthesized with either different B isotope compositions or with different B concentrations (up to 2.45 wt% B₂O₃). In order to explore the effect of water on transport properties, nominally dry and hydrous glasses (3.1–3.9 wt% H₂O) were investigated. The produced glasses were combined to diffusion couples to study self-diffusion and chemical diffusion of B in the same chemical system. Experiments were conducted using an internally heated pressure vessel and a rapid-heat and rapid-quench cold-seal pressure vessel at 100 MPa Ar pressure in the temperature range of 850–1250 °C for 1.6–97.3 h. In chemical diffusion experiments, the flow of B is compensated by a combined opposing flow of all other cations and F, without changing their element ratios. Thus, effective binary diffusion coefficients are obtained. Similar activation energies were determined for B self-diffusion and chemical diffusion under nominally dry conditions (196 ± 6 kJ/mol and 200 ± 17kJ/mol, respectively). Many hydrous experiments show a tilted to strongly deformed interface after the run, which was probably formed by advective flow of the low-viscosity melts during heating under pressure. Based on the successful experiments, an activation energy of 138 ± 8 kJ/mol was estimated for chemical diffusion of B in hydrous melts. We show that B diffusivity correlates with the Eyring diffusivity and, therefore, with the melt viscosity. The stable isotopes of B fractionate kinetically along the diffusion profiles due to the faster diffusivity of ¹⁰B over ¹¹B. This isotope fractionation can be quantified with an empirical isotope fractionation factor (β) of 0.032 ± 0.002. This effect is small, but significant and was previously not observed in experimental studies. While it is insensitive to the water contents used in our study, it seems to be strongly dependent on the experimental boundary conditions, such as the ratios of B between the diffusion couple halves and the B enrichment in the melt. Solid-state NMR experiments reveal that the majority of B is in trigonal coordination in our melts, which effectively weakens the melt structure. This further suggests that coordination differences are unlikely to be the driving force for B isotope fractionation between melt and fluid during late-stage fluid exsolution in pegmatitic systems.