Bench-Scale Development and Testing of Rapid PSA for CO2 Capture

James A. Ritter & The Team

2013 NETL CO2 Capture Technology Meeting Pittsburgh, PA, July 8, 2013

Overall Project Objectives design, develop and demonstrate a bench-scale process

for the efficient and cost effective separation of CO2 from flue gas using Pressure Swing Adsorption (PSA)

goal to reduce energy consumption, capital costs, and environmental burdens with novel PSA cycle/flow sheet designs

applicable to both large (500-1000 MW) and small (5-50 MW) capacity power plants, and industries with 10 to 100 times less CO2 production

Process simulations and experiments; structured adsorbent material development, CFDs and

experiments; and complete flow sheet analyses being used for demonstrating and validating the concepts.

Grace (Hoefer)

Catacel (Cirjak)

USC (Ritter & Ebner)

Battelle (Saunders & Swickrath)

materials characterization,

and process modeling and

experimentation

technology development and

process integration

thin film materials

development and characterization

The Team investigation

specification

validation

PSA Technology Advantages established, very large scale technology for other

applications needs no steam or water; only electricity tolerant to trace contaminants; possibly with use

of guard or layered beds zeolite adsorbent commercial and widely

available increase in COE lower than other capture

technologies beds can be installed under a parking lot

PSA Technology Challenges energy intensive, but better than today’s amines;

possibly overcome by novel designs today, very large beds required implies large

pressure drop more power; possibly overcome by structured adsorbents and faster cycling

large footprint; possibly overcome by underground installation and faster cycling smaller beds

high capitol cost; possibly overcome by faster cycling smaller beds

Key PSA Technology Project Challenge although a commercial tri-sieve zeolite could be

used today in an efficient PSA cycle, it would only minimize to some extent the pressure drop issues, but not the adsorbent attrition and mass transfer issues

key challenge is to develop a structured adsorbent around an efficient PSA cycle that exhibits a high enough packing density to allow the fastest possible cycling rate ( smallest possible beds), while improving pressure drop and mass transfer issues and eliminating attrition issues

Where are we going?

• increase working capacity 10 fold (herculean) • operate at 1/10th cycle time (achievable) • known as rapid PSA

Scale of PSA System for CO2 Capture from 500 MW Power Plant

Is it possible to achieve a 1/10th volume reduction?

although rapid PSA offers potential for a low-cost solution for CO2 capture, the extent of size

reduction achievable is, at the moment, unknown

QuestAir H-6200 Rapid PSA-Installed at ExxonMobil Facility

H2 Production Rapid PSA ~ 12,000 Nm3/h/module

H2 Production Conventional PSA ~ 20,000 Nm3/h

Two of Questair’s modules do 20% better than this 6-bed PSA system

and are much smaller.

A 500 MW plant produces ~ 33,000 Nm3/h at > 30 times

lower pressure!

Where are we now after completing first year?

Preliminary Technical and Economic Feasibility Study

Overall Outcome

Significant Outcomes from Year 1 developed PSA cycle and process flow sheet with less than 35%

LCOE increase; based on completed preliminary technical and economic feasibility study

demonstrated zeolite crystals can be coated onto basic metal structure with at least 50 mm thick coating; suggests it may be possible to achieve even 100 to 150 mm coatings, if needed

demonstrated Catacel core structures can be made with up to 400 cells per square inch (cpsi); makes goal of achieving 600 cpsi, possibly even 800 cpsi, within reach

demonstrated needed limit of < 20 kPa/m pressure drop through 400 cpsi core at very high velocities up to 25 m/s; pressure drop limit utilized in preliminary technical and economic feasibility study

Significant Outcomes from Year 1 Predicted pressure drop through Catacel core nearly

quantitatively using CFD model with no adjustable parameters; paves way to fabricate even more optimum core structures using computational tools

Demonstrated, via PSA process simulation, possibly lowest energy, highest feed throughput PSA cycle for CO2 capture; amazing when considering bulk density reduced from 710 kg/m3 (typical for packed bed of zeolite beads) to 400 kg/m3 (entirely feasible with Catacel core)

PSA cycle boasts feed throughput of around 3,000 L(STP)/hr/kg and separations energy < 18 kJ/mol CO2 captured

How did we get to this point?

CO2 Isotherm on Adsorbent and Dual Process Langmuir Fit

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 20 40 60 80 100 120

CO

2 Loa

ding

[mol

/kg]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

0.010

0.100

1.000

10.000

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

CO

2 Loa

ding

[mol

/kg]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

Dual Process Langmuir (DPL) Isotherm

1, 1, 2, 2,

1, 2,1 20

, , ,

,0, ,

1 1

exp

s si i i i i i

ii i i isite site

s s stj i j i j i

j ij i j i

n Py b n Py bn

Py b Py b

n n n T

Bb b

T

− −

= + + + = +

=

N2 Isotherm on Adsorbent and Dual Process Langmuir Fit

0.00

0.10

0.20

0.30

0.40

0 20 40 60 80 100 120

N2 L

oadi

ng [m

ol/k

g]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

0.0001

0.0010

0.0100

0.1000

1.0000

0.01 0.10 1.00 10.00 100.00 1000.00

N2 L

oadi

ng [m

ol/k

g]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

1, 1, 2, 2,

1, 2,1 20

, , ,

,0, ,

1 1

exp

s si i i i i i

ii i i isite site

s s stj i j i j i

j ij i j i

n Py b n Py bn

Py b Py b

n n n T

Bb b

T

− −

= + + + = +

=

Dual Process Langmuir (DPL) Isotherm

O2 Isotherm on Adsorbent and Dual Process Langmuir Fit

0.00

0.05

0.10

0.15

0 20 40 60 80 100 120

O2 L

oadi

ng [m

ol/k

g]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

0.0000

0.0001

0.0010

0.0100

0.1000

1.0000

0.01 0.10 1.00 10.00 100.00 1000.00

O2 L

oadi

ng [m

ol/k

g]

Pressure [kPa]

T = 25 oC T = 50 oC T = 75 oC

1, 1, 2, 2,

1, 2,1 20

, , ,

,0, ,

1 1

exp

s si i i i i i

ii i i isite site

s s stj i j i j i

j ij i j i

n Py b n Py bn

Py b Py b

n n n T

Bb b

T

− −

= + + + = +

=

Dual Process Langmuir (DPL) Isotherm

Zeolite Coated Metal Foil preliminary fabrication coated on flat foil coupon at 30 mg/in2

coating passed Catacel adhesion test goal: to make coating 100 – 150 µm thick

50 µm

50 µm

20 µm

19

Rapid Adsorbent Characterization

commercial zeolites – activation at 350 oC

overnight in N2 – cycling at 90 oC

• 2 min adsorption in 15% CO2-N2

• 2 min desorption in 100% N2

– PT = 1 atm

BalanceBalance

Perkin-Elmer TGA-7

To ExhaustTo Exhaust

TGA Controller

Timer

Flow meters Sample plate

Balance wireGlass tube

Furnace w/thermocouple

Syringe Pump(H20 Injection)

Feed

Balance Gas

BalanceBalance

Perkin-Elmer TGA-7Perkin-Elmer TGA-7

To ExhaustTo Exhaust

TGA ControllerTGA Controller

Timer

Flow meters Sample plate

Balance wireGlass tube

Furnace w/thermocouple

Syringe Pump(H20 Injection)

Feed

Balance Gas

TGA Runs at 70 oC Cycle: 100 s Stream with CO2 /100 s Pure N2

0 1 2 3 4 5 6 7Time (min)

PowderWashcoatBead

-0.4-0.20.00.20.40.60.81.01.21.4

0 1 2 3 4 5 6 7

Rel

ativ

e L

oadi

ng (m

ol/k

g)

Time (min)

PowderWashcoatBead

15% CO2 in N2/100 % N2 100% CO2 /100 % N2

Working Capacities of washcoat between 50 and 100% higher than commercial beads!

Corrugated Catacel Cores 1” x 6” x 400 cells/in2

goal: to increase corrugation to 800 cpsi

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35

dP/d

l (KP

a/m

)

Interstitial Velocity (m/s)

One 1.0" 289 CPSI coreOne 1.0" 400 CPSI coreTwo 1.0" 289 CPSI coresTwo 1.0" 400 CPSI coresOne 1.5" 289 CPSI coreCFDErgun Eq., pelletized bed349.4 g, 4.32 mm GB in 3.934" L, 1.524" D bed

Structured and Beaded Media Pressure Drop

Pressure Drop Apparatus Qmax = 1000 SLPM

∆Pmax = 30, 70 or 140 in H20

goal: ∆Pmax < 20 kPa/m at design velocity of 20 m/s

Volumetric Frequency Response Apparatus

5x10-5 to 10 Hz 100 g adsorbent

80 oC and 0.2 atm

V3 V4 Gas Inlet

To Vacuum

V2 Pressure

Transducer

Differential Pressure

Transducer

Thermocouple

Bellows

Linear Volumetric Displacement Transducer

Water Jacket

Constant Temperature

Bath

Acentric Cam

Sample Chamber

Volumetric Frequency Response Schematic

0

5

10

15

20

25

30

35

Inte

nsity

744 Torr185 Torr102 TorrMacropore DiffusionMicropore DiffusionMacropore Convection

0

10

20

30

40

1.0E-05 1.0E-03 1.0E-01 1.0E+01

Phas

elag

Frequency, Hz

744 Torr185 Torr102 TorrMacropore DiffusionMicropore DiffusionMacropore Convection

Volumetric Frequency Response Results

Mass Transfer Models Macropore Diffusion Macropore Convection Micropore Diffusion

System and Conditions CO2-commercial zeolite beads T = 25 oC P = 102, 185, or 744 torr f = 5x10-5 to 10 Hz

Correlation of Mass Transfer Models with Experimental Data

only macropore diffusion model unequivocally fits the data over the

pressure range investigated

1-Bed Rapid PSA Apparatus 0.1 to 2 Hz

100 g adsorbent 80 oC and 0.1 atm

TV-15

PT-LP

LP

FM-HP

TV-17

HP PT-HP

PT-VAC

HP CnD

VAC

TV-16

PT-LPP

PT-LR

PT-EQ3

PT-EQ2

PT-EQ1

PT-HR

PT-HE EQ

PT-F

PT-BED

BED

FM-F

FM-HR

FM-LPP

FM-LP

FM-LR

TV-6 SV-6

TV-1 SV-1

TV-2 SV-2 TV-3 SV-3

TV-4 SV-4

TV-5 SV-5 HE EQ

HR

F

HP LR

HP CnD

TV-12 SV-12

TV-7 SV-7

TV-8 SV-8 TV-9 SV-9

TV-10 SV-10

TV-11 SV-11

LE EQ 1

LR

LP F

LP HR

LP CoD LPP

TV-13 SV-13

LE EQ 2

TV-14 SV-14

LE EQ 3

1-Bed Rapid PSA Schematic

Two-Step CO2 and N2 Cycling Experiments

PH ↑ PL

PH ↓ PL

Pure CO2 or N2 Effluent

Counter-Current Blowdown Pressurization

Pure CO2 or N2 Feed

Information Obtained a) Determination of Valve Cv

b) Determination of excluded volume

c) Validation of Single Component Isotherms

d) Validation Adsorption/Desorption Mass Transfer Coefficients

Pure Gas Cycling in 1-Bed Rapid PSA System CO2 on Beaded Zeolite at 22 oC

0

25

50

75

100

125

150

0 10 20 30

Pres

sure

[kPa

]

Time [sec]

Feed Tank

Model Bed

Vacuum Tank

ts = 3 s

0

25

50

75

100

125

150

0 5 10 15 20

Pres

sure

[kPa

]

Time [sec]

0

25

50

75

100

125

150

0 10 20 30

Pres

sure

[kPa

]

Time [sec]

0

25

50

75

100

125

150

0 2 4 6 8 10

Pres

sure

[kPa

]

Time [sec]

0

25

50

75

100

125

150

0 1 2 3 4 5

Pres

sure

[kPa

]

Time [sec]

ts = 3 s ts = 2 s

ts = 1 s ts = 0.5 s

Pure Gas Cycling in 1-Bed Rapid PSA System CO2 on Beaded Zeolite at 22 oC

kLDF = 1.87 s-1

P(Torr)

hA(kJ/K/s)

Dp/Rp2

(s-1)kLDF

(s-1)

744 1.7e-4 3.32 1.94

185 1.7e-4 3.32 0.38

102 1.7e-4 3.32 0.18

Comparison of Mass Transfer Coefficients CO2 on Beaded Zeolite

Volumetric Frequency Response Apparatus (25 oC)

1-Bed Rapid PSA System (22 oC)

Snapshot of Multi-Bed PSA System

Typical Cycle Steps for PSA Operation

PH

T

PH PH ↓

PI1

PI2 ↓

PL

PL PL ↓

PI1

PI1 ↓

PH

T CO2 Feed Gas CO2 Rich Product

CO2 Depleted Product

Feed Heavy Reflux

Equalization Counter- Current

Depressurization

Light Reflux

Equalization Light Product

Pressurization

α β

Co-Current Depressurization

PH ↓

PI2

γ

0 > α, β, γ > 1

b

c d e

0 > a, b, c, d, e > 1

PI1 ↓

PH

Feed Gas

Pressurization

a

PSA Process Conditions for DAPS* Process Conditions PH = 120 kPa PL = 5 kPa TF = 75 C h = 0.0 W/m2 K (adiabatic) tc = 120 s θ = 2,600 – 3,100 L(STP)/kg/hr Structured Bed Properties Lb = 0.125 m db = 0.09848 m ρb = 400 kg/m3

εb = 0.64

Feed Composition (Dry) yCO2 = 0.1592 yN2 = 0.8029 yO2 = 0.0379

Mass Transfer Coefficients kCO2 = 10.0 s-1

kN2 = 1.0 s-1 kO2 = 1.0 s-1 * DAPS: dynamic adsorption process simulator

DAPS Results of Bench Scale PSA Process

17.35

17.40

17.45

17.50

17.55

17.60

17.65

17.70

17.75

17.80

17.85

84

86

88

90

92

94

96

98

100

2500 2700 2900 3100 3300

Ener

gy [K

J/m

ol C

O2]

CO

2 Pur

ity &

Rec

over

y [%

]

Throughput [L(STP)/hr kg ]

Purity Recovery Energy

DOE Crite ria

this is a low energy, high

feed throughput PSA cycle for CO2 capture

that meets the DOE criteria,

especially when considering the bed density is

only 400 kg/m3

Motivation to Compare Solid Amine to Zeolite sorbents for post-combustion CO2 capture zeolites have sufficient working capacity for CO2

not H2O tolerant: must be removed prior to PSA unit! solid amine sorbents commercial amines grafted or immobilized within

large pores of a high surface area support like silica gel

have sufficient working capacity for CO2 H2O tolerant: will it pass through the PSA unit with

N2?

solid amine sorbents have NOT been studied extensively for CO2 capture from flue gas by PSA

CARiACT G10 Solid Amine Sorbent* substrate: CARiACT G10 silicon

dioxide (Fuji Silysia) surface area: 300 m2/g pore volume: 1.3 ml/g particle size: 75-150 µm

polyethylenimine (PEI) (MN 423 Aldrich)

40 wt% PEI physically adsorbed (immobilized) onto G10

*Gray et al., Energy Fuels, 23 (2009) 4840.

Snapshot of Multi-Bed PSA System

Typical Cycle Steps for PSA Operation

PH

T

PH PH ↓

PI1

PI2 ↓

PL

PL PL ↓

PI1

PI1 ↓

PH

T CO2 Feed Gas CO2 Rich Product

CO2 Depleted Product

Feed Heavy Reflux

Equalization Counter- Current

Depressurization

Light Reflux

Equalization Light Product

Pressurization

α β

Co-Current Depressurization

PH ↓

PI2

γ

0 > α, β, γ > 1

b

c d e

0 > a, b, c, d, e > 1

PI1 ↓

PH

Feed Gas

Pressurization

a

Comparison of PSA Process Performance PEI vs Zeolite

for the same process performance and conditions (much different PSA cycle), zeolite beds are 10X smaller than PEI beds with PEI consuming 2X

the energy => need amine with faster desorption kinetics

PEI Zeolitetcyc (s) 300 120

Feed Throughput (L(STP)/kg/hr) 224 2870

CO2 Recovery (%) 91.0 91.8CO2 Purity (%) 95.2 95.8

Energy (kJ/mol CO2

Recovered) 34.2 17.6

Conclusions metal foil coated with commercial zeolite and corresponding

low pressure drop corrugated structure showing much promise for CO2 capture from flue gas

frequency response and 1-bed rapid cycling experiments both show very fast mass transfer rates of CO2 in beaded zeolite

very low energy, very high feed throughput PSA cycle configuration developed using validated DAPS

novel hybrid adsorption process flow sheet resulted in < 35% COE increase

shorter PSA cycle times showing potential to significantly reduce column size and thus plant footprint

PEI solid amine sorbent showing potential in PSA process; it may allow H2O to pass through bed with N2; need better kinetics to make PSA process performance more like zeolites

Acknowledgements

Thank You!

?

Funding provided by DOE/NETL and

SAGE is greatly appreciated!