Development of cyclic adsorption process for CO capture

Separation Processes Laboratory PCCC-3

Development of cyclic adsorption process for CODevelopment of cyclic adsorption process for CO2

capture: Process modeling and optimization

Marx D., Joss L., Hefti M., Gazzani M., Mazzotti M.Institute of Process Engineering, ETH Zurich

Adsorption-based capture processesAdsorption-based capture processesPressure 30-60 barCO2 fraction 10-40%Separation from CO2 fraction 10 40%H2 fraction 30-60%

CO, O2, N2, H2O...

pressurized syngas

Adsorption-based separation:

• Very high purity of the less retained

PSAH2

CO2 +

y g p ycomponent

• Very low energy requirement• Non-volatility of the sorbent

impurities∆

ads

Pressure Swing

|| || || 8 Sep 2015Matteo Gazzani, [email protected] 2

Pdesorption Padsorption

Adsorption-based capture processesAdsorption-based capture processesPressure 30-60 barCO2 fraction 10-40%Separation from CO2 fraction 10 40%H2 fraction 30-60%

CO, O2, N2, H2O...

pressurized syngas

PSAH2

CO2 +impurities

∆ad

s

Pressure Swing


Pdesorption Padsorption

Adsorption-based capture processesAdsorption-based capture processesPressure 30-60 barCO2 fraction 10-40%Separation from

Pressure 1 barCO2 fraction 6-15%Post combustionCO2 fraction 10 40%

H2 fraction 30-60%

CO, O2, N2, H2O...

pressurized syngas

CO2 fraction 6 15%N2 fraction 85-94%

Post-combustioncapture

H2O ,O2,...

PSAH2

CO2 +TSA

CO2

N2

impuritiesCO2

Sw

ing

∆ad

s

∆ad

sTadsorption

Tem

per.

S

Pressure Swing Tdesorption


Pdesorption Padsorption Pprocess

Adsorption-based capture processesAdsorption-based capture processes

Q

Hot fluegasCO2, N2, O2, H2O120-180°C Cold fluegas

TSADrying

Qrec

N2to atmosphere TSA

H2OCO2to storage

to atmosphere

Filippi et al. U.S. Patent 7 153 344 (2002)to storage

Motivation:• Heat recovery from low grade heat sources and no need of compression y g p

energy make temperature-based swing adsorption attractive for CO2 post-combustion capture.

• Gentle separation typical of adsorption process is complemented by low


primary energy demand.

Talk o tline

Process design &

Talk outline

Equilibria Fixed bed Modeling Lab pilot Process design & Optimization

AC MOF

Mix

Rubotherm MSB

mass and energy balances

Isothermsexperimental

validationEOS

|| || ||Separation Processes Laboratory 8 Sep 2015Matteo Gazzani, [email protected] 6

Talk o tline

ModelingCharacterization Validation Process design Improvements

Talk outline

Process design &

Approach for robust process design:

Equilibria Fixed bed Modeling Lab pilot Process design & Optimization

AC MOF

Mix

Rubotherm MSB



validationEOS

Casas et al., Adsorption 2012,18, 143-161

Schell et al., Adsorption 2012, 18, 49-65


Marx et al., Adsorption 2014, 20, 493-510

Hefti et al., Micropo. Mesopor. Mat. , 215, 2015, 215-228

Material characteri ation and selection


Material characterization and selectionPure-component CO2 isotherms Z lit ZSM 5 Z lit 13 XEquilibria Zeolite ZSM-5 Zeolite 13-X

Isotherm model• Sips model

Multi-component N2-CO2 isotherms

Rubotherm MSB CO2

Binary isotherms• Extended Sips

CO2

N2N2


22

Approach for rob st process design


Approach for robust process design

Process design &Equilibria Fixed bed Modeling Lab pilot Process design & Optimization

AC MOF

Mix

Rubotherm MSB



validationEOS

Schell et al., IEC&R 2013, 52, 8311-8322

M t l IEC&R 2015 54 6035 6045


Marx et al., in preparation

Marx et al., IEC&R 2015, 54, 6035-6045

Modeling: equations


Modeling: equations

1. Mass balances species i voids icu P T

t b b L( ) 0i i i ic n uc yD c

t t z z z

adsorbent (1 ) i

in sTwT

Accumulation Convection

pi

i i in k a n nt

3. Constitutive equationsLinear driving force

Dispersion

1. Non linear adsorption isotherm:

2. Energy balances

wL i b

2T h R T T h R T T

*i( , , )

iin n p T y

2. Equation of State: ideal gas L i w w o w amb

w w

h R T T h R T Tt C a

Acc Exchange column - wall

Exchange wall - outside

3. Pressure: Ergun equation

Lt g b s b ads t g b w b L

1 i

2( ) 0n

jj

j

n hT p T TC C C uC H T T Kt t z t R z z


Heat of adsorptionAccumulation Convection Exchange column - wall

Conductivity

Modeling: equations


Modeling: equations

1. Mass balances species i voids icu P T

t b b L( ) 0i i i ic n uc yD c

t t z z z

adsorbent (1 ) i

in sTwT

pi

i i in k a n nt

3. Constitutive equationsLinear driving force

1. Non linear adsorption isotherm:

2. Energy balances

wL i b

2T h R T T h R T T

*i( , , )

iin n p T y

Mass transfer CO2 kCO2= 0.1 s-1

M t f N k 0 5 L i w w o w ambw w

h R T T h R T Tt C a Mass transfer N2 kN2= 0.5

Internal heat transfer hL= 33 WInternal heat transfer, static hL

0= 22External heat transfer, hW= 220

Lt g b s b ads t g b w b L

1 i

2( ) 0n

jj

j

n hT p T TC C C uC H T T Kt t z t R z z

te a eat t a s e , W 0


Modeling: validation



Column length 1.2 [m]Internal radius 12.5 x 10-3 [m]External radius 15.0 x 10-3 [m]Heat capacity wall 4 x 106 [J / Km]

313X material density 2359 [kg/m3]Particle density 1085 [kg/m3]Bed density 652 [kg/m3]Particle diameter 1.6-2.0 x 10-3 [m]Heat capacity sorbent 920 [J / Kkg]

Rig modifications for TSA experiments 2 jacketed columns 2 thermostats (Hüber Kältmachinenbau DE)

Main features gas piping 5 thermocouples along the column 2 mass flow controllers for feeding control

|| || ||

2 thermostats (Hüber Kältmachinenbau, DE) Automatic valves at inlet/outlet of the jacket

2 mass flow controllers for feeding control Mass spectrometer for product analysis

8 Sep 2015Matteo Gazzani, [email protected] 12




Different comparisons between pmodel and test rig results were carried out:1. Breakthrough experiments1. Breakthrough experiments

2. Heating (material regeneration)

and cooling stepsg p

3. Whole TSA cycle








3. Whole TSA cycle








3. Whole TSA cycleWaste Simplest TSA cycle for CO2 extraction: (N2) 3‐steps

1. Feeding and adsorption of CO2

CO2Dryfeed

t ime

2. Heating for material regeneration3. Cooling for column preparationFeed:


Heating CoolingAdsorptionFeed:CO2‐N2 binary (12%‐88%), fixed flow ratetads theat tcool




Different comparisons between

CO2 purity > 90%

pmodel and test rig results were carried out:1. Breakthrough experiments1. Breakthrough experiments



3. Whole TSA cycleWaste (N2)

Considered simulated cycle for validation

100125150T°C

CO2Dryfeed

t ime


Experimental results for same cycle parameters

Heating CoolingAdsorptiontads theat tcool

Approach for rob st process designApproach for robust process design

Process design &Equilibria Fixed bed Modeling Lab pilot Process design & Optimization

AC MOF

Mix

Rubotherm MSB



validationEOS

Casas et al., Sep. Pur. Technol. 2013, 112, 34-48


Joss et al., in preparation

Joss et al., IEC&R 2015, 54, 3027-3038

Process synthesis: boundary conditions



GasHot fluegasCO N O H O

Boiler

Gas cleaning(Dust,

SOx and

CO2, N2, O2, H2O120-180°C

Cold fluegasSOx and Nox)

Conventional power plant or Drying

Qrec

Cold fluegas

N2p pcombustion process TSA

Drying

H2OCO2

2to atmosphere

Compression and storage

Assumptions:• Binary CO2+N2 (12%-88%) mixture entering

the capture section*


• Energy for drying: 8 MJth/kgH2O*

* Perry’s Chemical Engineers’ Handbook




GasHot fluegasCO N O H O

Boiler

Gas cleaning(Dust,

SOx and

CO2, N2, O2, H2O120-180°C

Cold fluegasSOx and NOx

Conventional power plant or Drying

Qrec

Cold fluegas

N2p pcombustion process

Drying

H2OCO2

2to atmosphere TSA

Compression and storage

Assumptions:• Binary CO2+N2 (12%-88%) mixture entering

the capture section*


• Energy for drying: 8 MJth/kgH2O*

* Perry’s Chemical Engineers’ Handbook

Process synthesis: cycle design



Cycle AWaste (N2)(N2)

CO2Dryfeed

Heating CoolingAdsorption

t ime

• 3-steps cycle• CO2 is produced as the bed

heated up• Simplest configurationSimplest configuration• Lowest number of variables





Cycle A Cycle BWaste (N2)(N2)

CO2Dryfeed

Adsorption Heating Cooling

t ime

Rinse Purge

• 5 steps cycle• CO2 is produced as the bed

heated up• CO rich recycle enhances purity• CO2 rich recycle enhances purity

and productivity• Make use of recycled N2 to push

CO2 out


• Increased number of variables




Cycle A Cycle B Cycle CWaste (N2)(N2)

CO2Dryfeed

Heat 2 CoolingAdsorption

t ime

Heat 1

• 4 steps cycle• First heating step to collect

impure CO2, which is then recycled to the feedrecycled to the feed

• Second heating step to produce the pure CO2product


Process synthesis: heuristic comparison


Process synthesis: heuristic comparison

Cycle A Cycle B Cycle CWaste (N2)

Waste (N2)

Waste (N2)

CO2DryfeedCO2

Dryfeed

CO2Dryfeed

t ime feed


t ime

Heat 1

feed


t ime

Rinse PurgeHeating CoolingAdsorption

t ime

• Lowest separation f

Rinse-purge combination: Dedicate heating time for performance • Purity and recovery

enhancement Additional step is fast:

I d ifi

evacuation:• Best purity among A, B and

CI d• Improved specific

productivity Compared to Cycle A, we should obtain an

• Improved recovery compared to A

Additional step is slow:• Reduced productivity and


should obtain an improvement in all indexes

• Reduced productivity and higher energy consumption

Process synthesis: separation performance



Recovery-Purity pareto curves can be established by varying the cycle timesy





theat tcool

Recovery-Purity pareto curves can be established by varying the cycle timesheat cool

Cycle A: 3-steps cycle:tads, theat and tcool

y

Waste (N2)

tads CO2Dryfeed

Heating CoolingAdsorption

t ime





Recovery-Purity pareto curves can be established by varying the cycle timestheat tcool

Cycle B: 5-steps cycletads, theat, tcool, tpurge= trinse

y

tpurge

Waste (N2)

tads CO2Dryfeed


t ime

Rinse Purge





Recovery-Purity pareto curves can be established by varying the cycle timestheat2 tcool

(R,P) = (0.8, 0.8)Cycle C: 4-steps cycletads, theat1, theat2, tcool

y

th t1Waste (N2)

theat1

CO2Dryfeed


t ime

Heat 1

tads

Cycle C outperforms A and B in term of separation performance


Process synthesis: energy and productivity



(R,P) = (0.8, 0.8)(R,P) (0.8, 0.8)

(R,P) = (0.8, 0.8)

Energy consumption-Productivity pareto curves are also established when varying cycle times.Simulations with (R,P)<(0.8,0.8) are not considered.





(R,P) = (0.8, 0.8)(R,P) (0.8, 0.8)

(R,P) = (0.8, 0.8)

Energy consumption-Productivity pareto curves are also established when varying cycle times.





(R,P) = (0.8, 0.8)(R,P) (0.8, 0.8)

(R,P) = (0.8, 0.8)

Energy consumption-Productivity pareto curves are also established when varying cycle times.





(R,P) = (0.8, 0.8)(R,P) (0.8, 0.8)

(R,P) = (0.8, 0.8)

Energy consumption-Productivity pareto curves are also established

Cycle B outperforms A and C in

when varying cycle times.

yterm of energy and productivity for a given separation target


Process design for CO capture


Process design for CO2 capture

(R,P) = (0.9, maxP=0.93)

PR Heuristic optim. Z13-X

Amines


Process improvement


Process improvementMaterials• Synthesis & engineering

Process design• Modeling

Process optimization• Mixed Integer Non Linear • Synthesis & engineering

• Working capacity• Stability to H2O• Formulation

• Modeling• Cycle configuration• Scheduling• Column/module design

• Mixed Integer Non‐Linear• Many degrees of freedom• Multi‐objective• Control • Formulation

• Characterization• Column/module design • Energy integration

• Control• Economics


Process improvement


Process improvementMaterials Process designProcess optimization

Challenges:M lti bj ti ti i ti f t fli ti bj ti ( d ti it )• Multi-objective optimization of two conflicting objectives (energy-productivity)

• Noisy and non-smooth objective function because of the PDE solver• Computational time of each function evaluation is considerable• Purity and recovery are non linear constraints• Purity and recovery are non-linear constraints

Implemented algorithm:• Model-based derivative-free algorithm:• Model-based, derivative-free algorithm:

revised version of the Multilevel Coordinate Search (MCS) algorithm


Process improvement



New feasible points are obtained:t f t t d hi h• pareto front moves towards higher

productivity.• Minimum energy consumption is

slightly affected

Heuristic optim. Z13-X

slightly affected.

Amines


Process improvement



Mg2(dobdc) (Mg-MOF-47): high CO2 capacity, high g 2 p y gselectivity, potential for TSA

Amines

Xiang et al. Nature Comm. 3:954 (2012)


Process improvement



Simulated counter‐current heat exchanger:

23 15 4

Low grade heat streamT i

Tout

Amines

heat streamT in

T5 > T4 > T3 > T2 > T1

Column 1 AdsorptionHeating CoolingColumn 1Column 2Column 3Column 4Column 5Column 6Column 7

AdsorptionHeating Cooling

Column 7Column 8Column 9Column 10

t cycle0


ConclusionsConclusions

Equilibria Fixed bed Modeling Lab pilot Process design & OptimizationOptimization

AC MOF

Mix

mass and

Rubotherm MSB


IsothermsEOS

experimentalvalidation




Schell et al., IEC&R 2013, 52, 8311-8322




Joss et al., IEC&R 2015, 54, 3027-3038

Marx et al., IEC&R 2015, 54, 6035-6045

ConclusionsConclusions

Equilibria Fixed bed Modeling Lab pilot Process design & OptimizationOptimization

AC MOF• A good commercial material for CO2 capture with a TSA process is already available

Mixprocess is already available.

• Separation performance is very promising; cycles can be tuned according to requirements.

mass and

• Careful cycle synthesis and selection are needed to obtain the best trade-off between energy consumption and productivity.

• Optimization new materials and heat integration make theRubotherm MSB


IsothermsEOS

experimentalvalidation

• Optimization, new materials and heat integration make the TSA a promising solution for 2nd generation CO2 post-combustion capture.




Schell et al., IEC&R 2013, 52, 8311-8322




Joss et al., IEC&R 2015, 54, 3027-3038

Marx et al., IEC&R 2015, 54, 6035-6045

��

� �� !"��#$� ��%� �&&�� &&��$��' ��(��)�&&� � ��

Backup slidesBackup slides

|| || ||

Breakthrough experiments on 13XBreakthrough experiments on 13X

Concentration and temperature profiles for the breakthrough experiment with an initial temperature of 25°C, a feed composition of yCO2 = 0.12, and a feed flow rate of 300 cm3/s.

|| || ||

Heating and coolingHeating and cooling

100

Heating

T

Cooling

feed = 12/88 CO2/N2

C] 80

100 Tset

x

feed 12/88 CO2/N2

pera

ture

[°C

60

x

Tem

p

40

x

h i i [ ]0 200 400 600 800

20

li i [ ]0 200 400 600 800

Tset

The full length of the steps was 1800 s for the heating step and 1200 s for the cooling step.Along with the temperatures measured in the bed (symbols) and the corresponding simulation

heating time [s] cooling time [s]

|| || ||

results (lines), the measured temperature of the heat exchange fluid at the jacket inlet and outlet is shown using dots

8 Sep 2015Matteo Gazzani, [email protected] 43

Pure component equilibria


Pure component equilibriaZeolite ZSM-5 Zeolite 13-X

O2

P/T conditions

• Pressure:

CO up to 10bar

• Temperature range:25 - 140°C

N2

N




Pure component equilibriaZeolite ZSM-5 Zeolite 13-X

P/T conditions

• Pressure:

O2

up to 10bar

• Temperature range:

CO

25 - 140°C

Isotherm modelSi d l• Sips model

N2

N




Pure component equilibriaZeolite 13-XZeolite ZSM-5

P/T conditions

• Pressure:C

1-10 bar


25°C

25 and 45°C

5°C

45




Pure component equilibriaZeolite 13-XZeolite ZSM-5

P/T conditions

• Pressure:C

1-10 bar


25°C

25 and 45°C

Binary isotherms

• Extended Sips

5°C

45


Process improvement



Simulated counter‐current heat exchanger:Trade‐off between CAPEX and OPEX

23 15 4

Low grade heat streamT i

Tout

heat streamT in

T5 > T4 > T3 > T2 > T1

Column 1 AdsorptionHeating CoolingColumn 1Column 2Column 3Column 4Column 5Column 6Column 7

AdsorptionHeating Cooling

Nsteps = 5

Column 7Column 8Column 9Column 10

t cycle0


Comparison with existing plants


Comparison with existing plantsMaterials Process designProcess optimization

TSA productivity is converted to volume based• Factor 2 assumed when

moving from tonnezeolite to m3

AminesBoundary Dam productivity:approximately 13 kgCO2/(m3h) using:using:

CO2 capacity: 3000 t/dayCO absorber 11 x 11 x 54 mCO2 absorber 11 x 11 x 54 mCO2 desorber 8 x 43 m


Documents

Development of cyclic adsorption process for CO capture