Supplementary
Material
for
:
Microbial Contribution Estimated by Clumped
Isotopologues (
13
CH
3
D and
12
CH
2
D
2
)
Characteristics in
a
CO
2
Enhanced
C oal
B ed
Methane
Reservoir
Xinchu Wang
1
, Biying Chen
1
, Guannan Dong
2
,
Naizhong Zhang
3
, W
eiyi
Liu
2
,
Jiaxu Han
1
, Cong-Qiang Liu
1,4*
, Si-Liang Li
1,4
, John M. Eiler
2
, Sheng Xu
1*
1
Institute of Surface
-Earth System Science, School of Earth System Science,
Tianjin University, Tianjin 300072, China
2
Division of Geological and Planetary Sciences, California Institute of
Technology, Pasadena, CA 91125, USA
3
Earth
-Life Science Institute, Tokyo Institute of Technology, Tokyo 152-8551,
Japan
4
Haihe Laboratory of Sustainable Chemical Transformations, Tianjin 300072,
China
−
Corresponding author: Cong-Qiang Liu; Sheng Xu
−
E-
mail: liucongqiang@tju.edu.cn; sheng.xu@tju.edu.cn
This PDF file includes:
1. Supplementary Texts
Text S1. Monitoring information of CBM wells after CO
2
injection
Text S2. Possible diffusion results: off
-equilibrium clumped isotope signals
Text S3. Re
-equilibration model
Text S4. Mixing model of methane bulk and clumped isotopes
Text S5. Constraints on equilibrated end members bulk isotopes
Text S6. Comparison of Qinshui CBM and Olla filed CCS
2. Supplementary Figures
Fig. S1. Global Carbon Capture and Storage sites.
Fig. S
2. The gas composition of study area CBM samples.
Fig. S
3. Geochemistry of CBM for validation clumped isotope varying.
(a)
δ
13
C-CH
4
versus T
-
Δ
13
CH
3
D and (b)
δ
D-CH
4
versus T
-
Δ
12
CH
2
D
2
.
Fig. S
4. Three
-dimensional spatial plots of methane bulk isotopes and gas wetness.
Fig. S
5. Mass-
18 methane isotopologues re
-equilibration models.
Fig
. S
6. Equilibrium and disequilibrium mixing scenarios.
3. Supplementary T
able
s
Table S1. The endmembers bulk and clumped isotopic composition in the mixing
scenario
.
Table S2. The
mixing
model
results in
two
mixing scenarios
.
Table S
3. Field sample bulk and clumped isotope results used in the diffusion results
comparison.
References
1. Supplementary Texts
Text S1. Monitoring information of CBM wells after CO
2
injection
Based on previous field data and observation, CO
2
injection provides greater
methane displacement efficiency (> 90%) compared to conventional production, which
only recovers about 52% methane (Zhou et al., 2013). At high injection pressure (> 8
MPa) conditions, CO
2
has the potential to drive CH
4
completely from the coal seams
(Zhang and Ranjith, 2019)
. In less than two months, 192.8 t
and 233.6 t
liquid CO
2
were
injected into the target CBM wells, respectively (Pan et al., 2018). Both wells had
significant increases in bottomhole pressure (BHP) after CO
2
injection (from 1.19 MPa
to 6.7 MPa and from 4 MPa to 7.0 MPa). With the restart of CBM production, the BHP
recovered again after two months, indicating the process of gas adsorption and
desorption in the coal matrix. The CO
2
content gradually decreases (~30% after 5
months) from the initial high CO
2
content (~ 70%), indicating that the injected CO
2
was
adsorbed onto the coal during wells shutdown (Wong et al., 2007)
. T
he available
information suggest that the subsequent CO
2
injections were essentially like the initial
CO
2
injection. The two injected CBM well were closed and not in production conditions
for a long time before sampling. The impact of CCS (if present) should not be limited
to the injection wells, but should be extended to the entire CBM production block.
Text S2. Possible diffusion results
: off-
equilibrium clumped isotope signals
Some samples are showing a deviation from the equilibrium line toward the upper
left (TS
-7
or TS
-9)
(Fig. 3)
. The results from bulk isotopes do not distinguish the
influence of post
-generation processes, and the gas composition may be influenced by
precursors from organic matter, making it difficult to give clear conclusions (Liu et al.,
2019).
The field samples TS
-6, TS
-7, TS
-5 and ZK
-2 are selected as possible diffusion
results, and the gas initial state is chosen to be the average value of TS
-5 and ZK
-2
(Table S
2)
, which have the resembling isotopologues composition with equilibrated
results.
These results can be explained by considering the fractional distillation process
based on a 1:1 mass dependency relationship
(Giunta et al., 2019; Young et al., 2017)
(Fig. 3)
. It is unlikely that the CBM samples have the same starting point
for
the
diffusion
process. These deviations may be due to absorption/desorption processes or
uncertainties associated with the initial state selection.
The migration processes during
CCS of CBM wells could contribute. The injection of CO
2
displaces CH
4
in the
desorbed coal and further promotes CH
4
diffusion. As a consequence of the injection,
the injected CO
2
and
in situ
CH
4
gas rapidly meet and mix
(Li and Fang, 2014)
.
Compared to CH
4
, coal has a higher sorption capacity for CO
2
. Therefore, the replaced
CH
4
is usually produced in a carry-over mode and CBM production increases.
Apart from the anthropogenic CO
2
injection, the Stage III of basin deposition may
have been accompanied by gas migration due to stratigraphic uplift during the
Yanshanian
(Wei et al., 2007)
, as indicated by the geochemical classification of Bernard
et al. (1976) and Whiticar (1999) in
Fig. 2
. While the observed elevated C
1
/C
2+
ratio
may be linked to the heterogeneity of the organic matter, in particular the Type II
or
III
kerogen, the potential influence of post
-generation effects cannot be entirely dismissed.
Evidence from clumped
isotopes suggests off
-equilibrium sample compositions (TS
-7,
TS-
9). Gas displacement
-migration
-dominated processes may be responsible for the
disequilibrium distribution.
Isotopologues artifacts may also be
present
, possibly related to kinetic processes
in other geologic settings. The K
loof
M
ine
and
Acqusasanta Terme have the upper
-left
deviation of the thermodynamic equilibrium line as a result of the enriched
∆
12
CH
2
D
2
(e.g.,
∆
13
CH
3
D < +2‰
,
∆
12
CH
2
D
2
> +15‰ for K
loof
Mine
)
(Fig.
3)
(Young et al.,
2017)
.
Text S3. Re-
equilibration
model
and
AOM Rayleigh equations
In the d
eep
-sea hydrothermal fluids system
, the methane from Rainbow and Von
Damm showed consistent equilibrium isotopologues with vented fluids' apparent
temperature, but the Lost City clumped isotope
-based temperatures are much higher
than those of vented fluids, which means methane
∆
13
CH
3
D large depleted
(Fig. 3)
(Labidi et al., 2020)
. These clumped isotope anomalies may result from individual
changes in the mass
-18 isotopologues, but the inconsistency of bulk and clumped
isotopes would also be expected. Outside of these two known equilibria
clumped
endmembers
, the a
biotic methanogenic processes occur, such as the Sabatier reaction,
in which
ruthenium converts CO
2
to CH
4
at temperatures below 100°C (Etiope and
Sherwood Lollar, 2013). The silane hydration reaction
(Si
5
C
12
H
36
and
H
2
O) also
proceeds at low temperatures
(Young et al., 2017), which would be accompanied by a
significant deficit of
∆
12
CH
2
D
2
. It seems that the low-
temperature intervals are likely
to result from the catalysis of methane bonds by low
-temperature fluids (60
– 90°C)
catalyzed by metal or clay minerals (Labidi et al., 2020)
. The caveat for abiotic methane
is few evidence in the study area for a noticeable methane contribution from low
-
temperature fluids. On the other hand, assuming a mass balance model in an open
system where a large amount of thermogenic methane transports
through the abiogenic
gas generation channel
s, advective processes have a larger scale compared to gas
mixing or diffusion, and should not be expected to cause sufficiently fractionation of
the two clumped
isotopes
(∆
13
CH
3
D and
∆
12
CH
2
D
2
) (Liu et al., 2023).
Additionally, the microbially derived methane gas has a significantly depleted
clumped
isotop
e composition
(Fig. S5)
, but mixing the depleted methane with the
known two end-
members
does not yield clumped isotope
signature
s in the study area.
To make mixing scenarios feasible, it is necessary to postulate a unique disequilibrium
endmember, exhibiting with
∆
13
CH
3
D de
pletion or
∆
12
CH
2
D
2
enric
hment (
∆
13
CH
3
D =
+2.20‰,
∆
12
CH
2
D
2
= +15.00‰), positioned towards the upper
-left of the equilibrium
line
(Fig. S5)
. Such a hypothetical endmember has not been reported in previous studies
and has no plausible reason to appear within the CBM study area. A
lso
, it is important
to consider the other nonlinear mixing effects of clumped isotopes
: there may be a slight
12
CH
2
D
2
enrichment resulting from mixing of different
δ
13
C and
δ
D materials, which
changes the stochastic distribution without intermolecular isotope exchange (Eiler,
2007; Eiler, 2013)
. It
is suggested that the post
-generation mixing and uplift of gases at
different maturity levels may lead to comparable enriched
∆
12
CH
2
D
2
clumped isotope
distributions in natural gas reservoirs (Xie et al., 2021)
. In the study area, coal seams
did not
exhibit a clear
range of maturity (
R
o
), but gas production in Shizhuang and
Yushe occurs at varying depths of 400 m to 1200 m (Pan et al., 2018)
, CBM from
different depth are possible to contribute.
It is necessary to explore the internal dynamics mechanisms
that lead to deviations
from the thermodynamic equilibrium line for both
clumped
methane isoto
pologues.
The
factors driving the methane C-
H bonds
re-ordering
are possibly from biocatalytic or
non-
biocatalytic processes (Labidi et al., 2020; Zhang et al., 2021)
. The re
-equilibration
is modelled based on the isotope exchange reactions:
12
13
13
12
34 34
1
CH D + CH
CH D + CH
k
→
(1)
12
12
12
12
3
3
22
4
2
CH D + CH D
CH D + CH
k
→
(2)
When isotope exchange rate constant ratio
k
r
is defined by
k
1
/
k
2
, the
k
r
represents
the degree of the re-
equilibration from the starting moment (
t
= 0) to the present state
(
t
). The “EQ” represents the calculated thermodynamic equilibrium state. It helps us to
simplify the change of two mass
-18 methane clumped isotopologues, both for bond re
-
ordering in abiogenic and biogenic systems (Giunta et al., 2021).
1
13
13
33
EQ
13
13
33
0
EQ
CH D -
CH D
CH D
-
CH D
kt
t
e
−
=
(3)
2
12
12
22
22
EQ
12
12
22
22
0
EQ
CH D
-
CH D
CH D
-
CH D
t
t
k
e
−
=
(4)
The
reservoir
ambient temperature is typically used as a proxy for endmember
(Tyne et al., 2021; Zhang et al., 2021)
. The c
lumped isotope-
based
temperatures in the
study area can reach as low as 69.9°C, within the interval of the geological cooling
settings temperature change (
ca.
250 –
60 °C)
(Chen et al., 2019; Zhang et al., 2018)
.
Here we consider two scenarios of re-
equilibration, both starting with the “ideal
thermogenic CBM” and ending with (1)
low temperature equilibrium represented by
the CBM reservoir temperature (~ 21°C:
∆
13
CH
3
D = +5.89‰,
∆
12
CH
2
D
2
= +20
.15‰),
or (2
) the sample (TS
-4:
∆
13
CH
3
D = +
4.55‰,
∆
12
CH
2
D
2
= +13.01‰) in the study area
closest to the low temperature equilibrium.
In both re
-equilibrium models,
k
r
= 1.5 – 2 for ambient temperature
ending and
k
r
= 5 for cooling ending essentially explain the clumped isotope distribution of the
samples within the study area
(Fig. S
3)
. The rate constant ratio
k
r
values can be equated
with the degree of difference in bond-
order rearrangement rates
(Giunta et al., 2021)
,
in which case
k
r
< 2 re
-equilibration
processes
should
be expected
to cover
the
off
-
equilibrium samples. It can be suggested that the methane bond
re-equilibration
could
be a plausible explanation, and that
prolonged stratigraphic cooling after the magmatic
event
may
not
be
the
major contributor.
Additionally, the initial clumped isotope compositions could be altered by
anaerobic oxidation of methane (AOM). The partial reversibility of AOM initial step
enzymatic reaction has been experimentally
(laboratory methanogen cultures)
demonstrated
(Ash et al., 2019; Ono et al., 2021)
. The methyl coenzyme M reductase
(Mcr) catalytically exchanges intracellular isotopologues with high reversibility
provide a convergence of
13
CH
3
D and
12
CH
2
D
2
to a specific equilibrium temperature
(Liu et al., 2023)
.The Rayleigh equations are used to help understand the effect of
certain geochemical reactions (e.g., AOM and other methane oxidation processes) on
isotopologue changes:
∆
13
CH
3
D =
∆
13
CH
3
D
init
+ (
13,2
γ
·
13
α
·
2
α
–
13
α
–
2
α
+ 1) ln
f
(5)
∆
12
CH
2
D
2
=
∆
12
CH
2
D
2,
init
+ (
2,2
γ
·
2
α
·
2
α
– 2·
2
α
+ 1)
ln
f
(6)
The clumped isotope fractionation can be described by the isotope fractionation
factors (
α
) and the kinetic clumped isotopologue factor (
γ
), where the
f
represents
residual methane or the reaction process (Ono et al., 2021; Wang et al., 2016; Whitehill
et al., 2017)
. By extrapolating the isotope fractionation to the open system (Liu et al.,
2023; Wang et al., 2016)
, only the fractionation trajectories change, while the
enrichment trends of the two clumped isotopes remain unchanged:
(
)
(
)
(
)
(
)
(
)
13,2
13
2
13
13
13
33
2
1
α α
1
ln
1+
α
1
1+
α
1
CH D
CH D
∆
+ ⋅⋅−
=
−
−
−
∆
γ
init
φ
φφ
(7)
(
)
(
)
(
)
2,2
2
2
2
12
1
2
2
2
2
22
1
α α
1
lH
n
1
CH
+
α
1
D
CD
∆
+⋅
−
∆
⋅
=
−
−
γ
, init
φ
φ
(8)
According to the fractionation factors (
1
·
13
α
,
2
α
and
13,2
γ
,
2,2
γ
) obtained from
previous microbial culture experiments, the two clumped isotopes would become
enriched due to AOM (if the
f
= 0.5, the
∆
13
CH
3
D is enriched with more than ~ 5
‰
)
(Ono et al., 2021).
Te x t S
4. Mixing model of methane bulk and clumped isotopes
The methane bulk and clumped isotopes mixing model
s are reported in previous
studies
(Labidi et al., 2020; Zhang et al., 2021)
, the detail processes are shown here.
The measured
δ
and
∆
values should be rewritten as an abundance ratio format
:
13
R
i
= (
δ
13
C
i
/ 1000 + 1) ×
13
R
VPDB
(9)
D
R
i
= (
δ
D
i
/ 1000 + 1) ×
D
R
SMOW
(10)
13CH3D
R
i
= (
∆
13
CH
3
D
i
/1000+1) × (
4 ×
13
R
i
×
D
R
i
)
(11)
12CH2D2
R
i
= (
∆
12
CH
2
D
2,
i
/1000+1) × (6 ×
D
R
i
×
D
R
i
)
(12)
The bulk and clumped isotope mixing models are conducted based on isotope
ratios (R) of each endmember (A, B):
13
R
m
=
f
×
13
R
A
+ (1 –
f
) ×
13
R
B
(13)
D
R
m
=
f
×
D
R
A
+ (1 –
f
) ×
D
R
B
(14)
13CH3D
R
m
=
f
×
13CH3D
R
A
+(1 –
f
) ×
13CH3D
R
B
(15)
12CH2D2
R
m
=
f
×
12CH2D2
R
A
+(1 –
f
) ×
12CH2D2
R
B
(16)
The subscript
m
is the mixture of A and B, the
f
value represents the fraction of
endmember A, the bulk and clumped isotope compositions of mixture can be expressed
as:
δ
13
C
m
=
f
× (
13
R
A
/
13
R
VPDB
– 1) +(1 –
f
)× (
13
R
B
/
13
R
VPDB
– 1)] ×1000
(17)
δ
D
m
= [
f
× (
D
R
A
/
D
R
SMOW
– 1) + (1 –
f
)× (
D
R
B
/
D
R
SMOW
– 1)] ×1000
(18)
∆
13
CH
3
D
m
= [ [
f
×
13CH3D
R
A
+ (1 –
f
)×
13CH3D
R
B
/ (4 ×
13
R
m
×
D
R
m
)] – 1]
×1000
(19)
∆
12
CH
2
D
2,
m
= [ [
f
×
12CH2D2
R
A
+ (1 –
f
)×
12CH2D2
R
B
/ (6 ×
D
R
m
×
D
R
m
)] –
1] ×1000
(20)
T
he calculation of the mixing model is based on four methane isotopologues
, and
the optimal model result is searched by integrating
f
values from 0 to 1 for 10
5
times
(fsolve function). The sklearn.metrics package is used to calculate the R
2
and RMSE
between the mixing model and actual results. The results are
shown in
T able
S2
.
Text S5
. Constraints on e
quilibrated end members bulk isotopes
From
our discussion in the main text, it is evident that the mixing should occur
between two equilibrated end
members. The clumped isotope composition
∆
13
CH
3
D
and
∆
12
CH
2
D
2
are determined by thermodynamic equilibrium
, but
the bulk isotope
composition should come from other independent constraints
(Turner et al., 2021).
The methane production from coal and shale involves the conversion of geological
macromolecules within source rock organic matter into low
-molecular weight organic
intermediate compounds (C
LMW
). Coal has a
high total organic carbon content and a
high yield of C
LMW
, which is thought to be the starting point of coal
-to-methane
conversion
(Vinson et al., 2017)
. T
he thermogenic CBM are assumed
to
have similar
bulk isotope compositions with C
LMW
, where
δ
13
C- CH
4
of –
25‰ and
δ
D-
CH
4
of –
160‰.
Furthermore, this value also represents the most negative bulk isotope
composition found within the study area, which is reasonable as a thermogenic end
member.
Turner
et al determined
the carbon isotopic equilibrium between CH
4
(g) and
CO
2
(g) using Path Integral Monte Carlo (PIMC) and
the hydrogen isotope equilibrium
based on experiments using
γ
-Al
2
O
3
and Ni catalysts from 3 to 200 °C
(Turner et al.,
2021)
. The
equilibrated
α
CO2-CH4
and
α
H2O-CH4
can be calculated based on
ambient
temperature (~ 21°C)
:
α
CO2-CH4
= 1.073 and
α
H2O
-CH4
= 1.206. In
our
another study
within the Qinshui project, t
he
carbon isotope of CO
2
(
δ
13
C- CO
2
) ranges from
0 to +25‰
and
the
hydrogen isotope of H
2
O (
δ
D-H
2
O)
range
s from
–85‰ to–75‰
in the study
area (Chen et al., under review).
Thus, the approximate estimation of the
equilibrium
CH
4
bulk isotopic range was determined based on these independent constraints (
Table
S1; Fig. S5).
Text S6. Comparison of Qinshui CBM and Olla filed CCS
For CO
2
bioconversion, the lack of positive evidence does not constitute disproof,
so it is worth considering whether we can place limits on the component of studied
methane that may come from conversion of the anthropogenic injected CO
2
. A reliable
method to determine whether CO
2
undergoes bioconversion into CH
4
is to compare it
with existing CCS regions. A notable study in this regard is conducted by Tyne et al. in
the Olla oilfield.
The understanding of the mechanisms of CO
2
trapping after injection
is critical for evaluating the safety and efficiency of geological sequestration of CO
2
(Leung et al., 2014; Pan et al., 2018). Although the main CO
2
sink is the dissolution
into pore fluids (formation water pH of 5 –
5.8) instead of the methanogenic bacteria
utilization based on natural gas field CCS experiences, methanogenesis through the
following reaction could be
significant in CCS areas (Gilfillan et al., 2009; Tyne et al.,
2021):
CO
2
+ 4H
2
→
CH
4
+ 2H
2
O
(21)
Tyne
et al. (2021) estimated that 13 –
19% of CO
2
was consumed by methanogenic
bacteria
in the Olla field after CO
2
injection in the 1980s, based on models of C stable
isotope, noble gas isotope, and clumped isotope data (Tyne et al., 2021)
. The Olla field
produced relatively positive
∆
13
CH
3
D of +5.66 ± 0.39‰ and
∆
12
CH
2
D
2
of +12.46 ±
1.07‰ signatures, which are distinct from the Nebo-
Hemphill clumped isotope results
used for comparison
(Fig. 3)
. Due to the uncertainty in the metabolic processes,
distinguishing between anthropogenic CO
2
reduction and
in situ
biogenic methane
isotopic endmembers remains challenging (Tyne et al., 2023).
When considering CO
2
conversion in the study area, a distinction of Qinshui CBM
from the Olla field is that the mixing of thermogenic methane and biogenic methane
should occur between equilibrated endmembers, where the ambient equilibrated
endmembers represent AOM and/or very
slow methanogenesis. On the one hand, if the
rate of methanogenesis is slow enough to support the equilibrated isotope exchange,
then we should anticipate that methane production from reduction of injected CO
2
will
be relatively slow (Gropp et al., 2022). On the other hand, a rapid biological
methanogenesis (e.g., rapid hydrogenotrophic methanogenesis, usually comes with
disequilibrium clumped isotope compositions) could not be consistent with our
observations (equilibrium clumped isotope distribution) if the metabolic reactions
exhibited incre
ased reversibility due to enzymatic catalysis (e.g., Mcr
-catalyzed
reactions during AOM)
(Ash et al., 2019; Giunta et al., 2019; Liu et al., 2023; Warr et
al., 2021)
.
2. Supplementary Figures
Fig. S1.
Global
Carbon Capture and Storage
sites. CCS sites data from National Energy
Technology Laboratory’
s (NETL) CCS Database
:
https://netl.doe.gov/carbon-
management/carbon
-storage/worldwide
-ccs
-database.
Fig. S2.
The gas composition of study area CBM samples. The
“other gas
” represents
gas components other than CO
2
and CH
4
, including alkanes such as ethane, propane,
N
2
, H
2
, etc.
Fig. S3.
Geochemistry of CBM
for validation
clumped isotope varying.
(a)
δ
13
C-
CH
4
versus
T-
Δ
13
CH
3
D and (
b)
δ
D-CH
4
versus
T-
Δ
12
CH
2
D
2
.
Fig
. S4.
Three
-dimensional spatial plots of methane bulk
isotopes (
δ
13
C-CH
4
and
δ
D-
CH
4
) and gas wetness. The green and blue scatter points indicate the projection of the
3D coordinate points in ZX and YZ coordinates, respectively. The gray line represents
the mixing scenarios
, the dashed line represents the unlikely mixing scenarios
, and the
solid line represents the most likely mixing scenario. The black arrows represent the
change from dry gas to wet gas. The ellipse represents the 95% confidence interval of
the CBM samples (except TS
-1- C).