Accepted Manuscript

Separation of CO2 / CH4 using Carbon Molecular Sieve (CMS) at low and high
pressure

Luis A.M. Rocha, Kari Anne Andreassen, Carlos Grande

PII: S0009-2509(17)30083-0
DOI: http://dx.doi.org/10.1016/j.ces.2017.01.071
Reference: CES 13418

To appear in: Chemical Engineering Science

Received Date: 26 August 2016

Revised Date: 24 January 2017
Accepted Date: 29 January 2017

Please cite this article as: L.A.M. Rocha, K. Anne Andreassen, C. Grande, Separation of CO2 / CH4 using Carbon
Molecular Sieve (CMS) at low and high pressure, Chemical Engineering Science (2017), doi: http://dx.doi.org/
10.1016/j.ces.2017.01.071

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 Separation of CO2 / CH4 using Carbon Molecular Sieve (CMS) at low
and high pressure

Luis A. M. Rocha §, Kari Anne Andreassen and Carlos Grande *

SINTEF Materials and Chemistry, Forskningsveien 1, 0373, Oslo, Norway

*Corresponding author: [email protected] ;

§
Present address: Faculty of Engineering, University of Porto. Portugal.

1
Abstract

Separation of CO2 from CH4-rich streams is an important separation for upgrading

biogas and natural gas. The main difference between these two sources of methane is the

pressure at which they are available. In order to design an adsorption process without

extrapolating data, it is necessary to measure the behaviour of the adsorbent in the current

conditions.

In this work, adsorption properties of CH4 and CO2 on a commercial carbon molecular sieve

(CMS) material are reported. Adsorption equilibrium of pure gases at 298 and 343 K were

measured until 70 bar for CH4 and 30 bar for CO2. In order to determine the performance of

the adsorbent using a real mixture, binary breakthrough curves (10% CO2 – 90% CH4) were

measured at 314 K at pressures from 5 to 70 bar. A mathematical model was used to interpret

the data obtained.

Keywords: carbon dioxide; natural gas; adsorption; diffusion; breakthrough curves; pressure

swing adsorption

2
1. Introduction

The production of sufficient energy with economic viability and environmental

sustainability is one of the main concerns nowadays (Shimekit and Mukhtar, 2012). Given its

importance as a fuel and as a source of hydrocarbons for petrochemical feed stocks, the use

of natural gas must therefore follow the aforementioned requirements. One of the main issues

regarding the handling and treatment of natural gas is the separation of impurities such as

carbon dioxide. The removal of CO2 is necessary for transportation in pipelines to avoid

corrosion. Typical content of CO2 to comply pipeline-quality natural gas is lower than 3%

and depends on the country – region.

Numerous methods have been studied and implemented for the separation of CO 2 from

CH4. Absorption with amines is the commercial state-of-the-art technology (Peters et al.,

2011; Shimekit and Mukhtar, 2012). Recently, cryogenic distillation (Shimekit and Mukhtar,

2012) and membrane processes (Peters et al., 2011; Shimekit & Mukhtar, 2012; Zhang et al.,

2013) are getting more attention under certain circumstances. Adsorption technologies such

as pressure swing adsorption (PSA) and temperature swing adsorption (TSA) have been

recently considered for methane purification from natural gas and have a great potential to

expand its utilization (Esteves et al. 2008; Shimekit and Mukhtar 2012; Grande 2012;

Tagliabue et al. 2009). However, those applications were considering relatively low pressures.

A full set of adsorption data to design and optimize a PSA unit that operates at pressures as

high as 70 bar or higher is not available in literature.

One of the most important choices for an adsorption process is the adsorbent that will be

employed for the separation. The adsorbent should exhibit appropriate selectivity, capacity,

service life and ease of regeneration. In CH4-CO2 applications at high pressures with

relatively low CO2 molar fractions, it is very important to limit the amount of methane

adsorbed (Grande and Blom 2012). If there is a significant amount of methane adsorbed, the

3
released heat due to CH4 adsorption will result in substantial temperature increase of the

adsorbent and thus in a reduced capacity towards CO2. A possible way to limit the amount of

methane adsorbed is using kinetic adsorbents where the pores of the material are small: CO 2

can adsorb relatively fast while CH4 will take longer time to penetrate the pores. But in any

case, the process will not be isothermal, due to the release of heat due to adsorption of CO2.

Carbon molecular sieve (CMS) is one type of kinetic adsorbent that operates based on

different rates of adsorption. CMS materials are normally prepared by making a hydrocarbon

cracking process over an activated carbon material, rendering pores that are partially blocked

at their entrance. By controlling the cracking, the aperture of the pores can be controlled. Like

the activated carbon used as precursor, CMS are bidisperse adsorbents, which means that its

pore structure is composed of macropores branched with micropores. The cracking process

will however generate extra resistances to diffusion, normally at the pore mouth of the

micropores (Haynes and Lucas, 2007).

Adsorption equilibrium and diffusion of methane and carbon dioxide has been already

reported in several CMS samples, both commercial and laboratory-prepared (Kapoor and Yang

1989; Jayaraman et al. 2002; Qinglin, Sundaram, and Farooq 2003; Rutherford et al. 2003;

Wang, Sward, and LeVan 2003; Bae and Lee 2005; Cavenati, Grande, and Rodrigues 2005;

Kim et al. 2006; Wahby et al. 2010; Silvestre-Albero et al. 2011; Arya et al. 2015). In most of

these publications, the equilibrium selectivity towards CO2 is not very high but the limitations

to diffusion of methane are considerable reason why an effective kinetic selectivity is high.

In this work we have made a complete study regarding the utilization of CMS KP407

(Japan Envirochemicals, Japan). We have measured adsorption equilibrium of pure methane

and pure carbon dioxide at temperatures of 298 and 343 K. Results were fitted with the

multisite Langmuir and dual-site Langmuir models.

Dynamic breakthrough experiments were carried out at 314.65 K using a fixed gas

4
composition: a pre-mixed gas bottle with 10% CO2 balanced by methane (Yara, Norway).

Experiments were made at 5, 10, 20, 40, 60 and 70 bar total pressure. A mathematical model

was used to interpret the data.

2. Experimental

Adsorption equilibrium of pure CH4 and CO2 was measured in a Belsorp Max instrument

(Japan). The measurements were carried out at 298 and 343 K. Methane with a purity of

>99.9995% and carbon dioxide with a purity of >99.9992% were employed without further

treatment. Degassing of the samples was performed at 593 K under vacuum for 12 h. The

transient variation of pressure of the first adsorption equilibrium point was recorded to obtain

the uptake curves and used to estimate the diffusion parameters. The equilibrium pressures of

the initial points were 0.02 bar for CH4 and 0.01 bar for CO2. For the initial measurement and

all measurements at pressures lower than 1.3 bar, it was considered that adsorption

equilibrium was achieved if the pressure variations were lower than 0.3% for intervals longer

than 9999 s. For higher pressures, the termination criteria is when pressure variations are

lower than 13 kPa over 9999 seconds. Under such criteria, each methane isotherm took over

15 days to be measured.

The fixed bed experiments were conducted in a fully equipped PSA unit. This unit is

generically composed of 4 columns and a set of different valves that allow the unit to run

different regeneration cycles. Since the equipment was used to run breakthrough curve

experiments, only one column was necessary. A simplified scheme used for the experiments

is shown in Figure 1. The unit was designed to run in a pressure range of 0.1 – 70 bar and 298

– 550 K. The unit allows the connection of diverse analytical methods for analysis of the

outlet components and includes an injection of a tracer gas (argon) to determine the existence

of flow variations. For the CO2/CH4 separation, mass spectroscopy (MS) was the method

used to identify the composition of the outlet stream.

5
 The column where adsorbent was packed for breakthrough curve measurements is made of

steel with the following dimensions: 0.56 m long and 0.0254 m external diameter. The

temperature changes inside the column were monitored by four thermocouples located at

0.12 m, 0.27 m, 0.42 m and 0.56 m from the feed inlet. The thermocouples are located in the

centre of the column. The columns are placed inside an oven with a fan. The breakthrough

curves were measured at different inlet pressures, ranging from 5 to 70 bar. At some pressure

level, experiments with two different flowrates were made. The operating conditions of the

experiments are presented in Table 1.

3. Mathematical modelling

3.1.Adsorption equilibrium

From the several mathematical models available in literature to describe adsorption

equilibria, we have selected the multi-site Langmuir (Nitta et al., 1984) and the dual-site

Langmuir models. Both are extensions of the Langmuir model. The multi-site Langmuir

(MSL) accounts for the variation of adsorbate size; the adsorbent has a limited number of

adsorption sites and that the adsorbate particles occupy a certain number of these sites. The

MSL model is described by:

(1)

where is the adsorption constant, is the maximum adsorption capacity and

represents the number of neighbouring sites that can be occupied by an adsorbate particle.

The adsorption constant has an exponential dependence with temperature described by:

(2)

where is the heat of adsorption, is the adsorption constant at infinite temperature,

is the ideal gas constant and is temperature. The multisite Langmuir isotherm corresponds
6
to the Langmuir isotherm when One major advantage of this model is that it has a

direct theoretical equation to describe adsorption equilibrium of multicomponent mixtures

based on pure gas parameters. The multicomponent extension of the model is given by:

(3)

Given the assumptions described above about the fixed number of sites on the adsorbent,

the following relation must be verified in multicomponent systems to ensure thermodynamic

consistency (Sircar, 1995):

(4)

The parameters were optimized with a code written in Python that uses the Nelder-Mead

optimization method for finding the minimum of the error function: square of residuals for

methane and average of residuals for carbon dioxide. The procedure using the square of

residuals did not render a good fitting of the CO2 isotherms reason why these isotherms were

fitted using the average of residuals as error function that enhances the error at lower

pressures.

The dual-site Langmuir model (DSL) assumes that there are two different sites of

adsorption in the adsorbent. The main advantage of this model compared to the traditional

Langmuir model is basically that it has more parameters which gives it more flexibility to fit

isotherms.

The adsorption constants are also described by exponential functions as described by

equation (2). This model also has a theoretical expression to calculate adsorption equilibrium

of multi-component mixtures based on pure component parameters. The multicomponent

extension is (Ritter et al., 2011):

(5)

7
3.2.Breakthrough experiments in fixed-bed column

The model used in this work to describe the breakthrough experiments was already

derived (Da Silva, Silva, and Rodrigues, 1999) and successfully applied to the CH4/CO2

separation with activated carbon (Grande and Rodrigues 2007). One of the assumptions of the

model is that the gas phase follows the ideal gas law, however the density of CH4, CO2 and of

the overall gas mixture start to deviate considerably from the ideal gas behaviour above 30

bar. Therefore, the BWR equation of state (Benedict, Webb, and Rubin, 1940) was used in the

simulations, and has been shown to adequately model this system (Grande et al. 2013). Thus,

the assumptions of the model are:

1. Mass, heat and momentum transport in the radial direction are negligible;

2. Enthalpy variation was assumed to follow ideal gas behaviour. This might be an

oversimplification that should be revised when operating under certain conditions with

strong deviations from ideal gas (Stegmaier, 2016).

3. The linear driving force model (LDF) is valid to describe mass transfer in the

macropores and micropores;

4. Constant cross section area, void fraction and adsorbent properties along the column;

5. Homogeneous temperature inside the particles.

The mass balance of component i is given by:

(6)

where ci is the concentration of component i in the fluid phase, is the total concentration

in the fluid phase, is the molar fraction of component i, ui is the interstitial velocity, Dax is

the axial dispersion coefficient, εb is the bed porosity, is the external mass transfer

coefficient for component i, Bi is the mass Biot number, t is time and z is the position along

the column. This equation is solved with the Danckwerts boundary conditions:

8
 (7)

(8)

The pore structure of carbon molecular sieves such as the one used in this work is made of

macropores branched with micropores. The adsorbent can thus be assumed to be bidisperse

(Yang, 1987) and a bi-LDF model can be used to reduce computational time without losing

accuracy in the solution. The diffusion in the macropores was determined with the following

equation:

(9)

where is the particle density, is the particle porosity and is the macropore

diffusion constant, given by:

(10
)

The diffusion in the micropores is given by the following equation, based on a LDF

approximation:

(11
)

where the concentration of the component i at the surface of the adsorbent and is

the micropore diffusion constant.

The Ergun equation is also assumed valid to calculate the pressure drop along the column:

(12
)

where is the total pressure, is the gas viscosity and is the gas density. As

mentioned, the BWR equation, given below, was used to model the behaviour of the gas

phase:

9
 (13
)

The expressions used to calculate the mixing parameters and the constants for CO 2 and CH4

are presented in Tables 2 and 3, respectively. The boundary condition for the pressure drop is

the following:

(14
)

where is the pressure at the exit of the column.

For the energy balance, a heterogeneous model was used, with an equation for the balance in

the gas phase, the solid phase and the column wall. The energy balance in the gas phase is

given by:

(15
)

where is the molar constant volumetric heat of component i, is the axial heat

dispersion coefficient, is the molar constant pressure specific heat of the gas mixture,

is the ideal gas constant, is the film heat transfer coefficient between the gas and the solid

phase, is the film heat transfer coefficient between the gas phase and the column wall,

is the radius of the wall, is the temperature of the gas, is the temperature of the

solid and is the temperature of the wall. The balance in the solid phase was determined

with this equation:

(16
)

10
where is the molar constant volumetric specific heat of component i adsorbed and

is the constant pressure specific heat of the adsorbent. Finally, the energy balance in the

column wall is given by:

(17
)

where is the wall density, is the specific heat of the adsorbent, is the global

external heat transfer coefficient, is the ambient temperature, is the ratio of the

internal surface area to the volume of the column wall, given by:

(18
)

where is the column internal diameter and is the thickness of the shell and is the

ratio of the logarithmic mean surface area of the column shell to the volume of the column

wall, given by:

(19
)

The boundary conditions for the energy balance are as follows:

(20
)

(21
)

The simulations were made with the following initial conditions:

(22
)

(23
)

(24

11
)

This model has been previously used for description of gas mixtures at lower pressures

(Da Silva, 1999). The same correlations were used for calculating all mass and energy

transfer parameters.

The mathematical model was solved using gPROMS (PSE Enterprise, UK). The second

order centred finite difference method with 200 elements was used.

4. Results and Discussion

4.1.Adsorption equilibrium and kinetics of pure gases

The adsorption equilibrium isotherms of CH4 and CO2 at 298 and 343 K are presented in

Figures 2 and 3. The adsorbent is more selective to carbon dioxide although the equilibrium

selectivity is not very high. This is expected due to the nature of a carbonaceous adsorbent.

Results are comparable with already available experimental results of CMS materials

(Cavenati, Grande, and Rodrigues 2004; Cavenati, Grande, and Rodrigues 2005; Grande and

Blom 2014). The heat of adsorption at zero coverage was determined only as a fitting

parameter. To make use of classical methods for determination of heat of adsorption (i.e.

Clausius Clapeyron equation) an isotherm at a third temperature would have been desired.

The solid lines in these figures represent the fitting of the Multi-site Langmuir (MSL) and

the dashed lines to the dual-site Langmuir (DSL). The parameters of the fitting are listed in

Tables 4 and 5. Both models show very good fitting of the isothermal experimental data. A

better description of the models and the experimental data is given in Figure 4 where the

Virial plot is presented for all the isotherms. As can be observed, the experimental data of

CH4 at 298 K is scatter at low pressures. It should be mentioned that each of these points took

around 1 day of gas-adsorbent contact. Since this process is entirely dominated by slow

12
diffusion, it can overcome the 9999 seconds used as criteria for achieving equilibrium (is not

possible to use longer times in the experimental unit used). If in that pre-specified time, the

pressure is not changing over 13 kPa, the equipment will automatically assume that

equilibrium was reached and will move to the next point.

The differences of the two models for fitting the experimental data is more evident in

Figure 4. The DSL model can describe better what happens in the low pressure range of CO2

and slightly better in the case of CH4 at higher pressures.

The diffusion of gases in the adsorbent is shown in Figure 5. The results clearly show that

carbon dioxide diffuses through the micropores at a much faster rate than methane.

Equilibrium is reached with CO2 after approximately 800 seconds (around 13 minutes), at a

time when the factor for CH4 is only at 0.07. Equilibrium for methane is reached only

after over 25 hours. In order to better visualize the effect of the resistance at the mouth of the

micropore, the results were shown in square root of time (Cavenati, Grande, and Rodrigues

2005). A smaller slope is clearly visible in the methane curve, indicating that the constriction

at the mouth of the micropore has an important contribution to the overall diffusion resistance.

For carbon dioxide, however, this effect is not observed. This difference in the kinetic

behaviour of the two components was previously reported and is due to the difference in the

molecular size of the molecules (CO2 diameter is 3.3 Å and CH4 diameter is 3.8 Å). From the

data in Figure 5, the CO2 micropore diffusion constant ( ) was 0.01 s-1 and the constant for

CH4 was 0.0001 s-1.

4.2.Breakthrough curves of binary mixture at different pressures

A comprehensive set of experimental data was measured to cover the performance of the

adsorbent in a wide range of pressures, being able to report use of this data at low pressures

(in the context of biogas upgrading) and at high pressures for natural gas conditioning. The

13
authors could not find multicomponent dynamic data (breakthrough curves for example) on

CMS materials at pressures higher than 30 bars.

Two examples, at 10 and at 70 bars, will be used to discuss the results. The results of all

other breakthrough curves will be reported as Supplementary Information.

The breakthrough results at 10 bar are shown in Figure 6 while the ones at 70 bar are

reported in Figure 7. Since mass and energy transfer are very interlinked, the exit molar

fraction and the temperature profiles inside the column are shown. Although the column is

placed in an oven with forced convection, the experimental data is collected under the format

non-adiabatic and non-isothermal. The rate of energy generation by adsorption is faster than

the rate at which heat is removed and thus there is a net increase of temperature of the

adsorbent during adsorption.

The shape of the breakthrough curve at 10 bar is similar to others previously reported

using CMS materials. Methane is only slightly adsorbed and carbon dioxide breakthrough is

relatively steep with some dispersion caused by thermal effects. The main difference between

the mathematical model and the experimental results is indeed in describing the thermal

change. In this model constant heat transfer parameters were used and in the experimental

results it seems that the heat transfer at the end of the column is more efficient so that

variable coefficients should be used.

The simulated amount adsorbed of both components during this breakthrough curve is

shown in Figure 8. Two different shapes on the adsorption of methane and carbon dioxide can

be observed. The shape of the evolution of loading in the case of CO 2 is typical while the

evolution of the amount adsorbed of methane is clearly limited by diffusion: the more time in

contact with the solid, the more methane is adsorbed. The position of the front of CO 2 is

important also to regulate the selectivity of the material.

A detailed look of Figure 6 will also show that the mathematical model seems to overshoot

14
the inlet molar fraction of CO2. Indeed, since the amount of methane adsorbed in the

adsorbent is very strongly diffusion limited, there is a second roll-up of CO2 caused by

additional adsorption of CH4 at longer times. This is shown in Figure 9, only for modelling

results. Indeed, during the experiments it was possible to see a reduction in the CO2 signal,

but is very difficult to know whether this is due to the real behaviour of the system or to a

slight drift of the MS signal.

The results of the breakthrough curve at 70 bar shown in Figure 7 are indeed slightly

different. At higher pressures and with a much higher content of CH4, the material still shows

kinetic selectivity, but the amount adsorbed of methane even in the initial moments of the

breakthrough curve, is already considerable. At equilibrium, the amount of methane adsorbed

is 1.92 mol/kg and only 1.08 mol/kg of CO2.

It has been previously reported that the diffusion of methane in CMS materials depends on

the concentration of methane. Indeed, what was observed in the experimental campaign is

that the diffusion of methane increases with pressure, while the diffusion of CO 2 remains

constant. For this reason, the diffusion of methane was considered as the only fitting

parameter of the breakthrough curves. This is also clearly observable in the temperature of

the column, where now two clear peaks can be observed. At higher pressures, since the

diffusion is faster, no CO2 roll-up can be seen, either in simulations or experimentally.

Experimentally it would not be possible to be observed because the small flowrate oscillation

of the pressure caused by the back-pressure regulator are much larger. The diffusion of

methane was a fitting parameter for the mathematical model. The dependence of the diffusion

constant at different pressures is shown in Figure 10. It does seem to show a linear increase

with pressure but it is very important to mention that since the temperature could not be

correctly described, the link between thermal and mass transfer can have an important effect

in determining a more accurate value.

15
 The results provided in this publication can allow the simulation of Pressure Swing

Adsorption (PSA) units in a wide range of pressures using CMS KP 407 as a CO 2-selective

material. It should be pointed however, that the lack of selectivity of CMS materials is more

detrimental at higher pressures where significant amount of CH 4 is co-adsorbed. In PSA units

this will result in longer cycles and probably larger thermal oscillations unless the heat waves

can be contained within the column. If the amount of carbon dioxide is higher than 10%, the

overall performance can be better, although a better control of the thermal effects should be

necessary.

5. Conclusions

Adsorption of carbon dioxide and methane in a carbon molecular sieve was evaluated for a

wide range of pressures: from 5 to 70 bar. Equilibrium data for both pure components was

studied at 298 and 343 K and fitted with the multisite Langmuir and dual-site Langmuir

models. While the equilibrium selectivity of this adsorbent was found to be low, the kinetic

selectivity is very high, with CO2 diffusing through the pores at a much higher rate than CH4.

This is particularly true at lower pressures since it was found that the diffusion of CH4

increases with the partial pressure of methane. The slow diffusion of methane at lower

pressures resulted in a double roll-up of gases exiting the column: methane displaced by

carbon dioxide (driven by equilibrium), followed by a CO2 roll up displaced by

slow-diffusing methane.

Results indicate that the overall kinetic selectivity of the CMS material is much higher at

lower pressures and thus this material should be better suited for biogas upgrading than for

natural gas conditioning at higher pressures.

Acknowledgment

The support of the Research Council of Norway through the CLIMIT program by the
16
SINTERCAP project (233818). This publication has been produced with support from the

BIGCCS Centre, performed under the Norwegian research program Centres for

Environment-friendly Energy Research (FME). The author acknowledges the following

partners for their contributions: ConocoPhillips, Gassco, Shell, Statoil, TOTAL, GDF SUEZ

and the Research Council of Norway (193816/S60).

Nomenclature

Number of neighbouring sites that can be occupied by and adsorbate -

particle in the multisite Langmuir model
Pellet specific area m-1
Biot number of component i -
Molar concentration of component i in the fluid phase mol·m-3
Average concentration of component i in the macropores mol·m-3
Total concentration in the fluid phase mol·m-3
Molar constant pressure specific heat of the gas mixture J·mol-1·K-1
Constant pressure specific heat of the adsorbent J·kg-1·K-1
Specific heat of the column wall J·kg-1·K-1
Molar constant volumetric specific heat of component i J·mol-1·K-1
Molar constant volumetric specific heat of component i adsorbed J·mol-1·K-1
Axial dispersion coefficient m2·s-1
Micropore diffusion coefficient m2·s-1
Macropore diffusion coefficient m2·s-1
Film heat transfer coefficient between gas and solid W·m-2·K-1
Film heat transfer coefficient between gas and the wall W·m-2·K-1
Adsorption constant bar-1
Adsorption constant at infinite temperature bar-1
Film mass transfer coefficient of component i m·s-1
Macropore diffusion constant of component i s-1
Micropore diffusion constant of component i s-1
Péclet number -
Partial pressure of component i Pa
Total pressure Pa
Maximum adsorption capacity mol·kg-1
Adsorbed phase concentration in equilibrium with the gas inside the mol·kg-1
particle for component i
Average adsorbed concentration for component i in the pellet mol·kg-1
Ideal gas constant J·mol-1·K-1
17
 Crystal radius m
Pore radius m
Radius of the wall m
time s
Gas temperature K
Solid temperature K
Wall temperature K
Ambient temperature K
Overall heat transfer coefficient W·m-2·K-1
Superficial gas velocity m·s-1
Axial distance along the column M

Greek letters

Ratio of the internal surface area to the volume of the column wall m-1
Ratio of the log mean surface to the volume of the column wall m-1
Isosteric heat of adsorption of I component J·mol-1
Porosity of the column -
Porosity of the pellet -
Axial heat dispersion coefficient W·m-2·K-1
Gas viscosity Pa·s
Gas density kg·m-3
Pellet density kg·m-3
-3
Column wall density kg·m

List of Acronyms

BPR Back-pressure regulator

MFC Mass flow controller
MS Mass spectrometer
PSA Pressure swing adsorption

18
References

Arya, A., Divekar, S., Rawat, R., Gupta, P., Garg, M. O., Dasgupta, S., Nanoti, A., Singh, R.,

Xiao, P., Webley, P. A. (2015). Upgrading biogas at low pressure by vacuum swing

adsorption. Industrial and Engineering Chemistry Research, 54, 404–413.

Bae, Y. S., Lee, C. H. (2005). Sorption kinetics of eight gases on a carbon molecular sieve at

elevated pressure. Carbon, 43, 95–107.

Benedict, M., Webb, G. B., Rubin, L. C. (1940). An Empirical Equation for Thermodynamic

Properties of Light Hydrocarbons and Their Mixtures I. Methane, Ethane, Propane and

n-Butane. The Journal of Chemical Physics, 8, 334–345.

Cavenati, S., Grande, C. A, Rodrigues, A. E. (2004). Adsorption equilibrium of methane,

carbon dioxide, and nitrogen on zeolite 13X at high pressures. Journal of Chemical and

Engineering Data, 49, 1095–1101.

Cavenati, S., Grande, C. A., Rodrigues, A. E. (2005). Upgrade of Methane from Landfill Gas

by Pressure Swing Adsorption. Energy & Fuels, 19, 2545–2555.

Da Silva, F.A. Cyclic Adsorption Processes: Application to Propane / Propylene Separation.

PhD Thesis. University of Porto, 1999.

Da Silva, F. A., Silva, J. A., Rodrigues, A. E. (1999). General package for the simulation of

cyclic adsorption processes. Adsorption, 5, 229–244.

Esteves, I. A. A. C., Lopes, M. S. S., Nunes, P. M. C., Mota, J. P. B. (2008). Adsorption of

natural gas and biogas components on activated carbon. Separation and Purification

Technology, 62, 281–296.

Grande, C. A. (2012). Advances in Pressure Swing Adsorption for Gas Separation. ISRN

Chemical Engineering, 2012, 1–13.

Grande, C. A., Blom, R. (2012). Utilization of dual - PSA technology for natural gas

upgrading and integrated CO2 capture. Energy Procedia, 26, 2–14.

19
Grande, C. A., Blom, R. (2014). Cryogenic Adsorption of Methane and Carbon Dioxide on

Zeolites 4A and 13X. Energy & Fuels, 28, 6688–6693.

Grande, C. A., Blom, R., Möller, A., Möllmer, J. (2013). High-pressure separation of

CH4/CO2 using activated carbon. Chemical Engineering Science, 89, 10–20.

Grande, C. A., Rodrigues, A. E. (2007). Biogas to fuel by vacuum pressure swing adsorption I.

Behavior of equilibrium and kinetic-based adsorbents. Industrial and Engineering

Chemistry Research, 46, 4595–4605.

Haynes, P. D., Lucas, S. K. (2007). Extension of a short-time solution of the diffusion

equation with application to micropore diffusion in a finite system. The ANZIAM

Journal, 48, 503–521.

Jayaraman, A., Chiao, A. S., Padin, J., Yang, R. T., Munson, C. L. (2002). Kinetic separation

of methane/carbon dioxide by molecular sieve carbons. Separation Science and

Technology, 37, 2505–2528.

Kapoor, A., Yang, R. T. (1989). Separation of Methane-Carbon Dioxide Mixture by

Adsorption on Molecular Sieve Carbon. Chemical Engineering Science, 44, 1723–1733.

Kim, M., Bae, Y., Choi, D., Lee, C. (2006). Kinetic Separation of Landfill Gas by a Two-Bed

Pressure Swing Adsorption Process Packed with Carbon Molecular Sieve :

Nonisothermal Operation. Industrial & Engineering Chemistry Research, 45,

5050–5058.

Nitta, T., Kuro-oka, M., Katayama, T., Shigetomi, T. (1984). Adsorption Isotherm of

Multi-Site Occupancy Model for Heterogeneous Surface. Journal of Chemical

Engineering of Japan, 17, 45–52.

Peters, L., Hussain, A., Follmann, M., Melin, T., Hägg, M. B. (2011). CO2 removal from

natural gas by employing amine absorption and membrane technology-A technical and

economical analysis. Chemical Engineering Journal, 172, 952–960.

20
Qinglin, H., Sundaram, S. M., Farooq, S. (2003). Revisiting transport of gases in the

micropores of carbon molecular sieves. Langmuir, 19, 393–405.

Ritter, J .A.; Bhadra, S. J.; Ebner, A. D. On the Use of the Dual-Process Langmuir Model for

Correlating Unary Equilibria and Predicting Mixed-Gas Adsorption Equilibria.

Langmuir, 2011, 27, 4700-4712.

Rutherford, S. W., Nguyen, C., Coons, J. E., Do, D. D. (2003). Characterization of Carbon

Molecular Sieves Using Methane and Carbon Dioxide as Adsorptive Probes. Langmuir,

19, 8335–8342.

Shimekit, B., Mukhtar, H. (2012). Natural Gas Purification Technologies–Major Advances for

CO2 Separation and Future Directions. Advances in Natural Gas Technology, 235–270.

Silvestre-Albero, J., Wahby, A., Sepúlveda-Escribano, A., Martínez-Escandell, M., Kaneko,

K., Rodríguez-Reinoso, F. (2011). Ultrahigh CO2 adsorption capacity on carbon

molecular sieves at room temperature. Chemical Communications, 47, 6840–6842.

Sircar, S. (1995). Influence of adsorbate size and adsorbent heterogeneity on IAST. AIChE,

41, 1135–1145.

Stegmaier, M. Nonideal Gas Simulation of Pressure Swing Adsorption Processes.

Fundamentals of Adsorption 12. 31st May 2016. Friedrichshafen, Germany.

Tagliabue, M., Farrusseng, D., Valencia, S., Aguado, S., Ravon, U., Rizzo, C., Corma, A.,

Mirodatos, C. (2009). Natural gas treating by selective adsorption: Material science and

chemical engineering interplay. Chemical Engineering Journal, 155, 553–566.

Wahby, A., Ramos-Fernández, J. M., Martínez-Escandell, M., Sepúveda-Escribano, A.,

Silvestre-Albero, J., Rodríguez-Reinoso, F. (2010). High-surface-area carbon molecular

sieves for selective CO2 adsorption. ChemSusChem, 3, 974–981.

Wang, Y., Sward, B. K., LeVan, M. D. (2003). New Frequency Response Method for

Measuring Adsorption Rates via Pressure Modulation: Application to Oxygen and

21
 Nitrogen in a Carbon Molecular Sieve. Industrial & Engineering Chemistry Research,

42, 4213–4222.

Yang, R. T. (1987). Gas Separation by Adsorption Processes. Boston:

Butterworth-Heinemann.

Zhang, Y., Sunarso, J., Liu, S., & Wang, R. (2013). Current status and development of

membranes for CO2/CH4 separation: A review. International Journal of Greenhouse Gas

Control, 12, 84–107.

22
Tables

Table 1. Experimental conditions used in the breakthrough curves.

Column properties

Length (m) 0.56

External diameter (m) 0.0254

Internal diameter (m) 0.0211

Bed voidage 0.377

Adsorbent properties

Shape Cylindrical

Diameter (m) 9.0 × 10-4

Length (m) 1.8 × 10-3

Particle density (kg·m-3) 1060

Inlet gases and operating conditions

Temperature (K) 314.65

Composition 90 % CH4, 10 % CO2

Pressure (bar) 5, 10, 20, 40, 55 and 70

Flowrate (mL/min) 517 and 1100 (298K, 1 atm)

Tracer gas Argon (100 ml/min)

23
 Table 2. Expressions used to calculate the mixing parameters of the BWR equation.

Table 3. Constants of the BWR equation for CH4 and CO2.

CH4 5 187.91 3.38 × 10-3 4.26 × 10-2 2.58 × 105 2.29 × 106 1.24 × 10-4 6 × 10-3

CO2 13.86 277.30 7.21 × 10-3 4.99 × 10-3 1.51 × 106 1.40 × 107 8.47 × 10-5 5.4 × 10-3

24
 Table 4. Fitting parameters of the multisite Langmuir model.

qm / mol·kg-1 k0 / Pa-1 -ΔH / J·mol-1 a qm·a

CH4 3.493 8.169 × 10-10 21 410 2.047 7.151

CO2 4.299 8.436 × 10-10 23 201 1.663 7.151

Table 5. Fitting parameters of the dual-site Langmuir model.

qm1 / mol·kg-1 k01 / Pa-1 -ΔH1 / J·mol-1 qm2 / mol·kg-1 k02 / Pa-1 -ΔH2 / J·mol-1

CH4 1.815 8.701 × 10-11 24287 1.226 1.629 × 10-7 9310

CO2 2.8442 1.560 × 10-10 25565 1.068 8.970 × 10-10 27948

25
Figure Captions

Figure 1. Simplified diagram of the experimental setup used for the breakthrough curve

experiments. Abbreviations: MFC: mass flow controller, BPR: back-pressure regulator,

MS: mass spectrometer

Figure 2. Adsorption equilibrium of carbon dioxide at 298 and 343 K on CMS KP 407.

Solid lines represent the fitting from MSL and dashed lines from DSL.

Figure 3. Adsorption equilibrium of methane at 298 and 343 K on CMS KP 407. Solid

lines represent the fitting from MSL and dashed lines from DSL.

Figure 4. Virial plot of adsorption equilibrium of CH4 (a) and CO2 (b) at 298 and 343 K

on CMS KP 407. Solid line corresponds to MSL and dashed lines to DSL models.

Figure 5. Diffusion of CH4 and CO2 at 298 K at low pressure (first adsorption

equilibrium point) in CMS KP 407.

Figure 6. Breakthrough curve of 10% CO2 – 90% CH4 in CMS KP 407 at 314.65 K and
10 bar. Solid lines correspond to the simulation using the adsorption equilibrium
described by the dual-site Langmuir model. (a) Exit molar fraction and (b) temperature
measured at four different positions along the column.

Figure 7. Breakthrough curve of 10% CO2 – 90% CH4 in CMS KP 407 at 314.65 K and

70 bar. Solid lines correspond to the simulation using the adsorption equilibrium

26
described by the dual-site Langmuir model. (a) Exit molar fraction and (b) temperature

measured at four different positions along the column.

Figure 8. Amount adsorbed of CH4 (a) and CO2 (b) in different positions of the column at

different times during breakthrough curve performed at 10 bar.

Figure 9. Simulated molar fraction of a "long-term" breakthrough curve for CH4 – CO2

mixture at 10 bars showing roll-up for methane and CO2.

Figure 10. Dependence of diffusion constant of methane with pressure based on fitting

the breakthrough curve experimental data.

27
Figures

Figure 1.

28
 4
298 K
3.5
Amount adsorbed [mol/kg]

3
343 K
2.5

1.5

0.5

0
0 5 10 15 20 25 30
Pressure [bar]

Figure 2.

29
 4

3.5
Amount adsorbed [mol/kg]

3
298 K
2.5
343 K
2

1.5

0.5

0
0 20 40 60 80
Pressure [bar]

Figure 3.

30
 2.5
(a)
2
1.5
343 K
1
0.5
ln(P/qads)

0
-0.5 298 K

-1
-1.5
-2
-2.5
0 1 2 3 4
Amount adsorbed [mol/kg]
3.5
(b)
3
2.5
343 K
2
ln(P/qads)

1.5
298 K
1
0.5
0
-0.5
-1
0 1 2 3 4
Amount adsorbed [mol/kg]

Figure 4.

31
 1

0.8 CO2

0.6
C/Co

CH4
0.4

0.2

0
0 50 100 150 200 250 300
sqrt(t) [s1/2]

Figure 5.

32
 1 0.1 (a)

0.8 0.08

0.6 0.06
yCH4

yCO2
0.4 0.04

0.2 0.02

0 0
0 1000 2000 3000 4000 5000
Time [seconds]

330
(b)
328
326
Temperature [K]

324
322
320
318
316
314
0 1000 2000 3000 4000 5000
Time [seconds]

Figure 6.

33
 1.00 0.1
(a)

0.80 0.08

0.60 0.06

yCO2
yCH4

0.40 0.04

0.20 0.02

0.00 0
0 2000 4000 6000 8000
Time [seconds]

330
328 (b)
326
Temperature [K]

324
322
320
318
316
314
312
0 2000 4000 6000 8000
Time [seconds]

Figure 7.

34
 1
amount adsorbed [mol/kg] 0.9 (a)
4000 s
0.8
0.7 3000 s
0.6
0.5 2000 s
0.4
0.3
1000 s
0.2
0.1
0s
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Axial position [m]

0.9
(b)
amount adsorbed [mol/kg]

0.8
4000 s
0.7
0.6
3000 s
0.5
0.4
0.3
2000 s
0.2
1000 s
0.1
0s
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Axial position [m]

Figure 8.

35
 1 0.1
CH4 molar fraction

0.8 0.08

CO2 molar fraction

0.6 0.06

0.4 0.04

0.2 0.02

0 0
1 10 100 1000 10000
Time [seconds]

Figure 9.

36
 0.0012

Methane diffusion constant [s-1] 0.001

0.0008

0.0006

0.0004

0.0002

0
0 20 40 60 80
Pressure [bar]

Figure 10.

37
Highlights
 Adsorption data for CH4 and CO2 is presented for pressures as high as 70 bar.
 Breakthrough curves were measured to assess the performance of the adsorbent.
 The adsorbent presents low equilibrium selectivity but high kinetic selectivity.
 The overall performance of the adsorbent is higher at lower pressures.

38