A critical evaluation of the boron isotope-pH proxy:: The accuracy of ancient ocean pH estimates

The boron isotope-pH technique is founded on a theoretical model of carbonate delta(11)B variation with pH that assumes that the boron isotopic composition of carbonates mirrors the boron isotopic composition of borate in solution (delta(11)B(carb) = delta(11)B(borate)). Knowledge of the fractionation factor for isotope exchange between boric acid and borate in solution (alpha(4-3)), the equilibrium constant for the dissociation of boric acid (pK(B)(*)), as well as the isotopic composition of boron in seawater (delta(11)B(sw)) are required parameters of the model. The available data suggests that both the value of alpha(4-3) and the history of delta(11)B(sw) are poorly constrained. However, if one assumes that delta(11)B(carb) = delta(11)B(borate), an empirical value for alpha(4-3) can be estimated from the results of inorganic carbonate precipitation experiments. This exercise yields an alpha(4-3) value of similar to0.974 in accordance with recent theoretical estimates, but substantially deviates from the theoretical value of 0.981 often used to estimate paleo-ocean pH. Re-evaluation of ocean pH using an alpha(4-3) value of 0.974 and published foraminiferal delta(11)B values for the Cenozoic yield pH estimates that are relatively invariant, but unrealistically high (similar to8.4-8.6). Uncertainty increases as foraminiferal 'vital effects' are considered and different models for secular changes in seawater delta(11)B are applied. The inability to capture realistic ocean pH possibly reflects on our understanding of the isotopic relationship between carbonate and borate, as well as the mechanism of boron incorporation in carbonates. Given the current understanding of boron systematics, pH values estimated using this technique have considerable uncertainty, particularly when reconstructions exceed the residence time of boron in the ocean. Copyright (C) 2005 Elsevier Ltd