1. 1 Shale-brine-CO 2 interactions and the long-term stability of shale caprock Anastasia Ilgen 1 T. Stewart 1, J. Feldman 1, J. Major 2, P. Eichhubl 2, M. Aman 2, N. Espinoza 2, C. Choens 1, T. Dewers 1, and P. Newell 1 1 Sandia National Laboratories 2 University of Texas at Austin March 3, 2016

2. Challenge 1: Sustaining Large Storage Rates Challenge 2: Using pore space with unprecedented efficiency Challenge 3: Controlling undesired or unexpected behavior Theme 1 Fluid-Assisted Geomechanics Chemo-mechanical coupling during fracture propagation Chemo-mechanical effects on reservoir rock weakening --- Coupled chemo- mechanical processes in shale caprock Pressure and fluid-driven fracture propagation in porous media using an adaptive finite element phase field model 2 Senior PersonnelStudents and Post-Docs Anastasia Ilgen: shale characterization, alteration, and geochemical modeling Peter Eichhubl: Crystal Geyser field study, geomechanical experiments Nicolas Espinoza: micromechanics: scratch and indentation Tom Dewers: micromechanics: nano-indentation, micro- pillar compression Pania Newell: modeling micromechanical data coupled with chemical effects Jon Major: Crystal Geyser field study, geomechanical experiments Michael Aman: micromechanics: scratch and indentation Charles Choens: micromechanics: nano-indentation, micro-pillar compression

3. 3 Establish quantitative relationships between chemical reactions triggered by the addition of supercritical CO 2 and changes in nano- to micro- scale mechanical properties of shale. ACTIVITY OBJECTIVES Storage Efficiency Improve sweep efficiency Predict solubility trapping Predict mineral trapping Enhance capillary (ganglion) trapping Controlling Emergence Prevent unwanted fracturing Prevent unexpected migration of CO 2 Sustaining Injectivity Control wellbore failure Enhance permeabilty/avoid precipitation during injection Guide injection limits CHALLENGES Control pathway development

4. scCO 2 Brine Shale Caprock Sandstone

5. 5 At geologic storage PT: CO 2 is supercritical (scCO 2 ). scCO 2 stimulates geochemical responses: acidification of parent brine, and dehydration of mineral surfaces. 1-3, 8 Experimental and field studies: geochemical reactions differ significantly for different rock assemblages and brine compositions. 7-9 Low-permeability caprocks (shale) are reactive at the higher end of the geologic carbon storage temperature range. 10, 11 Dissolution and secondary mineral precipitation control the evolution of porosity and permeability 8, with potential impact on the caprock integrity, and CO 2 leakage. 10, 12 [1] DePaolo et al., 2013[4] Kobos et al., 2011[7] Bickle et al., 2013[10] Liu et al., 2012 [2] Marini, 2006[5] Steele-MacInnis et al., 2012[8] Jun et al., 2012[11] Kaszuba et al., 2003 [3] Kharaka and Cole, 2011[6] Gilfillan et al., 2009[9] Lu et al., 2012 [12] Harvey et al., 2012 CO 2 injection Volume Pore pressure Stress Temperature Chemistry Fluid displacement Deformation Mineral alteration Shale Geochemical response triggered by the injection of CO 2

6. Coupled chemo-mechanical processes in shale caprock 6 Mancos Cretaceous Woodford Devonian Eau Claire Cambrian Laboratory experiments on shale samples at conditions typical of GCS to understand time- dependent geochemical reactions. Geochemical modeling for data interpretation. Nano- and micro-mechanical characterization to understand chemical effects on mechanical properties in heterogeneous shale caprock. Field Study on shale and sandstone to understand fractures in natural CO 2 reservoir caprock: Crystal Geyser analog site. Mechanical characterization to understand chemical effects on mechanical properties in shale caprock.

7. Mineral alteration 7 Dissolution of feldspars and phyllosilicate minerals, dissolution and re-precipitation of carbonate and clay minerals [1]. [1] Liu et al., 2012 [2] Gaus et al., 2010 [3] Garcia et al., 2012 [4] Fu et al., 2009 Fe 2.5 Mg 2.5 Al 2 Si 3 O 10 (OH) 8 + 2.5CaCO 3 + 5CO 2 → → 2.5FeCO 3 + 2.5MgCa(CO 3 ) 2 + Al 2 Si 2 O 5 (OH) 4 + SiO 2 + 2H 2 O [2] siderite dolomite kaolinite chalcedony chlorite calcite 2NaAlSi 3 O 8(s) + H + + 0.5H 2 O → AlO(OH) (s) + 0.5Al 2 Si 2 O 5 (OH) 4(s) + 2Na + + 5SiO 2 albiteboehmitekaolinite [4] CaCO 3(s) + 2H + → Ca 2+ + H 2 CO 3 calcite KMg 2.87 Si 3.07 Al 1.23 O 10 (OH) 2 + H 2 CO 3 + Na + + Cl - → → (K,H 3 O)(Al,Mg) 2 (Si,Al) 4 O 10 [(OH) 2,(H 2 O)] + SiO 2(am) + AlOOH (s) + + NaCl + K + + Mg 2+ + Al 3+ +H 4 SiO 4 + HCO 3 [3] phlogopite diaspore illite

8. Shale alteration in brine-CO 2 mixtures 8 Stirred reactors pressurized with CO 2 Control reactors – N 2 or ambient atm Powdered shale (A BET = 8.3 m 2 g -1 ) + brine Sample brine and solids at time intervals Analysis by IC, ICP-MS, and XRD Geochemical modeling 2500 psi CO 2, 90 °C 100 psi CO 2, 90 °C fCO 2

9. 9 Shale alteration in brine-CO 2 mixtures Mancos Woodford Synthetic brine* pH 7.44I = 1.45 M Cl - 1589mg/L NO 3 - 4.1 mg/L SO 4 2- 47251mg/L K + 20.5 mg/L Ca 2+ 484 mg/L Na + 19000 mg/L Mg 2+ 2700 mg/L Fe 2+ 2 mg/L Estimated solubility of CO 2 in brine = 0.84 M** Synthetic brine** pH6.5I = 0.26 M Cl - 8651mg/L HCO 3 - 1100mg/L SO 4 2- 82 mg/L Ca 2+ 65 mg/L Na + 5910 mg/L Mg 2+ 20 mg/L * *Developed based on U.S. Geological Survey National Produced Waters Geochemical Database v2.1, Woodford shale * Developed based on the dataset from “Natural Contamination from the Mancos Shale” Report, US DOE, ESL−RPT−2011−01. ** Using Duan and Sun (2003) data for NaCl.

10. Aqueous chemistry: Mancos 10 pH is buffered by the dissolution of CO 2 into brine, and dissolution of carbonates. Decrease in TDS in low and high pCO 2 reactors, and control reactors. Positive Na-K correlation indicates no cation exchange. Ca and Sr are more soluble in high pCO 2 vs. control reactors. Very low (microM) Fe - Fe is more soluble in high p-CO 2 -pressurized reactor. Al – below the detection limit. CO 2 N2N2 Low pCO 2 High pCO 2 CO 2 atm

11. Geochemical modeling: Mancos 11 Low pCO 2 Model DL Geochemical models capture major trends, but not everything (yet!). Calcite dissolution/precipitation and gypsum precipitation are the predominant mineralogical changes on the time scale of the experiment. Model predicts: minor amount of brucite in N 2 - pressurized reactor, and magnesite and gibbsite– in CO 2 -pressurized reactor. Low pCO 2 Gypsum Calcite Gibbsite Magnesite

12. XRD, XRF: Mancos 12 Reactor MgAlSiSKCaFe CO 2 1.787.6653.1611.324.2312.557.61 N2N2 2.627.0654.74.562.6114.4111.61 XRF

13. Mancos shale: Conclusions 13 Alteration by CO 2 -brine mixture maybe causing net decrease in density and hardness. KAl 3 Si 3 O 10 (OH) 2(s) + 4H 2 O → 3Al(OH) 3(s) + 3SiO 2 + K + + OH - muscovite gibbsite CaCO 3(s) + 2H + → Ca 2+ + H 2 CO 3 CaMg(CO 3 ) 2(s) + 4H + → Ca 2+ + Mg 2+ + 2H 2 CO 3 calcite dolomite gypsum Ca 2+ + SO 4 2- → CaSO 4 (s) + 2H + calcite Ca 2+ + H 2 CO 3 → CaCO 3 (s) + 2H + Mg 2+ + H 2 CO 3 → MgCO 3 (s) + 2H + magnesite CO 2 injection Volume Shale Chemistry Mineral alteration MineralCalciteAlbiteDolomiteMuscovite ρ, g cm -3 2.712.622.842.82 Hardness373.5 - 42 - 2.5 MineralCalciteGypsumMagnesiteGibbsite ρ, g cm -3 2.712.332.42 Hardness3243 Laboratory study

14. Fracture Toughness: Scratch Test Fracture Toughness (MPa√m) Load (kg) Mancos Shale Woodford Shale 4mm

15. Micromechanical Properties 15 Micron scale laboratory experiments on unaltered and chemically altered shales to assess changes in mechanical properties like hardness, Young’s Modulus, and unconfined compressive strength due to chemical reactions likely to occur with scCO 2 in GCS. Results can be applied to assess long-term caprock stability. Nano-indentationMicropillar compression An AFM based tool, this utilizes a rigid punch to indent a machined smooth surface of a sample to calculate hardness and Young’s modulus. Automation allows for high density surface coverage to map changes in properties. Different indentation depths test different volumes of materials. Analogous to standard geomechanics tests, pillars are machined into sample surface using FIB, and deformed with a rigid punch in SEM to measure UCS and Young’s modulus. The small size of the pillars allows the testing of individual mineral constituents to determine mechanical properties. Dewers et al., 2010 Bennett et al., 2015 Kiener et al., 2011 Tang & Ngan, 2007 CopperGothic Shale Woodford Shale

16. Computational model 16 Discrete Element Method (DEM) Particle Fracture Model –Poly-ellipsoidal DEM Particle Fracture Model (bonding for chemistry effect) Finite Element Method (FEM) –XFEM-Presto (Sierra Mechanics) –Kayenta as constitutive model Wredenberg and Larsson, 2007 Rizo, 2013, University of Pittsburgh, PhD Thesis Zhang et al., 2015, LLNL

17. Fractures in CO 2 caprocks: Crystal Geyser analog site Active on 10 2 - 10 5 year time scales 5 cm Field study

18. Compositional changes of mudrock

19. Field fracture characterization in natural CO 2 systems 19

20. Double torsion testing: Fracture toughness, subcritical crack index n (SCI) 20 Field study

21. 21 Establish quantitative relationships between chemical reactions triggered by the addition of supercritical CO 2 and changes in nano- to micro-scale mechanical properties of shale. ACTIVITY ACCOMPLISHMENTS Storage Efficiency Improve sweep efficiency Predict solubility trapping Predict mineral trapping Enhance capillary (ganglion) trapping Controlling Emergence Prevent unwanted fracturing Prevent unexpected migration of CO 2 Sustaining Injectivity Control wellbore failure Enhance permeabilty/avoid precipitation during injection Guide injection limits CHALLENGES Control pathway development Quantified alteration of shale by brine-CO 2 mixtures in laboratory. Complimentary micromechanical characterization is underway. Quantified geomechanical changes in reservoir rock and caprock at Crystal Geyser analog site. Future work

22. Acknowledgements 22