Krista S. Walton

Yoshiaki Kawajiri, Matthew J. Realff, David S. Sholl, Ryan P. Lively

Stephen J. DeWitt, Rohan Awati, Jongwoo Park, Eli Carter, Hector Rubiera Landa

Georgia Institute of Technology

School of Chemical & Biomolecular Engineering

Atlanta, GA 30332

Enabling 10 mol/kg swing capacity in post-

combustion CO2 capture processes

BM Sanderson, BC O’Neill, C Tebaldi, Geophysical Research Letters 2017

RP Lively, MJ Realff. AIChE J. 2016

Adsorption (and membranes) are materials-enabled separations2

SH Pang, CW Jones et al., J. Am. Chem. Soc., 2017, 139, 3627-3630

Adsorption (and membranes) are materials-enabled separations3

Pressure (torr)

CO

2a

dso

rbe

d (

mm

ol/g

)

SH Pang, CW Jones et al., J. Am. Chem. Soc., 2017, 139, 3627-3630

Mmen-Co2(dobpdc)

Pressure (torr)

25°C

75°C

Adsorption (and membranes) are materials-enabled separations4

TM McDonald, JR Long et al., Nature, 2015, 519, 303-308

Pressure (torr)

CO

2a

dso

rbe

d (

mm

ol/g

)

SH Pang, CW Jones et al., J. Am. Chem. Soc., 2017, 139, 3627-3630

Connecting materials to engineering solutions—fibers lead the way5

WJ Koros, RP Lively. AIChE J. 2012, 58(9)

Connecting materials to engineering solutions—fibers lead the way6

2 µm

10,000 m2 / m3

module volume

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

Module

make-up

Sorbent dispersed in

polymer solution

WJ Koros, RP Lively. AIChE J. 2012, 58(9)

Connecting materials to engineering solutions—fibers lead the way7

10,000 m2 / m3

module volume

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

Module

make-up

Sorbent dispersed in

polymer solution

WJ Koros, RP Lively. AIChE J. 2012, 58(9)

Rapid thermal swing adsorption—amines/hollow fiber sorbents8

Mesoporous silica / PEI

Swing capacity and cycle time are

key for driving down capital costs of

adsorption-based CO2 capture

systems!

Key question: Can we increase swing

capacity by 10x and reduce cycle time by 5x

to dramatically drive down adsorbent costs?

Y Fan, CW Jones et al., Int. J. Greenhouse Gas Control 2014, 21, 61-71

Costs dominated by costs of adsorbent

Rapidly cycled pressure swing adsorption using MOFs9

Cycle times of ~20 seconds are common for industrial RCPSA (>5x faster than RTSA)

JM Simmons, T Yildirim et al., Energ. Env. Sci., 2011, 4(6), 2177-2185

Rapidly cycled pressure swing adsorption using MOFs10

Cycle times of ~20 seconds are common for industrial RCPSA (>5x faster than RTSA)

0.0 0.5 1.0 1.5 2.0 2.50

10

20

30

40

50

CO

2 U

pta

ke (

mol/kg)

Pressure (bar)

213 K

228 K

243 K

258 K

273 K

Pressure Swing Adsorption

ΔP = 1.9 bar

Sub-Ambient ΔNCO2

~ 40 mol/kg

210 220 230 240 250 260 270 2800

5

10

15

20

25

30

35

40

45

N

CO

2 (

mol/kg)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

Pdes = 0.5 bar

Pdes = 1.0 bar

J Park, RP Lively, DS Sholl, J. Mater. Chem. A. 2017, 5, 12258-12265

Rapidly cycled pressure swing adsorption using MOFs11

Cycle times of ~20 seconds are common for industrial RCPSA (>5x faster than RTSA)

𝑝𝐶𝑂2𝑎𝑑𝑠 = 2 𝑏𝑎𝑟

𝑝𝐶𝑂2𝑑𝑒𝑠 = 0.1 𝑏𝑎𝑟

J Park, RP Lively, DS Sholl, J. Mater. Chem. A. 2017, 5, 12258-12265

Enabling 10 mol/kg swing capacities via flue gas pretreatment12

Air Liquide Sub-Ambient Membrane System

Hollow fiber membrane

D Hasse, S Kulkarni et al., Energy Procedia, 2013, 37, 993-1003

Enabling 10 mol/kg swing capacities via flue gas pretreatment13

Air Liquide Sub-Ambient Membrane System

D Hasse, S Kulkarni et al., Energy Procedia, 2013, 37, 993-1003

Enabling 10 mol/kg swing capacities via flue gas pretreatment14

Air Liquide Sub-Ambient Membrane System

Key parameters: swing capacity & selectivity

Sub-Ambient Adsorption System

➢ DOE guideline: 90% CO2 removal from

flue gas

➢ Total heat integration with no external

cold (i.e., refrigerant) or hot (i.e., steam)

utility

➢ Costs between $35-$45/tonne CO2

➢ Parasitic loads of 18-30%D Hasse, S Kulkarni et al., Energy Procedia, 2013, 37, 993-1003

Enabling 10 mol/kg swing capacities: Potential MOF candidates15

0 1 2 3 4 50

2

4

6

8

10

12

14

16

18

Up

take

(m

mo

l/g

)

Pressure (bar)

223 K

233 K

243 K

253 K

263 K

273 K

HKUST-1 CO2 Isotherms

0 1 2 3 4 50

2

4

6

8

10

Up

take

(m

mo

l/g

)

Pressure (bar)

223 K

233 K

243 K

253K

263 K

273 K

UiO-66 CO2 Isotherm

0 1 2 3 4 50

5

10

15

20

25

30

Up

take

(m

mo

l CO

2/g

)

Pressure (Bar)

223 K

233 K

243 K

253 K

263 K

273 K

MIL-101(Cr) CO2 Isotherms

220 230 240 250 260 270 2800

5

10

15

20

CO2 single component

N

CO

2 (

mo

l/kg

)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

Pdes = 0.5 bar

Pdes = 1.0 bar

210 220 230 240 250 260 270 2800

1

2

3

4

5

6

7P

ads = 2.0 bar

Pdes

= 0.1 bar

UiO-66

UiO-66-NH2

UiO-66-(NH2)2

UiO-66-CH2NH

2

N

CO

2 (

mo

l/kg

)

Temperature (K)

210 220 230 240 250 260 270 2800

3

6

9

12

15

18

Cu-BTC

N

CO

2 (

mol/kg)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

[1] MJ Cliffe, AL Goodwin et al., Nature Comm, 2014, 5

[1] [3]

[3] L Hamon, GD Weireld et al., J. Am. Chem. Soc. 2009, 131, 8775-8777

[2]

[2] A Zukal, J Jagiello et al., Catal. Today 2015, 243, 69-75

UiO-66 HKUST-1 MIL-101(Cr)

Enabling 10 mol/kg swing capacities: Potential MOF candidates16

0 1 2 3 4 50

2

4

6

8

10

12

14

16

18

Up

take

(m

mo

l/g

)

Pressure (bar)

223 K

233 K

243 K

253 K

263 K

273 K

HKUST-1 CO2 Isotherms

0 1 2 3 4 50

2

4

6

8

10

Up

take

(m

mo

l/g

)

Pressure (bar)

223 K

233 K

243 K

253K

263 K

273 K

UiO-66 CO2 Isotherm

0 1 2 3 4 50

5

10

15

20

25

30

Up

take

(m

mo

l CO

2/g

)

Pressure (Bar)

223 K

233 K

243 K

253 K

263 K

273 K

MIL-101(Cr) CO2 Isotherms

220 230 240 250 260 270 2800

5

10

15

20

CO2 single component

N

CO

2 (

mo

l/kg

)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

Pdes = 0.5 bar

Pdes = 1.0 bar

210 220 230 240 250 260 270 2800

1

2

3

4

5

6

7P

ads = 2.0 bar

Pdes

= 0.1 bar

UiO-66

UiO-66-NH2

UiO-66-(NH2)2

UiO-66-CH2NH

2

N

CO

2 (

mo

l/kg

)

Temperature (K)

210 220 230 240 250 260 270 2800

3

6

9

12

15

18

Cu-BTC

N

CO

2 (

mol/kg)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

[1] MJ Cliffe, AL Goodwin et al., Nature Comm, 2014, 5

[1] [3]

[3] L Hamon, GD Weireld et al., J. Am. Chem. Soc. 2009, 131, 8775-8777

[2]

[2] A Zukal, J Jagiello et al., Catal. Today 2015, 243, 69-75

UiO-66 HKUST-1 MIL-101(Cr)

High stability

Low cost / high scalability

Low swing capacity

Low stability

Low cost / high scalability

High swing capacity

Good stability

Moderate cost

High swing capacity

Complexities of developing engineering solutions for post-

combustion CO2 capture (next 4 slides)

17

Scaling-up MOF contactors

• Synthesis

• Stability

• Etc.

Transport Limitations

• Heat effects

• Pressure drop

• Etc.

Contaminants

• Acid gases

• Water

• Etc.

Systems Engineering

• Optimized cycles

• Process Economics

• Etc.

Complexities: Scaling up MOF fiber sorbent contactors18

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

MOF

Module

make-up

UiO-66 or MIL-101(Cr)

dispersed in polymer

solution

Scale-up of MOFs—Inmondo Tech

Complexities: Scaling up MOF fiber sorbent contactors19

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

MOF

Module

make-up

UiO-66 or MIL-101(Cr)

dispersed in polymer

solution

3 μm

Unpublished data

Complexities: Scaling up MOF fiber sorbent contactors20

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

MOF

Module

make-up

UiO-66 or MIL-101(Cr)

dispersed in polymer

solution

3 μm

Unpublished data

Surface areas of fiber sorbents are

equivalent to (mass frac. MOF ) x

(MOF surface area)

Complexities: Scaling up MOF fiber sorbent contactors21

Peristaltic Pump

Swagelok®

Module

w/ fibers

Liquid Reservoir

Cu(NO3)2 BTC + Cu(NO3)2

ZnO (Zn, Cu)

HDSHKUST-1

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

ZnO

Module

make-up

ZnO dispersed in

polymer solution

BR Pimentel, RP Lively et al., Ind. Eng. Chem. Res. 2017, 56(17), 5070-5077

Complexities: Scaling up MOF fiber sorbent contactors22

Peristaltic Pump

Swagelok®

Module

w/ fibers

Liquid Reservoir

Spinneret

Quench bath

Fiber collectionHollow fiber

sorbent spinning

ZnO

Module

make-up

ZnO dispersed in

polymer solution

BR Pimentel, RP Lively et al., Ind. Eng. Chem. Res. 2017, 56(17), 5070-5077

∆𝑃 𝐿(𝑃𝑎/𝑐𝑚

)

Superficial velocity (cm/s)

Complexities: Transport Limitations23

Unpublished data

Te

mp

era

ture

Enthalpy

ADSORPTION

DESORPTION

Concept of Phase Change Material

for PSA Heat Management

∆𝑃 𝐿(𝑃𝑎/𝑐𝑚

)

Superficial velocity (cm/s)

Tem

pera

ture

[K

]

Complexities: Transport Limitations24

Sorbent-loaded

porous polymer

matrix

µPCM

Melt/

Freeze

Sorption/

desorption

enthalpy

Fiber Module

Phase Change

Material

Impermeable

microcapsule

Unpublished data

μPCM

UiO-6610 μm

Complexities: Transport Limitations25

Sorbent-loaded

porous polymer

matrix

µPCM

Melt/

Freeze

Sorption/

desorption

enthalpy

Fiber Module

Phase Change

Material

Impermeable

microcapsule

Unpublished data

μPCM

UiO-6610 μm

Complexities: Transport Limitations26

Unmodulated

0 100 200 300 400 500 600 700 800 900

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0.22

0.24

Modulated Fibers

Unmodulated FibersF

ibe

r C

ap

acity (

mm

ol/g

)

Flow Rate (sccm)

Breakthrough Capacity at 5% leakage(238K, 16 bar)

Bre

akth

rough c

apacity

at 5%

le

akage (

mm

ol/g)

Unpublished data

9x

Unpublished data

Complexities: Contaminants and Stability27

Unpublished data

Complexities: Contaminants and Stability28

0 2 4 6

-2

0

2

4

6

8

10

12

14

16

18

Bed P

ressure

(bar)

Time (Hours)

Bed Pressure

Fiber sorbents were cycled in

CO2/N2 for 4 weeks (~4000

cycles)

Complexities: Cycle optimization and systems engineering29

Unpublished data

Complexities: Cycle optimization and systems engineering30

Unpublished data

Conclusions and perspectives31

Key question: Can we increase swing capacity by 10x and reduce cycle time by 5x to dramatically

drive down adsorbent costs?

• Combining RCPSA cycles with appropriate metal-organic frameworks in sub-ambient conditions

results in highly productive adsorption systems (i.e., tonne CO2/tonne adsorbent-day)

220 230 240 250 260 270 2800

5

10

15

20

CO2 single component

N

CO

2 (

mo

l/kg

)

Temperature (K)

Pads = 2.0 bar

Pdes = 0.1 bar

Pdes = 0.2 bar

Pdes = 0.3 bar

Pdes = 0.5 bar

Pdes = 1.0 bar

• Significant “real world” complexities exist, but hollow fiber sorbent platform provides solutions to

many of these (scalability, transport limitations, etc.)

• Costs in the range of $35-$45/tonne CO2 may be achievable using these materials in this

process concept, but significant work remains

Process Scope—Key Topics, BP3 (Jan 18-Dec 18)

Eight major activity areas for BP2: Task 15.0: Process flowsheet refinement —Ongoing, 80% complete

Task 16.0: Generate >250 g/quarter of UiO-66 and spin fibers —Ongoing, 80% complete

Task 17.0: Construct/test RCPSA system for dirty gas testing—Ongoing, 50% complete

Task 18.0: Model Validation for fiber module —Complete

Task 19.0: Monolithic Fiber sorbent stability in dirty gases —Ongoing, 25% complete

Task 20.0: Composite (PCM containing) fiber testing in sub-ambient PSA—Complete

Task 21.0: Sub-ambient Technical Feasibility Study —Ongoing, 50% complete

Task 22.0: Large module testing in sub-ambient PSA —Task Initiated this quarter

32

Collaborators and funding 33

Collaborators

• Yoshiaki Kawajiri (GT)

• Ryan Lively (GT)

• Matthew Realff (GT)

• David Sholl (GT)

• Eli Carter

Walton Lab 2018