- Email: [email protected]
Assessing Chemo-mechanical Behavior Induced by CO2-Water-rock Interactions in Clay-rich Fault Gouges
Available online at www.sciencedirect.com
ScienceDirect Procedia Earth and Planetary Science 17 (2017) 292 – 295
15th Water-Rock Interaction International Symposium, WRI-15
Assessing chemo-mechanical behavior induced by CO2-water-rock interactions in clay-rich fault gouges Elisenda Bakkera, John P. Kaszubab,1, Suzanne J.T. Hangxa a Department of Earth Sciences, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, the Netherlands Department of Geology & Geophysics and School of Energy Resources, University of Wyoming, 1000 E. University Avenue, Laramie, WY 82071, USA
b
Abstract Ensuring sealing integrity is key for long-term (>10 000 years) storage of CO2. Fault reactivation and leakage can potentially be impacted by (often slow) CO2-water-rock interactions. We used the predictive power of geochemical modelling coupled to mechanical experiments on simulated fault gouge material to study the impact of long-term CO2-water-rock interactions on fault friction and stability. Of particular interest were the conditions (mineralogical composition, temperature, slip velocity) that may result in stable (velocity-strengthening) vs. unstable (velocity-weakening), and thus potentially seismogenic, behavior. Our preliminary results show that this approach poses a promising avenue for assessing fault stability. © 2017 2017Published The Authors. Published B.V. by Elsevier B.V. by ThisElsevier is an open access article under the CC BY-NC-ND license Peer-review under responsibility of the organizing committee of WRI-15. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of WRI-15 Keywords: CO2 storage; fault slip; storage integrity; fluid-rock interactions
1. Introduction Long-term geological storage is considered as one of the most immediately practical options to mitigate global CO2 emissions. However, changes in stress regime, possibly combined with CO2-water-rock interactions, can potentially reactivate pre-existing faults1. Experiments have shown that short-term interaction of CO2 with fault gouge does not lead to immediate changes in fault stability2. However, long-term reaction-induced mineralogical changes in fault gouge composition may affect the frictional and transport properties of the fault, leading to slip or the formation of leakage pathways. Generally, fault rocks consisting of phyllosilicate minerals show low friction coefficients3, 4, i.e. they require lower shear stresses to induce sliding. Carbonates and sulfates have demonstrated the *Corresponding author. Tel.: +1-307-766-6065; fax: +1307-766-6679. E-mail address: [email protected]
1878-5220 © 2017 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of WRI-15 doi:10.1016/j.proeps.2016.12.060
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295
potential for unstable fault sliding, i.e. potential for induced seismicity upon slip, under specific temperature and chemical conditions5, 6. Formation of these minerals due to CO2-fluid-rock interactions may therefore affect fault stability and require investigation when assessing potential CO 2 storage sites. Many current and potential CO2 storage sites are or will be capped by shale-rich formations7. We used geochemical models to investigate how the mineralogy of a clay-bearing fault gouge is affected by long-term (10010000 years) reaction with CO2-charged water. The predicted compositions form the basis for simulated fault gouges, for which the fault frictional behavior is tested. As an analogue for a typical clay-rich fault gouge, we study material derived from the Opalinus Claystone (Mont Terri Underground Rock Laboratory, Switzerland, courtesy of Swisstopo). In this short paper, we present preliminary results of this chemo-mechanical study on fault friction. Nomenclature (a-b) µ Pf σn T τ
rate parameter indicating stable, velocity-strengthening (i.e. (a-b) > 0) vs. unstable, velocity-weakening (i.e. (a-b) < 0) fault friction behavior [-] sliding coefficient of friction [-] pore fluid pressure [MPa]; when a pore pressure is present the effective stress acting on a medium can be calculated as follows: σeff = σ - Pf, where σ is the applied stress applied normal stress acting on the fault plane [MPa] temperature shear stress acting along the fault plane [MPa]
2. Methods 2.1. Starting material: Opalinus Claystone and model shale X-ray diffraction (XRD) analysis (this study and in the literature8) showed that Opalinus Claystone (OPA) mainly consists of phyllosilicates, quartz and calcite. The unreacted starting material (column 1, Table 1) was crushed and sieved to a grain size of less than 35 μm to simulate a fine-grained gouge material typically found in the core of a fault2. Simulated, CO2-exposed fault gouges were prepared based on the modelled compositions by combining the individual minerals, also crushed and sieved to < 35 μm grain size prior to mixing. 2.2. Geochemical modelling Geochemical models were developed using The Geochemist’s Workbench9. Two separate models, one equilibrating porewater with fine-grained OPA gouge material and a second equilibrating CO2-saturated water with a sandstone reservoir, were constructed. In a subsequent kinetic model, the CO2-saturated reservoir water displaced the OPA porewater and reacted with OPA gouge material for 10,000 years. Fault-valve behaviour10 was emulated by periodically displacing reacted porewater in the fault with fresh, CO2-saturated reservoir water. We modelled two time intervals of fluid displacement, 10 and 1000 years, representing a permeable and an impermeable fault, respectively. Each new batch of fluid reacted with OPA fault gouge at 100°C. Table 1 presents the mineralogy of simulated OPA fault gouge before reaction with CO2-saturated water and mineralogic changes induced by reaction with CO2-saturated water. Mechanical experiments were performed for each of these three mineral assemblages. 2.3. Mechanical experiments Velocity-stepping (5.43-0.22-1.086-10.86-1.086-3-10-30-100 μm/s) shear experiments were performed under wet conditions (Pf = 25 MPa), at different temperatures, and an effective normal stress of 50 MPa, using a direct shear assembly within a triaxial apparatus (see for details11). After each velocity sequence the temperature was increased. The friction coefficient is defined as µ = τ/σneff. Rate- and state-friction theory states that a positive (a–b)-value means the material is velocity-strengthening and will slip a-seismically, while for a negative (a–b)-value the material is said to be velocity-weakening and could potentially, though not necessarily, slip seismically12.
293
294
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295 Table 1. Unreacted and predicted mineralogical compositions of Opalinus Claystone fault gouge. R.T. = residence time of CO2-saturated water in fault. Minerals [vol-%] Main minerals Quartz Calcite Muscovite Chamosite Phengite Total clay minerals Illite Illite-smectite mixture Chlorite Kaolinite Trace minerals Dolomite/ankerite Siderite Pyrite Organic carbon K-feldspar Albite TOTAL
XRD analysis (this study; n = 2)
Pearson et al. (2003)8
Modelled gouge (unreacted)
Modelled gouge (1000 yrs R.T.)
Modelled gouge (10 yrs R.T.)
20.7-41.5 15.1-22 23.8-42.2 8.5-12.8 < 13.5
6-24 5-28
32 14.2 19.5
33.9 4.4 19.9
14.9
58-76 16-40 5-20 4-20 15-33
72.7 9.3 (smectite) 6.2 17.8
4.2 (smectite)
0.2-2 1-4 0.6-2 <0.1-1.5 1-3.1 0.6-2.2
0.9 (dolomite)
15.6 (dolomite)
36.9 (dolomite)
100
100
100
44.8 (smectite) 4.0
22
3. Fault friction results The friction behavior of the three fault gouge compositions show an initial increase in μ to a peak strength, followed by yielding and strain-weakening or strain-hardening behavior (Fig. 1a). Final µ-values at each velocitysequence slightly increase with temperature and displacement. At all three temperature-stages, the modelled fault gouge assuming a 1000 year residence time is stronger than the unreacted OPA sample, while the fault gouge modelled with a 10-year residence time exhibited lower friction coefficients. Note that the total carbonate content (i.e. calcite and dolomite) increases with decreasing residence time (see Table 1). At room temperature and 100°C all three samples exhibited stable, velocity-strengthening behavior, as illustrated by near neutral to positive (a-b)values (see Fig. 1b). At 150°C, OPA and the 10-yr modelled gouge still show stable slip behavior. In contrast, the 1000-yr modelled gouge shows (a-b)-values nearing potentially unstable slip behavior, i.e. near zero, but still positive, (a-b)-values. 4. Implications and conclusions Our chemical and frictional results for the modelled, CO2-exposed clay-rich fault gouges provide insight into the impact of CO2-fluid-rock interactions on fault stability. The two scenarios tested in our experiments reflect two endmembers of a ‘fault valve’ system, i.e. a system of episodic fluid displacement along a fault zone. We studied a relatively impermeable fault (1000 years residence time for fluid in the fault) and a relatively permeable fault (10 years residence time) exposed to CO2-saturated reservoir brine. For a long residence time, the geochemical simulations predict dolomite, quartz and kaolinite precipitation at the expense of calcite, smectite and chlorite, while for a short residence time smectite and dolomite precipitate at the expense of quartz, kaolinite and muscovite. The change in mineralogical composition can explain the frictional behavior seen in our experiments, as kaolinite and dolomite are considered to be frictionally strong minerals6, 13, compared to smectite and calcite4, 5. This may explain the ~25% increase in frictional strength observed for this simulated fault gouge, with respect to unreacted OPA, i.e. the fault require a higher shear stress to induce slip, at a fixed effective normal stress.
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295
Fig. 1. A) Typical evolution of friction coefficient with shear displacement obtained in the sliding experiments. Velocity steps are indicated in μm/s, with the location of a change in velocity indicated at the top of the curves via tick marks. B) (a-b)-values for each of the three fault gouges.
In contrast, if fluid penetration and replacement occurs at a more rapid time scale of 10 years, the precipitation of frictionally weak smectite suggests that the fault may become easier to reactivate for a similar state of stress. It should be noted that the formation of carbonate minerals in the longer residence time scenario could mean that the fault may behave unstably upon reactivation, given that the amount of carbonate minerals present in the fault is sufficient and that the temperature is sufficiently elevated. Our preliminary study suggests that predicting long-term fault behavior is not necessarily straightforward, as it depends not only on the friction strength but also the velocitydependence of the minerals present in the gouge. However, coupling geochemical simulations with laboratory experiments poses a promising avenue to increase the predictive capability for assessing long-term fault behavior in a CO2 storage system. References 1. Zoback MD, Gorelick SM, Earthquake triggering and large-scale geologic storage of carbon dioxide. Proc. Nat. Acad. Sci. 2012; 109: 1016410168. 2. Samuelson J, Spiers CJ, Fault friction and slip stability not affected by CO2 storage: Evidence from short-term laboratory experiments on North Sea reservoir sandstones and caprocks. Int. J. Greenhouse Gas Control 2012; 11S: S78-S90. 3. Ikari MJ, Saffer DM, Marone C, Frictional and hydrologic properties of clay-rich fault gouge. J. Geophys. Res. 2009; 114: B05409. 4. Tembe S, Lockner DA, Wong T-F, Effect of clay content and mineralogy on frictional sliding behavior of simulated gouges: Binary and ternary mixtures of quartz, illite, and montmorillonite. J. Geophys. Res. 2010; 115: B03416. 5. Verberne BA, Spiers CJ, Niemeijer AR, De Bresser JHP, De Winter DAM, Plümper O, Frictional Properties and Microstructure of CalciteRich Fault Gouges Sheared at Sub-Seismic Sliding Velocities. Pure Appl. Geophys. 2014; 171: 2617-2640. 6. Pluymakers AMH, Niemeijer AR, Spiers CJ, Frictional properties of simulated anhydrite-dolomite fault gouge and implications for seismogenic potential. J. Struct. Geol. 2016; 84: 31-46. 7. Michael K, Golab A, Shulakova V, Ennis-King J, Allinson G, Sharma S, Aiken T, Geological storage of CO2 in saline aquifers--A review of the experience from existing storage operations. Int. J. Greenhouse Gas Control 2010; 4: 659-667. 8. Pearson FJ, Arcos D, Bath A, Boisson, Fernández AM, Gäbler H-E, Gaucher E, Gautschi A, Griffault L, Hernán P, Waber HN, Mont Terri Project – Geochemistry of Water in the Opalinus Clay Formation at the Mont Terri Rock Laboratory, in Reports of the Federal Office for Water and Geology (FOWG), Geology Series. 2003. p. 321. 9. Bethke CM, Yeakel S, The Geochemist's Workbench Release 10.0: Reaction Modeling Guide. 2014, Champaign: Aqueous Solutions, LLC. 10. Sibson RH, Conditions for fault-valve behavior, in Deformation Mechanisms, Rheology and Tectonics, RJ Knipe and EH Rutter, Editors. 1990, Geological Society Special Publication. p. 15-28. 11. Niemeijer AR, Spiers CJ, Velocity dependence of strength and healing behaviour in simulated phyllosilicate-bearing fault gouge. Tectonophysics 2006; 427: 231-253. 12. Ruina A, Slip instability and state variable friction laws. J. Geophys. Res. 1983; 88: 10359-10370. 13. Crawford BR, Faulkner DR, Rutter EH, Strength, porosity, and permeability development during hydrostatic and shear loading of synthetic quartz-clay fault gouge. J. Geophys. Res. 2008; 113.
295
ScienceDirect Procedia Earth and Planetary Science 17 (2017) 292 – 295
15th Water-Rock Interaction International Symposium, WRI-15
Assessing chemo-mechanical behavior induced by CO2-water-rock interactions in clay-rich fault gouges Elisenda Bakkera, John P. Kaszubab,1, Suzanne J.T. Hangxa a Department of Earth Sciences, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, the Netherlands Department of Geology & Geophysics and School of Energy Resources, University of Wyoming, 1000 E. University Avenue, Laramie, WY 82071, USA
b
Abstract Ensuring sealing integrity is key for long-term (>10 000 years) storage of CO2. Fault reactivation and leakage can potentially be impacted by (often slow) CO2-water-rock interactions. We used the predictive power of geochemical modelling coupled to mechanical experiments on simulated fault gouge material to study the impact of long-term CO2-water-rock interactions on fault friction and stability. Of particular interest were the conditions (mineralogical composition, temperature, slip velocity) that may result in stable (velocity-strengthening) vs. unstable (velocity-weakening), and thus potentially seismogenic, behavior. Our preliminary results show that this approach poses a promising avenue for assessing fault stability. © 2017 2017Published The Authors. Published B.V. by Elsevier B.V. by ThisElsevier is an open access article under the CC BY-NC-ND license Peer-review under responsibility of the organizing committee of WRI-15. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of WRI-15 Keywords: CO2 storage; fault slip; storage integrity; fluid-rock interactions
1. Introduction Long-term geological storage is considered as one of the most immediately practical options to mitigate global CO2 emissions. However, changes in stress regime, possibly combined with CO2-water-rock interactions, can potentially reactivate pre-existing faults1. Experiments have shown that short-term interaction of CO2 with fault gouge does not lead to immediate changes in fault stability2. However, long-term reaction-induced mineralogical changes in fault gouge composition may affect the frictional and transport properties of the fault, leading to slip or the formation of leakage pathways. Generally, fault rocks consisting of phyllosilicate minerals show low friction coefficients3, 4, i.e. they require lower shear stresses to induce sliding. Carbonates and sulfates have demonstrated the *Corresponding author. Tel.: +1-307-766-6065; fax: +1307-766-6679. E-mail address: [email protected]
1878-5220 © 2017 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of WRI-15 doi:10.1016/j.proeps.2016.12.060
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295
potential for unstable fault sliding, i.e. potential for induced seismicity upon slip, under specific temperature and chemical conditions5, 6. Formation of these minerals due to CO2-fluid-rock interactions may therefore affect fault stability and require investigation when assessing potential CO 2 storage sites. Many current and potential CO2 storage sites are or will be capped by shale-rich formations7. We used geochemical models to investigate how the mineralogy of a clay-bearing fault gouge is affected by long-term (10010000 years) reaction with CO2-charged water. The predicted compositions form the basis for simulated fault gouges, for which the fault frictional behavior is tested. As an analogue for a typical clay-rich fault gouge, we study material derived from the Opalinus Claystone (Mont Terri Underground Rock Laboratory, Switzerland, courtesy of Swisstopo). In this short paper, we present preliminary results of this chemo-mechanical study on fault friction. Nomenclature (a-b) µ Pf σn T τ
rate parameter indicating stable, velocity-strengthening (i.e. (a-b) > 0) vs. unstable, velocity-weakening (i.e. (a-b) < 0) fault friction behavior [-] sliding coefficient of friction [-] pore fluid pressure [MPa]; when a pore pressure is present the effective stress acting on a medium can be calculated as follows: σeff = σ - Pf, where σ is the applied stress applied normal stress acting on the fault plane [MPa] temperature shear stress acting along the fault plane [MPa]
2. Methods 2.1. Starting material: Opalinus Claystone and model shale X-ray diffraction (XRD) analysis (this study and in the literature8) showed that Opalinus Claystone (OPA) mainly consists of phyllosilicates, quartz and calcite. The unreacted starting material (column 1, Table 1) was crushed and sieved to a grain size of less than 35 μm to simulate a fine-grained gouge material typically found in the core of a fault2. Simulated, CO2-exposed fault gouges were prepared based on the modelled compositions by combining the individual minerals, also crushed and sieved to < 35 μm grain size prior to mixing. 2.2. Geochemical modelling Geochemical models were developed using The Geochemist’s Workbench9. Two separate models, one equilibrating porewater with fine-grained OPA gouge material and a second equilibrating CO2-saturated water with a sandstone reservoir, were constructed. In a subsequent kinetic model, the CO2-saturated reservoir water displaced the OPA porewater and reacted with OPA gouge material for 10,000 years. Fault-valve behaviour10 was emulated by periodically displacing reacted porewater in the fault with fresh, CO2-saturated reservoir water. We modelled two time intervals of fluid displacement, 10 and 1000 years, representing a permeable and an impermeable fault, respectively. Each new batch of fluid reacted with OPA fault gouge at 100°C. Table 1 presents the mineralogy of simulated OPA fault gouge before reaction with CO2-saturated water and mineralogic changes induced by reaction with CO2-saturated water. Mechanical experiments were performed for each of these three mineral assemblages. 2.3. Mechanical experiments Velocity-stepping (5.43-0.22-1.086-10.86-1.086-3-10-30-100 μm/s) shear experiments were performed under wet conditions (Pf = 25 MPa), at different temperatures, and an effective normal stress of 50 MPa, using a direct shear assembly within a triaxial apparatus (see for details11). After each velocity sequence the temperature was increased. The friction coefficient is defined as µ = τ/σneff. Rate- and state-friction theory states that a positive (a–b)-value means the material is velocity-strengthening and will slip a-seismically, while for a negative (a–b)-value the material is said to be velocity-weakening and could potentially, though not necessarily, slip seismically12.
293
294
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295 Table 1. Unreacted and predicted mineralogical compositions of Opalinus Claystone fault gouge. R.T. = residence time of CO2-saturated water in fault. Minerals [vol-%] Main minerals Quartz Calcite Muscovite Chamosite Phengite Total clay minerals Illite Illite-smectite mixture Chlorite Kaolinite Trace minerals Dolomite/ankerite Siderite Pyrite Organic carbon K-feldspar Albite TOTAL
XRD analysis (this study; n = 2)
Pearson et al. (2003)8
Modelled gouge (unreacted)
Modelled gouge (1000 yrs R.T.)
Modelled gouge (10 yrs R.T.)
20.7-41.5 15.1-22 23.8-42.2 8.5-12.8 < 13.5
6-24 5-28
32 14.2 19.5
33.9 4.4 19.9
14.9
58-76 16-40 5-20 4-20 15-33
72.7 9.3 (smectite) 6.2 17.8
4.2 (smectite)
0.2-2 1-4 0.6-2 <0.1-1.5 1-3.1 0.6-2.2
0.9 (dolomite)
15.6 (dolomite)
36.9 (dolomite)
100
100
100
44.8 (smectite) 4.0
22
3. Fault friction results The friction behavior of the three fault gouge compositions show an initial increase in μ to a peak strength, followed by yielding and strain-weakening or strain-hardening behavior (Fig. 1a). Final µ-values at each velocitysequence slightly increase with temperature and displacement. At all three temperature-stages, the modelled fault gouge assuming a 1000 year residence time is stronger than the unreacted OPA sample, while the fault gouge modelled with a 10-year residence time exhibited lower friction coefficients. Note that the total carbonate content (i.e. calcite and dolomite) increases with decreasing residence time (see Table 1). At room temperature and 100°C all three samples exhibited stable, velocity-strengthening behavior, as illustrated by near neutral to positive (a-b)values (see Fig. 1b). At 150°C, OPA and the 10-yr modelled gouge still show stable slip behavior. In contrast, the 1000-yr modelled gouge shows (a-b)-values nearing potentially unstable slip behavior, i.e. near zero, but still positive, (a-b)-values. 4. Implications and conclusions Our chemical and frictional results for the modelled, CO2-exposed clay-rich fault gouges provide insight into the impact of CO2-fluid-rock interactions on fault stability. The two scenarios tested in our experiments reflect two endmembers of a ‘fault valve’ system, i.e. a system of episodic fluid displacement along a fault zone. We studied a relatively impermeable fault (1000 years residence time for fluid in the fault) and a relatively permeable fault (10 years residence time) exposed to CO2-saturated reservoir brine. For a long residence time, the geochemical simulations predict dolomite, quartz and kaolinite precipitation at the expense of calcite, smectite and chlorite, while for a short residence time smectite and dolomite precipitate at the expense of quartz, kaolinite and muscovite. The change in mineralogical composition can explain the frictional behavior seen in our experiments, as kaolinite and dolomite are considered to be frictionally strong minerals6, 13, compared to smectite and calcite4, 5. This may explain the ~25% increase in frictional strength observed for this simulated fault gouge, with respect to unreacted OPA, i.e. the fault require a higher shear stress to induce slip, at a fixed effective normal stress.
Elisenda Bakker et al. / Procedia Earth and Planetary Science 17 (2017) 292 – 295
Fig. 1. A) Typical evolution of friction coefficient with shear displacement obtained in the sliding experiments. Velocity steps are indicated in μm/s, with the location of a change in velocity indicated at the top of the curves via tick marks. B) (a-b)-values for each of the three fault gouges.
In contrast, if fluid penetration and replacement occurs at a more rapid time scale of 10 years, the precipitation of frictionally weak smectite suggests that the fault may become easier to reactivate for a similar state of stress. It should be noted that the formation of carbonate minerals in the longer residence time scenario could mean that the fault may behave unstably upon reactivation, given that the amount of carbonate minerals present in the fault is sufficient and that the temperature is sufficiently elevated. Our preliminary study suggests that predicting long-term fault behavior is not necessarily straightforward, as it depends not only on the friction strength but also the velocitydependence of the minerals present in the gouge. However, coupling geochemical simulations with laboratory experiments poses a promising avenue to increase the predictive capability for assessing long-term fault behavior in a CO2 storage system. References 1. Zoback MD, Gorelick SM, Earthquake triggering and large-scale geologic storage of carbon dioxide. Proc. Nat. Acad. Sci. 2012; 109: 1016410168. 2. Samuelson J, Spiers CJ, Fault friction and slip stability not affected by CO2 storage: Evidence from short-term laboratory experiments on North Sea reservoir sandstones and caprocks. Int. J. Greenhouse Gas Control 2012; 11S: S78-S90. 3. Ikari MJ, Saffer DM, Marone C, Frictional and hydrologic properties of clay-rich fault gouge. J. Geophys. Res. 2009; 114: B05409. 4. Tembe S, Lockner DA, Wong T-F, Effect of clay content and mineralogy on frictional sliding behavior of simulated gouges: Binary and ternary mixtures of quartz, illite, and montmorillonite. J. Geophys. Res. 2010; 115: B03416. 5. Verberne BA, Spiers CJ, Niemeijer AR, De Bresser JHP, De Winter DAM, Plümper O, Frictional Properties and Microstructure of CalciteRich Fault Gouges Sheared at Sub-Seismic Sliding Velocities. Pure Appl. Geophys. 2014; 171: 2617-2640. 6. Pluymakers AMH, Niemeijer AR, Spiers CJ, Frictional properties of simulated anhydrite-dolomite fault gouge and implications for seismogenic potential. J. Struct. Geol. 2016; 84: 31-46. 7. Michael K, Golab A, Shulakova V, Ennis-King J, Allinson G, Sharma S, Aiken T, Geological storage of CO2 in saline aquifers--A review of the experience from existing storage operations. Int. J. Greenhouse Gas Control 2010; 4: 659-667. 8. Pearson FJ, Arcos D, Bath A, Boisson, Fernández AM, Gäbler H-E, Gaucher E, Gautschi A, Griffault L, Hernán P, Waber HN, Mont Terri Project – Geochemistry of Water in the Opalinus Clay Formation at the Mont Terri Rock Laboratory, in Reports of the Federal Office for Water and Geology (FOWG), Geology Series. 2003. p. 321. 9. Bethke CM, Yeakel S, The Geochemist's Workbench Release 10.0: Reaction Modeling Guide. 2014, Champaign: Aqueous Solutions, LLC. 10. Sibson RH, Conditions for fault-valve behavior, in Deformation Mechanisms, Rheology and Tectonics, RJ Knipe and EH Rutter, Editors. 1990, Geological Society Special Publication. p. 15-28. 11. Niemeijer AR, Spiers CJ, Velocity dependence of strength and healing behaviour in simulated phyllosilicate-bearing fault gouge. Tectonophysics 2006; 427: 231-253. 12. Ruina A, Slip instability and state variable friction laws. J. Geophys. Res. 1983; 88: 10359-10370. 13. Crawford BR, Faulkner DR, Rutter EH, Strength, porosity, and permeability development during hydrostatic and shear loading of synthetic quartz-clay fault gouge. J. Geophys. Res. 2008; 113.
295