Adsorption modelling as

a tool to estimate
transport properties

Moises Bastos-Neto
Universidade Federal do Ceará

Leipziger Symposium on dynamic sorption 2019 May 14th 2019

Advanced Sorbent Materials
on the Way to Application
 Outline

Introduction – Concepts
 Aims and basics
 Adsorption
 Adsorbents
 Column dynamics
Modelling – Theoretical Background
 Definitions and terminology
 Momentum, Material and Energy Balances
 Equilibrium theory
 Adsorption kinetics
Assessing mass transfer
 Simple fit to breakthrough curves
 From uptake curves
 From calorimetry
Final remarks
 Outline

Introduction – Concepts
 Aims and basics
 Adsorption
 Adsorbents
 Column dynamics
Modelling – Theoretical Background
 Definitions and terminology
 Momentum, Material and Energy Balances
 Equilibrium theory
 Adsorption kinetics
Assessing mass transfer
 Simple fit to breakthrough curves
 From uptake curves
 From calorimetry
Final remarks
 Column dynamics
Designing a process

H2
N2

CO

CH4

CO2

1.0
relative concentration

0.8

H2/N2/CO/ desorption 0.6

CH4/CO2
0.4
feed gas Tailgas

0.2

0.0
0 200 400 600 800 1000
time / s
What is your aim?

Synthesize better sorbents?

•Material
•Shape
•Properties

Design process units?

• Size
• Material
• Control

Optimize operations?
• Costs
• Maintenance
• Facilities

Understand the phenomena?

• Effects
• Simplifications
• Improvements
 Necessary knowledge

• Which column size?

• Which flowrate?
• Is it reversible?
• How long do cycles last?
• Should columns be thermostated?
• Which operating conditions maximize purity, recovery
from the feed, and minimize energy /solvent
consumption?

etc
Basics

Conservation equations
(mass, energy, momentum, electric charge)

Equilibrium laws at the interface(s)

Constitutive laws

Kinetic laws of heat/mass transfer and reaction

Initial and boundary conditions

Optimization criterion
 Classification of systems
• Nature of equilibrium relationship
• Linear isotherm
• Favorable isotherm
• Unfavorable isotherm
• Thermal effects
• Isothermal
• Near isothermal
• Concentration level
• Trace systems
• Nontrace systems
• Flow model
• Plug flow
• Dispersed flow
• Complexity of kinetic model
• Negligible transfer resistance
• Single transfer resistance
• Multiple transfer resistance
 Adsorbents
• Types
• Structures
• Homogeneous
• Porous
• Bidisperse
• Properties
• Adsorption capacity
• Selectivity
• Kinetics
• Stability
• Mechanical
• Thermal
• Chemical
 Outline

Introduction – Concepts
 Aims and basics
 Adsorption
 Adsorbents
 Column dynamics
Modelling – Theoretical Background
 Definitions and terminology
 Momentum, Material and Energy Balances
 Equilibrium theory
 Adsorption kinetics
Assessing mass transfer
 Simple fit to breakthrough curves
 From uptake curves
 From calorimetry
Final remarks
An accurate process simulator is an important
tool for learning, designing and optimization
purposes.

Alírio E. Rodrigues
 Definitions and terminology
• Concentration profiles – Ci(z) at a given t

• Concentration histories – Ci(t) at a given z

Alírio E. Rodrigues (2014)

 Definitions and terminology
• Overall balance

Alírio E. Rodrigues (2014)

 Definitions and terminology
• Concentration profile at t = tbt

Alírio E. Rodrigues (2014)

 Adsorption-based processes

In general, one is interested in re-using the adsorbent for a relatively

large numbers of cycles. Industrial sep processes alternate two
steps:
 Adsorption: fluid phase is enriched with the weakly adsorbed
species (raffinate)

 Desorption: fluid phase is enriched with the strongly adsorbed

components (extract) and the adsorbent is regenerated to be
used in another cycle (by temperature, pressure, pH or
concentration swings)

Adsorption Desorption

A+B B (+A) A (+B)

A (+B)
Breakthrough of mixtures
Modelling adsorption processes

Equilibrium

Purity

Feed Kinetics

Simulation
Model Cycle time

Co-
Heat data
adsorption

Adsorber
dimensions

Hydrodynamics
Modelling adsorption processes
 Modelling a fixed bed

Transport Phenomena
To model the dynamic behavior of an adsorption
column is a problem far from trivial.
 Momentum balance

Pressure drop in packed beds:

150  g 1   1.751    g

2
p
   2
z  3 d p2  3 dp

Blake-Kozeny equation Burke-Plummer equation

Laminar Flow Turbulent Flow

 Material balance

infinitesimal cross sectional cut of the adsorbent column


Continuity – General Form: CT z, t     Fz, t   s z, t   0
t

m m
Considering the interparticle volume: CT      Cg
VT V

CT C g
Flux – Accounting convective and dispersive effects: F   CT  Dax or F    C g   Dax
z z

Q
Material “removal” rate: s   where Q is the amount of the species leaving the bulk phase in the control volume
t

mP  m A
Concentration inside the particle: Q 
VT

Defining the adsorbent (particle) volume: V A  VT 1    and the pore volume: VP  V A  P

m m   m m 
Then: Q  1      P  A   1      P P  A 
 VA VA   VP VA 
 Material balance
mS
Defining the particle density:  s 
VA

 m 

Thus: Q  1      P C g   s A   1      P C g   s q 
 mS 

Q
Deriving to obtain s  
t


We get: s 1      P Cg   s q 
t

1
q t    qr , t   dV
VA
Where q is defined as the average specific amount adsorbed:
VA 0

C
Substituting equations:

 Cg        Cg   Dax g    1      P Cg   s q   0
t  z  t

Three reasonable assumptions are very often made:

(i) bed porosity is homogeneous and constant along the bed
(ii) particle porosity is the same for every adsorbent particle and
(iii) the gas flows in only one dimension – axially
C g  C g   2C g  C g q 
    Dax  1       P  s 0
t z z 2
 t t 
 Energy Balance

infinitesimal cross sectional cut of the adsorbent column

Analogously to the Continuity:


E z, t     Fz, t   s z, t   0
t

Volumetric sensible heat in the control volume:

m
E cT C cT
VT

Temperature changes in the given control volume is represented by the temperature changes in the gas and in the
solid phases
mg ~
For the gas phase: E g  c g Tg   C g c~g Tg
VT

For the solid: E s  ms cˆ s Ts  1     s cˆ s Ts

VT
 Energy Balance

Summing up and differentiating:

E Tg Ts
 E g  E s   C g c~g  1     s cˆ s
t t t

Considering identical temperature profiles for the fluid and

solid phase in the adsorbent column operating at cyclic
steady state:
T g  Ts

E Tg
Then:   C g c~g  1    s cˆ s 
t t

Heat is transported through the adsorbent bed along with the fluid flow and dispersed analogously to the mass.
The dispersion term can be simplified and evaluated by applying Fourier’s method of separation of variables. Thus,
the energy flux can be written as:
Tg
F  E conv  E disp    C g c~g Tg   
z

Applying the same assumptions as before:

(i) bed porosity is homogeneous and constant along the bed
(ii) particle porosity is the same for every adsorbent particle and   Fz, t    c~ C T    T C g   C Tg     Tg
2

g g 
z z z  z 2
g g g
(iii) the gas flows in only one dimension – axially 
 Energy Balance
Heat is generated in the system through adsorption and removed by conduction through the walls and later by
convection with the environment.

 q  4 hw
s  1    s   H    Tg  Tw 
 t  di

Substituting and arranging:

Tg   Tg   2Tg ~  C T    C Tg   4 hw T  T 

 C c~g  1     s cˆ s    c~g C g Tg   Cg     c g g
t z z  z 
g
z 2 z
g g g w
  di
or
 2Tg   q  4 hw  Tg q
   1    s   H    Tg  Tw   0   C g c~g  1     s cˆ s   1    s  H  0
z 2  t  di  t t
  

Additional equation - heat transfer from the wall to the environment:

Tw
 w cˆ w   w hw Tg  Tw    wL U g Tw  T 
t
 Summarizing

Material Balance
C g ,i

1      qi
 Dax
 2 C g ,i
 u
C g ,i
0
t  t z z
b 2

accumulation dispersion convection

Energy Balances
1     H  q    2Tg   1     c  
 Tg
 
Tg 4hw
 Tg  Tw   0
 g PG 
c u c
 t z 2   t z
b b PS g PG
 di
generation dispersion accumulation convection transfer to wall

Tw
 whw Tg  Tw    wC pw   wLU g Tw  T   0
t
gas to wall wall to environment A.M. Ribeiro, Chem. Eng. Sci. 63 (2008)
accumulation D.M. Ruthven, Principles of Adsorption (1984)
D. Bathen and M. Breitbach, Adsorptionstechnik (2001)
 Assumptions and Simplifications

• No gradients in the radial direction?

• Plug flow with axial mass dispersion?
• Mass transfer into the particle in accordance to the linear driving force
(LDF) model?
• Thermal equilibrium between the gas and the adsorbent?
• Adiabatic operation?
• Constant heat transfer coefficients?
• Homogeneous porosity along the bed?
• No pressure drop?

A.M. Ribeiro, Chem. Eng. Sci. 63 (2008)

D. Bathen and M. Breitbach, Adsorptionstechnik (2001)
 Equilibrium law
At interfaces:  
qi*  f C g ,i

favorable unfavorable linear

d 2 qi* d 2 qi* d 2 qi*

0 0 0
dC g2,i dC g2,i dC g2,i

rectangular with an inflection

D.M. Ruthven, Principles of Adsorption (1984)

W. Kast, Adsorption aus der Gasphase (1987)
 Influence of the equilibrium
Primary influence

The shape and nature of the breakthrough curve are strongly influenced by the
equilibrium

1.0 1.0

0.8 0.8
relative concentration

relative concentration
0.6 0.6

0.4 0.4
Max. Adsorption Capacity
Adsorption Affinity
qmax0 K0
0.2 qmax0 + 20% 0.2 K0 + 20%
qmax0 - 20% K0 - 20%

0.0 0.0
0 20 40 60 80 100 0 20 40 60 80 100
time / min time / min

Kp
q   q max
1 K  p
 Influence of the equilibrium
Increasing the partial pressure
Flow [Norm] P y Pi Partial Flow
mL/min bar % bar mL/min
100 10 10 1.0 10
150 15 10 1.5 10
200 20 10 2.0 10
300 30 10 3.0 10

1.0 3.0
10 bar
15 bar Methane
20 bar AC CarboTech D 55/2 PSA
0.8 30 bar 2.5

-1
amount adsorbed / mmol.g
relative concentration

2.0
0.6

1.5
0.4
1.0

0.2 10°C
methane 0.5
AC CarbTech D 55/2 PSA 20°C
300 K 40°C
0.0 0.0
0 20 40 60 80 100 120 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
time / min pressure / bar

Bastos-Neto et al., Chem. Ing. Tech., 83 (2011)

 Equilibrium theory
“Simple is beautiful (and useful)”
 Isothermal operation
Equilibrium reached instantaneously in each point of the bed: qi  qi
*

 Plug flow
 Negligible pressure drop
 Negligible dispersion and mass transfer effects

the material balance

C g ,i  ui C g ,i   2 C g ,i  C g ,i q 
    Dax  1       P  s   0
t z z 2  t t 

becomes

C g ,i C g ,i  1    qi
*
ui    0
z t    t
 Equilibrium theory

considering q *  f (C g ,i )

C g ,i 1   C g ,i
then ui  1  f ' (C g ,i )  0
z    t

C g ,i
z t
since  z
t c
C g ,i
z t

z ui
uc  
it results in t 1  De Vault’s Equation
c 1 f ' (C g ,i )

Adsorption as a wave phenomenon
The velocity of propagation of a concentration C, i.e. uc, is inversely proportional to the local slope of the isotherm f’(C)

A.E. Rodrigues and D. Tondeur, Percolation Processes: Theory and Applications (1981)
D. DeVault, J. Am. Chem. Soc. 65 , 532 (1943)
 Equilibrium theory

unfavorable isotherms

Concentration profiles Ci (z) at a given t

Dispersive Front

As Ci the local slope of the isotherm f’(Ci) and uc

Higher concentrations travel at lower velocities

 Equilibrium theory

favorable isotherms: “shock wave”

Concentration profiles Ci (z) at a given t

Compressive (shock) Front

As Ci the local slope of the isotherm f’(Ci) and uc

Higher concentrations travel at higher velocities

 Equilibrium theory

favorable isotherms: “shock wave”

z ui
u sh  
t 1   qi
c 1
 C g ,i

“ideal” situation
real situation
Lc
u sh 
t st

Vc  1   qi 
t st  1  
Q   C g ,i 
 Influence of the equilibrium

favorable unfavorable linear

d 2 qi* d 2 qi* d 2 qi*

0 0 0
dC g2,i dC g2,i dC g2,i

Ci Ci
 If non-idealities are present
Axial dispersion
1.0
Axial Dispersion Coefficient Dax,2
Dax,1
0.8
relative concentration

Dax,3
0.6 Dax,1 < Dax,2< Dax,3

0.4

0.2

0.0
0 20 40 60 80 100
time / min
 If non-idealities are present
Mass transfer resistances
1.0

0.8
relative concentration

0.6

0.4
Mass Transfer Coefficient
kLDF
0.2 kLDF x 0.2
kLDF x 5.0

0.0
0 20 40 60 80 100
time / min
 If non-idealities are present
Axial dispersion and mass transfer resistances are to be considered

For favorable isotherms, the

concentration front disperses up
to a certain extent and assumes a
CONSTANT PATTERN BEHAVIOR

For unfavorable and linear

isotherms, the concentration front
disperses continuously as it moves
along the bed and hence follows a
PROPORTIONATE PATTERN
BEHAVIOR
What about kinetics?
How to assess the phenomenon?
How relevant can it be to the process?
 Intraparticle kinetics
Remembering the material balance q ' q
s 
t t
C g   C g   2C g  C g q ' 
    Dax  1       P  s   0 q'  adsorbed amount
t z z 2  t t  adsorbent mass

q   adsorbed amount
adsorbent volume

Homogeneous q 1   2 q 
 Dh  2 r 
particle t r r  r 

Porous C p q 1   2 q 
particle    D p ,e  2 r 
r t r r  r 
 Intraparticle kinetics
Adsorbed amount as a function of radius and time
q  f r , t 

q t   q r , t   dV
1 VA q
Averaging:
VA 
0

q
  Rp 0 Rp
Linear Driving Force (LDF) approach:  k LDF  q *  q
t

E. Glueckauf, Trans. Far. Soc., 51(11), (1955)

 LDF model

The LDF model works in practice!

Adsorptive separation process models require several sets of

averaging of local kinetic properties, which are often lost during a
series of integration processes.

The overall adsorption kinetics for a heterogeneous adsorbent can

be described by a heterogeneous-LDF model, even though the
kinetics in each adsorption site is Fickian.

S. Sircar, Ind. Eng. Chem. Res., 45 (16), (2006)

 Outline

Introduction – Concepts
 Aims and basics
 Adsorption
 Adsorbents
 Column dynamics
Modelling – Theoretical Background
 Definitions and terminology
 Momentum, Material and Energy Balances
 Equilibrium theory
 Adsorption kinetics
Assessing mass transfer
 Simple fit to breakthrough curves
 From uptake curves
 From calorimetry
Final remarks
 Estimating kLDF

“Simple Fit” - Experiment vs Simulation

C g   C g   2C g  C g q ' 
    Dax  1       P  s   0
t z z 2  t t 
1.0
0.1 MPa
relative concentration (C/C0)

0.8 1 MPa
0.5 MPa
2 MPa

q
0.6 4 MPa
t

 k LDF  q *  q 
0.4

KÖSTROLITH 5ABF
0.2 nitrogen
model
0.0
0.0 1.5 3.0 4.5 6.0 7.5 9.0
3
time / 10 s
 Estimating kLDF

From uptake curves

Convenience: during the measurement of equilibrium

isotherms

• Continuous measurement of mass variation for each pressure step

• Mass and energy balances are used to estimate the mass transfer coefficient

• Relatively simple, but reliable

• Restricted to “non-instantaneous” adsorption

• LDF approach
 Estimating kLDF

From uptake curves

Equilibrium model
qm ,i (bi P )1/ ni
qi* 
1  (bi P )1/ ni

Adsorption kinetics
q  D
t

 k LDF  q *  q  kLDF 
r 2

Energy Balance

T q
ms cˆ p , s  ms (H )  have A (T  T )
t t

Clausis-Clapeyron Natural convection coefficient

H  28 kJ mol 1 h  70 W m ² K
 Estimating kLDF

From uptake curves

film and macropore resistances are negligible

q
 k LDF
t
 Estimating kLDF

From uptake curves

CO2 and N2 uptake curves

Estimated coefficients correspond to the minimum value fitting to experiments

R.M. Siqueira et al., Chem. Eng. Technol. 41(8), (2018)

 Estimating kLDF

From uptake curves

Breakthrough curves experiments were measured to validate the simulation model
using the kLDF values estimated from gravimetric experiments.

Model parameters used for fixed bed simulations

Adsorbent mass [kg] Adsorbent mass [kg] 0.155

Bed density 𝜌𝐿 [kg m-3] Bed density 𝜌𝐿 [kg m-3] mads Vcol
Bed porosity 𝜀𝐿 [-] Bed porosity 𝜀𝐿 [-]

1  Vˆp  Vˆs   L  1

Particle density 𝜌𝑝 [kg m-3] Particle density 𝜌𝑝 [kg m-3]  L 1   L 

Particle porosity 𝜀𝑝 [-] Particle porosity 𝜀𝑝 [-] 1  Vˆs
 Vˆ  Vˆ 
s po 1

Heat transfer coefficient [W m² Heat

K] transfer coefficient [W m² K] 100
-1 -1 -1 -1
] specific heat 𝑐𝑝,𝑠 [J kg K ]
Solid specific heat 𝑐𝑝,𝑠 [J kg KSolid 820
-1
] specific heat 𝑐𝑝,𝑤 [J kg-1 K-1]
Wall specific heat 𝑐𝑝,𝑤 [J kg-1 KWall 477
Wall density 𝜌𝑤 [kg m-3] Wall density 𝜌𝑤 [kg m-3] 786
2ui rp 1 0.7  2
s-1] a) mass dispersion 𝐷𝑎𝑥 [m² s-1] a) Dax 
Axial mass dispersion 𝐷𝑎𝑥 [m² Axial ;   0.5Dax 
Pe Pe Re Sc P
ax ax
Axial heat dispersion  𝑎𝑥 [W m-1
] dispersion  𝑎𝑥 [W m-1 K-1]
K-1heat
Axial  (7  0.5 Pr Re)
kg kg

1 1
 Estimating kLDF

From uptake curves

Simulated results using the estimated coefficients showed good agreement with the
experimental breakthrough data.

R.M. Siqueira et al., Chem. Eng. Technol. 41(8), (2018)

 Estimating kLDF

From uptake curves

Different coefficient values were used to evaluate the influence of the kLDF on the
relative concentration curve shape and the temperature history.

R.M. Siqueira et al., Chem. Eng. Technol. 41(8), (2018)

 Estimating kLDF

From uptake curves

Temperature histories are important to cross-check the method to estimate kLDF

R.M. Siqueira et al., Chem. Eng. Technol. 41(8), (2018)

Using adsorption-related heat effects and heat
transport to estimate mass transfer
 Estimating kLDF
Tian-Calvet microcalorimeter

𝑑𝑄𝑟𝑒𝑣 𝑑𝑝
Δℎ𝑎𝑑𝑠,𝑇,𝑛 = + 𝑉𝑑
𝑑𝑛𝜎 𝑇,𝐴
𝑑𝑛𝜎 𝑇,𝐴
 Estimating kLDF
Discontinuous procedure:

 dQrev   dp 
    Vc     hadsT , n
 dn T , A  dn T , A
Vcalorimetric cell
nadsorbed
 Estimating kLDF
Heat peaks

8
ሶ
𝑄𝑎𝑑𝑠 = ෍ න 𝑄𝑑𝑡
𝑖=1

ሶ
𝑄𝑑𝑒𝑠𝑜𝑟 = න 𝑄𝑑𝑡

𝐸𝑟𝑒𝑔𝑇 = 𝑄𝑎𝑑𝑠 − 𝑄𝑑𝑒𝑠𝑜𝑟

D.A.S. Maia et al., Chem. Eng. Res. Des. 136 (2018)

 Estimating kLDF
Modelling – Defining the system
𝑛ሶ 𝑠 𝑛ሶ 𝑒

Vv, Cv
Vd, Td, P

dosing cell Pv Tv
dead T
volume
P
3 defined parts (volumes):
• Dosing cell
calorimetric cell with
• Dead volume
temperature control
• Calorimetric cell

Mass and Energy balances for each part

 Estimating kLDF
Modelling – Assumptions

• Ideal gas behavior

• The dosing cell and the dead volume are under isothermal and
non-adiabatic operation
• The pressure in the dead volume is the same of the calorimetric
cell
• Two approaches for mass transfer: Linear Driving Force e
Diffusion in a spherical particle
 Estimating kLDF
Dosing cell

𝑛ሶ 𝑠

𝑑𝐶𝑣 𝑛ሶ 𝑠 Vv, Cv
=−
𝑑𝑡 𝑉𝑣

Pv Tv
𝑑𝑇𝑣 ℎ𝑣 𝐴𝑣
𝐶𝑣 𝑐𝑝 = 𝑇01 − 𝑇𝑣
𝑑𝑡 𝑉𝑣

𝑃𝑣 = 𝐶𝑣 𝑅𝑇𝑣 (EOS)
 Estimating kLDF
Dead volume
𝑛ሶ 𝑠 𝑛ሶ 𝑒

𝑑𝐶𝑑 Vd, Td, P

𝑉 = 𝑛ሶ 𝑠 − 𝑛ሶ 𝑒
𝑑𝑡 𝑑 dead volume

𝑑𝑇𝑑 𝑑𝑃
𝑉𝑑 𝑐𝑝 𝐶𝑑 − 𝑉𝑑 = 𝑛ሶ 𝑠 𝑐𝑝 𝑇𝑣 − 𝑇𝑑 − ℎ𝑐 𝐴𝑑 𝑇𝑑 − 𝑇01
𝑑𝑡 𝑑𝑡

𝑃𝑑 = 𝐶𝑑 𝑅𝑇𝑑 (EOS)

𝑃𝑑 = 𝑃
 Estimating kLDF
Calorimetric cell calorimetric cell with temperature control
𝑛ሶ 𝑒
𝑑𝐶𝑐 𝑑𝑞ത
𝑉𝑐 = 𝑛ሶ 𝑒 − 𝑚𝑠
𝑑𝑡 𝑑𝑡

T
𝑑𝑇 𝑑𝑃 𝑑𝑇 𝑑 𝑞ത 𝑑𝑇
𝑉𝑐 𝑐𝑝 𝐶 − 𝑉𝑐 + 𝑚𝑠 𝑐𝑝𝑠 + 𝑚𝑠 −∆𝐻 + 𝑚𝑠 𝑐𝑝 𝑞ത P
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
= 𝑛ሶ 𝑒 𝑐𝑝 𝑇𝑑 − 𝑇 − ℎ𝑐 𝐴𝑐 𝑇 − 𝑇𝑐 (𝑅1)

𝑃 = 𝐶𝑅𝑇 (EOS)

𝑃 = 𝑓(𝑡) Measured – Needed for the solution

 Estimating kLDF
Calorimetric cell wall – Energy Balance

R1 < R < R2:

𝑑𝑇𝑐 1 𝜕 𝜕𝑇𝑐
𝜌𝑐1 𝑐𝑐2 = 𝐾𝑐1 𝑅
𝑑𝑡 𝑅 𝜕𝑅 𝜕𝑅

R2 <R < R3:

𝑑𝑇𝑐 1 𝜕 𝜕𝑇𝑐
𝜌𝑐2 𝑐𝑐2 = 𝐾𝑐2 𝑅
𝑑𝑡 𝑅 𝜕𝑅 𝜕𝑅

Boundary conditions:
𝑑𝑇𝑐
ℎ𝑐 𝑇 − 𝑇𝑐 (𝑡, 𝑅1) = −𝐾𝑐1 (𝑡, 𝑅1) 𝑇𝑐 𝑡, 𝑅3 = 𝑇02
𝑑𝑡

Initial condition: 𝑇𝑐 0, 𝑅 = 𝑇02

 Estimating kLDF
Two approaches
1. Linear Driving Force

𝑑 𝑞ത
= 𝑘𝐿𝐷𝐹 𝑞∗ − 𝑞ത
𝑑𝑡

2. Diffusion

𝑑𝑞𝑝 𝜕𝑞𝑝 2 𝜕𝑞𝑟

= 𝐷𝑐 +
𝑑𝑡 𝜕𝑟 𝑟 𝜕𝑟

Boundary conditions:
𝜕𝑞𝑝
𝑡, 0 = 0 𝑞𝑝 𝑡, 𝑟𝑝 = 𝑞𝐸 (𝑃)𝜌𝑝
𝜕𝑟

Initial condition: 𝑞𝑝 0, 𝑟 = 𝑞𝐸 (𝑃𝑖 )𝜌𝑝

 Estimating kLDF
Heat flux and total heat

The heat flux out of the cell is given by

𝜕𝑇𝑐
𝑄1 = −𝐾𝑐1 𝐴𝑐0 𝑡, 𝑅2
𝜕𝑅

The total heat is calculated as follows

∞
𝑄𝑡𝑜𝑡𝑎𝑙 = න 𝑄1 𝑑𝑡
0
 Estimating kLDF
Experimental procedure

• Heat of adsorption is determined prior to each run according to the

equation for the total heat
𝑫𝒄
• The kinetic parameters are then fitted 𝒌𝑳𝑫𝑭 or 𝑹𝟐𝒄

Resistance transitions
Relationship between and the mass transfer resistances (film, macro
and micropores)

1 𝑅𝑝 𝑞0 𝑞0 𝑅𝑝2 𝑅𝑐2
= + +
𝑘𝐿𝐷𝐹.𝑖 3𝑘𝑓,𝑖 𝐶0 15𝜀𝑝 𝐷𝑝,𝑖 𝐶0 15𝐷𝑐,𝑖
 Estimating kLDF
AC Norit RB4 – CO2 adsorption 1.6

1.4
1.8 -1
1.6
1.2 kLDF =0.13 s

heat rate (mWatt)

adsorbed amount (mmol/g)

1.4 1.0

1.2 0.8
1.0 0.6
0.8 experimental
0.4
0.6 simulation
0.2
0.4
experimental 0.0
0.2 Langmuir fit 0 500 1000 1500 2000 2500 3000
0.0 time (s)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
pressure (bar) 1.0
0.9
34
0.8
enthalpy of adsorption (kJ/mol)

2 -1
32
Dc/R = 0.0084 s

heat rate (mWatt)

0.7
30
0.6
28
0.5
26
24 0.4

22 0.3
experimental
20 0.2
simulation
18 Peak 3 0.1
16 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 0 500 1000 1500 2000 2500 3000
adsorbed amount (mmol/g) time (s)
 Estimating kLDF
AC Norit RB4 – CO2 adsorption 1.6

1.4
1.8 -1
1.2
kLDF =0.13 s
1.6

heat rate (mWatt)

adsorbed amount (mmol/g)

1.4 1.0
1.2 0.8
1.0
0.6
0.8
experimental
0.4
0.6 simulation
0.4 0.2
experimental
0.2 Langmuir Fit 0.0
0 500 1000 1500 2000 2500 3000
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 time (s)
pressure (bar) 1.6

1.4 2 -1
34 Dc/R = 0.009 s
1.2
enthalpy of adsorption (kJ/mol)

heat rate (mWatt)

32
30 1.0
28
0.8
26
24 0.6
22 0.4 experimental
20 simulation
0.2
18
16 Peak 5 0.0
0 500 1000 1500 2000 2500 3000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
adsorbed amount (mmol/g)
time (s)
 Estimating kLDF
AC Norit RB4 – Comparing with the uptake measurements

Sample Method kLDF (s-1)

Uptake 0.1
AC Norit RB4
Calorimetry 0.13
R.M. Siqueira et al., Chem. Eng. Technol. 41(8), (2018)
 Estimating kLDF
AC Norit RB4 – Sensibility of the method
1.1

1.0

0.9

0.8

0.7
 DcR2 = 0,0041 1/s
0.6
 DcR2 = 0,0041/5 1/s
 DcR2 = 0,0041/10 1/s
KLDF.Q1

0.5

0.4

0.3

0.2

0.1

0.0
1000 2000 3000 4000

-0.1 Time

KLDF.Q1 KLDF.Q1 KLDF.Q1

 Estimating kLDF
ACPX series – kLDF x PSD

Sample kLDF (1/s) Dc/R2 (1/s) Ratio

ACPX 22 0.075 0.004 19.7
ACPX 41 0.120 0.009 13.3
ACPX 76 0.136 0.009 14.9
P. A. S. Moura et al., Adsorption 24, (2018)
 Estimating kLDF
Zeolite 13X
0.7 0.7

0.6 -1 0.6
KLDF = 0.022 s 2
Dc/R = 0.0013 s
-1
0.5 0.5

Heat Rate (mWatt)

Heat Rate (mWatt)

0.4 0.4

0.3 0.3

Experiment Experiment
0.2 0.2
Simulation Simulation
0.1 0.1

0.0 0.0
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 0 250 500 750 1000 1250 1500 1750 2000 2250 2500
Time (s) Time (s)

https://www.explainthatstuff.com/zeolites.html

Sample Method Dc/R2

binderless ZLC at 313 K 0.0009-0.0012

13X zeolite Calorimetry at 298 K 0.0013

J. A. C. Silva, K. Schumann, A. E. Rodrigues, Microporus Mesoporous Materials 158, (2012)

 Outline

Introduction – Concepts
 Aims and basics
 Adsorption
 Adsorbents
 Column dynamics
Modelling – Theoretical Background
 Definitions and terminology
 Momentum, Material and Energy Balances
 Equilibrium theory
 Adsorption kinetics
Assessing mass transfer
 Simple fit to breakthrough curves
 From uptake curves
 From calorimetry
Final remarks
 Final remarks
What about breakthrough curves of systems with non-conventional
adsorption isotherms?

PX isotherm
MIL-53 (Al)
Remy et al., 2011

CO2 isotherms
YO-MOF
Mulfort et al., 2010

IUPAC classification for condensable

vapors (e.g. N2 at 77 K)
 Final remarks
A zoo of breakthrough curves
 Final remarks
Xylenes adsorption in MOFs
 Final remarks
MIL-47 (V) – Materiaux de l’Institute Lavoisier
Octahedral metallic clusters VO4(OH)2 connected by terephthalic acid linkers
 Final remarks
MIL-53 (Al) – Flexible MOF: “Breathing effect”

Octahedral metalic cluster AlO4(OH)2 connected by terephthalic acid linkers

 Final remarks
CS 1: OX/EB breakthrough curves in MIL-53

Finsy et al., Chem. Eur. J., 15, 7724 – 7731, (2009)

 Final remarks
From Rietveld refinement of in-situ DRX of OX adsorption in MIL-53 (Al), it was
found that…
 Final remarks
Fixed bed model equations (LDF)

ISOTHERM EQUATIONS
Final remarks
 Final remarks
CS2: MIL-47 (V) – a rigid MOF
 Final remarks
PX and MX isotherms in MIL-47 at 343 K

• No breathing
• Similar isotherms, although PX shows an inflexion point at 10-3 and 10-2 bar
 Final remarks
Same trick applied to provide a mathematical description of PX isotherm
Final remarks

Finsy et al., JACS (2008) 130, 7110

Remy et al., Langmuir (2011) 27, 13064
Final remarks

Finsy et al., JACS (2008) 130, 7110

Remy et al., Langmuir (2011) 27, 13064
 Final remarks
Coming back to the equilibrium theory...
 Final remarks

• Classical concepts such as phase equilibrium

and transport phenomena have been revisited
and applied to the description of adsorption
dynamics in a fixed bed

• The correct analysis of batch adsorption data

should provide scalable (design) parameters that
will be useful not only for process design and
optimization, but also to plan experiments in
fixed bed in lab scale
 Final remarks

We hope this is a small brick in bridging the gap

between different approaches of adsorption
scientists from more fundamental and more
applied backgrounds
 Our team

Faculty

Diana Azevedo Célio Cavalcante Jr. Eurico Torres Moises Bastos-Neto Enrique Vilarrasa-Garcia

Post-Doc Fellows

D Sc

M Sc

Ba
Laboratório de Pesquisa em Adsorção e Captura de CO2
(Laboratory of Adsorption Research and CO2 Capture)
Universidade Federal do Ceara
Campus do Pici – bloco 731
Fortaleza - Brazil

email:
[email protected]

Thank you for your attention!