Dissolution and carbonation
of activated serpentine
for combined capture and storage
Mischa Werner1, Lilian Gasser1, Daniela Zingaretti2,
Hari Subrahmaniam1, Marco Mazzotti1
TCCS-6
Accelerated Carbonation for Environmental and Materials Engineering
Trondheim, Norway, 14 – 16 June 2011
1) Institute
of Process Engineering, ETHZ
2) University of Rome, Thor Vergata
Outline
 Flue gas mineralization – the concept
 Experimental setup and material
 Capture via mineralization:
 EQ3/6 equilibrium simulations
 Dissolution experiments
 Dissolution model
 Storage via mineralization:
 EQ3/6 equilibrium simulations
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
2
Flue gas mineralization within a CCS system
spent flue gas
CO2 in
flue gas
silicate
Capture plant
Grinding
pure CO2
Leaching
Precipitation
carbonate
Geological
storage
Traditional approach
Novel approach
CO2 in
flue gas
silicate
CO2 capture &
grinding (wet)
CO2 capture
& leaching
Precipitation
carbonate
spent flue gas
after Verduyn et al. 2009, TCCS5, Trondheim
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
3
Flue gas mineralization within a CCS system
spent flue gas
CO2 in
flue gas
silicate
Capture plant
Grinding
pure CO2
Leaching
Precipitation
carbonate
Geological
storage
Novel approach
CO2 in
flue gas
silicate
CO2 capture &
grinding (wet)
CO2 capture
& leaching
Precipitation
carbonate
spent flue gas
after Verduyn et al. 2009, TCCS5, Trondheim
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
4
Mineralization plant
to vent
high-P CO2
buffer tank
front pressure
regulator
mass flow
controller
gas analysis
(MS)
preheater
back pressure
condenser
regulator
stirrer
compressor
inlet mass flow
pump
meter
make-up
gas
100% CO2
gas feed
90% N2
10% CO2
make-up
liquid
Key:
regular capillary
heated capillary
isolated from rack
by flexible connections
Thursday, June 16, 2011
sample
withdrawal
Raman probe
and make-up
30-180oC
cooler
reactor
oil bath
high capacity low readability
balance
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
outlet
pump
pH probe
liquid analysis
(IC)
to sample collector
(& offline analysis)
5
Mineralization plant
PCO2 = 11 – 150 bar
high-P CO2
buffer tank
mass flow
controller
preheater
gas feed
90% N2
10% CO2
make-up
liquid
Key:
regular capillary
heated capillary
isolated from rack
by flexible connections
Thursday, June 16, 2011
Raman probe
inlet mass flow
pump
meter
make-up
gas
100% CO2
30-180oC
gas analysis
(MS)
back pressure
condenser
regulator
stirrer
compressor
Gas phase:
100% CO2
front pressure
regulator
to vent
cooler
outlet
pump
pH probe
liquid analysis
(IC)
reactor
oil bath
high capacity low readability
balance
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
to sample collector
(& offline analysis)
6
Mineralization plant
PCO2 = 0.36 – 4 bar
high-P CO2
buffer tank
Synthetic flue gas:
gas analysis
10%vol CO2 in N2
(MS)
front pressure
regulator
mass flow
controller
preheater
back pressure
condenser
regulator
stirrer
compressor
gas feed
90% N2
10% CO2
make-up
liquid
Key:
regular capillary
heated capillary
isolated from rack
by flexible connections
Thursday, June 16, 2011
Raman probe
inlet mass flow
pump
meter
make-up
gas
100% CO2
to vent
30-180oC
cooler
outlet
pump
pH probe
liquid analysis
(IC)
reactor
oil bath
high capacity low readability
balance
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
to sample collector
(& offline analysis)
7
Activated serpentine




Mg2.42Fe0.58Si2O5(OH) 4
dry ground to <125 μm
thermal activation at 600°C
1.5 mol H2O per mole
serpentine removed
sub 20 µm
Thursday, June 16, 2011
sub 125 µm
 fractionation by dry
sieving
 fines removal w/ EtOH
 micropores (2-3 nm)
measured (BET method)
20 – 63 µm
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
63 – 125 µm
8
Capture part: EQ3/6 equilibrium simulations
CO2 solubility as function of T and [Mg2+], at different PCO2 levels
using flue gas (Ptot = 10PCO2):
using pure CO2:
PCO2 =
PCO2 =
4 bar
1.5 bar
0.36 bar
120 bar
60 bar
11 bar
 If flue gas used: CO2 solubility
predominately sensitive to [Mg2+]
 If pure CO2 used: low T favorable,
highest PCO2 not worth the while
 even at low PCO2 high CO2 loads possible
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
9
Act. serpentine dissolution experiments
Mg2.42Fe0.58Si2O5(OH)4
Magnesium concentration and conversion at 30oC:
flue gas
pure CO2
80
pCO2= 11 bar
p
CO2
0.6
= 1.5 bar
= 60 bar
Mg
2+
[mM]
p
CO2
0.8
0.4
0.2
0
0
100
pCO2= 0.36 bar
Conversion [%]
1
60
40
pCO2= 0.36 bar
pp == 0.36
bar
1.5 bar
CO2
CO2
20
2
4
6
8
time [h]
10
12
14
0
0
0.36
bar
ppCO2
== 11
1.5
bar
bar
CO2
pp == 11
0.36
1.5
bar
bar
bar
60
bar
CO2
CO2
2
4
8
6
time [h]
10
12
14
 Dissolution using flue gas similarly effective as with pure CO2
 Complex dissoltuion mechanism: Drastic slowdown after fast initial phase
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
10
Olivine vs activated serpentine dissolution
T = 30oC, PCO2 = 0.36 bar, flue gas mode, 20-63 μm fraction
act. serp and olivine in solution [µM]
200
act. serp, [Mg2+] based
olivine, [Mg2+] based
150
100
19 h runtime
50
43 h runtime
0
0
1
time: t/ttot
 Olivine so far “best candidate” Mg-silicate (most reactive)*
*e.g. O’Connor et al,. 2005, Albany Research Center
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
11
Dissolution model for natural serpentine
L0
Natural serpentine
dissolves acc. to
Shrinking core
model*
ash
D
time
 Population balance:
 Dissolution rate:
n = particle population, PSD
∂n ∂Dn
−
=
0
D = dL/dt, dissolution rate
∂t ∂L
D = f ( L, L0 )
dc
dt
 Solute mass balance: V =
dm
− Qc
dt
not constant, diffusion controls
c = solute concentration
m = mass of particle population
*e.g. Teir et al., 2007, Int J Miner Process 83;
Van Essendelft et al., 2009/2010, Ind Eng Chem Res 48/49
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
12
Dissolution model for activated serpentine
 Presence of micropores and altered crystal structure cause
complex, incongruent dissolution behavior
 Mg2.42Fe0.58Si2O5(OH)4  Silica detected in reactor solution:
T=30oC
Mg, 0.36 bar
Mg, 1.5 bar
Mg, 11 bar
Mg, 60 bar
Si, 0.36 bar
Si, 1.5 bar
Si, 11 bar
Si, 60 bar
0.6
0.4
Mg
2+
and silica [mM]
0.8
0.2
0
0
100
80
Conversion [%]
1
60
Mg, 0.36 bar
Mg, 1.5 bar
Mg, 11bar
Mg, 60bar
Si, 0.36 bar
Si, 1.5 bar
Si, 11 bar
Si, 60 bar
T=30oC
40
20
2
Thursday, June 16, 2011
4
6
8
time [h]
10
12
14
0
0
2
4
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
6
8
time [h]
10
12
14
13
Dissolution model for activated serpentine
 Presence of micropores and altered crystal structure cause
complex, incongruent dissolution behavior
 Mg2.42Fe0.58Si2O5(OH)4  Iron absent in reactor solution:
1
T=60oC
Fe, p
=1bar, original sample
Fe, p
=1bar, acidified&heated
CO2
Fe [µM]
CO2
 Preferential leaching of Mg2+
0
-1
0
2
Thursday, June 16, 2011
4
6
8
time [h]
10
12
14
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
14
Dissolution model for activated serpentine
 Presence of micropores and altered crystal structure cause
complex, incongruent dissolution behavior
 Back to single particle model to explore the role of:
 diffusion
 counterdiffusion
 ash layer
 unreacted core
 pores
 reactants (H+)
 products (Mg2+, silica)
 PSD discretized to perform preliminary simulations
 Single experiments can be fitted nicely, but fitting multiple
exeriments yet an issue
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
15
Dissolution model for activated serpentine
Measured and modelled Mg2+ profile and conversion:
1
T=90oC, PCO2 = 0.36bar, flue gas
100
measured
model
Conversion [%]
80
0.6
Mg
2+
[mM]
0.8
0.4
0.2
0
0
T=90oC, PCO2 = 0.36bar, flue gas
60
40
20
2
Thursday, June 16, 2011
4
6
8
time [h]
10
12
14
0
0
measured
model
2
4
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
6
8
time [h]
10
12
14
16
Serpentine dissolution experiments
T=30oC
[mM]
0.4
0
0
1
0.36 bar
1.5 bar
11 bar
60 bar
0.6
0.4
4
6
8
time [h]
10
12
14
0.6
0
0
0.4
2
4
6
8
time [h]
10
12
0
0
14
100
80
80
80
40
20
0
0
Mg,
Mg,
Mg,
Mg,
2
4
Thursday, June 16, 2011
6
8
time [h]
10
60
40
20
0.36 bar
1.5 bar
11bar
60bar
12
Conversion [%]
100
Conversion [%]
100
60
14
0.36 bar
1.5 bar
11 bar
60 bar
0.2
0.2
2
Mg,
Mg,
Mg,
Mg,
0.8
Mg
Mg
Mg
0.8
0.2
Conversion [%]
Mg,
Mg,
Mg,
Mg,
2+
0.6
2+
[mM]
0.8
1
0.36 bar
1.5 bar
11 bar
60 bar
[mM]
Mg,
Mg,
Mg,
Mg,
T=90oC
2+
1
T=60oC
0
0
Mg,
Mg,
Mg,
Mg,
2
4
6
8
time [h]
10
0.36 bar
1.5 bar
11 bar
60 bar
12
2
4
6
8
time [h]
10
12
14
60
40
20
14
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
0
0
Mg,
Mg,
Mg,
Mg,
2
4
6
8
time [h]
10
0.36 bar
1.5 bar
11 bar
60 bar
12
14
17
Storage part: EQ3/6 equilibrium simulations
Thermodynamic driving force (Q/K) as function of T and [Mg2+]:
PCO2= 0.36 bar, flue gas
PCO2= 60 bar, pure CO2
hydromagnesite
hydromagnesite
magnesite
nesquehonite
nesquehonite
magnesite
 Higher T favors carbonate precipitation
 High PCO2 not beneficial, flue gas conditions preferable!
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
18
Concluding remarks
 CO2 capture into aqueous solution promoted by presence of
Mg2+ from the dissolution process
 Activated serpentine dissolution fast even at low PCO2, thereby
preferential leaching of Mg2+
 Higher T promotes Mg leaching and reduces carbonate solubility
 Thermodynamic driving force for precipitation higher under flue
gas conditions
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
19
Acknowledgements
Funding support by:
• Shell Global Solutions Int. Amsterdam, special thanks to
Dr. Marcel Verduyn and Gert van Mossel
• Swiss National Fund for Science
• Competence Center Energy and Mobility, CH
• Competence Center Environment and Sustainability, CH
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
20
Thanks for your attention
 Flue gas mineralization – the concept
 Experimental setup and material
 Capture via mineralization:
 EQ3/6 equilibrium simulations
 Dissolution experiments
 Dissolution model
 Storage via mineralization:
 EQ3/6 equilibrium simulations
Thursday, June 16, 2011
Mischa Werner - D-MAVT - ETH Zurich - werner@ipe.mavt.ethz.ch
21