Article pubs.acs.org/est
CaCO3 Precipitation, Transport and Sensing in Porous Media with In Situ Generation of Reactants George Redden,*,† Don Fox,‡ Chi Zhang,§ Yoshiko Fujita,‡ Luanjing Guo,⊥ and Hai Huang‡ †
Chemical and Biological Engineering, Montana State University, Bozeman, Montana 59717, United States Idaho National Laboratory, Idaho Falls, Idaho 83415, United States § Department of Earth and Environmental Sciences, Rutgers University, Newark, New Jersey 07102, United States ⊥ Department of Chemical Engineering, University of Utah, Salt Lake City, Utah 84112, United States ‡
S Supporting Information *
ABSTRACT: Ureolytically driven calcite precipitation is a promising approach for inducing subsurface mineral precipitation, but engineered application requires the ability to control and predict precipitate distribution. To study the coupling between reactant transport and precipitate distribution, columns with defined zones of immobilized urease were used to examine the distribution of calcium carbonate precipitation along the flow path, at two different initial flow rates. As expected, with slower flow precipitate was concentrated toward the upstream end of the enzyme zone and with higher flow the solid was more uniformly distributed over the enzyme zone. Under constant hydraulic head conditions the flow rate decreased as precipitates decreased porosity and permeability. The hydrolysis/precipitation zone was expected to become compressed in the upstream direction. However, apparent reductions in the urea hydrolysis rate and changes in the distribution of enzyme activity, possibly due to CaCO3 precipitate hindering urea transport to the enzyme, or enzyme mobilization, mitigated reaction zone compression. Co-injected strontium was expected to be sequestered by coprecipitation with CaCO3, but the results suggested that coprecipitation was not an effective sequestration mechanism in this system. In addition, spectral induced polarization (SIP) was used to monitor the spatial and temporal evolution of the reaction zone.
■
INTRODUCTION
We investigated the temporal and spatial distribution of calcium carbonate precipitated within porous media columns where bicarbonate ions were generated in situ by urea hydrolysis. The relevant reactions are as follows:
Engineered stimulation of subsurface mineral precipitation has been proposed for remediation of metal contaminants through coprecipitation and encapsulation,1−4 permeability modification in porous and fractured media, and strengthening of the geotechnical properties of soils.5−8 However, forming mineral precipitates where contaminant treatment is needed requires choosing appropriate strategies for introducing and mixing reactants to create favorable precipitation conditions. Solution− solution interfaces can be created where longitudinal or lateral dispersion facilitates molecular-level (diffusion-controlled) mixing. Alternatively, reactants can be generated in situ at solid−solution interfaces, which improves mixing because of the high solid/solution ratios in the subsurface. In all cases the problem is challenging because of the complex coupling between reactions, changes in subsurface media properties, processes controlling transport of reactants and products9−14 and the difficulty of monitoring subsurface reaction fronts. Advancement of numerical simulation and monitoring methods for such problems relies on data documenting the spatial and temporal evolution of physical and chemical conditions in porous media. © 2013 American Chemical Society
urease
(NH 2)2 CO(urea) + 3H 2O ⎯⎯⎯⎯⎯→ 2NH+4 + HCO−3 + OH− (1)
HCO−3 + Ca 2 + + OH− → H 2O + CaCO3(s)
(2)
Production of bicarbonate was catalyzed by urease enzyme that was adsorbed in a defined section of the column. Ca2+ and Sr2+ were provided with urea in the column influent. This model system is a simplified analog to an approach proposed for in situ remediation of 90Sr or other divalent metals by coprecipitation in calcite.15 In the field, calcium may be present at sufficient concentrations and the abundance and distribution of indigenous ureolytic activity can be manipulated.16,17 Granular solid silica gel was selected for the porous media Received: Revised: Accepted: Published: 542
July 5, 2013 November 27, 2013 December 2, 2013 December 2, 2013 dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
modifications (see Figure SI-1 in the Supporting Information). Lexan columns were packed with a 17 cm interval of silica gel with adsorbed urease between 15 cm (bottom) and 16 cm (top) sections of urease-free silica gel for a total porous medium length of 48 cm. Pairs of opposing ports, one for liquid sampling and the other for SIP measurements, were installed at nine locations along the length of the packed medium, 5 cm apart. All ports were covered with nylon screen (200 mesh) to prevent medium disturbance during sampling as well as insertion of platinum wire electrodes (5 × 0.05 cm). Coiled platinum electrodes (54 × 0.05 cm) were installed at both ends of the column, separated from Teflon diffusers by 80 mesh nylon screens. (Ag/AgCl electrodes were used previously20 but soluble Ag+ was sufficient to deactivate the urease enzyme.) Upward constant-head flow of the reactant solutions was supplied by Mariotte bottles, purged with nitrogen and fitted with Ascarite II vents to minimize PCO2 in solution, and positioned to maintain positive hydrostatic pressure. Flow rates were monitored periodically. The conservative tracer LiBr was injected using a multisyringe pump. A peristaltic pump and pulse dampener were used with the NaCl (50 mM) background solution for pre- and postexperiment column flushing. An LED light box (Displays2go.com), with a stable spectrum and no flicker, positioned behind the column provided backlighting for automated transmitted light imaging of pH indicator color changes and precipitate formation using a digital camera (Nikon DX2; 50 mm lens). Experimental Procedure. Constant head experiments were conducted under two different initial conditions: (1) “slow” flow with Darcy velocity 0.2 cm min−1 (equivalent to 5 mL min−1), pore volume 448 mL, residence time 90 min; and (2) “fast” flow with Darcy velocity 0.6 cm min−1 (equivalent to 5 mL min−1), pore volume 443 mL, residence time 30 min. Prior to each series, columns were flushed with 2 L of NaCl (50 mM, pH 6) to remove residual mobile fines. Each series began with a 300 mL injection of LiBr (22.5 mM; pH 6) tracer and collection of effluent samples. Next, urea (30 mM) solutions were introduced to characterize the urease zone in the absence of precipitation. After confirmation of enzyme performance (visual observation of pH indicator dye and solution chemistry data), injection of urea (30 mM) with Ca2+ (30 mM) and Sr2+ (0.3 mM) began. During both reactant injection stages, fluid samples (7 mL) were periodically withdrawn from the side ports, with sequential sampling starting at the downstream (upper) end to minimize upstream perturbation. Conductivity and pH were measured immediately and aliquots for measurement of urea, NH4+, Ca2+, and Sr2+ were acidified (H2SO4) and stored at 4 °C until analysis. Urea+Ca+Sr injection ceased after approximately 35 L (corresponding to 5 and 2 days, for the slow and fast flow cases, respectively), after which the columns were flushed with 1 pore volumes (PV) of NaCl (50 mM; pH adjusted to 8.6 using Na2CO3 to mitigate CaCO3 dissolution) to remove residual reactants. A 300 mL LiBr (22.5 mM; pH adjusted to 8.6 using Na2CO3) slug injection was then performed for the fast flow experiment as a postexperiment tracer test. Postexperiment tracer injection was not performed for the slow flow experiment due to inadvertent introduction of air during the NaCl flushing. The tracer breakthrough profiles were used to estimate effective porosity and dispersion coefficients using the code CXTFIT, which allows estimation of such transport parameters from laboratory or field data obtained under steady state flow conditions.25
because of its ability to adsorb urease and because silica is a plausible analog to some natural mineral surfaces; important differences include high effective surface area due to the nanoscale porosity (which facilitates urease adsorption) and the grain angularity compared to other media such as sand. During the experiments solution chemistry was monitored along the length of the column, and in one experiment spectral induced polarization (SIP) was also used for real-time monitoring of physicochemical evolution within the column. Field-scale inducement of CaCO3 precipitation is likely to involve the creation of reaction fronts with distinctive transitions in electrolyte and solid phase compositions. Noninvasive geophysical methods, such as SIP, for sensing the progress and distribution of physical/chemical changes that occur in the reaction fronts will be extremely valuable for assessing and optimizing treatment efficiencies. Recent studies have demonstrated SIP’s utility for detecting changes in solid−fluid and fluid−fluid interface properties in porous media, including spatiotemporal monitoring of calcium carbonate precipitation processes.18−20 SIP is sensitive to (1) ionic strength of fluids in interconnected pore space, speciation and pH;21,22 (2) changes in total surface area due to precipitation;18 and (3) changes in surface charge due to production of a mineral phase different from the original porous medium.20 The objective of the experiments was to investigate the temporal and spatial formation of CaCO3 precipitates in a system with in situ reactant generation and dynamic coupling between precipitation and permeability, and possible changes in strontium mobility. Experiments were conducted at two flow rates. Under constant hydraulic head, mineral formation sufficient to reduce porosity should result in reduced flow velocity. If in situ reactant (CO32‑) formation and precipitation are kinetically controlled the reaction zone should become compressed toward the upstream end. A higher initial flow rate should result in elongation of the initial reaction zone because of the shorter residence time. Although the principles governing reactions in flow-through fixed-bed catalytic reactors are well established in chemical engineering (Bird et al.23), the coupling between solute transport and reactions forming mineral precipitates in porous media receives less attention and often in the context of undesirable outcomes, such as scale formation and reservoir damage.
■
MATERIALS AND METHODS Materials. All chemicals were ACS reagent grade from Sigma-Aldrich, Fisher or J.T. Baker. Solutions were prepared using degassed (Nold DeAerator by Geokon) and 0.2 μm filtered Nanopure water (18 MΩ-cm; Barnstead). Phenol red (0.01 mM) was added to all reactant solutions to facilitate observation of pH changes; the color transitions from yellow at pH 6.5 to red or violet by pH 8.2. Silica gel (Sigma-Aldrich Davisil, grade 636, 35−60 mesh/250−500 μm, pore size 60 Å, 480 m2 g−1) was saturated, degassed and titrated to neutral pH as described previously.20 Urease was immobilized on the silica gel (2.2 mg urease mL−1 dry silica gel) by adding dry enzyme (Sigma-Aldrich Type-IX from Jack Bean; nominal activity 50,600 units per gram) to the silica gel with mixing. Previous characterization of urease immobilized in a similar fashion indicated that >90% of the enzyme activity (measured by ammonia production using the indophenol method24) was associated with the solid phase. Column Construction and General Operation. Column construction was as described previously,20 with minor 543
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
Figure 1. Changes in pore water composition, measured in extracted samples, for the urea only injection stage. Flow rates are Darcy velocities. Arrows indicate flow direction. Shading indicates the enzyme zone.
Figure 2. Change in pore water composition, measured in extracted samples, for the urea, Ca2+and Sr2+ injection stage. Flow rates are Darcy velocities. Arrow indicates direction of flow. Shading indicates the enzyme zone.
SIP measurements were made on the “fast” flow rate experiment, following the data collection and analysis protocols described previously.20,26 Measurements over the frequency range 0.1−1 kHz were acquired simultaneously between
adjacent potential electrode pairs for eight different zones using a multichannel broadband complex conductivity instrument. The current was injected through the coiled platinum electrodes installed at both ends of column. The impedance 544
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
Figure 3. (a) Transmitted light images (backlighting) showing pH increase (violet) and increased opacity due to formation of CaCO3 precipitate. (b) Reflected light image showing extent of CaCO3 precipitate (white). Flow rates are Darcy velocities. Arrows indicate direction of flow.
through the enzyme zone. In the fast flow experiment urea persisted for a longer distance into the enzyme zone. The ratio of urea hydrolyzed to ammonium produced was 1:1 rather than the expected stoichiometric ratio of 1:2. Sorption of ammonium to the silica gel, measured under conditions representative of the experiments, amounted to only 4% at pH 7 and 9% at pH 9.2. Loss of volatile ammonia prior to sample analysis, and biodegradation due to microbial contamination, were considered, however, losses via these mechanisms would not be expected to result in the relatively constant downstream concentration profiles observed. An explanation for the discrepancy is not known. Figure 2 shows the pH, urea, ammonium and soluble calcium profiles following transition to the solution containing urea with Ca2+ and Sr2+. After 1.5 PV for the 0.2 cm/min case and 1.3 PV for the 0.6 cm/min case a drop in pH was evident at the start of the enzyme zone, which is consistent with CaCO3 precipitation (eq 2). The pH decrease in the precipitation zone propagated downstream and in the fast flow case resulted in a new steady state pH of approximately 8.2 by 61 PV. In the slow flow case, the pH within and beyond the urease zone was similar after 40 PV, but at 57 PV the pH had decreased to 7.9−8.0. We noted that the measured pH prior to the urease zone had also decreased, which might reflect some degree of urea hydrolysis in the influent reservoir. The predicted pH, based on complete urea hydrolysis and precipitation of CaCO3(calcite), was 7.9. With the introduction of Ca2+ and Sr2+, urea removal continued in the urease zone in both cases. However, it appeared that urea hydrolysis activity within the enzyme zone gradually decreased over time. A number of processes could be responsible. The enzyme could have degraded, although in separate tests (not shown) adsorbed enzyme was stable for periods longer than the experiment durations. Adsorption is known to enhance enzyme stability in some cases.30 Another possibility is inhibition by the ammonium product;31 however this mechanism would not produce a gradual decline in enzyme activity unless the effect is delayed. There is also a possibility that the enzyme was partially mobile and thus the enzyme loading in the zone could have decreased. In the fast flow experiment there was evidence for additional urea removal and ammonium production beyond the enzyme zone. Another potential explanation for the reduction in urea hydrolysis is that the enzyme was becoming coated by CaCO3 thus hindering
magnitude and phase shift in different zones were recorded relative to a high precision reference resistor following a swept sine wave current injection every 20 min without flow interruption. The measured magnitude and phase can be converted into real (energy loss) and imaginary (energy storage) parts of complex conductivity with correction for the geometry of the sample holder. After flow cessation, the porous medium columns were destructively sampled (cut in 5 mm thick slices) to determine the spatial distribution of precipitates. The material was rinsed three times with 95% EtOH, dried for 24 h at 105 °C, and cooled in a desiccator. A fraction (∼0.5 g) of each sample was weighed and digested in 6 mL of 10% HNO3 for Ca2+ and Sr2+ analysis. Other fractions were reserved for imaging by scanning electron microscopy (SEM). Analytical Methods. Ca, Sr, and Br were measured by inductively coupled plasma with mass spectrometry according to standard protocols. Urea and NH4+ were measured by ion chromatography.27 SEM was performed using an FEI Q650 field emission gun microscope in low vacuum mode with an accelerating voltage of 20 kV and a large field detector for imaging; on some samples energy dispersive X-ray spectrometry (EDS) was performed for qualitative element mapping. Geochemical Speciation Calculations. Equilibrium calculations were performed using The Geochemist’s Workbench28 and the thermo.dat database29 (http://www.gwb.com/ thermo.php).
■
RESULTS AND DISCUSSION Pore Water Chemistry. The pH, urea and ammonium profiles for the columns during the urea-only injection stage are shown in Figure 1 for the two flow rate conditions, at four selected sampling intervals (data from additional PV intervals are omitted for clarity, but are consistent with the trends shown). The pH increased at the start of the adsorbed urease zone to approximately 9.1, which is the value expected for complete hydrolysis of the 30 mM urea in the influent. If injection of the urea solution had continued, we expect that the pH would have eventually reached 9.1 downstream of the urease zone. The slope of the pH rise reflected the combined processes of kinetically constrained urea hydrolysis and OH− sorption by silica gel.20 Figure 1 shows the expected decrease in urea and associated increase in ammonium. At the slower flow rate, urea was completely removed approximately halfway 545
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
Figure 4. Profiles for solid phase Sr and Ca (a and c) from silica gel samples at the conclusion of the slow and fast flow rate experiments, and profiles of soluble Sr (b and d) for selected pore volumes during the slow and fast flow rate experiments. Arrows indicate direction of flow.
This occurred even with the increasing solid-solution contact time as the flow rates decreased. The difference in precipitate distribution between the slow and fast flow experiments is apparent in Figure 3 and Figure 4a and c. In the slow flow case the solid is concentrated toward the upstream end of the enzyme zone; in the fast flow case the solid appears to be more uniformly distributed over the span of the enzyme zone. This difference between the two cases was expected. The hydrolysis and precipitation reactions can progress further toward completion over a shorter distance along the flow path when the flow rate is lower and the corresponding solution residence time is longer. This, of course, also assumes that the local urease activity is not significantly diminished. For the slow flow rate experiment, a precipitation wedge appeared that advanced further downstream than the main precipitation front. Although we cannot fully account for this feature, there is a possibility that a permeability reduction at the upstream boundary of the precipitation zone generated a fast flow path. Other chemical and geophysical differences between the two column experiments were consistent with expectations based on pseudo 1-D flow conditions, but the potential effect of evolving fast flow paths may prove to be important in future work. SEM images of silica gel excavated from the enzyme zones at the conclusion of the experiments (Figure SI-2 in the Supporting Information) revealed several interesting results. Precipitates seemed to form preferentially at grain−grain contact points, which was consistent with strong cementation observed after the experiments. In addition, the absence of a uniform CaCO3 coating (as illustrated in Figure SI-2) suggests that the effective enzyme activity would not have been reduced
mass transfer of substrate (urea) to the enzyme and release of carbonate ion to the pore fluid. This process would counteract the compression of the reaction zone expected based solely on a reduction in flow rate. Precipitate Distribution and Morphology. Analyses for soluble Ca2+ during the experiments (Figure 2), the photographic record of the column (Figure 3), and analyses of final solid phase (acid-extractable) calcium (Figure 4a,c), all indicated that CaCO3(s) deposition initiated at the upstream edge of the urease zone and decreased downstream, including the region with urease-free silica gel. Rate limited precipitation is expected to cause supersaturated conditions to persist downstream of the enzyme zone. As noted earlier, the pH in samples extracted from the column was higher than expected for complete hydrolysis of urea and equilibrium precipitation of calcite (pH 8.2 vs 7.9). However, sharp declines in acidextractable calcium were observed within the enzyme zone for the slower flow rate, and at the downstream end of the enzyme zone for higher flow. Initial nucleation of CaCO3 is likely to be more rapid where urea hydrolysis sustains higher supersaturation states. The downstream conditions are consistent with lower nucleation and precipitation rates. Although mobilization of some urease was possible, gel samples excavated from the post urease section after preliminary tests of column performance (not reported here) showed no detectable urease activity. Pore water samples collected within and downstream of the enzyme zone in later stages of the experiments were supersaturated with respect to calcium carbonate and were metastable in the extracted samples. Samples that were initially clear (no scattering of light from a laser pointer) would quickly become cloudy upon stirring with a concurrent decrease in pH. 546
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
Figure 5. For the fast flow rate case, a) real (σ’) and b) imaginary conductivity (σ”) changes in the preurease (Zones 1 and 2), urease (Zone 3, 4, and 5), and posturease column sections (Zone 6, 7, and 8) for the two injection stages: urea alone and urea with Ca2+ and Sr2+. The demarcation between the two injection stages is shown by the vertical line. The complex conductivity (σ’ and σ”) data were collected at 1 Hz.
reported to be on the order of 0.0435,36 with a large range attributed to variations in experimental conditions. Higher DSr values appear to be possible when CaCO3 precipitation is facilitated by microbial processes.37,38 Although immobilized Sr and Ca appear to be somewhat correlated in Figure 4, the estimated DSr is approximately 7 × 10−5 and 4−7 × 10−5 for the low and high flow rate experiments respectively; significantly smaller than values reported from well-controlled experiments.36 Although some Sr may have been coprecipitated, our solid phase extractions would include adsorbed plus precipitated Sr. The results suggest that Sr was not significantly immobilized through coprecipitation in our experiments. Rather, Sr might have been adsorbed to the surface of CaCO3, temporarily coprecipitated in a metastable surface layer, or could be isolated from the mobile fluid phase as Sr adsorbed to the silica gel that was covered by precipitated CaCO3. Carroll et al.39 demonstrated that Sr2+ sorption to silica gel increases as the pH increases from 6 to 10. In a real subsurface environment with 90Sr contamination, where the urea-based remediation strategy might be applied, both Sr and Ca would be pre-existing in the system prior to urea introduction and would be present in various exchange sites. Although Sr2+ was introduced in the influent together with Ca2+ and urea in our experimental approach, the apparent outcome could be relevant, where coprecipitation may be less significant than isolating Sr2+ under layers of precipitated CaCO3. Geophysical Monitoring. The SIP measurements on the fast flow experiment provided the ability to detect, in real time, the changes in physicochemical solution properties from urea hydrolysis and the subsequent precipitation of calcium carbonate. The real part of complex conductivity (σ’, primarily controlled by charge transfer through interconnected fluid filled pore spaces; 1 Hz values shown in Figure 5a) and the imaginary part (σ”, associated with charge polarization within the electrical double layer at the grain/electrolyte interfaces; 1 Hz values shown Figure 5b) were both consistent with the chemistry changes measured in the extracted samples. During the urea hydrolysis only phase, as NH4+ and OH− were produced, we detected a significant increase in both σ’ and σ” sequentially from upstream to downstream within the urease zone (Zones 3, 4, and 5), and the changes of σ’ followed the changes of fluid conductivity σf (Figure SI-3 in the Supporting Information), indicating that fluid conductivity was the
for the remaining exposed silica surfaces. Preferential precipitation at grain−grain contacts could occur if hindered mixing at grain−grain contacts results in higher concentrations of surface generated carbonate and OH− ions. Gebrehiwet et al.32 observed that, in attempts to prepare metastable solutions for studying calcite precipitation rates, CaCO3 nucleation could be very rapid at high carbonate:calcium ratios. Once precipitate growth commences following nucleation, the local saturation state would be lowered thereby reducing the likelihood for precipitation on surfaces bounding the main pore channels. Another observation was that precipitates in the low flow experiment had a morphology consistent with that reported for vaterite (e.g., Han et al.33) while some precipitates from the high flow rate experiments appeared dominated by rhombohedral crystals typical of calcite (Refer to Figure SI-2 a−d in the Supporting Information). One speculative explanation for a difference in CaCO3 phases could be that the longer residence time in the slow flow experiments produced higher states of supersaturation. A consequently higher rate of CaCO 3 precipitation can favor the formation of vaterite over the more stable calcite.34 Strontium Mobility. Figure 4b and d show the evolution of Sr2+ concentration profiles during the experiments along with the moles of solid phase Ca and Sr extracted from the silica gel media at the conclusion of the experiments. Although initial reductions in both soluble Sr and Ca concentrations were observed at the entrance to the urease zone, and Sr also appeared in the final Ca-dominated acid-extracted phase, evolution of the Sr concentration profiles was more consistent with breakthrough of a sorbing solute than with coprecipitation with CaCO3(s). Equilibrium calculations assuming complete conversion of urea to carbonate and ammonium also did not predict precipitation of strontianite even though it is moderately less soluble than calcite. If coprecipitation was significant, loss of Sr from solution would be expected to follow the trend for removal of Ca by precipitation. By the end of the experiments the effluent Sr concentrations, unlike Ca, had rebounded to approximately 90% of the inlet concentration in the slow flow case and 80% of the inlet concentration in the fast flow case. The distribution coefficient DSr for Sr in calcite according to the distribution relationship DSr = (XSr CCa)/ (XCa CSr) (where XMe is the mole fraction of metal in the solid phase and CMe is the concentration of free metal in solution) is 547
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
removal, ammonium production, and soluble calcium depletion, appeared to broaden slightly in both cases. Moreover, as noted above, the real-time SIP data for the fast flow rate case also suggested that precipitation reactions occurred downstream of the urease zone albeit at reduced rates. Our observations serve as a reminder that in real systems the engineered deposition of CaCO3 (and manipulation of Sr speciation, if that is the objective) by urea-based schemes depends on the distributions of physical and chemical/ biochemical properties, and these will likely change over time particularly when urease activity is generated by microorganisms. Our findings may appear to be inconsistent with other reports where strontium coprecipitation in calcium carbonate was found to be significant. However, our results should not be construed as implying that in situ precipitation will not be an effective remediation tool. The important message is that the complexity of process coupling in systems with nonideal mixing can produce results that differ from expectations depending on the specific set of conditions and the approach to reactant mixing. In addition to a direct precipitation mechanism, strontium retardation at the field scale could also occur by permeability modification and consequent alteration of flow paths, depending on how the precipitate is distributed at multiple scales. The results illustrate how close coupling of multiple, rate-limiting processes and the manner of reactant introduction and mixing can affect (1) the distribution of reaction products, (2) changes in media properties, and (3) our ability to correctly interpret data from point samples. The results also highlight the value of real-time monitoring methods and the potential applicability of SIP for observing dynamic reaction fronts.
dominant control on bulk conductivity. This was also observed in previous studies.19,20 We attribute the increase of σ” in the urease zones during the ureolysis only stage to the increase of surface (interfacial) conduction in response to pore fluid chemical changes including the sorption of hydroxide as pH continued to rise20 and ion exchange associated with production of NH4+.19 Both species presumably contribute to surface complexation. The arrival of Ca2+ and Sr2+ (an increase in ion abundance) was evident in the σ’ data at the beginning of the second injection stage. With the onset of precipitation, both σ’ and σ” dropped. The drop in σ” is attributed at least in part to decreasing surface conduction due to the consumption of OH− (eq 2) and formation of a less negatively charged surface (calcium carbonate) compared to the silica gel. It was interesting to note that σ” also decreased in Zone 6, and to a lesser extent Zone 7; both zones were downstream of the urease zone. The onset of the σ” decreases in Zones 6 and 7 occurred after the onset of the decreases in Zones 3, 4 and 5. Unexpectedly, σ” appeared to increase in zones 1, 2, and 8 after approximately 600 min. This had not been observed in previous similar studies of ureolytically driven calcite precipitation19,20 and we suspect that it was an artifact introduced during the post-acquisition data processing to reduce the signal noise associated with the high conductivity of the pore solution. The time-lapse geophysical data revealed that the SIP technique can capture subtle changes in surface chemistry and associated interfacial properties during mineral precipitation in porous media. The polarization phenomena indicated by the electrical data are due to the electrochemical polarization occurring at the EDL of mineral particles which would be influenced by ion transport during adsorption/precipitation reactions. Establishing a link between electrical signatures and specific reactive transport events advances our ability to interpret geophysical measurements on systems undergoing geochemical transformations and provides necessary support for the development of field applications of these methods to remotely monitor complex processes associated with engineered treatments. Dynamic Coupling between Reactions and Transport. As noted earlier, in the slow flow case the precipitate deposition was concentrated toward the upstream end of the enzyme zone while in the fast flow case the solid was more evenly distributed over the enzyme zone. Our original hypothesis was that as precipitation reduced porosity and permeability, we would observe a compression of the reaction zone, and this compression would be more pronounced in a system with a slower flow rate. Permeability decreased in both columns and was greater in the slow flow column, as indicated by a decrease in saturated hydraulic conductivity (Ksat) calculated using Darcýs Law; Ksat (m/s) decreased from 3.8 × 10−4 to 2.73 × 10−4 (28% decrease) in the slow flow column and from 4.14 × 10−4 to 3.18 × 10−4 (23% decrease) in the fast flow column. The initial effective porosity (n) in both columns, estimated using the bromide tracer breakthrough curves with CXTFIT, was 0.35. In the fast flow column, n decreased to 0.27 by the end of the experiment. Although a final effective porosity could not be determined for the slower flow column due to the presence of air in the column, the Ksat changes suggest that effective porosity also decreased in the slow flow column. However, the solution chemistry profiles (Figure 2) did not show a compression in the reaction zone for either flow rate condition, and instead the reaction zone, as defined by urea
■
ASSOCIATED CONTENT
S Supporting Information *
Three figures are available in the Supporting Information. The figures are a schematic of the experimental apparatus, SEM images of calcium carbonate associated with silica gel, and plots of fluid sample conductivity data from the high flow experiment. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Author Contributions
G.R. was primarily responsible for the experimental plan, manuscript preparation and contributed to the experimental work. D.F. was responsible for experimental design, the majority of column experiments, data collection and sample analysis. C.Z was responsible for the geophysical data collection and analysis. L.G., H.H., and Y.F. contributed to experimental planning in the context of model development. Y.F. contributed to planning and manuscript preparation. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank James Henriksen for assistance with urease characterization, Joanna Taylor for assistance with analytical measurements, and Tammy Trowbridge for SEM imaging. This 548
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
Environmental Science & Technology
Article
Geophysical monitoring and reactive transport modeling of ureolytically-driven calcium carbonate precipitation. Geochem. Trans. 2011, 12. (20) Zhang, C.; Slater, L.; Redden, G.; Fujita, Y.; Johnson, T.; Fox, D. Spectral induced polarization signatures of hydroxide adsorption and mineral precipitation in porous media. Environ. Sci. Technol. 2012, 46 (8), 4357−4364. (21) Skold, M.; Revil, A.; Vaudelet, P., The pH dependence of spectral induced polarization of silica sands: Experiment and modeling. Geophys. Res. Lett. 2011, 38. (22) Vaudelet, P.; Revil, A.; Schmutz, M.; Franceschi, M.; Begassat, P., Induced polarization signatures of cations exhibiting differential sorption behaviors in saturated sands. Water Resour. Res. 2011, 47. (23) Byron Bird, R.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena, Chapter 18; Wiley, 2007. (24) Scheiner, D. Determination of ammonia and Kjeldahl nitrogen by indophenol method. Water Res. 1976, 10 (1), 31−36. (25) Toride, N.; Leij, F. J.; van Genuchten., M. Th. The CXTFIT Code for Estimating Transport Parameters from Laboratory or Field Tracer Experiments, Version 2.0; U.S. Salinity Laboratory, USDA, ARS: Riverside, CA, 1995. (26) Versteeg, R. J.; Richardson, A. N.; Rowe, T. Web-accessible scientific workflow system for performance monitoring. Environ. Sci. Technol. 2006, 40 (8), 2692−2698. (27) Mace, K. A.; Duce, R. A. Determination of Urea in Atmospheric Aerosols and Natural WatersA Cation Exchange Method. Int. J. Environ. Anal. Chem. 2002, 82 (6), 341−352. (28) Bethke, C.; Yeakel, S. The Geochemist’s Workbench (Release 8.0), GWB Essentials Guide; Hydrology Program, University of Illinois: Urbana, IL, 2009 (29) Wolery, T. J.; Jackson, K. J.; Bourcier, W. L.; Bruton, C. J.; Viani, B. E.; Knauss, K. G.; Delany, J. M. Current status of the EQ3/6 software package for geochemical modeling. ACS Symp. Ser. 1990, 416, 104−116. (30) Gianfreda, L.; Rao, M. A.; Violante, A. Adsorption, activity and kinetic-properties of urease on montmorillonite, aluminum hydroxide and al(oh)x-montmorillonite complexes. Soil Biol. Biochem. 1992, 24 (1), 51−58. (31) Fidaleo, M.; Lavecchia, R. Kinetic study of enzymatic urea hydrolysis in the pH range 4−9. Chem. Biochem. Eng. Qu. 2003, 17 (4), 311−318. (32) Gebrehiwet, T.; Redden, G.; Fujita, Y.; Beig, M.; Smith, R. The effect of the CO32− to Ca2+ Ion activity ratio on calcite precipitation kinetics and Sr2+ partitioning. Geochem. Trans. 2012, 13 (1), 1−11. (33) Han, Y. S.; Hadiko, G.; Fuji, M.; Takahashi, M. Crystallization and transformation of vaterite at controlled pH. J. Cryst. Growth 2006, 289 (1), 269−274. (34) Spanos, N.; Koutsoukos, P. G. Kinetics of precipitation of calcium carbonate in alkaline pH at constant supersaturation. Spontaneous and seeded growth. J. Phys. Chem. B 1998, 102 (34), 6679−6684. (35) Tesoriero, A. J.; Pankow, J. F. Solid solution partitioning of Sr2+, Ba2+, and Cd2+ to calcite. Geochim. Cosmochim. Acta 1996, 60 (6), 1053−1063. (36) Curti, E. Coprecipitation of radionuclides with calcite: Estimation of partition coefficients based on a review of laboratory investigations and geochemical data. Appl. Geochem. 1999, 14 (4), 433−445. (37) Fujita, Y.; Redden, G. D.; Ingram, J. A.; Cortez, M. M.; Ferris, F. G.; Smith, R. W. Strontium incorporation into calcite generated by bacterial ureolysis. Geochim. Cosmochim. Acta 2004, 68 (15), 3261− 3270. (38) Lauchnor, E. G.; Schultz, L. N.; Bugni, S.; Mitchell, A. C.; Cunningham, A. B.; Gerlach, R. Bacterially induced calcium carbonate precipitation and strontium coprecipitation in a porous media flow system. Environ. Sci. Technol. 2013, 47 (3), 1557−1564. (39) Carroll, S. A.; Roberts, S. K.; Criscenti, L. J.; O’Day, P. A. Surface complexation model for strontium sorption to amorphous silica and goethite. Geochem. Trans. 2008, 9.
research was supported by the Subsurface Biogeochemical Research Program, Office of Biological and Environmental Research, U.S. Department of Energy, under contract DEAC07-05ID14517.
■
REFERENCES
(1) Warren, L. A.; Haack, E. A. Biogeochemical controls on metal behaviour in freshwater environments. Earth-Sci. Rev. 2001, 54 (4), 261−320. (2) Shelobolina, E. S.; Konishi, H.; Xu, H.; Roden, E. E. U(VI) sequestration in hydroxyapatite produced by microbial glycerol 3phosphate metabolism. Appl. Environ. Microbiol. 2009, 75 (18), 5773− 5778. (3) Roden, E. E. Immobilization of strontium during iron biomineralization coupled to dissimilatory hydrous ferric oxide reduction. Geochim. Cosmochim. Acta 2002, 66, 2823−2839. (4) Fujita, Y.; Taylor, J. L.; Wendt, L. M.; Reed, D. W.; Smith, R. W. Evaluating the potential of native ureolytic microbes to remediate a 90Sr contaminated environment. Environ. Sci. Technol. 2010, 44 (19), 7652−7658. (5) DeJong, J. T.; Fritzges, M. B.; Nusslein, K. Microbially induced cementation to control sand response to undrained shear. J. Geotech. Geoenviron. Eng. 2006, 132 (11), 1381−1392. (6) National Research Council. Geological and Geotechnical Engineering in the New Millennium: Opportunities for Research and Technological Innovation; National Research Council, 2006. (7) Hafez, I. T.; Paraskeva, C. A.; Klepetsanis, P. G.; Koutsoukos, P. G. Sand consolidation with calcium phosphate−polyelectrolyte composites. J. Colloid Interface Sci. 2011, 363 (1), 145−156. (8) Fryar, A. E.; Schwartz, F. Hydraulic-conductivity reduction, reaction-front propagation, and preferential flow within a model reactive barrier. J. Contam. Hydrol. 1998, 32 (3−4), 333−351. (9) Saripalli, K. P.; Meyer, P. D.; Bacon, D. H.; Freedman, V. L. Changes in hydrologic properties of aquifer media due to chemical reactions: A review. Crit. Rev. Environ. Sci. Technol. 2001, 31 (4), 311− 349. (10) Rege, S. D.; Fogler, H. S. Competition among flow, dissolution, and precipitation in porous-media. AIChE J. 1989, 35 (7), 1177−1185. (11) Reis, J. C.; Acock, A. M. Permeability reduction models for the precipitation of inorganic solids in berea sandstone. In Situ 1994, 18 (3), 347−368. (12) Steefel, C. I.; Lasaga, A. C. A coupled model for transport of multiple chemical-species and kinetic precipitation dissolution reactions with application to reactive flow in single-phase hydrothermal systems. Am. J. Sci. 1994, 294 (5), 529−592. (13) Wiesner, M. R.; Grant, M. C.; Hutchins, S. R. Reduced permeability in groundwater remediation systems: Role of mobilized colloids and injected chemicals. Environ. Sci. Technol. 1996, 30 (11), 3184−3191. (14) Tartakovsky, A. M.; Redden, G.; Lichtner, P. C.; Scheibe, T. D.; Meakin, P. Mixing-induced precipitation: Experimental study and multiscale numerical analysis. Water Resour. Res. 2008, 44, 6. (15) Warren, L. A.; Maurice, P. A.; Parmar, N.; Ferris, F. G. Microbially mediated calcium carbonate precipitation: Implications for interpreting calcite precipitation and for solid-phase capture of inorganic contaminants. Geomicrobiol. J. 2001, 18 (1), 93−115. (16) Burbank, M. B.; Weaver, T. J.; Green, T. L.; Williams, B. C.; Crawford, R. L. Precipitation of calcite by indigenous microorganisms to strengthen liquefiable soils. Geomicrobiol. J. 2011, 28 (4), 301−312. (17) Fujita, Y.; Taylor, J. L.; Gresham, T. L. T.; Delwiche, M. E.; Colwell, F. S.; McLing, T. L.; Petzke, L. M.; Smith, R. W. Stimulation of microbial urea hydrolysis in groundwater to enhance calcite precipitation. Environ. Sci. Technol. 2008, 42 (8), 3025−3032. (18) Wu, Y. X.; Hubbard, S.; Williams, K. H.; Ajo-Franklin, J., On the complex conductivity signatures of calcite precipitation. J. Geophys. Res.: Biogeosci. 2010, 115. (19) Wu, Y. X.; Ajo-Franklin, J. B.; Spycher, N.; Hubbard, S. S.; Zhang, G. X.; Williams, K. H.; Taylor, J.; Fujita, Y.; Smith, R. 549
dx.doi.org/10.1021/es4029777 | Environ. Sci. Technol. 2014, 48, 542−549
