The effect of burning on the dissolution behaviour and silicon and oxygen isotope composition of phytolith silica

The δ 30 Si and δ 18 O values of opal-A precipitated in plants (silica phytoliths) have been shown to be useful for paleoenvironmental reconstructions. Here, the effects of burning and partial dissolution of phytoliths on their isotopic compositions and dissolution behaviour were examined. Phytoliths were heated to 700 °C and then dissolution experiments were conducted in batch reactors under a range of pH (4–8) and temperature (4–19 °C) conditions. Heating caused a −2.6 ‰ shift in phytolith δ 18 O values. NMR results suggest that heating reduces the number of surface vicinal silanols, which likely results in the formation of strained Si O Si bonds which incorporate oxygen from 18 O-depleted hydroxyl groups. During dissolution, the δ 18 O of burned phytoliths increased by up to 3.5 ‰ (average 1.8 ‰) until 15–45 % saturation was reached, and then adsorption of silica on the surface of the solid began to reduce the δ 18 O value of solid silica despite a net dissolution. The maximum increase in δ 18 O during dissolution of burned phytoliths is 1.8 ‰ smaller than previously observed for unburned silica subjected to partial dissolution under the same conditions. Heating did not cause a significant change in δ 30 Si values, and partial dissolution of burned phytoliths caused a slight increase in δ 30 Si values that was smaller in magnitude than for unburned phytoliths. Dissolution of burned phytoliths progressed more slowly than dissolution of fresh phytoliths in low pH and temperature conditions, but was faster than the dissolution of fresh phytoliths when pH > 6 and temperature = 19 °C. We propose that because fewer hydrolysis sites exist on the surface of burned phytoliths that the isolated silanols that remain after heating are difficult to deprotonate at low pH resulting in slower dissolution. However, at higher pH the breakage of strained Si O Si bonds in burned phytoliths may explain their higher dissolution rate relative to fresh phytoliths. We recommend caution in using the δ 18 O values of soil phytoliths in paleoclimate reconstructions as they can be altered during both heating and partial dissolution. For phytolith assemblages collected from archaeological hearths or grasslands prone to wildfires, the shift towards lower δ 18 O values caused by heating would result in overestimations of temperature using paleothermometer equations. Care must be taken to identify alteration by dissolution or burning, which may not always be visually evident.