Manuscript Details

Manuscript number SEPPUR_2019_3658

Title Using Transient Breakthrough Experiments for Screening of Adsorbents for

Separation of C2H4/CO2 Mixtures

Article type Full Length Article

Abstract
The recovery of C2H4 from gaseous reactor effluents from processes such as oxidative coupling of methane (OCM),
and biomass gasification (at relatively low temperatures (770 - 880 °C) is becoming of increasing industrial and
economic importance. The reactor effluents are N2/H2/CO/CO2/CH4/C2H6/C2H4 mixtures, typically containing less
than 15% C2H4. For recovery of C2H4, pressure swing adsorption (PSA) technology using a selective adsorbent
offers an energy-efficient alternative to the more conventional separation schemes such as amine absorption and
cryogenic distillation The major objective of this investigation is to screen commercially available cation exchanged
zeolites (13X, CaX, NaY, 5A, 4A) and activated carbon (AC) to determine the most suitable adsorbent. For all these
materials, the adsorption strengths of CO2, and C2H4 are significantly higher than that of other gaseous constituents;
consequently, the C2H4/CO2 separation selectivity is the key to the efficacy of any adsorbent. The variety of
adsorbents were screened using transient breakthrough experiments with feed mixtures using different C2H4/CO2
ratios. On the basis of the breakthroughs, the adsorbents could be distinguished in three different categories: (i) 13X
and 4A are selective to CO2, (ii) CaX, NaY, and 5A are virtually non-selective, and (iii) AC is selective to C2H4 over
the entire range of feed compositions. The experimental breakthrough results are also compared with simulations
using published unary isotherm data, along with the Ideal Adsorbed Solution Theory (IAST) for determination of
mixture adsorption equilibrium. This comparison demonstrates that screening adsorbents solely on the basis of IAST
calculations is likely to be misleading. This article underscores the need for performing transient breakthrough
experiments with realistic feed gas mixtures for process modelling and development purposes.

Keywords ethylene recovery; oxidative coupling of methane; biomass gasification; transient

breakthrough; zeolites; activated carbon; mixture adsorption; non-ideality;
selectivity reversal;

Manuscript category Adsorption processes for separation including pressure swing, temperature
swing

Corresponding Author Rajamani Krishna

Corresponding Author's Van 't Hoff Institute for Molecular Sciences

Institution

Order of Authors Ilona van Zandvoort, Erik-Jan Ras, Robbert de Graaf, Rajamani Krishna

Suggested reviewers Matthias Thommes, Alvaro Orjuela, Ryan Lively, Andreas Moeller, Jens Möllmer,
Jens-Uwe Repke, Martin HARTMANN
Submission Files Included in this PDF
File Name [File Type]
cover letter.doc [Cover Letter]

highlights.doc [Highlights]

graphical_abstract.pdf [Graphical Abstract]

Screening text.docx [Manuscript File]

Figure 01.pdf [Figure]

Figure 02.pdf [Figure]

Figure 03.pdf [Figure]

Figure 04.pdf [Figure]

Figure 05.pdf [Figure]

Figure 06.pdf [Figure]

SupplementaryMaterial.pdf [e-Component]

To view all the submission files, including those not included in the PDF, click on the manuscript title on your EVISE
Homepage, then click 'Download zip file'.

Research Data Related to this Submission

There are no linked research data sets for this submission. The following reason is given:
The Supplementary Material (e-component) contains all the necessary research data
 Dear Editor

Our submission concerns the recovery of C2H4 from CO2-bearing gaseous mixtures emanating from

(a) the OCM process, (b) producer gas from biomass gasification. Use of PSA technology for C2H4

recovery requires C2H4-selective sorbents, because C2H4 in the feed is < 15%. The selective adsorption

of C2H4 from CO2 is not discussed widely in the published literature. We present a screening strategy

based on transient breakthrough experiments for screening cation-exchanged zeolites and activated

carbon (AC). Remarkably, the best C2H4-selective sorbent is AC. We also show that use of IAST

calculations alone can lead to misleading results. Our paper is of both fundamental and practical

interest.

Regards

Rajamani Krishna

1
Highlights

 Recovery of small amounts of C2H4 from CO2-bearing mixtures is important

 For use in PSA technology, we require C2H4-selective sorbents

 Sorbent screening is performed by transient breakthrough experiments

 Activated carbon is the best choice for C2H4/CO2 separations

 Use of IAST calculations for sorbent screening is not reliable

 Raffinate:
C2H6
H2/CH4/C2H6/
purge
CO/CO2
PSA

Feed: H2/CH4/CO Extract:

/CO2/C2H4/C2H6 C2H4+(CO2,C2H6)
Using Transient Breakthrough Experiments for
Screening of Adsorbents for Separation of
C2H4/CO2 Mixtures

Ilona van Zandvoort, Erik-Jan Ras, Robbert de Graaf

Avantium Chemicals B.V.
Zekeringstraat 29
1014 BV Amsterdam, The Netherlands

and

Rajamani Krishna*
Van ‘t Hoff Institute for Molecular Sciences
University of Amsterdam
Science Park 904
1098 XH Amsterdam, The Netherlands
email: [email protected]

1
 Abstract
The recovery of C2H4 from gaseous reactor effluents from processes such as oxidative coupling of

methane (OCM), and biomass gasification (at relatively low temperatures (770 - 880 °C) is becoming of

increasing industrial and economic importance. The reactor effluents are

N2/H2/CO/CO2/CH4/C2H6/C2H4 mixtures, typically containing less than 15% C2H4. For recovery of

C2H4, pressure swing adsorption (PSA) technology using a selective adsorbent offers an energy-efficient

alternative to the more conventional separation schemes such as amine absorption and cryogenic

distillation The major objective of this investigation is to screen commercially available cation

exchanged zeolites (13X, CaX, NaY, 5A, 4A) and activated carbon (AC) to determine the most suitable

adsorbent. For all these materials, the adsorption strengths of CO2, and C2H4 are significantly higher

than that of other gaseous constituents; consequently, the C2H4/CO2 separation selectivity is the key to

the efficacy of any adsorbent. The variety of adsorbents were screened using transient breakthrough

experiments with feed mixtures using different C2H4/CO2 ratios. On the basis of the breakthroughs, the

adsorbents could be distinguished in three different categories: (i) 13X and 4A are selective to CO2, (ii)

CaX, NaY, and 5A are virtually non-selective, and (iii) AC is selective to C2H4 over the entire range of

feed compositions.

The experimental breakthrough results are also compared with simulations using published unary

isotherm data, along with the Ideal Adsorbed Solution Theory (IAST) for determination of mixture

adsorption equilibrium. This comparison demonstrates that screening adsorbents solely on the basis of

IAST calculations is likely to be misleading. This article underscores the need for performing transient

breakthrough experiments with realistic feed gas mixtures for process modelling and development

purposes.

Keywords: ethylene recovery; oxidative coupling of methane; biomass gasification; transient

breakthrough; zeolites; activated carbon; mixture adsorption; non-ideality; selectivity reversal;

2
1 Introduction

Ethene (C2H4, commonly referred to as ethylene) a valuable petrochemical feedstock, is primarily

produced on a large scale by naphtha cracking. Due to increasing demand and the desire to replace fossil

fuels, alternative sources of C2H4 have received increased attention in recent years; two such

alternatives are discussed below.

In the oxidative coupling of methane (OCM) process, two CH4 molecules are coupled to yield C2H4.

The reactor effluent consists of H2/CO/CO2/CH4/C2H6/C2H4 mixtures typically containing < 10% C2H4.

The conventional schemes for recovery of C2H4, involving amine absorption and cryogenic distillation,

are energy intensive. Use of pressure swing adsorption (PSA) technology for C2H4 recovery offers the

potential for development of an energy-efficient OCM process [1-3]. The optimal process scheme for

implementation of PSA technology requires use of a C2H4-selective sorbent in which the C2H4-free

raffinate is recycled back to the reaction train; see Figure 1. CO2 is removed from the C2H4-rich extract

stream and subsequently distilled to obtain C2H4 of the required purity. C2H6, the bottoms product from

the distillation tower, is used as purge gas in the PSA section and recycled to the OCM reaction unit.

Biomass gasification at relatively low temperatures (770 - 880 °C) results in a syngas mixture, which

also contains CO2, C2H6, along with the valuable C2H4 [4]. Recovery of C2H4 by selective adsorption

could significantly contribute to the economic viability of syngas production from biomass.

For most available adsorbents, CO2, and C2H4 are more strongly adsorbed than H2, CH4, CO, and

C2H6; consequently, the efficacy of any PSA scheme, such as that shown in Figure 1, requires the

sorbent to have a high C2H4/CO2 separation selectivity. Bachman et al.[1] have recently reported that

Mn2(m-dobdc) (m-dobdc4- = 4,6-dioxido-1,3-benzenedicarboxylate), a metal-organic framework (MOF)

featuring a high density of unsaturated Mn2+ sites, offers a C2H4/CO2 separation selectivity of about 10.

In the context of the OCM process scheme (cf. Figure 1), Bachman et al.[1] conclude “unique metal-

adsorbate interactions facilitated by Mn2(m-dobdc) render this material an outstanding adsorbent for the

3
capture of ethylene from the product mixture, enabling this potentially disruptive alternative process for

ethylene production”.

The primary objective of this article is examine whether any commercially available adsorbents could

offer comparable C2H4/CO2 separation capabilities, obviating the need for development of novel MOFs.

Prior to embarking on an extensive experimental campaign, we use literature data on unary isotherms,

along with the Ideal Adsorbed Solution Theory (IAST) of Myers and Prausnitz [5] to compare the

performance of Mn2(m-dobdc) with different commercially available sorbents such as cation exchanged

zeolites (13X, CaX, NaY, 5A, 4A) and activated carbon (AC). For this purpose, the required unary

isotherm data are culled from literature sources; the isotherm data, along with the details of IAST

methodology, are provided in the Supplementary material accompanying this publication. It must be

noted that the unary isotherm data for CO2, and C2H4 in AC, used in the IAST calculations are for

heterogeneous microporous activated carbon (Type BPL, 6/16 mesh, manufactured by the Pittsburgh

Chemical Company) as reported by Reich et al.[6]

In Figure 2, the IAST calculations of the adsorbed phase mole fraction of C2H4 for binary

C2H4(1)/CO2(2) mixtures, x1, is plotted as a function of the mole fraction of C2H4 in the bulk gas phase

mixture, y1. The larger the departure from the parity line, x1  y1 , the more selective is adsorption from

the mixture. The data for 5A, 4A, 13X, NaY, and CaX lie below the parity line, indicating that these

materials preferentially adsorb CO2. Zeolite 5A, the material with the largest deviation from the parity

line, has the highest selectivity to CO2, whereas the CO2/C2H4 selectivity of CaX and NaY are close to

unity. For selective adsorption of C2H4, the best performing material is Mn2(m-dobdc), in agreement

with the calculations in Bachman et al.[1] The IAST calculations for activated carbon (AC) also

displays selectivity towards C2H4.

In order to determine the reliability of IAST estimations for C2H4/CO2 mixture adsorption, as shown

in Figure 2, transient breakthrough experiments were performed. The secondary objective of this article

is to highlight some shortcomings of the adopting a sorbent screening strategy relying only on IAST

calculations of selectivities of mixture adsorption.

4
2 Transient breakthrough experiments

Transient breakthrough experiments were performed in a Flowrence set-up, described in earlier works

[7, 8], with multiple packed tubes (4 mm i.d., 300 mm height) that can be operated sequentially to test

different sorbent materials for the same feed mixture under isothermal conditions; the experimental

details are summarized in the Supplementary Material accompanying this publication. The sorbent beds

are packed with six different commercial materials: 13X (Aldrich), CaX (Siliporite), NaY (CBV 100

CY, Zeolyst), 4A (= NaA = LTA-4A, Acros), 5A (= CaA =LTA-5A, Acros), and Activated carbon

(GCN 3070 Cabot corp.), that were crushed and sieved to obtain 150-250 μm particles. At the start of

each run, the materials are dried with flow of 25 N mL/min N2 per tube at 473 K for 2 h. During the

duration of the transient adsorption and desorption process, the packed bed is maintained under

isothermal conditions at 313 K. The fixed-bed tube is first flushed, i.e. equilibrated, with pure N2 at the

specified total pressure, before injection of the feed mixture at time t = 0. The feed to each tube consists

of C2H4/CO2/N2/Ar mixtures using different C2H4/CO2 ratios; N2 forms about 58% and serves as diluent

in order to maintain nearly constant flow conditions and reduce axial dispersion. About 2% Ar in the

feed serves as inert internal tracer to monitor the start of the component breakthroughs. All the

experiments reported in this article are conducted at 1 bar absolute pressure and 313 K.

For each sorbent, four different C2H4/CO2 ratios in the feed mixture are chosen: Run 1 (C2H4/CO2 

3), Run 2 (C2H4/CO2  1.5), Run 3 (C2H4/CO2  0.8), and Run 4 (C2H4/CO2  0.5). Figures S6 – S11

provide a summary of the experimental breakthroughs for each of the six sorbents. As illustration,

Figure 3, and Figure 4 provide comparisons of the breakthroughs for Runs 1, and Run 4, respectively,

for each sorbent. For 13X, CaX, NaY, 5A, and AC, the breakthrough characteristics are reasonably

sharp, indicating that diffusional limitations are of negligible importance. It is worth noting that the

breakthroughs of C2H4, and CO2 occur at practically the same time for CaX, and NaY. For 4A zeolite,

the breakthroughs have distended characteristics, indicative of intra-crystalline diffusional limitations.

5
The breakthroughs with 5A are remarkable because in Run 1, C2H4 breaks through before CO2, whereas

in Run 4, CO2 breaks through earlier than C2H4.

3 Numerical analysis of transient breakthrough experiments

First, we seek comparisons with the IAST calculations in Figure 2 by constructing a x1 vs y1 diagram

using the methodology in earlier works [9, 10]. Let mads represent the mass of adsorbent, expressed in

kg, packed into the tube that is fed with the feed mixture at a constant flow rate of Q m3 s-1. The uptake

of C2H4, expressed as moles per kg of adsorbent in the packed tube, can be determined from a material

balance
tss
ct Q
q1 =
mads y
0
1, feed  y1,exit  dt (1)

In eq (1), ct represents the total molar concentration of the entering feed mixture at 1 bar, and 313 K.

The upper limit of the integral, tss, is the time required to reach steady-state. Analogously, the uptake of

CO2 is
tss
cQ
q2  t
mads y
0
2, feed  y2,exit  dt (2)

The integrals in eqs (1), and (2) can be numerically evaluated using a quadrature formula. In our

analysis, we found that the use of the Simpson’s rule provided results of good accuracy. Combining eqs

(1), and (2) we can determine the mole fraction of C2H4, x1 = q1  q1  q2  , essentially invoking the

assumption that the mixture can be considered to be a pseudo-binary due to the relatively poor

adsorptivity of both Ar and N2 present in the feed mixture. Figure 5 presents the results of this foregoing

numerical analysis in which the adsorbed phase mole fraction of C2H4, x1 , is plotted as a function of the

mole fraction of C2H4 in the feed mixture, y1 , treated as a pseudo-binary.

In agreement with the IAST estimations in Figure 2, the sorbents 4A, and 13X are CO2-selective, and

AC is C2H4-selective. The data for CaX, and NaY lie close to the x1  y1 parity line in Figure 5; this

6
indicates selectivities are close to unity for both these adsorbents. Indeed, examination of Figure 3, and

Figure 4 show that the breakthroughs of C2H4, and CO2 occur at practically the same time for CaX, and

NaY. In contrast to the expectations of the IAST estimates in Figure 2, the data for 5A zeolite lie close

to the parity line, indicating selectivities close to unity. Remarkably, 5A zeolite exhibits a tendency for

selectivity reversal for y1  0.5 ; the reasons for such selectivity reversal can be traced to the non-

idealities in mixture adsorption, as detailed in earlier work [8].

4 Transient breakthrough simulations

Transient breakthroughs in fixed bed adsorbers are influenced by adsorption selectivities, uptake

capacities, and intra-particle diffusional influences [10-12]. Therefore, we performed breakthrough

simulations for direct comparison with the experimental breakthroughs. For an n-component gas

mixture flowing through a fixed bed adsorber maintained under isothermal, isobaric, conditions, the

molar concentrations in the gas phase at any position and instant of time are obtained by solving the

following set of partial differential equations for each of the species i in the gas mixture [13-19]

 2 ci (t , z )  ci (t , z )   v(t , z )ci (t , z )  1     q i (t , z )
 Dax      0; i  1, 2,...n (3)
 z2 t z  t

In equation (3), t is the time, z is the distance along the adsorber,  is the bed voidage, Dax is the axial

dispersion coefficient, v is the interstitial gas velocity, and q i (t , z ) is the spatially averaged molar

loading within the crystallites of radius rc, monitored at position z, and at time t [14] . Ruthven et al.[15]

state “when mass transfer resistance is significantly greater than axial dispersion, one may neglect the

axial dispersion term and assume plug flow”. The assumption of plug flow is invoked in all the

simulation results presented in this article.

The radial distribution of molar loadings, qi, is obtained from a solution of a set of differential

equations describing the transient uptake within a spherical crystallite of radius rc

7
 qi (r , t ) 1  2

t
 2
r r
r Ni   (4)

The intra-crystalline fluxes Ni, in turn, are related to the radial gradients in the molar loadings by

qi
N i    Ði ; i  1, 2..n (5)
r

In eq (5), Ði is the Maxwell-Stefan (M-S) diffusivity for interaction of species i with the material

framework. The use of eq (5) essentially implies that we are ignoring the influence of thermodynamic

coupling effects [16-18].

At any time t, the component loadings at the surface of the particle qi (rc , t )  qi is in equilibrium with
*

*
the bulk phase gas mixture [16]; the loadings qi are determined by use of the IAST. The spatial-

averaged component loading within the crystallites of radius rc is calculated using

3 rc
q i (t ) 
rc3 
0
qi (r , t )r 2 dr (6)

At time, t = 0, the inlet to the adsorber, z = 0, is subjected to a step input of the feed gas mixture, with

inlet partial pressures pi0, and this step input is maintained till the end of the adsorption cycle when

steady-state conditions are reached.

t  0; pi (0, t )  pi 0 ; ci (0, t )  ci 0 (7)

Combination of the discretized partial differential equations along with the algebraic IAST equations

describing mixture adsorption equilibrium results in a set of differential-algebraic equations, which are

solved using a sparse matrix solver based on the semi-implicit Runge-Kutta method; further numerical

details are provided in the Supplementary material.

To match the experiments with 4A zeolite, intra-crystalline diffusional influences must be accounted

for in the breakthrough simulations. The best match was obtained by taking the values of Ði rc2 = 0.1

s-1 for Ar, N2, and CO2; For C2H4, the inter-cage hopping is significantly hindered by cations, and the

choice Ði rc2 = 0.01 s-1 yielded the best fits with experiments. For all other sorbents, intra-particle

diffusional influences may be neglected. For Run 1, and Run 4, the continuous solid black lines in
8
Figure 3, and Figure 4 represent the breakthrough simulations; more detailed comparisons for each of

the four runs with different C2H4/CO2 ratios are provided in Figures S13 – S18.

For 13X, CaX, and NaY zeolites, the breakthrough simulations show the correct qualitative features

as the experimental data. However, the match is not quantitatively perfect; the simulations tend to

predict a larger gap between the breakthrough times of C2H4 and CO2. This indicates that the unary

isotherms anticipate more CO2-selective separations than is achieved in practice.

For 5A zeolite the agreement between experimental data and breakthrough simulations is poor; the

selectivity reversal observed in Run 4 in favor of C2H4 is not anticipated by the breakthrough

simulations. In earlier work [8] it has been established that the non-idealities in mixture adsorption

equilibrium is the root cause of selectivity reversals. The selectivity reversal with 5A zeolite is also

confirmed in the breakthrough experiments of García et al. [3] for 20/80 C2H4/CO2 and 40/4/49/7

CH4/C2H6/CO2/C2H4 feed mixtures,

For AC, the transient breakthrough simulations show the same qualitative features as the experimental

data but the quantitative match is poor because unary isotherm data used in the simulations were based

on the literature data for microporous activated carbon (Type BPL, 6/16 mesh, manufactured by the

Pittsburgh Chemical Company) and the experiments were performed with Cabot carbon.

5 Transient desorption: experiments and simulations

For AC, the sorbent of choice, C2H4 is recovered in nearly purified form from extract phase of the

PSA unit. To demonstrate the feasibility of this recovery, transient desorption experiments were

conducted for each of the four runs 1, 2, 3, and 4. In the desorption experiments, the equilibrated bed is

flushed with a constant flow of pure N2, and the product compositions monitored by GC analysis. For

the sake of completeness, desorption experiments were conducted for all six sorbents, not just for AC;

the entire set of desorption experiments are presented in Figures S19-S24. As illustration, the data for

desorption for the equilibrated Run 4 are indicated by the symbols in Figure 6. The continuous solid

9
lines in Figure 6 are the simulations of the desorption transience using the IAST for calculations of

mixture adsorption equilibrium; the match with the experiments with AC are poor, as expected, because

the unary isotherms are not for the Cabot AC used in the experiments.

For 13X and 4A zeolites, the data in Figure 6a,d show that purified CO2 can be recovered during the

later stages of the transient desorption process using. The data in Figure 6b,c reinforce the conclusion,

drawn earlier, that both CaX and NaY are non-selective to either constituent; neither constituent is

available in pure form at the exit of the desorption cycle.

The desorption data for 5A shows that despite the selectivity reversal in favor of C2H4 is observed in

the adsorption cycle (cf. Figure 4e), neither of the constituent can be recovered in pure form during the

desorption cycle; see Figure 6e. Evidently, the selectivity reversal in favor of C2H4 is not strong enough

to enable pure C2H4 recovery as raffinate.

The data in Figure 6f provide confirmation of the ability of AC to produce nearly pure C2H4 for later

stages of the desorption, t > 12 minutes.

6 Conclusions

Transient breakthrough experiments with C2H4/CO2 feed mixtures in fixed beds packed with six

different sorbents 13X, CaX, NaY, 5A, 4A, and AC were carried out to determine the most suitable

adsorbent for selective adsorption of C2H4. The results indicate that AC is the most suitable adsorbent

for this purpose. Though AC exhibits a lower C2H4/CO2 selectivity than the tailor-made Mn2(m-dobdc),

its commercial availability obviates the need for materials development. Of the other sorbents 4A, and

13X are CO2-selective, and suitable for CO2 capture of refinery off-gases containing C2 hydrocarbons.

The sorbents CaX, NaY, and 5A zeolites are unable to produce purified products in either the rafffinate

(adsorption cycle) or extract (desorption cycle) phases The results obtained also indicate that sorbent

screening strategies relying on IAST calculations of mixture adsorption equilibrium, based on literature

10
data on isotherms may be misleading. For 13X, NaY, CaX, and 5A zeolites, the actual separations are

poorer than anticipated in the IAST calculations in Figure 2.

7 Acknowledgement

Rijksdienst voor Ondernemend Nederland (RVO), Topsector Energie is acknowledged for financial

support under project code TEBE117010.

Supplementary material (e-component) associated with this article can be found in the online

version, at doi:10.1016/j.seppur.XXXX.XX.XXX.

11
8 Nomenclature

Latin alphabet

ci molar concentration of species i, mol m-3

ci0 molar concentration of species i in fluid mixture at inlet, mol m-3

ct total molar concentration of gas mixture, mol m-3

Dax axial dispersion coefficient, m2 s-1

Ði Maxwell-Stefan diffusivity for molecule-wall interaction, m2 s-1

n number of species in the mixture, dimensionless

L length of packed bed adsorber, m

mads mass of adsorbent in packed tube, kg

Ni molar flux of species i with respect to framework, mol m-2 s-1

pi partial pressure of species i, Pa

pt total system pressure, Pa

Q volumetric flow rate, m3 s-1

qi molar loading species i, mol kg-1

r radial direction coordinate, m

rc radius of crystallite, m

R gas constant, 8.314 J mol-1 K-1

T absolute temperature, K

v interstitial gas velocity in packed bed, m s-1

xi mole fraction of species i in adsorbed phase, dimensionless

yi mole fraction of species i in bulk fluid mixture, dimensionless

z distance along the adsorber, m

12
Greek letters

 voidage of packed bed, dimensionless

 framework density, kg m-3

Subscripts

i,j components in mixture

t referring to total mixture

13
9 References

[1] J.E. Bachman, D.A. Reed, M.T. Kapelewski, G. Chachra, D. Jonnavittula, G. Radaelli, J.R.
Long, Enabling alternative ethylene production through its selective adsorption in the metal–
organic framework Mn2(m-dobdc), Energy Environ. Sci. 11 (2018) 2423-2431.
(https://doi.org/10.1039/c8ee01332b)
[2] G. Radaelli, G. Chachra, D. Jonnavittula, Low-Energy, Low-Cost Production of Ethylene by
Low- Temperature Oxidative Coupling of Methane, U.S. DOE/EERE/ Advanced Manufacturing
Office (AMO), https://www.osti.gov/servlets/purl/1414280,
[3] L. García, Y.A. Poveda, G. Rodríguez, E. Esche, H.R. Godini, G. Wozny, J.-U. Repke, A.
Orjuela, Adsorption Separation of Oxidative Coupling of Methane Effluent Gases. Miniplant
Scale Experiments and Modeling, J. Nat. Gas Sci. Eng. 61 (2019) 106-118.
(https://doi.org/10.1016/j.jngse.2018.11.007)
[4] C.M. Van der Meijden, Development of the MILENA Gasification Technology for the
Production of Bio-SNG, Ph.D. Thesis, Ph.D. Dissertation, Eindhovent University of
Technology, Eindhoven, 2010.
[5] A.L. Myers, J.M. Prausnitz, Thermodynamics of Mixed Gas Adsorption, A.I.Ch.E.J. 11 (1965)
121-130.
[6] R. Reich, W.T. Zlegler, K.A. Rogers, Adsorption of Methane, Ethane, and Ethylene Gases and
Their Binary and Ternary Mixtures and Carbon Dioxide on Activated Carbon at 212-301 K and
Pressures to 35 Atmospheres, Ind. Eng. Chem. Process Des. Dev. 19 (1980) 336-344.
[7] I. van Zandvoort, G.P.M. van Klink, E. de Jong, J.C. van der Waal, Selectivity and Stability of
Zeolites [Ca]A and [Ag]A towards Ethylene Adsorption and Desorption from Complex Gas
Mixtures, Microporous Mesoporous Mat. 263 (2018) 142-149.
[8] I. van Zandvoort, J.K. van der Waal, E.-J. Ras, R. de Graaf, R. Krishna, Highlighting non-
idealities in C2H4/CO2 mixture adsorption in 5A zeolite, Sep. Purif. Technol. 227 (2019) 115730.
(https://doi.org/10.1016/j.seppur.2019.115730)
[9] H. Yu, X. Wang, C. Xu, D.-L. Chen, W. Zhu, R. Krishna, Utilizing transient breakthroughs for
evaluating the potential of Kureha carbon for CO2 capture, Chem. Eng. J. 269 (2015) 135-147.
[10] R. Krishna, Methodologies for Evaluation of Metal-Organic Frameworks in Separation
Applications, RSC Adv. 5 (2015) 52269-52295.
[11] R. Krishna, Screening Metal-Organic Frameworks for Mixture Separations in Fixed-Bed
Adsorbers using a Combined Selectivity/Capacity Metric, RSC Adv. 7 (2017) 35724-35737.
(https://doi.org/10.1039/C7RA07363A)
[12] R. Krishna, Methodologies for Screening and Selection of Crystalline Microporous Materials in
Mixture Separations, Sep. Purif. Technol. 194 (2018) 281-300.
(https://doi.org/10.1016/j.seppur.2017.11.056)
[13] R. Krishna, R. Baur, Modelling Issues in Zeolite Based Separation Processes, Sep. Purif.
Technol. 33 (2003) 213-254.
[14] R. Krishna, The Maxwell-Stefan Description of Mixture Diffusion in Nanoporous Crystalline
Materials, Microporous Mesoporous Mater. 185 (2014) 30-50.
[15] D.M. Ruthven, S. Farooq, K.S. Knaebel, Pressure swing adsorption, VCH Publishers, New
York, 1994.
[16] R. Krishna, Tracing the Origins of Transient Overshoots for Binary Mixture Diffusion in
Microporous Crystalline Materials, Phys. Chem. Chem. Phys. 18 (2016) 15482-15495.

14
[17] R. Krishna, Highlighting the Influence of Thermodynamic Coupling on Kinetic Separations with
Microporous Crystalline Materials, ACS Omega 4 (2019) 3409-3419.
(https://doi.org/10.1021/acsomega.8b03480)
[18] R. Krishna, Maxwell-Stefan Modelling of Mixture Desorption Kinetics in Microporous
Crystalline Materials, Sep. Purif. Technol. 229 (2019) 115790.
(https://doi.org/10.1016/j.seppur.2019.115790)
[19] R. Krishna, J.R. Long, Screening metal-organic frameworks by analysis of transient
breakthrough of gas mixtures in a fixed bed adsorber, J. Phys. Chem. C 115 (2011) 12941-
12950.

15
10 List of Figures

Figure 1. OCM process flow scheme with C2H4-selective adsorbent in the PSA adsorber beds; this

flow scheme is adapted from Radaelli et al.[2]

Figure 2. IAST calculations for adsorption of binary C2H4(1)/CO2(2) mixtures in different sorbents at

313 K and total pressure of 1 bar. The adsorbed phase mole fraction of C2H4, x1, is plotted as a function

of the mole fraction of C2H4 in the bulk gas phase mixture, y1. See Supplementary Material for

information on the unary isotherm fit parameters and details of IAST calculations.

Figure 3. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in (a) 13 X, (b) CaX, (c) NaY,

(d) 4A, (e) 5A, and (f) AC. The total pressure is 1 bar, and temperature T = 313 K. The feed mixture

composition corresponds to Run 1 (C2H4/CO2  3). The continuous solid black lines are breakthrough

simulations using the IAST for determination of mixture adsorption equilibrium. The % N2 in the outlet

gas can be determined by the taking the sum of the mole % = 100.

Figure 4. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in (a) 13 X, (b) CaX, (c) NaY,

(d) 4A, (e) 5A, and (f) AC. The total pressure is 1 bar, and temperature T = 313 K. The feed mixture

composition corresponds to Run 4 (C2H4/CO2  0.5). The continuous solid black lines are breakthrough

simulations using the IAST for determination of mixture adsorption equilibrium. The % N2 in the outlet

gas can be determined by the taking the sum of the mole % = 100.

16
 Figure 5. Analysis of the experimental breakthroughs with different sorbents. The adsorbed phase

mole fraction of C2H4, x1, is plotted as a function of the mole fraction of C2H4 in the bulk gas phase

mixture, y1. The data for 5A zeolite includes the entire data set obtained at 1 bar and reported in earlier

work [8].

Figure 6. Transient desorption data for the equilibrated bed in Run 4 for (a) 13 X, (b) CaX, (c) NaY,

(d) 4A, (e) 5A, and (f) AC. The total pressure is 1 bar, and temperature T = 313 K. The continuous solid

black lines are desorption simulations using the IAST for determination of mixture adsorption

equilibrium. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

17
 Figure 1

H2/CH4/C2H6/CO/CO2

Methanation

CH4+O2

Purified
C2H4

OCM C2H6 C2H6 C2H4/C2H6

Distillation
C2H6
Raffinate:
purge
H2/CH4/C2H6/
CO/CO2 C2H4/
Drying C2H6
PSA

CO2 removal
CO2

H2/CH4/CO/CO2/ Extract:
C2H4/C2H6 C2H4+(CO2,C2H6)
 Figure 2

C2H4(1)/CO2(2)
1.0
mole fraction of C2H4 in adsorbed phase, x1

40 kPa; 313 K; IAST

0.8

y1=x1
0.6
Mn2(m-dobdc)
AC
0.4 CaX
NaY
13X
0.2
4A
5A
0.0
0.0 0.2 0.4 0.6 0.8 1.0

mole fraction of C2H4 in bulk gas phase, y1

 Figure 3
a b
25 Ar/C2H4/CO2/N2; 20 C2H4
13X; 313 K, CO2
% component in outlet gas

% component in outlet gas

20 100 kPa Ar
15 simulation
C2H4
15 CO2
Ar 10 Ar/C2H4/CO2/N2;
10 simulation CaX; 313 K,
100 kPa
5
5

0 0
0 5 10 15 20 25 30 0 5 10 15 20

time, t / min
c d time, t / min

C2H4
20
CO2
20
Ar
% component in outlet gas

% component in outlet gas

C2H4
simulation 15 CO2
15
Ar
simulation
Ar/C2H4/CO2/N2; 10
10
NaY; 313 K,
Ar/C2H4/CO2/N2;
100 kPa
LTA-4A; 313 K, 100 kPa
5 5

0 0
0 5 10 15 20 25 30 0 10 20 30 40

e time, t / min
f time, t / min

25
20 C2H4
CO2
20
% component in outlet gas

% component in outlet gas

C2H4 Ar
15 simulation
CO2
15 Ar/C2H4/CO2/N2;
Ar
AC; 313 K,
simulations 10 100 kPa
10 N2/Ar/CO2/C2H4
LTA-5A; 313 K, 100 kPa
5
5

0 0
0 10 20 30 40 50 0 5 10 15 20 25

time, t / min time, t / min

 Figure 4
a Ar/C2H4/CO2/N2;
b 30

13X; 313 K, C2H4

100 kPa 25 CO2
30
% component in outlet gas

% component in outlet gas

Ar
20 simulation

20 15
C2H4
Ar/C2H4/CO2/N2;
CO2
10 CaX; 313 K,
Ar
10 100 kPa
simulation
5

0 0
0 5 10 15 20 0 2 4 6 8 10 12 14 16

time, t / min
c d 30
time, t / min

C2H4
30
CO2 25
% component in outlet gas

% component in outlet gas

Ar C2H4
simulation 20 CO2
20 Ar
15 simulation
Ar/C2H4/CO2/N2;
NaY; 313 K,
10
10 100 kPa
Ar/C2H4/CO2/N2;
5 LTA-4A; 313 K, 100 kPa

0 0
0 2 4 6 8 10 12 0 5 10 15 20 25 30

e time, t / min
f time, t / min

C2H4
30 CO2 30 C2H4
% component in outlet gas

% component in outlet gas

Ar CO2
Ar, simulation
Ar
20 CO2 simulation
simulation
20
C2H4 simulation
Ar/C2H4/CO2/N2;
N2/Ar/CO2/C2H4
AC; 313 K,
10 5A; 313 K, 10 100 kPa
100 kPa

0 0
0 5 10 15 20 25 0 5 10 15 20

time, t / min time, t / min

 Figure 5

1.0 C2H4(1)/CO2(2)
mole fraction of C2H4 in adsorbed phase, x1

313 K; from
breakthroughs
0.8

y1=x1
0.6 AC
5A
NaY
0.4 CaX
13X
4A
0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0

mole fraction of C2H4 in bulk gas phase, y1

 Figure 6
a 100
b 100
CO2
C2H4 Run 4
CaX; 313 K,
% component in outlet gas

% component in outlet gas

10 simulation 100 kPa
10

1
0.1 CO2
Run 4
C2H4
13X; 313 K,
100 kPa simulation
0.01 0.1
0 10 20 30 40 0 5 10 15 20 25 30

time, t / min time, t / min

c 100 d 100
CO2 CO2
C2H4 C2H4
simulation
% component in outlet gas

% component in outlet gas

10 simulation
10

1 Run 4
4A; 313 K,
100 kPa
0.1 Run 4
NaY; 313 K,
100 kPa 0.1
0.01
0 5 10 15 20 25 0 10 20 30 40 50

time, t / min time, t / min

e 100 f 100
CO2
Run 4
5A; 313 K, C2H4
% component in outlet gas

% component in outlet gas

100 kPa 10 simulation

10

1 CO2 Run 4
C2H4 0.1 AC; 313 K,
100 kPa
C2H4 simulation
CO2 simulation
0.1 0.01
0 5 10 15 20 25 0 5 10 15 20 25 30

time, t / min time, t / min

Using Transient Breakthrough Experiments for
Screening of Adsorbents for Separation of
C2H4/CO2 Mixtures

Ilona van Zandvoort, Erik-Jan Ras, Robbert de Graaf

Avantium Chemicals B.V.
Zekeringstraat 29
1014 BV Amsterdam, The Netherlands

and

Rajamani Krishna*
Van ‘t Hoff Institute for Molecular Sciences
University of Amsterdam
Science Park 904
1098 XH Amsterdam, The Netherlands
email: [email protected]

S1
Table of Contents
1 Preamble ......................................................................................................................... 4
2 Unary Isotherms............................................................................................................. 5
2.1 Isotherm fits for LTA-5A zeolite ................................................................................................. 5

2.2 Isotherm fits for 13X zeolite ........................................................................................................ 5

2.3 Isotherm fits for CaX zeolite ........................................................................................................ 6

2.4 Isotherm fits for NaY zeolite ....................................................................................................... 6

2.5 Isotherm fits for LTA-4A zeolite ................................................................................................. 7

2.6 Isotherm fits for Activated Carbon .............................................................................................. 7

2.7 Isotherm fits for Mn2(m-dobdc) ................................................................................................... 8

2.8 List of Tables for Unary Isotherms .............................................................................................. 9

3 IAST calculations of mixture adsorption .................................................................. 16

3.1 Brief outline of theory ................................................................................................................ 16

3.2 IAST calculations for C2H4/CO2 mixture adsorption equilibrium ............................................. 18

3.3 List of Figures for IAST calculations of mixture adsorption ..................................................... 20

4 Methodology for transient breakthrough simulations ............................................. 22

4.1 List of Figures for Methodology for transient breakthrough simulations.................................. 25

5 Transient breakthrough experiments vs simulations ............................................... 27

5.1 The Flowrence set-up for transient breakthrough experiments ................................................. 27

5.2 Experimental breakthrough campaigns ...................................................................................... 28

5.3 Numerical analysis of experimental breakthroughs ................................................................... 29

5.4 Experimental breakthroughs versus simulations........................................................................ 30

5.5 List of Figures for Transient breakthrough experiments vs simulations ................................... 32

6 Transient desorption: experiments vs simulations ................................................... 47

6.1 List of Figures for Transient desorption: experiments vs simulations ....................................... 48

S2
7 Nomenclature ............................................................................................................... 54
8 References ..................................................................................................................... 57

S3
Preamble

1 Preamble

This Supporting Information accompanying the article Using Transient Breakthrough Experiments for

Screening of Adsorbents for Separation of C2H4/CO2 Mixtures provides:

(a) Unary isotherm data sources, and fits for various sorbents

(b) Brief summary of the IAST for calculation of mixture adsorption equilibrium

(c) Methodology used for transient breakthroughs in fixed bed adsorbers

(d) Flowrence set-up, and experimental breakthrough comparisons

(e) Detailed comparisons of experimental breakthroughs with transient breakthrough simulations

For ease of reading, the Supplementary Material is written as a stand-alone document.

S4
Unary Isotherms

2 Unary Isotherms

2.1 Isotherm fits for LTA-5A zeolite

The unary isotherm data for C2H4 in LTA-5A zeolite as reported in Table 2 and Table 3 of Mofarahi

and Salehi1 at temperatures of 283 K, 303 K, and 323 K were fitted with the dual-site Langmuir model

bA p b p
q  q A,sat  qB ,sat B (S1)
1  bA p 1  bB p

with T-dependent parameters bA, and bB

E  E 
bA  bA0 exp A ; bB  bB 0 exp B  (S2)
 RT   RT 

The fitted parameter values are presented in Table S1.

The unary isotherm data for CO2 in LTA-5A zeolite as reported in Table A1 of Mofarahi and

Gholipour2 at temperatures of 273 K, 283 K, 303 K, 323 K, and 343 K were fitted with the dual-site

Langmuir model; the parameter fits are presented in Table S2.

The unary isotherm data for N2 in LTA-5A zeolite, as reported in Table 3 and Table 4 of Bakhtyari

and Mofarahi3 at temperatures of 273 K, 283 K, 303 K, 323 K, and 343 K were fitted with excellent

accuracy with the single-site Langmuir model

bp  E 
q  qsat ; b  b0 exp  (S3)
1  bp  RT 

The fitted parameter values are presented in Table S3.

In the breakthrough simulations, the isotherm fits for N2 in LTA-5A zeolite were assumed to also hold

for Ar.

2.2 Isotherm fits for 13X zeolite

The unary isotherm data for CO2, and C2H4 in 13X (= NaX) zeolite at temperatures of 279 K, 293 K,

and 308 K are reported in Table I of Costa et al.4 Their experimental data were fitted with the dual-site

S5
Unary Isotherms

Langmuir model with T- dependent parameters, as described by eqs (S1), and (S2); the fit parameters

are reported in Table S4.

The unary isotherm data for N2 in 13 X zeolite at temperatures of 298 K, 308 K, and 323 K are

reported in Table 3 of Cavenati et al.5 Their experimental data were fitted with the dual-site Langmuir

model with T-dependent parameters, as described by eqs (S1), and (S2); the fit parameters are reported

in Table S5. The idea of using T-dependent isotherm fits is that extrapolation to 313 K can be done for

IAST calculations of selectivities and breakthroughs.

In the breakthrough simulations, the isotherm fits for N2 in 13X zeolite were assumed to also hold for

Ar.

2.3 Isotherm fits for CaX zeolite

The unary isotherm fit parameters for CO2, and C2H4 in CaX zeolite at temperatures of 298.15 K,

308.15 K, and 318.15 K are reported in Supplementary Tables 9, and 10 of Bachman et al.6 Their fit

parameters, however, do not seem to be consistentwith the raw experimental data at 298.15 plotted in

Figure 2. Therefore, the unary fit parameters were refitted in order to match the raw experimental data.

The re-fitted parameters, using the 1-site Langmuir model with T-dependent parameters; these are

reported in Table S6.

In the breakthrough simulations, the isotherm fits for N2 and Ar are assumed to be the same as for

13X zeolite.

2.4 Isotherm fits for NaY zeolite

The unary isotherm data for CO2, and C2H4 in NaY zeolite at temperatures of 305 K, and 353 K are

presented in Figure 2 of Choudhary et al.7 These data were re-fitted using the 1-site Langmur model

with T-dependent parameters; the fit parameters are reported in Table S7 The idea of using T-dependent

isotherm fits is that interpolation to 313 K can be done for IAST calculations of selectivities and

breakthroughs.

S6
Unary Isotherms

In the breakthrough simulations, the isotherm fits for N2 and Ar are assumed to be the same as for

13X zeolite.

2.5 Isotherm fits for LTA-4A zeolite

The unary isotherm fit parameters for CO2, and C2H4 in LTA-4A (=CaA) zeolite at temperatures of

293.15 K, and 323.15 K are reported in Table 4 of Romero-Pérez and Aguilar-Armenta.8 Their

tabulated data for CO2 were re-fitted with the dual-site Langmuir model with T-dependent parameters,

as described by eqs (S1), and (S2); the fit parameters are reported in Table S8.

For C2H4 , their data were re-fitted with the 1-site Langmuir model with T-dependent parameters:

bp  E 
q  qsat 
; b  b0 exp   (S4)
1  bp  RT 

The fit parameters are reported in Table S9.

In the breakthrough simulations, the isotherm fits for N2 and Ar are assumed to be the same as for

LTA-5A zeolite.

2.6 Isotherm fits for Activated Carbon

The unary isotherm data for CO2, and C2H4 for heterogeneous microporous activated carbon (Type

BPL, 6/16 mesh, manufactured by the Pittsburgh Chemical Company) are reported by Reich et al.9 The

reported data at temperatures of 212.72 K, 260.2 K, and 301.4 K, reported in their Table 1, were fitted

with the dual-site Langmuir model with T-dependent parameters, as described by eqs (S1), and (S2); the

fit parameters are reported in Table S10. The idea of using T-dependent isotherm fits is that

extrapolation to 313 K can be done for IAST calculations of selectivities and breakthroughs.

The unary isotherm data for N2, and Ar for cylindrical activated carbon supplied by Kuraray Chemical

company (Coal-derived activated carbon; 2GA-H2J) by Park et al.10 The reported data at temperatures

of 293 K, 308 K, and 323 K, in their Tables 9 and 11, were fitted with the 1-site Langmuir model with

T-dependent parameters; the fit parameters are reported in Table S10.

S7
Unary Isotherms

2.7 Isotherm fits for Mn2(m-dobdc)

The unary isotherm fit parameters for CO2, and C2H4 in Mn2(m-dobdc) (m-dobdc4- = 4,6-dioxido-1,3-

benzenedicarboxylate) at temperatures of 298.15 K, 308.15 K, and 318.15 K are reported in

Supplementary Tables 1, and 2 of Bachman et al.6 These data were re-fitted with the dual-site Langmuir

model with T- dependent parameters, as described by eqs (S1), and (S2); the fit parameters are reported

in Table S11.

S8
Unary Isotherms

2.8 List of Tables for Unary Isotherms

Table S1. Dual-site Langmuir parameter fits for C2H4 in LTA-5A zeolite. These parameters are based

on the unary isotherm data reported in Table 2 and Table 3 of Mofarahi and Salehi1 at temperatures of

283 K, 303 K, and 323 K.

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

C2H4 2.5 6.98E-08 19 0.75 4.18E-18 67

Table S2. Dual-site Langmuir parameter fits for CO2 in LTA-5A zeolite. These parameters are based

on the unary isotherm data reported in Table A1 of Mofarahi and Gholipour2 at temperatures of 273 K,

283 K, 303 K, 323 K, and 343 K.

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

CO2 1.5 4.5E-10 23.5 2.5 2.99E-12 49

S9
Unary Isotherms

Table S3. Single-site Langmuir parameter fits for N2 in LTA-5A zeolite. These parameters are based

on the unary isotherm data reported in Table 3 and Table 4 of Bakhtyari and Mofarahi3 at temperatures

of 273 K, 283 K, 303 K, 323 K, and 343 K.

qsat b0 E
mol kg-1 Pa 1 kJ mol-1

N2 2.5 1.68E-09 16.6

S10
Unary Isotherms

Table S4. Dual-site Langmuir parameter fits for CO2 and C2H4 in 13X zeolite. These parameters are

based on the unary isotherm data at temperatures of 279 K, 293 K, and 308 K as reported in Table I of

Costa et al.4

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

CO2 2.0 4.813E-08 23 2.5 1.129E-12 42

C2H4 1.35 3.403E-05 4.6 1.4 8.508E-17 70

Table S5. Dual-site Langmuir parameter fits for N2 in 13X zeolite. These parameters are based on the

unary isotherm data at temperatures of 298 K, 308 K, and 323 K as reported in Table 3 of Cavenati et

al.5

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

N2 3.0 4.075E-09 13 6.0 4.681E-10 13

S11
Unary Isotherms

Table S6. 1-site Langmuir parameter fits for CO2 and C2H4 in CaX zeolite. These re-fitted parameters

are based on the data of Bachman et al.6

qsat b0 E
mol kg-1 Pa 1 kJ mol-1

CO2 3.5 2.0E-08 22

C2H4 5.4 3.6E-09 25

Table S7. 1-site Langmuir parameter fits for CO2 and C2H4 in NaY zeolite. These parameters are

based on the unary isotherm data from Figure 2 of Choudhary et al.7

qsat b0 E
mol kg-1 Pa 1 kJ mol-1

CO2 6.3 6.36E-10 30

C2H4 4.6 4.967E-09 25.3

S12
Unary Isotherms

Table S8. Dual-site Langmuir parameter fits for CO2 in LTA-4A zeolite. The T-dependent parameters

were obtained by re-fitting the data reported in Table 4 of Romero-Pérez and Aguilar-Armenta.8

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

CO2 1.3 6.075E-12 40 2.0 6.722E-12 50

Table S9. 1-site Langmuir-Freundlich parameter fits for C2H4 in LTA-4A zeolite. The T-dependent

parameters were obtained by re-fitting the data reported in Table 4 of Romero-Pérez and Aguilar-

Armenta.8

bp  E 
q  qsat 
; b  b0 exp  
1  bp  RT 

qsat b0 E 
mol kg-1 Pa  kJ mol-1 dimensionless

C2H4 2.3 6.287E-11 35 1.26

S13
Unary Isotherms

Table S10. Dual-site Langmuir parameter fits for CO2 and C2H4 in Activated Carbon (BPL), obtained

by fitting of the experimental data of Reich et al.9 Single-site Langmuir parameter fits for N2 and Ar in

Activated Carbon (Kuraray), obtained by fitting of the experimental data of Park et al.10

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

CO2 3.5 5.617E-10 22.5 7.6 5.893E-11 22.6

C2H4 3.6 1.299E-09 24 4.4 9.630E-11 21.5

N2 3.2 1.768E-09 16.5

Ar 4.35 1.681E-09 15.2

S14
Unary Isotherms

Table S11. Dual-site Langmuir parameter fits for CO2, and C2H4 in Mn2(m-dobdc), obtained by re-

fitting of the data reported in Supplementary Tables 1, and 2 of Bachman et al.6

Site A Site B

qA,sat bA0 EA qB,sat bB0 EB

mol kg-1 Pa 1 kJ mol-1 mol kg-1 Pa 1 kJ mol-1

CO2 13.4 2.031E-11 27 6.2 7.627E-11 34

C2H4 2.45 3.435E-11 47 4.0 1.619E-11 42

S15
IAST calculations of mixture adsorption

3 IAST calculations of mixture adsorption

3.1 Brief outline of theory

Within microporous crystalline materials, the guest molecules exist in the adsorbed phase. The Gibbs

adsorption equation11 in differential form is

n
Ad   qi di (S5)
i 1

The quantity A is the surface area per kg of framework, with units of m2 per kg of the framework of

the crystalline material; qi is the molar loading of component i in the adsorbed phase with units moles

per kg of framework; i is the molar chemical potential of component i. The spreading pressure  has

the same units as surface tension, i.e. N m-1.

The chemical potential of any component in the adsorbed phase, i, equals that in the bulk fluid phase.

If the partial fugacities in the bulk fluid phase are fi, we have

di  RTd ln f i (S6)

where R is the gas constant (= 8.314 J mol-1 K-1).

Briefly, the basic equation of Ideal Adsorbed Solution Theory (IAST) theory of Myers and

Prausnitz12 is the analogue of Raoult’s law for vapor-liquid equilibrium, i.e.

f i  Pi 0 xi ; i  1,2,...n (S7)

where xi is the mole fraction in the adsorbed phase

qi
xi  (S8)
q1  q2  ...qn

and Pi 0 is the pressure for sorption of every component i, which yields the same spreading pressure, 

for each of the pure components, as that for the mixture:

S16
IAST calculations of mixture adsorption

P10 P0 P0
A q0 ( f ) 2
q0 ( f ) 3
q0 ( f )
 1 df   2 df   3 df ... (S9)
RT 0 f 0
f 0
f

A
where qi0 ( f ) is the pure component adsorption isotherm. The units of , also called the adsorption
RT

potential,13 are mol kg-1.

The unary isotherm may be described by say the 1-site Langmuir isotherm

bf bf
q 0  f   q sat ;  (S10)
1  bf 1  bf

where we define the fractional occupancy of the adsorbate molecules,   q0  f  qsat . The superscript 0

is used to emphasize that q 0  f  relates the pure component loading to the bulk fluid fugacity. More

generally, the unary isotherms may need to be described by, say, the dual-site Langmuir model

bA f b f
q 0 ( f )  q A, sat  qB , sat B (S11)
1  bA f 1  bB

Each of the integrals in Equation (S9) can be evaluated analytically:

Pi0
q0 ( f )
 
f 0 f df  qA,sat ln 1  bA  Pi   qB,sat ln 1  bB  Pi 
0 0
 
Pi0
(S12)
q0 ( f )   f    fi 
f 0 f df  qA,sat ln 1  bA  xii    qB , sat ln 1  bB 
  xi
 

 

The right hand side of equation (S12) is a function of Pi 0 . For multicomponent mixture adsorption,

each of the equalities on the right hand side of Equation (S9) must be satisfied. These constraints may

be solved using a suitable equation solver, to yield the set of values of P10 , P20 , P30 ,.. Pn0 , all of which

satisfy Equation (S9). The corresponding values of the integrals using these as upper limits of

A
integration must yield the same value of for each component; this ensures that the obtained solution
RT

is the correct one.

The adsorbed phase mole fractions xi are then determined from

S17
IAST calculations of mixture adsorption

fi
xi  ; i  1,2,...n (S13)
Pi 0

A key assumption of the IAST is that the enthalpies and surface areas of the adsorbed molecules do

not change upon mixing. If the total mixture loading is qt , the area covered by the adsorbed mixture is

A
with units of m2 (mol mixture)-1. Therefore, the assumption of no surface area change due to
qt

A Ax Ax Ax
mixture adsorption translates as  0 10  0 20   0 n0 ; the total mixture loading is qt is
qt q1 P1 
q2 P2  
qn Pn  
calculated from

1
qt  q1  q 2 ...  q n 
x1 x x (S14)
0 0
 0 2 0  ....  0 n 0
q ( P1 ) q 2 ( P2 )
1 q n ( Pn )

in which q10 ( P10 ) , q20 ( P20 ) ,… qn0 ( Pn0 ) are determined from the unary isotherm fits, using the sorption

pressures for each component P10 , P20 , P30 ,.. Pn0 that are available from the solutions to equations

Equations (S9), and (S12).

The entire set of equations (S7) to (S14) need to be solved numerically to obtain the loadings, qi of the

individual components in the mixture.

3.2 IAST calculations for C2H4/CO2 mixture adsorption equilibrium

In order to gauge the potential of different adsorbents for separating C2H4(1)/CO2(2) mixtures we use

the IAST calculations. Figure S1a presents IAST calculations for adsorption of binary C2H4(1)/CO2(2)

mixtures in different sorbents at 313 K and total pressure of 40 kPa, wherein the adsorbed phase mole

fraction of C2H4 (y-axis) is plotted as a function of the mole fraction of C2H4 in the bulk gas phase

mixture (x-axis). The dotted line represents the parity line, indicating no selective separation potential.

The larger the departure from the x1  y1 , parity line, the more selective is mixture adsorption. Figure

S1b presents calculations for the adsorption selectivity, Sads, defined by

S18
IAST calculations of mixture adsorption

q1 q2
S ads  (S15)
y1 y2

where q1 and q2 are the molar loadings of the components 1, and 2 in the adsorbed phase in equilibrium

with a bulk gas phase mixture with mole fractions y1 and y2.

The data for 5A, 4A, 13X, NaY, and CaX lie below the x1  y1 parity line, indicating that these

materials preferentially adsorb CO2. LTA-5A, the material with the largest deviation from the x1  y1

parity line, has the highest selectivity to CO2. The data for CaX, and NaY lie very near the x1  y1

parity line, and the preferential selectivity toward CO2 is only marginal. For selective adsorption of

C2H4, the best performing material is Mn2(m-dobdc), as was also claimed in the paper by Bachman et

al.6 The IAST calculations for activated carbon (AC) also displays selectivity towards C2H4.

In order to determine the reliability of IAST predictions for C2H4/CO2 mixture adsorption, we resort

to transient breakthrough experiments, buttressed by transient breakthrough simulations incorporating

the IAST.

S19
IAST calculations of mixture adsorption

3.3 List of Figures for IAST calculations of mixture adsorption

a
1.0
mole fraction of C2H4 in adsorbed phase, x1
0.8

0.6

0.4

0.2 y1=x1
Mn2(m-dobdc)
AC
0.0 NaY
CaX
C2H4(1)/CO2(2)
13X
40 kPa; 313 K; IAST 4A
b 5A

10
C2H4/CO2 adsorption selectivity

0.1
0.0 0.2 0.4 0.6 0.8 1.0

mole fraction of C2H4 in bulk gas phase, y1

Figure S1. IAST calculations for adsorption of binary C2H4(1)/CO2(2) mixtures in different sorbents

at 313 K and total pressure of 40 kPa. (a) The adsorbed phase mole fraction of C2H4, x1, is plotted as a

function of the mole fraction of C2H4 in the bulk gas phase mixture, y1. (b) C2H4(1)/CO2(2) adsorption

selectivity is plotted as a function of the y1.

S20
IAST calculations of mixture adsorption

S21
Methodology for transient breakthrough simulations

4 Methodology for transient breakthrough simulations

We describe below the simulation methodology used to perform transient breakthrough calculations

for fixed bed adsorbers (see schematics in Figure S2, and Figure S3). The simulation methodology is the

same as used in our earlier publications.14-17 For an n-component gas mixture flowing through a fixed

bed maintained under isothermal, isobaric, conditions, the molar concentrations in the gas phase at any

position and instant of time are obtained by solving the following set of partial differential equations for

each of the species i in the gas mixture16, 18-20

 2 ci (t , z )  ci (t , z )   v(t , z )ci (t , z )  1     q i (t , z )
 Dax      0; i  1, 2,...n (S16)
 z2 t z  t

In equation (S16), t is the time, z is the distance along the adsorber,  is the framework density,  is

the bed voidage, Dax is the axial dispersion coefficient, v is the interstitial gas velocity, and q i (t , z ) is

the spatially averaged molar loading within the crystallites of radius rc, monitored at position z, and at

time t. The time t = 0, corresponds to the time at which the feed mixture is injected at the inlet to the

fixed bed. Prior to injection of the feed mixture, N2 gas flows through the fixed bed. In this model

described by equation (S16), the effects of all mechanisms that contribute to axial mixing are lumped

into a single effect axial dispersion coefficient Dax . Ruthven et al.20 state that more detailed models that

include radial dispersion are generally not necessary. They also make the following remark “when mass

transfer resistance is significantly greater than axial dispersion, one may neglect the axial dispersion

term and assume plug flow”. All of the analysis and breakthrough simulations were carried out using the

plug flow assumption.

The radial distribution of molar loadings, qi, within a spherical crystallite, of radius rc, is obtained

from a solution of a set of differential equations describing the uptake

qi (r , t ) 1  2

t
 2
r r
r Ni  (S17)

The intra-crystalline fluxes Ni, in turn, are related to the radial gradients in the molar loadings by

S22
Methodology for transient breakthrough simulations

qi
Ni    Ði ; i  1, 2, n (S18)
r

In eq (S17) Ði is the Maxwell-Stefan (M-S) diffusivity for interaction of species i with the material

framework.

For all times t ≥ 0, the exterior of the crystal is brought into contact with a bulk gas mixture at partial

pressures pi 0 that is maintained constant till the crystal reaches thermodynamic equilibrium with the

surrounding gas mixture. At any time t, the component loadings at the surface of the particle

qi (rc , t )  qi* is in equilibrium with the bulk phase gas mixture with partial pressures pi 0 . In the general

case, the component loadings are calculated using the Ideal Adsorbed Solution Theory (IAST) of Myers

and Prausnitz.12

At any time t, during the transient approach to thermodynamic equilibrium, the spatial-averaged

component loading within the crystallites of radius rc is calculated using

3 rc
q i (t ) 
rc3 0
qi (r , t )r 2 dr (S19)

Summing equation (S19) over all n species in the mixture allows calculation of the total average

molar loading of the mixture within the crystallite

n
q t (t , z )   q i (t , z ) (S20)
i 1

 q i (t , z )
The term in equation (S16) is determined by solving the set of equations (S17), and (S19),
t

and (S20). At any time t, and position z, the component loadings at the outer surface of the particle

qi (rc , t , z ) is in equilibrium with the bulk phase gas mixture with partial pressures pi (t , z ) in the bulk

gas mixture. In the general case, the component loadings qi (rc , t , z ) are calculated using the Ideal

Adsorbed Solution Theory (IAST) of Myers and Prausnitz.12

S23
Methodology for transient breakthrough simulations

Ði
If the value of is large enough to ensure that intra-crystalline gradients are absent and the entire
rc2

crystallite particle can be considered to be in thermodynamic equilibrium with the surrounding bulk gas

phase at that time t, and position z of the adsorber

q i (t , z )  qi (t , z ) (S21)

The interstitial gas velocity is related to the superficial gas velocity by

u
v (S22)


At time, t = 0, the inlet to the adsorber, z = 0, is subjected to a step input of the n-component gas

mixture and this step input is maintained till the end of the adsorption cycle when steady-state

conditions are reached.

t  0; pi (0, t )  pi 0 ; u (0, t )  u0 (S23)

where u0  v0 is the superficial gas velocity at the inlet to the adsorber.

Typically, the adsorber length is divided into 100 slices, and each spherical crystallite was discretized

into 20 equi-volume slices. The results thus obtained were confirmed to be of adequate accuracy.

Combination of the discretized partial differential equations (PDEs) along with the algebraic equations

describing mixture adsorption equilibrium (IAST, or mixed-gas Langmuir model, as appropriate),

results in a set of differential-algebraic equations (DAEs), which are solved using BESIRK.21 BESIRK

is a sparse matrix solver, based on the semi-implicit Runge-Kutta method originally developed by

Michelsen,22 and extended with the Bulirsch-Stoer extrapolation method.23 Use of BESIRK improves

the numerical solution efficiency in solving the set of DAEs. The evaluation of the sparse Jacobian

required in the numerical algorithm is largely based on analytic expressions.18 Further details of the

numerical procedures used in this work, are provided by Krishna and co-workers;18, 19, 24, 25 interested

readers are referred to our website that contains the numerical details.19

S24
Methodology for transient breakthrough simulations

4.1 List of Figures for Methodology for transient breakthrough simulations

L = length of packed bed

Entering fluid mixture

Exiting fluid mixture

distance along bed, z

Two different discretization strategies for adsorbent particle

r=0 r = rc r=0 r = rc

Equi-volume slices Equi-distant slices

Figure S2. Two different discretization schemes for a single spherical crystallite.

S25
Methodology for transient breakthrough simulations

L = length of packed bed

Entering gas mixture

Exiting gas mixture

100-200 slices

distance along bed, z

Equi-volume grid within
crystal

20-150 slices

Figure S3. Discretization scheme for fixed bed adsorber.

S26
Transient breakthrough experiments vs simulations

5 Transient breakthrough experiments vs simulations

5.1 The Flowrence set-up for transient breakthrough experiments

A Flowrence set-up is modified for the transient breakthrough experiments (Figure S4) such that the

feed selector valve selects one reactor (= fixed-bed adsorber) tube, which is fed with the adsorption gas

mixture. Meanwhile, all other reactors are fed with nitrogen (the desorption gas). A selector valve in the

effluent is used to lead the effluent gas from the selected reactor to the mass spectrometer (Hiden

Analytical HPR-20 QIC). The selected reactor is fed with the mixed gas feed and continuously

monitored by the MS (10 s interval).26 A second selector valve is used to analyze the gas stream from

one of the desorbing reactors (under N2 flow) with a compact GC (Interscience), which is equipped with

TCD and FID detectors to detect CO2 and C2H4, respectively.

The Flowrence has four heated reactor blocks (40 - 300 °C), each containing four packed tubes of

560 mm height, 6 mm OD and 4 mm ID that can be pressurized (0-10 barg). The isothermal zone was

determined to be 350 mm. During a run, one blank and 15 sorbent materials can be tested. At the bottom

of the reactor a diluent gas (N2) can be mixed with the effluent, which is used to dilute the gas flow

before analysis, increase the flow rate and reduce the axial dispersion effects. At the start of each run,

the materials are dried in the reactor under 25 NmL/min N2 per reactor at 473 K for 2 h. All the

experiments reported here were conducted at 1 bar absolute pressure, and 313 K. See also the schematic

in Figure S5.

The sorbent bed is 4 mm in diameter, and 30 cm height. The sorbent beds are packed with commercial

materials: 13X (Aldrich), [Na]A (=LTA-4A = 4A, Acros), [Ca]A (=LTA-5A = 5A, Acros), [Ca]X

(Siliporite), NaY (CBV 100 CY, Zeolyst), and Activated carbon (GCN 3070 Cabot Corp.), that were

crushed and sieved to obtain 150-250 μm particles.

A layer of SiC (inert diluent, particle size of 150 μm) was loaded in the reactors first to make sure that

the sorbent bed is located in the isothermal zone. SiC is also used to fill up the interstitial void space

between adsorbent particles and top of the reactors up to a height of 55 cm.

S27
Transient breakthrough experiments vs simulations

A lecture bottle (Messer) containing 30 vol% C2H4, 5 vol % Ar and 65 vol% N2 was used to

determine C2H4 adsorption. CO2 was mixed with the gas feed via a separate mass flow controller and

the total feed flow was kept at 25 mL/min per reactor.

The fixed-bed tube is first flushed with pure N2 at the specified total pressure, before injection of the

feed mixture, at time t = 0. Ar is used as an inert internal standard to monitor the start of the adsorption

experiment and calculate gas concentrations. In the desorption experiments, the equilibrated bed is

flushed with a constant flow of pure N2, and the product compositions monitored by GC analysis.

In the MS, ions are produced by electron ionization (EI), separated by a quadrupole analyzer and

detected by a secondary electron multiplier (SEM). The raw signal was monitored at the selected m/z

values, corrected for spectral overlap by taking into account the relative abundancies of the different

peaks. The data was subsequently normalized to the signal of N2 (relative sensitivity = 1). The selected

m/z values and their relative sensitivities (RS) were calibrated using known concentrations of the gases.

The percentage of each component is calculated based on the total normalized response. It should be

noted that the molecular ion peaks of N2, and C2H4 are observed at m/z 28. Therefore, m/z 27 was used

to monitor the C2H4 concentration in the effluent gas.

5.2 Experimental breakthrough campaigns

The feed to each tube consists of C2H4/CO2/N2/Ar mixtures using different C2H4/CO2 ratios; N2 forms

about 58%, and Ar about 2%. Argon serves as inert internal tracer to signal the start of the component

breakthroughs; N2 serves as diluent in order to maintain nearly constant flow conditions and reduce

axial dispersion. All the experiments reported here are conducted at 1 bar absolute pressure, and 313 K.

The sorbents tried were cation-exchanged zeolites (13X, CaX, NaY, 4A, 5A) and activated carbon (from

Cabot Corp.). All of the experiments were conducted at a temperature of 313 K and 1 bar operating

pressure. Initially, the fixed bed was equilibrated by flow of pure N2. At time t = 0, the packed tube was

fed with C2H4/CO2/Ar/N2 mixtures of varying C2H4/CO2 ratios. For each sorbent, four different

C2H4/CO2 ratios in the feed mixture were chosen: Run 1 (C2H4/CO2  3), Run 2 (C2H4/CO2  1.5), Run

S28
Transient breakthrough experiments vs simulations

3 (C2H4/CO2  0.8), and Run 4 (C2H4/CO2  0.5). The experimental results for the different sorbents are

presented in Figure S6 (13X), Figure S7 (CaX), Figure S8 (NaY) , Figure S9 (LTA-4A), Figure S10

(LTA-5A), and Figure S11 (AC). In each of these Figures, the data for Runs 1, 2, 3 and 4 are indicated,

respectively, by a, b, c, d. In all the experimental campaigns, steady-state conditions are established at

the end of each experimental campaign. It is noteworthy that the breakthroughs obtained with LTA-4A

display distended characteristics, indicative of strong intra-crystalline diffusion limitations. For all other

sorbents, the breakthroughs obtained are sharp, indicating that intra-particle diffusion limitations are of

negligible importance.

5.3 Numerical analysis of experimental breakthroughs

We first undertake a simple numerical analysis of the experimental breakthroughs in order to compare

with the IAST calculations presented in Figure S1, using the methodology in earlier works.15, 27 A brief

outline of the analysis is presented below.

Let mads represent the mass of adsorbent, expressed in kg, packed into the tube that is fed with the

C2H4(1)/CO2(2) feed mixture at a constant flow rate of Q m3 s-1. The uptake of C2H4, expressed as

moles per kg of adsorbent in the packed tube, can be determined from a material balance
tss
ct Q
q1 =
mads y
0
1, feed  y1,exit  dt (S24)

In eq (S24), ct represents the total molar concentration of the entering feed mixture at 1 bar, and 313

K. The upper limit of the integral, tss, is the time required to reach stead-state. Analogously, the uptake

of CO2 is
tss
ct Q
q2 
mads y
0
2, feed  y2,exit  dt (S25)

The integrals in eqs (S24), and (S25) can be determined numerical using a quadrature formula. In our

analysis, we found that the use of the Simpson’s rule provided results of good accuracy. Combining eqs

(S24), and (S25) we can determine the mole fraction of C2H4, x1 = q1  q1  q2  , essentially invoking the

S29
Transient breakthrough experiments vs simulations

assumption that the mixture can be considered to be a pseudo-binary due to the poor adsorptivity of

bothAr and N2 present in the feed mixture. Figure S12 presents the results of this foregoing numerical

analysis in which the adsorbed phase mole fraction of C2H4 (y-axis), x1 , is plotted as a function of the

mole fraction of C2H4 in the feed mixture, y1 , treated as a pseudo-binary.

Comparison with the IAST calculations presented in Figure S1, a number of observations emerge.

Firstly, in agreement with the IAST predictions, the sorbents LTA-4A, and 13X are CO2-selective. The

data for CaX, and NaY zeolites show that both these adsorbents are practically non-selective to either

component as the data lie close to the y  x parity. Indeed, examination of Figure S7 (CaX), and Figure

S8 (NaY) show that the breakthroughs of C2H4, and CO2 occur at practically the same time. Contrary to

the expectations of the IAST, 5A exhibits a tendency for selectivity reversal for y1  0.5 ; the reasons for

such selectivity reversal can be traced to the non-idealities in mixture adsorption, as detailed in earlier

work.28

The most important conclusion to emerge is that AC is a suitable sorbent for selective adsorption of

C2H4 from CO2-bearing mixtures.

5.4 Experimental breakthroughs versus simulations

Transient breakthrough simulations were performed in order to compare with the experimental

breakthroughs. In all the breakthrough simulations reported in this article, the IAST was used to model

mixture adsorption equilibrium. The breakthroughs obtained with 4A have distended characteristics. To

match the experiments, intra-crystalline diffusional influences must be accounted for in the

breakthrough simulations. The best match was obtained by taking the values of Ði rc2 = 0.1 s-1 for Ar,

N2, and CO2; For C2H4, the intra-cage hopping is more hindered by cations, and the value Ði rc2 = 0.01

s-1 yielded the best fits. Figure S13 compares the breakthrough experiments with simulations for the

complete set of four runs; the match is very good.

S30
Transient breakthrough experiments vs simulations

For 13X, CaX, NaY, 5A, and AC, the breakthrough characteristics are sharp, and intra-crystalline

diffusional limitations are of negligible importance. The comparisons of the breakthrough experiments

with simulations are presented in Figure S14, Figure S15, Figure S16, Figure S17, and Figure S18.

For 13X zeolite, the breakthrough simulations show the correct qualitative features as the

experimental data; see Figure S14. However, the match is not quantitatively perfect. The simulations

tend to predict a larger gap between the breakthrough times of C2H4 and CO2; this indicates that the

unary isotherms anticipate more selective separations than is achieved in practice.

For CaX, and NaY zeolites (see Figure S15, and Figure S16), tthe breakthrough simulations show the

right qualitative trends as the experimental data, the gap between the breakthrough times of C2H4 and

CO2 is slightly larger than observed experimentally. Indeed, the experiments, show that C2H4 and CO2

breakthrough at nearly the same time. The unary isotherm data, along with IAST, for CaX and NaY

anticipate a somewhat higher CO2/C2H4 selectivity than is realized in practice.

For 5A zeolite (see Figure S17), the agreement between experimental data and breakthrough

simulations is significantly worse, the selectivity reversal observed in Run 5 is not anticipated by the

breakthrough simulations using IAST for calculations of the mixture adsorption equilibrium. In earlier

work28 it has been established that the non-idealities in mixture adsorption equilibrium is the root cause

of selectivity reversals.

For AC, the transient breakthrough simulations show the same qualitative features as the experimental

data but the quantitative match is poor because unary isotherm data used in the simulations were based

on the literature data for microporous activated carbon (Type BPL, 6/16 mesh, manufactured by the

Pittsburgh Chemical Company).

S31
T
Transient breakthroug
gh experimeents vs simulations

55.5 List of Figurees for Tran

nsient brea
akthrough
h experiments vs sim
mulations

Figure S4. Schematic of

o experimen
ntal set up fo
or transient bbreakthroughhs.

S32
Transient breakthrough experiments vs simulations

Ar/N2/C2H4/CO2 N2
N2
Feed, 1 bar, 313 K Feed, 1 bar, 313 K
473 K
4 mm
i.d.

Adsorption

Desorption
Drying
300 mm

2h

phase

phase
height

GC/MC GC/MC
analysis analysis

Figure S5. Schematic of experimental procedure for transient breakthroughs.

S33
Transient breakthrough experiments vs simulations

a 30 b 30
C2H4 C2 H4
25 CO2 25 CO2

% component in outlet gas

% component in outlet gas

Ar Ar
20 20

15 Ar/C2H4/CO2/N2; 15
13X; 313 K,
Ar/C2H4/CO2/N2;
10 100 kPa 10 13X; 313 K,
100 kPa
5 5

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30

time, t / min time, t / min

c 30
d
C2H4 C2 H4
CO2 30
% component in outlet gas

25
% component in outlet gas

CO2
Ar Ar
20
20
15 Ar/C2H4/CO2/N2;
13X; 313 K, Ar/C2H4/CO2/N2;
10 100 kPa 13X; 313 K,
10
100 kPa
5

0 0
0 5 10 15 20 25 0 5 10 15 20

time, t / min time, t / min

Figure S6. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in 13X zeolite at 313 K and

total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can

be determined by the taking the sum of the mole % = 100.

S34
Transient breakthrough experiments vs simulations

a b 25
25
C2H4
C2H4
CO2
CO2 20
20

% component in outlet gas

Ar
% component in outlet gas

Ar

15 15

Ar/C2H4/CO2/N2; Ar/C2H4/CO2/N2;
10 CaX; 313 K, 10 CaX; 313 K,
100 kPa 100 kPa

5 5

0 0

0 5 10 15 20 0 5 10 15 20

time, t / min time, t / min

c d 30
30 C2 H4 C2 H4
CO2 % component in outlet gas 25 CO2
25
% component in outlet gas

Ar Ar
20
20

15 Ar/C2H4/CO2/N2;
15 Ar/C2H4/CO2/N2; CaX; 313 K,
CaX; 313 K,
100 kPa
10
10 100 kPa

5 5

0 0
0 5 10 15 20 0 2 4 6 8 10 12 14 16

time, t / min time, t / min

Figure S7. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in CaX zeolite at 313 K and

total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can

be determined by the taking the sum of the mole % = 100.

S35
Transient breakthrough experiments vs simulations

a b
25 25
C2H4
C2H4 CO2
20 CO2 20

% component in outlet gas

% component in outlet gas

Ar
Ar
15 15
Ar/C2H4/CO2/N2;
Ar/C2H4/CO2/N2;
NaY; 313 K,
NaY; 313 K, 10 100 kPa
10
100 kPa

5 5

0 0

0 5 10 15 20 25 30 0 5 10 15 20 25

c time, t / min d time, t / min

20 30
C2H4
CO2 25 C2H4
% component in outlet gas

CO2
% component in outlet gas

15 Ar
20 Ar

10 15 Ar/C2H4/CO2/N2;
Ar/C2H4/CO2/N2;
NaY; 313 K,
NaY; 313 K, 10 100 kPa
5 100 kPa

0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12

time, t / min time, t / min

Figure S8. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in NaY zeolite at 313 K and

total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can

be determined by the taking the sum of the mole % = 100.

S36
Transient breakthrough experiments vs simulations

a b
20 20
% component in outlet gas

% component in outlet gas

C2H4
15 CO2 15 C2H4
Ar CO2
Ar
10 10
Ar/C2H4/CO2/N2;
LTA-4A; 313 K, 100 kPa Ar/C2H4/CO2/N2;
5 5 LTA-4A; 313 K, 100 kPa

0 0
0 10 20 30 40 0 10 20 30 40

time, t / min time, t / min

c d 30
20 Ar/C2H4/CO2/N2; C2H4
LTA-4A; 313 K, 100 kPa 25
CO2
% component in outlet gas
% component in outlet gas

Ar
15 20
Ar/C2H4/CO2/N2;
LTA-4A; 313 K, 100 kPa
15
10
C2H4
CO2 10
5 Ar
5

0 0
0 10 20 30 40 0 10 20 30 40

time, t / min time, t / min

Figure S9. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in LTA-4A zeolite at 313 K and

total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can

be determined by the taking the sum of the mole % = 100.

S37
Transient breakthrough experiments vs simulations

a b
25
25

20
% component in outlet gas

% component in outlet gas

C2H4 20
CO2
15 C2 H4 N2/Ar/CO2/C2H4
Ar 15
CO2 5A; 313 K,
Ar 100 kPa
10 10

5 5

0 0
0 10 20 30 40 0 10 20 30 40
time, t / min time, t / min
c 20
d
C2 H4
CO2
30
Ar C2 H4
% component in outlet gas
% component in outlet gas

15
CO2
Ar
20
10
N2/Ar/CO2/C2H4
N2/Ar/CO2/C2H4
5A; 313 K,
5A; 313 K,
100 kPa 10
5 100 kPa

0 0
0 5 10 15 20 25 30 0 5 10 15 20

time, t / min time, t / min

Figure S10. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in LTA-5A zeolite at 313 K

and total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. These data are taken from

earlier work.28 The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S38
Transient breakthrough experiments vs simulations

a 25
b 25 Ar/C2H4/CO2/N2;
C2H4 AC; 313 K,
C2H4
CO2 100 kPa
20 20 CO2

% component in outlet gas

% component in outlet gas

Ar
Ar

15 Ar/C2H4/CO2/N2; 15
AC; 313 K,
100 kPa
10 10

5 5

0 0

0 5 10 15 20 0 5 10 15 20

time, t / min time, t / min

c 30 d 40

C2H4
25 CO2
% component in outlet gas
% component in outlet gas

30
Ar
20
C2H4 Ar/C2H4/CO2/N2;
15 20
CO2 AC; 313 K,
Ar Ar/C2H4/CO2/N2; 100 kPa
10
AC; 313 K,
10
100 kPa
5

0 0
0 5 10 15 20 25 0 5 10 15 20

time, t / min time, t / min

Figure S11. Experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in Activated Carbon (AC) at

313 K and total pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the

outlet gas can be determined by the taking the sum of the mole % = 100.

S39
Transient breakthrough experiments vs simulations

1.0

mole fraction of C2H4 in adsorbed phase, x1

y1=x1
0.8
AC
5A
NaY
0.6
CaX
13X
0.4 4A

0.2 C2H4(1)/CO2(2)
313 K;
from breakthroughs
0.0
0.0 0.2 0.4 0.6 0.8 1.0

mole fraction of C2H4 in bulk gas phase, y1

Figure S12. Analysis of the experimental breakthroughs with different sorbents. The adsorbed phase

mole fraction of C2H4 (y-axis) is plotted as a function of the mole fraction of C2H4 in the bulk gas phase

mixture (x-axis). The data for 5A zeolite includes the entire data set obtained at 1 bar and reported in our

earlier work.28

S40
Transient breakthrough experiments vs simulations

a b
20 20

% component in outlet gas

% component in outlet gas

C2H4
15 CO2 15 C2H4
Ar CO2
simulation Ar
10 10
simulation
Ar/C2H4/CO2/N2;
LTA-4A; 313 K, 100 kPa
5 5

Ar/C2H4/CO2/N2;
0 0 LTA-4A; 313 K, 100 kPa

0 10 20 30 40 0 10 20 30 40

time, t / min time, t / min

c d 30
20
25
% component in outlet gas
% component in outlet gas

C2H4
15 20 CO2
Ar
C2H4 15 simulation
10
CO2
Ar 10
simulation Ar/C2H4/CO2/N2;
5
5 LTA-4A; 313 K, 100 kPa
Ar/C2H4/CO2/N2;
0 LTA-4A; 313 K, 100 kPa 0
0 10 20 30 40 0 5 10 15 20 25 30

time, t / min time, t / min

Figure S13. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in LTA-4A zeolite at 313 K and total

pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be

determined by the taking the sum of the mole % = 100.

S41
Transient breakthrough experiments vs simulations

a b
30
Ar/C2H4/CO2/N2;
25 13X; 313 K,
Ar/C2H4/CO2/N2; 25 100 kPa
13X; 313 K,

% component in outlet gas

% component in outlet gas

20 100 kPa
20
C2H4
15 CO2 15 C2H4

Ar CO2
10 simulation 10 Ar
simulation
5 5

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30

time, t / min time, t / min

c d
Ar/C2H4/CO2/N2;
30 Ar/C2H4/CO2/N2;
13X; 313 K,
13X; 313 K,
30 100 kPa
% component in outlet gas

25
% component in outlet gas

100 kPa

20
C2H4
20
15 CO2 C2H4
Ar CO2
10 simulation Ar
10
simulation
5

0 0
0 5 10 15 20 25 0 5 10 15 20

time, t / min time, t / min

Figure S14. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in 13X zeolite at 313 K and total pressure of

1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be determined by

the taking the sum of the mole % = 100.

S42
Transient breakthrough experiments vs simulations

a b
25
C2H4 C2H4
20 CO2
20 CO2

% component in outlet gas

Ar
% component in outlet gas

Ar
simulation simulation
15
15

Ar/C2H4/CO2/N2; Ar/C2H4/CO2/N2;
10 CaX; 313 K, 10 CaX; 313 K,
100 kPa 100 kPa

5 5

0 0

0 5 10 15 20 0 5 10 15 20

time, t / min time, t / min

c d 30
30 C 2H 4
C2H4 25 CO2
% component in outlet gas

25 CO2 Ar
% component in outlet gas

Ar simulation
20
20 simulation

15
15 Ar/C2H4/CO2/N2;
Ar/C2H4/CO2/N2; CaX; 313 K,
CaX; 313 K, 10 100 kPa
10
100 kPa
5 5

0 0
0 5 10 15 0 2 4 6 8 10 12 14 16

time, t / min time, t / min

Figure S15. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in CaX zeolite at 313 K and total pressure of

1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be determined by

the taking the sum of the mole % = 100.

S43
Transient breakthrough experiments vs simulations

a b
C2H4 25
C2H4
CO2 CO2
20

% component in outlet gas

Ar 20
% component in outlet gas

Ar
simulation simulation
15
15

Ar/C2H4/CO2/N2; Ar/C2H4/CO2/N2;
10 10 NaY; 313 K,
NaY; 313 K,
100 kPa 100 kPa

5 5

0 0
0 5 10 15 20 25 30 0 5 10 15 20

time, t / min time, t / min

c 30
d
C2H4
C2H4 30
25 CO2
CO2
% component in outlet gas
% component in outlet gas

Ar
Ar
20 simulation
simulation
20
15 Ar/C2H4/CO2/N2; Ar/C2H4/CO2/N2;
NaY; 313 K, NaY; 313 K,
10 100 kPa 100 kPa
10

0 0
0 2 4 6 8 0 2 4 6 8 10 12

time, t / min time, t / min

Figure S16. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in NaY zeolite at 313 K and total pressure of

1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be determined by

the taking the sum of the mole % = 100.

S44
Transient breakthrough experiments vs simulations

a b
25
25

% component in outlet gas 20

% component in outlet gas

C2H4 20
CO2
15 C2H4 N2/Ar/CO2/C2H4
Ar 15
CO2 5A; 313 K,
simulations
Ar 100 kPa
10 10 simulations

5 5

0 0
0 10 20 30 40 50 0 10 20 30 40
time, t / min time, t / min
c d
C2H4
30
CO2 C2H4
30 CO2
25 Ar

% component in outlet gas

% component in outlet gas

Ar simulation Ar
CO2 simulation Ar, simulation
20
C2H4 simulation 20 CO2 simulation

15 C2H4 simulation
N2/Ar/CO2/C2H4 N2/Ar/CO2/C2H4
10 5A; 313 K, 5A; 313 K,
10
100 kPa 100 kPa
5

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25

time, t / min time, t / min

Figure S17. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in 5A zeolite at 313 K and total pressure of 1

bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be determined by

the taking the sum of the mole % = 100.

S45
Transient breakthrough experiments vs simulations

a b
20 C2H4
20
Ar/C2H4/CO2/N2; CO2
AC; 313 K, Ar

% component in outlet gas

% component in outlet gas

15 100 kPa simulation Ar/C2H4/CO2/N2;

C2H4 15
AC; 313 K,
CO2
100 kPa
Ar
10 simulation 10

5 5

0 0

0 5 10 15 20 25 0 5 10 15 20 25

time, t / min time, t / min

c Ar/C2H4/CO2/N2;
d Ar/C2H4/CO2/N2;
AC; 313 K,
30 AC; 313 K, 100 kPa
100 kPa
30
% component in outlet gas

25
% component in outlet gas

20 C2H4
C2H4
20 CO2
CO2
15 Ar
Ar
simulation
simulation
10
10

0 0
0 5 10 15 20 25 0 5 10 15 20 25

time, t / min time, t / min

Figure S18. Comparison of transient breakthrough simulations (continuous black solid lines) with

experimental breakthroughs for Ar/C2H4/CO2/N2 mixtures in Activated Carbon at 313 K and total

pressure of 1 bar for (a) Run1, (b) Run 2, (c) Run 3, and (d) Run 4. The % N2 in the outlet gas can be

determined by the taking the sum of the mole % = 100.

S46
Transient desorption: experiments vs simulations

6 Transient desorption: experiments vs simulations

For each of the Runs 1, 2, 3, and 4 the equilibrated beds for the six sorbents (presented earlier in

Figure S6 (13X), Figure S7 (CaX), Figure S8 (NaY) , Figure S9 (LTA-4A), Figure S10 (LTA-5A), and

Figure S11 (AC)) were desorption by flow of pure N2. See also the schematic in Figure S5. The

experimental data for % components in the outlet gas, determined by GC/MS are shown (symbols),

respectively for the six sorbents, in Figure S19, Figure S20, Figure S21, Figure S22, Figure S23, and

Figure S24.

S47
Transient desorption: experiments vs simulations

6.1 List of Figures for Transient desorption: experiments vs simulations

a b
CO2
CO2 10 C2H4
10
C2H4 simulation
% component in outlet gas

% component in outlet gas

simulation

0.1
Run 2
Run 1 13X; 313 K,
13X; 313 K, 100 kPa
100 kPa

0.1 0.01
0 10 20 30 40 0 10 20 30 40 50

time, t / min time, t / min

c d 100
NaY; 313 K, Run 4
100 kPa CO2 13X; 313 K,
10 C2H4 100 kPa
% component in outlet gas

% component in outlet gas

simulation 10

1
1

Run 3 0.1 CO2

13X; 313 K, C2H4
100 kPa
0.1 simulation
0.01
0 10 20 30 40 0 10 20 30 40

time, t / min time, t / min

Figure S19. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated 13X zeolite beds for (a) Run1, (b) Run 2, (c) Run 3, and

(d) Run 4. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S48
Transient desorption: experiments vs simulations

a b
CO2 CO2

10 C2 H4 10 C2H4
% component in outlet gas

% component in outlet gas

simulation simulation

Run 1 Run 2
1 CaX; 313 K, 1 CaX; 313 K,
100 kPa 100 kPa

0 5 10 15 20 25 30 0 5 10 15 20 25 30

time, t / min time, t / min

c d
NaY; 313 K,
100 kPa CO CO2
2
C2 H4
% component in outlet gas

% component in outlet gas

C2 H4
10 simulation
10 simulation

Run 3 Run 4
1 CaX; 313 K, CaX; 313 K,
100 kPa 100 kPa
0.1
0 5 10 15 20 25 30 0 5 10 15 20 25 30

time, t / min time, t / min

Figure S20. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated CaX zeolite beds for (a) Run1, (b) Run 2, (c) Run 3,

and (d) Run 4. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S49
Transient desorption: experiments vs simulations

a b
CO2 CO2
10
C2H4 C2H4
10
% component in outlet gas

% component in outlet gas

simulation simulation

Run 2
Run 1
1 NaY; 313 K,
NaY; 313 K,
100 kPa
100 kPa
0.1
0 10 20 30 40 0 10 20 30 40

time, t / min time, t / min

c d 100
NaY; 313 K,
10 100 kPa CO2 CO2
C2H4 C2H4
% component in outlet gas

% component in outlet gas

simulation 10
simulation

1
1

0.1
Run 3 0.1 Run 4
NaY; 313 K, NaY; 313 K,
100 kPa 100 kPa

0.01 0.01
0 5 10 15 20 0 5 10 15 20 25 30

time, t / min time, t / min

Figure S21. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated NaY zeolite beds for (a) Run1, (b) Run 2, (c) Run 3,

and (d) Run 4. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S50
Transient desorption: experiments vs simulations

a b
CO2
CO2
10 C2H4 10
C2H4
% component in outlet gas

% component in outlet gas

simulation
simulation

1 1

Run 1
4A; 313 K, Run 2
100 kPa 4A; 313 K,
100 kPa
0.1 0.1
0 5 10 15 20 25 30 0 10 20 30 40 50

time, t / min time, t / min

c d
NaY; 313 K, CO2
100 kPa CO2
C2H4
10 C2H4 10
% component in outlet gas

% component in outlet gas

simulation
simulation

1 1 Run 4
4A; 313 K,
100 kPa

Run 3
4A; 313 K,
0.1 0.1
100 kPa

0 5 10 15 20 25 0 5 10 15 20 25 30

time, t / min time, t / min

Figure S22. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated 5A zeolite beds for (a) Run1, (b) Run 2, (c) Run 3, and

(d) Run 4. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S51
Transient desorption: experiments vs simulations

100

Run 4
5A; 313 K,

% component in outlet gas

100 kPa
10

1 CO2
C2H4
C2H4 simulation
CO2 simulation
0.1
0 5 10 15 20 25

time, t / min

Figure S23. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated 5A zeolite beds for Run 4. The % N2 in the outlet gas

can be determined by the taking the sum of the mole % = 100.

S52
Transient desorption: experiments vs simulations

a b
100 100
CO2
C2H4
CO2
% component in outlet gas

% component in outlet gas

10 simulation 10 C2H4
simulation

1 1

Run 1
AC; 313 K, Run 2
0.1 100 kPa 0.1 AC; 313 K,
100 kPa

0.01 0.01
0 5 10 15 20 25 30 0 10 20 30 40

time, t / min time, t / min

c 100 d 100
NaY; 313 K, CO2
100 kPa
Run 3 C2H4
AC; 313 K,
% component in outlet gas

% component in outlet gas

10 100 kPa 10 simulation

1 1

Run 4
CO2 AC; 313 K,
0.1 0.1 100 kPa
C2H4
simulation

0.01 0.01
0 5 10 15 20 25 0 5 10 15 20 25 30

time, t / min time, t / min

Figure S24. Comparison of transient desorption simulations (continuous black solid lines) with

experimental data for desorption of equilibrated AC beds for (a) Run1, (b) Run 2, (c) Run 3, and (d)

Run 4. The % N2 in the outlet gas can be determined by the taking the sum of the mole % = 100.

S53
Nomenclature

7 Nomenclature

Latin alphabet

A surface area per kg of framework, m2 kg-1

bi Langmuir parameter, Pa 1

ci molar concentration of species i, mol m-3

ct total molar concentration in mixture, mol m-3

ci0 molar concentration of species i in fluid mixture at inlet to adsorber, mol m-3

Ði M-S diffusivity of component i for molecule-pore interactions, m2 s-1

Ðax axial dispersion coefficient, m2 s-1

E energy parameter, J mol-1

L length of packed bed adsorber, m

n number of species in the mixture, dimensionless

pi partial pressure of species i, Pa

pt total system pressure, Pa

Pi 0 sorption pressure, Pa

qA molar loading species A, mol kg-1

qi,sat molar loading of species i at saturation, mol kg-1

qt total molar loading of mixture, mol kg-1

r radial coordinate, m

rc radius of crystallite, m

R gas constant, 8.314 J mol-1 K-1

Sads adsorption selectivity, dimensionless

t time, s

S54
Nomenclature

T absolute temperature, K

u superficial gas velocity in packed bed, m s-1

v interstitial gas velocity in packed bed, m s-1

Vp pore volume, m3 kg-1

xi mole fraction of species i in adsorbed phase, dimensionless

yi mole fraction of species i in bulk fluid mixture, dimensionless

z distance along the adsorber, and along membrane layer, m

Greek letters

 voidage of packed bed, dimensionless

i molar chemical potential, J mol-1

 spreading pressure, N m-1

 framework density, kg m-3

 time, dimensionless

Subscripts

i,j components in mixture

i referring to component i

t referring to total mixture

sat referring to saturation conditions

Superscripts

0 referring to pure component loading

excess referring to excess parameter

S55
Nomenclature

S56
References

8 References

(1) Mofarahi, M.; Salehi, S. M. Pure and Binary Adsorption Isotherms of Ethylene and Ethane
on Zeolite 5A. Adsorption 2013, 19, 101-110.
(2) Mofarahi, M.; Gholipour, F. Gas Adsorption Separation of CO2/CH4 System using Zeolite
5A. Microporous Mesoporous Mater. 2014, 200, 47-54.
(3) Bakhtyari, A.; Mofarahi, M. Pure and Binary Adsorption Equilibria of Methane and Nitrogen
on Zeolite 5A. J. Chem. Eng. Data 2014, 59, 626-639.
(4) Costa, E.; Calleja, G.; Jimenez, A.; Pau, J. Adsorption Equilibrium of Ethylene, Propane,
Propylene, Carbon Dioxide, and Their Mixtures in 13X Zeolite. J. Chem. Eng. Data 1991, 36, 218-224.
(5) Cavenati, S.; Grande, C. A.; Rodrigues, A. E. Adsorption Equilibrium of Methane, Carbon
Dioxide, and Nitrogen on Zeolite 13X at High Pressures. J. Chem. Eng. Data 2004, 49, 1095-1101.
(6) Bachman, J. E.; Reed, D. A.; Kapelewski, M. T.; Chachra, G.; Jonnavittula, D.; Radaelli, G.;
Long, J. R. Enabling alternative ethylene production through its selective adsorption in the metal–
organic framework Mn2(m-dobdc). Energy Environ. Sci. 2018, 11, 2423-2431.
https://doi.org/10.1039/c8ee01332b.
(7) Choudhary, V. R.; Mayadevi, S.; Singh, A. P. Sorption Isotherms of Methane, Ethane,
Ethene and Carbon Dioxide on NaX, NaY and Na-mordenite Zeolites. J. Chem. Soc. Faraday Trans.
1995, 91, 2935-2944.
(8) Romero-Pérez, A.; Aguilar-Armenta, G. Adsorption Kinetics and Equilibria of Carbon
Dioxide, Ethylene, and Ethane on 4A(CECA) Zeolite. J. Chem. Eng. Data 2010, 55, 3625-3630.
(9) Reich, R.; Zlegler, W. T.; Rogers, K. A. Adsorption of Methane, Ethane, and Ethylene Gases
and Their Binary and Ternary Mixtures and Carbon Dioxide on Activated Carbon at 212-301 K and
Pressures to 35 Atmospheres. Ind. Eng. Chem. Process Des. Dev. 1980, 19, 336-344.
(10) Park, Y.; Moon, D.-K.; Kim, Y.-H.; Ahn, H.; Lee, C.-H. Adsorption isotherms of CO2, CO,
N2, CH4, Ar and H2 on activated carbon and zeolite LiX up to 1.0 MPa. Adsorption 2014, 20, 631-647.
https://doi.org/10.1007/s10450-014-9608-x.
(11) Ruthven, D. M. Principles of Adsorption and Adsorption Processes. John Wiley: New
York, 1984.
(12) Myers, A. L.; Prausnitz, J. M. Thermodynamics of Mixed Gas Adsorption. A.I.Ch.E.J.
1965, 11, 121-130.
(13) Siperstein, F. R.; Myers, A. L. Mixed-Gas Adsorption. A.I.Ch.E.J. 2001, 47, 1141-1159.
(14) Krishna, R. Screening Metal-Organic Frameworks for Mixture Separations in Fixed-Bed
Adsorbers using a Combined Selectivity/Capacity Metric. RSC Adv. 2017, 7, 35724-35737.
https://doi.org/10.1039/C7RA07363A.
(15) Krishna, R. Methodologies for Evaluation of Metal-Organic Frameworks in Separation
Applications. RSC Adv. 2015, 5, 52269-52295.
(16) Krishna, R. The Maxwell-Stefan Description of Mixture Diffusion in Nanoporous
Crystalline Materials. Microporous Mesoporous Mater. 2014, 185, 30-50.
(17) Krishna, R. Tracing the Origins of Transient Overshoots for Binary Mixture Diffusion in
Microporous Crystalline Materials. Phys. Chem. Chem. Phys. 2016, 18, 15482-15495.

S57
References

(18) Krishna, R.; Baur, R. Modelling Issues in Zeolite Based Separation Processes. Sep. Purif.
Technol. 2003, 33, 213-254.
(19) Krishna, R.; Baur, R. Diffusion, Adsorption and Reaction in Zeolites: Modelling and
Numerical Issues. http://krishna.amsterchem.com/zeolite/, University of Amsterdam, Amsterdam, 1
January 2015.
(20) Ruthven, D. M.; Farooq, S.; Knaebel, K. S. Pressure swing adsorption. VCH Publishers:
New York, 1994.
(21) Kooijman, H. A.; Taylor, R. A dynamic nonequilibrium model of tray distillation columns.
A.I.Ch.E.J. 1995, 41, 1852-1863.
(22) Michelsen, M. An efficient general purpose method of integration of stiff ordinary
differential equations. A.I.Ch.E.J. 1976, 22, 594-597.
(23) Bulirsch, R.; Stoer, J. Numerical treatment of ordinary differential equations by
extrapolation methods. Numer. Math. 1966, 8, 1-14.
(24) Krishna, R.; van Baten, J. M. Investigating the potential of MgMOF-74 membranes for CO2
capture. J. Membr. Sci. 2011, 377, 249-260.
(25) He, Y.; Krishna, R.; Chen, B. Metal-Organic Frameworks with Potential for Energy-
Efficient Adsorptive Separation of Light Hydrocarbons. Energy Environ. Sci. 2012, 5, 9107-9120.
(26) van Zandvoort, I.; van Klink, G. P. M.; de Jong, E.; van der Waal, J. C. Selectivity and
Stability of Zeolites [Ca]A and [Ag]A towards Ethylene Adsorption and Desorption from Complex Gas
Mixtures. Microporous Mesoporous Mat. 2018, 263, 142-149.
(27) Yu, H.; Wang, X.; Xu, C.; Chen, D.-L.; Zhu, W.; Krishna, R. Utilizing transient
breakthroughs for evaluating the potential of Kureha carbon for CO2 capture. Chem. Eng. J. 2015, 269,
135-147.
(28) van Zandvoort, I.; van der Waal, J. K.; Ras, E.-J.; de Graaf, R.; Krishna, R. Highlighting
non-idealities in C2H4/CO2 mixture adsorption in 5A zeolite. Sep. Purif. Technol. 2019, 227, 115730.
https://doi.org/10.1016/j.seppur.2019.115730.

S58