Experimental constraints on the effect of phosphorous and boron on Nb and Ta ore formation

Evolved granites and pegmatites are important sources for critical metals, including niobium (Nb) and tantalum (Ta). The ore deposits that are commonly associated with these granitic rocks are commonly enriched in phosphorous (P) and boron (B), yet there are few experimental constraints on the influence of these fluxing compounds on ore forming processes. Here, the effects of phosphorous (P), boron (B), temperature and melt alkalinity on the solubility of manganocolumbite (MnNb₂O₆) and manganotantalite (MnTa₂O₆) in evolved, fluid-saturated granitic melts were studied. Solubility experiments were performed at 100 MPa and 800 to 1000 °C, using three haplogranitic melt compositions with alumina saturation indices (ASI) of 0.8 (peralkaline), 1.0 (subaluminous) and 1.2 (peraluminous). The possible effect of P and the influence of B in P-rich granitic systems were investigated by adding 4 wt% P₂O₅ and 4 wt% P₂O₅ + 4 wt% B₂O₃, respectively, to the starting compositions. The addition of P decreased the solubility of manganocolumbite and manganotantalite in peralkaline granitic melts. In agreement with previous studies, we suggest that P was present as network forming alkali phosphate complexes such as M_XPO_Y in the peralkaline melt structure (M = K, Na etc.), resulting in the observed effect of P at ASI = 0.8. The presence of P had a positive effect on the solubility of manganocolumbite and manganotantalite in subaluminous granitic melts, which became less significant in peraluminous systems. This may reflect the formation of complexes such as Al_XPO_Y and (Ta,Nb)_XPO_Y in the melt. Empirical relationships describing the effect of P on manganocolumbite and manganotantalite solubility in the aforementioned haplogranitc melts at 100 MPa and 800 °C were determined. The addition of B to P-rich peralkaline melts had a minor effect on manganocolumbite solubility, but resulted in an increase of manganocolumbite and manganotantalite solubility in subaluminous systems. For peraluminous melts, a minor increase of the solubility of manganocolumbite and manganotantalite with the addition of B was typically observed. We propose that the addition of B to the P-enriched melts resulted in the formation of network-forming boro-phosphate complexes until the B/P molar ratio in the melts reached unity. Once the B/P molar ratio was greater than one, polymerizing MBO₂ complexes were probably formed in peralkaline melts, while B acted presumably as a network modifier in subaluminous and peraluminous systems with B/P > 1. In combination with literature data, we determined similar dissolution enthalpies for manganocolumbite and manganotantalite for given a ASI (and pressure), where the dissolution enthalpy increased with ASI from ∼60 kJ/mol at ASI ≈ 0.6 to ∼150 kJ/mol at ASI ≈ 1.2. Cooling is thus likely to be of greater importance for magmatic mineralization in peraluminous compared to peralkaline systems. The primary influence of B and P in peraluminous systems is to act as fluxes, lowering the crystallization temperature of melts, which allows for the crystallization of primary magmatic tantalite.