Stack A G Raiteri P Gale J D (2012) J Am Chem Soc 134 11-14

Mechanisms and Rates of Reaction at Mineral-Water Interfaces from the Atomic to Pore Scales Blending

Simulation Theory and ExperimentStack Andrew G

bull New computational methodologies are enhancing our ability to determine the mechanisms of reactions too slow to be observed using direct simulation

bull Rigorous quantitative comparison to neutron and X-ray scattering allows us to validate our computational models

bull Understanding the atomic-level reaction mechanisms allow us to build more robust predictive theories to predict growth rates of minerals

bull Small Angle Neutron and X-ray scattering allow us to observe into which pores precipitation preferentially occurs and model rates of precipitation

bull The improved fundamental understanding resulting from these activities will substantially enhance our predictive capability for mineral growth and dissolution reactions in a variety of energy-related geochemical systems

Introduction

Summary

Acknowledgements and Collaborations

Mesoscale Growth Rates

Reaction Mechanisms on Surfacesbull Metadynamics and other ldquorare event theoriesrdquo

allow us to use atomistic computational simulation to probe reactions too slow to be observed through direct simulation

bull We discovered the mechanisms of attachmentand detachment of aqueous barium ion at thebarite (BaSO4) surface in contact with water

bull Conventional treatment of mineral reactions often include only a single reaction step shy we find multiple steps correspond to energetically and structurally distinct intermediate states

bull Activation energies for rate limiting steps thatare different for attachment and detachment precisely match macroscopic experimental estimates for growth and dissolution

How do we know the predicted reaction mechanisms are correct

bull To be reliable molecular Dynamics (MD) models must be calibrated to the structural thermodynamic and kinetic properties of the systems of interest Stack A G Gale J DRaiteri P (in press) Elements

bull Quasi-Elastic Neutron Scattering (QENS) yields the water exchange rates of the mineral surfaces Neutron Scattering with Isotopic Substitution (NDIS) allows us to obtain coordination environments for aqueous ions and X-ray Reflectivity (XR) is used for mineral-water interface structureQENS Data Water adsorbed to barite

nanoparticlesMD probability isosurfaces for water

adsorbed to the barite 001Comparison of QENS diffusional

motions and MD water exchange rates

Ba S O(sulfate) O(water) high prob O(water) low prob

The Effects of Poresbull Small Angle Neutron and X-ray scattering (SANS SAXS) allow us to monitor changes as

a function of pore size during precipitation

bull A model porous material was used controlled pore glass (CPG) which contains 75 Aring nanopores and intergranular spaces (macropores) This was compared to a natural limestone sample

bull Fit SAXS curve with Core-Shell model to model occlusion of a nanopore with CaCO3

during precipitation reaction

bull Peak growth rate is consistent with predictive model (above)SAXS Data amp Model CPG Functionalized with Anhydride-Terminated Self-

Assembled Monolayer under SI = 075 [Ca2+][CO32-] = 90

large changes in nanopores

small changes in macropores

time

matches model

Prec

ipita

tion

CessationLag

Preliminary SANS Data CaCO3

precipitation in limestone

bull Collaborators Lawrence M Anovitz Gernot Rother Eugene Mamontov Hsiu-Wen Wang Lukas Vlcek David J Wesolowski (ORNL) Jacquelyn N Bracco (ORNLWright StateUniv) Alejandro Fernandez-Martinez (LBNLUniv Grenoble) Glenn A Waychunas Carl ISteefel (LBNL) Paolo Raiteri Julian D Gale (Curtin Univ Australia)

bull Research sponsored by the Division of Chemical Sciences Geosciences and BiosciencesOffice of Basic Energy Sciences US Department of Energy and the Center for Nanoscale Control of Geologic CO2 an Energy Frontier Research Center

Changes in scattering with different solutions

bull Once the atomic-scale reaction mechanisms are known their concepts need to be incorporated into models capable of explaining macroscopic mineral reaction rates

bull Flow-through Atomic Force Microscopy (AFM) is used to measure rates of monomolecular step velocities and densities (equivalent to the intrinsic reaction rate of surface sites and the reactive site density)

bull A new theory consistent with the atomic-level reaction mechanisms discussed above isbeing actively developed to predict these growth rates

Figure 4 Poisoning of gro e by strontium19 Left) AFstrontium Center) Model for why str tium poisons growth of calcite Right)monomolecular step velocities A [Sr ][Ca2+] ratio of ≳ 1 substantially reduces gr

Note this document may contain some elements that are not fully accessible to users with disabilities If you need assistance accessing any information in this document please contact ORD_Webmasterepagov

Feasibility of In Situ Sequesteration of Toxic Metals in Flowback Water from Hydraulic Fracturing Andrew G Stack Chemical Sciences Division Oak Ridge National Laboratory

Introduction Case Study Calcite growth rates

bull Barium (Ba2+) is a toxic metal present in someproduced oil and gas field waters at levels of several thousand mgL62 far exceeding the EPA Maximum Contaminant Limit (MCL) of 2 mgL

constrained

(PA) contain up to 18000 pCiL dissolved radium (Ra2+)6 far exceeding the MCL of 5 pCiL

bull Radium is the primary Naturally Occurring Radioactive Material (NORM) in these fluids34 Some produced water from wells in the Marcellus shale

bull Subsurface bull uids produced as wastewater in Macroscopic growth rates and morphologies are driven by atomic-scale reactions athydrofracturing activities particularly from shale mineral-water interfaces but reaction mechanisms are poorly understood

produced toxic and radioactive metal contaminants bull Subsurface environments also affect these reactions through interfacial phenomena such as barium and radium and transport of reactants within a porous media these effects are similarly not well

bull These hinder predictive model development for processes in the subsurfaceTraditional models are often not able to make predictions outside of the conditions to

bull We are focusing on identifying reaction mechanisms for mineral growth and use theseto build a predictive model for the growth of sparingly-soluble ionically-bonded

bull These minerals are important in energy-related environmental and industrial settings

fl

formations are resulting in significant amounts of

1-5

which they were calibrated

minerals (CaCO3 BaSO4)

Figure 1 Barite scale in a pipe used to carry oil bull Barite (BaSO4) lsquoscalesrsquo in wellhead and borehole (httptheoildrumcom) For example mineral trapping during carbon sequestration forms carbonate minerals

equipment can also be radioactive due to radium and the engineered precipitation of minerals in the subsurface is a possible method to incorporation2sequester toxic metal contaminants such as barium strontium-90 and radium-226

bull Current practice is to treat the contaminants in produced waters on the surface but municipal and industrial water treatment plants are not currently equipped to deal withsome of the contaminants7 Instead of above-ground treatment a possible strategy may beto induce precipitation of mineral phases containing the contaminants directly in the subsurface reducing treatment of flowback water at the surface

What are the sources of the metals bull Barite is present in Marcellus shale and related formations as nodules veins replacement

crystals Its sulfur and oxygen isotope ratios donrsquot match seawater Its source could be from hydrothermal springs during early diagenesis and has been remobilized8

bull Radium is a product of radioactive decay of uranium which in turn is enriched in shaleslikely caused by the shalersquos reducing environment (uranium reduction causes precipitation of UO2) and high organic matter (which complexes uranium)

Can we induce precipitation to reduce toxic metals bull Barium and radium sulfates have very low solubilities

(Figure 2) bull Radium readily substitutes for barium in barite The

mobility of Ra2+ is entirely dominated by a disordered (RaBa)SO4 barite phase10

bull Evaporation followed by crystallization has been used as an above-ground treatment strategy but not for hydraulic fracturing flowback water11 This method has high energy costs and large capital costs5

bull These costs may be reduced substantially for in situ sulfate mineral precipitation This is due to the low

Figure 2 Logarithm of the solubility solubilities of barium and radium sulfates and that product (Ksp) of sulfate minerals as a the flowback waters are high in cations but low in function of temperature Barite and RaSO4 sulfate indicating sulfate concentration is limiting have very low solubilities9 precipitation (Table 1) Table 1 Example Flowback Water bull In situ sequestration has been proposed for otherCompositions5 types of contamination eg strontium and uranium

1213

bull The key for in situ precipitation is the rate of reactionA precipitation that occurs too rapidly after injectionwill clog porosity near well screens and reducepermeability1415 a precipitation rate that occurs tooslowly will not remove the contaminants before theflowback water is brought to the surface This willlead to scale formation and require above-groundtreatment

bull Precipitation rate depends on both saturation stateand the aqueous cation-to-anion ratio (Figure 3)1617

bull Traditional geochemical models donrsquot account forthis they only consider saturation state

Step Velocity (vs)

Step

Density

(ρ)v =

SI = log aCa2+ aCO2+

Ksp

AFM Growth Hillock on a Calcite macrSurface

Crystal Growth Theory Net Growth Rates

vh Measured Data and Model Fit Rsur f ace = Vm

= mSI + nlog([Ca2+][CO2]) + b3

kCa[Ca2+]kCO3 [CO2]

3a

VmkCa[Ca2+]+ kCO3 [CO2]

3

3

h Vm

a

kCa kCO3

Figure 3 New mineral precipitation model 18 Left) Growth rates of CaCO3 are measured on single crystals using theatomic force microscope (AFM) The velocities and densities of monomolecular steps are measured Center) These are combined into a model that predicts rates of growth per unit surface area Right) Measured data points and model prediction along with 95 Confidence Intervals and growth rate in porous media

bull Process-based precipitation models can successfully predict the dependence of growth rateon changing cation-to-anion ratio

bull Other ions in solution can poison growth of certain phases eg strontium inhibits growthof CaCO3 when the [Sr2+][Ca2+] ≳ 1 (Figure 4) Will flowback water compositions inhibit growth of (RaBa)SO4

EFfect of strontium on owth rates

wth of calcit M image of single crystal surface exposed to on

2+

bull How do the components of the fracturing fluid affect precipitation rates These includeviscosity modifiers scale inhibitors proppants etc Scale inhibitors in particular will likelynot be possible to use for an in situ precipitation technique How will this affect scaleformation of other phases (such as CaCO3)

The effects of pores on precipitation bull Precipitation in porous media may be affected by the size of the pores in which the

precipitation is occurring (Figure 5) Permeability will be most impacted by precipitation in larger pores How will precipitation be affected in shales

Figure 5 Small Angle X-ray and Neutron Scattering (SAXS SANS) of CaCO3 precipitation in pores20

a) SAXS data in controlled poreglass (CPG) an amorphous silica with well defined nanopores and intergranular spaces (macropores) b) CPG functionalized with a anhydride-terminated self assembled monolayer known to promote nucleation Precipitation is observed in nanopores c) Modeling of results from b whose maximum rate matches AFM model (Figure 4) d) SANS of precipitation in limestone which behaves similarly to b

Summary bull While preliminary data is encouraging many questions remain before the feasibility of in

situ precipitation can be established These include the effects of other dissolved species inthe flowback water the composition of the flowback water itself and the effects of pores

References 1) Veil JA Puder MG Elcock D Redweik RJ (2004) Argonne National Laboratory

Report2) Haluszczak LO Rose AW Kump LR (2012) Geochim Cosmochim Acta DOI

[101016japgeochem201210002]3) Rowan EL Engle MA Kirby CS Kraemer TF (2011) USGS Scientific Investigations

Report 2011-5135

4) Smith KP (1992) Argonne National Laboratory Report ANLEAIS-75) Gregory KB Vidic RD Dzombak DA (2011) Elements 7 181-1866) Urbina I (2011) New York Times Feb 26 20117) Ferrar K J Michanowicz D R Christen C L et al (2013) Environ Sci Technol DOI

[101021es301411q]8) Nuelle L M Shelton K L (1986) Econom Geol 81 1405-14309) FrenierWW Ziauddin M Formation Removal and Inhibition of Inorganic Scale in

the Oilfield Environment 2008

10) Martina A J Crusiusb J McNeea J J Yanfulc E K (2003) Appl Geochem 181095-1110

11) Minnich K (2011) Proceedings of the Technical Workshops for the HydraulicFracturing Study Water Resources Management US EPA

12) Wright K E Hartmann T Fujita Y (2011) Geochem Trans 12813) Gebrihiwet T A Redden G D Fujita Y et al (2012) Geochem Trans 13114) Hornberger G M Mills A L Herman J S (1992) Water Resourc Res 28 915-93815) Borgia A Pruess K Kneafsey T J et al (2012) Geothermics 44 13-22

16) Tartakovsky A M Redden G Lichtner P C et al(2008) Water Resour Res 44W06S04

17) Stack A G Grantham M C (2010) Cryst Growth Des 10 (3) 1409-141318) Bracco J N Stack A G Steefel C I (in review) Environ Sci Technol19) Bracco J N Grantham M C Stack A G (2012) Cryst Growth Des 12 3540-354820) Stack A G Fernandez-Martinez A Rother G et al (in prep)