Evaluation of capillary electrophoresis for determining the concentration of dissolved silica in geothermal brines

The determination of silica concentrations in geothermal brines is widely recognized as a difficult analytical task due to its complex chemical polymerization kinetics that occurs during sample collection and chemical analysis. Capillary electrophoresis (CE) has been evaluated as a new reliable analytical method to measure silica (as silicates) in geothermal brines. Synthetic and geothermal brine samples were used to evaluate CE methodology. A capillary electrophoresis instrument, Quanta 4000 (Waters-Millipore) coupled with a Waters 820 workstation was used to carry out the experimental work. The separation of silicates was completed in approximately 5.5 min using a conventional fused-silica capillary (75 microm i.d. x 375 microm o.d. x 60 cm total length). A hydrostatic injection (10 cm for 20 s at 25 degrees C) was employed for introducing the samples. The carrier electrolyte consisted of 10 mM sodium chromate, 3 mM tetradecyltrimethyl-ammonium hydroxide (TTAOH), 2 mM sodium carbonate, and 1 mM sodium hydroxide, adjusted to a pH 11.0 +/- 0.1. Silicates were determined using an indirect UV detection at a wavelength of 254 nm with a mercury lamp and with a negative power supply (-15 kV). A good reproducibility in the migration times (%R.S.D. approximately 1.6%) based on six non-consecutive injections of synthetic brine solutions was obtained. A linear response between silica concentration and corrected peak area was observed. Ordinary (OLR) and weighted (WLR) linear regression models were used for calculating silica concentrations in all samples using the corresponding fitted calibration curves. The analytical results of CE were finally compared with the most probable values of synthetic reference standards of silica using the Student's t-test. No significant differences were found between them at P = 0.01. Similarly, the atomic absorption spectrometry (AAS) results were also compared with the most probable concentrations of the same reference standards, finding significant differences at P = 0.01.