Calcium carbonate precipitation during subsurface injection of RO-brine

Calcium carbonate precipitation during subsurface injection of RO-brine: The effect on the hydraulic conductivity

Hamer, F.W.

Rietveld, L. (mentor)
Heijman, B. (mentor)
van der Meer, W. (mentor)
Bakker, M. (mentor)

Civil Engineering and Geosciences

Water Management

Sanitary Engineering

2016-07-29

The increase in salinity of groundwater is a global problem (Van Weert et al. 2009). In the west of the Netherlands, seepage of saline and brackish groundwater into layers of freshwater increases the salinity in polders and around drinking water wells (Grakist et al. 2002, Olsthoorn 2008). Systems like PURO and the Freshkeeper abstract brackish groundwater and turn it into drinking water. In both systems, the salty groundwater is desalinised with a reverse osmosis membrane and the leftover brine with a high salt content is injected in a deeper layer with saline groundwater. A too high salt content will result in precipitation of crystals like calcium carbonate, which can clog the injection well. To prevent clogging, the production of freshwater should be balanced with the supersaturation of the brine. In addition, inhibitors can reduce or even completely hinder precipitation from supersaturated solutions. This research looked at the effect of the degree of supersaturation of calcium carbonate on the precipitation and clogging in a porous medium with a series of column experiments. In addition, it looked at the inhibition of calcium carbonate precipitation by humic acids with batch experiments. A Literature research and modelling with Phreeqc in combination with Python supported the experiments. Whenever the grains in the column were cemented together with calcium carbonate that has filled the pores, a clear decrease in the hydraulic conductivity could be observed. Both cementation and a decrease in hydraulic conductivity were found in the top of the column only. The presence of calcium carbonate crystals can trigger the precipitation of calcium carbonate from a metastable solution. In the growth models of Plummer et al. (1978), Schagen et al. (2008) and Wolthers et al. (2012) the rate of precipitation is linearly related with the surface area of calcium carbonate crystals. Column simulations with those models hence show that with more calcium carbonate seed crystals present in the column, the location of precipitation shifts more to the very start of the bed. Following the equation of Carman-Kozeney, precipitation focused on a short distance is much more disastrous for the hydraulic conductivity than when the same amount is distributed over a longer distance. Also, in radial flow, clogging close to the injection well will require a much higher pressure for injection than clogging further away from the well. Both the results of the column experiments and the equation of Carman-Kozeney show that a clear exponential increase in hydraulic resistance can only be observed after a period where calcium carbonate has already been precipitating. It can make the effect of precipitating calcium carbonate on the hydraulic conductivity deceiving: this relationship between conductivity and decreasing porosity together with the precipitation rate and increase in surface area accelerating each other, make that during the initial period of precipitation clogging is unnoticed. At relatively low SIcalcite’s – observed in column experiments with SIcalcite’s of 1.0 and 1.23 and calcium carbonate seeds – the presence of calcium carbonate seeds could trigger a metastable solution to become unstable. When there were no calcium carbonate crystals present, experiments with an SIcalcite of 1.0, 1.14 and 1.23 remained metastable – yet, during one contradicting set of experiments without calcite seeds and an SIcalcite of 1.14 precipitation and cementation did occur. Hence, the soil characteristics are crucial for the stability of a metastable solution. The dominant precipitationprocesses at those SIcalcite’s are growth and 2D-nucleation. These processes both rely on the presence of calcium carbonate. Nucleation becomes more dominant at higher values of SIcalcite, when the solution is no longer metastable. In the column experiments with an SIcalcite of 1.5, the dominant processes were homogeneous and heterogeneous nucleation. In addition, crystals that were found prior to the sand bed, show that homogeneous nucleation had also taken place. Consistent with the column experiments, the upper limit of metastability that was found in the batch experiments also lies somewhere between an SIcalcite of 1.0 and 1.5. Humic acids can inhibit the precipitation of calcium carbonate at high enough concentrations. Homogeneous nucleation in solutions with an SIcalcite of 1.5 was delayed, the rate was reduced and the final total precipitation was less in presence of 5.35mgC/l or more. 5.35 mgC/l Of humic acids also inhibited nucleation in solution with an SIcalcite of 2.0. However, only a small reduction of the final total precipitation could be observed. At lower SIcalcite, no nucleation had taken place (SIcalctie of 0.5) or no clear trend could be observed among different concentrations of humic acids (SIcalcite of 1.0).

http://resolver.tudelft.nl/uuid:5d735825-2032-41a0-bc6e-f6e5bfefe3d6

Student theses

master thesis

(c) 2016 Hamer, F.W.