Skip to Main content Skip to Navigation

Experimental weathering rates of aluminium-silicates

Abstract : The chemical weathering of primary rocks and minerals in natural systems has a major impact on soil development and its composition. Chemical weathering is driven to a large extent by mineral dissolution. Through mineral dissolution, elements are released into groundwater and can readily react to precipitate secondary minerals such as clays, zeolites, and carbonates. Carbonates form from divalent cations (e.g. Ca, Fe and Mg) and CO2, and kaolin clay and gibbsite formation is attributed to the weathering of aluminium-rich minerals, most notably the feldspars. The CarbFix Project in Hellisheiði SW-Iceland aims to use natural weathering processes to form carbonate minerals by the re-injection of CO2 from a geothermal power plant back into surrounding basaltic rocks. This process is driven by the dissolution of basaltic rocks, rich in divalent cations, which can combine with injected CO2 to form and precipitate carbonates. This thesis focuses on the dissolution behaviour of Stapafell crystalline basalt, which consists of three major phases (plagioclase, pyroxene, and olivine) and is rich in divalent cations. Steady-state element release rates from crystalline basalt at far-from-equilibrium conditions were measured at pH from 2 to 11 and temperatures from 5° to 75° C in mixed-flow reactors. Steady-state Si and Ca release rates exhibit a U-shaped variation with pH, where rates decrease with increasing pH at acid condition but increase with increasing pH at alkaline conditions. Silicon release rates from crystalline basalt are comparable to Si release rates from basaltic glass of the same chemical composition at low pH and temperatures ≥25°C but slower at alkaline pH and temperatures ≥50°C. In contrast, Mg and Fe release rates decrease continuously with increasing pH at all temperatures. This behaviour is interpreted to stem from the contrasting dissolution behaviours of the three major minerals comprising the basalt: plagioclase, pyroxene, and olivine. Element release rates estimated from the sum of the volume fraction normalized dissolution rates of plagioclase, pyroxene, and olivine are within one order of magnitude of those measured in this study. In addition, these experimental results show that during injection of CO2-charged waters with pH close to 3.6, crystalline basalt preferentially releases Mg and Fe relative to Ca to the fluid phase. The injection of acidic CO2-charged fluids into crystalline basaltic rocks may therefore favour the formation of Mg and Fe carbonates rather than calcite at acidic to neutral conditions. Plagioclase is the most abundant phase in crystalline basalts and thus influences strongly its reactivity. Plagioclase dissolution rates based on Si release show a common U-shaped behaviour as a function of pH where rates decrease with increasing pH at acid condition but increase with increasing pH at alkaline conditions. Constant pH plagioclase dissolution rates increase with increasing anorthite content at acid conditions, in agreement with literature findings. Interpretation and data fitting suggests that plagioclase dissolution rates are consistent with their control by the detachment of Si-rich activated complexes formed by the removal of Al from the mineral framework. Most notably, compared with previous assumptions, plagioclase dissolution rates are independent of plagioclase composition at alkaline conditions, e.g. anorthite-rich plagioclase dissolution rates increase with increasing pH at alkaline conditions. At such conditions rapid plagioclase dissolution rates likely dominate divalent metal release from crystalline basalts to the fluids phase due to its high Ca content. Gibbsite is commonly the first mineral formed during low temperature dissolution of plagioclase. Gibbsite is an aluminium-hydroxide that is found in various soils as well as the dominant phase in many bauxite ores. Gibbsite precipitation rates were measured in closed system reactors at alkaline condition, both at 25 °C and 80 °C as a function of fluid saturation state. Analyses of the solids demonstrate that gibbsite precipitation occurred in all experiments. The comparison of gibbsite precipitation to the dissolution rates of plagioclase at pH 11 shows that the rates are close to equal. The precipitation rates of gibbsite, however, decrease faster with decreasing pH than plagioclase dissolution rates. As such it seem likely that plagioclase dissolution is faster than gibbsite precipitation at near to neutral pH, and the relatively slow rate of gibbsite precipitation influences plagioclase weathering in many Earth surface systems. Kaolinite is commonly the second secondary mineral formed during low temperature dissolution of plagioclase. Kaolinite precipitation rates were measured in mixed flow reactors as a function of fluid saturation state at pH=4 and 25 °C. In total eight long-term precipitation experiments were performed in fluids mildly supersaturated with respect to kaolinite, together with a known quantity of cleaned low defect Georgia Kaolinite as seeds. Measured kaolinite precipitation rates are relatively slow compared with plagioclase dissolution rates. This observation suggests that kaolinite formation during weathering is limited by its precipitation rates rather than by the availability of aqueous species sourced from plagioclase dissolution. Taken together the results of this study provide some of the fundamental scientific basic for predicting the rates and consequences of crystalline basalt and plagioclase dissolution at both the Earth's surface and during the near surface injection of CO2 as part of carbon storage efforts. Results indicate that although gibbsite precipitation rates are relatively rapid, the relatively slow precipitation rates of kaolinite may be the process controlling the formation of this mineral at the Earth's surface. This observation highlights the need to further quantify this secondary mineral precipitation rates at conditions typical at the Earth's surface. Moreover, as the composition of divalent metals released from crystalline basalts varies significantly with pH, CO2 carbonation in basalt should yield a systematic variation in the identity of carbonate and zeolite minerals precipitated with distance from the injection site. This latter conclusion can be tested directly as part of the currently on-going CarbFix project in Hellisheiði, Iceland.
Complete list of metadatas
Contributor : Williams Exbrayat <>
Submitted on : Monday, January 20, 2014 - 3:07:40 PM
Last modification on : Friday, January 8, 2021 - 10:12:03 AM
Long-term archiving on: : Saturday, April 8, 2017 - 8:39:13 PM


  • HAL Id : tel-00933440, version 1



Snorri Gudbrandsson. Experimental weathering rates of aluminium-silicates. Geochemistry. Université Paul Sabatier - Toulouse III, 2013. English. ⟨tel-00933440⟩



Record views


Files downloads