Evaluation of the Potential for CO2
Sequestration in Deep Formations Beneath Ohio
Beverly
Saylor (Department of Geological Sciences,
Case Western Reserve University, Cleveland, Ohio, USA)
Ohio Geological Society meeting
12 November 2001
We're
iterested in taking CO2 from power plants and injecting it into deep
formations and see it, hopefully, staying there - the goal is to help prevent
emission of a key greenhouse gas.
Ohio
has a good potential for CO2 sequestration, especially in the coal
producing areas of eastern Ohio and in deep, extensive oil aquifers that are
not producing anymore. CO2 injection will also help in enhanced oil
recovery from producing wells.
Enhanced
oil recovery (EOR) by miscible CO2 flood, using the WAG approach
(water after gas) - repeated gas, then water, then gas, then water injection
method. CO2 maintains pressure in the reservoir and decreases
the viscosity of the oil, and geometric traps will store the CO2.
Existing
EOR using CO2 has as a goal of getting as much CO2 back
out of the system as possible, since the CO2 has to be paid
for. The approach discussed here is a bit different.
Coal-bed
methane recovery (CBMR) by CO2 flood is another potential
operation. The CO2 will maintain pressure, desorb CH4,
and adsorption to coal stores the CO2.
Need
pure-as-possible CO2 from stacks to do EOR (N2 in there,
for example, has less effective results). CBMR doesn’t require pure CO2.
A
third option for CO2 sequestration is injecting CO2 into
deep saline aquifers. Saylor hasn’t ruled out this approach, but is
moving away from this idea. Examples in Ohio include the Mt. Simon
Sandstone [Cambrian] and Rose Run Sandstone [Cambrian] (where it isn’t an
oil/gas reservoir). The mechanisms for CO2 sequestration in
this situation are hydrodynamic trapping (versus geometric trapping above), and
mineral trapping of CO2 as carbonate. The formation waters in
these reservoirs are moving downdip (we hope) toward Pennsylvania. The CO2
dissolves into the saline waters or mixes with the saline waters or mineralizes
out as silicate minerals dissolve and carbonate minerals crystallize.
Permanent mineral trapping of CO2 - would this work, though
and how long would this take after injection? Deep saline aquifers have
high storage volumes, but this option is way down the line.
Need
to look at the EOR/CBMR approaches first to get a feel for how CO2
sequestration works there. Looking at this problem at the basin scale and
at the microscale.
Porous
flow - true paths of particles are crooked, and often fluid gets trapped at
corners or at blind ends - studying how this relates to mineral trapping of CO2.
Looking
at the distribution of Ohio power plants and oil/gas production areas - several
plants are in eastern Ohio, but several are in west-central and southwestern
Ohio, away from o/g reservoir areas.
CO2
floods in EOR has been going on for ~20 years (~1978 start with gusto in the
Permian Basin, for example). Ohio has a good potential for EOR and CO2
storage. Also potential (??) for enhanced gas recovery.
Evaluating
the Rose Run Sandstone of Ohio for this idea - EOR is possible and so is
aquifer storage for CO2.
Significance
of mineral-brine-CO2 reactions - we’re interested in the integrity
of the seal (especially considering that CO2 is acidic) and the
permeability near the injection site (will we clog up the injection site with
mineral formation? This is not much of a problem yet with carbonate reservoirs,
but need more research on the problems of this in sandstone reservoirs.
Get precipitation of clays and calcite along pore throats), the extent of
mineral trapping, and the capacity & duration of storage.
Why
study the Rose Run? It is the most shallow of the deep aquifers in
Ohio. Also looking at the Mt. Simon as a potential aquifer. The
Kerbel Sandstone [Cambrian] might work. The Rose Run is the shallowest,
but it is not the simplest. Studying the heterogeneity of the Rose Run
will help in understanding the potential of CO2 sequestration in
Ohio.
The
Rose Run consists of carbonate platform sediments with sand coming in from the
craton from the west that interfinger with the carbonates. The Rose Run
is the equivalent of the upper sandy member of the Gatesburg Formation
[Cambrian] of Pennsylvania.
Looked
at exposures of the Rose Run (= upper sandy member of the Gatesburg) at Tyrone,
Pennsylvania, which is near State College/Penn State. See a succession of
lithofacies: planar-laminated sandstones, trough-cross-laminated sandstones, ooid
grainstones, burrowed ooid packstones and wackestones, microbial bioherms,
laminites, and dolosiltstones interbedded with sandstone.
Planar-laminated sandstone facies - commonly with sharp erosive bases with rippled
tops, rip-up clasts, tidal zone deposition, almost channel-like.
Trough-cross-laminated sandstone facies
Ooid grainstone and burrowed ooid packstone facies - the burrowed parts are well mottled. These
have faint planar laminations and some trough x-stratification. The ooid
grainstones are often atop sandstones - often have mixed transition between the
two units - the sand grains and the ooids look alike at outcrop and are easily
confused.
Microbial bioherm facies
- a tremendous variety here, ranging from stromatolites to thrombolites (up to
1 meter tall and 4 meters across), surrounded by ooid grainstone facies.
Get everything in-between stromatolites and thrombolites also. The
microbial bioherm facies is closely associated with the ooid grainstone
facies. The stromatolites here aren’t analogous to Shark Bay
stromatolites, but rather are comparable with subtidal stromatolites seen in
the Bahamas (not Shark Bay type).
Laminite facies -
mm-laminated, slightly wavy and crinkly, with desiccation cracks; supratidal
mudcracked microbial mats (see Andros Island analogues).
Ooid packstone & wackestone facies
Sandy dolosiltite facies
- common facies in cores, but more islated in the Pennsylvania outcrops.
Has burrows - crinkly burrows with sand fillings.
Subtidal
cycles - these facies do form cycles - repetitively stacked successions: light
gray sandstone (with sharp erosive base and rip-up calsts & carbonate mud
mixed in) to darker gray laminated grainstone to burrowed-mottled packstone
& repeat.
Sands
could be interpreted as channels - but Saylor doesn’t like this idea.
What’s
causing the sandstone/carbonate intercalations is the key to understanding
these cycles.
The
sandstone-carbonate subtidal cycle may be due to sea level change.
Looking
at the Tyrone outcrops, looking for Enterline’s 1991 thesis core from Ashtabula
County and a Coshocton County core, and hopefully as many other cores as
possible.
The
lower parts of the Tyrone outcrops have repeated subtidal cycles and the upper
parts of the Tyrone outcrops have tidal flat cycles (laminites with interbedded
sandstones and burrowed recrystallized wackestone - finer-grained seds. that
lack ooids). These Tyrone outcrops cycles are parasequences (shallow-up
successions in general, apparently). There's hummocky
cross-stratification present in the dolostiltite above the lower Gatesburg
sandy member.
Enterline’s
1991 Ashtabula County core has more tidal flat facies represented. The
examined Coshocton County core is further toward the land - more sand and fewer
carbonates - silt even in where you have carbonates - backing up from the
intertidal areas.
The
Bahamas are not a good analogue to this Rose Run system (the Bahamas lack
siliciclastics). Florida is better, but it still has problems being a
good analogue.
The
Rose Run has mixed lithologies at the meter-scale and at the thin-section
scale. It has silicified ooids and sand. Dissolution clears out the
carbonate portions of ooids (have quartz nuclei, often) and have quartz
precipitated in its place. Not a lot of feldspar - what little there is
has been dissolved away, accounting for some of the porosity in the Rose Run.
Did
a correlation of the five Rose Run sandstone layers to see the geometries of
these sandstone/carbonate intervals.
Mineral-brine-CO2
reactions:
1)
Carbonate aquifer:
CO2 + H2O + CaCO3
[calcite] --> Ca2+ + 2HCO3-
2)
Siliciclastic aquifer, alkali phases
2KAlSi3O8 [K-feldspar] + CO2
+ H2O -->
2K+ + HCO3- + Al2Si2O5(OH)4
[illite] + 4SiO2 [quartz]
3)
Siliciclastic aquifer, alkali Earth phases
CaAl2Si3O8
[plagioclase feldspar] + CO2 + 2H2O -->
CaCO3 [calcite] + Al2Si2O5(OH)4
[illite]
It
will take time to dissolve feldspars (too long?) & precipitate out clay.
There's
also glauconitic sandstone in the Rose Run - need to see how that reacts...
Lab
experiments are testing the speed of these reactions. Then, will do
modeling of these reactions.
Particle
image velocimetry - using see-through rock models, seeing flow path
geometries. This model also describes Leisgang banding. Using a
laser to see what’s happening at the pore scale. Laser-based imaging
needs something that reflects light - have optically transparent particles as
the framework but need reflective micron-scale particles as the flow medium for
laser light reflection. Can track individual particles - can determine
velocities, flow paths, etc.
Summary - Potential
for aquifer and reservoir sequestration is high in Ohio (it’s as good here as
its going to be anywhere). Good potential for EOR. Acid waters and
dissolution are expected. Mineral trapping as calcite is important.
Knowing speed of reactions is critical.
See
Russell Kansas Project - using CO2 in EOR as part of an ethanol
production system (from Carr).
The
Rose Run Sandstone looks like, in general, a progradational succession, but
with something else also going on, maybe.