1

Professor Sally M. BensonEnergy Resources Engineering Department

Executive Director, Global Climate and Energy ProjectStanford University

Los Alamos, NM, June 30, 2008

Carbon Dioxide Capture and Sequestration in Saline Aquifers: Fundamental

Studies of Multi-Phase Flow of CO2 and Brine

2

Topics

• Motivation and importance of CCS• Sequestration options• Rationale for a focus on saline aquifers• Key questions for saline aquifer

storage• Multi-phase flow investigations for CO2

and brine• Conclusions

3

Gas, 19.70%

Oil, 39.50%

Coal, 40.50%

Other, 0.30%

CO2 Emissions from Fossil Fuels

Power, 10539

Cement, 932

Refineries, 798

Iron and Steel, 646 Other, 46260% of global

fossil fuel emissions come

from large stationary sources

CO2 Emissions (Mt/year)

40.5% of global emissions come from coal… this is not expected to change any

time soon.Global Emissions 27,136 Mt (2005)

4

Carbon Dioxide Capture and Geologic Sequestration is a Way to Reduce Emissions

CaptureCapture UndergroundUndergroundInjectionInjection

PipelinePipelineTransportTransportCompressionCompression

5

CCS Could Make a Large Contribution to Reducing CO2 Emissions

From IPCC, 2007:WG III

650 CO2-eq 550 CO2-eq 450 CO2-eq

0

5

10

15

20

25

30

1970

1990

2010

2030

2050

2070

2090

Emis

sion

s (G

tC-e

q)

0

5

10

15

20

25

30

1970

1990

2010

2030

2050

2070

2090

Emis

sion

s (G

tC-e

q)0

5

10

15

20

25

30

1970

1990

2010

2030

2050

2070

2090

Emis

sion

s (G

tC-e

q)

Sinks

Non-CO2

Other

Fuel switch

CCS

Biofuels

Nuclear, renewable

Efficiency

Under a range of GHG stabilization scenarios – CCS is expected to contribute ~ 20% to emissions reductions

6

Options for CO2 Sequestration

7

Why Saline Aquifers?Widespread Global Distribution

Saline aquifers in sedimentary basins are widely distributed and co-located with many CO2 sources.

IPCC, 2005

8

Why Saline Aquifers?Largest Capacity

~ 1041000Saline aquifers

2003–15Coal seams (ECBM)

900a675aOil and gas fields

Upper Estimate of Global Storage

Capacity (GtCO2)

Lower Estimate of Global Storage

Capacity (GtCO2)

Reservoir TypeFrom IPCC Special Report

a. Estimates would be 25% larger if undiscovered reserves were included.

Current U.S. Saline Aquifer Capacity Estimates (U.S. DOE)

920 to 3,380 Gt CO2

IPCC, 2005

9

Some Key Questions for Saline Aquifer Storage

1. What kind of reservoir seals are needed to contain CO2?

2. What fraction of the pore space can be used for storage?

3. How far will the CO2 move from the injection site?4. What is the long term fate of CO2?5. Where does the displaced brine go?6. What is the risk that CO2 will leak out of the

storage reservoir?

10

Multi-Phase Flow DynamicsKey to Answering Questions

GravityViscous and

capillary forces Heterogeneity Structure

Answering these questions depends on the complex interplay of viscous, capillary, buoyancy forces, heterogeneity and structure of the aquifer.

2.2 mm

Micro-tomogram of a CO2 and water-filled rock: From L. Tomutsa, LBNL

11

Hierarchy of Space and Time ScalesPore-Scale

Core-Scale

Pilot-Scale

Field-Scale

Primary Focus of Our Current Research

12

Core-Scale Experiments I

• Relative Permeability Measurements• Displacement Efficiency

13

Relative Permeability CurvesR

elat

ive

Per

mea

bilit

y

Brine Saturation

CO2 Brine

14

Early Experimental Results: Small-Scale CO2 Saturation Variations

5% CO2 10% CO2

20% CO2 50% CO2 80% CO2

90% CO2 100% CO2

CO2 Saturation:0% 100%50% 75%25%

Sub-corescale saturation variations generally overlooked in relative permeability measurements.

15

Homogeneous Simulations

10%CO2

90%CO2

100%CO2

Variable Φ, k Simulations

CO2 Saturation:0% 70%

Simulated CO2 SaturationsConstant Pc Produces Homogeneous CO2 Saturations

PorosityLab Data

16

Fitting Capillary Pressure CurveP

c (P

a)

Brine Saturation

1000

10,000

100,000

Simulation Input Curve*

*Silin et al. (submitted, 2

Hg Injection Data Curve

2000 8000Pcap (Pa)given 20% CO2

Pcap = 4500PaPcap =

4500Pa

Φiki√Pc,i ∝

17

10%CO2

90%CO2

100%CO2

Variable Φ, k Simulations

CO2 Saturation:0% 70%

Lab Data

Simulated CO2 SaturationsVariable Pc Produces Small-scale CO2 Saturation Variations

Variable Pc Simulations

18

Capillary Pressure CurveP

cap (P

A)

Brine Saturation

CO2 Saturation:

0% 70%10% CO210% CO290% CO290% CO2100% CO2100% CO2

Avg. Pc Φ=.22 k=301mD Pc Envelope

19

CO

2 S

atur

atio

n

Length of Core (cm)

3.09.0

mL/min

0.3

0.065

0.03

0.001

Buckley-Leverett

0 87621 53 4

Flow-Rate Dependent Displacement Efficiency

Consequence of Variable Capillary Pressure

Simulated Displacement Efficiency

20

Tantalizing Results

0

0.2

0.4

0.6

0.8

0.1 10 1000

Pore Volumes Injected

CO

2 Sat

urat

ion 0.22m/day 1.1m/day 5.5m/d 11m/day

16.5m/d

But the experiment was flawed.

Data from Liviu Tomuta, LBNL, 2006.

21

Core-Scale Experiments II

• Relative Permeability Measurements• Displacement Efficiency• Flow Rate Dependence of Both of These

22

Schematic of Multi-Phase Flow Apparatus

R R

R

D

A2 B1 B2 C

R

PR

A1T°= 5°C

T°res

T°res

T°room

T°roomCO

2ta

nk

RR R R

R

pressure transducer

relief valve

manual on/off valve

electric on/off valve

filter

check valve

separator

CO2 Brine

confiningpressurePres

3-way valve

backpressurePpore

core holder

CO2

brine

R R

R

D

A2 B1 B2 C

R

PR

A1T°= 5°C

T°res

T°res

T°room

T°roomCO

2ta

nk

RR R R

R

pressure transducer

relief valve

manual on/off valve

electric on/off valve

filter

check valve

separator

CO2 Brine

confiningpressurePres

3-way valve

backpressurePpore

core holder

CO2

brine

23

Multi-Phase Flow Laboratory

Replicate in situ conditions- Pressure- Temperature- Brine composition

24

● Sample = Berea Sandstone

q ΔPA

Lk

μ1=

known coefficient

A = 40.53 cm2

L = 15.24 cmμ = 0.54 cp

● Absolute permeability:- Injection of brine (10 000 ppm NaCl ≈ 10 g/L)

T°= 50°C , Ppore = 12.4 MPa

- Measure ΔP as a function of the Flow Rate q

kav= 430.3 ± 7.0 mD

Petrophysical Characterization: Absolute Permeability

6

ΔPpr

essu

re d

rop

(psi

)

q Flow rate (mL/min)

15.24 cm

5.08

cm

25

→ Using the CT scanner we take the● DRY images when the core is completely dry● BRINE SATURATED images when the core is full of brine

→ each slice is● 3 mm thick ● 169*169 pixels (pixel size = 0.3*0.3 mm)● gap of 2.08 mm between 2 consecutives images

3 mm

0.3 mm

1 voxel

Petro-physical Characterization:Porosity

147.32142.24

outlet

inlet0

5.0810.16

15.2420.32

25.430.48

35.56distance from the inlet (mm)

airbrine

drysatbrine

CTCTCTCT

−−

=Φ

26Permeability (mD)

Kozeny – Carman relation

2

3

1 )(Sk

φφ−

=

• Porosity and permeability have important spatial variations due to the pore-scale structure of the rock sample.

• Slice-averaged values are relatively constant along the core.

Petrophysical Characterization:Sub-Core Scale Property Variations

27

● Co-injection of supercritical CO2 and brineat reservoir conditions:

T°= 50°CPpore = 12.4 MPa

Experimental Conditions

CO2CO2 saturated brine

Supercritical

μ = 0.046 cPd = 0.28 g/cm3

Liquid

μ = 0.558 cPρ = 0.990 g/cm3

Viscosity ratio μ brine / μ CO2 = 12.1

Density ratio ρbrine / ρCO2 = 3.5

Bond number ~ 2.2.10-3

Capillary number ~ [2.10-6 -10-5]

● Physical properties of CO2 and brineat reservoir conditions:

28

Experimental Procedure

● At a given total Flow Rate :- the core is initially saturated with brine- CO2 and brine are injected at a given fractional flow

- wait until steady state is reached (8 to 10 pore volumes)stabilization of pressure drop and saturation

- measure ΔP, saturation

-increase the proportion of CO2 ( )

)brine(FR)CO(FR)CO(FRf

)brine(FR)CO(FR)brine(FRf

CO

brine

+=

+=

2

2

2

2

2COf

)brine(FR)CO(FR +2

● Run the same procedure at different total flow rates: 2.6, 1.2 and 0.5 mL/min

29

0.88 2.06 3.38 5.29

0.315 0.387 0.3952COS 0.394

# porevolumes

Injection of 100% CO2 @ 2mL/minINLETOUTLET

CO2

The steady state is reachedafter ~ 5 pore volumes

To be sure to always reach steadystate ΔP and are measured afterhaving injected 8-10 pore volumes of fluid.

2COS●

●

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0CO2 saturation

CO

2sa

tura

tion

nb of pore volumes

30

Saturation Maps: Total Flow Rate = 2.6 mL/min

0.25 0.33 0.50

0.60 0.72 1

2COf

0.060 0.152 0.205

0.273 0.341 0.430

2COS

2COf

2COS

INLETOUTLET

CO2

brine

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0CO2 saturation

top

31

0.25 0.33 0.50

0.60 0.79 1

2COf

0.054 0.133 0.209

0.227 0.269 0.380

2COS

2COf

2COS

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0CO2 saturation

14

Saturation Maps: Total Flow Rate = 1.2 mL/min

INLETOUTLET

CO2

brine

top

32

0.16 0.30 0.42 0.60

0.84 1

2COf

0.019 0.045 0.177

0.233 0.244 0.271

2COS

2COf

2COS

0.74

0.105

Saturation Maps: Total Flow Rate = 0.5 mL/min

INLETOUTLET

CO2

brine

top

33

Permeability (mD)

2.6

0.4302COS 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0CO2 saturation

1.2

0.380

0.5

0.271

Flow rate(mL/min)

• High permeability paths correspond to high CO2 saturation close to inlet

• Elsewhere the correlation between permeability & saturation is poor due to unconnected high permeability pathways

• The low porosity regions act as local capillary barriers

INLETOUTLET

CO2

brine

Explanation for Saturation Variations

34

The higher the proportion of CO2, the higher the CO2 saturation.

Lower values when the flow rate decreases.

●

●

2.6 mL/min 1.2 mL/min

0.5 mL/min

CO

2sa

tura

tion

Distance from INLET (cm)

CO

2sa

tura

tion

Distance from INLET (cm)

CO

2sa

tura

tion

Distance from INLET (cm)

CO2 Saturation Profiles at Different Fractional Flows

CO

2sa

tura

tion

35

Flow Rate Dependent CO2 Saturations

At any given fractional flow the CO2 saturation is a function of the flow rate

The higher the flow rate, the higher the CO2saturation

Results not consistent with classical multi-phase flow theory where saturation – and thus relative permeability – are independent of flow rate

●

●

●

CO

2sa

tura

tion

Flow rate (mL/min)

36

Flow Rate Dependent Relative Permeability Curves

• Same general trend

• Global shift between the two curves

• The lower the flow rate, the lower the relative permeability to brine and CO2

Brine saturation

Rel

ativ

e pe

rmea

bilit

y

End Point Saturations

37

Flow Rate Dependant End-Point Brine Saturation

● Strong variations of the end point brine saturation

● Unusually high saturation end-point values

● Asymptotic value for high flow rates

Flow rate (mL/min)

End

poin

t brin

e sa

tura

tion

38

Why Might This Be Important?

Rel

ativ

e pe

rmea

bilit

y

• Relative permeability curves (drainage) are typically treated as single-valued functions in numerical and analytical models– These data suggest this may not

be correct• Plume size and capacity depend

on relative permeability curves– These data suggest that plumes

will be bigger and capacity will be lower than estimates based on relative permeability measurements made at high flow rates

• Volume averaging will have a large effect on simulation results– Tendency to overestimate CO2

saturations

39

Core-Scale Multiphase Flow Experiments

Multiphase Flow Theory

PetrophysicalCharacterization

NumericalSimulation

What’s Next?

• Continue to investigate flow-rate dependence with more experiments on additional rocks

• Tests with vertical orientation

• Reliable methods for permeability mapping

• Reliable methods for capillary pressure mapping

• Theoretical foundation for observed multi-phase flows

• Parametric formulation for rate dependent relative permeability curves

• History matching experiments

• Sensitivity analysis• Simulation artifacts• Implication for large

scale processes

40

Acknowledgements

• Post Doc – Jean-Christophe Perrin• Research Associate – Ljuba Miljkovic• Graduate Students – Michael Krause, Chia-

Wei Kuo, Ethan Chabora• GCEP sponsors

– ExxonMobil, Schlumberger, Toyota, GE

• ERE Department– Tony Kovscek, Louis Castanier, Roland Horne