Modeling CO2–H2S–Water–Rock Interactions at Williston Basin Reservoir Conditions

University ofNorth Dakota

Grand Forks

EERCEnergy & Environmental Research Center®

Putting Research into Practice

Rock Unit Selection

To evaluate potential chemical and physical reactions between CO2 and selected Williston Basin rock units, samples representing three different formations were tested in bench-scale laboratory experiments. Numerical modeling of geochemical reactions was performed and verified with laboratory results. The samples were chosen as a part of the Williston Basin characterization effort, based on both availability and on the likelihood of future exposure to injected CO2. In previous work, five different formations were subjected to initial evaluation, and powdered rocks were used in order to obtain initial results. A further, more focused investigation of three different formations was undertaken: Madison Group (limestone- and dolostone- dominated samples), Broom Creek (dolomite and calcite cemented sandstone), and Pennsylvanian–Tyler Formation (sandstone with carbonates and clays). All Williston Basin samples were obtained through the North Dakota Geological Survey’s Core Library, located on the campus of the University of North Dakota.

Experimental Setup and Conditions

On the left: The combined mineralogical analysis of initial (unexposed) sample (blue color), sample exposed to CO2 (dark green), and sample exposed to CO2 and H2S (orange). On the right: The exposed water composition analysis for metals compared to numerical modeling.

The refined QEMSCAN image of the Broom Creek Formation rock, which illustrates the spatial distribution of iron content within the sample. On the left: unreacted sample. On the right: sample after exposure to CO2.

Summary This preliminary geochemical assessment Madison Group, Broom Creek Formations, and Tyler Formations of Williston Basin indicated that there are no immediate operational concerns for both cases of injection scenarios (pure CO2 and sour gas). The analysis of obtained reaction products suggests that 1) there is no strong evidence for higher degradation of samples exposed to a mixture of CO2 and H2S if compared to the pure CO2 stream; however, 2) carbonate rocks seem to be more unstable if exposed to the acid gas if compared to pure CO2 3) if H2S is present in the stream, it seems to be more dominant in the reactions; and 4) reactivity of the sample is strongly driven by its mineralogy.

The mineralogical analysis performed with various analytical tools (x-ray fluoresence, XRD, and QEMSCAN) required verification with numerical modeling tools. Often, the error in instrument tolerance, small-scale sample heterogeneity, or measurement error can be corrected by thermodynamic modeling suggestions.

Authors: Yevhen I. Holubnyak, Steven B. Hawthorne, Blaise A. F. Mibeck, David J. Miller, Jordan M. Bremer, James A. Sorensen, Edward N. Steadman, and John A. Harju

Abstract A series of laboratory experiments, field observations, and numerical modeling of geochemical reactions have been conducted to determine the chemical kinetics of potential mineral dissolution and/or precipitation caused by the injection of carbon dioxide (CO2) and sour gas. Kinetic experiments were conducted using core samples from potential Williston Basin storage formations and pure mineral samples (e.g., calcite, dolomite, siderite, etc.) obtained from vendors. Samples were analyzed using x-ray diffraction (XRD) and QEMSCAN® techniques. Two sample sets consisting of 16 samples each, under the same experimental conditions, were “soaked” for a period of 4 weeks at 2500 psi (172 bar) and 176°F (80°C) in synthetically generated brine. Over that time period, one set was exposed to pure carbon dioxide and the other to a mixture of CO2 (88 mol%) and H2S (12 mol%). The initial XRD mineralogical analysis of selected samples indicates the presence of the following minerals: anhydrite, calcite, dolomite, forsterite, halite, illite, magnetite, and quartz. The main objectives of this work were 1) to determine possible mineral reactions of the Madison Group, Broom Creek Formations, and Tyler Formations of Williston Basin, North Dakota with CO2 and sour gas; 2) to identify potential operational concerns; 3) to compare differences in mineral reactions between pure CO2 and sour gas injection scenarios; and 4) to adjust kinetic reaction rates for geochemical modeling tools with experimental observations. This work was performed by the Energy & Environmental Research Center through the Plains CO2 Reduction Partnership, one of the U.S. Department of Energy National Energy Technology Laboratory’s Regional Carbon Sequestration Partnerships.

CO2 Partial Pressure 2250 psi/155 bar

Temperature 158°F/70°C

Gas Mixture1) CO2 – 100 mol%2) CO2 – 67.3 mol% H2S – 32.7 mol%

Mass of Sample 10–15 g

Type of Sample Core plugs

Saturation Conditions Synthetic brineNaCI, 10% by weight

Time of Exposure 4 weeks

Mississippian Mission Canyon Limestone sample collected from a depth of 8140 ft (2481 m) saturated with brine (NaCl, 10 %) and exposed to pure supercritical CO2 and a mixture of supercritical CO2 (88 mol%) and H2S (12 mol%) under a pressure of 2100 psi (145.4 bar) and temperature of 176°F (80° C). The left side of the figure represents vacuum-dried-after-exposure samples compared with the original specimen; and the right side illustrates samples saturated in fluid after experiment completion.

Results

0

10

20

30

40

50

60

70

Ankerite Quartz Calcite Dolomite Hematite Magnetite

Rela

tive

Wei

ght,

%

Blank CO2 CO2+H S

1

10

100

1000

10000

Al Ca Mg Si

mg/

L

CO2 – Lab CO2 + H2S – Lab

CO2 – Modeling CO2 + H2S – Modeling

BackgroundFe-MineralNon-Fe Mineral

10.0 µr1000.0 µr

White River GrpGolden Valley Fm

Wood Mountain Fm

Ravenscrag Fm

Hell Creek FmFox Hills Fm

Frenchman Fm

Pierre Fm

PierreFm

Judith River Fm

Eagle Fm

Whitemud Fm

Niobrara Fm Niobrara FmCarlile Fm Carlile Fm

Greenhorn FmBelle Fourche Fm

Belle Fourche FmSecond White Specks

Fish Scales FmWestgate FmViking FmJoli Fou Fm

Mowry FmNewcastle FmSkull Creek FmInyan Kara Fm

Col

orad

oG

roup

Swift Fm

Rierdon FmPiper Fm

Spearfish Fm

Mannville GroupSuccess FmMasefield FmRierdon FmUpper Watrous Fm

Oil Fields

Lower Watrous Fm

Minnekahta FmOpeche Fm

Missing

Tyler FmAmsden Fm

Otter FmKibbey Fm Charles

FmMissionCanyon

Fm

Poplar MbrRatcliffe MbrMidale MbrFrobisher MbrAlida MbrTilston MbrSouris Valley

Charles Fm

Mission Canyon

Lodgepole Fm Lodgepole

Bakken Fm Bakken Fm

Mad

ison

Gro

up

Mad

ison

Gro

up

Min

nelu

saG

roupBroom Creek Fm

Fort Union Grp

Pennsylvanian

Mississippian

359

318

299

251200

146

66.5

1.8

Eastend FmBearpaw FmJudith River Fm

Milk River FmFirst White Speckled Shale

Coal Seams

Coal SeamsSaline Formations

Oil FieldsSaline Formations

Oil FieldsSaline Formations

Quaternary

Tertiary

Cretaceous

Jurassic

Triassic

Permian

Age Units YBP(Ma) USA (ND) Canada (SK)

Rock Units (Groups, Formations) Potential Sequestration

Targets

Cen

ozoi

cM

esoz

oic

Pal

eozo

icPha

nero

zoic

Dak

ota

Gro

up

Support Rackwith Samples

Heater

Ther

moc

oupl

e

Thermocouple

A

A

0.125

Support Rack (Cut)2.5”

13”

Original SampleAfter Exposure

to CO2

After Exposureto CO2

After Exposureto CO2 + H2S

After Exposureto CO2 + H2S

WILLISTON BASIN