Advanced Adsorbents with High Selectivity and Enhanced ...

Advanced Adsorbents with High Selectivity and Enhanced

Capacity for Use in Next Generation Separation Processes

Shamsur Rahman

Master of Applied Science in Mechanical Engineering Bachelor of Engineering in Mechanical Engineering (Honours)

This thesis is presented for the degree of

Doctor of Philosophy

at

The University of Western Australia

Department of Chemical Engineering

Fluid Science and Resources Division

Australian Centre for LNG Futures

2020

PhD Thesis Shamsur Rahman 1

Summary

The use of natural gas, the cleanest fossil fuel, continues to grow for global energy production with

increasing environmental concerns and stricter regulations on carbon emissions. For a large fraction

of transcontinental trade, natural gas is transported in tankers as liquefied natural gas (LNG). Since

CO2 can freeze at low temperatures during the liquefaction process and excessive N2 content in the

LNG can lead to hydrostatic instabilities, both N2 and CO2 concentrations must be reduced significantly

(CO2<50 ppmv and N2<1 mol%) from their typical levels (e.g. 10 % CO2 and 5+ % N2) in the feed gas to

an LNG plant. The conventional method used to remove CO2 from natural gas involves absorption by

aqueous amine solutions, which is an energy‐intensive process requiring high temperatures to

regenerate the amine solvents. Nitrogen is generally removed from natural gas by the process of

cryogenic distillation, which is also energy intensive and only cost effective at large scales. Alternative

separation processes are sought by the natural gas industry to more efficiently handle these key

impurities.

Pressure swing adsorption (PSA) is a promising technology that is both cost‐effective and energy‐

efficient and can potentially replace amine absorption and cryogenic distillation as separation

processes for CO2 and N2, respectively, in LNG production. However, this technology relies on the

availability of adsorbents with high capacity and selectivity for the target gas molecules specific to the

separation application and the gas mixture being considered. Zeolites and activated carbon are two

of the common materials used as adsorbents in PSA cycles. These conventional materials are limited

in their ability to achieve the required separations because their selectivities and/or capacities are

usually insufficient for cost‐effective use in large‐scale separations. To overcome this limitation, a

novel zeolite‐based adsorbent, with significantly enhanced adsorption capacity, was developed from

a commercially available variant of a faujasite type zeolite Y.


The synthesis procedure involved treating zeolite powder, premixed with 10% by mass of activated

carbon, in an alkaline medium which caused an expansion of the intra‐crystalline zeolite cavities

through desilication leading to a gain in pore volume. This expansion is likely to allow a larger number

of gas molecules to be accommodated inside the zeolite and provide favourable conditions for

multilayer adsorption. The role of the exterior carbon layer is to possibly facilitate in keeping the gas

molecules entrapped inside the adsorbent material and prevent from exiting at high pressure. The

resulting material, termed a carbon enhanced zeolite (CEZ), demonstrated a 265% gain in pure

component isothermal N2 adsorption capacity compared to the parent zeolite at 35 bar and 30 °C.

Flexible porous materials, in particular some metal‐organic frameworks (MOFs), have been reported

to have very high selectivity for CO2/CH4 mixtures and are therefore widely considered as potential

adsorbents for removal of CO2 from natural gas using PSA. Despite their high selectivity, a major

obstacle that remains in designing a separation process using these advanced adsorbents is that

classical isotherm models such as the Langmuir, Toth and the BET models are unable to quantify the

adsorption isotherms of these materials, especially along the part of the isotherm curve that

corresponds to structural transitions. A novel isotherm model was therefore developed to

accommodate these distinctly shaped isotherms exhibited by flexible adsorbents.

The model was successfully regressed to both adsorption and desorption branches of stepped‐

isotherms using a wide range of experimental data from the literature as well as measurements taken

at the laboratory over a wide range of temperatures from 233 K – 323 K. The model parameters were

physically meaningful with uncertainties usually several orders of magnitude lower than the regressed

values. It was also shown that the model can be used to quantify the onset and magnitude of

hysteresis in flexible adsorbents as a function of both pressure and temperature. Finally, the functional

form was incorporated in a commercial simulation software as an isotherm model and PSA cycles to

predict the performance of a well‐known flexible MOF in separating a CH4/CO2 mixture were

successfully executed.


To summarize, this research project has made two key contributions, involving material development

and numerical modelling, to the field of adsorptive separation of gas mixtures. By developing carbon

enhanced zeolite as an adsorbent with significantly enhanced N2 and CH4 adsorption capacities, the

experimental results of this project open new doors to synthesis of low‐cost high capacity zeolite‐

based adsorbents with applications not only in gas separation but also in gas storage. The pressure‐

induced LJM model, on the other hand, enables process simulations with high‐selectivity flexible

adsorbents and is a major step forward towards the design of separation processes that can make full

use of the selective properties of these materials.


Thesis Declaration

I, Shamsur Rahman, certify that:

This thesis has been substantially accomplished during enrolment in the degree.

This thesis does not contain material which has been submitted for the award of any other degree or

diploma in my name, in any university or other tertiary institution.

No part of this work will, in the future, be used in a submission in my name, for any other degree or

diploma in any university or other tertiary institution without the prior approval of The University of

Western Australia and where applicable, any partner institution responsible for the joint‐award of this

degree.

This thesis does not contain any material previously published or written by another person, except

where due reference has been made in the text and, where relevant, in the Declaration that follows.

The work(s) are not in any way a violation or infringement of any copyright, trademark, patent, or

other rights whatsoever of any person.

This thesis contains published work and/or work prepared for publication, some of which has been

co‐authored.

25/05/2020


Authorship Declaration

This thesis contains published work and work prepared for publication.

The content of Appendix C has been published in the following refereed journal:

Du, T., Fang, X., Liu, L., Shang, J., Zhang, B., Wei, Y., Gong, H., Rahman, S., May, E.F., Webley, P.A. and

Li, G.K., 2018. An optimal trapdoor zeolite for exclusive admission of CO2 at industrial carbon capture

operating temperatures. Chemical Communications, 54(25), pp.3134‐3137.

My contribution to the above publication was limited to: equipment procurement and maintenance,

implementing upgrades to the experimental apparatus, setting up and running experiments, and data

analysis and interpretation.

Co‐author signatures:

Tao Du: ____ ____ Date: _______________________

Xin Fang: ____ ____ Date: _______________________

Liying Liu: ____ ____ Date: _______________________

Jin Shang: ____ ____ Date: _______________________

Bin Zhang: ____ ____ Date: _______________________

Yichao Wei: ____ ____ Date: _______________________

He Gong: ____ ____ Date: _______________________

Paul A. Webley: ____ ____ Date: _______________________

Gang (Kevin) Li: ____ ____ Date: _______________________

26/05/2020

26/05/2020

26/05/2020

26/05/2020

26/05/2020

26/05/2020


The content of Appendix D has been submitted to the following refereed journal for publication:

Shamsur Rahman, Arash Arami‐Niya, Xiaoxian Yang, Gongkui Xiao, Gang (Kevin) Li and Eric F. May,

2020. Temperature Dependence of Adsorption Hysteresis in Flexible Metal Organic Frameworks.

Communications Chemistry, Submitted.

The above manuscript is based predominantly on the work presented in Chapter 2 of this thesis.

Co‐author signatures:

Arash Arami‐Niya: _ _ Date: _______________________

Xiaoxian Yang: _ _ Date: _______________________

Gongkui Xiao: _ _ Date: _______________________

Gang (Kevin) Li: _ _ Date: _______________________

Student signature

Shamsur Rahman: Date: 25/05/2020

Coordinating Supervisor signature

Eric F. May: ____ ____ Date: _______________________


Table of Contents

Summary ................................................................................................................................................. 1

Thesis Declaration ................................................................................................................................... 4

Authorship Declaration ........................................................................................................................... 5

List of Figures ........................................................................................................................................ 11

List of Tables ......................................................................................................................................... 23

Acknowledgements ............................................................................................................................... 26

Chapter 1 Introduction and Motivation ................................................................................................ 27

1.1 The rise of natural gas as a fuel ...................................................................................................... 27

1.1.1 Natural gas in Australia ............................................................................................................ 29

1.1.2 Environmental regulations ....................................................................................................... 30

1.2 The challenge of separating gas mixtures ....................................................................................... 31

1.3 Adsorptive separation ..................................................................................................................... 33

1.3.1 The need for adsorbents with high capacity ............................................................................ 34

1.3.2 The need for adsorbents with high selectivity ......................................................................... 35

1.3.3 The need for reliable isotherm models .................................................................................... 36

1.4 The objectives of this research ....................................................................................................... 37

Chapter 2 Structural Transitions in Metal Organic Frameworks .......................................................... 40

Foreword ........................................................................................................................................... 40

2.1 Introduction .................................................................................................................................... 40

2.2 Development of the pressure‐induced LJM model ........................................................................ 43

2.3 Modelling stepped breathing isotherms ......................................................................................... 44

2.3.1 CH4 on Fe(bdp) ......................................................................................................................... 45

2.3.2 A thermodynamic model for hysteresis ................................................................................... 51

2.3.3 CH4 on Co(bdp) ......................................................................................................................... 55

2.4 Experimental work .......................................................................................................................... 56

2.4.1 Hysteresis in non‐breathing adsorbents .................................................................................. 61

2.5 Further modelling ........................................................................................................................... 63

2.5.1 CO2 on mmen‐Fe2(dobpdc) and mmen‐Co2(dobpdc) .............................................................. 63

2.5.2 CO2 on amino‐MIL‐53(Al) ......................................................................................................... 64

2.6 Process simulation .......................................................................................................................... 66

2.6.1 Pressure swing adsorption simulation of an equimolar CH4/CO2 mixture .............................. 67

2.7 Concluding remarks ........................................................................................................................ 72

Chapter 3 Effect of Structural Manipulations on Selectivity and Capacity of Zeolite Y ........................ 74


3.1 Introduction .................................................................................................................................... 74

3.2 Zeolite structures ............................................................................................................................ 75

3.2.1 Role of the cation on the adsorption mechanism ................................................................... 76

3.2.2 Ionic Liquidic Zeolites (ILZs) ..................................................................................................... 78

3.2.3 Organic cations ........................................................................................................................ 79

3.2.4 Synthesis of TMAY .................................................................................................................... 79

3.3 Characterisation of existing zeolite‐based adsorbents ................................................................... 82

3.3.1 Adsorption measurements ...................................................................................................... 82

3.3.2 Effect of Si/Al ratio ................................................................................................................... 83

3.3.3 Regression to isotherm models ............................................................................................... 85

3.3.4 TMAY vs parent zeolite ............................................................................................................ 89

3.3.5 Effect of degassing ................................................................................................................... 90

3.4 Novel zeolite‐based adsorbents ...................................................................................................... 91

3.4.1 Synthesis .................................................................................................................................. 91

3.4.2 Effect of intermediate cations ................................................................................................. 92

3.4.3 Effect of alkalinity .................................................................................................................... 94

3.4.4 Effect of H+ cation .................................................................................................................... 96

3.5 Novel carbon‐based adsorbents ..................................................................................................... 98

3.5.1 Acid functionalisation ............................................................................................................ 102

3.5.2 Acid base reaction approach .................................................................................................. 107

3.5.3 Alkaline treatment ................................................................................................................. 109

3.5.4 APS functionalisation ............................................................................................................. 110

3.5.6 Carbon Enhanced Ionic Liquidic Zeolites ............................................................................... 112

3.6 Concluding remarks ...................................................................................................................... 114

Chapter 4 Carbon Enhanced Zeolite .................................................................................................. 116

4.1 Introduction .................................................................................................................................. 116

4.2 Synthesis ....................................................................................................................................... 117

4.3 Characterisation ............................................................................................................................ 118

4.3.1 Chemical composition ............................................................................................................ 118

4.3.2 X‐ray Diffraction ..................................................................................................................... 119

4.3.3 Surface area and pore volume ............................................................................................... 120

4.3.4 Sorption experiments ............................................................................................................ 121

Enhancement in N2 adsorption capacity ..................................................................................... 122

Enhancement in CH4 adsorption capacity ................................................................................... 123

Selectivity .................................................................................................................................... 125

4.4 Possible mechanism ...................................................................................................................... 125


4.5 Further characterisation ............................................................................................................... 130

4.5.1 Scanning Electron Microscopy (SEM) .................................................................................... 130

4.5.2 Scanning Transition Electron Microscopy (STEM) ................................................................. 132

4.5.3 Pore size distribution ............................................................................................................. 133

4.6 Concluding remarks ...................................................................................................................... 135

Chapter 5 Conclusions and Future Research ...................................................................................... 136

5.1 Conclusions ................................................................................................................................... 136

5.2 Recommendations for future research ......................................................................................... 139

Appendices .......................................................................................................................................... 140

Appendix A .......................................................................................................................................... 141

A1 ‐ Development of the pressure‐induced LJM model ................................................................. 141

A2 ‐ Parameters for the pressure‐induced LJM model ................................................................... 144

A3 ‐ Summary of best‐fit parameters of the pressure‐induced LJM model and their associated

statistical uncertainties ................................................................................................................... 145

A4 ‐ Residual plots ........................................................................................................................... 146

A5 ‐ Simulation of the dynamic breakthrough of an equimolar CO2/CH4 mixture ......................... 148

A6 ‐ Regression of CH4 and CO2 adsorption isotherms on Zeolite 13X using the Toth Isotherm

model .............................................................................................................................................. 151

A7 ‐ Simulation parameters ............................................................................................................ 151

Appendix B .......................................................................................................................................... 154

B1 ‐ Effect of carbon‐content .......................................................................................................... 154

B2 ‐ Effect of NaOH concentration ................................................................................................. 155

B3 ‐ Toth fittings .............................................................................................................................. 156

B4 – Pore size distribution (based on pore volume) ....................................................................... 159

B5 – High resolution TEM images ................................................................................................... 160

Appendix C .......................................................................................................................................... 164

C1 ‐ Foreword .................................................................................................................................. 164

C2 ‐ The Quest for a Trapdoor Zeolite for Exclusive Admission of CO2 at Industrial Working

Temperatures .................................................................................................................................. 164

CS ‐ Supporting Information (SI) ..................................................................................................... 173

CS1 ‐ Experimental section ......................................................................................................... 173

CS2 ‐ Characterisations ............................................................................................................... 173

CS3 ‐ DFT calculations ................................................................................................................. 174

CS4 ‐ Multicomponent breakthrough experiments .................................................................... 177

Appendix D .......................................................................................................................................... 178

D1 ‐ Foreword ................................................................................................................................. 178


D2 ‐ Temperature Dependence of Adsorption Hysteresis in Flexible Metal Organic Frameworks 178

DS ‐ Supporting Information (SI) ..................................................................................................... 190

DS1 ‐ Derivation of the LJMY model for stepped sorption isotherms ........................................ 190

DS2 ‐ Results of LJMY‐Langmuir model regression to MOF isotherm data ................................ 193

DS3 ‐ Measurements of CO2 sorption on ZIF‐7 ........................................................................... 201

DS4 ‐ Temperature dependence and hysteresis of structural transition parameters ................ 206

DS5 ‐ Pressure Vacuum Swing Adsorption simulation of a CH4/CO2 mixture with ZIF‐7 ............ 209

Appendix E .......................................................................................................................................... 215

E1 – Data table for experimental data presented in Figure 2.1 and Figure 2.13 ............................ 215

E2 – Data table for experimental data presented in Figure 2.3...................................................... 218

E3 – Data table for experimental data presented in Figure 2.8...................................................... 226

E4 – Data table for experimental data presented in Figure 2.9 and Figure 2.10 ............................ 232

E5 – Data table for experimental data presented in Figure 2.11 ................................................... 238

E6 – Data table for all experimental isotherm data presented in Chapter 3 .................................. 244

E7 – Data table for all experimental isotherm data presented in Chapter 4 .................................. 264

References .......................................................................................................................................... 278


List of Figures

Figure 1.1. Comparison of growth of natural gas consumption and production in 2018 with the

previous 10‐year average annual growth rate and the major countries which contributed to the growth

[1]. ......................................................................................................................................................... 27

Figure 1.2. Comparison of the historical and projected use of different fuels to generate electricity in

the US from 2010 – 2050 [3]. ................................................................................................................ 28

Figure 1.3. LNG export projects in operation in Western Australia since 2015 and the growth in the

state’s annual LNG export capacity in million tonnes [5]. .................................................................... 29

Figure 1.4. Conventional gas processing operations in a typical natural gas plant with a cryogenic

process for liquefied natural gas production. MCHE = main cryogenic heat exchanger [12]. ............. 31

Figure 1.5. (a) Schematic representation of a two‐bed cyclic adsorption process with (b) pressure swing

adsorption (PSA) and temperature swing adsorption (TSA) cycles shown on a typical isotherm [12]. 34

Figure 2.1. Deviation of the classical Toth model from isothermal experimental data [60] at 303K for

the adsorption of (a) CO2 and (b) CH4 on amino‐MIL‐53 (Al). ............................................................... 42

Figure 2.2. 3D depiction of (a) solid‐state structure of the collapsed phase of Fe(bdp) under vacuum at

298K and (b) idealized average solid‐state structure of the expanded phase of Fe(bdp) under 50 bar

pressure at 298K [48]. ........................................................................................................................... 46

Figure 2.3. Demonstration of the pressure‐induced LJM model’s ability to quantify breathing isotherms

for CH4 (a) adsorption and (b) desorption on Fe(bdp) at seven temperatures reported in reference

[48]. Solid and hollow symbols denote, respectively, adsorption and desorption experimental data.

Solid curves represent the pressure‐induced LJM model. .................................................................... 47


Figure 2.4. Plots of P0 vs. temperature for adsorption (breathing‐in) and desorption (breathing‐out).

The plots indicate that P0 = f (t) is linear for this data set. The two straight lines through the plots meet

at a point, where the breathing‐in transition pressure, P0,ads, becomes equal to the breathing‐out

transition pressure, P0,des. ..................................................................................................................... 49

Figure 2.5. (a) – (e) Adsorption and desorption isotherms for CH4 on Fe(bdp) at temperatures of (a)

323K, (b) 248K obtained from the fitting to experimental data and (c) 175K, (d) 115K, and (e) 100K

projected by the pressure‐induced LJM model. The adsorption and desorption transition pressures,

P0,ads and P0,des are shown in (a) where they are farthest apart and (d) where they are equal at the

point of intersection. (f) Area enclosed by each pair of adsorption and desorption isotherms calculated

graphically for each of the temperatures shown in (a) – (e) plus four additional temperatures. ........ 50

Figure 2.6. Energy paths followed during adsorption/desorption cycles when (a) T ≤ Ti where hysteresis

is at its low asymptote as the energy paths during adsorption and desorption are closest to each other,

(b) T > Ti giving rise to hysteresis as the energy path during adsorption is longer than that during

desorption and (c) T >> Ti where the effect of hysteresis is larger as the difference between the energy

paths during adsorption and desorption are greater. .......................................................................... 53

Figure 2.7. 3D depiction of the solid‐state transformation of Co(bdp) from collapsed phase at 0 bar

(Low 4) to expanded phase at 30 bar (High 4) [48]. ............................................................. 55

Figure 2.8. The pressure‐induced LJM model regressed on isothermal (a) adsorption and (b) desorption

of CH4 on Co(bdp). Experimental data has been collected from reference [48]. The model’s ability to

quantify breathing isotherms at all five reported temperatures is demonstrated for both (a) adsorption

and (b) desorption where solid and hollow symbols denote experimental data and the solid curves

represent the pressure‐induced LJM model. (c) Plots for ln (P0) vs. 1/T for adsorption (breathing‐in)

and desorption (breathing‐out) showing that both sets of points fall on a straight line and justifying


the use of the Clausius‐Clapyron equation and predicting the existence of a point where P0,ads = P0,des.

.............................................................................................................................................................. 56

Figure 2.9. The pressure‐induced LJM model regressed on low pressure isothermal (a) adsorption and

(b) desorption of CO2 on ZIF‐7. Measurements were made using a Micromeritics ASAP2020

instrument. (c) Clausius‐Clapyron plots for ln (P0) vs. 1/T for isothermal adsorption (breathing‐in) and

desorption (breathing‐out) of CO2 on ZIF‐7 between 233K and 247K showing that both sets of points

fall on a straight line and predicting the existence of a point where P0,ads = P0,des. ............................... 59

Figure 2.10. (a) – (d) Results of adsorption/desorption experiments of CO2 on ZIF‐7 illustrating the

diminishing effect of hysteresis with decreasing temperature. The area between the adsorption and

desorption isotherms can be used as a measure of hysteresis at each temperature. (e) Area enclosed

by each pair of adsorption and desorption isotherms calculated graphically for temperatures up to

283K. ..................................................................................................................................................... 61

Figure 2.11. Demonstration of the pressure‐induced LJM model’s ability to quantify breathing

isotherms for the adsorption of CO2 on (a) mmen‐Fe2(dobpdc) and (b) mmen‐CO2(dobpdc) at three

temperatures reported by McDonald et al [64]. Solid symbols denote experimental data and solid

curves represent the pressure‐induced LJM model. ............................................................................ 64

Figure 2.12. 3D structure of amino‐MIL‐53(Al) in (a) np and (b) lp phases. Color legend – yellow: Al,

red: O, gray: C, and blue: N [80]. .......................................................................................................... 65

Figure 2.13. Demonstration of the pressure‐induced LJM model’s ability to quantify double‐stepped

breathing isotherms for CO2 (a) adsorption and (b) desorption on amino‐MIL‐53(Al) at two

temperatures reported in reference [60]. Solid symbols denote experimental data, and solid curves

represent the pressure‐induced LJM model. ........................................................................................ 66


Figure 2.14. Illustration of CO2 and CH4[ [(a) and (b) respectively] isothermal working capacity of amino‐

MIL‐53 (Al) for a PSA cycle swing pressure of 6.4 – 20 bar that corresponds to the adsorbent’s CO2

breathing region. Zeolite 13X has a much smaller CO2 working capacity[(c)] in the same pressure range

and also a lower CH4 working capacity[(d)]. Hence, under identical conditions, amino‐MIL‐53 (Al) is

expected to produce superior CO2/CH4 separation performance compared to Zeolite 13X when used

as an adsorbent in a PSA process.......................................................................................................... 70

Figure 2.15. Comparison between amino‐MIL‐53(Al) and Zeolite 13X as adsorbents used in a two‐bed

PSA simulation with an equimolar CH4/CO2 feed. Variation in product stream composition with time

show that a far superior separation performance can be achieved using amino‐MIL‐53(Al) as the

adsorbent. The pressure‐induced LJM model was used as the isotherm model to characterise the

adsorbent bed for amino‐MIL‐53(Al) while the classical Toth model was used for Zeolite 13X. ......... 71

Figure 3.1. Framework of (a) zeolite X/Y and (b) zeolite A [25]. ........................................................... 76

Figure 3.2. Structure of Tetramethylammonium Chloride (TMA‐Cl) [94]. ........................................... 80

Figure 3.3. (a) Structure of Faujasite zeolite NaX/NaY [89]. (b) Structure of TMA‐X or TMA‐Y

synthesised through ion exchange [95]. ............................................................................................... 80

Figure 3.4. Micromeritics ASAP 2020 used for single component adsorption analyses. ..................... 82

Figure 3.5. Adsorption isotherms at different temperatures for TMAY – (a) CBV100 with CH4, (b)

CBV720 with CH4, (c) CBV100 with N2 and (d) CBV720 with N2. ........................................................... 84

Figure 3.6. The Langmuir isotherm model (line) fitted to adsorption data (circles) at ‐15 °C, 0 °C and 30

°C for TMAY – (a) CBV100 with CH4, (b) CBV720 with CH4, (c) CBV100 with N2 and (d) CBV720 with N2.

.............................................................................................................................................................. 86


Figure 3.7. Adsorption isotherms of TMAY and parent zeolite NaY at (a) 0 °C with CH4, (b) 0 °C with N2,

(c) 30 °C with CH4 and (d) 30 °C with N2................................................................................................ 90

Figure 3.8. Adsorption isotherms of degassed and not degassed TMAY at (a) 0 °C with CH4, (b) 30 °C

with CH4. ............................................................................................................................................... 91

Figure 3.9. (a) Shaking water bath used for synthesis temperatures below 90 °C (b) Teflon autoclave

and electric hotplate used for synthesis temperatures above 90 °C. ................................................... 92

Figure 3.10. Adsorption isotherms of TMAY, TMA‐Mg‐Y and NaY at (a) 0 °C with CH4, (b) 0 °C with N2,

(c) 30 °C with CH4, and (d) 30 °C with N2. .............................................................................................. 93

Figure 3.11. Adsorption isotherms of TMAY, TMA‐Base‐Y and NaY at (a) 0 °C with CH4 and (b) 0 °C with

N2. .......................................................................................................................................................... 94

Figure 3.12. Adsorption isotherms of TMAY, TMAOH‐NaY and NaY at (a) 0 °C with CH4, (b) 0 °C with N2,

(c) 30 °C with CH4, and (d) 30 °C with N2. .............................................................................................. 96

Figure 3.13. Adsorption isotherms of TMAY, TMAOH‐HY600 and TMAOH‐HY400 at (a) 0 °C with CH4,

and (b) 0 °C with N2. .............................................................................................................................. 97

Figure 3.14. Adsorption isotherms of TMAY and TMAOH‐HY400 at (a) 30 °C with CH4 and (b) 30 °C with

N2. .......................................................................................................................................................... 98

Figure 3.15. Diagram showing mesoporous zeolite templated by (a) carbon nanoparticles, (b) carbon

nanotubes, (c) carbon aerogel and (d) 3D ordered mesoporous 3DOm carbon [31]. ....................... 100

Figure 3.16. Schematic showing the influence of Si/Al ratio on the desilication effect of MFI zeolites

treated in NaOH solution and the associated mechanism of mesopore formation [103]. ............... 101

Figure 3.17. Adsorption isotherms of carbon (activated carbon Norit RB3), Acid‐Carbon1 and TMA‐

Carbon1 at (a) 0 °C with CH4 and (b) 0 °C with N2. .............................................................................. 104


Figure 3.18. Adsorption isotherms of carbon (activated carbon Norit RB3), Acid‐Carbon2 and TMA‐

Carbon2 at (a) 0 °C with CH4 and (b) 0 °C with N2. .............................................................................. 104

Figure 3.19. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐Carbon3 at (a) 0

°C with CH4 and (b) 0 °C with N2. ........................................................................................................ 106

Figure 3.20. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐Carbon4 at (a) 0

°C with CH4 and (b) 0 °C with N2. ........................................................................................................ 107

Figure 3.21. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐AcidBaseCarbon

at (a) 0 °C with CH4 and (b) 0 °C with N2. ............................................................................................ 109

Figure 3.22. Adsorption isotherms of activated carbon and Base‐Carbon at (a) 0 °C with CH4 and (b) 0

°C with N2. ........................................................................................................................................... 110

Figure 3.23. Schematic showing pore opening of functionalised carbon surface under increased

oxidation temperature and duration [108]. ....................................................................................... 111

Figure 3.24. Adsorption isotherms of activated carbon, TMA‐APS‐Carbon, and APS‐Carbon at (a) 0 °C

with CH4 and (b) 0 °C with N2. ............................................................................................................. 112

Figure 3.25. Adsorption isotherms of activated carbon, Carbon‐E‐TMAY1, TMAY and NaY at (a) 0 °C

with CH4, (b) 0 °C with N2, (c) 30 °C with CH4 and (d) 30 °C with N2. .................................................. 113

Figure 4.1. Results of Thermogravimetric analysis conducted on (a) NaY, (b) C‐R2030 and (c) CEZ. The

red line shows the change in mass of the sample with the corresponding change in temperature (blue

line). .................................................................................................................................................... 119

Figure 4.2. X‐ray Diffraction (XRD) patterns for NaY and CEZ. ............................................................ 120


Figure 4.3. Setup of the Sieverts apparatus used to conduct high pressure sorption experiments. The

sample cell is wrapped in aluminium foil for insulation and submerged into the temperature control

bath during experiments to maintain isothermal conditions, as shown in the inset. ........................ 122

Figure 4.4. Single component N2 adsorption isotherms for CEZ, C‐R2030 and NaY obtained using a

custom‐built computer‐controlled Sieverts apparatus under isothermal conditions at (a) 0 °C, (b) 15 °C

and (c) 30 °C. ....................................................................................................................................... 123

Figure 4.5. Single component CH4 adsorption isotherms for CEZ, C‐R2030 and NaY obtained using a

custom‐built computer‐controlled Sieverts apparatus under isothermal conditions at (a) 0 °C, (b) 15 °C

and (c) 30 °C. ....................................................................................................................................... 124

Figure 4.6. A model of the Faujasite unit cell. Near the centre of each line segment is an oxygen atom.

Silicon and aluminium atoms alternate at the tetrahedral intersections, except that Si substitutes for

Al at about 4% of the Al positions. Exchangeable cations are found in site I at the centre of a D6R ; site

II at the centre of the S6R, or displaced from this point into a supercage; sites I’ and II’ lie in the sodalite

cavity, on the opposite sides of the corresponding six‐rings from sites I and II, respectively; site III on

a twofold axis opposite a four‐ring inside the supercage and site III’ off site III. The eighth sodalite

cavity was removed from this diagram for simplicity and it is orthogonal to the surface of this paper.

Both figure and caption text are from Ahmed et al. [119]. ................................................................ 126

Figure 4.7. Schematics showing (a) the arrangement of Faujasite frameworks in a typical regular

pyramid‐shaped NaY particle and (b) the relative size of N2 molecules and a path through which they

can enter the supercage in order to be adsorbed. Only monolayer adsorption is possible as volume

inside the cage is insufficient to accommodate additional molecules for multilayer adsorption. ..... 127

Figure 4.8. Schematic showing (a) possible effect of NaOH treatment on the pore structure within a

regular pyramid‐shaped NaY particle when no carbon is added. Material is removed by desilication

causing a series of walls in between the supercages to be washed away, joining together multiple


supercages to create a significantly wider cavity with larger pore volume and smaller surface area. (b)

Schematic showing the passage of gas molecules through zeolite cavities in absence of exterior walls.

............................................................................................................................................................ 128

Figure 4.9. Schematic showing (a) the effect of NaOH treatment when the zeolite is premixed with

carbon which causes the exterior of the zeolite particle to be possibly enclosed by porous carbon and

(b) multilayer adsorption taking place inside the zeolite cavity as the exterior carbon film enables

additional molecules to become trapped inside the enlarged cavity. ............................................... 129

Figure 4.10. SEM image of NaY. The texture of the particles is clean and smooth with sharp edges.

Pyramid‐shaped particles typical of zeolite Y can be clearly identified. ............................................. 131

Figure 4.11. SEM image of CEZ. The surface has become rough and coarse with terrains created by

desilication. Particles appear to be stacked on top of each other. Pyramid‐shaped particles can still be

detected. ............................................................................................................................................. 131

Figure 4.12. STEM image of CEZ with (a) no EDS mapping as well as EDS mapping for (b) C, (c) Si, (d) Al,

(e) Na and (f) O. Dark green areas in image (b) are more around the exterior of the particles confirming

that CEZ particles are surrounded by carbon. .................................................................................... 133

Figure 4.13. Pore size distribution based on surface area of CEZ, NaY and C‐R2030. ........................ 134

Figure A1. Distribution of unit cells within the MOF structure. When the pressure is (a) greater than

the breathing pressure, more unit cells are in an expanded state than in the collapsed state, i.e, 0.5

< ΦP (P > P0) ≤ 1, (b) equal to the breathing pressure, number of unit cells in expanded state is the

same as the number of unit cells in collapsed state, i.e., ΦP (P = P0) = 0.5, and (c) smaller than the

breathing pressure, fewer unit cells are in expanded state than in collapsed state, i.e, 0 ≤ ΦP (P < P0) <

0.5. ...................................................................................................................................................... 143


Figure A2. Residuals (nmeas ‐ ncalc) of the Pressure‐induced LJM model’s regression on experimental

data for the adsorption of CH4 on febdp reported by Mason et al. [48]. ........................................... 146


data for the desorption of CH4 on febdp reported by Mason et al. [48]. ........................................... 146


data for the adsorption of CH4 on cobdp reported by Mason et al. [48]. .......................................... 147


data for the desorption of CH4 on cobdp reported by Mason et al. [48]. .......................................... 147

Figure A6. Deviation of the Toth model from the pressure‐induced LJM model in comparison to

measured isothermal data at 303K for the adsorption of amino‐MIL‐53 (Al) on (a) CO2 and (b) CH4.

............................................................................................................................................................ 148

Figure A7. Comparison between the classical Toth model and the pressure‐induced LJM model used

as isotherm models in ASPEN Adsorption for performing column breakthrough simulations of an

equimolar CO2/CH4 mixture using amino‐MIL‐53 (Al) as the adsorbent at 303K with a feed pressure of

(a) 1.013 bar, (b) 11 bar and (c) 30 bar. ............................................................................................. 150

Figure A8. The Toth isotherm model regressed to (i) CH4 and (ii) CO2 adsorption data reported by

Cavenati et al. [81]. Symbols denote experimental data and curves represent the Toth model. ...... 151

Figure B1. Effect of carbon‐content. ................................................................................................... 154

Figure B2. Effect of NaOH concentration. ........................................................................................... 155

Figure B3. Toth Fitting of CEZ on N2. ................................................................................................... 156

Figure B4. Toth Fitting of C‐R2030 on N2. ........................................................................................... 157


Figure B5. Toth Fitting of NaY on N2. .................................................................................................. 157

Figure B6. Toth Fitting of CEZ on CH4. ................................................................................................. 158

Figure B7. Toth Fitting of C‐R2030 on CH4. ......................................................................................... 158

Figure B8. Toth Fitting of NaY on CH4. ................................................................................................ 159

Figure B9. Pore size distribution based on pore volume of CEZ, NaY and C‐R2030. .......................... 160

Figure B10. TEM image of CEZ. ........................................................................................................... 161

Figure B11. TEM image of CEZ. ........................................................................................................... 161

Figure B12. High Resolution TEM image of CEZ. ................................................................................. 162




Figure C1. (a) The 8MR and door‐keeping potassium cation movement during CO2 adsorption. (b)

Energy profiles for the potassium cation of r1KCHA and r3KCHA. (c) SEM images for r1.9KCHA. (d) XRD

patterns of raw fly ash, intermediate and resultant r1.9CHA. (The scales of the atoms in this illustration

are not in portion). .............................................................................................................................. 167

Figure C2. CO2, N2 and CH4 adsorption isotherms on r1.9KCHA at (a) 273, (b) 303 and (c) 333 K, and (d)

breakthrough curve for r1.9KCHA at 348 K and total pressures of 1 bar. .......................................... 170

Figure CS1. SEM micrograph of fly ash. .............................................................................................. 174

Figure CS2. Isobar of as‐synthesized r1.9KCHA at 0.5 bar. ................................................................. 174


Figure CS3. 3D structures of chabazite. .............................................................................................. 176

Figure CS4. Structure changes of potassium chabazite systems of r1KCHA and r3KCHA during CO2

adsorption process. The migration of the door‐keeping potassium cation (marked by the white cross)

at different steps was tracked to elucidiate the pore opening procedure in these two types of trapdoor

chabazites with different cation density. ............................................................................................ 176

Figure D1. Exemplar hysteretic sorption isotherms for CH4 on the MOFs (a) Fe(bdp) and (b) Co(bdp)

measured by Mason et al.[48] , illustrating the key parameters in eq (D1). ...................................... 180

Figure D2. Transition pressure ptr (a, c) and transition width, , (b, d) vs temperature, T, for the

sorption of CH4 on Fe(bdp) and Co(bdp) extracted from the data of Mason et al [48] with eq (D1).

Statistical uncertainties are shown as error bars in all panels but are smaller than the symbol for ptr in

(a) and (c). ........................................................................................................................................... 182

Figure D3. Equilibrium CO2 capacities, q, measured in adsorption (filled symbols) and desorption

(empty symbols) for ZIF‐7 at pressures, p, to 0.1 MPa, and eight temperatures T from (233 to 293) K.

Curves represent fits of the LJMY‐Langmuir model (eq. (D1) and described in the SI). ..................... 184

Figure D4. (a) Variation of transition pressures, ptr, with temperature and (b) relative to the value ptr0

at 303 K value for CO2 on ZIF‐7. (c) Variation of ptr with temperature and (d) with the transition

pressure measured in desorption, des . ..................................................................................... 185

Figure DS1. Hysteretic sorption isotherms for CH4 on the Fe(bdp) measured by Mason et al. [48]

Symbols: experimental data; curves: calculated with LJMY‐Langmuir model. (left) Adsorption or

desorption branch at each temperature was fitted separately. (right) The adsorption and desorption

branches at each temperature were fitted simultaneously. .............................................................. 198

Figure DS2. Hysteretic sorption isotherms for CH4 on the Co(bdp) measured by Mason et al. [48]

Symbols: experimental data; curves: calculated with LJMY‐Langmuir model. (left)Adsorption or


desorption branch at each temperature was fitted separately. (right) The adsorption and desorption

branches at each temperature were fitted simultaneously. .............................................................. 199

Figure DS3. Equilibrium CO2 capacities, q, measured in adsorption (filled symbols) and desorption

(empty symbols) for ZIF‐7. Curves represent fits of the LJMY‐Langmuir model with adsorption and

desorption branched at each temperature fitted simultaneously. .................................................... 200

Figure DS4. Temperature dependence and hysteresis of the transition width parameter, , reported

in Tables DS1 to DS3 when each branch is fit separately to eq (D1). ................................................. 208

Figure DS5. Graphical layout of the PVSA model built with Aspen Adsorption. ................................ 209

Figure DS6. Isotherm data and LJMY‐Langmuir fits for CO2 and CH4 on ZIF‐7 at 303 K. ..................... 211

Figure DS7. Simulated methane content of light product (P1 stream in Figure DS5) produced from an

equimolar feed of CH4 + CO2 using ZIF‐7 in a 2‐bed, 4‐step PVSA process with phigh = 2 bar. For every

desorption pressure (plow) considered the result of using each of the three isotherm models

(adsorption branch, desorption branch or full hysteresis) is shown. ................................................. 212


List of Tables

Table 1.1. Comparison of CO2 emissions per unit of energy produced from the major fossil fuels [10].

.............................................................................................................................................................. 30

Table 3.1. Structural and chemical composition of CBV100 and CBV720 parent zeolites. .................. 83

Table 3.2. Parameters for the Langmuir isotherm model. ................................................................... 86

Table 3.3. Comparison of enthalpy of adsorption values obtained in this work with values reported in

.............................................................................................................................................................. 88

Table 3.4. Structural and chemical composition of HY400 and HY600 parent zeolites. ....................... 97

Table 3.5. Comparison of CH4 capacity and CH4/N2 selectivity of Carbon‐E‐TMAY1 with competitor

adsorbents for CH4/N2 separation. ..................................................................................................... 114

Table 4.1. Results of ICP tests conducted on NaY and CEZ. Si/Al ratio was calculated from the results.

............................................................................................................................................................ 118

Table 4.2. Surface area measurements of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption.

Manufacturer's surface area for NaY (parent zeolite of TMAY ‐ CBV100) is available in Table 3.1. .. 121

Table 4.3. Pore volume measurements of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption. .. 121

Table 4.4. CH4/N2 Selectivity of CEZ, Carbon‐R2030 and NaY. ............................................................ 125

Table C1. Comparison of the performance of the adsorbing materials with highest CO2 selectivities.

............................................................................................................................................................ 171

Table CS1. The gas selectivities and capacities on chabazite. ............................................................ 177


Table DS1. Values and statistical standard uncertainties, u, of best fit parameters determined by

regression of the LJMY‐Langmuir model (eq (D1) or eq (DS3)) to the data of Mason et al. [48] for

sorption of CH4 on Fe(bdp), together with the standard error of the fit. ........................................... 194


regression of the LJMY‐Langmuir model (eq (D1) or eq (DS3)) to the data of Mason et al. [48] for

sorption of CH4 on Co(bdp), together with the standard error of the fit. .......................................... 195


regression of the LJMY‐Langmuir model (eq (D1) or eq(DS3)) to the data measured in this work and by

Arami‐Niya et al. [59] at 303 K for the sorption of CO2 on ZIF‐7, together with the standard error of the

fit. ........................................................................................................................................................ 196


regression of the LJMY‐Langmuir model (eq (D1) or eq(DS3)) to the data of Couck et al. [162] for

sorption of CO2 on CH4 on MIL‐53(Al), together with the standard error of the fit. .......................... 197

Table DS5. Equilibrium capacities, q, of CO2 on ZIF‐7 measured in adsorption and desorption as a

function of pressure, p, and temperature, T. The relative combined standard uncertainty of sorption

capacity, uc(qi)/qi, measured with this ASAP2020 was estimated previously [131, 161] to be 1.4 %201

Table DS6. PVSA cycle step sequence and timing ............................................................................... 210

Table DS7. Simulation parameters ...................................................................................................... 210

Table DS8. Values of the LJMY‐Langmuir sorption isotherm parameters for ZIF‐7 at 303 K used to

simulate a PVSA separation of an equimolar CO2 + CH4 mixture, based on the data reported by Aram‐

Niya et al. [163] and by Yang et al. [164]. ........................................................................................... 210

Table E1. Data points extracted graphically from Couck et al. [60] .................................................... 215


Table E2. Data points extracted graphically from Mason et al. [48] .................................................. 218

Table E3. Data points extracted graphically from Mason et al. [48] .................................................. 226

Table E4. Data table for experimental data presented in Figure 2.9 and Figure 2.10 ........................ 232

Table E5. Data points extracted graphically from McDonald et al. [64] ............................................. 238

Table E6. Data table for all experimental isotherm data presented in Chapter 3 .............................. 244

Table E7. Data table for all experimental isotherm data presented in Chapter 4 .............................. 264


Acknowledgements

I am profoundly indebted to the individuals whose contributions facilitated the timely completion of

this work. I would like to thank my Principal and Coordinating Supervisor Professor Eric May for

providing guidance, selecting key topics, reviewing written work and helping achieve milestones

during the entire duration of this research project. Despite being very busy with teaching

commitments and managing a large research group, Eric was always available to participate in

discussions, many of which were long and took place during the busiest times of the semester. He was

always patient in answering questions and actively contributed by providing valuable feedback.

I would also like to thank my Co‐Supervisor Professor Gang (Kevin) Li for his continuous suggestions

and feedback that helped achieve some of the key outcomes of this work. Kevin provided hands‐on

training with laboratory equipments and experimental procedures during the early stages of this work

and conducted tutorials and learning sessions that helped me build a strong foundation on the

fundamentals of adsorption. Kevin brings a wealth of knowledge in the field of adsorption which has

been invaluable to this research work.

In addition I would also like to express my gratitude to my parents, Mr Sayedur Rahman and Mrs

Suraiya Akhter, for their motivation and encouragement to continue my work through some of the

toughest times we went through as a family. Finally, I would like to acknowledge my daughter,

Nusaibah Shams Rahman and my wife Sharmin Alam for the immeasurable sacrifices they made to

accommodate my PhD in their lives. They were severely deprived of my time and involvement in

numerous family activities, yet they never complained. Without their support, it would have been

impossible to complete this work.

Despite the assistance of all the people mentioned and many others not mentioned, the errors that

still remain are my own.

This research has been supported by an Australian Government Research Training Program (RTP)

Scholarship.


Chapter 1

Introduction and Motivation

1.1 The rise of natural gas as a fuel

With rapid urbanisation and increasingly widespread use of technology around the world, global

energy consumption continues to grow. In 2018, world energy consumption was 2.9% higher than the

previous year and almost double the average annual growth rate of 1.5% over the last 10 years [1].

The largest contribution to this increase came from natural gas which accounted for more than 40%

of the growth [1, 2]. Globally natural gas consumption rose by 5.3% over the previous year while

production grew by 5.2% [1, 2]. These figures were the fastest annual growth rates recorded for

natural gas in the last 30 years and more than double the average 10‐year annual growth rate, as

shown in figure 1.1 [1].

Figure 1.1. Comparison of growth of natural gas consumption and production in 2018 with the previous 10‐year average annual growth rate and the major countries which contributed to the

growth [1].


The largest contribution to the increase in natural gas consumption was driven by growing demand in

power generation to replace coal‐fired power stations, especially in the US and China [3]. In the US,

15 gigawatts of coal‐fired power generation capacity was retired in 2018 while in China strong

environmental policies encouraging coal‐to‐gas switching have triggered a number of large scale

projects leading to a remarkable 18% increase in natural gas consumption [1, 3]. Figure 1.2 below

shows the historical as well as a projected decline in coal‐fired electricity generation in the US from

2010 – 2050 and the corresponding rise in natural gas as a source of power generation.

Figure 1.2. Comparison of the historical and projected use of different fuels to generate electricity in the US from 2010 – 2050 [3].

To meet the growing demand, it is inevitable that the supply of natural gas also has to increase

significantly. In 2018, the largest contribution to the increase in production came from the US shale

gas reserves in Mercellus, Haynesville and Permian. In addition, continuous rapid expansion in

liquefied natural gas (LNG) made a major contribution to the overall supply [1‐3]. A number of existing


LNG plants in Australia, US and Russia achieved substantial increase in production volumes while a

number of new plants went into production [1].

1.1.1 Natural gas in Australia

In recent years, Australia has played a major role in meeting the global demand for natural gas. In

FY2018‐19, LNG was Australia’s second largest export commodity by value after iron ore and had by

far the highest annual growth in both value and volume amongst all resource and energy commodities

[4]. A major portion of Australia’s LNG exports come from Western Australia (WA), which as of October

2019, has 65.5 trillion cubic feet (tcf) of proven (2P) offshore conventional natural gas reserves and an

additional 68.1 tcf of 2C conventional gas resources [5]. In 2018, WA contributed to 66% of Australia’s

total LNG exports and 14% of global LNG supply. Figure 1.3 shows the number of WA LNG export

projects in operation since 2015 and the growth in annual export capacity each of these contributes

to. Amongst these, the Prelude FLNG plant made its first LNG shipment in June 2019.

Figure 1.3. LNG export projects in operation in Western Australia since 2015 and the growth in the state’s annual LNG export capacity in million tonnes [5].


1.1.2 Environmental regulations

Over the last few decades there has been growing concern over the impact of climate change around

the world. A number of accords and agreements have been signed setting targets and limits for

signatory countries to reduce greenhouse gas emissions. The most well‐known amongst these are the

Kyoto Protocol (1997), the Copenhagen Accord (2009), and the Paris Agreement (2016), whose main

objective is to ensure that global temperature increase by the end of the century is kept well‐below 2

degrees Celsius above pre‐industrial levels [6]. To achieve this target, strict regulatory measures have

been undertaken by governments across the world to cut greenhouse gas emissions from major

industries such a power generation, manufacturing and chemical processing [7, 8].

The primary greenhouse gas whose atmospheric concentration is increasing at the highest rate as a

direct consequence of human activities is carbon dioxide [9]. Reducing CO2 emissions has therefore

been a key target for environmental regulations being implemented around the world. Table 1.1

below shows the comparison of CO2 emissions per unit (BTU) of energy produced from the major fossil

fuels including coal, oil and natural gas [10].

Table 1.1. Comparison of CO2 emissions per unit of energy produced from the major fossil fuels [10].

Fuel Pounds of CO2 emitted per million British Thermal Unit (BTU) of energy produced

Coal (anthracite) 228.6

Coal (bituminous) 205.7

Coal (lignite) 215.4

Coal (subbituminous) 214.3

Diesel fuel and heating oil

161.3

Gasoline (without ethanol)

157.2

Propane 139.0

Natural gas 117.0


As shown in the table above, amongst all fossil fuels, natural gas has the lowest CO2 emissions per unit

of energy produced. It is therefore considered to be the cleanest fossil fuel and has been a major

driver, along with clean air issue, of programs such as coal‐to‐gas switch in China and US, as described

above. For this same reason, rapid growth in the demand and consumption of natural gas is expected

to continue in the coming years.

1.2 The challenge of separating gas mixtures

As discussed above, one of the essential factors facilitating the growth of global supply and trade of

natural gas is LNG. This is because many of the large natural gas reserves are located far away from

the global markets and transporting in shipping vessels is far more economically viable than

constructing pipelines across oceans [11]. The operational schematic of a conventional LNG processing

plant is shown below in figure 1.4.

Figure 1.4. Conventional gas processing operations in a typical natural gas plant with a cryogenic process for liquefied natural gas production. MCHE = main cryogenic heat exchanger [12].


As shown above, before a feed of natural gas can enter the liquefaction unit of an LNG processing

plant, its impurity contents, mainly carbon dioxide and nitrogen, must be reduced significantly to

deliver a high concentration methane stream. Most feed gas specifications require CO2 contents to

be less than 50 ppmv and N2 concentration to be under 1% [13, 14]. This is because CO2 freezes at

cryogenic temperatures to form dry ice which can cause blockage and N2 has no heating value and can

cause stratification and rollover of the liquid product during shipping. However, most natural gas

reserves around the world contain much higher impurities, typically over 3% CO2 and 4% N2 [13].

In order to meet the feed gas specifications of LNG plants, CO2 and N2 need to be selectively removed

from natural gas. The process of separating these components from a mixture of several gases is not

trivial. Conventionally, CO2 is removed using the method of amine absorption where the mixture is

treated with aqueous amine solution and the resulting CO2‐reach acidic by‐product is rejected through

acid gas rejection units while the remaining hydrated contents are passed through dehydration loops

before delivering into the liquefaction plant [15]. N2 rejection takes place at low temperatures through

the process of cryogenic distillation which uses the difference in relative volatility of CH4 and N2 at

high pressure conditions to separate N2 + CH4 mixtures [16, 17].

Both amine absorption and cryogenic distillation are energy‐intensive processes and raise the unit

cost of LNG production. For amine absorption, gas processing costs range from USD 0.18 to USD 0.31

per Mscf product [18]. The major contributions come from the additional power consumption

required to (i) maintain a continuous sorbent circulation rate determined by the loading capacity of

the sorbent and the amount of CO2 that needs to be removed and (ii) regenerate the amine sorbent

[13]. For cryogenic distillation, the main added operating cost arises from the additional power needed

to extend the high pressure conditions, required for N2/CH4 separation, to the nitrogen rejection unit

[16, 19]. Typical power ratings of conventional cryogenic distillation columns have been reported to

range between 360 MW and 386 MW [20].


1.3 Adsorptive separation

Due to the additional operating costs associated with amine absorption and cryogenic distillation,

alternative cost‐effective gas separation methods are sought by the natural gas industry. In particular,

adsorption‐based separation methods are considered as possible cost‐effective solutions [21‐24]. The

application of adsorption as a method for separating mixtures into two or more streams, each

enriched with a valuable component to be recovered, began in the early 1950s [25]. Initially used for

the recovery of aromatic hydrocarbons, adsorption based processes became popular in the 1970s for

separating close boiling point crude oil components. With increasing energy prices, adsorption, which

in principle was the more economic separation method, competed with distillation – the conventional

method for separating crude oil fractions [25].

The two most promising adsorptive separation methods that can potentially replace amine absorption

and cryogenic distillation in the LNG industry include pressure swing adsorption (PSA) and

temperature swing adsorption (TSA). Both these processes use cycles consisting of adsorption to load

gas onto the adsorbent material and desorption to regenerate the adsorbents. To separate gas

mixtures, PSA relies on the difference in adsorption capacity at different pressures under isothermal

conditions while TSA makes uses of the difference in capacity at different temperatures under isobaric

conditions. A schematic representation of a two‐bed cyclic separation process and the capacity

difference used by PSA/TSA are shown in figure 1.5.

Under isothermal conditions, adsorption occurs when the pressure is increased and desorption occurs

when the pressure is lowered again. The PSA method therefore uses depressurisation to regenerate

the adsorbent bed and is able to operate with short cycle times, in the order of minutes or even

seconds, provided that the kinetics/mass‐transfer properties of the adsorbent being used is able to

support rapid loading/unloading [26]. This means, with the suitable adsorbent, a PSA process can

deliver faster separation using a smaller amount of adsorbent compared to an equivalent TSA process

[12].


Figure 1.5. (a) Schematic representation of a two‐bed cyclic adsorption process with (b) pressure swing adsorption (PSA) and temperature swing adsorption (TSA) cycles shown on a typical isotherm [12].

1.3.1 The need for adsorbents with high capacity

Success of any commercial scale adsorption process depends crucially on the availability of suitable

adsorbents in tonnage quantities at economic cost [25]. This need for adsorbents has stimulated, over

the last 40 years, fundamental research that led to the development of a range of commercially

available materials such as activated carbon, silica gel, molecular sieves, zeolites, metal organic

frameworks (MOFs), covalent organic frameworks (COFs) etc. Adsorption‐based separation processes

are heavily dependent on the adsorptive characteristics of these adsorbents. Some of the key

desirable properties include:

High capacity for the target gas

High selectivity for the target gas mixture

High mass transfer coefficient to ensure sustainable adsorption kinetics

Regenerability to ensure continuous performance upon repeated load/unload cycles

(a) (b)


Adsorption capacity is the amount of gas that can be loaded onto the solid adsorbent upon reaching

steady‐state equilibrium at a certain pressure under isothermal conditions. In a PSA process, the

adsorbent needs to have a high capacity for the target gas so that large amounts of it can be captured

in each cycle, leading to a smaller quantity of adsorbent material needed or a lesser number of cycles

required to achieve the desired separation. Either way, this means the associated operating costs

needed for the separation process are reduced.

Wide range of studies are available where capacity of common adsorbents such as zeolites and carbon

have been measured to determine and compare their suitability for use in PSA processes [27‐30].

Contemporary research has also focused on implementing changes within existing adsorbent

structures to achieve enhanced capacity [31‐33] and developing emerging adsorbent materials such

as metal organic frameworks (MOFs) with increased adsorption capacities [34‐37]. While adsorptive

properties of MOFs are often more desirable than zeolites and carbon, the availability and production

cost of MOFs are usually not as favourable. Yilmaz et al [38] provide a detail analysis of factors that

affect large scale synthesis of MOFs. The main considerations are cost and availability of raw

materials/reagents and the space‐time‐yield (STY) of synthesis. The cost of production is usually

significantly higher than adsorbents such as zeolites and carbon, even when the reagents are available

in abundant quantities. However, favourable separation characteristics can mean that a MOF is a

preferred adsorbent for certain applications, despite its high synthesis cost.

1.3.2 The need for adsorbents with high selectivity

Selectivity is a measure of an adsorbent’s relative affinity towards a particular gas over another gas in

a mixture of two or more gases. Some factors that affect selectivity include the relative molecular

diameters of the gas molecules in comparison to the size of the adsorbent’s pore opening and the

electrostatic interaction between the gas and the solid materials. When exposed to a mixture of two

gases, an adsorbent material will adsorb higher amount of the gas it is more selective towards and

lower amount of the gas it is less selective towards. Hence, to capture nitrogen from a typical mixture


of methane and nitrogen present in natural gas, under identical conditions, a PSA process using an

adsorbent with a higher N2/CH4 selectivity will require fewer cycles and less time to achieve its target

separation in comparison to a PSA process using the same amount of another adsorbent with a lower

N2/CH4 selectivity. Thus, as with higher capacity, using adsorbents with higher selectivity for the target

gas also leads to savings in operating costs as the same separation performance can be delivered

within a shorter time or using smaller quantity of adsorbent.

Achieving highly selective adsorption has been a key driver behind the development and evolution of

materials such as metal organic frameworks (MOF’s) and covalent organic frameworks (COF’s) [39‐

43]. For large scale processes such as LNG, the preference is usually to use adsorbents such as zeolites

and carbon which can mostly be obtained from natural sources with minimal treatment. However,

substantial difference in capacity, selectivity and kinetics can also make synthetic materials such as

MOFs and COFs economically more favourable than zeolites and carbon.

1.3.3 The need for reliable isotherm models

Our understandings of the governing relationships that dictate the interaction between the solid

surface of an adsorbent and the gas molecules being adsorbed can be described in functional forms

generally known as isotherm models. Classical isotherm models, such as the Langmuir [44], BET [45],

and Toth [46] models have been derived by applying thermodynamic equations of state to the

adsorption equilibrium while making certain assumptions in regards to the phase of the adsorbed

layer. These models have been generally able to quantify the adsorption behaviour of most common

adsorbents such as zeolites, carbon, silica etc over a wide range of temperature and pressure.

However, some well‐known emerging materials such as MIL‐53 (metal organic framework) and ZIF‐7

(zeolitic imidazole framework), which have demonstrated extraordinary adsorption capacity and

selectivity, exhibit adsorption behaviours very different from those of carbon or zeolites. As a result

of the flexibility in their structure, these materials undergo a phenomenon known as “breathing”

which cause a step‐shape in their adsorption/desorption isotherms. Classical isotherm models are


generally unable to capture this behavior and are therefore not suitable for these emerging

adsorbents.

In an adsorptive separation process such as PSA, isotherm models are important for obtaining reliable

process simulation results which are used to design and optimise the performance of the PSA under

different operating conditions. Hence, in order to utilise the adsorptive characteristics of emerging

flexible adsorbents such as breathing MOF’s in PSA processes, an isotherm model capable of

quantifying adsorption/desorption behaviour of these materials within a wide range of temperature

and pressure is required.

1.4 The objectives of this research

This research project aims to support the implementation of novel pressure swing adsorption

separation technologies in the natural gas industry through the following objectives:

Develop an isotherm model capable of quantifying changes in adsorption capacity in response

to pressure‐induced structural transitions in flexible adsorbents

Correlate hysteresis in breathing adsorbents with temperature and validate experimentally

Develop a simple method to substantially enhance nitrogen adsorption capacity of a

commercially available Faujasite type Zeolite Y

Identify the structural modifications in the novel adsorbent and propose a mechanism to

explain how these changes cause the observed enhancements in capacity

In Chapter 2, an isotherm model capable of quantifying adsorption/desorption behaviour, including

hysteresis, caused by structural transitions in flexible metal organic frameworks, is developed. The

proposed model uses a functional form that incorporates a step change in adsorption capacity without

introducing any discontinuities. The model is validated by regressing experimental isotherm data,

collected from the literature as well as from laboratory measurements, for a number of well‐known

flexible adsorbents over a wide range of temperature and pressure. Using regression parameters, the


model is then used to correlate the magnitude of hysteresis in flexible adsorbents with temperature,

leading to conditions where hysteresis reaches a limiting asymptotic value. Finally, to demonstrate

the model’s suitability for use in industrial process simulations, a commercial software package is used

to incorporate the proposed functional form as an isotherm model to define the adsorption bed for

simulating the separation of a binary gas mixture, first using a dynamic column breakthrough

apparatus and then using PSA. The results obtained are compared with experimental data from the

literature as well as with results from simulations conducted under identical conditions but using a

classical counterpart of the proposed model.

Chapter 3 focuses on material development where the effects of a range of chemical treatments on

the methane and nitrogen adsorption capacity of some common naturally sourced variants of zeolite

and carbon are studied experimentally. Measurements of isothermal adsorption capacities are

presented for a number of novel zeolite‐based and carbon‐based adsorbents synthesised using

different wet chemistry procedures. Structural properties such as BET surface area and chemical

compositions such as silicon‐to‐aluminum ratio and the type of cation that affect the adsorption in

zeolites are discussed with reference to the observed changes in capacity. Thermodynamic properties

such as enthalpy of adsorption are also measured and compared in the different variants of zeolites

used in the study. In addition to capacity, changes in selectivity due to the presence of organic cations

such as tetramethyl ammonium are also studied in both zeolites and carbon. A number of methods to

successfully functionalise carbon surfaces in order to introduce organic cations are presented and the

resulting changes in methane/nitrogen selectivity are analysed and explained using characterisation

techniques such as scanning electron microscopy. The theory and literature review pertaining to this

concept has also been discussed in detail.

In Chapter 4, a simple laboratory method to synthesise a novel zeolite‐based adsorbent producing

remarkable enhancement in nitrogen adsorption capacity is presented. This material, a carbon

enhanced zeolite (CEZ), was synthesised by encapsulating a readily available variant of a naturally


occurring faujasite type zeolite into an external mesoporous carbon film. A wide range of

characterisation techniques including inductively coupled plasma analysis, x‐ray diffraction, nitrogen

sorption at 77K, scanning electron microscopy and scanning transition electron microscopy are used

to identify the structural changes responsible for capacity enhancement in this material. Furthermore,

a mechanism explaining the observed gain in adsorption capacity is proposed and validated by the

results of the characterisation experiments.

Finally, in Chapter 5, the key outcomes of this work are summarised. The relevance of the findings to

the natural gas industry are clearly identified and how the results can be used to address a number of

industrial challenges are described. Future research directions to further progress the outcomes of

this work are also suggested.


Chapter 2

Structural Transitions in Metal Organic Frameworks

Foreword

This chapter forms the basis of a manuscript that has been prepared for submission to the Journal of

American Chemical Society (JACS). The manuscript has been reproduced in Appendix D.

2.1 Introduction

The development and study of flexible porous materials, in particular metal‐organic frameworks

(MOFs), has been a major focus in the adsorption community in recent years. Interest in these

materials, from a process perspective, can be attributed to their potential use in a wide range of

applications in gas separation, gas storage, and catalysis [47‐53]. Owing to the numerous different

methods by which these structures have been designed and engineered for their desired applications

[54, 55], the reported adsorption characteristics of most MOF’s are fundamentally different from

those exhibited by conventional adsorbents such as molecular sieves, zeolites and activated carbon.

In flexible MOFs, the host‐guest interaction affects the host’s pore characteristics and the associated

pore accessibility by the guest molecule. The interaction may be classified as one of two mechanisms:

(1) a crystallographic phase transition that results in unit cell volume change (breathing or swelling

effect [56, 57]), or (2) a linker rotation and sub‐net sliding where there is no change in the unit cell

volume (gate‐opening). In both cases, the ability of these “soft materials” to change structure means

their observed adsorption characteristics are fundamentally different to those exhibited by

conventional adsorbents such as molecular sieves, zeolites and activated carbon[54, 55]. One of the

most obvious differences is the drastic increase in sorption capacity that occurs when the number of

accessible adsorption sites in the framework suddenly changes in response to external stimuli such as

temperature, pressure or the adsorption/desorption of guest molecules[56, 57]. This feature of


flexible MOFs allows the adsorption of guest molecules seemingly larger than the nominal

crystallographic pore diameter to suddenly increase. Such materials offer, in principle, a very high

selectivity and could potentially be very useful in the development of new gas separation technologies.

Although adsorption/desorption behaviours of several flexible MOFs have been studied extensively,

quantifying the shape of the isotherms and predicting the phase transition pressure for a given host‐

guest molecule system still remains challenging. An isotherm model with a consistent and physically

meaningful set of parameters is essential for reliable simulation of any adsorption‐based industrial

process utilizing the breathing MOF, such as the gas separation process that often motivates the

material’s development. Use of classical isotherm models such as the Langmuir, Toth and the BET

models, which work well for conventional adsorbents, are likely to produce erroneous process

simulation results for flexible adsorbents, especially near the region of the isotherm that corresponds

to the structural transition where the classical model deviates from the experimental data, as shown

in figure 2.1. Some developers of flexible MOFs have attempted to use combinations of classical

functional forms to regress breathing isotherm data; however, these resulted in piece‐wise definitions

of classical adsorption models being stitched together with inconsistent parameters produced by

fitting to different parts of the breathing isotherm [48, 58, 59]. Arami‐Niya et al. [59] used the

Langmuir‐Freundlich model to regress experimental isotherm data for the adsorption of CH4 and CO2

on ZIF‐7. However, in addition to difficulties in fitting, this model requires an excessive number of

fitting parameters and is cumbersome for use in process simulation design. Moreover, this approach

does not readily lend itself to insight about the nature of the material’s behaviour or predictions of

key properties at conditions far from the measured data.


Figure 2.1. Deviation of the classical Toth model from isothermal experimental data [60] at 303K for

the adsorption of (a) CO2 and (b) CH4 on amino‐MIL‐53 (Al).

Recently, Li et al. [61] proposed an approach for quantifying temperature‐regulated guest admission

in adsorbents such as trapdoor zeolites based on the energy distribution of oscillators within an

adsorbent structure. This method, termed as the Li‐Jensen‐May (LJM) approach, led to a functional

form capable of quantitatively describing observed adsorption isobars for materials that exhibit

temperature‐dependent adsorption. In this work, we have used an adaptation of the LJM approach

to propose an empirical isotherm model capable of providing a more coherent description of the

observed breathing isotherms, avoiding piece‐wise definition or any discontinuity. Experimental

isotherm data for different flexible MOFs from the literature were successfully regressed to this new

functional form, termed as the “pressure‐induced LJM model”. For isotherms with hysteresis, this

model can be used to quantify desorption together with adsorption. This means hysteresis associated

with breathing transitions can be analyzed as a temperature dependent effect and consequently leads

to the prediction of a temperature at which hysteresis becomes very small. Finally, the model is used

as an isotherm model to simulate first the binary breakthrough and then the pressure swing

adsorption separation of a gas mixture using a commercial software package to demonstrate the

model’s suitability for use in industrial process simulations. The results obtained are compared with

experimental data from literature as well as with results from simulations conducted under identical

conditions but using a classical counterpart of the proposed model.


2.2 Development of the pressure‐induced LJM model

The LJM approach originally proposed to quantify temperature‐regulated guest admission exhibited

by adsorbents such as trapdoor zeolites where, contrary to normal physiosorption, an adsorption

isobar features a point of maximum capacity beyond which the capacity decreases with decreasing

temperature [61]. In these materials, admission to adsorption sites occurs through interactions

between guest molecules and oscillating pore‐keeping cations that obstruct the entrance to the

interior of a unit cell. Admission of the guest molecules is based on its interaction with the door‐

keeping cation (oscillator). At a certain condition, the guest molecule induces temporary deviations in

the cation from its original position, providing sufficient space for the guest molecule to enter the pore

that contains the majority of adsorption sites. For the oscillator to deviate from its preferred site and

allow a guest molecule to be admitted, the combined guest‐host system must have enough thermal

energy to overcome the energy barrier associated with the temporary deviation of the host that is

sufficient to allow the guest access to the pore.

Structural change in flexible MOFs is a highly guest dependent transition between collapsed or narrow

pore (np) and expanded or large pore (lp) phases that is characterised by gate‐opening and gate‐

closing pressures [62]. Thus, analogous to temperature‐regulated guest admission in trapdoor

adsorbents, breathing effect in MOFs can be defined as a pressure dependent phenomenon [56, 57,

63]. Hence, a functional form for pressure‐induced adsorption can be derived using a similar approach

used for deriving the temperature‐induced model. An expression for the adsorption capacity of a

breathing adsorbent obtained by incorporating the LJM form into the classical Toth model is given in

equation Eq. 2.1 below. The full derivation is available in Appendix A1.

1,

√1 (Eq. 2.1)

In equation Eq. 2.1, np is the adsorption capacity, , is the theoretical maximum value that the

breathing pressure can have when temperature becomes large, k is the breathing pressure coefficient,


βP is a measure of the distribution width, is the fraction of external surface sites and defects always

available for adsorption, is the maximum monolayer adsorption capacity, b0 is the gas‐solid affinity

coefficient, ΔH is the enthalpy of adsorption, R is the gas constant, m is the surface heterogeneity

coefficient and T and P respectively are the temperature and pressure, as shown in equation Eq. 2.1,

to the parameters present within the standard adsorption equation. A summary of all model

parameters considered in this work, including their units is provided in Appendix A2.

The pressure‐induced form of the LJM model, as described by equation Eq. 2.1, was used in the next

sections to quantify isothermal adsorption capacity of a number of breathing MOFs that exhibit

stepped isotherms as a result of phase transition of these materials. The experimental data used for

this quantitative study were obtained from multiple sources published in the open literature [48, 60,

64]. In order to validate the model, attempts were made to fit different forms of breathing isotherms

reported in the literature. These are 1‐ CH4 adsorption/desorption measurements on Fe(bdp) and

Co(bdp) in a temperature range of 248‐323 K and pressures up to 7000 kPa, reported by Mason et

al.[48]; 2‐ CO2 adsorption/desorption measurements on mmen‐Fe2(dobpdc) and mmen‐Co2(dobpdc)

in a temperature range of 298‐323 K and pressures up to 100 kPa [64]; 3‐ CO2 adsorption isotherms

on amino‐MIL‐53(Al) at 288‐303 K and pressures up to 2900 kPa [60]. For amino‐MIL‐53(Al)),

experimental data for the column breakthrough separation of a CH4/CO2 mixture using were also

reported [60], allowing for comparisons to be made with simulations performed using the pressure‐

induced LJM model to quantify isothermal breathing behavior. In addition, we conducted experiments

in our own laboratory to measure isothermal adsorption/desorption of CO2 on ZIF‐7 up to a pressure

of 100 kPa at temperatures ranging from 233 K – 253 K and made attempts to regress the pressure‐

induced LJM model on the obtained data.

2.3 Modelling stepped breathing isotherms

The stepped breathing isotherm, observed with many flexible adsorbents, refers to an adsorption

isotherm where the amount of gas adsorbed is small at low pressures but rises sharply as the pressure


approaches P0 and thereafter continues to rise further before forming a plateau [48, 56, 57, 63]. This

isotherm shape has been attributed to a two‐stage adsorption process caused by the framework’s

transition from np to lp phases. As the physical dimensions and characteristics of the pores and cages

are different in the two phases, two types of adsorption sites become available for guest molecules

[65]. The first stage in the adsorption isotherm where uptake is low takes place because at low

pressures the framework is mostly in the np phase. In this phase, most of the unit cells are in a

collapsed stage making the pore surface area and volume available for adsorption to be low. The

energy released due to adsorption in this stage is used to overcome the interfacial energy penalty

required to create new surface area and cause pore expansion (np lp) whereby surface area and

volume increase dramatically making additional cavities and adsorption sites accessible to the gas

molecules for the second stage of adsorption to take place [66]. The specific shape of the isotherm in

the two adsorption stages depends on types of adsorption sites present in the framework being

considered. In the following examples, we investigate the compatibility of the pressure‐induced LJM

model in quantifying different forms of stepped isotherms exhibited by a number of breathing MOFs.

2.3.1 CH4 on Fe(bdp)

Mason et al. conducted a set of CH4 volumetric adsorption and desorption experiments for the metal‐

organic framework Iron(II)‐1,4‐benzenedipyrazolate, abbreviated as Fe(bdp) that shows noticeably

high capacity for CH4 adsorption. 3‐D schematics of this compound in both collapsed and expanded

forms are shown in figure 2.2. Measurements of isothermal excess adsorption and desorption

capacities at low and high pressures (up to 7000 kPa) are reported and cover a wide range of

temperatures from 248 K to 323 K [48]. The broad hysteretic desorption of CH4 from Fe(bdp) is used

to suggest an application for practical gas storage since the gas adsorbed at high pressure can be

stored at lower pressures. However, in a gas separation process by pressure swing adsorption (PSA),

a significant adsorption‐desorption hysteresis loop may affect the low‐pressure release stage. It has

been suggested that significant hysteretic desorption seen in stepped breathing isotherms of hybrid

MOFs is caused mainly by the strong interactions between the adsorbed guest molecules and the host


[67]; nevertheless, there is still no practical tool to predict the hysteretic desorption pressure of these

materials in the presence of a guest molecule and at a certain temperature.

Figure 2.2. 3D depiction of (a) solid‐state structure of the collapsed phase of Fe(bdp) under vacuum at 298K and (b) idealized average solid‐state structure of the expanded phase of Fe(bdp) under 50

bar pressure at 298K [48].

Figure 2.3 (a) and (b) show the pressure‐induced LJM model represented by equation Eq. 2.1 regressed

on experimental data respectively for isothermal adsorption and desorption of CH4 on Fe(bdp)

reported by Mason et al. [48]

(a) Collapsed Fe(bdp)

(b) Expanded Fe(bdp)


Figure 2.3. Demonstration of the pressure‐induced LJM model’s ability to quantify breathing isotherms for CH4 (a) adsorption and (b) desorption on Fe(bdp) at seven temperatures reported in

reference [48]. Solid and hollow symbols denote, respectively, adsorption and desorption experimental data. Solid curves represent the pressure‐induced LJM model.

The fitting exercise performed above shows that the regressed pressure‐induced LJM model can

provide a good representation of the measured data. The quality of the regression is validated by the

ability of the model to fit a total of 14 isotherms (seven each for adsorption and desorption)

encompassing a range of temperatures between 248 K and 323 K. The regression had an R2 of 0.996

for adsorption and 0.990 for desorption with values of all parameters, listed in Appendix A3, being

physically meaningful. The statistical uncertainties of regression in the parameter values are also listed

in Appendix A3. In particular, the model predicts an enthalpy of adsorption, ‐ΔH, the value of 12.8

kJ/mol, which is near the 12.7 kJ/mol (expanded) value reported in the experimental work [48]. The

value of the other classical parameter n∞ was above 20 mol/kg which was expected as the material

was designed to have a high CH4 capacity. The Toth parameter m had a value of 1 indicating that the

classical part of the model was essentially the Langmuir model.

A unique feature of this modeling approach lies in its ability to connect desorption with adsorption for

materials that exhibit hysteresis. In the absence of hysteresis, a desorption isotherm should, in theory,

be quantifiable by the same parameter values that describe the corresponding adsorption isotherm

[25]. Hence, when hysteresis is present, it is necessary to identify the parameters that would be

affected by hysteresis. Among the parameters used in the above model, only P0,ref, k and βP are

(b)


affected by hysteresis as ε is a measure the adsorption on the external surface and the other

parameters are inherited from the classical adsorption equations which are derived for cycles with no

hysteresis. Therefore, only these three parameters were allowed to vary during the regression of the

desorption isotherms while all other parameters were fixed at values obtained from the regression of

the adsorption isotherms. Thus, the fit presented in figure 2.3(b) is essentially a three‐parameter fit.

Following a similar approach used by Du et al. [66], the natural logarithm of the breathing pressure P0

obtained from the above regression was plotted against 1/T for both adsorption and desorption, as

shown in figure 2.4. The plot shows that the points fall on a straight line with a negative slope for both

adsorption (breathing‐in) and desorption (breathing‐out), justifying the use of the Clausius‐Clapyron

equation to model breathing pressure as a function of temperature. A detailed derivation of this

approach that aligns with Du et al. [66] has been provided in Appendix A1.

Another remarkable observation lies in the fact that the slopes of the two straight lines in figure 2.4

differ with the adsorption branch being steeper than the desorption branch. This means that the two

lines can be extended to meet at a unique point, which we term as the point of intersection. This point

corresponds to a temperature, Ti, at which the breathing‐in transition pressure, P0,ads, and the

breathing‐out transition pressure, P0,des, become equal to each other. For CH4 adsorption/desorption

on Fe(bdp), this temperature is Ti ~ 115K, which is just above the liquefaction temperature of CH4 at

atmospheric pressure, making it extremely difficult to conduct an adsorption/desorption experiment

in the gas phase.

The model parameter k is equivalent to the slope of the straight lines shown in figure 2.4 and is termed

as the breathing pressure coefficient. For materials that exhibit hysteresis, this quantity is different

for adsorption as opposed to desorption. Hence, in subsequent discussions, either kads or kdes will be

used instead of k while referring to the breathing pressure coefficient to differentiate between

adsorption and desorption, respectively.


Figure 2.4. Plots of P0 vs. temperature for adsorption (breathing‐in) and desorption (breathing‐out). The plots indicate that P0 = f (t) is linear for this data set. The two straight lines through the plots

meet at a point, where the breathing‐in transition pressure, P0,ads, becomes equal to the breathing‐out transition pressure, P0,des.

To further study hysteresis as a function of temperature, CH4 adsorption and desorption isotherms for

Fe(bdp) predicted by the pressure‐induced LJM model at nine different temperatures were plotted.

Seven of these plots are shown in figure 2.5 (a) – (e). Among these, the plots at 323 K and 248 K

correspond to the fittings to experimental data presented in figure 2.3 above, while the plots at 175

K, 115 K and 100 K are theoretical isotherms predicted by the pressure‐induced LJM model. The plot

at 115 K corresponds to the point of intersection between the breathing‐in and the breathing‐out

branches shown in figure 2.4. Considering the area enclosed between the adsorption and the

desorption isotherms as a measure of hysteresis, the plots clearly show a trend of diminishing

hysteresis with decreasing temperature. In order to quantify hysteresis, this enclosed area is

calculated graphically for each pairs of isotherms and plotted against temperature in figure 2.5 (f). The

plot confirms that hysteresis does in fact decrease with temperature and appears to approach an

asymptote at temperatures below 115 K, which is the temperature at which the transition pressure

during adsorption, P0,ads, becomes equal to the transition pressure during desorption P0,des. An

exponential trendline can be fitted through the points, as shown, confirming that the magnitude of

hysteresis approaches an asymptotic value at low temperatures.

y = ‐1065.7x + 11.438

y = ‐935.19x + 10.3

0

2

4

6

8

10

0.003 0.004 0.005 0.006 0.007 0.008 0.009

ln(P

0)

1/T (1/K)

Adsorption

Desorption

Point of intersection


Figure 2.5. (a) – (e) Adsorption and desorption isotherms for CH4 on Fe(bdp) at temperatures of (a) 323K, (b) 248K obtained from the fitting to experimental data and (c) 175K, (d) 115K, and (e) 100K projected by the pressure‐induced LJM model. The adsorption and desorption transition pressures, P0,ads and P0,des are shown in (a) where they are farthest apart and (d) where they are equal at the

point of intersection. (f) Area enclosed by each pair of adsorption and desorption isotherms calculated graphically for each of the temperatures shown in (a) – (e) plus four additional

temperatures.

(a) (b)

(c) (d)

(e) (f)


2.3.2 A thermodynamic model for hysteresis

Hysteresis is explained by pore opening np lp transitions being different to the reverse pore closing

lp np transitions. This is attributed to the fact that, once expanded to lp state under pressure, the

material remains in lp state even when the pressure is reduced and does not contract back to np state

until the pressure is sufficiently reduced and reaches a threshold contraction pressure. From

thermodynamic considerations, a host free energy penalty needs to be paid for the creation of new

surface when expansion occurs from np lp during adsorption. For the reverse transition from lp

np, no new surface needs to be created and hence no host free energy penalty is required.

Schneeman et al. [57] depict this as an energy barrier between the np and lp phases that can be used

to describe hysteresis qualitatively. A detailed discussion on the free energy method to analyse

structural transition behaviour is presented by Numaguchi et al. [68] while a full derivation of the host

free energy equations is provided by Coudert et al. [69]. Using a similar approach, we attempt to draw

host energy vs deformation diagrams for isothermal adsorption/desorption cycles of breathing

adsorbents in figure 2.6. The first diagram in figure 2.6(a) shows the scenario where we expect to see

hysteresis remain constant at a low asymptotic value as the temperature is less than or equal to the

temperature, Ti, at the point of intersection. In this case, the material is already predominantly in the

lp phase even before any adsorption. However, a small fraction of the framework still remains in the

np phase regardless of how low the temperature is. Thus only a small host free energy penalty is

required to create the extra surface and expand these few collapsed cells during adsorption. During

desorption, no host free energy penalty is required and hence the energy paths taken during

adsorption and desorption become very close to each other giving rise to a minimal amount of

hysteresis.

Figure 2.6(b) shows the energy path followed during adsorption and desorption when the

temperature is greater than Ti. A larger fraction of unit cells is now in the np phase and therefore, a

higher host free energy penalty is required for np lp transition during adsorption to create the


additional surface in the lp phase. However, during desorption, the shortest energy path is followed

for the lp np transition as there is no need for the creation of any additional surface and hence no

energy barrier. Thus, the breathing‐out pressure becomes smaller than the breathing‐in pressure

giving rise to hysteresis.

The situation when the temperature is even higher is shown in figure 2.6 (c). In this case, a larger

fraction of unit cells is in the np phase and hence a larger surface needs to be created for np lp

transition during adsorption requiring a larger host free energy penalty. However, the reverse lp np

transition during desorption still follows the shortest energy path requiring no host free energy

penalty. Therefore, the difference between the breathing‐in and breathing‐out pressures is higher,

manifesting in a larger hysteresis.

To summarise, the hysteresis observed in a MOF’s stepped isotherm is due to the energy penalty

associated with increasing host surface area as np lp. No such energy barrier needs to be overcome

for the reverse lp np structural change: hence the isotherm’s desorption branch corresponds to

thermodynamic equilibrium, while the adsorption branch reflects the system accessing metastable

states. This indicates that the np lp structural transition is entropically driven with the lp state being

less‐ordered than the np state. Consequently, the np lp transition should be endothermic and can

only take place when the magnitude of the enthalpy of adsorption is greater than the magnitude of

the enthalpy associated with the np lp transition [66, 70]. The reverse transition lp np should be

exothermic.


Figure 2.6. Energy paths followed during adsorption/desorption cycles when (a) T ≤ Ti where hysteresis is at its low asymptote as the energy paths during adsorption and desorption are closest to each other, (b) T > Ti giving rise to hysteresis as the energy path during adsorption is longer than

that during desorption and (c) T >> Ti where the effect of hysteresis is larger as the difference between the energy paths during adsorption and desorption are greater.

Hos

t fre

e en

ergy

Deformation

lp np

Host free energy penalty

(b) T > Ti

Hos

t fre

e en

ergy

Deformation

lp np Host free energy penalty

(c) T >> Ti

Hos

t fre

e en

ergy

Deformation

lp

(a) T ≤ Ti

Host free energy penalty

np


The prediction of Ti is an interesting and useful artefact of the pressure‐induced LJM model. Although

it does not directly incorporate the physical mechanism causing the hysteresis, it is useful for

predicting the extent to which hysteresis can be minimized in a particular breathing material. This can

become an important design parameter should a particular process using the adsorbent consider

hysteresis as undesirable.

The breathing pressure coefficient, k, is a measure of the effect of temperature on the phase transition

of flexible adsorbents. As can be seen in figure 2.4, the slope of breathing‐out pressures (lp np) vs.

temperature is smaller than that for the breathing‐in pressures (np lp) vs. temperature. That is:

kdes < kads

This means that,

When T > Ti, then P0,ads > P0,des

When T < Ti, then P0,ads < P0,des

When T = Ti, then P0,ads = P0,des

Structurally, the difference in kads and kdes implies that the effect of temperature on phase transitions

between np and lp phases are not the same during adsorption and desorption. When P0,ads > P0,des, the

transition from collapsed to expanded state (np lp) is retained even when the stress is relaxed, and

the pressure falls below the breathing‐in pressure which caused the transition in the first place. Only

when the stress is further relaxed, and the pressure becomes sufficiently lower and reaches the

breathing‐out pressure does the reverse transition from the expanded to the collapsed state (lp np)

take place. However, as the temperature is reduced, the framework’s tendency to retain its expanded

state begins to diminish. Thus, the difference between the breathing‐in and breathing‐out pressure

also decreases and the adsorption and the desorption branches converge. Consequently, kdes < kads

becomes a necessary condition for hysteresis to appear in breathing isotherms.


2.3.3 CH4 on Co(bdp)

As another example, a similar approach was taken to model CH4 adsorption and desorption on another

flexible MOF, Cobalt‐1,4‐benzenedipyrazolate, abbreviated as Co(bdp) [48]. 3‐D schematics of this

compound in both collapsed and expanded forms are shown in figure 2.7. The pressure‐induced LJM

model regressed on experimental isotherms and the Clausius‐Clapyron plots are presented in figure

2.8. The model is able to quantitatively describe all adsorption and desorption isotherms at all five

reported temperatures between 273 K and 323 K, as shown in figure 2.8(a) and (b) with a regression

R2 of 0.998 for adsorption and 0.994 for desorption. As with the previous regression, the ‐ΔH value of

14 kJ/mol predicted by the model was in close agreement with the 13.5 kJ/mol (expanded) value

reported in the experimental work [48]. The values of the other classical and LJM parameters are also

meaningful and listed in Appendix A3. For CH4 adsorption on Co(bdp), the temperature below which

hysteresis remains at its low asymptotic range is predicted to be ~ 175K.

Figure 2.7. 3D depiction of the solid‐state transformation of Co(bdp) from collapsed phase at 0 bar (Low ) to expanded phase at 30 bar (High ) [48].


Figure 2.8. The pressure‐induced LJM model regressed on isothermal (a) adsorption and (b) desorption of CH4 on Co(bdp). Experimental data has been collected from reference [48]. The

model’s ability to quantify breathing isotherms at all five reported temperatures is demonstrated for both (a) adsorption and (b) desorption where solid and hollow symbols denote experimental data and the solid curves represent the pressure‐induced LJM model. (c) Plots for ln (P0) vs. 1/T for

adsorption (breathing‐in) and desorption (breathing‐out) showing that both sets of points fall on a straight line and justifying the use of the Clausius‐Clapyron equation and predicting the existence of

a point where P0,ads = P0,des.

2.4 Experimental work

As discussed above, the pressure‐induced LJM model predicts that at a sufficiently low temperature,

hysteresis in an adsorption/desorption isotherm should become very small and approach a low

asymptotic value for materials that exhibit breathing behavior. In order to validate this prediction,

y = -969.03x + 10.72

y = -579.14x + 8.4942

5

6

7

8

0.003 0.0035 0.004 0.0045 0.005 0.0055 0.006

ln (

P0)

1/T (1/K)

(c)

Point of intersection


adsorption/desorption experiments were carried out on a breathing adsorbent. The material chosen

was a synthesized [59] zeolitic imidazole framework composed of Zn metal clusters connected with

1H‐benzimidazole linkers, abbreviated as ZIF‐7 [66, 71, 72].

Figure 2.9 (a) and (b) show the pressure‐induced LJM model regressed on equilibrium capacity

measurements, obtained using a Micromeritics ASAP2020 instrument, for the adsorption and

desorption of pure (>99.97%) CO2 on ZIF‐7 under isothermal conditions for five temperatures between

233K and 253K and pressures up to 1 bar. Contrary to the previous cases, using the adsorption

enthalpy for desorption in this case did not produce a good fit. Hence the enthalpy of adsorption was

used as separate fitting parameter for adsorption and desorption. This value was ‐ΔH = 28 kJ/mol for

adsorption and ‐ΔH = 28.9 kJ/mol for desorption. These values are in close proximity to the 27.6 kJ/mol

value reported by Arami‐Niya et al. [59] obtained by fitting the Langmuir‐Freundlich model to

measurements from high pressure experiments up to 40 bar. The fits indicate that the model is able

to quantitatively describe both adsorption and desorption isotherms at these temperatures. The

regression R2 was 0.994 for adsorption and 0.97 for desorption. The values of the other classical and

LJM parameters are also meaningful and listed in Appendix A3.

The Clausius‐Clapyron plot for ln(P0) vs 1/T is shown in figure 2.9 (c). This plot shows that the slope of

the breathing‐out branch is just slightly lower than the breathing‐in branch meaning that considerable

extrapolation of the lines are required to reach the point of intersection, which corresponds to a

temperature of 109K. Thus, although the original intent was to conduct measurements at the point of

intersection, this plot shows that the temperature at that point is outside the limits of our

experimental apparatus.

In order to study hysteresis, four sets of experimental adsorption and desorption isotherms were

compared with those obtained from the pressure‐induced LJM model for temperatures between 233K

and 247K. This comparison is shown in figure 2.10 (a) – (d). The area between the adsorption and

desorption isotherms can be used as a measure of hysteresis at each temperature. Both the shape of


this area as well as the onset of hysteresis appear to be in agreement with the model. This area

becomes progressively smaller as the temperature is reduced, as predicted by the LJM model.

Although the lowest temperature (233K) allowed by our apparatus for performing the experiment was

well above the predicted Ti of 109K for this material, we still observed significant reduction in

hysteresis. Figure 2.10 (e) shows magnitudes of this enclosed area calculated graphically for each pairs

of isotherms and plotted against temperature. An exponential trendline can be fitted through these

points which also confirms that hysteresis decreases with temperature and reaches an asymptotic

value. We consider this observed trend of diminishing hysteresis with decreasing temperature to be

in agreement with the pressure‐induced LJM model’s prediction.


Figure 2.9. The pressure‐induced LJM model regressed on low pressure isothermal (a) adsorption and (b) desorption of CO2 on ZIF‐7. Measurements were made using a Micromeritics ASAP2020

instrument. (c) Clausius‐Clapyron plots for ln (P0) vs. 1/T for isothermal adsorption (breathing‐in) and desorption (breathing‐out) of CO2 on ZIF‐7 between 233K and 247K showing that both sets of points

fall on a straight line and predicting the existence of a point where P0,ads = P0,des.



Figure 2.10. (a) – (d) Results of adsorption/desorption experiments of CO2 on ZIF‐7 illustrating the diminishing effect of hysteresis with decreasing temperature. The area between the adsorption and desorption isotherms can be used as a measure of hysteresis at each temperature. (e) Area enclosed by each pair of adsorption and desorption isotherms calculated graphically for temperatures up to

283K.

2.4.1 Hysteresis in non‐breathing adsorbents

Morishige and Tateishi [73] conducted single component isothermal adsorption/desorption of Ar, O2

and CO2 on SBA‐16 ordered mesoporous material, which is a non‐breathing solid adsorbent with

cagelike ink‐bottle pores, over a wide range of temperatures and pressures. In remarkable contrast to

our observations with flexible adosrbents, they reported hysteresis in these materials to become


larger at lower temperatures and diminish and eventually disappear completely at higher

temperatures for all three gases [73]. They suggested that adsorption and desorption in an ink‐bottle

pore is similar to the process of disappearance and formation of a gas bubble in a liquid droplet

confined to the pore. For the formation and disappearance of these vapor bubbles, an energy barrier

needs to be overcome, which gives rise to hysteresis. At higher temperatures, it is easier to overcome

this energy barrier and hence hysteresis diminishes.

Neimark et al [74] presented results from a study involving capillary condensation hysteresis resulting

from low temperature sorption of Ar and N2 in nanoporous cylindrical channels of MCM‐41 type

mesoporous siliceous molecular sieves. For Ar, isothermal adsorption/desorption data obtained from

experiments at 77.9K, 93.9K and 99.9K were compared with results of nonlocal density function theory

calculations. In this case too, the hysteresis loop decreased in size with temperature, as evident from

both the experimental data and DFT calculations.

The examples above, when combined with our observations for Fe(bdp), Co(bdp), and ZIF‐7, show that

the effect of temperature on hysteresis depends very much on the mechanism of the

adsorption/desorption process taking place. Temperature appears to have an opposite correlation

with hysteresis in flexible adsorbents compared to those reported for rigid adsorbents. The reason for

this, we think, is that in flexible adsorbents, the main cause of hysteresis is structural change in the

solid itself, whereas in rigid materials adsorption hysteresis is caused mainly by the interaction of the

fluid molecules with the solid surface.

The only data we were able to find in the literature where hysteresis had a temperature correlation

similar to the one we observed for Fe(bdp), Co(bdp), and ZIF‐7 were those reported in 1989 by Burgess

et al [75], who provided a review of adsorption hysteresis of CO2, N2O and Xe on a number of porous

materials. In most of these materials, the size of the hysteresis loop was reported to decrease with

temperature. However, for adsorption of N2O on Vycor glass, hysteresis increased with temperature.

Likewise, for adsorption of CO2 on porous glass at temperatures below 217K, the hysteresis loop was


also reported to expand with temperature [75]. For this material, hysteresis peaked at ~ 217K,

decreased as the temperature was further increased and completely disappeared at temperatures

above 259K. Low pressure hysteresis in these materials was attributed to swelling of the adsorbent

which, in many cases, required long periods of degassing at high temperature, in order to return to

their original form.

2.5 Further modelling

2.5.1 CO2 on mmen‐Fe2(dobpdc) and mmen‐Co2(dobpdc)

After modeling the adsorption of CH4 on two flexible MOFs successfully, an attempt was made at

modeling breathing behavior with a different gas. For this purpose, experimental data reported by

McDonald et al [64]. that covered CO2 adsorption on two MOFs, namely Iron(II)‐4,4'‐dioxidobiphenyl‐

3,3'‐dicarboxylate‐N,N'‐dimethylethylenediamine, abbreviated as mmen‐Fe2(dobpdc), and Cobalt‐

4,4'‐dioxidobiphenyl‐3,3'‐dicarboxylate‐N,N'‐dimethylethylenediamine, abbreviated as mmen‐

Co2(dobpdc), was used. Isothermal adsorption capacity measurements were reported at three

temperatures from 298 K to 323 K for pressures up to 110 kPa, as well as calculated values for the

enthalpy of adsorption.

For CO2 adsorption on these materials, no corresponding desorption data was reported. Therefore, it

was not possible to determine values for P0,des and Ti. The isotherm fits are presented in figure 2.11,

which confirms that the model is able to successfully quantify CO2 adsorption on both of these MOFs.

The regression R2 was 0.984 and 0.973, respectively, for mmen‐Fe2(dobpdc) and mmen‐Co2(dobpdc).

The enthalpy of adsorption predicted by the model was 51.8 kJ/mol for CO2 adsorption on mmen‐

Fe2(dobpdc) which was in rough agreement with the reported value of 57 kJ/mol obtained from

density function theory (DFT) calculations [64]. For CO2 adsorption on mmen‐Co2(dobpdc), the ‐ΔH

value predicted by the model was 42 kJ/mol which was also in approximate agreement with the

reported value of 47 kJ/mol obtained from DFT calculations [64]. The values of all other parameters

are also meaningful and are listed in Appendix A3.


Figure 2.11. Demonstration of the pressure‐induced LJM model’s ability to quantify breathing

isotherms for the adsorption of CO2 on (a) mmen‐Fe2(dobpdc) and (b) mmen‐CO2(dobpdc) at three temperatures reported by McDonald et al [64]. Solid symbols denote experimental data and solid

curves represent the pressure‐induced LJM model.

2.5.2 CO2 on amino‐MIL‐53(Al)

The next dataset used for regressing the pressure‐induced LJM model is the adsorption and desorption

of CO2 on amino‐MIL‐53(Al) reported by Couck et al. [60] The solid‐state 3D structures of this material

in np and lp phases are shown in figure 2.12 (a) and (b) respectively. Amino‐MIL‐53(Al) is a well‐known

MOF and has been studied in great detail [58, 60, 76‐79]. This example was chosen because of high

adsorption capacity of amino‐MIL‐53 at low pressures (in the first stage of adsorption) that makes the

shape of the isotherm different from the previous examples. There appears to be a minor low‐pressure

phase transition taking place around 0.3 bar and the second breathing transition, which seems to be

the main np lp phase transition, is apparent at higher pressure. Consequently, a sharp increase in

the amount of gas adsorbed is featured at low pressures followed by a near‐flat region (saturation of

the preliminary cavities) and then a subsequent second sharp increase in the uptake. Once the

pressure becomes sufficiently high and reaches a threshold, the familiar np lp transition takes place

following similar principles described for stepped breathing.


Figure 2.12. 3D structure of amino‐MIL‐53(Al) in (a) np and (b) lp phases. Color legend – yellow: Al, red: O, gray: C, and blue: N [80].

Experimental data for CO2 adsorption and desorption on amine‐functionalised amino‐MIL‐53(Al) was

used to investigate the pressure‐induced LJM model’s ability to quantify this type of isotherm shape.

Since desorption data is available, it is possible to determine the point of intersection of the breathing‐

in and breathing‐out pressures with the same approach used for CH4 on Fe(bdp) and Co(bdp). Results

of the model’s regression to experimental data are shown by the fits in figure 2.13. As before, the

desorption isotherms were regressed using only three parameters, βP, k and Po,ref while all other

parameters were fixed at values obtained from the adsorption fit. The regression R2 was 0.995 and

0.992 for the adsorption and desorption fits respectively. Since only two isotherms are available for

each regression, the uncertainties in the parameter values are expected to be higher than those

obtained from the previous fits. However, as a validation, the enthalpy of adsorption value of 36.7

kJ/mol predicted by the model is still in agreement with the reported value of 38.4 kJ/mol [60]. Other

parameter values are also meaningful and listed in Appendix A3. The predicted value for the

temperature at the point of intersection where P0,ads = P0,des was Ti ~ 6K. Since the

adsorption/desorption data is available only at two temperatures, only two points are available for

plotting the breathing‐in and breathing‐out straight lines in the breathing pressure vs. temperature

graph, resulting in higher uncertainties in the predicted values.

(a) np phase (b) lp phase


Figure 2.13. Demonstration of the pressure‐induced LJM model’s ability to quantify double‐stepped breathing isotherms for CO2 (a) adsorption and (b) desorption on amino‐MIL‐53(Al) at two

temperatures reported in reference [60]. Solid symbols denote experimental data, and solid curves represent the pressure‐induced LJM model.

2.6 Process simulation

Simulation has become an integral part of the design exercise for major industrial processes involving

adsorption. However, to simulate any adsorption process, an isotherm model that is capable of

successfully quantifying the adsorption isotherms for the adsorbent being used is required. Since no

such isotherm model is available to date for breathing MOFs, no reliable simulation can be performed

for the entire range of pressures and temperatures at which breathing behaviour is exhibited.

Of course, a classical form of the isotherm model such as the Langmuir or the Toth model can still be

used to generate process simulation results. However, the validity of those results would be

questionable at temperatures and pressures at which the adsorption isotherms for the breathing MOF

deviates from the classical model. A similar approach is seen in the work of Gomez et al. [79] where

the Dubinin‐Astakhov (D‐A) model is used to introduce adjustments to the exponents of the classical

adsorption model and subsequently generate a series of simulation results for binary breakthrough

separation of a CO2/CH4 mixture using amino‐MIL‐53 (Al) as the adsorbent. However, the D‐A model’s

isotherm fits with experimental data are rather poor, even for the single temperature presented.

Importantly, the model fails to capture the distinct S‐Shaped feature of the breathing isotherm. Thus,


the validity of all simulation results presented using the D‐A model as the isotherm model becomes

questionable, especially near breathing conditions where the model deviates the most from

experimental data.

The pressure‐induced LJM model developed and presented in this work can be used as an isotherm

model for performing process simulations. As this model can successfully predict both the single‐S

shaped and the double‐stepped isotherms as demonstrated above in the fits with experimental data,

it is expected that, for breathing MOFs, the results of the simulation will be significantly more reliable

than any simulation that uses a classical adsorption model as the isotherm model. To demonstrate

this, the model was first incorporated into the commercial simulation package ASPEN Adsorption by

encoding the governing equations into the software so that the model was now available as a user‐

defined isotherm model that could be selected to characterise any adsorption bed/column in the

simulation flowsheet. Two sets of simulations were then performed. The first simulation, as presented

and discussed in Appendix A5, was the dynamic breakthrough of an equimolar CH4/CO2 mixture using

a classical isotherm model and how the results differed when the pressure‐induced LJM model was

used. An equimolar composition was chosen to align with breakthrough experiments for the same

mixture conducted by Couck et al. [60] so that the simulation results could be compared with

experimental results reported in the literature. The second simulation was a pressure swing

adsorption simulation as discussed below.

2.6.1 Pressure swing adsorption simulation of an equimolar CH4/CO2 mixture

Pressure swing adsorption (PSA) is a technology widely used for separating gas mixtures. In recent

years, PSA applications have been a major motivating factor in the developments of hybrid adsorbents

including flexible materials such as breathing MOFs. Many of these materials have been designed with

the objective of providing superior separation performance when compared to traditional adsorbents

such as zeolite or carbon molecular sieve. One such conventional adsorbent is Zeolite 13X.


Figure 2.14 (a) and (b) show pure component adsorption isotherms for CO2 and CH4 respectively on

amino‐MIL‐53(Al) at 303 K reported by Couck et al. [60] along with the regressed pressure‐induced

LJM model. If this material is used as an adsorbent in a PSA process at 303 K, the difference between

the CO2 and CH4 working capacity is highest when the cycle swing pressure is chosen to be around 6.4

and 20 bar, which corresponds approximately to the pressure across the transition region, as shown

in figure 2.14 (a) and figure 2.14 (b).

Pure component CO2 and CH4 adsorption isotherms for Zeolite 13X at the same temperature, obtained

by regressing the Toth model to published experimental data at three temperatures, are shown in

figure 2.14 (c) and (d). The regressions and the resulting parameters of the Toth model are available

in Appendix A6. The CO2 working capacity of Zeolite 13X is considerably smaller than that of amino‐

MIL‐53(Al), as shown in figure 2.14 (c). The CH4 working capacity is also slightly lower, as shown in

figure 2.14 (d). Thus, a PSA process using a cycle swing pressure of 6.4 ‐ 20 bar at 303K with amino‐

MIL‐53(Al) as the adsorbent is expected to produce superior separation of a CH4/CO2 mixture

compared to a PSA process using Zeolite 13X as the adsorbent.

In order to demonstrate that using the pressure across the breathing region of a flexible MOF as the

cycle swing pressure of a PSA process can produce superior separation, a two‐bed PSA process was

simulated, using the commercial simulation package ASPEN Adsorption, to separate an equimolar

mixture of CH4 and CO2 using amino‐MIL‐53(Al) as the adsorbent. The pressure‐induced LJM model

was used as a user‐defined isotherm model for both beds. The same simulation was then repeated

using Zeolite 13X as the adsorbent and the classical Toth model as the isotherm model to define the

adsorption beds. The ASPEN Adsorption flowsheet used for the simulations is shown in Appendix A7.

For amino‐MIL‐53(Al), the parameter values obtained from the regression of the pressure‐induced

LJM model on published adsorption data as discussed above and reported in Appendix A3 were used

as isotherm parameters. For Zeolite 13X, isotherm parameters obtained from the regression of the

Toth model, as shown in Appendix A6, on published CO2 and CH4 adsorption data [81] was used.


The column dimensions and the values of the adsorbent properties used are listed in Appendix A7.

These values have been collected from the literature for both amino‐MIL‐53(Al) and Zeolite 13X [79,

82].

One major limitation of the simulation package that makes it particularly unsuitable for modelling

processes involving flexible adsorbents is that it only allows fixed values for the inter‐ and intra‐particle

voidage. For breathing MOFs, of course, both these values are expected to change during the

adsorption/desorption processes as the MOF structure undergoes transitions from collapsed to

expanded states and vice versa. Another limitation is that for each bed, only one set of isotherm

parameters can be used for each gas meaning that the same parameters are used to model both

adsorption and desorption (bed regeneration) , i.e, hysteresis is not accounted for. In spite of these

limitations, the simulation was carried out.

The parameters for the PSA cycle are given in Appendix A7. A feed composition of 50% CH4 + 50% CO2

was used. Initial compositions of the product and the beds were set as equal to the feed.

Figure 2.15 shows the results of the PSA simulation as a plot of the composition of the product stream

against time. Amino‐MIL‐53(Al), being CO2 selective in the chosen swing pressure range and having

higher CO2 working capacity compared to Zeolite 13X, rapidly brings the CO2 mole fraction in the

product stream to less than 0.1 resulting in far superior separation performance compared to Zeolite

13X, which has a much smaller CO2 working capacity in the given pressure range.


Figure 2.14. Illustration of CO2 and CH4[ [(a) and (b) respectively] isothermal working capacity of amino‐MIL‐53 (Al) for a PSA cycle swing pressure of 6.4 – 20 bar that corresponds to the adsorbent’s CO2 breathing region. Zeolite 13X has a much smaller CO2 working capacity[(c)] in the same pressure range and also a lower CH4 working capacity[(d)]. Hence, under identical conditions, amino‐MIL‐53 (Al) is expected to produce superior CO2/CH4 separation performance compared to Zeolite 13X when

used as an adsorbent in a PSA process.


Figure 2.15. Comparison between amino‐MIL‐53(Al) and Zeolite 13X as adsorbents used in a two‐bed PSA simulation with an equimolar CH4/CO2 feed. Variation in product stream composition with time show that a far superior separation performance can be achieved using amino‐MIL‐53(Al) as the adsorbent. The pressure‐induced LJM model was used as the isotherm model to characterise the adsorbent bed for amino‐MIL‐53(Al) while the classical Toth model was used for Zeolite 13X.


2.7 Concluding remarks

An empirical model has been proposed to mathematically express adsorption and desorption

isotherms of gate‐opening flexible adosrbents. The model was tested successfully by regressing

literature and experimental data for breathing isotherms for the adsorption and desorption of CH4

and CO2 by a number of flexible adsorbents over a wide range of temperature. The model uses

considerably fewer fitting parameters compared to other models that have attempted to quantify

breathing isotherms and is particularly useful for process design. Uncertainties in fitting parameters

were significantly lower compared to the values of the parameters themselves, except for the cases

where experimental data was available only at a few temperatures. The enthalpies of adsorption

predicted by the model were in agreement with the reported values for all the gas‐adsorbent pairs

used in the regressions.

The model is consistent with existing explanations of the breathing mechanism and includes

quantification of hysteresis. For breathing isotherms with hysteresis, the model predicts that

hysteresis diminishes with decreasing temperature and reaches a very small asymptotic value when

the temperature corresponds to the point of intersection of the breathing‐in transition pressure line

and the breathing‐out transition pressure line where P0,ads = P0,des. As the temperature is further

reduced, no significant change in the magnitude of hysteresis is expected. The temperature, Ti, at the

point of intersection is thought of as being the temperature at which the fraction of np unit cells in

the framework reaches a minimum. This fraction remains unchanged as the temperature is further

reduced suggesting that, at low pressures, a small portion of the framework remains in np phase

regardless of how low the temperature is. In order to expand these cells during adsorption, an energy

barrier must be overcome to create the extra surface for np lp transition, resulting in hysteresis.

Hence, hysteresis in these materials does not disappear completely but is more likely to remain steady

at an asymptotic value at temperatures below Ti.


The model is particularly useful for simulating industrial adsorption processes involving breathing

adsorbents. This was previously not possible as existing models were incapable of quantifying

isothermal adsorption/desorption capacities for materials that exhibit breathing behaviour.

There are observed variations in the breathing behaviour, for example, those including negative

adsorption [83], which the current form of the model is unable to quantify. Further understanding of

the unusual structural transitions that cause these exceptions is required before conducting a

quantitative analysis. However, the model’s ability to address the most common types of breathing

isotherms has been well‐demonstrated in this work, and upon further development, it is expected

that a wider array of breathing phenomena can be accounted for.

In the next chapter, we shift our focus away from flexible MOF’s towards more conventional rigid

adsorbents namely zeolites and activated carbon. We detail a series of experimental attempts to alter

structures of these materials with the intention of improving their adsorptive separation performance.

We demonstrate how selectivity and capacity of these adsorbents are affected by certain chemical

treatment procedures and examine the results in light of specific gas separation applications.


Chapter 3

Effect of Structural Manipulations on Selectivity and

Capacity of Zeolite Y

3.1 Introduction

Pressure swing adsorption (PSA) is a cost‐effective and energy‐efficient technology that can

potentially replace conventional cryogenic distillation as a method to separate N2 from natural gas

during LNG production. There is a strong interest in the natural gas industry to adopt and implement

this new technology due to the high power consumption associated with maintaining the conditions

for distillation‐based separation processes [12, 16, 19]. However, for cost‐effective CH4/N2 separation,

PSA relies on the availability of adsorbents with high CH4 capacity and high CH4/N2 selectivity [27, 28,

30]. This desirable combination is unfortunately not found in readily available adsorbents. While some

recent developments involving metal organic frameworks (MOFs) have reported high CH4/N2

selectivity [84‐86], the CH4 capacity of these materials is significantly lower than those of existing

commercially available materials such as activated carbon [28].

One of the goals of this research project was therefore to develop novel materials that combine both

desirable properties of capacity and selectivity. The approach involved identifying structural

characteristics that play key roles in the adsorption mechanisms of existing materials and developing

synthesis methods that implement structural modifications in order to achieve enhancement in

CH4/N2 separation. To determine the effect of the performed modifications, a series of structural

characterisations and adsorption analyses were conducted on parent materials, final products and on

intermediate entities obtained at different stages of synthesis.


3.2 Zeolite structures

For this work, the base materials of primary interest were zeolites. These are microporous adsorbents

consisting of a crystalline framework of SiO4 and AlO4 arranged in a regular assemblage forming an

open‐lattice structure requiring cations, usually Na+ or K+, to balance the negative charges induced on

the framework by the aluminum atoms [25]. Regularly arranged tetrahedral building units form an

open crystal lattice within the framework containing uniform pores with no distribution in pore size –

a unique feature that distinguishes zeolites from other adsorbents such as carbon. These uniform

pores allow gas molecules of dimensions smaller than the pore size to penetrate the lattice during

adsorption while preventing larger molecules from entering the pores. During desorption, the

previously adsorbed molecules are usually readily released from the pores allowing the adsorbent

material to be regenerated for reuse [25, 87, 88].

Due to the presence of aluminum, the zeolite framework is inherently negatively charged as each

aluminum atom induces one negative charge on the framework. In order to balance this negative

charge, cations are required to make zeolite structures stable. The cations commonly present in

zeolites are Na+ and K+. Along with pore size and other structural features, these cations also play an

important role in determining adsorptive properties of zeolites. Figure 3.1 shows crystal framework

structure of two common zeolites, namely zeolite X/Y (Faujasite) and zeolite A (Linde Type A).


(a) (b)

Figure 3.1. Framework of (a) zeolite X/Y and (b) zeolite A [25].

3.2.1 Role of the cation on the adsorption mechanism

Physical adsorption involves both van der Waals forces (dispersion‐repulsion) and electrostatic forces

arising from polarization, dipole and quadrupole interactions. While dispersion‐repulsion forces are

always present, for adsorbents such as zeolites which have an ionic structure, the electrostatic

contributions are also significant [25]. The total adsorption potential, ΦTotal, is a measure of the

strength of all interactions taking place and is given by [89]:

ΦTotal = φadsorbate – adsorbate + φadsorbate – adsorbent (Eq. 3.1)

where φadsorbate – adsorbate is the potential for interactions between the gas molecules and φadsorbate

– adsorbent is the potential for interactions between the gas molecules and the adsorbent surface.

Assuming that φadsorbate – adsorbate << φadsorbate – adsorbent, the adsorption potential can be

approximated as:


ΦTotal = φadsorbate – adsorbent (Eq. 3.2)

Expanding φadsorbate – adsorbent to include the contribution of all the individual forces, ΦTotal

becomes:

ΦTotal = φD + φR + φInd + φFμ + φFQ (Eq. 3.3)

where φD and φR represent the interaction potentials due to the dispersion and repulsion forces

respectively, φInd represents the potential due to interaction between the electric field and induced

dipole, φFμ represents the potential due to interaction between the electric field and permanent dipole

and φFQ represents the potential due to interaction between the electric field and quadrupole.

Expanding further, φInd can be expressed as

12

2 (Eq. 3.4)

where F is the electric field and α is the polarizability, defined as the ease of distortion of the electron

cloud of an atom or molecule – a property specific to the type of gas molecule being adsorbed [90].

CH4, for example, has a polarizability almost 1.5 times higher than N2 [91] Thus, CH4 molecules will

generate higher interaction potential, φInd, than N2 molecules when interacting with the same

adsorbent.

The cation present in the zeolite framework is an entity that significantly affects the interaction

potential φInd. A cation directly affects the electric field generated by the zeolite and, being positively

charged, also has the ability to polarize or distort the electron cloud of a target gas molecule. A

measure of this cationic ability is a quantity known as polarizing power which varies significantly with

different types of cations [91]. Thus the cation is an important element of the zeolite structure that

can be used to manipulate adsorption characteristics. The ability of the cation to influence adsorptive

properties was taken advantage of in this work by adopting a process called ion‐exchange where the


cation of a zeolite is replaced by another target cation, thereby altering adsorption characteristics

according to requirements.

3.2.2 Ionic Liquidic Zeolites (ILZs)

For a binary mixture at a fixed temperature and pressure, the equilibrium selectivity, α1/2, of

component 1 over component 2 s given by [28, 92]

//

/ (Eq. 3.5)

where and are equilibrium adsorption capacities and and are equilibrium mole fractions

in the gas phase of components 1 and 2 respectively. For single component adsorption of a pure gas,

equation Eq. 3.5 simplifies to

/ (Eq. 3.6)

The Na+ and K+ ions commonly present in zeolites are small in size meaning that the charge is

concentrated in a small volume leading to high charge density. If these ions can be replaced by larger

monovalent cations, then the charge would be distributed over a larger spatial dimension and the

charge density would be lower. This would increase the cation’s “power” to distort the electron cloud

of a large inbound gas molecule such as CH4 and interact with it more strongly. Thus the interaction

potential φInd, between the adsorbent and the CH4 molecule can increase, resulting in considerably

higher adsorption capacity and selectivity for CH4 compared to another gas molecule such as N2 which

is small in size and thereby provides less room for its electron cloud to be distorted. This mechanism

has been supported by ab initio density functional theory (DFT) calculations used to estimate

interaction energies between gas molecules and adsorbent surfaces [93]. The proposed synthesis

route for replacement of zeolite cations is through an ion exchange process, where an ionic compound

in liquid state is made to react with a zeolite producing a novel adsorbent material, termed Ionic

Liquidic Zeolite (ILZ). Using this method, it was demonstrated in previous work conducted in our


research group that it is possible to design and implement modifications to zeolite structures which

will result in enhanced separation performance for CH4/N2 gas mixtures [93].

3.2.3 Organic cations

Organic cations are generally bulky in size and consequently possess higher polarization power than

small inorganic ions such as Na+ or K+ [93]. One such cation, tetramethylammonium+ (TMA+) found in

ionic form in its commercially available salt tetramethylammonium chloride (TMA‐Cl), is considered

as a potential candidate for cation replacement in zeolites. The structure of this salt is shown in figure

3.2. By performing ion exchange with this salt, the Na+ cations in the Faujasite zeolite NaX or NaY can

be replaced by TMA+ resulting in the Ionic Liquidic Zeolite TMAX or TMAY. Faujasite is chosen as the

parent zeolite because it contains a large central cavity pore volume with sufficient room for the TMA+

cation. Figure 3.3(a) shows the structure of NaX or NaY while figure 3.3(b) shows a schematic of the

TMAX or TMAY structure synthesised through ion exchange between Na+ and TMA+ ions.

3.2.4 Synthesis of TMAY

With respect to separation of CH4/N2 gas mixtures, TMAY has been the most promising ILZ synthesised

so far [93]. Its synthesis procedure involved treating a commercially available NaY zeolite in powder

form (NaY ‐ CBV100 supplied by Zeolyst International) with 0.5 M aqueous TMA‐Cl solution at 40 °C

for 5 hours in a shaking water bath set at 120 oscillations per minute. The resulting solid was filtered

and washed twice with deionized water before repeating the treatment for a second round of ion‐

exchange. After two‐rounds of treatment, the samples were air‐dried in an oven at 80 °C for 12 hours.

The resulting solid powder was then pelletised by first compressing into compact discs in a hydraulic

press and then cutting into small pellets.


Figure 3.2. Structure of Tetramethylammonium Chloride (TMA‐Cl) [94].

(a) (b)

Figure 3.3. (a) Structure of Faujasite zeolite NaX/NaY [89]. (b) Structure of TMA‐X or TMA‐Y synthesised through ion exchange [95].

Initial experiments with TMA+ & NaX and TMA+ & NaY showed that it is indeed possible to synthesise

novel adsorbent materials by ion exchange, as described above [93]. However, only about 30% of the

Na+ cations present in NaY could be replaced with TMA+ while the proportion was much lower with

NaX. This observation is attributed to the fact that the inner cavity of both NaX and NaY are occupied

with Na+ ions making it considerably difficult for the larger TMA+ cations to enter the cavity through

the 12‐member oxygen ring (12MR) aperture of Faujasite. NaY, with a higher Si/Al ratio than NaX, has

fewer Na+ ions and hence more room for TMA+ to enter the cavity and replace the Na+ ions.


The first step in this experimental study was therefore to repeat the synthesis procedure with other

available variants of the Faujasite zeolite with increasingly higher Si/Al ratios. It should be noted,

however, that with increasing Si/Al ratios the negative charges on the framework decreases resulting

in fewer TMA+ cations that are attracted. The goal therefore was to determine the optimum Si/Al ratio

in Faujasites that will allow the highest number of TMA+ cations to be introduced into their inner

cavity.

Adsorption analyses were performed alongside the synthesis experiments in order to determine single

component CH4 and N2 adsorption capacity and selectivity of the parent zeolites as well as each new

adsorbent material synthesised. Comparisons were made in order to evaluate the adsorption

performance of the novel materials relative to each other as well as their parent zeolites.

Microscopic imaging studies were also performed on the synthesised novel materials as well as their

parent zeolites in order to detect structural modifications that have been implemented. Facilities at

the UWA Centre for Microscopy, Characterisation and Analysis (CMCA) were used for imaging‐based

studies.


3.3 Characterisation of existing zeolite‐based adsorbents

A number of Ionic Liquidic Zeolite (ILZ) materials were synthesised in our research group prior to the

start of this project [93, 96]. In this work, a series of characterisation experiments were initially

conducted to determine the optimum composition and experimental conditions of the previously

synthesised adsorbents to use as a basis for further developments.

3.3.1 Adsorption measurements

Single component adsorption analyses were conducted using a commercial adsorption analyser –

Micromeritics ASAP2020. This instrument has an automatic degas functionality where adsorbents are

subjected to predefined temperature ramping under vacuum. A schematic of the instrument is shown

in figure 3.4.

Figure 3.4. Micromeritics ASAP 2020 used for single component adsorption analyses.

The gases used for single component adsorption analysis in the ASAP were supplied by BOC Australia.

As stated by the supplier the functional purities of these gases were CH4 99.995% and N2 99.999%. In

this work, these gases are referred to as pure CH4 and pure N2.

Sample (Degas)

Sample (Analysis)

Glycol Bath

Temperature Control


3.3.2 Effect of Si/Al ratio

Adsorption analyses were conducted on two ILZ variants, termed TMAY‐CBV100 and TMAY‐CBV720,

which were synthesised prior to the start of this doctoral project [97]. Both of these materials were

synthesised using commercial Zeolite Y powders purchased from Zeolyst International, Netherlands.

The relevant information regarding the parent materials, as provided by the manufacturer [98], are

listed in Table 3.1 below.

Table 3.1. Structural and chemical composition of CBV100 and CBV720 parent zeolites.

The H+ cation present in the parent zeolite of TMAY – CBV720 is smaller in size compared to the Na+

cation present in the parent zeolite of TMAY – CBV100. This means the cations occupy less cavity space

in CBV720 than they do in CBV100 leading to the expectation that there should be more room for

bulky TMA+ cations in the CBV720 cavity than in the CBV100 cavity allowing for more TMA+ cations to

be introduced during ion exchange. However, the surface area of CBV720 is smaller than that of

CBV100 implying that the effects of surface area on adsorption are less favourable in CBV720. Both

ILZs were degassed for 5h at 180 °C under vacuum before performing adsorption analyses. Figure 3.5

shows adsorption isotherms of TMAY – CBV100 and TMAY – CBV720 for CH4 and N2 at different

temperatures. The data used in the plots are presented in tabular form in Appendix D1.

Si/Al ratio of parent zeolite

Nominal cation form in parent zeolite

BET surface area of parent zeolite (m2/g)

TMAY – CBV100 2.55 Na+ 900

TMAY – CBV720 15 H+ 780


Figure 3.5. Adsorption isotherms at different temperatures for TMAY – (a) CBV100 with CH4, (b) CBV720 with CH4, (c) CBV100 with N2 and (d) CBV720 with N2.

As shown by the plots above, at all temperatures both CH4 and N2 adsorption capacities are higher for

TMAY‐CBV100 than TMAY‐CBV720. This can be attributed to the fact that CBV100 surface area is

higher than that of CBV720. Also, Na+ is larger in size than H+ meaning that it may have a stronger

ability to distort electron clouds of CH4 molecules, as explained in section 3.2.2. Thus, the interaction

potential between the guest molecules and the Na+ cation is likely to be higher than that between the

guest molecules and the H+ cation. This is also an indication that both ILZs, even after ion exchange

with TMA+, still contain a significant portion of their original cations whose roles are apparent in their

adsorption performance.


3.3.3 Regression to isotherm models

The Langmuir Isotherm model is widely used in adsorption studies to describe the adsorption of gas

molecules onto solid adsorbents. This model is based on the assumption that adsorption takes place

on the surface of the solid adsorbents only as a monolayer of gas molecules on a fixed number of

vacant sites which are of equal size and shape and can each hold a maximum of one molecule [25, 44].

A constant amount of heat is released when each molecule is adsorbed, making adsorption an

exothermic process. Furthermore, dynamic equilibrium exists between the adsorbed gas molecules

and the free gas molecules such that

where A(g) is an unadsorbed gas molecule, B(s) is a vacant site on the adsorbent surface and AB

represents an adsorbed gas molecule on site B.

Based on the above assumptions, the Langmuir model can be mathematically expressed as follows:

(Eq. 3.7)

(Eq. 3.8)

where q is the quantity adsorbed, qm is the monolayer adsorption capacity, b0 is the gas‐solid affinity

coefficient, Q is the enthalpy of adsorption, P is the pressure, T is the Temperature and R is the gas

constant. Figure 3.6 below shows the Langmuir model fitted to the CH4 and N2 adsorption isotherms

reported above for TMAY‐CBV100 and TMAY‐CBV720.


Figure 3.6. The Langmuir isotherm model (line) fitted to adsorption data (circles) at ‐15 °C, 0 °C and 30 °C for TMAY – (a) CBV100 with CH4, (b) CBV720 with CH4, (c) CBV100 with N2 and (d) CBV720 with

N2.

The values of the Langmuir fitting parameters are listed in Table 3.2 below.

Table 3.2. Parameters for the Langmuir isotherm model.

Adsorbent Gas qm (mol/kg) b0 (1/Pa) Q (kJ/mol) Fit – Average R2

TMAY‐CBV100 CH4 3.02 1.17x10‐6 18.5 0.9994

TMAY‐CBV100 N2 0.996 5.02x10‐6 14 0.9975

TMAY‐CBV720 CH4 1.04 1.92x10‐5 12 0.9993

TMAY‐CBV720 N2 0.85 1.27x10‐4 5.4 0.9884


The enthalpy of adsorption, also known as the isosteric heat of adsorption, is defined as the ratio of

the infinitesimal change in the adsorbate enthalpy to the infinitesimal change in the amount adsorbed

[89]. This is an important parameter that gives a measure of energy favourability of adsorption

processes. An adsorption process with a higher enthalpy of adsorption is more energetically

favourable than an adsorption process with a lower enthalpy of adsorption. As this quantity is one of

the parameters (Q) for the Langmuir Isotherm Model, values for enthalpy of adsorption can be

extracted from Langmuir fits to experimental data. These values obtained from the experiments

described above are listed in Table 3.3 along with values reported in the literature.


Table 3.3. Comparison of enthalpy of adsorption values obtained in this work with values reported in

the literature.

Among the adsorption processes in this particular experiment, the highest Q value of 18.5 kJ/mol was

obtained for CH4 adsorption on TMAY‐CBV100. This value is comparable to the reported value of 18.9

for CH4 adsorption on NaY with a Si/Al ratio close to that of the parent zeolite CBV100. This suggests

Adsorbent Si/Al Gas This work Literature (kJ/mol)

(kJ/mol) Talu et

al. [99] Salem et al. [100]

Maurin et al. [101]

TMAY‐CBV100 2.55 CH4 18.5

TMAY‐CBV720 15 CH4 12

NaY 2.2 CH4 18.9

MgY 2.2 CH4 19.5

CaY 2.2 CH4 22.3

Zeolite 13X 1.18 CH4 17.53

Active Carbon AS ‐ CH4 14.35

TMAY‐CBV100 2.55 N2 14

TMAY‐CBV720 15 N2 5.5

NaX 1 N2 19 ± 0.11

KX 1 N2 14 ± 0.08

CaX 1 N2 27 ± 0.16

Zeolite 13X 1.18 N2 13.91

Active Carbon AS ‐ N2 11.95


that replacing 30% [93] of Na+ ions in the parent zeolite does not have significant effect on the

enthalpy of adsorption. However, it is notable that TMAY‐CBV720 has a much lower enthalpy of

adsorption for both CH4 and N2 compared to that of TMAY‐CBV100. This is most likely due to the fact

that CBV100 has a larger BET surface area than CBV720. The lowest enthalpy value of 5.5 kJ/mol was

obtained for N2 adsorption on TMAY‐CBV720 implying that this was the least energy favourable

adsorption process.

As TMAY‐CBV100 produced more desirable results compared to TMAY‐CBV720, the CBV100 variant of

Zeolite Y was chosen as the parent material for further synthesis of ILZs. Henceforth, unless otherwise

mentioned, TMAY‐CBV100 will be referred to simply as “TMAY” and the parent zeolite NaY‐CBV100

will be referred to as “NaY’.

3.3.4 TMAY vs parent zeolite

Figure 3.7 shows CH4 and N2 adsorption isotherms for both TMAY and the parent Zeolite NaY at two

different temperatures. At both temperatures, TMAY has higher CH4 adsorption capacity and lower

N2 capacity compared to NaY. Both enhancement in CH4 capacity and suppression in N2 capacity are

attributed to the presence of TMA+ cation.

TMAY is used as the base reference material in this project. All subsequent results are compared with

TMAY data in order to determine whether or not enhancement in adsorptive separation has been

achieved over the performance of TMAY.


Figure 3.7. Adsorption isotherms of TMAY and parent zeolite NaY at (a) 0 °C with CH4, (b) 0 °C with N2, (c) 30 °C with CH4 and (d) 30 °C with N2.

3.3.5 Effect of degassing

All results reported in this section are from adsorption analyses of materials which have been

subjected to rigorous degassing procedures at elevated temperatures under vacuum. In order to

determine the effect of degassing on the adsorptive performance of ILZs, single component

adsorption analysis using pure CH4 was conducted at two temperatures, 0 °C0 °C and 30 °C0 °C, on a

TMAY sample that has not been degassed and the isotherms were compared to those from a TMAY


sample that has been degassed for 5h at 180 °C under vacuum. The results of this comparison are

shown in figure 3.8.

Figure 3.8. Adsorption isotherms of degassed and not degassed TMAY at (a) 0 °C with CH4, (b) 30 °C with CH4.

As illustrated by the plots above, if not degassed properly, TMAY completely loses its ability to adsorb

CH4. A likely explanation of this phenomenon is that the zeolite adsorbs moisture from the atmosphere

which causes the pores to be completely occupied resulting in the prevention of adsorption. These

results align with reported effects of moisture on the adsorption behaviour of zeolites [102]. As such,

before performing adsorption analyses, all TMA containing compounds were degassed at 180 °C for

5h under vacuum. NaY was degassed at 300 °C for the same duration.

3.4 Novel zeolite‐based adsorbents

3.4.1 Synthesis

Synthesis of novel materials were carried out using standard wet chemistry procedures that involved

homogeneous mixing under controlled temperatures. For procedures that require temperatures

below 90 °C, a shaking water bath with controllable temperature and oscillations per minute, as shown

in figure 3.9(a) was used. For temperatures above 90 °C, a Teflon coated autoclave, with a magnetic


stirrer placed inside, was used on an electric hot plate with controllable temperature and rpm. This

arrangement is shown in figure 3.9(b). Synthesis products were filtered using filter papers and air dried

in an oven at temperatures ranging between 50 °C and 80 °C.

(a) (b)

Figure 3.9. (a) Shaking water bath used for synthesis temperatures below 90 °C (b) Teflon autoclave

and electric hotplate used for synthesis temperatures above 90 °C.

3.4.2 Effect of intermediate cations

As the Mg2+ cation is divalent, fewer Mg2+ cations are needed to balance the negative charges on the

Zeolite framework compared to Na+ ions. Furthermore, as it has been reported that using smaller sized

organic cations result in higher ion exchange rate [93], our expectation was that Mg2+, being smaller

in size than Na+, should also favour increased TMA+ introduction. Compared to that of NaY, there

should be more room in the cavity of an MgY zeolite for introduction of a bulky cation such as TMA+.

Based on this hypothesis, attempt was made to synthesize a new variant of ILZ with Mg2+ as the

intermediate cation. The synthesis involved a three‐step ion exchange approach with Mg2+

introduction taking place in the first step and TMA+ introduction taking place in the second and third

steps. NaY powder was first treated with 0.5 M aqueous MgCl2 for 8h at 50 °C in a shaking water bath.

The resulting suspension was filtered, washed with DI water and dried before being treated twice with

0.5 M aqueous TMA‐Cl for another 6h at 40 °C, filtering, washing and drying the product after each

Autoclave Teflon container Hot plate


treatment. The resulting material is referred to as TMA‐Mg‐Y. CH4 and N2 adsorption isotherms for

TMA‐Mg‐Y at two temperatures, 0 °C and 30 °C, are shown below in figure 3.10.

Figure 3.10. Adsorption isotherms of TMAY, TMA‐Mg‐Y and NaY at (a) 0 °C with CH4, (b) 0 °C with N2, (c) 30 °C with CH4, and (d) 30 °C with N2.

At 0 °C, CH4 adsorption capacity is clearly enhanced for TMA‐Mg‐Y compared to the parent zeolite

NaY, although the enhancement is not as much as that produced by TMAY. At 30 °C, no noticeable

enhancement is seen for TMA‐Mg‐Y. N2 adsorption is suppressed in TMA‐Mg‐Y at both temperatures.

At 0 °C, N2 suppression is close to that of TMAY while at 30 °C, it is less. As a consequence, both CH4

capacity and CH4/N2 selectivity are lower in TMA‐Mg‐Y than in TMAY. These results indicate that


amount of TMA+ ions that have been introduced is less in TMA‐Mg‐Y than in TMAY. A possible

explanation could be that Mg2+ ions are preferred as cations in the zeolite cavity and are therefore

difficult to replace, even after repeated ion exchange treatments.

3.4.3 Effect of alkalinity

Alkaline medium has been reported to cause the formation of mesopores on the zeolite surface [103].

Presence of mesopores should also facilitate the introduction of bulky cations such as TMA+ during

ion‐exchange. In order to test the effect this has on adsorption, two new ILZs were synthesised. The

first material, referred to as TMA‐Base‐NaY, was synthesised using a two‐step process in which the

parent zeolite NaY was first treated with 0.5 M aqueous NaOH for 24h at 80 °C followed by filtration,

drying and a second ion‐exchange treatment with 0.5 M aqueous TMA‐Cl for 6h at 40 °C, also followed

by filtration and drying.

CH4 and N2 adsorption isotherms of this material at 0 °C are shown in figure 3.11, along with TMAY for

comparison. Both CH4 and N2 adsorption isotherms for TMA‐Base‐NaY are almost identical to the

isotherms for TMAY, indicating that there is no apparent effect of this particular alkaline environment

on the adsorption performance of this ILZ.

Figure 3.11. Adsorption isotherms of TMAY, TMA‐Base‐Y and NaY at (a) 0 °C with CH4 and (b) 0 °C with N2.


The second material, referred to as TMAOH‐NaY, was synthesised using a one‐step process in which

the parent zeolite NaY was treated with 0.5 M aqueous TMAOH for 6h at 40 °C followed by filtration

and drying. The TMAOH solution used in the treatment was prepared by diluting commercially

available TMAOH pentahydrate crystals [(CH3)4N(OH) ∙5H2O] sourced from Sigma‐Aldrich Australia.

CH4 and N2 adsorption isotherms for this material at two temperatures, 0 °C and 30 °C, are shown in

figure 3.12 below. At 0 °C, both CH4 and N2 adsorption isotherms of TMAOH‐NaY almost overlap with

that of TMAY.. Thus, no significant change in capacity or selectivity can be observed with TMAOH‐NaY.

At 30 °C, CH4 capacity of TMAOH‐NaY is lower than that of TMAY and N2 capacity of TMAOH‐NaY is

higher than that of TMAY, resulting in a loss of CH4/N2 selectivity. Thus, TMAOH‐NaY does not perform

as well as TMAY for separation of CH4/N2 at 30 °C.


Figure 3.12. Adsorption isotherms of TMAY, TMAOH‐NaY and NaY at (a) 0 °C with CH4, (b) 0 °C with N2, (c) 30 °C with CH4, and (d) 30 °C with N2.

3.4.4 Effect of H+ cation

As discussed in section 3.2.2, TMAY‐CBV720, an ILZ synthesised from a parent zeolite containing H+ as

the cation, did not perform as well as TMAY. However, this material had a Si/Al ratio 7 times higher

than that of TMAY. In order to determine whether lower Si/Al ratio zeolites containing H+ cations

would have any enhancement in performance, two new zeolites, HY400 and HY600 (manufacturer’s

brand names: HY – CBV400 and HY – CBV600 respectively), with lower Si/Al ratios were used as parent

materials for new ILZ synthesis. The relevant properties of these two HY zeolites, as provided by the

manufacturer Zeolyst International, Netherlands, are mentioned in Table 3.4.


Table 3.4. Structural and chemical composition of HY400 and HY600 parent zeolites.

Parent Zeolite Si/Al ratio Nominal cation form in parent zeolite

BET surface area of parent zeolite (m2/g)

HY400 2.55 H+ 730

HY600 2.6 H+ 660

The ion exchange process for synthesizing ILZs using the above materials involves an acid base reaction

approach anticipated as:

TMAOH + HY TMAY + H2O (R2)

HY400 was treated with 0.5 M aqueous TMAOH for 6h at 40 °C and the suspension was filtered,

washed with DI water and dried. The resulting material was termed as TMAOH‐HY400. Using the same

method with the other parent zeolite, a second variant, termed as TMAOH‐HY600, was synthesised.

CH4 and N2 adsorption isotherms at 0 °C are shown in figure 3.13 for both materials along with TMAY

for comparison.

Figure 3.13. Adsorption isotherms of TMAY, TMAOH‐HY600 and TMAOH‐HY400 at (a) 0 °C with CH4, and (b) 0 °C with N2.

At 0 °C, both CH4 and N2 capacities are highest for TMAY, followed by TMAOH‐HY600 and TMAOH‐

HY400. However, the ratio of CH4 capacity to N2 capacity is higher for TMAOH‐HY400 compared to


that of TMAY and TMAOH‐HY600, implying that CH4/N2 selectivity is also higher for TMAOH‐HY400

compared to the other two adsorbents.

Using equation Eq. 3.6, at 0 °C and 100kPa, CH4/N2 selectivity for TMAOH‐HY400, TMAOH‐HY600 and

TMAY respectively are 4.25, 3.92 and 3.91.

Since TMAOH‐HY400 produced considerably higher CH4/N2 selectivity than TMAY at 0 °C, it is of

interest to see how this adsorbent performs at 30 °C. Figure 3.14 shows CH4 and N2 adsorption

isotherms for TMAOH‐HY400 and TMAY at 30 °C.

Figure 3.14. Adsorption isotherms of TMAY and TMAOH‐HY400 at (a) 30 °C with CH4 and (b) 30 °C with N2.

As with 0 °C, both CH4 and N2 capacities are still higher for TMAY, but the difference between the two

adsorbents’ capacities, especially for N2, is lower. Using equation Eq. 3.6 to calculate the CH4/N2

selectivity at 30 °C and 100kPa, gives 4.99 for TMAY and 4.03 for TMAY‐HY400. Thus, although TMAY‐

HY400 supersedes TMAY in terms of selectivity at 0 °C, it fails to do the same at 30 °C.

3.5 Novel carbon‐based adsorbents

The microporous structure of zeolites is highly beneficial for selective adsorption based on the size of

the gas molecules – a process known as molecular sieving. However, the sole presence of a


microporous network also imposes a significant limitation on the diffusion of larger molecules into

and from the active sites confined within the zeolite crystals [31]. As a result, guest molecules are

unable to access all the active sites and a limitation is imposed on the adsorption capacity.

Furthermore, due to restricted access and low diffusion coefficients, the adsorption kinetics are

hindered, even when the micropores are larger than the size of the guest molecules.

In order to solve the diffusion limitation problem associated with conventional microporous zeolites,

great efforts have been made over the last decade on finding ways to improve the pore architecture

of zeolites. Among new materials developed, the most promising are zeolites featuring hierarchical

porosity with at least two levels of pore size – the inherent zeolite microporosity and an additional

level of induced mesoporosity. This new architecture can be created either by generating intra‐

crystalline mesopores inside the microporous zeolite crystals or by inducing inter‐crystalline

mesopores in between the nano‐sized zeolite crystals. In hierarchical zeolites, the essential feature of

selective molecular sieving is provided by the intracrystalline micropores and efficient mass transfer

is facilitated by the increased diffusivity and reduced diffusion path length due to the presence of the

mesopore structures [31].

The method reported to be most successful in synthesising hierarchical zeolites involves using carbon

based mesoporous materials such as carbon nanoparticles, nanotubes, nanofibers, aerogels or

ordered mesoporous carbons as a template matrix for growing zeolite crystals. Several examples of

carbon‐templated zeolites synthesised using different types of carbon materials as templates are

depicted in figure 3.15.


Figure 3.15. Diagram showing mesoporous zeolite templated by (a) carbon nanoparticles, (b) carbon nanotubes, (c) carbon aerogel and (d) 3D ordered mesoporous 3DOm carbon [31].

A second route for synthesizing hierarchical zeolites is through demetallization involving either

dealumination or desilication where mesopore structures are introduced into pre‐synthesised zeolite

crystals through selective removal of aluminium or silicon atoms from the zeolite framework. Figure

3.16 illustrates the mechanism in one such method that involves desilication. Depending on the Si/Al

ratio in the zeolite, an alkaline medium can extract Si atoms from the zeolite framework causing the

formation of mesopores. This effect is most prominent in zeolites with high Si/Al ratio because the Al

atoms prevent the Si atoms they are bound to through oxygen from being extracted. However, as

most zeolites contain more Si than Al (Si/Al > 1), an alkaline environment can cause, although to a

lesser extent, desilication to take place even in low Si/Al zeolites, resulting in the formation of

mesopores in the zeolite framework [103].

This work aimed to develop a zeolite architecture consisting of both micropore and mesopore

structures in order to significantly enhance adsorbent performance for gas separation applications. A

synthesis route similar to the demetallization procedure described above was employed, using a

commercially available Faujasite type zeolite as the first parent material. However, the method was


modified to introduce mesoporous activated carbon into the zeolite surface simultaneously with

demetallization. This process is fundamentally different from the carbon‐templated zeolite synthesis

described above as pre‐synthesised stable zeolites are used as the starting material as opposed to

growing zeolite crystals on a carbon‐template. The resulting materials, termed as Carbon Enhanced

Zeolites (CEZs), were subjected to characterisation, adsorption analyses and imaging studies and will

be discussed in detail in Chapter 4.

Attempt was then made to extend this principle in synthesizing ILZs with a similar architecture

consisting both micropores and mesopores. This was done by adding a simultaneous ion‐exchange

process for TMA+ introduction during the demetallization and carbon introduction processes

described above for CEZ synthesis. The resulting materials were termed as Carbon Enhanced Ionic

Liquidic Zeolites (CE‐ILZs) and also subjected to characterisation, adsorption analyses and imaging

studies. All adsorbents were degassed for 12h at 200 °C under vacuum prior to conducting adsorption

analyses.

Figure 3.16. Schematic showing the influence of Si/Al ratio on the desilication effect of MFI zeolites treated in NaOH solution and the associated mechanism of mesopore formation [103].


3.5.1 Acid functionalisation

Activated carbon, a low‐cost material with a large surface area and a wide pore size distribution, has

considerable merit in industrial gas adsorption applications [28]. It is known for higher CH4 adsorption

capacity compared to that of zeolites [28]. However, the CH4/N2 selectivity of this material is lower

than ILZs such as TMAY [93]. In an effort to combine the selectivity of TMAY with the capacity of

activated carbon, a series of synthesis procedures were carried out in attempts to produce carbon‐

based materials with selectivity values similar to those of ILZs, while retaining or enhancing the CH4

adsorption capacity of activated carbon. The synthesis methods were based on using commercial

activated carbon as parent materials and performing surface functionalisation that would enable ion

exchange leading to TMA introduction, as discussed in section 3.2.3.

As carbon is a non‐ionic (covalent) compound, it is not possible to introduce cations directly into the

carbon structure, in the manner it is done with zeolites. In order to perform ion‐exchange reactions

to introduce cations into carbon, it is first necessary to functionalise the carbon surface. The most

common method to functionalise carbon, as reported in the literature, is through treatment with a

concentrated acid [104, 105]. It has been reported that treating the carbon with nitric acid, a strong

oxidizing agent, results in carboxyl groups being implanted on the carbon surface according to the

reaction [106]:

(R1)

As the carboxyl group contains an H+ ion, it should be possible in theory to replace it with a TMA+ ion.

Furthermore, presence of ionic carboxyl groups on the carbon surface means that electrostatic

interaction between the adsorbent and specific gas molecules is now enhanced.


It should be noted that this approach is a novel attempt to improve adsorption capacity of carbon by

introducing organic cations to the carbon surface. Acid treatment is used to functionalise the carbon

surface to enable possible introduction of cations. Acid treatment alone is not expected to produce

capacity enhancements. The literature cited above [106] describes acid treatment as a method for

functionalising carbon surface.

Norit RB3, a commercial activated carbon manufactured by Cabot Corporation, Boston, USA, was

treated with 35% HNO3 for 6h at 90 °C before filtering out the liquid and adding fresh HNO3 to continue

the treatment for an additional 3h. The residue was filtered, washed with DI water and dried. A portion

of this intermediate compound, termed Acid‐Carbon1, was then treated with 0.5 M aqueous TMA‐Cl

for 6h at 40 °C before being filtered, washed and dried again. The final product is referred to as TMA‐

Carbon1. Using the same method but changing the HNO3 concentration to 60% and the TMA‐Cl

concentration to 5M, a second variant, termed TMA‐Carbon2, was synthesised. The intermediate acid‐

treated compound for this second variant is referred to as Acid‐Carbon2. The anticipated reaction

scheme is as follows:

Step 1 – Acid treatment of carbon:

Step 2 – TMA exchange of acid treated carbon:

+ 2TMA‐Cl + 2 HCl

TMA

TMA


CH4 and N2 adsorption isotherms at 0 °C for both variants are shown in figure 3.17 and figure 3.18

along with the baseline for the activated carbon parent material.

Figure 3.17. Adsorption isotherms of carbon (activated carbon Norit RB3), Acid‐Carbon1 and TMA‐Carbon1 at (a) 0 °C with CH4 and (b) 0 °C with N2.

Figure 3.18. Adsorption isotherms of carbon (activated carbon Norit RB3), Acid‐Carbon2 and TMA‐Carbon2 at (a) 0 °C with CH4 and (b) 0 °C with N2.


As illustrated by the plots, acid treatment reduces both CH4 and N2 adsorption capacity of activated

carbon and the capacities are further reduced after treatment with TMA‐Cl. When the acid

concentration is increased, the suppressions in capacities are even greater, both due to the acid

treatment and the TMA‐Cl treatment. A possible hypothetical explanation of this effect could lie in the

corrosive nature of the HNO3 used. The acid may cause damages to the pore structure of the carbon

and effectively reduce the surface area. Accordingly, using higher acid concentration can cause more

severe damage leading to greater reduction in adsorption capacity. This hypothesis can be tested by

conducting 77K N2 sorption experiments to measure BET surface area of the parent carbon as well as

the acid‐treated and the TMA‐exchanged carbons.

Another explanation could be that the carboxyl groups that functionalise the carbon surface through

acid treatment block the entrance to the pores and the bulky TMA+ ions that are introduced through

ion exchange cause further blockage. This theory also explains the greater reduction in capacity with

increasing acid concentration.

To avoid the detrimental effect of the acid on the carbon, attempt was made to perform TMA‐

exchange directly on the carbon, skipping the acid‐treatment step. Accordingly, the same activated

carbon parent material was treated with 5M TMA‐Cl for 6h at 40 °C and the resulting compound,

referred to as TMA‐Carbon4 was filtered, washed with DI water and dried. Figure 3.19 shows CH4 and

N2 adsorption isotherms for this material along with those for the parent carbon.


Figure 3.19. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐Carbon3 at (a) 0 °C with CH4 and (b) 0 °C with N2.

As shown in the plots, both CH4 and N2 isotherms are almost identical implying that, in absence of

acid, TMA‐Cl has no apparent effect on the carbon.

Since acid treatment is necessary for carbon functionalisation but hinders adsorptive performance,

attempt was made to synthesise another variant of acid‐functionalised carbon material using a

minimum acid concentration and perform TMA‐exchange with maximum TMA‐Cl concentration. This

time, the activated carbon was treated with 9% HNO3 for 6h at 90 °C before filtering out the liquid and

adding fresh HNO3 to continue the treatment for an additional 3h. The residue, referred to as Acid‐

Carbon4, was filtered, washed with DI water and dried. A portion of this intermediate compound was

then treated with 19M aqueous TMA‐Cl for 6h at 40 °C before being filtered, washed and dried again.

The end product is referred to as TMA‐Carbon4. CH4 and N2 adsorption isotherms for this material

along with those for the base carbon material are shown in figure 3.20.


Figure 3.20. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐Carbon4 at (a) 0 °C with CH4 and (b) 0 °C with N2.

As shown in the plots, both CH4 and N2 capacities are lower for TMA‐Carbon4 compared to the parent

carbon, indicating that, the modified treatment procedure is also unable to produce enhancements in

adsorptive performance of activated carbon.

3.5.2 Acid base reaction approach

It is thought that the reaction described in section 3.5.1 is not favourable because the product acid

HCl is stronger than the reactant carboxylic acid. The scheme was therefore modified to a more

favourable acid‐base reaction, as suggested in the literature [107], followed by an ion‐exchange

process, as described below:

Acid base step:

+ 2 NaOH + 2 H2O

Ion exchange step:

Na

Na


+ 2 TMA‐Cl + 2 NaCl

In order to implement this modified scheme, Acid‐Carbon4, whose synthesis is described above in

section 3.5.1, was treated with 0.5 M NaOH for 1h at room temperature. After filtering, washing with

DI water and drying, the residue was treated with 0.5 M TMA‐Cl for 6h at 40 °C and once again filtered,

washed and dried. CH4 and N2 adsorption isotherms for this material, referred to as TMA‐Acid Base

Carbon, together with those for the parent carbon are shown in figure 3.21.

TMA

TMANa

Na


Figure 3.21. Adsorption isotherms of carbon (activated carbon Norit RB3) and TMA‐AcidBaseCarbon at (a) 0 °C with CH4 and (b) 0 °C with N2.

The plots show that the modified treatment procedure involving acid base reaction has suppressed

both CH4 and N2 adsorption capacities of activated carbon by almost the same amount resulting in no

net gain in either selectivity or capacity. Thus, the modified procedure also fails to produce any

noticeable enhancement in the performance of activated carbon as an adsorbent.

3.5.3 Alkaline treatment

As neither the acid treatment procedure nor the modified procedure involving acid base reaction

produced desirable results, it was of interest to see if treatment with a base alone has any effect on

the adsorption performance of activated carbon. Norit R2030, another commercially available

activated carbon also manufactured by Cabot Corporation was used for this analysis. This material was

treated with 4M NaOH for 6h at 90 °C followed by filtration, washing with DI water and drying. Figure

3.22 shows CH4 and N2 adsorption isotherms for this material, referred to as Base‐Carbon, along with

those for the parent material.


Figure 3.22. Adsorption isotherms of activated carbon and Base‐Carbon at (a) 0 °C with CH4 and (b) 0 °C with N2.

As can be seen from the plots above, the isotherms for the parent carbon and the base‐treated carbon

are identical, indicating that base‐treatment has no apparent effect on adsorption characteristics of

activated carbon.

3.5.4 APS functionalisation

Another method suggested for functionalizing carbon surface involves the use of acidic ammonium

persulfate (APS) as the oxidizing agent [108‐111]. Prolonged APS treatment at high temperature is

reported to enable pore reopening after carbon functionalisation, as depicted in figure 3.23.


Figure 3.23. Schematic showing pore opening of functionalised carbon surface under increased oxidation temperature and duration [108].

In order to implement this method, activated carbon Norit R2030 was treated with 1M APS in 2M

aqueous H2SO4 in a Teflon autoclave for 24h at 120 °C. The residue was filtered, washed with DI water,

and dried. A portion of this material, referred to as APS‐Carbon, was then treated with 0.5 M NaOH

for 24h at 50 °C in order to carry out an acid base reaction and subjected again to filtration, washing

and drying. Finally, in order to perform ion exchange, the residual material was treated with 0.5 M

TMA‐Cl before carrying out the last round of filtration, washing and drying. CH4 and N2 adsorption

isotherms at 0 °C for the final product, referred to as TMA‐APS‐Carbon and the intermediate product,

APS‐Carbon, are shown in figure 3.24 along with the isotherms for the parent carbon material.


Figure 3.24. Adsorption isotherms of activated carbon, TMA‐APS‐Carbon, and APS‐Carbon at (a) 0 °C with CH4 and (b) 0 °C with N2.

As illustrated by the plots, APS treatment reduces both CH4 and N2 adsorption capacities for activated

carbon. However, TMA‐exchange causes the capacities to increase again, although they are still lower

than the capacities of the parent materials. The increase in CH4 capacity due to the presence of TMA+

was seen in the case of ILZs and can be explained by increased interaction between CH4 and TMA+.

However, the increase in N2 capacity after TMA exchange is a phenomenon that was not observed

with ILZs and is difficult to explain at this stage. In any case, as this effect causes loss in selectivity, the

results of this experiment show that the modified synthesis route involving APS functionalisation

followed by TMA exchange is also not successful in enhancing adsorptive performance of activated

carbon with respect to CH4/N2 separation.

3.5.6 Carbon Enhanced Ionic Liquidic Zeolites

As a further attempt to improve CH4/N2 selectivity of Carbon enhanced zeolites, a modified synthesis

procedure involving a one‐step reaction scheme was employed. A mixture of 10% activated carbon

Norit R2030 and 90% NaY by mass was treated with 0.5 M aqueous TMAOH for 24 h at 80 C. This

scheme is anticipated to enable simultaneous execution of three processes: (i) TMA‐exchange, (ii)


formation of mesopores due to the alkaline environment and (iii) carbon introduction. The residue

was filtered, washed with DI water and dried. Figure 3.25 shows CH4 and N2 adsorption isotherms of

this material, termed Carbon‐E‐TMAY1, at 0 °C and 30 °C.

Figure 3.25. Adsorption isotherms of activated carbon, Carbon‐E‐TMAY1, TMAY and NaY at (a) 0 °C with CH4, (b) 0 °C with N2, (c) 30 °C with CH4 and (d) 30 °C with N2.

At 0 °C, CH4 adsorption capacity of Carbon‐E‐TMAY1 is 20% higher than that of TMAY and 79% higher

than NaY while N2 adsorption capacity of Carbon‐E‐TMAY1 is 15% higher than TMAY but 18% lower

than NaY. At 30 °C, CH4 adsorption capacity of Carbon‐E‐TMAY1 is 8% higher than TMAY and 51%

higher than NaY while N2 adsorption capacity of Carbon‐E‐TMAY1 is 30% higher than TMAY but 41%


lower than NaY. Table 3.5 below shows a relative comparison of CH4 capacity and CH4/N2 selectivity

for four materials: activated carbon R2030, Carbon‐E‐TMAY1, TMAY and the parent zeolite NaY.

Table 3.5. Comparison of CH4 capacity and CH4/N2 selectivity of Carbon‐E‐TMAY1 with competitor adsorbents for CH4/N2 separation.

Temperature

& Pressure

0 °C, 1 bar 30 °C, 1 bar

Adsorbent CH4 Capacity(mmol/g)

CH4/N2

selectivity CH4 Capacity (mmol/g)

CH4/N2 selectivity

Activated Carbon R2030 1.22 2.93 0.75 3.12

Carbon‐E‐TMAY1 1.03 4.12 0.54 4.13

TMAY 0.86 3.91 0.5 4.99

NaY 0.57 1.84 0.35 1.6

At 0 °C, both CH4 capacity and CH4/N2 selectivity of Carbon‐E‐TMAY1 exceeds that of TMAY, making

Carbon‐E‐TMAY1 the first material to outperform TMAY in terms of both capacity and selectivity at 0

°C, although Carbon‐E‐TMAY1 is still unable to exceed the selectivity of TMAY at 30 °C. CH4 Capacity

of activated carbon is the highest at both temperatures but in terms of selectivity, it falls behind

both TMAY and Carbon‐E‐TMAY1.


In this chapter, we have demonstrated, experimentally, that critical structural features such as pore

architecture, chemical composition, crystal structure, surface morphology and presence of organic

cations play vital roles in determining characteristic properties of adsorbents. We determined the

optimum Si/Al ratio for Zeolite Y that produces the most desirable results for CH4/N2 separation and

used this to synthesise a number of novel adsorbents using NaY as the parent material. We also

confirmed that using an intermediate cation such as Mg2+ does not enhance the separation

performance of the final product. Similarly using a parent zeolite with a smaller cation such as H+ also

does not produce any desired enhancements. Thus, we established that a parent zeolite Y with a Si/Al


ratio of 2.55 and Na+ as the parent cation is likely to produce the most desirable CH4/N2 separation

enhancements when subjected to TMA‐exchange.

We also synthesised a number of novel carbon‐based adsorbents in order to analyse the effect of

treating activated carbon with oxidising agents such as HNO3 and APS and reducing agents such as

NaOH. Although a number of synthesis routes have been reported in the literature for combining

carbon with zeolite, the method proposed in this work is different in the sense that it uses

commercially available zeolite Y and activated carbon as the raw material and does not require zeolite

crystals to be synthesised in the laboratory.

Structural features within adsorbent materials can be manipulated to alter capacity and selectivity;

however, the changes implemented do not always produce desired results. For example, our

expectation was that performing TMA‐exchange on acid‐functionalised carbon will result in enhanced

CH4 adsorption and suppressed N2 adsorption leading to higher CH4/N2 selectivity. However, the

results obtained showed an opposite effect. Likewise, separation performance of CE‐ILZ was similar

to that of ILZ. Significant enhancement was not observed even after introduction of carbon into the

TMAY framework.

In the next Chapter, we demonstrate that massive gains in both CH4 and N2 capacity were observed

unexpectedly upon introduction of carbon solely on the NaY framework, leading to the formation of

Carbon Enhanced Zeolite (CEZ). This was an interesting observation which has been investigated in

detail in Chapter 4 to understand how adsorption occurs on this material and what role carbon plays

in its enhancement.


Chapter 4

Carbon Enhanced Zeolite

4.1 Introduction

Significant amount of work has been done in recent years to develop methods for combining carbon

with zeolite – a crystalline aluminosilicate material containing a highly ordered network of micropores

and a narrow distribution of pore size. The majority of this work has been driven by industrial

application of zeolites in catalysis with a desire to overcome diffusion limitation by using carbon to

introduce mesopores either inside or in between zeolite crystals, thereby allowing faster mass transfer

of large reactant and product molecules. Wei et al. [31] provided a comprehensive review of the

different methods, primarily based on using carbon templates to grow zeolite crystals, that are used

for this purpose.

A method for synthesising carbon‐zeolite composites, directed towards application in adsorptive

separation of gas mixtures, was described by Zhang et al. [112]. This method involves subjecting

naturally occurring Elutrilithe to carbonization, CO2‐activation and hydrothermal treatment with

NaOH followed by surface modifications and calcination. Samples with different carbon‐contents were

synthesised and the material containing the largest proportion of carbon was reported to have the

highest CH4/N2 selectivity. Capturing CH4 from low grade coal bed methane was suggested as a

possible application for this adsorbent.

Coal fly ash, a by‐product of coal‐fired electric power plants, can also be converted into carbon‐zeolite

composites. Primary objective of doing this is to simply convert an environmentally harmful material

into a product that is less hazardous and can potentially be used in industrial applications such as

adsorption and catalysis. Miyake et al. [113] and Jha et al. [114] described a method based on fusion

with NaOH followed by hydrothermal treatments that can be used to convert coal fly ash to carbon‐


zeolite composites. The resulting products were reported to be able to adsorb toxic Pb2+, Cu2+, Cd2+

and Ni2+ metal ions from solutions with a suggested application in industrial waste water purification.

In this work, we present a method, considerably simpler than the ones described above, to enclose

zeolite crystals in a mesoporos carbon layer, distinguishing that the approach is notably different from

growing the crystals on a carbon template as described by Wei et al. [31]. Our synthesis route involves

subjecting commercially available Zeolite Y and activated carbon to standard wet chemistry

procedures to enclose the zeolite in a mesoporous carbon film and does not require in‐situ

preparation of zeolite crystals. Motivated by widespread use of zeolites in adsorption processes, we

demonstrate using single‐component CH4 and N2 sorption experiments that the resulting product has

undergone significant enhancement in adsorption capacity compared to both the original zeolite and

the carbon from which it was synthesised. We then focus on understanding the cause of this

enhancement by studying the structural changes that took place within the material. In order to do

this, we employ a series of characterisation techniques, the findings from which are presented and

discussed in the following sections.

4.2 Synthesis

Three adsorbent materials were considered in this work. The first is a commercially available zeolite Y

(Faujasite), NaY CBV100, obtained from Zeolyst International in powder form. This material, referred

to as NaY in this work, was used in all experiments without any physical alteration. The second

adsorbent is a commercially available activated carbon, Norit R‐2030, obtained from Cabot

Corporation in extruded‐rod form. Prior to use in experiments, this material, referred to as C‐R2030,

was crushed into fine particles in a pneumatic press with an applied pressure of 20,000 psi.

The third material was synthesised in the laboratory using the two commercially available adsorbents

described above as parent materials. A mixture of 10% C‐R2030 and 90% NaY by mass was treated

with a 3M aqueous NaOH solution at 80 °C for 24 hours in a shaking water bath set to 120 rpm. The


residue was then collected by filtration and air‐dried at 50 °C for 24 hours. The resulting material is

referred to as carbon enhanced zeolite (CEZ) and used in all subsequent experiments described in this

work. Additional products were also synthesised by varying the carbon content of the initial mixture

as well as the concentration of the NaOH solution used in the treatment. Results of sorption

experiments conducted on these additional materials that illustrate the effect of carbon content and

NaOH concentration were used to determine the optimum levels mentioned above and are presented

in Appendix B1 and B2.

4.3 Characterisation

4.3.1 Chemical composition

Results of Inductively Coupled Plasma (ICP) tests conducted on NaY and CEZ are shown in Table 4.1.

The results show a decrease in Si content and increase in Al and Na content in CEZ compared to the

parent NaY. Desilication in zeolites is expected to be caused by NaOH treatment, as reported by Groen

et al. [103]. Due to desilication, some of the Si is removed from the zeolite. As a consequence,

percentage of Al in the framework increases leading to an increased concentration of negative charges

requiring a corresponding counter increase in positively charged Na+. The Si/Al ratio of 2.54 for NaY is

in close agreement with the Si/Al ratio of 2.55 provided by the supplier.

Table 4.1. Results of ICP tests conducted on NaY and CEZ. Si/Al ratio was calculated from the results.

% by mass Ratio

Na (Mw = 23) Al (Mw =26.98) Si (Mw = 28.09) Si/Al

NaY 6.72 8.69 23 2.54

CEZ 8.29 10.24 17.5 1.64

Thermogravimetric Analysis (TGA) was conducted to determine the carbon‐content of NaY, C‐R2030

and CEZ. The results are shown in figure 4.1. As expected, the results confirmed that NaY does not

contain any carbon while C‐R2030 contains about 89% carbon by mass that can combust in air. The

remaining 11% is likely to be ash and impurities. The mass of CEZ dropped by 14% when combusted


in air. Taking into consideration the 11% ash and impurities in C‐R2030, the carbon content of CEZ

becomes 12%.

Figure 4.1. Results of Thermogravimetric analysis conducted on (a) NaY, (b) C‐R2030 and (c) CEZ. The red line shows the change in mass of the sample with the corresponding change in temperature

(blue line).

4.3.2 X‐ray Diffraction

Figure 4.2 shows results of X‐ray Diffraction (XRD) analysis on CEZ and NaY. Both materials produce

XRD patterns typical of zeolite Y. However, the pattern for CEZ has a left‐shift compared to NaY,

indicating pore expansion. This left‐shift is more pronounced at higher angles.


Figure 4.2. X‐ray Diffraction (XRD) patterns for NaY and CEZ.

4.3.3 Surface area and pore volume

BET surface area and micropore area of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption

conducted individually on all three materials using a Micromeritics ASAP2020 instrument are shown

in Table 4.2. The mesopore area was calculated by subtracting the micropore area from the BET

surface area. Among the three materials, C‐R2030 has the largest BET surface area while CEZ has the

smallest. This means that a loss in surface takes place during the synthesis process of CEZ, implying

that some of the walls in between the crystal cavities may have been broken. This effect would reduce

the surface area but would create larger cavities with higher volumes. Also, the contribution of

mesopores in the surface area of CEZ is almost double that of NaY. The additional mesopores could

have been created not only by joining together of smaller cavities to form bigger ones but also due to

the presence of mesoporous carbon in CEZ.


Table 4.2. Surface area measurements of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption. Manufacturer's surface area for NaY (parent zeolite of TMAY ‐ CBV100) is available in Table 3.1.

Adsorbent BET Surface Area (m2/g) Micropore Area (m2/g)

Mesopore area as a % of BET Area

NaY 648 605 6.6%

CEZ 601 529 12.0%

C‐R2030 716 575 19.7%

Total pore volume of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption is shown in Table 4.3. As

expected, the pore volume of CEZ is higher than that of NaY, further indicating that intracrystalline

walls have been broken causing larger cavities to be formed.

Table 4.3. Pore volume measurements of NaY, CEZ and C‐R2030 obtained from 77K N2 sorption.

Adsorbent Total pore volume at 1 bar (cm3/g)

NaY 0.368

CEZ 0.406

C‐R2030 0.394

4.3.4 Sorption experiments

Single component N2 and CH4 sorption experiments were conducted on CEZ, NaY and C‐R2030 at three

temperatures under isothermal conditions over a pressure range up to 70 bars. As this pressure range

is well‐outside the operating limits of the ASAP2020, these high pressure sorption experiments were

conducted using a custom‐built computer‐controlled Sieverts apparatus [115] available at the

Hydrogen Storage Research Laboratory of Curtin University. The principal components of this setup is

outlined below in figure 4.3. The gas used was purchased from BOC Australia and was certified to

contain 99.999% N2 and 99.995% CH4 respectively.


Figure 4.3. Setup of the Sieverts apparatus used to conduct high pressure sorption experiments. The sample cell is wrapped in aluminium foil for insulation and submerged into the temperature control

bath during experiments to maintain isothermal conditions, as shown in the inset.

Enhancement in N2 adsorption capacity

Single component N2 adsorption isotherms for CEZ, C‐R2030 and NaY at 0 °C, 15 °C and 30 °C are

shown respectively in figures 4.4 (a), (b) and (c). At all three temperatures CEZ displays significantly

higher N2 capacity than its parent NaY as well as the C‐R2030 additive. At a pressure of 35 bar, for

example, the capacity compared to NaY is enhanced by 150% at 0 °C, 167% at 15 °C and 265% at 30

°C. This is a remarkable observation given that CEZ contains only 12% carbon and the XRD pattern

shows typical characteristics of zeolite Y. Furthermore, the capacity of CEZ even exceeds that of the

C‐R2030 additive. At 35 bar, capacity enhancement over C‐R2030 is 25% at 0 °C, 21% at 15 °C and 19%

at 30 °C. Interestingly, the enhancement over NaY increases at higher temperature while that over C‐


R2030 decreases with increasing temperature. As the pressure is further increased, the CEZ isotherm

reaches a maximum and bends over, a characteristic feature of excess adsorption at high pressure

[116],[117].

Figure 4.4. Single component N2 adsorption isotherms for CEZ, C‐R2030 and NaY obtained using a custom‐built computer‐controlled Sieverts apparatus under isothermal conditions at (a) 0 °C, (b) 15

°C and (c) 30 °C.

Enhancement in CH4 adsorption capacity

Isotherms for single component adsorption of CEZ, C‐R2030 and NaY at 0 °C, 15 °C and 30 °C are shown

respectively in figures 4.5 (a), (b) and (c). Similar to N2 adsorption, CH4 capacity of CEZ is also higher

than both C‐R2030 and NaY at all three temperatures. At a pressure of 30 bar, the enhancement in

CH4 capacity over NaY is 12% at 0 °C, 21% at 15 °C and 69% at 30 °C. Although the enhancement is


smaller than that observed with N2, the trend is still the same, meaning that the gain in capacity

increases at higher temperatures. With respect to C‐R2030, the enhancement in CH4 capacity at 30

bar is 11% at 0 °C and 8% at 15 °C. At 30 °C, no gain in capacity is observed as the CEZ and C‐R2030

isotherms almost overlap. Thus, as with N2, enhancement in CH4 capacity over C‐R2030 also diminishes

with increasing temperature. Hence, CEZ displays similar temperature‐dependent enhancement

trends with both N2 and CH4 but the magnitude of enhancement is greater with N2.

Figure 4.5. Single component CH4 adsorption isotherms for CEZ, C‐R2030 and NaY obtained using a custom‐built computer‐controlled Sieverts apparatus under isothermal conditions at (a) 0 °C, (b) 15

°C and (c) 30 °C.


Selectivity

CH4/N2 selectivity of CEZ, Carbon‐R2030 and NaY at 30 bar calculated by applying Eq. 3.6 to the single

component isotherm data presented in Figure 4.4 and Figure 4.5 are shown in Table 4.4.

Table 4.4. CH4/N2 Selectivity of CEZ, Carbon‐R2030 and NaY.

CH4/N2 selectivity

Temperature & Pressure

0 °C, 30 bar 15 °C, 30 bar 30 °C, 30 bar

CEZ 1.3 1.3 1.3

C‐R2030 1.4 1.5 1.6

NaY 3.0 3.2 3.0

Although CEZ remains a CH4‐selective adsorbent like its parent materials, CH4/N2 selectivity of CEZ is

significantly lower than that of NaY and slightly lower than that of Carbon‐R2030. Consequently it is

also lower than the selectivity of TMAY and Carbon‐E‐TMAY1 at 1 bar which are shown in Table 3.5.

This is expected as the enhancement in N2 capacity in CEZ is greater than the enhancement in CH4

capacity. This indicates that the structural changes implemented in CEZ favours an increased

adsorption of smaller molecules compared to larger ones. We hypothesise that this enhancement

could possibly be driven by an increase in cavity volume leading to multilayer adsorption. However,

one concern that remains is that this hypothesis does not fully align with the general expectation that

small molecule adsorption will be favoured by pores which are comparable in size with the diameter

of that molecule.

4.4 Possible mechanism

A schematic showing the well‐known Faujasite crystal framework of zeolite Y is shown in figure 4.6.

Adsorption in enabled primarily by the presence of the large supercage located in the crystal interior.

This supercage is reported to have a window width (pore opening) of 7.4 Å [118] and an internal width

of 12.5 Å [88].


Figure 4.6. A model of the Faujasite unit cell. Near the centre of each line segment is an oxygen atom. Silicon and aluminium atoms alternate at the tetrahedral intersections, except that Si

substitutes for Al at about 4% of the Al positions. Exchangeable cations are found in site I at the centre of a D6R ; site II at the centre of the S6R, or displaced from this point into a supercage; sites I’ and II’ lie in the sodalite cavity, on the opposite sides of the corresponding six‐rings from sites I and II, respectively; site III on a twofold axis opposite a four‐ring inside the supercage and site III’ off site III. The eighth sodalite cavity was removed from this diagram for simplicity and it is orthogonal to

the surface of this paper. Both figure and caption text are from Ahmed et al. [119].

The arrangement of Faujasite frameworks in a typical regular pyramid‐shaped NaY particle is shown

in figure 4.7 (a). The schematic represents a two‐dimensional cross‐section along the vertex of the

regular pyramid. Figure 4.7 (b) shows relative size of N2 molecules and a path through which the

molecules can enter the supercage in order to be adsorbed. Only monolayer adsorption takes place

as the fixed volume within the supercage is unable to accommodate sufficient molecules for multilayer

adsorption.

7.4 Å


Figure 4.7. Schematics showing (a) the arrangement of Faujasite frameworks in a typical regular pyramid‐shaped NaY particle and (b) the relative size of N2 molecules and a path through which they can enter the supercage in order to be adsorbed. Only monolayer adsorption is possible as volume inside the cage is insufficient to accommodate additional molecules for multilayer adsorption.

Figure 4.8 (a) shows a possible effect of NaOH treatment on the pore structure within a regular

pyramid‐shaped NaY particle when no carbon is added. As discussed above and supported by ICP

results, NaOH removes materials from within the zeolite through desilication. The resulting effect

causes a series of walls in between the supercages to be washed away, which essentially joins multiple

supercages to create a significantly wider cavity with larger pore volume and smaller surface area.

However, NaOH has a destructive effect not only on the interior walls of the zeolite particle but also

on the exterior walls, meaning that gas molecules can now easily pass through the cavities, as shown

in figure 4.8 (b), without getting adsorbed as there is no external surface to provide resistance. For

some cavities, it is possible that only internal walls are affected by NaOH and the exterior walls are

left intact. This results in additional monolayer adsorption leading to slight enhancement over

untreated NaY, as reported in Appendix B1.

The BET surface area and pore volume results discussed above support the notion of gain in volume

and loss in surface area.

(a) (b)


Figure 4.8. Schematic showing (a) possible effect of NaOH treatment on the pore structure within a regular pyramid‐shaped NaY particle when no carbon is added. Material is removed by desilication causing a series of walls in between the supercages to be washed away, joining together multiple supercages to create a significantly wider cavity with larger pore volume and smaller surface area. (b) Schematic showing the passage of gas molecules through zeolite cavities in absence of exterior

walls.

Figure 4.9 (a) shows the effect of NaOH treatment when the zeolite is premixed with carbon. This time,

we hypothesise that the exterior of the zeolite particle is possibly enclosed by a mesoporous carbon

film. Although NaOH is able to break and wash away silicon walls within the zeolite, it does not have

the same destructive effect on carbon. As a result the exterior carbon remains intact and is able to

provide a porous surface for gas molecules trapped inside the zeolite cavity to be adsorbed on.

Furthermore, with a larger cavity, the volume is now sufficient to accommodate enough molecules to

allow multilayer adsorption, as shown in figure 4.9 (b). Availability of sufficient pore volume as a

prerequisite for multilayer adsorption is in agreement with the BET model [45, 120].

(a) (b)


Figure 4.9. Schematic showing (a) the effect of NaOH treatment when the zeolite is premixed with carbon which causes the exterior of the zeolite particle to be possibly enclosed by porous carbon and (b) multilayer adsorption taking place inside the zeolite cavity as the exterior carbon film

enables additional molecules to become trapped inside the enlarged cavity.

This mechanism can also be used to explain the greater magnitude of enhancement observed with N2

than with CH4. A N2 molecule has a diameter of 3.64 Å [121] while a CH4 molecule has a diameter of

3.99 Å [122]. This means that, under identical conditions, a CEZ particle can accommodate more N2

molecules than CH4 molecules. A greater number of N2 molecules can therefore undergo multilayer

adsorption compared to CH4 molecules.

(a)

(b)


It should be noted that the above mechanism requires an optimum carbon to zeolite ratio as well

NaOH concentration. If too much carbon is added, the separate adsorption on carbon alone becomes

more pronounced than the combined effect of carbon and zeolite described above. If NaOH

concentration is too high, the extent of destruction is greater and the overall pore architecture is lost.

These effects are demonstrated in the sorption results presented in Appendix B2.

4.5 Further characterisation

4.5.1 Scanning Electron Microscopy (SEM)

Figure 4.10 and 4.11 show images obtained from scanning electron microscopy (SEM) conducted

respectively on NaY and CEZ using a Zeiss 1555 VP‐FESEM instrument. For NaY, regular pyramid‐

shaped structures, a characteristic feature of Zeolite Y, can be seen. In addition, as expected, the

surface is clean and smooth with sharp edges. For CEZ, however, the morphology is very different. The

surface has become very coarse and rough. The particles have lost their regular shapes, contain more

fragments and appear to be stacked together in blocks. Terrains and tracks created by desilication are

visible along the surface of CEZ. Also, upon close observation, a few pyramid‐shaped structures can

still be identified. But these particles have lost their sharp edges and the pointed vertex. A pyramid‐

shaped NaY particle in its original form has been marked in figure 4.10. A corresponding CEZ particle

produced after the possible transformations mentioned above has been marked in figure 4.11.


Figure 4.10. SEM image of NaY. The texture of the particles is clean and smooth with sharp edges. Pyramid‐shaped particles typical of zeolite Y can be clearly identified.

Figure 4.11. SEM image of CEZ. The surface has become rough and coarse with terrains created by desilication. Particles appear to be stacked on top of each other. Pyramid‐shaped particles can still

be detected.

A typical pyramid‐shaped

NaY particle

A CEZ particle, possibly

wrapped in carbon,

produced from a pyramid‐

shaped NaY particle


4.5.2 Scanning Transition Electron Microscopy (STEM)

A Scanning Transition Electron Microscopy (STEM) image of CEZ obtained using a FEI‐Titan G2 80‐200

TEM instrument is shown in figure 4.12 (a). Energy‐dispersive X‐ray Spectroscopy (EDS) mappings for

C, Si, Al, Na and O on the same STEM image frame are shown respectively in figures 4.12 (b), (c), (d),

(e) and (f). The EDS mappings show the presence of known elements in CEZ and cover the same area

and shape as the unmapped STEM image, confirming that all the particles seen in the STEM image are

CEZ particles. The mapping for carbon has a light green background as a carbon‐coated grid is used to

hold the sample. However darker green areas covering the same shapes shown in the other mappings

is clearly distinguishable and appear to be around exterior surface of the particles, confirming the

presence of carbon in CEZ. We consider this to further support our hypothesis that CEZ particles are

possibly enclosed by carbon films. Additional High resolution TEM images are available in Appendix

B5.

Background

carbon

Pyramid‐shaped

CEZ particle,

possibly wrapped

in carbon film

Pyramid‐shaped

CEZ particle

(a) (b)


Figure 4.12. STEM image of CEZ with (a) no EDS mapping as well as EDS mapping for (b) C, (c) Si, (d) Al, (e) Na and (f) O. Dark green areas in image (b) are more around the exterior of the particles

confirming that CEZ particles are surrounded by carbon.

4.5.3 Pore size distribution

Figure 4.13 shows results of pore size distribution based on surface area of NaY, CEZ and C‐R2030

obtained from 77 K N2 sorption experiments conducted using a Micromeritics ASAP2020 instrument.

NaY, as expected, has a narrow distribution of pore width ranging from 9 to 11 Å. The largest

contribution in surface area comes from 9 Å pores while some contribution is also made by 10 Å and

(c) (d)

(e) (f)


11 Å pores. In CEZ, some 9 Å pores are still present; however, the 10 Å and 11 Å pores have completely

disappeared and 13 Å pores, not originally present in NaY, are detected. This observation is consistent

with our hypothesis and suggests that the 10 Å and 11 Å pores present in NaY have been widened

through desilication to form 13 Å pores in CEZ, as described in the proposed mechanism. Pores of 15,

18 and 19 Å are also detected in CEZ. These pores were present in C‐R2030 as well and are consistent

with the presence of carbon on the exterior surface of CEZ particles, as per the mechanistic

description.

Results of pore size distribution based on pore volume also confirms the findings discussed above and

are presented in Appendix B4.

Figure 4.13. Pore size distribution based on surface area of CEZ, NaY and C‐R2030.



In this work we have demonstrated a method to enhance gas adsorption capacity of a narrow‐pore

commercially‐available zeolite, which is a Faujasite type Zeolite Y abundant in nature. The method is

considerably simpler than existing procedures to alter pore‐architecture within zeolites. It uses NaOH

treatment to widen intra‐crystalline cavities through desilication, and carbon to enclose zeolite

crystals so that gas molecules can be trapped inside – a possible mechanism hypothesised based on

results of characterisation experiments. For small molecules such as N2, our hypothesis suggests that

capacity is enhanced due to CEZ’s ability to enable multilayer adsorption as more volume is now

available to accommodate higher number of molecules. The enhancement in capacity has been

validated by conducting high pressure isothermal sorption experiments with single component N2 and

CH4. The expansion in pore width through desilication has been confirmed by ICP analysis, XRD

patterns and pore size distribution. The possible encapsulation of zeolite crystals by carbon has been

supported by TGA, pore size distribution and EDS mapping of STEM images.


Chapter 5

Conclusions and Future Research

5.1 Conclusions

The production and use of natural gas as an alternative to oil and coal continues to grow around the

world due to increasing environmental awareness and concerns. Some of the major sources of natural

gas in the world are either surrounded by oceans or thousands of kilometres away from the large

export markets. For these countries, the only feasible way to become global exporters is to have the

ability and a cost‐effective mechanism to convert billions of cubic metres of natural gas readily into

liquefied natural gas (LNG) so that it can be transported in tankers to fulfil shipment orders from

anywhere in the world. However, before a raw stream of natural gas can even enter the processing

stages of an LNG plant, the impurity contents, mainly CO2 and N2, must be reduced significantly to

meet the feed gas specifications. Conventionally, amine absorption and cryogenic distillations are the

two methods used to remove CO2 and N2 respectively from natural gas. As both of these methods are

energy‐intensive and increases the unit cost of LNG production due to high power consumptions, the

natural gas industry is considering a new process called Pressure Swing Adsorption (PSA) as an

alternative separation technology that is both cost‐effective and energy‐efficient. PSA has been

successfully used as a N2‐rejection technology in a number of industrial applications such as small scale

natural gas processing (feed rates of up to 15 MMscfd) [21], purification of biogas [22, 23] and

enrichment of coal mine methane [24]. In this work, we addressed two major requirements for a

successful cost‐efficient implementation of this technology in large scale LNG processing: (i) abundant

supply of adsorbents with high capacity and selectivity for the target gas components and (ii) reliable

process simulations to optimise the design parameters of a PSA cycle.


In Chapter 2, we presented a novel isotherm model, with physically meaningful parameters, that is

able to successfully capture the S‐shaped breathing isotherms exhibited by next generation

adsorbents such as MOF’s and ZIF’s with high selectivity for target gas mixtures. We demonstrated

that this model can be applied to both adsorption and desorption breathing using a wide range of

experimental data from the literature as well as our own laboratory. Furthermore we showed that this

model can be used to quantify hysteresis in breathing adsorbents including the onset and magnitude

of hysteresis as a function of both pressure temperature. We validated the model experimentally over

a wide range of temperatures reaching as low as 233K. To our knowledge no similar work involving

hysteresis has been reported before in the literature. Finally, we incorporated the functional form as

an isotherm model in a commercial simulation software and successfully executed PSA cycles to

predict the performance of a breathing MOF in separating a CH4/CO2 mixture using PSA.

In Chapter 3, we demonstrated experimentally that adsorptive properties of porous materials can be

altered by making changes to their pore architecture and chemical composition. Using zeolite Y as the

parent material we studied the effect of Si/Al ratio, parent and intermediate cations and alkalinity of

the synthesis medium on pure component CH4 and N2 capacities and resulting CH4/N2 selectivity of

the Ionic Liquidic Zeolite (ILZ) products synthesised using a previously established ion‐exchange

procedure. Among the different variants of Zeolite Y tested, the parent zeolite with Si/Al ratio of 2.55

and parent cation of Na+ produced ILZ’s with the highest CH4/N2 selectivity.

In Chapter 3, we also examined the effect of a number of different treatment procedures described in

the literature on CH4 and N2 adsorption capacities of activated carbon. These include surface

functionalisation treatments with nitric acid and ammonium persulfate (APS), treatment in an alkaline

medium and a two‐step treatment procedure in acidic and alkaline mediums. We performed extensive

characterisation experiments on the parent, intermediate and end products to determine the effect

of each stage of synthesis on adsorption. Finally, we developed a novel experimental technique to

introduce carbon into the frameworks of both zeolite Y and the ILZ TMAY. Although this did not


produce significant enhancements in the CH4/N2 selectivity of the resulting adsorbent over the already

achieved selectivity of TMAY, substantial gains in both CH4 and N2 capacity were observed when

carbon was introduced into the NaY framework using this method. We named this material Carbon

Enhanced Zeolite (CEZ) and pursued further research to understand the observed results.

In Chapter 4, we presented Carbon Enhanced Zeolite as a next generation high capacity adsorbent and

studied the structural changes through a series of advanced characterisation techniques including

inductively coupled plasma (ICP) analysis, X‐ray diffraction, Pore size distribution, Scanning Electron

Microscopy (SEM), Scanning Transition Electron Microscopy (STEM) and Energy‐dispersive X‐ray

Spectroscopy (EDS) mapping of STEM images. Based on this study, it was deduced that both carbon

and the alkaline synthesis medium play complimentary roles in the observed capacity enhancement.

NaOH treatment causes an expansion of the intra‐crystalline cavities through desilication leading to a

gain in pore volume which, in turn, enables increased multilayer adsorption. The role of carbon is to

enclose the zeolite crystals in an exterior porous layer which, during adsorption, facilitates higher

numbers of gas molecules to be “trapped” inside the adsorbent material. In this chapter, we

demonstrated how this proposed mechanism is supported by the results of the characterisation

experiments.

To conclude, this research work has made two contributions in the field of adsorptive separation of

gas mixtures. The first involves numerical modeling where a new isotherm model capable of enabling

process simulations with highly selective breathing MOF’s is developed. This is a major progress

towards the design and implementation of efficient and cost‐effective separation processes using

these flexible adsorbents. The second contribution is in material development where a novel method

to enclose naturally abundant zeolite Y in an external porous carbon layer resulting in major

enhancements in CH4 and N2 capacity has been identified. Supported by extensive characterisations,

the mechanisms for these enhancements were studied and explained in detail. This is a significant

achievement in the development of adsorbent materials and can lead to synthesis of low‐cost high


capacity zeolite‐based adsorbents with potential applications not only in gas separation but also in gas

storage.

5.2 Recommendations for future research

The findings of this research could be used as a baseline for further development of gas separation

processes through pressure swing adsorption (PSA). Process simulation results obtained by using the

pressure‐induced LJM model as an isotherm model for flexible adsorbents can subsequently be used

to determine the optimum design parameters of a PSA cycle to deliver the best separation

performance from these next generation adsorbents. Further experimental work can also be

conducted to extend the pressure‐induced LJM model to other flexible MOF’s, ZIF’s, COF’s etc. over a

wider range of temperature and pressure. High pressure adsorption experiments are currently being

conducted within the FSR group to further regress ZIF‐7 adsorption/desorption data to the pressure‐

induced LJM model.

Carbon enhanced zeolite (CEZ) developed in this work could be the subject of further research on this

material’s potential application in gas storage. The method used in its synthesis can be applied to

achieve substantial gain in methane capacity of naturally occurring faujasites. Further work can be

carried out to determine the effect of this method on other large pore zeolites for use in gas storage

applications. In addition, the range of temperature and pressure used in the isothermal

adsorption/desorption experiments can be extended to determine the effect of low temperatures and

high pressures on the observed gains. Furthermore, since the method produced capacity gains for

both methane and nitrogen, other gases such as carbon dioxide should also be included in the

experiments.


Appendices


Appendix A

A1 ‐ Development of the pressure‐induced LJM model

The LJM approach for temperature‐induced guest admission is based on the underlying assumption

that, because the system of oscillators present within a trapdoor adsorbent possesses a large number

of degrees of freedom, the energy distribution of these oscillators can be modelled as a normal

distribution [123, 124]. The LJM model for temperature‐induced guest admission can then be

summarized, mathematically, as [61]:

Φ 1√

(Eq. A1)

1 (Eq. A2)

where Φ is the cumulative probability that a given oscillator will admit a guest molecule, T0 the

threshold temperature defined as the temperature at which half of the oscillators present in the

system possess an energy greater than the energy barrier, β is a parameter that characterises the

width of the system‐specific distribution in temperature, is the fraction of external surface sites and

defects always available for adsorption, n is the adsorption capacity, is the maximum monolayer

adsorption capacity, b0 is the gas‐solid affinity coefficient, ΔH is the enthalpy of adsorption, R is the

gas constant, m is the surface heterogeneity coefficient and T and P respectively are the temperature

and pressure. It should be noted that the term outside the first square brackets in equation Eq. A2 is

essentially the classical Toth model with the conventional parameters , ΔH, b0 and m. The LJM

contribution therefore manifests in the term within the square brackets in equation Eq. A1, and

contains only two additional parameters if is taken to be approximately zero or is independently

estimated or measured.


Breathing effect in flexible MOFs is a highly guest dependent transition between collapsed or narrow

pore (np) and expanded or large pore (lp) phases that is characterised by gate‐opening and gate‐

closing pressures [62]. Thus, analogous to temperature‐regulated guest admission in trapdoor

adsorbents, breathing effect in MOFs can be defined as a pressure dependent phenomenon [56, 57,

63]. Similar to the threshold temperature required to overcome an energy barrier, we define a

threshold transition pressure, P0, as the pressure at which half of the unit cells present within the

framework transforms from np to lp state. We term this transition pressure as the “breathing”

pressure. By assuming that the ensemble of flexible unit cells constitutes a system with a large number

of degrees of freedom [125], the fraction of unit cells in the expanded state can be modeled as a

normal distribution as shown in figure A2. Accordingly, for a given pressure P the cumulative fraction,

ΦP, of unit cells that are in expanded state and will, therefore, admit guest molecules under isothermal

conditions is given by:

1√

(Eq. A3)

where βP is a parameter that characterises the width of the system‐specific distribution in pressure.


Figure A1. Distribution of unit cells within the MOF structure. When the pressure is (a) greater than the breathing pressure, more unit cells are in an expanded state than in the collapsed state, i.e, 0.5 < ΦP (P > P0) ≤ 1, (b) equal to the breathing pressure, number of unit cells in expanded state is the same as the number of unit cells in collapsed state, i.e., ΦP (P = P0) = 0.5, and (c) smaller than the breathing pressure, fewer unit cells are in expanded state than in collapsed state, i.e, 0 ≤ ΦP (P < P0)

< 0.5.

Although the admission of guest molecules in flexible MOFs is regulated by pressure, the

transformation from collapsed to expanded states also greatly depends on temperature. Hence it

should be expected that the breathing pressure, P0, would also be a function of temperature:

(Eq. A4)

The exact nature of the function f (T) is dictated by the interaction between a particular flexible

structure and the guest molecules being admitted. Du et al. analyzed the temperature‐dependent

structural change of a zeolitic imidazole framework (ZIF‐7) that exhibits breathing behavior during CO2

adsorption and showed that a plot of the natural logarithm of the empirical transition pressures (P0)

versus 1/T produces a straight line with a negative slope [66]. This implies that a Clausius‐Clapeyron

form, as shown below, can be used to model P0 as a function of 1/T:

P = P0

ΦP = 0.5

P > P0

0.5 < ΦP ≤ 1

P < P0

0 ≤ ΦP < 0.5

ΦP = Fraction of unit cells in expanded state

Rejection of guest molecules

(1- ΦP) = Fraction of unit cells in collapsed state

Admission of guest molecules

(a)

(b)

(c)

ΦP represents the shaded area under the curve


ln ln , (Eq. A5)

where k is the slope of the ln vs 1/T plot and , is the theoretical maximum value that the

breathing pressure can have when T becomes large. Upon rearrangement, equation Eq. A5 becomes:

, exp (Eq. A6)

In this work, equation Eq. A6 is used to substitute P0 in equation Eq. A3 to obtain as a function of

both pressure and temperature:

1,

√ (Eq. A7)

Equations Eq. A5 – Eq. A7 can then be incorporated into a standard adsorption model, such as the

classical Toth equation [126, 127], to obtain the pressure‐induced adsorption capacity, np, of a

breathing MOF as:

1,

√1 (Eq. A8)

A2 ‐ Parameters for the pressure‐induced LJM model

Parameter Physical Implication Units

βP A measure of the distribution width kPa

P0,ref Theoretical maximum value of the breathing pressure when temperature becomes large

kPa

k Breathing pressure coefficient (when is linear) kPa/K

Maximum monolayer adsorption capacity mol/kg

b0 Gas‐solid affinity coefficient 1/kPa

ΔH Enthalpy of adsorption J/mol

m Surface heterogeneity coefficient ‐

Fraction of external surface sites accessible to guest molecules ‐


A3 ‐ Summary of best‐fit parameters of the pressure‐induced LJM model and their

associated statistical uncertainties

Parameter Values and Uncertainties

Adsorbent Gas Mode (mol/kg)

b0 (1/kPa)

ΔH(J/mol)

m P0,ref (kPa)

βP (kPa)

k (kPa/K)

No. of observ.

R‐sq

Fe(bdp) CH4 Adsorption 21.244 +/‐

0.1874

1.398x10‐6 +/‐

2.9289x10‐8

12773+/‐

0.00014

1fixed

0.09282+/‐

0.006529

90276+/‐

5.45x10‐5

243.99 +/‐

0.00058

1054+/‐

0.0166

242 0.996

Fe(bdp) CH4 Desorption ” ” ” ” ” 37222+/‐

2750.9

169.66 +/‐

15.321

991.93+/‐

20.601

102 0.987

Co(bdp) CH4 Adsorption 16.267 +/‐

0.14923

1.1102x10‐6 +/‐

2.3214x10‐8

14028+/‐

4.19x10‐5

1fixed

0.081978+/‐

0.0071382

45258+/‐

4.867x10‐5

316.81 +/‐

0.00047

969.03+/‐

0.007253

159 0.997

Co(bdp) CH4 Desorption ” ” ” ” ” 4886.5+/‐

633.67

107.39 +/‐

14.65

579.14+/‐

37.78

90 0.99

mmen‐ Fe2(dobpdc)

CO2 Adsorption 4.05fixed

2.82x10‐10 fixed

51816fixed

0.7fixed

0.2286fixed

6.801x1010

+/‐ 7.327x10‐9

1.308 +/‐

0.20609

6990.5+/‐

1.5702

101 0.984

mmen‐ Co2(dobpdc)

CO2 Adsorption 4.53fixed

6.22x10‐9

fixed 42039fixed

0.7fixed

0.2140fixed

1.481x1011

+/‐ 3.068x10‐9

6.2274 +/‐

0.91038

6874.2+/‐

1.415

139 0.973

amino‐ MIL‐53(Al)

CO2 Adsorption 6.5232 +/‐

9.15x10‐6

4.9115x10‐9 +/‐

5.9229x10‐10

36669+/‐

5.28x10‐18

1fixed

0.32574+/‐

0.0095042

3x105

+/‐ 1.158x10‐8

561.7 +/‐

1.52x10‐6

1612.5+/‐

1.185x10‐5

53 0.993

amino‐ MIL‐53(Al)

CO2 Desorption ” ” ” ” ” 5.8148x105

+/‐ 7.843x105

494.46 +/‐

53.621

2073.4+/‐

400.69

39 0.99

ZIF‐7 CO2 Adsorption 1.9318 +/‐

0.0217

3.2516e‐07 +/‐

3.5518e‐08

29089fixed

1fixed

0.2fixed

7.28e+06 +/‐

1.3344e‐07

0.54632 +/‐

0.10041

3494.6+‐

0.0039923

37 0.994

ZIF‐7 CO2 Desorption ” ” 29869+/‐

107.1

” ” 1.525e+06+/‐

63418

0.46 fixed

3324fixed

66 0.97

For CO2 adsorption on mmen‐Fe2(dobpdc) and mmen‐Co2(dobpdc) and CH4 adsorption/desorption on

amino‐MIL‐53(Al), MATLAB was unable to produce a best‐fit regression by allowing all seven

parameters to vary. Therefore, , b0, ΔH and were fixed at values obtained by iteration and a three‐

parameter non‐linear regression was carried out in MATLAB.

For an overwhelming majority of the fits, uncertainties in fitting parameters are significantly smaller

than the parameter values being reported. Parameter values whose uncertainties are greater than

100% are highlighted in red. These are for CH4 adsorption and desorption on amino‐MIL‐53(Al) for

which isothermal data is available at only one temperature causing the fitting parameters to have high

uncertainties.


A4 ‐ Residual plots

Figure A2. Residuals (nmeas ‐ ncalc) of the Pressure‐induced LJM model’s regression on experimental data for the adsorption of CH4 on febdp reported by Mason et al. [48].

Figure A3. Residuals (nmeas ‐ ncalc) of the Pressure‐induced LJM model’s regression on experimental data for the desorption of CH4 on febdp reported by Mason et al. [48].


Figure A4. Residuals (nmeas ‐ ncalc) of the Pressure‐induced LJM model’s regression on experimental data for the adsorption of CH4 on cobdp reported by Mason et al. [48].

Figure A5. Residuals (nmeas ‐ ncalc) of the Pressure‐induced LJM model’s regression on experimental data for the desorption of CH4 on cobdp reported by Mason et al. [48].


A5 ‐ Simulation of the dynamic breakthrough of an equimolar CO2/CH4 mixture

A dynamic breakthrough separation of an equimolar binary mixture of CO2/CH4 using amino‐MIL‐

53(Al), for which experimental breakthrough data is available from Couck et al. [60] for comparison,

was then simulated. The second simulation under identical conditions was then performed using the

classical Toth model as the isotherm model. The experiment conditions such as feed composition,

pressure, temperature, flow rate and adsorption column dimensions were taken to correspond to the

experiment reported by Couck et al. [60] The simulation parameters such as, inter‐ and intra‐particle

voidage, adsorbent bulk density, radius, and mass transfer coefficients were taken from Gomez et al.

[79] These along with the experiment conditions are listed in section‐7 below. Isotherm parameters

were taken from the pressure‐induced LJM model’s regression of measured CO2 and CH4 adsorption

data [60] for amino‐MIL‐53 (Al), as discussed above.

The pressure‐induced LJM model fitted on amino‐MIL‐53 (Al) CO2 and CH4 adsorption isotherm data

from Couck et al. [60] at 303K is shown in figure A6 (a) and (b) respectively below. Also, the

corresponding Toth model, without the LJM modifications, is also plotted. As expected, the Toth

isotherm deviates from the LJM isotherm in the breathing region for both CO2 and CH4. To

demonstrate the effect of this deviation on simulation, three pressures: 1 bar, 11 bar, and 30 bar are

chosen.

Figure A6. Deviation of the Toth model from the pressure‐induced LJM model in comparison to measured isothermal data at 303K for the adsorption of amino‐MIL‐53 (Al) on (a) CO2 and (b) CH4.


Results of dynamic breakthrough simulation of an equimolar binary mixture of CO2/CH4 using amino‐

MIL‐53(Al) at feed pressures of 1 bar, 11 bar and 30 bar are shown in figures A7 (a), (b) and (c)

respectively.

At 1 bar, as there are deviations between the CO2 isotherms predicted by LJM and Toth models,

deviations in the breakthrough curves predicted by the two models, as shown in figure A7 (a) are also

expected. The Toth model predicts significantly higher adsorption capacity for CO2 than LJM does and

hence it takes much longer for the adsorption bed to become saturated with CO2 before it is released

in the product stream. The CO2 breakthrough curve predicted by Toth is therefore shifted to the right

of the CO2 breakthrough curve predicted by LJM. As for CH4, the Toth capacity is also higher but not

as much as CO2. Therefore, the Toth breakthrough curve is also shifted to the slight right of the LJM

breakthrough curve.

At 11 bar, the deviations between the isotherms predicted by the two models are larger, with Toth

predicting higher capacity for both CO2 and CH4. As a result, the extent of the right‐shift of the Toth

breakthrough curves compared to the LJM curves is greater, than at 1 bar, for both gases. This is shown

in figure A7 (b). At 30 bar, there is no difference between the adsorption capacities predicted by the

two models and hence both breakthrough curves are identical (figure A7 (c)).

A qualitative comparison with experimental breakthrough data for amino‐MIL‐53 (Al) reported by

Couck et al. [60] shows that breakthroughs of CH4 and CO2 take place at approximately 25 and 100

seconds respectively which are in agreement with the breakthrough times predicted by the simulation

using LJM as the isotherm model. The experimental data, however, also shows roll‐up effect [128]

which ASPEN Adsorption is unable to simulate.


Figure A7. Comparison between the classical Toth model and the pressure‐induced LJM model used as isotherm models in ASPEN Adsorption for performing column breakthrough simulations of an

equimolar CO2/CH4 mixture using amino‐MIL‐53 (Al) as the adsorbent at 303K with a feed pressure of (a) 1.013 bar, (b) 11 bar and (c) 30 bar.


One limitation of the simulation package in regards to modeling processes involving flexible

adsorbents is that it only allows fixed values for the inter‐ and intra‐particle voidage. For breathing

MOFs, of course, both these values are expected to change during adsorption as the MOF structure

undergoes transitions from the collapsed to the expanded state.

A6 ‐ Regression of CH4 and CO2 adsorption isotherms on Zeolite 13X using the Toth

Isotherm model

Figure A8. The Toth isotherm model regressed to (i) CH4 and (ii) CO2 adsorption data reported by Cavenati et al. [81]. Symbols denote experimental data and curves represent the Toth model.

Adsorbent Gas (mol/kg) b0 (1/kPa) ‐ΔH (J/mol) m

Zeolite 13X CH4 10.1 1.55 x10‐06 15600 0.608

Zeolite 13X CO2 12.9 1.02 x 10‐09 58116 0.199

Summary of parameters used to regress the Toth isotherm model to CH4 and CO2 adsorption data reported by Cavenati et al. [81].

A7 ‐ Simulation parameters

(a) Values of adsorbent properties used in ASPEN Adsorption to characterise the adsorption beds:

Parameter amino‐MIL‐53(Al) [79] Zeolite 13X [129‐131]

Inter‐particle voidage (Ei) 0.573 0.348

Intra‐particle voidage (Ep) 0.146 0.3696

Bulk solid density (RHOs) (kg/m3) 385 1120

Particle radius (Rp) (m) 0.00005 0.0007

Mass Transfer Coefficient (MTC) – CH4 (1/s) 0.5 0.29

Mass Transfer Coefficient (MTC) – CO2 (1/s) 0.8 1.1


(b) Dynamic Breakthrough and PSA simulation parameters used in ASPEN Adsorption:

Dynamic Breakthrough PSA

Feed Composition 0.5 CH4 + 0.5 CO2 0.5 CH4 + 0.5 CO2

Feed Pressure (bar) 1.013, 11, 30 20.045

Height of adsorbent layer (m) 0.3 0.3 (both beds)

Internal diameter of adsorbent layer (m) 0.00216 0.03

Temperature (K) 303 303.15

(c) Dynamic breakthrough simulation flowsheet:

(d) (i) PSA simulation flowsheet:

F 1

B 1

P 1

S 1

S 2


(ii) PSA cycle steps:

Step No. Description Duration

Step 1 Bed 1 adsorption Bed 2 regeneration (purge)

1 minute 1 minute

Step 2 Bed 1 blow down Bed 2 repressurisation

Until Bed 1 pressure falls below 6.5 bar Until Bed 2 pressure exceeds 20 bar

Step 3 Bed 2 adsorption Bed 1 regeneration (purge)

1 minute 1 minute

Step 4 Bed 2 blow down Bed 1 repressurisation

Until Bed 2 pressure falls below 6.5 bar Until Bed 1 pressure exceeds 20 bar


Appendix B

B1 ‐ Effect of carbon‐content

Different products were synthesised following the same method as CEZ but varying the carbon‐

content of the initial mixture while keeping the NaOH concentration fixed at 3M and all other

conditions the same. Figure B1 shows the single‐component N2 adsorption capacity of these products

under isothermal conditions at 0 °C for pressures up to 1 bar, obtained using a Micromeritics

ASAP2020 instrument. Results for parent, untreated NaY and C‐R2030 are also shown on the same

plot.

As shown in figure B1, the product with no carbon‐content has higher N2 capacity than the product

with 20% carbon‐content as well as the parent NaY and C‐R2030. However, the product with 10%

carbon‐content has the highest N2 capacity amongst all adsorbents shown in this plot. This was

therefore chosen as the optimum carbon to zeolite ratio for CEZ.

Figure B1. Effect of carbon‐content.


B2 ‐ Effect of NaOH concentration

Different products were synthesised following the same method as CEZ but varying the concentration

of NaOH used in the treatment while keeping the carbon‐to‐zeolite ratio of the initial mixture fixed at

10%C‐R2030+90%NaY and all other conditions the same. Figure B2 shows the single‐component N2

adsorption capacity of these products under isothermal conditions at 0 °C for pressures up to 1 bar,

obtained using a Micromeritics ASAP2020 instrument.

As shown in figure B2, N2 capacity increases as the concentration of NaOH is increased from 2M to 3M

but falls sharply, to even lower than the 2M treatment, as NaOH concentration is further increased to

4M. NaOH. Thus, the product treated with 3M NaOH has the highest N2 capacity amongst these three

adsorbents and was therefore chosen as the optimum NaOH concentration for CEZ.

Figure B2. Effect of NaOH concentration.


B3 ‐ Toth fittings

Fittings produced by regressing the classical Toth isotherm model to results of pure component N2

adsorption experiments conducted on CEZ, C‐R2030 and NaY at temperatures of 0 °C, 15 °C and 30 °C

are shown respectively in figures B3, B4 and B5. Adsorption data until the maximum point of each

isotherm, prior to any decrease in capacity due to excess adsorption were used for the fits. The

Regression R2 for these fits were 0.996, 0.998 and 0.987 respectively for CEZ, C‐R2030 and NaY.

Figure B3. Toth Fitting of CEZ on N2.


Figure B4. Toth Fitting of C‐R2030 on N2.

Figure B5. Toth Fitting of NaY on N2.

Fittings produced by regressing the classical Toth isotherm model to results of pure component CH4

adsorption experiments conducted on CEZ, C‐R2030 and NaY at temperatures of 0 °C, 15 °C and 30 °C


are shown respectively in figures B6, B7 and B8. The Regression R2 for these fits were 0.995, 0.999 and

0.989 respectively for CEZ, C‐R2030 and NaY.

Figure B6. Toth Fitting of CEZ on CH4.

Figure B7. Toth Fitting of C‐R2030 on CH4.


Figure B8. Toth Fitting of NaY on CH4.

B4 – Pore size distribution (based on pore volume)

Figure B9 shows results of Pore size distribution based on pore volume of NaY, CEZ and C‐R2030

obtained from 77K N2 sorption experiments conducted using a Micromeritics ASAP2020 instrument.

These results provide identical findings to those obtained from surface‐area based Pore size

distribution results.


Figure B9. Pore size distribution based on pore volume of CEZ, NaY and C‐R2030.

B5 – High resolution TEM images

TEM images of CEZ obtained using a FEI‐Titan G2 80‐200 TEM instrument are shown in figures S10‐

S11. These images show particle shapes that confirm the SEM images reported earlier. In addition,

High Resolution TEM images, clearly showing zeolite lattice fringe patterns, obtained using the same

instrument are shown in figures S12‐S14. The small fragments seen in figures S12 and S13 may be

pieces of carbon which broke‐off from the encapsulating carbon film and became embedded on the

exterior surface of zeolite particles.


Figure B10. TEM image of CEZ.

Figure B11. TEM image of CEZ.


Figure B12. High Resolution TEM image of CEZ.






Appendix C

C1 ‐ Foreword

This article has been published in Chemical Communications, vol. 54. pp. 3134‐3137, 2018. It is

reproduced in this appendix as it involves motivational works carried out in the early phase of this

research project, demonstrating how a zeolite’s chemical composition can play a major role in

selectivity.

C2 ‐ The Quest for a Trapdoor Zeolite for Exclusive Admission of CO2 at Industrial

Working Temperatures

Tao Du, Xin Fang, Liying Liu, Jin Shang, Bin Zhang, Yichao Wei, He Gong, Shamsur Rahman,

Eric F. May, Paul A. Webley, Gang (Kevin) Li

High purity molecular trapdoor chabazite with an optimal Si/Al ratio of 1.9 was prepared from fly

ash. Gas adsorption isotherms and binary breakthrough experiments show dramatically large

selectivity of CO2 over N2 and CH4, among the highest for physisorption based carbon capture at

industrial operation temperatures.

Adsorption based separation of carbon dioxide from important industrial gas mixtures such as flue gas

and natural gas has gained worldwide research interest owing to increased environmental and

economic incentives [132‐134]. High CO2 selectivity and low manufacture cost are among the key

merits to the timely implementation of new carbon capture adsorbents by industry. Conventionally,

depending on the difference between gas molecules’ properties, adsorption selectivity essentially

relies on three separation mechanisms in physisorption (the preferred regime due to low energy

consumption for regeneration), namely equilibrium [135], kinetics [136] and size‐sieving [137]. Carbon

dioxide is known for its large multi‐poles and polarizability and able to bind preferentially to

adsorbents, giving higher equilibrium capacity than light components such as N2 and CH4 [138].

However, equilibrium selectivity is often limited and rapidly reduced with increasing pressure because


of the linear trend of the isotherms of light components [132]. Likewise, kinetic separations frequently

suffer from reduced mass transfer rates and compromised bed capacity due to short cycle times. Size‐

sieving may claim perfect separations but not readily available when the size difference between the

two molecules are negligible [139].

In contrast, a new physisorption mechanism namely, “molecular trapdoor” [133] which were recently

discovered in small pore zeolites, provides an alternative pathway to make ultra‐high selective

adsorbents. Molecular trapdoor is based on the ability of certain gas molecules to reversibly and

temporarily open up the cation blocked 8MRs (eight‐membered oxygen rings) of ~0.38×0.38 nm pore

apertures the only entrance for the admission of guest molecules into small pore zeolites, such as K‐

CHA, Cs‐CHA, and RHO [133, 139]. Specifically, strong guest molecules (such as CO2) having sufficient

interaction to deviate the door‐keeping cations are able to enter the zeolite supercage, whereas weak

molecules (such as N2 or CH4) are excluded. Therefore, record high selectivities, 93 for CO2/CH4

separation and 80 for CO2/N2 separation, have been achieved at 273 K and 100 kPa with a r1KCHA

trapdoor zeolite (i.e. potassium chabazite with Si/Al ratio of 1) in 2012 [133]. However, one limitation

of the trapdoor effect is that adsorption must be conducted well below the threshold admission

temperature T0; above this threshold, the door‐keeping cation will gain sufficient thermal energy and

become mobilized to allow for non‐discriminative admission of molecules and fail the trapdoor effect.

Recent process demonstration using pressure swing adsorption (PSA) technology for carbon capture

with kilogram‐scale trapdoor chabazites indicated that CO2 recovery and purity dropped dramatically

above 291 K column temperature using an r2.2KCHA [140]. Given most post‐combustion carbon

capture processes deal with relatively hot flue gases (313 ‐ 363 K even after washing tower) and the

adsorption step of a PSA cycle is strongly exothermic due to adiabatic operations [141, 142], there is

a strong demand to develop a trapdoor zeolite that can exclusively adsorb CO2 at industrially practical

temperatures.


The main objective of this work is to find such a trapdoor zeolite that rejects N2 and CH4 but adsorbs

CO2 with a working temperature suitable for gas industry, e.g. from sub‐ambient up to 348 K.

Our strategy to elevate the threshold admission temperature of CH4 and N2 is to increase the energy

barrier ΔE required for admitting the guest molecules through the cation blocked 8MR doorway. Here

we use density functional theory (DFT) calculations to demonstrate the ΔE for a given gas‐zeolite

system has a strong dependence on the density of cations in the trapdoor zeolite. In a typical

potassium chabazite system (see figure CS3 for detailed 3D structure) [143], we studied the admission

process of CO2 gas as an example in the scenarios of two different Si/Al ratios, namely 3 and 1, with

the corresponding chabazite denoted as r3KCHA and r1KCHA, respectively. It is obvious that lower the

Si/Al ratio, higher the cation density. As shown in figure C1b, the admission energy barrier of ΔE(CO2‐

r3KCHA) is 22% lower than that of ΔE(CO2‐r1KCHA). This is because more cations in the chabazite

supercage can substantially increase the space hindrance for the movement of the door‐keeping

cation and the guest molecules, and in the meantime high charge density in the aluminosilicate

framework will make the cations less mobile due to increased charge repulsion. It is thus expected

that reducing the Si/Al ratio in KCHA could effectively increase the threshold admission temperature

T0. In a typical r2.2KCHA, T0 for CH4 and N2 is 266 K and 254 K, respectively [134]. However, attempts

to produce functional trapdoor chabazite with Si/Al ratio below 2 were not successful due to inherent

drawbacks of the synthesis method [144] in which alumina hydroxide slurry was used to enrich the Al

content of the chabazite and consequently inhomogeneous Si Al exchange and presence of

amorphous alumina in product are unavoidable.


Figure C1. (a) The 8MR and door‐keeping potassium cation movement during CO2 adsorption. (b) Energy profiles for the potassium cation of r1KCHA and r3KCHA. (c) SEM images for r1.9KCHA. (d) XRD patterns of raw fly ash, intermediate and resultant r1.9CHA. (The scales of the atoms in this

illustration are not in portion).

Herein, we report an ultra‐low Si trapdoor chabazite (Si/Al = 1.9) with unparalleled CO2/CH4 and

CO2/N2 adsorption selectivity at industrially practical working temperatures up to 348 K. Importantly,

this r1.9KCHA was synthesized from coal fly ash through a template‐free hydrothermal method which

is very different from conventional procedures reported by others using either organic structure‐

directing agents or inter‐zeolite conversion from Y procurers [145, 146]. Briefly, selected fly ash with

Si and Al content close to low silica chabazite was reacted with excessive KOH powder through a high

temperature fusion process at 923 K for 1 h to break up mullite and quartz phases (figure C1d) and


the resultant mixture was treated under hydrothermal condition for 4 days to produce the target

r1.9KCHA (see Supporting Information below for complete procedure).

The morphology and structure of the starting material and the end product was examined by scanning

electron microscope (SEM) and X‐ray diffraction (XRD), respectively. As shown in figure CS1, the fly

ash is comprised of particles with a wide size distribution from sub‐micron to 30 micron, in the shape

of spherical beads, irregular lumps and short needles which is mainly attributed to mullite and quartz,

according to literature [145] and their characteristic XRD patterns (figure C1d). In contrast, the as‐

synthesized chabazite (figure C1c) presents a walnut‐like shape with a uniform diameter of

approximately 2‐5 micron [147]. High resolution SEM (figure C1c) reveals the “walnuts” are intergrown

multi‐crystals of chabazite. It is worth mentioning that the morphology of our chabazite is quite

different from those reported before through intercrystal conversion process [138] which exhibited a

similar morphology to the precursor zeolite Y. Meanwhile, amorphous material was hardly noticeable

in the SEM images of the as‐synthesized chabazite. High purity high crystalline chabazite phase was

clearly evidenced by the characteristic XRD (figure C1d).

The Si/Al ratio of the as‐synthesized chabazite was verified to be 1.9 by XRF analysis. Apart from the

dominant potassium, other metallic minor components from the fly ash were also partially retained

in the chabazite giving a product composition of K9.79Fe0.44Ca0.57Al12.25Si23.75O72, suggesting the product

chabazite has mixed cations. Note that potassium is an effective door‐keeping cation in chabazite, and

the minimum number of potassium cation required for a chabazite to keep the trapdoor effect is 9

per unit cell [133], which is less than the number in the chabazite synthesized from this work,

suggesting its potential trapdoor capability.

The adsorption properties of our r1.9KCHA were analyzed with single component N2, CH4 and CO2,

respectively. The N2 adsorption on this chabazite at 77K was found to be close to that of nonporous

powered materials, giving rise to negligible BET surface area, which is typical to trapdoor zeolites. At

temperatures of 273, 303 and 333 K, the adsorption of both N2 and CH4 still remains negligible, with


uptake signals falling in the noise level of the measurement instrument (ASAP 2010, Micromeritics).

Even at 333 K temperature and 100 kPa partial pressure, the adsorption capacity of N2 and CH4 on

r1.9KCHA is still below 0.02 mmol/g, about 25‐30 times smaller than that of r2.2KCHA reported in our

earlier work [133], indicating the threshold admission temperature T0 for N2 and CH4 have been

substantially elevated to above 343 K.

On the contrary, a considerable amount of CO2 was adsorbed onto r1.9KCHA. At 100 kPa, the

adsorption capacity at 273K, 303K and 333K is 0.81 mmol/g, 1.49 mmol/g, and 1.54 mmol/g,

respectively, showing a positive dependence on temperature: greater equilibrium capacity at higher

temperature. This trend is opposite to that of the normal Arrhenius type physisorption. Certainly, the

equilibrium adsorption of CO2 will not increase monotonically with temperature, and it peaks around

333 K and then declines with further increase of temperature. A “bell shaped” isobar of CO2 adsorption

is observed from figure CS2, which is signatory to trapdoor effect previously reported for weak

molecules like CH4 and N2 in r2.2KCHA [133, 134]. However, it is rare to see temperature regulated

admission of CO2 molecules, because CO2 is able to interact strongly with door‐keeping cations due to

its large quadrupole moment (13.4×10‐40 C∙m2) and polarizability (29.1×10‐25 cm‐3). This study suggests

if temperature is low enough, it is possible to restrict the pore access of trapdoor zeolites for strong

molecules like CO2, potentially enabling sieving of polar molecules.


Figure C2. CO2, N2 and CH4 adsorption isotherms on r1.9KCHA at (a) 273, (b) 303 and (c) 333 K, and (d) breakthrough curve for r1.9KCHA at 348 K and total pressures of 1 bar.

Based on the single component adsorption isotherms, one can boldly anticipate that the r1.9KCHA

will exclusively adsorb of CO2 but reject CH4 and N2 in a gas mixture in the temperature region of 273‐

333 K. However, such extrapolation can be risky as multicomponent selectivity is not always

guaranteed from equilibrium information of single components, regardless the model of prediction

(e.g. IAST [148] or extended Langmuir).

To evaluate the real separation selectivity of CO2 against CH4 and N2 on our r1.9KCHA, multicomponent

column breakthrough experiments were conducted using gas mixtures of CO2/N2 (or CH4):50/50 at

temperatures ranging from 303 up to 348 K, at 1 and 3 bar, respectively. In a typical run, a stainless

steel column containing 25.8 g of tightly packed r1.9KCHA pellets (~1 mm in diameter) was initially

flushed by He and, then switched to the flow of target gas mixtures at the same pressure and the gas

composition and the mass flowrate at the outlet of the column were recorded as a function of elution

time to give the so‐called breakthrough curve (refer to Supporting Information below for full


procedure) [96]. As shown by figure 2d, instantaneous elution of N2 occurred right after the gas switch

along with the flush of the prefilled He, while CO2 was detected at the outlet after a substantially

longer time, indicative of negligible loading of N2 but a substantial uptake for CO2. Remarkably, the

actual CO2/N2 and CO2/CH4 selectivities under these conditions are enormously large (summarized in

Table CS2), e.g. 688 for the case in figure 2d, as determined from mass balance calculations, which are

consistent with our earlier anticipations from single component isotherms.

Table C1. Comparison of the performance of the adsorbing materials with highest CO2 selectivities.

Compounds Selectivity at 1 bar CO2 uptake at

partial pressure (mmol/g)

Temp. (K)

Ref. CO2/CH4 CO2/N2

r1.9KCHA 50/50, 583a 50/50,90 (303 K)a

50/50,688 (348 K)a

1.01 (51 kPa, 303 K) 0.83 (51 kPa, 348 K)

303~348 this work

r1KCHA 15/85, 79a 1.59 (17.4 kPa) 293 2 r2CsCHA 15/85, 109a 1.63 (17.4 kPa) 293 2 MIL‐53(Cr) 50/50, 3.2a 2.25 (500 kPa) 303 20 porph(Cl‐

)@MOM‐11(Mn2+)

50/50, 11.9b

2.86 (101 kPa) 298 21

CBZ 50/50, 13.2b

50/50, 100b 1.56 (51 kPa) 273 22

Mg‐DOBDC 10/90, 235b 3.84 (10 kPa) 298 23 UTSA‐16 10/90, 58b 1.86 (10 kPa) 298 23 Ni‐MOF‐74 15/85, 30b 3.84 (15 kPa) 298 24,25 HKUST‐1 15/85, 101b 2.64 (15 kPa) 298 26 ZIF‐78 15/85, 30b 0.75 (15 kPa) 298 27,28 mmen‐Cu‐BTTri 15/85, 165b 2.16 (15 kPa) 298 29

a selectivity were obtained from breakthrough exp., b selectivity were calculated from single‐

component isotherms.

In CO2 capture applications, a high selectivity for CO2 over the other components of the gas mixture is

essential [135]. The comparison of the performance of as‐synthesized chabazite with common CO2

adsorbing materials are listed in Table C1. However, the selectivity factors obtained from

multicomponent breakthrough experiments are quite limited. Many of the selectivity numbers

reported are calculated from the pure‐component adsorption isotherms. Furthermore, the CO2

selecitivities over N2 are frequently determined at or below 298 K. It would be more realistic to obtain

ideal adsorbents for post‐combustion carbon capture processes, if the adsorbents can exhibit a high


CO2 over N2 selectivity in the range of 313‐3363 K. As shown in Table C1, the selectivities of CO2 over

N2 and CH4 on our prepared chabazite are dramatically large, which are the highest for CO2 over N2

and CH4 among all the physisorbents to the best of our knowledge. It is worth mentioning that such

high CO2/N2 selectivity was achieved at 348 K, within the temperature range of real flue gases.

Apparently, as‐synthesized chabazite with unparalleled CO2 selectivities, could be used as an ideal

adsorbent for CO2/N2 and CO2/CH4 separation.

In summary, we have discovered a novel approach to produce low silica chabazites from fly ash. The

as prepared chabazite was well‐crystallized exhibiting typical walnut‐like shape morphology with a

uniform size distribution. Adsorption measurements showed dramatically large selectivities of CO2

over N2 and CH4 on our chabazite, suggesting promising performance for the separation of real gas

mixtures such as flue gas and natural gas streams. Additionally, this work demonstrates a new

pathway of recycling fly ash for the manufacture of high value added products.


CS ‐ Supporting Information (SI)

CS1 ‐ Experimental section

The fly ash used in this study was obtained from Shanlu Power Plant, China, and the composition (in

wt %) is as follows: Si (22.6 wt %), Al (21.6 wt %), Fe (1.4 wt %), and Ca (2.4 wt %).

The synthesis procedure for chabazite involved two steps viz. fusion and hydrothermal treatment. In

a typical procedure, 5 g of fly ash was fused with KOH solid in a tube furnace at 923 K for 1 hour, using

the KOH/fly‐ash mass ratios of 2.5.Then, 5 g of the resulting product was dissolved in 20 ml water with

vigorously stirring for 30 min, followed by ageing of 1 h. Finally, the mixture was transferred into an

autoclave and heated at 368 K for 4 days under static conditions. After cooling to the room

temperature, the resultant solid was filtered, washed three times with deionized water, and dried at

373 K overnight.

CS2 ‐ Characterisations

The crystalline properties of the raw materials and the synthesized samples were determined by X‐ray

diffraction (XRD) using a Shimadzu X‐ray diffractometer, with a scanning rate of 2°/min from 4° to 60°.

FE‐SEM (Field Emission Scanning Electron Microscopy) analysis was conducted by employing a ZEISS

scanning electron microscope operated at 15 kV.


Figure CS1. SEM micrograph of fly ash.

280 300 320 340

0.4

0.6

0.8

1.0

1.2

1.4

1.6

CO2

Gauss fitted isobar

CO

2 ads

orpt

ion

capa

city

(m

mol

/g)

Temperature (K)

Figure CS2. Isobar of as‐synthesized r1.9KCHA at 0.5 bar.

CS3 ‐ DFT calculations

All results were calculated using the Vienna Ab initio Simulation Package (VASP) [149] with the

projector augmented waves (PAW) approach [150] . The cut‐off energy of the plane wave basis‐set

was 405 eV. A gamma point only k‐point mesh was used for one unit cell of chabazite (including three

double six‐ring prisms or one and a half supercavities). Such cut‐off energy and k‐point mesh have


been tested to ensure the total energy value convergence within 1 meV/atom. The atomic positions

were optimized with the conjugate gradient method until the forces acting on atoms were below

0.015 eV/Å, as suggested by Göltl and Hafner [143] . To account for the van der Waals interactions,

we also adopted the DFT‐D3 functional (with IVDW=11). We applied the nudged‐elastic‐band (NEB)

method for energy barrier calculations.

To understand the effect of Si/Al ratio (cation density) on the threshold gas admission temperature,

we determined and compared the energy barrier associated with the migration of door‐keeping

potassium cation (K+) in r3KCHA and r1KCHA using density functional theory calculations. Such energy

barrier should rely on the affinity of the K+ to both the starting point and ending point, respectively of

the migration pathway. For the same type of cation site (i.e., 8MR, 6MR, or 4MR site), the affinity can

be different due to the difference in the number of aluminium atoms contained at this site. A cation

site with more aluminum atoms in the ring (thus more negative charge) imparts higher affinity to the

K+ sitting at this site. A zeolite with a Si/Al ratio other than one would have various types of aluminium

distribution in the zeolite framework and thus various affinities to the cation for the same type of

cation site. In our case, r1KCHA has sole aluminum distribution while r3KCHA has varied aluminum

distribution. Thus when we consider K+ migration pathway for opening the “door” (i.e., K+ moves from

8MR site to 4MR site [134]), we need to consider the combinations of different 8MRs and 4MRs in the

case of r3KCHA. Apart from aluminum distribution, the number of K+ adjacent to the ending point of

migration (4MR site) will also affect the energy barrier associated with cation migration process, given

that these adjacent K+ will repel the incoming K+. In the case of r3KCHA, the effect of adjacent K+ is

absent since totally 9 K+ occupies all nine 8MR in one unit cell and there is no excess K+ sitting at 4MR

site, while there exist excess K+ cations to occupy 4MR sites in r1KCHA. Considering the above two

factors – variable aluminium distribution and “adjacent K+ cations” and to simply the discussion, we

only study the cation migration path associated with the lowest possible energy barrier. For r1KCHA,

we consider a K+ migrating from an 8MR site to an adjacent 4MR site with no other adjacent K+ and

the energy barrier is determined to be 1.22 eV (see Fig. 1a). For r3KCHA, we consider a K+ migrating


from an 8MR site (with only one Al in the ring) to an adjacent 4MR site (with 2 Al in the ring) and the

energy barrier is determined to be 0.96 eV (see Fig. 1b). We can conclude that potassium chabazite

with lower Si/Al ratio imparts higher energy barrier for the migration of door‐keeping K+ and thus

higher threshold gas admission temperature. This explains why N2 and CH4 are excluded by r1KCHA

at ambient temperature but admitted by r3KCHA.

Figure CS3. 3D structures of chabazite.

Figure CS4. Structure changes of potassium chabazite systems of r1KCHA and r3KCHA during CO2 adsorption process. The migration of the door‐keeping potassium cation (marked by the white cross)

at different steps was tracked to elucidiate the pore opening procedure in these two types of trapdoor chabazites with different cation density.


CS4 ‐ Multicomponent breakthrough experiments

Binary breakthrough experiments were examined by a dynamic column breakthrough apparatus [96],

which consists of a stainless steel adsorption column (130 mm long and 22.2 mm internal diameter).

Feed gas was flown to the column controlled by four mass flow controllers (MFCs) where a four way

valve controls whether helium or a combination of helium, methane (or CO2) and nitrogen. The

effluent gas flow rate was measured by an orifice type mass flow meter (MFM) which was then

corrected for the gas composition.

Table CS1. The gas selectivities and capacities on chabazite.

Run # Temperature

(oC)

Pressure

(bar)

CO2 Adsorption

(mmol/g)

CH4 Adsorption

(mmol/g)

N2 Adsorption

(mmol/g) Selectivity

#1 30 1 1.0126 n.a. 0.0112 90.41

#2 60 3 0.9654 n.a. 0.0088 109.70

#3 75 1 0.8266 n.a. 0.0012 688.83

#4 30 1 0.9320 0.0016 n.a. 582.50


Appendix D

D1 ‐ Foreword

The following manuscript has been submitted to Communications Chemistry for publication. It is

reproduced here as it is based on the work reported in Chapter 2.

D2 ‐ Temperature Dependence of Adsorption Hysteresis in Flexible Metal Organic

Frameworks

Shamsur Rahman, Arash Arami‐Niya, Xiaoxian Yang, Gongkui Xiao, Gang (Kevin) Li, Eric F. May

“Breathing” and “gating” are striking phenomena exhibited by flexible Metal Organic Frameworks

(MOFs) in which their pore structures transform upon external stimuli. These effects are often

associated with eminent steps and hysteresis in sorption isotherms. Despite significant mechanistic

studies, the accurate description of stepped isotherms and hysteresis remains a barrier to the

promised applications of flexible MOFs in molecular sieving, storage and sensing. Here we

investigate the temperature dependence of structural transformations in three flexible MOFs and

present a new isotherm model to consistently analyse the transition pressures and step widths. The

transition pressure reduces exponentially with decreasing temperature as does the degree of

hysteresis (c.f. capillary condensation). The MOF structural transition enthalpies range from +6 to

+31 kJmol‐1 revealing that the adsorption‐triggered transition is entropically driven. Pressure swing

adsorption process simulations based on flexible MOFs that utilise the model reveal how isotherm

hysteresis can effect separation performance.

The development and study of flexible porous materials, and metal‐organic frameworks (MOFs) in

particular, has been a major focus of the adsorption community over recent years. Interest in these

materials can be attributed to their potential use in a wide range of applications for gas separation,

gas storage, and catalysis [47‐53, 151]. The host‐guest interaction in flexible MOFs affects the host’s

pore characteristics and the associated accessibility by the guest molecule. The interaction may be


classified as either: (1) a crystallographic phase transition causing a unit cell volume change (breathing

effect), or (2) a linker rotation and sub‐net sliding with no change in the unit cell volume (gate‐

opening). In both cases, the ability of these “soft materials” to change structure means their observed

adsorption characteristics are fundamentally different to that of conventional adsorbents [54, 55, 83].

The most obvious difference is the drastic increase in uptake that occurs when the number of

accessible adsorption sites in the framework suddenly changes in response to external stimuli such as

the adsorption/desorption of guest molecules [56, 57, 152]. This feature of flexible MOFs allows the

adsorption of guest molecules seemingly larger than the nominal crystallographic pore diameter to

suddenly increase. Such materials offer, in principle, a very high selectivity and could be very useful in

the development of gas separation technologies [153]. However, studies of MOF‐based applications

are limited by the lack of isotherm models able to accurately describe their abrupt changes in sorption

capacities.

Structural transitions triggered by gas phase pressure are guest‐dependent phenomena that cause the

MOF to change from a non‐porous (np) phase (narrow‐pores in breathing MOFs or closed‐pores in

gating MOFs) to a porous (lp) phase (expanded, large pores in breathing MOFs or open‐pores in gating

MOFs) [56, 57, 69, 154, 155]. This structural change occurs when a critical transition pressure, ptr, is

reached, causing a step‐change in the isotherm characterized by a finite transition‐width, (in Pa).

Moreover, the phase change is hysteretic [77, 156, 157] with the transition pressure measured along

the adsorption branch of the isotherm (increasing pressure), , being greater than that measured

along the desorption branch, , with the difference ptr ( relevant to the degree

of hysteresis typically significantly larger than . In this work, we investigated the dependence of ptr

on temperature for three different MOFs and explored their asymptotic behavior at low temperatures.

Fundamental simulations capable of modelling responsive adsorption processes in flexible porous

materials have been developed [158]. These simulations combine DFT calculations of adsorption in slit

pores with Helmholtz energy descriptions of the crystal as function of pore width, and cannot be


readily fit to experimentally‐measured isotherms for real MOF samples, or used in simulations of

pressure swing adsorption processes based on such materials. Here we utilise a new emprical model

for flexible adsorbents that can efficiently extract salient features of measured hysteretic MOF

sorption isotherms and evaluate their dependence on temperature.

Results

The wide‐ranging data of Mason et al. [48] for methane sorption on Iron(II)‐1,4‐benzenedipyrazolate

(Fe(bdp)) and Cobalt‐1,4‐benzenedipyrazolate (Co(bdp)) provide excellent examples of MOF

hysteretic stepped sorption isotherms. Figure 1 shows the CH4 sorption capacity data along the two

extreme isotherms [48] for each MOF. The number of data acquired along each branch of each

hysteretic sorption isotherm is sufficient to clearly resolve the characteristic parameters ptr, ptr, and

as well as their respective temperature dependences.

Figure D1. Exemplar hysteretic sorption isotherms for CH4 on the MOFs (a) Fe(bdp) and (b) Co(bdp) measured by Mason et al.[48] , illustrating the key parameters in eq (D1).

To extract these parameters from stepped sorption isotherms, we developed an empirical model that

can be regressed to capacity data, q, measured along each branch, the basis and derivation of which

is detailed in the Supplementary Information (SI).

(D1a)

1√

(D1b)

(D1c)


Here, qLangmuir is a Langmuir isotherm function with two adjustable parameters, K and Qm,

characterizing the initial slope and limiting sorption amount, respectively, at pressures below the np

lp phase transition. The function describes the stepped part of the sorption isotherm using

three additional adjustable parameters, ptr, and Qstep, where the latter represents the increase in

capacity after the phase transition (above that of qLangmuir). As detailed in the SI, the LJMY model can

be derived by assuming the sorption site energies made available by the MOF’s structural phase

transition have a Gaussian distribution. It is analogous to the LJM model developed by Li et al. [159]

which describes the temperature‐regulated admission and release of gases in microporous trapdoor

materials. To describe a branch of a hysteretic stepped sorption isotherm for a given MOF, the LJMY

model may be combined with any classical sorption isotherm model (e.g. Sips, Toth) as required to

best describe the data.

Equation (DD1) was regressed to each branch of the sorption isotherms for Fe(bdp) and Co(bdp)

measured by Mason et al. [48] o determine the five parameters, ptr, , Qstep, K and Qm. The best‐fit

LJMY parameter values, their statistical uncertainties and the fit’s standard error are listed for CH4

sorption on Fe(bdp) and Co(bdp) in Tables DS1 and DS2, respectively. Figure D2 shows the

temperature dependence and hysteresis of the best‐fit parameters ptr and for these MOFs. Figures

D2(a) and D2(c) show that the transition pressures measured on the adsorption and desorption

branches have statistically significant differences in their temperature dependence, with

varying more rapidly in both cases than . In constrast, the degree of hysteresis in the transition

width is much smaller, with no statistically significant difference between (ads) and (des) observed for

Fe(bdp) at any temperature. For Co(bdp) a small difference in the temperature dependence of (ads)

and (des) is apparent, although at 323 K the (largest) difference is only about twice the combined

statistical uncertainties of the fit parameters.


Figure D2. Transition pressure ptr (a, c) and transition width, , (b, d) vs temperature, T, for the sorption of CH4 on Fe(bdp) and Co(bdp) extracted from the data of Mason et al [48] with eq (D1). Statistical uncertainties are shown as error bars in all panels but are smaller than the symbol for ptr

in (a) and (c).

Figure D2 shows that the width of the MOF’s structural hysteresis, ptr, decreases with reducing

temperature. This is opposite to the behavior observed for hysteresis associated with capillary

condensation in classical adsorbents with a rigid structure, which becomes larger at lower

temperatures [73, 160].

Simple extrapolations of the apparently linear trends in and with temperature exhibited

in figure D2 suggest that the hysteresis in the sorption isotherms might disappear (e.g. ptr 0)

around 210 K and 230 K for Fe(bdp) and Co(bdp), respectively. Measuring sorption capacities

accurately at such temperatures is very challenging and so an alternative flexible adsorbent was

sought to investigate the asymptotic behaviour of structural hysteresis with temperature. Analysis of

the data published by Arami Niya et al.[59] for CO2 sorption on a zeolitic imidazolate framework (ZIF‐

7) above 273 K suggested that ptr 0 as T 265 K, a significantly more accessible temperature.


Accordingly, ZIF‐7 was synthesized following Arami‐Niya et al. [59]. The adsorption and desorption of

pure CO2 on the ZIF‐7 was measured at eight temperatures between 233 K and 293 K at pressures up

to 0.1 MPa. Figure D3 shows the measured data together with the fits of the LJMY‐Langmuir model

for each isotherm; these are also listed in Table DS3 (fit parameters) and Table DS5 (data). At the

lowest temperatures measured, ZIF‐7 exists in the lp phase even when the CO2 pressure is around 2

kPa. Moreover the hysteresis for CO2 on ZIF‐7 becomes progressively smaller as the temperature is

reduced, consistent with the observations of ptr for Fe(bdp) and Co(bdp). However, ptr does not

become zero at any of the temperatures measured for ZIF‐7.


Figure D3. Equilibrium CO2 capacities, q, measured in adsorption (filled symbols) and desorption (empty symbols) for ZIF‐7 at pressures, p, to 0.1 MPa, and eight temperatures T from (233 to 293) K.

Curves represent fits of the LJMY‐Langmuir model (eq. (DD1) and described in the SI).


At the four lowest temperatures shown in figure D3 for CO2 on ZIF‐7, ptr and are both less than

3 kPa. This makes high resolution measurements of the step in the sorption isotherm difficult and

determination of ptr, , Qstep, K and Qm by regressing each branch separately can be challenging. In

such cases, simultaneously fitting both branches of the hysteretic sorption isotherm as described in

the SI can be helpful.

Figure D4. (a) Variation of transition pressures, ptr, with temperature and (b) relative to the value ptr0 at 303 K value for CO2 on ZIF‐7. (c) Variation of ptr with temperature and (d) with the transition

pressure measured in desorption, .

Figure D4 shows that over a wide temperature range the transition pressures do not vary linearly but

are instead described by the Clausius‐Clapeyron equation.

ln∆

(D2)

Here, is the transition pressure (for either adsorption or desorption), at a reference temperature,

T0; R is the universal gas constant; and Htr is the enthalpy associated with the material’s structural

transition (np lp). Regression of eq (DD2) to the measured for CO2 on ZIF‐7 gives

∆ = (30.1 1.1) kJmol‐1 while fitting to gives ∆ = (31.4 1.6) kJmol‐1. Thus while


there is an offset between the transition pressures measured in adsorption and desorption, (reflected

by different values of for the two branches) their dependence on temperature is the same.

However over the temperature ranges measured by Mason et al. [48], ∆ = (9.1 0.2) kJmol‐1

and ∆ = (8.3 0.1) kJmol‐1 for CH4 on Fe(bdp), while ∆ = (9.1 0.3) kJmol‐1 and

∆ = (5.6 0.5) kJmol‐1 for CH4 on Co(bdp). The SI presents the results of fits to isotherms

measured for CO2 on amino‐MIL‐53 (Al) [60], from which values of ∆ = 15 kJmol‐1 and

∆ = 19 kJmol‐1 can be estimated. In general, the temperature dependence of and

could depend on the MOF, adsorbate and/or the temperature range considered.

The asymptotic temperature dependence of ptr follows from eq (DD2). Whereever ≡ ∆

∆ is finite, the transition pressure hysteresis will vary according to

δ exp 1 (D3)

where the temperature dependence of is given by eq (DD2). When 0 (e.g. CO2 on ZIF‐

7)

δ δ exp ∆

(D4a)

δ 1 (D4b)

Here δ is the difference in the transition pressures on the two branches at the reference

temperature. Equation (D4) explains the exponential trend shown in figure D4(c) and the linear

correlation in figure D4(d). Equations (D3) and (D4) also enable the asymptotic behaviour of the

transition pressure hysteresis observed for Fe(bdp) and Co(bdp) to be predicted reliably, in contrast

to the simplistic linear extrapolations with temperature shown in Figure D2.

Discussion

As discussed by Schneeman et al. [57], the hysteresis observed in a MOF’s stepped isotherm is due to

the energy penalty associated with increasing interfacial area as np lp. No such energy barrier


needs be overcome for the reverse lp np structural change: hence the isotherm’s desorption

branch corresponds to thermodynamic equilibrium, while the adsorption branch reflects the system

accessing metastable states. The positive Htr values indicate that the MOF’s structural transition is

entropically driven with the lp state being less‐ordered than the np state. This is consistent with

previous but less precise estimates by microcalorimetry: for CO2‐driven structural transitions,

Llewellyn et al. [70] estimated +20 kJmol‐1 for MIL‐53 (Cr) and Du et al. [66] estimated +7 kJmol‐1 for

ZIF‐7. The latter is significantly smaller than the value obtained here likely because of the difficulty

separating the exothermic adsorption process from the endothermic structural transition.

Interpreting the non‐zero value of ≡ ∆ ∆ observed for CH4 on Co(bdp) is more

difficult. Potentially, non‐zero Htr values might reflect a combination of effects (e.g. structural

transition plus pore filling/emptying, hysteresis in crystal lattice strain) occurring simultaneously

and/or asymmetrically along the isotherm’s two branches.

The SI shows how the LJMY‐Langmuir model can be used to more reliably simulate PVSA separation

processes utilizing flexible MOFs. To achieve optimal separation performance, the transition pressures

and widths on both branches of the isotherm must be considered appropriately when specifying the

PVSA cycle’s high and low pressures. Additionally, use of only a single branch (desorption or

adsorption) in the simulation will produce an overly optimistic prediction of the separation

performance, relative to that calculated with the fully hysteretic isotherm.

Conclusions

The temperature dependence and hysteresis of stepped sorption isotherms associated with structural

transitions was studied for three flexible MOFs: Fe(bdp) with CH4, Co(bdp) with CH4, and ZIF‐7 with

CO2. A five‐parameter model developed to describe stepped sorption isotherms enabled the transition

pressures, ptr, widths, , and hysteresis ptr ( , to be consistently and robustly

analysed. In contrast to capillary condensation, ptr increases with temperature with a Clausius‐

Clapeyron dependence. At low temperatures, ptr approaches an asymptotic limit proportional to


. The MOFs’ structural phase transitions are entropically driven with enthalpy changes ranging

from (5 to 31) kJmol‐1. Curiously, the structural transition enthalpy for CH4 on Co(bdp) is 60 % larger

in adsorption than desorption, possibly due to multiple effects occurring simultaneously or

asymmetrically along the two isotherm branches.

Methods

Gas adsorption measurements. A volumetric measurement system (model ASAP2020 by

Micromeritics) was used to measure ZIF‐7 sorption capacity for CO2 along multiple isotherms [161].

Before the isotherm measurements, the ZIF‐7 sample [59] was thoroughly dehydrated and degassed

by heating stepwise to 473 K under high vacuum overnight. The degassed sample was cooled to room

temperature and backfilled with helium. Adsorption isotherms for CO2 on ZIF‐7 were measured in the

temperature range of 233–293 K at pressures up to 120 kPa. The adsorption temperatures were

controlled by a homemade heating jacket connected to a thermostatic liquid bath (Ultra‐Low

Refrigerated Circulators JULABO FP88) filled with either silicone oil (for T > 283 K) or ethanol

(T 283 K). The temperature of the bath fluid was measured by a 100 platinum resistance

thermometer.

The equilibration required at each pressure was determined automatically by the ASAP2020 system

with the criterion being a relative pressure change of less than 0.01% over an interval of 15 seconds.

For the adsorption measurements on ZIF‐7, the equilibration time varied from 5 minutes to 15 hours

across all temperatures, with an average value of 92 minutes. For the desorption measurements, the

equilibration time ranged from 5 minutes to about 4 days, with an average value of 2 hours. The

longest equilibration times occurred at low temperatures (233 K and 238 K) and pressures less

than about 1 kPa. The relative combined standard uncertainty of sorption capacity, uc(qi)/qi, measured

with this ASAP2020 was estimated previously [131, 161] to be 1.4 %.


Acknowledgements

This research was funded by the Australian Research Council through IC150100019 and DP190100983.

S.R was supported by an Australian Government Research Training Program Scholarship.

Author Contributions

S.R., A.A., G.L. and E.F.M. conceived and designed the experiments. S.R. and A.A. performed the

synthesis and measured the adsorption isotherms. S.R. A.A. and E.F.M. interpreted the adsorption

data and discussed the findings in this paper. X.Y., G.L. and E.F.M. derived the stepped adsorption

isotherms model. G.X. and S.R. performed the Aspen Adsorption simulations. S.R., A.A., G.L. and E.F.M.

wrote the paper. All authors discussed the results and commented on the manuscript.


DS ‐ Supporting Information (SI)

DS1 ‐ Derivation of the LJMY model for stepped sorption isotherms

The equation for qLJMY is based on a conceptual model, which assumes that the increased capacity

associated with the step is a result of a localized structural phase transition in the adsorbent material

(MOF), triggered by the presence of an adsorbed phase. While the model may be applied to regress

hysteretic stepped isotherms for either so‐called breathing or gating MOFs, the conceptual description

detailed below applies to gating MOFs, where in the absence of an adsorbate the crystal is in a narrow‐

pore (np) state [57]. Before the coverage of the adsorbed phase is sufficient, the interfacial energy

penalty associated with the MOF surface forces the crystal to minimize its area [69]. However, as the

amount of adsorbed phase in a local region of the surface increases with gas pressure, the interfacial

free energy of the MOF in that local region decreases. At some threshold value, the reduction in the

interfacial free energy is sufficient to drive a localized crystalline phase transition of the MOF, with the

resulting structure having a much larger surface area: this may be called the large‐pore (lp) state [69].

The spatial extent of the crystalline transition is limited by the amount of adsorbed phase available to

lower the interfacial energy of the MOF surface. Further increases in gas phase pressure and

adsorption allow the MOF crystal transition to propagate further throughout the material until the

sample reaches its maximum possible surface area.

The conceptual model also assumes that the threshold gas pressure required to induce structural

transition in the MOF varies throughout the sample and is described by a Gaussian distribution with a

mean and standard deviation of ptr and , respectively. One possible explanation for this relates to

the distribution of pore length scales within the np state. These varying length scales influence the gas

pressure, p, required locally for an amount of adsorption sufficient to trigger the structural transition.

Pores in the sample with lower threshold pressures trigger a local np to lp transition earlier along the

adsorption isotherm, and the variation between pores is the reason that the entire sample does not

undergo the transition at a single pressure. This Gaussian distribution would essentially describe the


probability density function, Pr(p), for MOF pores in the np state to have undergone the structural

transition.

√exp

(DS1)

The total adsorbed capacity at a given pressure associated with the MOF structural transition is

determined by integrating this probability density function

1√

(DS2)

where Qstep is the maximum sorption capacity increase resulting from the MOF structural change. To

describe a branch of a hysteretic stepped sorption isotherm for a given MOF, the LJMY model may be

combined with any classical sorption isotherm model (e.g. Sips, Toth) as required to best describe the

data. Here we chose the Langmuir function to minimise the number of adjustable parameters being

regressed: the combined function shown in eq. (D1a) of the main text is therefore the LJMY‐Langmuir

model.

To regress both branches of a hysteretic isotherm simultaneously a dummy independent variable, y,

may be introduced into the fit function. During the regression, the value of y determines along which

branch of the isotherm the data were measured. It can be convenient to set y = 0 as corresponding to

the desorption branch and y = 1 as corresponding to the adsorption branch. This allows quantities that

take on different values for the two branches, (e.g. and ) and must therefore be

represented by two separate parameters, to be expressed in terms of y. For example, to

simultaneously regress eq (D1) to both branches of a stepped sorption isotherm while constraining

four of the parameters (, Qstep, K and Qm) to have the same value on each branch, the capacity

function can be defined

, 1

√

(DS3)


where the additional adjustable parameter, , has been introduced. This allows the transition

pressures along the desorption and adsorption branches to be identified as:

(DS4a)

1 (DS4b)

Fundamental simulations capable of modelling responsive adsorption processes in flexible porous

materials have been developed [158], which allow for a strained framework to exist during the

transition between the np and lp phases as the unit cell parameters change [48, 59]. The parameters

extracted by fitting the LJMY isotherm to experimental data could in principle be related to the

predictions of fundamental models that couple pore width with DFT calculations and can account for

the effect of strain in the crystal. This might provide a pathway to efficiently quantifying the effect of

strain on observed isotherms.

Of the parameters in the LJMY model, the transition width, , is the most likely to be influenced by

the effects of strain, particularly in those cases where there is hysteresis in its value between the

adsorption and desorption branch. For example differences in strain caused by the np to lp transition

from those caused by the lp to np transition might manifest themselves in a non‐zero = (ads) ‐

(des) as observed for Co(bdp).

However, the LJMY model does not make any direct reference to crystal strain. Any connection

between the LJMY parameters extracted by fitting to sorption isotherm data and more fundamental

simulations that do include crystal strain will need to be inferred by analysing the predictions of those

simulations. Additionally, the LJMY model cannot treat stepped isotherms where the observed

sorption capacity does not increase monotonically with pressure, for example as observed for

materials that exhibit negative gas adsorption [83]. The primary purpose of the LJMY model is to serve

as a useful tool for analysing experimental sorption isotherms measured for particular combinations


of guests and flexible adsorbents, and representing those isotherms in simulations of pressure swing

adsorption processes based on those materials.

DS2 ‐ Results of LJMY‐Langmuir model regression to MOF isotherm data

Tables DS1 to DS3 show the best‐fit parameters obtained from regressing the LJMY‐Langmuir model

to the MOF isotherm data. For each isotherm, three sets of parameters are listed: a fit to each branch

separately and the results of fitting both branches simultaneously with eq (DS3).

Fitting both branches simultaneously with eq (DS3) constrains , Qstep, Qm and K to have the same

value in adsorption and desorption, preventing any investigation of hysteresis in these quantities.

Particularly for the Fe(bdp) and Co(bdp) data, different values of Qstep on the two isotherm branches

are needed to accurately represent the measured data across all conditions, especially when the

hysteresis in transition pressure becomes at higher temperatures. Appreciably worse fits are achieved

in these cases if the branches are fit simulatenously.

A possible physical explanation for the hysteresis in Qstep suggested by fitting the branches separately

is that at higher temperatures the properties of the adsorbed phase do not lower the MOF’s interfacial

free energy as effectively and hence significantly more adsorption is necessary to induce the np to lp

structural transition.

However, care should be taken not to over‐interpret the best‐fit value of Qstep obtained from a fit to a

single isotherm branch, because it is correlated with Qm and Kp. In contrast, p(tr) and , which are of

primary interest in this work, are robustly orthogonal to the other parameters in the model; their

values and uncertainties do not change significantly if the branches are fit simultaneously, although

their statistical uncertainties inherently increase if the fit quality deteriorates. Accordingly to more

accurately investigate the dependence of p(tr) and on temperature and isotherm branch, the values

obtained by fitting the independent branches were used in the main text.


Table DS1. Values and statistical standard uncertainties, u, of best fit parameters determined by regression of the LJMY‐Langmuir model (eq (D1) or eq (DS3)) to the data of Mason et al. [48] for sorption of CH4 on Fe(bdp), together with the standard error of the fit.

T / K ptr / MPau(ptr) / MPa

s / MPa u(s) / MPa

Qstep / mol×kg‐1

u(Qstep) / mol×kg‐1

Qm / mol×kg‐1

u(Qm) / mol×kg‐1

K / MPa‐1 u(K) / MPa‐1

Standard error / mol×kg‐1

Adsorption

248 1.257 0.006 0.05 0.012 9.62 0.54 32.6 14 0.052 0.033 0.54

261 1.615 0.008 0.066 0.014 8.82 0.45 359.4 3100 0.003 0.028 0.56

273 1.953 0.012 0.091 0.018 8.24 0.36 986.9 30000 0.001 0.030 0.55

285 2.337 0.014 0.116 0.021 7.98 0.37 1100.0 34000 0.001 0.026 0.57

298 2.715 0.017 0.16 0.026 8.1 0.4 1014.4 53000 0.001 0.038 0.59

311 3.17 0.025 0.261 0.039 8.44 0.48 1559.4 150000 0.000 0.038 0.62

323 3.493 0.018 0.275 0.028 8.01 0.32 850.7 88000 0.001 0.055 0.4

Desorption

248 0.738 0.008 0.061 0.013 7.15 0.52 25.5 5.3 0.119 0.045 0.5

261 0.904 0.008 0.072 0.014 6.47 0.5 30.7 10 0.077 0.040 0.42

273 1.074 0.015 0.087 0.029 5.56 0.66 36.1 21 0.059 0.049 0.45

285 1.258 0.017 0.13 0.029 5.83 0.64 177.1 825 0.008 0.040 0.37

298 1.499 0.024 0.143 0.04 4.87 0.71 150.2 704 0.009 0.046 0.31

311 1.671 0.027 0.183 0.049 4.6 0.52 1866.9 380000 0.0005 0.093 0.29

323 1.865 0.064 0.23 0.1 3.9 1.2 2060.4 4200000 0.0002 0.47 0.35

Adsorption and desorption branched fitted simultaneously

248 0.742 0.008 0.063 0.012 8.21 0.43 28.2 0.083 0.70 0.02 0.58

261 0.907 0.009 0.08 0.013 7.94 0.4 84.9 0.017 0.79 0.02 0.58

273 1.083 0.018 0.115 0.022 7.91 0.48 889.4 0.0012 0.81 0.03 0.63

285 1.278 0.02 0.143 0.023 7.72 0.42 1094.4 0.0009 0.83 0.03 0.6

298 1.508 0.034 0.182 0.037 7.43 0.65 1096.9 0.0007 0.80 0.04 0.67

311 1.726 0.042 0.231 0.039 7.12 0.43 1369.4 0.0005 0.83 0.05 0.65

323 1.944 0.052 0.266 0.049 6.91 0.55 978.8 0.0005 0.80 0.05 0.61


Table DS2. Values and statistical standard uncertainties, u, of best fit parameters determined by regression of the LJMY‐Langmuir model (eq (D1) or eq (DS3)) to the data of Mason et al. [48] for sorption of CH4 on Co(bdp), together with the standard error of the fit.

T / K ptr / MPa u(ptr) / MPa

s / MPa u(s) / MPa

Qstep / mol×kg‐1

u(Qstep) / mol×kg‐1

Qm / mol×kg‐1

u(Qm) / mol×kg‐1



Adsorption

273 1.317 0.009 0.095 0.016 6.73 0.31 284.3 1900 0.0034 0.023 0.33

285 1.591 0.014 0.148 0.02 6.17 0.23 552.5 8900 0.0015 0.025 0.35

298 1.906 0.021 0.192 0.033 5.58 0.33 916.3 37000 0.0008 0.033 0.46

311 2.185 0.028 0.258 0.048 5.34 0.42 955.7 49000 0.0007 0.036 0.48

323 2.443 0.031 0.336 0.058 5.64 0.52 995.7 65000 0.0006 0.038 0.46

Desorption

273 0.641 0.011 0.066 0.02 3.98 0.54 19.2 4.4 0.1374 0.065 0.4

285 0.745 0.018 0.087 0.033 3.33 0.62 19.7 5.5 0.1231 0.07 0.38

298 0.833 0.022 0.109 0.043 2.8 0.65 20.4 6.2 0.1101 0.065 0.34

311 0.894 0.031 0.102 0.061 1.77 0.63 23.4 9.6 0.0844 0.056 0.28

323 0.945 0.035 0.134 0.071 1.52 0.53 25.4 11.1 0.0700 0.049 0.22

Adsorption and desorption branched fitted simultaneously

273 0.662 0.011 0.113 0.018 5.87 0.35 44.8 0.029 1.00 0.04 0.44

285 0.78 0.017 0.163 0.025 5.51 0.35 401.3 0.0026 1.04 0.05 0.44

298 0.901 0.024 0.232 0.035 5.58 0.42 718.8 0.0011 1.12 0.06 0.48

311 1.024 0.038 0.318 0.055 5.24 0.5 993.2 0.0007 1.13 0.09 0.51

323 1.129 0.055 0.4 0.073 5.02 0.57 1018.2 0.0006 1.13 0.11 0.53


Table DS3. Values and statistical standard uncertainties, u, of best fit parameters determined by regression of the LJMY‐Langmuir model (eq (D1) or eq(DS3)) to the data

measured in this work and by Arami‐Niya et al. [59] at 303 K for the sorption of CO2 on ZIF‐7, together with the standard error of the fit.

T / K ptr / MPa u(ptr) / MPa

s / MPa u(s) / MPa Qstep /

mol×kg‐1 u(Qstep) / mol×kg‐1

Qm / mol×kg‐1

u(Qm) / mol×kg‐1



Adsorption

303 0.07925 0.00041 0.01303 0.00046 1.21 0.03 1.21 0.03 6.50 0.28 0.01

293 0.05183 0.00031 0.0096 0.00061 1.06 0.06 1.06 0.06 33.29 11 0.02

283 0.03128 0.00035 0.00547 0.00054 0.88 0.04 0.88 0.04 19.93 3.2 0.03

273 0.02305 0.00007 0.00395 0.00011 0.98 0.01 0.98 0.01 42.77 1.5 0.01

253 0.00767 0.00004 0.00354 0.00007 1.56 0.02 1.56 0.02 25.79 2.4 0.0005

247 0.00695 0.00001 0.0029 0.00002 1.51 0.01 1.51 0.01 37.11 2.3 0.001

244 0.004 0.0011 0.00069 0.00104 0.93 0.36 0.93 0.36 132.22 71 0.01

238 0.00315 0.00003 0.00053 0.0001 1.29 0.05 1.29 0.05 157.78 12 0.01

233 0.001883 7E‐07 0.00001 0 1.31 0.03 1.31 0.03 272.33 26 0.01

Desorption

303 0.04971 0.0019 0.00789 0.002 1.2 0.03 1.20 0.03 7.49 0.49 0.01

293 0.03067 0.00016 0.00455 0.00023 0.87 0.02 0.87 0.02 21.43 1.7 0.01

283 0.01893 0.00007 0.00203 0.00012 0.88 0.02 0.88 0.02 36.03 2 0.01

273 0.014746 0.000003 0.000083 0.000002 0.87 0.02 0.87 0.02 69.22 3 0.01

253 0.005193 0.000001 0.000035 0.000001 0.92 0.02 0.92 0.02 192.53 5.4 0.01

247 0.0043 0.00006 0.00019 0.00009 1.36 0.08 1.36 0.08 76.37 22 0.02

244 0.00219 0.00001 0.00101 0.00002 1.29 0.01 1.29 0.01 160.49 5.8 0.001

238 0.00129 0.00012 0.00074 0.00017 1.53 0.32 1.53 0.32 95.84 280 0.09

233 0.00115 0.00015 0.0007 0.00021 1.66 0.33 1.66 0.33 22.59 400 0.14

Adsorption and Desorption branched fitted simultaneously

293 0.03031 0.00045 0.00712 0 0.98 0.03 0.84 24 0.72 0.03 0.03

283 0.0187 0.00029 0.00311 0 0.9 0.03 0.93 29 0.67 0.03 0.04

273 0.01293 0.00026 0.00274 0 0.97 0.03 0.86 56 0.82 0.04 0.03

253 0.00421 0.00007 0.00105 0 1.02 0.02 0.91 160 1.92 0.08 0.01

247 0.00353 0.00005 0.00001 0 1.03 0.01 0.91 190 0.79 0.03 0.02

244 0.00218 0.00002 0.00104 0 1.2 0.01 0.73 210 1.01 0.02 0.01

238 0.00128 0.00008 0.00071 0 1.43 0.1 0.52 179 1.51 0.16 0.05

233 0.00117 0.00013 0.00066 0 1.41 0.17 0.50 410 1.41 0.36 0.07


Table DS4. Values and statistical standard uncertainties, u, of best fit parameters determined by regression of the LJMY‐Langmuir model (eq (D1) or eq(DS3)) to the data of

Couck et al. [162] for sorption of CO2 on CH4 on MIL‐53(Al), together with the standard error of the fit.

T / K ptr / MPa u(ptr) / MPa / MPa u() / MPa Qstep /

molkg‐1 u(Qstep) /

molkg‐1 Qm /

molkg‐1 K / MPa‐1

Standard error

/ molkg‐1

CO2 – Adsorption

303 1.506 0.017 0.317 0.023 4.35 0.12 2.17 11.25 0.13

288 1.101 0.016 0.232 0.025 4.15 0.12 2.09 18.04 0.17

CO2 – Desorption

303 0.606 0.049 0.312 0.054 4.86 0.52 1.08 31.59 0.21

288 0.406 0.015 0.08 0.02 3.88 0.31 2.19 28.40 0.24

CH4 – Adsorption

303 1.028 0.01 0.149 0.015 0.71 0.03 5.37 0.17 0.02

CH4 – Desorption

303 0.895 0.076 0.292 0.051 1.28 0.21 6.71 0.08 0.04

Regression of eq (D2) to the values of ptr measured for CO2 on MIL‐53(Al) at 303 K and 288 K gives ∆ = 15 kJmol‐1 and ∆ = 19 kJmol‐ 1


Figure DS1. Hysteretic sorption isotherms for CH4 on the Fe(bdp) measured by Mason et al. [48] Symbols: experimental data; curves: calculated with LJMY‐Langmuir model. (left) Adsorption or

desorption branch at each temperature was fitted separately. (right) The adsorption and desorption branches at each temperature were fitted simultaneously.


Figure DS2. Hysteretic sorption isotherms for CH4 on the Co(bdp) measured by Mason et al. [48] Symbols: experimental data; curves: calculated with LJMY‐Langmuir model. (left)Adsorption or

desorption branch at each temperature was fitted separately. (right) The adsorption and desorption branches at each temperature were fitted simultaneously.


Figure DS3. Equilibrium CO2 capacities, q, measured in adsorption (filled symbols) and desorption (empty symbols) for ZIF‐7. Curves represent fits of the LJMY‐Langmuir model with adsorption and

desorption branched at each temperature fitted simultaneously.


DS3 ‐ Measurements of CO2 sorption on ZIF‐7

The ZIF‐7 was synthesized following the method detailed by Arami Niya et al. [163]. A Micromeritics

ASAP2020 instrument was used to measure the equilibrium capacities tabulated below:

Table DS5. Equilibrium capacities, q, of CO2 on ZIF‐7 measured in adsorption and desorption as a function of pressure, p, and temperature, T. The relative combined standard uncertainty of sorption capacity, uc(qi)/qi, measured with this ASAP2020 was estimated previously [131, 161] to be 1.4 %

Adsorption Desorption

T / K p / kPa q / mol.kg‐1 T / K p / kPa q / mol.kg‐1

233 0.094 0.020 233 19.210 1.868

233 0.463 0.084 233 9.539 1.808

233 1.134 0.147 233 5.116 1.735

233 1.867 0.333 233 3.872 1.696

233 3.965 1.646 233 3.119 1.663

233 26.087 1.879 233 1.912 1.573

233 1.038 0.591

233 0.491 0.445

238 0.096 0.017 238 39.755 1.893

238 0.479 0.064 238 28.915 1.869

238 1.028 0.098 238 19.337 1.835

238 1.997 0.178 238 9.868 1.767

238 2.998 0.729 238 5.175 1.679

238 4.300 1.551 238 3.994 1.637

238 16.122 1.812 238 3.102 1.592

238 20.958 1.838 238 2.082 1.463

238 30.621 1.872 238 1.017 0.457

238 39.755 1.893 238 0.511 0.390

244 0.558 0.055 244 40.803 1.859

244 1.232 0.163 244 38.209 1.853

244 3.319 0.483 244 35.721 1.847

244 5.304 1.356 244 34.969 1.845


244 39.251 1.853 244 32.530 1.839

244 42.612 1.862 244 29.975 1.831

244 44.745 1.867 244 27.494 1.822

244 44.951 1.868 244 25.008 1.812

244 22.490 1.800

244 20.024 1.787

244 17.508 1.771

244 15.046 1.752

244 12.100 1.722

244 9.142 1.679

244 6.890 1.632

244 4.797 1.564

244 3.918 1.484

244 2.791 1.134

247 4.974 0.454 247 100.039 1.913

247 10.819 1.519 247 93.244 1.908

247 24.376 1.750 247 88.228 1.904

247 27.335 1.766 247 83.229 1.899

247 30.383 1.780 247 80.034 1.896

247 31.882 1.787 247 75.074 1.890

247 35.129 1.800 247 70.081 1.884

247 38.003 1.809 247 65.106 1.878

247 40.932 1.818 247 59.964 1.871

247 44.045 1.827 247 55.965 1.865

247 46.949 1.834 247 52.933 1.860

247 50.009 1.841 247 50.102 1.855

247 53.055 1.847 247 47.128 1.849

247 56.070 1.853 247 43.960 1.842

247 60.043 1.861 247 40.955 1.835

247 65.084 1.869 247 37.966 1.828


247 69.974 1.876 247 34.983 1.819

247 75.050 1.883 247 31.953 1.810

247 80.019 1.890 247 29.014 1.800

247 85.041 1.896 247 26.007 1.787

247 89.999 1.902 247 23.014 1.773

247 95.047 1.908 247 20.021 1.756

247 100.039 1.913 247 17.062 1.735

247 14.035 1.709

247 11.091 1.674

247 8.179 1.625

247 4.770 1.524

253 4.815 0.378 253 100.030 1.895

253 14.433 1.639 253 93.259 1.889

253 44.292 1.805 253 88.280 1.884

253 48.597 1.817 253 83.272 1.879

253 51.649 1.824 253 80.031 1.876

253 53.014 1.828 253 75.071 1.870

253 56.135 1.834 253 69.952 1.863

253 60.061 1.841 253 65.050 1.856

253 65.001 1.850 253 60.116 1.848

253 70.046 1.858 253 55.948 1.841

253 75.052 1.865 253 52.936 1.835

253 80.067 1.872 253 49.967 1.829

253 85.052 1.878 253 46.957 1.822

253 90.027 1.884 253 43.974 1.815

253 95.052 1.889 253 40.949 1.807

253 100.030 1.895 253 37.978 1.798

253 34.949 1.789

253 32.008 1.778

253 28.987 1.765


253 26.027 1.751

253 23.014 1.734

253 20.008 1.714

253 17.054 1.690

253 14.065 1.658

253 11.106 1.618

253 8.206 1.560

253 5.227 1.273

253 2.576 0.323

253 1.278 0.222

273 2.617 0.067 273 100.014 1.703

273 7.219 0.217 273 93.423 1.695

273 13.795 0.345 273 88.327 1.688

273 15.425 0.383 273 83.368 1.680

273 19.720 0.604 273 79.937 1.674

273 24.830 1.121 273 75.120 1.665

273 43.801 1.561 273 70.046 1.655

273 46.423 1.574 273 65.139 1.643

273 51.191 1.593 273 60.011 1.630

273 54.956 1.606 273 55.162 1.617

273 60.003 1.622 273 49.984 1.601

273 65.044 1.636 273 44.970 1.583

273 70.072 1.648 273 39.990 1.562

273 75.042 1.659 273 35.004 1.537

273 80.049 1.669 273 30.047 1.508

273 85.036 1.679 273 25.073 1.472

273 90.052 1.687 273 20.207 1.423

273 95.026 1.696 273 14.809 1.153

273 100.014 1.703 273 9.719 0.363

273 3.995 0.208


273 2.581 0.163

273 0.919 0.105

283 4.887 0.119 283 99.522 1.595

283 8.593 0.173 283 79.641 1.559

283 11.262 0.206 283 60.039 1.508

283 13.821 0.239 283 54.698 1.491

283 16.983 0.282 283 52.825 1.484

283 19.874 0.328 283 50.112 1.474

283 22.848 0.386 283 46.964 1.461

283 25.602 0.470 283 43.971 1.448

283 29.106 0.699 283 40.970 1.434

283 32.381 0.999 283 37.976 1.418

283 38.553 1.243 283 34.985 1.401

283 40.825 1.296 283 32.013 1.381

283 44.605 1.373 283 29.029 1.358

283 46.934 1.410 283 26.131 1.326

283 50.024 1.447 283 22.671 1.228

283 53.027 1.471 283 19.514 0.940

283 55.990 1.488 283 17.398 0.535

283 59.988 1.504 283 14.576 0.351

283 80.031 1.558 283 11.034 0.258

283 99.522 1.595 283 7.865 0.195

283 4.868 0.137

293 5.084 0.113 293 99.399 1.582

293 7.853 0.145 293 79.814 1.540

293 11.803 0.182 293 60.472 1.480

293 13.776 0.199 293 54.912 1.456

293 16.960 0.225 293 52.791 1.446

293 19.961 0.249 293 50.005 1.430

293 22.931 0.274 293 47.033 1.409


293 25.957 0.300 293 44.076 1.379

293 28.942 0.329 293 41.068 1.337

293 31.968 0.361 293 38.154 1.279

293 34.945 0.396 293 35.325 1.189

293 37.904 0.438 293 32.593 1.039

293 40.890 0.493 293 28.611 0.675

293 43.795 0.568 293 23.254 0.387

293 46.576 0.689 293 21.973 0.364

293 49.771 0.876 293 19.693 0.324

293 54.347 1.084 293 16.907 0.279

293 56.536 1.151 293 13.998 0.238

293 59.389 1.223 293 11.034 0.201

293 82.688 1.519 293 8.116 0.166

293 99.399 1.582 293 5.154 0.126

DS4 ‐ Temperature dependence and hysteresis of structural transition parameters

In the absence of structural information about a given adsorbent, observations regarding the

temperature dependence of any hysteresis in the sorption isotherm could help identify the nature of

the material. For example, the degree of isotherm hysteresis increases with decreasing temperature

when the hysteresis is caused by capillary condensation in the mesopores of a rigid material. This is in

contrast to the increase in isotherm hysteresis with increasing temperature associated flexible

adsorbent structural transformations. Potentially, investigating the temperature dependence and

resulting hysteresis of the isotherm’s transition pressures and/or transition widths could provide

additional insight into the nature of the material’s structural transformation.

For example, from eq (D2) if the ratio changes more with increasing temperature than

, then ∆ is larger than ∆ . For CO2 on ZIF‐7 over the temperature range

considered in this work, these enthalpy changes for each branch are statistically equivalent. For CH4

on Co(bdp), ∆ are 60 % larger than ∆ , while for CO2 on MIL‐53(Al), ∆ is 22 % smaller


than ∆ . Further research is needed to investigate how these differences in ≡ ∆

∆ correspond to the specific nature of the structural transformation.

Such information may also be encoded within the transition width’s temperature dependence and

hysteresis. Figure DS4 shows the variation in this parameter for both branches of the sorption

isotherms measured for CH4 on Fe(bdp), CH4 on Co(bdp) and CO2 on ZIF‐7. No hysteresis in the

transition width is resolvable for CH4 on Fe(bdp) while for Co(bdp) the difference in between the

two branches is clear and increases with temperature. For CO2 on ZIF‐7 below 247 K the values of

determined for the adsorption and desorption branches are essentially statistically equivalent: at

higher temperatures a sustained difference in the range (3 to 5) kPa is observed. Collectively, the

results from these three MOFs suggest there might be a temperature below which is the same for

both branches of the isotherm, while at higher temperatures a hysteresis in this parameter becomes

increasingly manifest. As suggested in Section DS1, differences in the value of obtained for each

branch might reflect differences in the crystal strain caused by the direction of the np‐lp transition.

Alternatively (or in addition) the magnitude and temperature dependence of might contain

information about the sample’s heterogeneity. Clearly, however, such speculation needs to be

investigated further by measuring hysteretic sorption isotherms over a wider range of T, with

additional MOF‐adsorbate combinations, and through fundamental simulations with methods such as

those of Evans et al. [158].


Figure DS4. Temperature dependence and hysteresis of the transition width parameter, , reported in Tables DS1 to DS3 when each branch is fit separately to eq (D1).

0.000.050.100.150.200.250.300.350.400.45

240 260 280 300 320 340

/ MPa

T / K

(a) CH4 on Fe(bdp)

ads

des

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

240 260 280 300 320 340

/ MPa

T / K

(b) CH4 on Co(bdp)

ads

des

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

230 250 270 290 310

/ MPa

T / K

(c) CO2 on ZIF‐7

ads

des


DS5 ‐ Pressure Vacuum Swing Adsorption simulation of a CH4/CO2 mixture with ZIF‐7

In principle, MOFs have significant potential for use in gas separation processes as a result of their

characteristic stepped isotherms, which could deliver benefits in terms of both selectivity and working

capacity. However, studies of process applications involving MOFs remain comparatively limited in

part because of the difficulty associated with the description of the stepped and hysteretic sorption

isotherms within the frameworks provided by process simulators. Here we demonstrate the use of

the LJMY‐Langmuir isotherm model within a simulation of a Pressure Vacuum Swing Adsorption

(PVSA) process separating an equimolar mixture of methane + carbon dioxide using ZIF‐7. The study

reveals the importance of rigorously considering the sorption isotherms, and particularly their

hysteresis, in both the design and simulation of the separation process.

A two‐bed, four‐step PVSA process was simulated numerically using the software package Aspen

Adsorption. A graphical layout of the simulation is shown in Figure DS5, and the step sequence is

listed in Table DS6.

Figure DS5. Graphical layout of the PVSA model built with Aspen Adsorption.


Table DS6. PVSA cycle step sequence and timing

Step 1 Step 2 Step 3 Step 4

Time (s) 60 60 60 60

Bed 1 Adsorption Blowdown Evacuate Repressurisation

Bed 2 Evacuate Repressurisation Adsorption Blowdown

The column dimensions, adsorbent properties, and process conditions are listed in Table DS7, while

the isotherm parameters for CO2 and CH4 on ZIF‐7 at 303 K are listed in Table DS8. The values of the

ZIF‐7 isotherm parameters were obtained by fits of eq (D1) to the data measured by Arami‐Niya et al.

[163] for CO2 and to the data reported recently by Yang et al. [164] for CH4. The isotherm data and

corresponding fits of the LJMY‐Langmuir models are shown in Figure DS6.

Table DS7. Simulation parameters

Parameter Value

Nodes per column 20

Height of column (m) 0.3

Diameter of column (m) 0.03

Inter‐particle void fraction 0.348

Intra‐particle void fraction 0.146

Bulk solid density (kg/m3) 385.0

Particle radius (m) 0.0007

Mass Transfer Coefficient – CH4 (s‐1) 1

Mass Transfer Coefficient – CO2 (s‐1) 1

Column and gas temperature (K) 303.15

phigh in column (bar) 2, 1.5

plow in column (bar) 1.0, 0.8, or 0.6

Feed flowrate (mol/s) 3.4 × 10‐3

Feed gas mole fraction composition 0.5 CH4 + 0.5CO2

Table DS8. Values of the LJMY‐Langmuir sorption isotherm parameters for ZIF‐7 at 303 K used to simulate a PVSA separation of an equimolar CO2 + CH4

mixture, based on the data reported by Aram‐Niya et al. [163] and by Yang et al. [164].

Parameter CO2 CH4

Adsorption Desorption Adsorption Desorption

ptr / bar 0.793 0.497 6.99 4.76

/ bar 0.130 0.079 0.73 0.61

Qstep / molkg‐1 1.20 1.20 1.00 1.00

K / molkg‐1bar‐1 0.65 0.75 0.20 0.32

Qm / molkg‐1 0.89 0.89 0.88 0.79


Figure DS6. Isotherm data and LJMY‐Langmuir fits for CO2 and CH4 on ZIF‐7 at 303 K.

A user‐defined isotherm model was implemented in Aspen Adsorption for both beds. The simulation

could be run using either (i) only the adsorption branch, (ii) only the desorption branch, or (iii) both

branches as part of a fully hysteretic model. In this third case, for a given step of the PVSA cycle the

appropriate branch of the isotherm was selected via a programming flag: during the “Adsorption” and

“Repressurisation” steps, the adsorption branch of the isotherm was selected, while for the

“Blowdown” and “Evacuate” steps, the desorption branch was selected. Sorption was assumed to be

non‐competitive: the equilibrium capacity of each component at a given node within the bed was

evaluated from the partial pressure of that component in the gas phase at that node. In principle,

competitive adsorption effects could be simulated using, for example, the Osmotic Framework

Adsorbed Solution Theory proposed by Coudert et al. [62].

To ensure complete utilisation of the step in the CO2 isotherm by the PVSA process, the CO2 partial

pressure needs to swing from to . The selectivity

inferred from the pure fluid isotherms reaches a maximum of about 7 at a (partial) pressure of

; at higher pressures the selectivity decreases as the CH4 uptake increases faster than

the CO2 uptake. To investigate the impact of varying the pressure swing limits in the PVSA cycle on the

separation performance, simulations were conducted for a high pressure of (2 and 1.5) bar, with


desorption pressures of either (1, 0.8, 0.6 or 0.4) bar; for an equimolar mixture, these values span the

partial pressure limits identified above. The methane mole fraction of the light product stream (P1 in

Figure DS5) was the metric used to assess changes in separation performance. These pressure swing

limits and the resulting separation performance were also analysed in the context of the model used

to represent the CO2 sorption isotherm: either fully hysteretic, just the adsorption branch or just the

desorption branch. Figure DS7 shows the light product methane mole fraction produced by a PVSA

cycle using ZIF‐7 with phigh = 2 bar at the four values of plow considered, when the simulation used each

of the three isotherm models.

Figure DS7. Simulated methane content of light product (P1 stream in Figure DS5) produced from an

equimolar feed of CH4 + CO2 using ZIF‐7 in a 2‐bed, 4‐step PVSA process with phigh = 2 bar. For every

desorption pressure (plow) considered the result of using each of the three isotherm models

(adsorption branch, desorption branch or full hysteresis) is shown.

There are three key features of the results shown in Figure DS7 that inform the design and simulation

of PVSA cycles using a flexible MOF like ZIF‐7 with a hysteretic stepped isotherm.

1. The separation performance predicted using the fully hysteretic isotherm model is typically

worse than the performance predicted using either branch alone.

0.0

0.2

0.4

0.6

0.8

1.0

1 0.8 0.6 0.4

Lig

ht

prod

uct

CH

4m

ole

frac

tion

Desorption pressure (bar)

Ads Hyst Des


2. Single‐branch isotherm models will predict significantly better separation performance than

the fully hysteretic model if the pressure swing limits phigh or plow are not both clear of the

partial pressure limits or , respectively.

3. If the pressure swing limits are consistent with the partial pressure limits required for a good

separation, the performance predicted using the branch with the smallest transition width, ,

will be the more (overly) optimistic.

There are two primary reasons for these features. The most important is the need for the pressure

swing limits to be beyond the partial pressure limits set by the locations and widths of the steps on

the two branches. The parameters in the LJMY isotherm model allow these to be readily identified, so

that phigh and plow can be chosen to take advantage of the full isotherm step. For example, if only the

adsorption branch is considered in the simulation and plow is chosen so that the partial pressure is

below but not below then an overly optimistic separation performance will be

predicted. This explains the high adsorption branch result observed for plow = (1 and 0.8) bar in Figure

DS7. The converse also applies: if phigh is chosen so that the partial pressure is above but not

above , then simulations based only on the desorption branch isotherm will predict

significantly better separation performance than those based on either the fully hysteretic model or

the adsorption branch model. The simulations conducted with phigh = 1.5 bar in this work produce

results with a similar pattern to that shown in Figure DS7, except that light product methane fractions

are lower for all cases, reflecting the reduction in selectivity during the adsorption step. However,

when plow was (0.6 or 0.4) bar, the desorption branch model predicted CH4 fractions around 0.8, similar

to those shown in Figure DS7. In contrast, the fully hysteretic and adsorption branch models predicted

CH4 fractions around 0.67 because phigh was not sufficient to clear the upper partial pressure limit of

.

The second reason for the three features identified is that sharper isotherm steps produce better

separation performance. During the adsorption stage of the PVSA cycle, the partial pressure of the


adsorbate varies spatially across the bed according to the gas‐phase composition profile. Narrower

steps in the adsorption isotherm will act to sharpen the gas composition front in the bed and improve

the separation performance. For ZIF‐7, so models based on the desorption branch will

always predict better performance than those based on the adsorption branch, provided the pressure

swing limits are appropriate. Imposing two different widths in the fully hysteretic model typically

results in a separation performance that is slightly worse than that predicted using the adsorption

branch model.


Appendix E

E1 – Data table for experimental data presented in Figure 2.1 and Figure 2.13

Table E1. Data points extracted graphically from Couck et al. [60]

Adsorbent Gas Temperature (K) Pressure (kPa) Quantity Adsorbed (mol/kg)

amino‐MIL‐53 (Al) CO2 288 16.30 0.075

amino‐MIL‐53 (Al) CO2 288 25.29 0.620

amino‐MIL‐53 (Al) CO2 288 29.76 0.794

amino‐MIL‐53 (Al) CO2 288 34.24 0.983

amino‐MIL‐53 (Al) CO2 288 56.49 1.171

amino‐MIL‐53 (Al) CO2 288 93.54 1.383

amino‐MIL‐53 (Al) CO2 288 143.91 1.503

amino‐MIL‐53 (Al) CO2 288 188.35 1.586

amino‐MIL‐53 (Al) CO2 288 228.34 1.631

amino‐MIL‐53 (Al) CO2 288 271.29 1.691

amino‐MIL‐53 (Al) CO2 288 336.46 1.774

amino‐MIL‐53 (Al) CO2 288 360.16 1.804

amino‐MIL‐53 (Al) CO2 288 437.17 1.901

amino‐MIL‐53 (Al) CO2 288 549.73 1.983

amino‐MIL‐53 (Al) CO2 288 672.65 2.088

amino‐MIL‐53 (Al) CO2 288 708.19 2.118

amino‐MIL‐53 (Al) CO2 288 802.99 2.268

amino‐MIL‐53 (Al) CO2 288 921.52 2.645

amino‐MIL‐53 (Al) CO2 288 1007.55 3.423

amino‐MIL‐53 (Al) CO2 288 1081.72 4.178

amino‐MIL‐53 (Al) CO2 288 1250.68 5.008

amino‐MIL‐53 (Al) CO2 288 1536.61 5.761

amino‐MIL‐53 (Al) CO2 288 1831.30 5.917

amino‐MIL‐53 (Al) CO2 288 2109.70 5.982

amino‐MIL‐53 (Al) CO2 288 2386.62 6.131

amino‐MIL‐53 (Al) CO2 288 2824.98 6.421

amino‐MIL‐53 (Al) CO2 288 2646.26 6.397

amino‐MIL‐53 (Al) CO2 288 2455.87 6.350

amino‐MIL‐53 (Al) CO2 288 2220.95 6.310

amino‐MIL‐53 (Al) CO2 288 1982.96 6.263

amino‐MIL‐53 (Al) CO2 288 1731.14 6.193

amino‐MIL‐53 (Al) CO2 288 1474.71 6.106

amino‐MIL‐53 (Al) CO2 288 1216.72 5.949

amino‐MIL‐53 (Al) CO2 288 958.68 5.674

amino‐MIL‐53 (Al) CO2 288 702.19 5.430

amino‐MIL‐53 (Al) CO2 288 448.58 4.699


amino‐MIL‐53 (Al) CO2 288 314.13 2.405

amino‐MIL‐53 (Al) CO2 288 186.46 1.784

amino‐MIL‐53 (Al) CO2 288 106.53 1.556

amino‐MIL‐53 (Al) CO2 288 61.94 1.367

amino‐MIL‐53 (Al) CO2 288 37.30 1.187

amino‐MIL‐53 (Al) CO2 288 18.80 0.990

amino‐MIL‐53 (Al) CO2 288 15.66 0.794

amino‐MIL‐53 (Al) CO2 288 14.06 0.629

amino‐MIL‐53 (Al) CO2 288 6.15 0.031

amino‐MIL‐53 (Al) CO2 303 14.86 0.257

amino‐MIL‐53 (Al) CO2 303 17.86 0.469

amino‐MIL‐53 (Al) CO2 303 28.26 0.680

amino‐MIL‐53 (Al) CO2 303 50.51 0.892

amino‐MIL‐53 (Al) CO2 303 90.52 1.050

amino‐MIL‐53 (Al) CO2 303 118.68 1.171

amino‐MIL‐53 (Al) CO2 303 163.13 1.314

amino‐MIL‐53 (Al) CO2 303 241.64 1.517

amino‐MIL‐53 (Al) CO2 303 317.19 1.645

amino‐MIL‐53 (Al) CO2 303 358.66 1.720

amino‐MIL‐53 (Al) CO2 303 520.10 1.900

amino‐MIL‐53 (Al) CO2 303 653.39 2.027

amino‐MIL‐53 (Al) CO2 303 896.27 2.214

amino‐MIL‐53 (Al) CO2 303 1034.00 2.334

amino‐MIL‐53 (Al) CO2 303 1085.83 2.386

amino‐MIL‐53 (Al) CO2 303 1269.52 2.785

amino‐MIL‐53 (Al) CO2 303 1325.89 3.321

amino‐MIL‐53 (Al) CO2 303 1466.69 4.038

amino‐MIL‐53 (Al) CO2 303 1683.12 5.314

amino‐MIL‐53 (Al) CO2 303 1709.78 5.374

amino‐MIL‐53 (Al) CO2 303 1980.86 5.916

amino‐MIL‐53 (Al) CO2 303 2154.14 6.111

amino‐MIL‐53 (Al) CO2 303 2254.86 6.268

amino‐MIL‐53 (Al) CO2 303 2524.40 6.462

amino‐MIL‐53 (Al) CO2 303 2582.15 6.492

amino‐MIL‐53 (Al) CO2 303 2776.15 6.596

amino‐MIL‐53 (Al) CO2 303 2909.43 6.678

amino‐MIL‐53 (Al) CO2 303 2396.34 6.395

amino‐MIL‐53 (Al) CO2 303 2221.74 6.311

amino‐MIL‐53 (Al) CO2 303 2059.88 6.234

amino‐MIL‐53 (Al) CO2 303 1860.48 6.157

amino‐MIL‐53 (Al) CO2 303 1670.59 6.034

amino‐MIL‐53 (Al) CO2 303 1478.38 5.865

amino‐MIL‐53 (Al) CO2 303 1285.65 5.681


amino‐MIL‐53 (Al) CO2 303 1082.82 5.459

amino‐MIL‐53 (Al) CO2 303 896.14 5.106

amino‐MIL‐53 (Al) CO2 303 697.64 4.601

amino‐MIL‐53 (Al) CO2 303 544.25 2.996

amino‐MIL‐53 (Al) CO2 303 343.52 1.886

amino‐MIL‐53 (Al) CO2 303 192.00 1.516

amino‐MIL‐53 (Al) CO2 303 107.33 1.283

amino‐MIL‐53 (Al) CO2 303 63.59 1.073

amino‐MIL‐53 (Al) CO2 303 41.35 0.933

amino‐MIL‐53 (Al) CO2 303 29.34 0.770

amino‐MIL‐53 (Al) CO2 303 22.91 0.654

amino‐MIL‐53 (Al) CO2 303 18.97 0.577

amino‐MIL‐53 (Al) CO2 303 16.69 0.492

amino‐MIL‐53 (Al) CH4 303 69.53 0.020

amino‐MIL‐53 (Al) CH4 303 143.84 0.117

amino‐MIL‐53 (Al) CH4 303 250.65 0.220

amino‐MIL‐53 (Al) CH4 303 397.98 0.337

amino‐MIL‐53 (Al) CH4 303 529.22 0.469

amino‐MIL‐53 (Al) CH4 303 669.80 0.585

amino‐MIL‐53 (Al) CH4 303 802.17 0.693

amino‐MIL‐53 (Al) CH4 303 930.52 0.908

amino‐MIL‐53 (Al) CH4 303 1042.50 1.245

amino‐MIL‐53 (Al) CH4 303 1178.11 1.497

amino‐MIL‐53 (Al) CH4 303 1315.88 1.666

amino‐MIL‐53 (Al) CH4 303 1440.11 1.781

amino‐MIL‐53 (Al) CH4 303 1571.45 1.866

amino‐MIL‐53 (Al) CH4 303 1734.86 1.973

amino‐MIL‐53 (Al) CH4 303 1932.04 2.081

amino‐MIL‐53 (Al) CH4 303 2120.54 2.165

amino‐MIL‐53 (Al) CH4 303 2300.53 2.257

amino‐MIL‐53 (Al) CH4 303 2488.55 2.342

amino‐MIL‐53 (Al) CH4 303 2684.11 2.426

amino‐MIL‐53 (Al) CH4 303 2886.58 2.480

amino‐MIL‐53 (Al) CH4 303 2691.96 2.442

amino‐MIL‐53 (Al) CH4 303 2467.01 2.380

amino‐MIL‐53 (Al) CH4 303 2221.73 2.303

amino‐MIL‐53 (Al) CH4 303 1971.92 2.195

amino‐MIL‐53 (Al) CH4 303 1719.91 2.065

amino‐MIL‐53 (Al) CH4 303 1461.34 1.919

amino‐MIL‐53 (Al) CH4 303 1188.60 1.696

amino‐MIL‐53 (Al) CH4 303 947.07 1.168

amino‐MIL‐53 (Al) CH4 303 691.61 0.662

amino‐MIL‐53 (Al) CH4 303 419.35 0.322


amino‐MIL‐53 (Al) CH4 303 217.98 0.104

amino‐MIL‐53 (Al) CH4 303 115.98 0.001

E2 – Data table for experimental data presented in Figure 2.3

Table E2. Data points extracted graphically from Mason et al. [48]


Fe(bdp) CH4 248 74 0.17

Fe(bdp) CH4 248 202 0.25

Fe(bdp) CH4 248 319 0.39

Fe(bdp) CH4 248 621 0.59

Fe(bdp) CH4 248 818 0.72

Fe(bdp) CH4 248 1014 0.96

Fe(bdp) CH4 248 1147 1.38

Fe(bdp) CH4 248 1211 3.79

Fe(bdp) CH4 248 1269 7.72

Fe(bdp) CH4 248 1604 11.15

Fe(bdp) CH4 248 1843 11.74

Fe(bdp) CH4 248 1997 12.11

Fe(bdp) CH4 248 2119 12.40

Fe(bdp) CH4 248 2278 12.78

Fe(bdp) CH4 248 2512 13.34

Fe(bdp) CH4 248 2756 13.86

Fe(bdp) CH4 248 3011 14.63

Fe(bdp) CH4 248 3266 15.05

Fe(bdp) CH4 248 3521 15.34

Fe(bdp) CH4 248 3771 15.57

Fe(bdp) CH4 248 4020 15.82

Fe(bdp) CH4 248 4265 16.06

Fe(bdp) CH4 248 4504 16.26

Fe(bdp) CH4 248 4753 16.45

Fe(bdp) CH4 248 5003 16.62

Fe(bdp) CH4 248 5470 16.84

Fe(bdp) CH4 248 5964 17.11

Fe(bdp) CH4 248 6453 17.36

Fe(bdp) CH4 248 6984 17.56

Fe(bdp) CH4 248 6088 17.18

Fe(bdp) CH4 248 5102 16.63

Fe(bdp) CH4 248 4073 15.83

Fe(bdp) CH4 248 3040 14.77

Fe(bdp) CH4 248 2541 13.62

Fe(bdp) CH4 248 2038 12.45


Fe(bdp) CH4 248 1784 11.81

Fe(bdp) CH4 248 1614 11.38

Fe(bdp) CH4 248 1423 10.86

Fe(bdp) CH4 248 1222 10.21

Fe(bdp) CH4 248 1021 9.51

Fe(bdp) CH4 248 925 9.11

Fe(bdp) CH4 248 867 8.82

Fe(bdp) CH4 248 825 8.52

Fe(bdp) CH4 248 783 7.85

Fe(bdp) CH4 248 741 5.98

Fe(bdp) CH4 248 688 3.26

Fe(bdp) CH4 248 615 1.55

Fe(bdp) CH4 248 562 1.05

Fe(bdp) CH4 248 525 0.88

Fe(bdp) CH4 248 413 0.58

Fe(bdp) CH4 261 64 0.18

Fe(bdp) CH4 261 202 0.25

Fe(bdp) CH4 261 345 0.27

Fe(bdp) CH4 261 637 0.39

Fe(bdp) CH4 261 823 0.49

Fe(bdp) CH4 261 1020 0.59

Fe(bdp) CH4 261 1222 0.76

Fe(bdp) CH4 261 1349 0.88

Fe(bdp) CH4 261 1466 1.25

Fe(bdp) CH4 261 1540 2.74

Fe(bdp) CH4 261 1599 5.71

Fe(bdp) CH4 261 1657 8.35

Fe(bdp) CH4 261 1779 9.78

Fe(bdp) CH4 261 1891 10.27

Fe(bdp) CH4 261 2008 10.55

Fe(bdp) CH4 261 2098 10.77

Fe(bdp) CH4 261 2278 11.14

Fe(bdp) CH4 261 2517 11.63

Fe(bdp) CH4 261 2778 12.05

Fe(bdp) CH4 261 3027 12.47

Fe(bdp) CH4 261 3272 12.89

Fe(bdp) CH4 261 3521 13.36

Fe(bdp) CH4 261 3771 13.69

Fe(bdp) CH4 261 4020 14.32

Fe(bdp) CH4 261 4270 14.55

Fe(bdp) CH4 261 4514 14.77

Fe(bdp) CH4 261 4998 15.16

Fe(bdp) CH4 261 5508 15.45


Fe(bdp) CH4 261 5970 15.67

Fe(bdp) CH4 261 6442 15.84

Fe(bdp) CH4 261 6947 16.06

Fe(bdp) CH4 261 6067 15.66

Fe(bdp) CH4 261 5081 15.10

Fe(bdp) CH4 261 4069 14.33

Fe(bdp) CH4 261 3544 13.41

Fe(bdp) CH4 261 3035 12.74

Fe(bdp) CH4 261 2537 11.88

Fe(bdp) CH4 261 2017 10.83

Fe(bdp) CH4 261 1774 10.24

Fe(bdp) CH4 261 1620 9.89

Fe(bdp) CH4 261 1418 9.36

Fe(bdp) CH4 261 1217 8.77

Fe(bdp) CH4 261 1122 8.40

Fe(bdp) CH4 261 1021 7.90

Fe(bdp) CH4 261 963 7.23

Fe(bdp) CH4 261 931 6.38

Fe(bdp) CH4 261 889 4.88

Fe(bdp) CH4 261 832 2.71

Fe(bdp) CH4 261 768 1.65

Fe(bdp) CH4 261 716 1.12

Fe(bdp) CH4 261 525 0.48

Fe(bdp) CH4 273 53 0.15

Fe(bdp) CH4 273 175 0.19

Fe(bdp) CH4 273 329 0.29

Fe(bdp) CH4 273 632 0.35

Fe(bdp) CH4 273 823 0.41

Fe(bdp) CH4 273 1025 0.47

Fe(bdp) CH4 273 1216 0.54

Fe(bdp) CH4 273 1418 0.64

Fe(bdp) CH4 273 1609 0.73

Fe(bdp) CH4 273 1742 1.01

Fe(bdp) CH4 273 1832 1.97

Fe(bdp) CH4 273 1901 4.69

Fe(bdp) CH4 273 1986 7.55

Fe(bdp) CH4 273 2098 9.13

Fe(bdp) CH4 273 2209 9.78

Fe(bdp) CH4 273 2305 10.03

Fe(bdp) CH4 273 2417 10.25

Fe(bdp) CH4 273 2539 10.44

Fe(bdp) CH4 273 2714 10.74

Fe(bdp) CH4 273 2916 11.04


Fe(bdp) CH4 273 3219 11.48

Fe(bdp) CH4 273 3617 12.34

Fe(bdp) CH4 273 3999 12.96

Fe(bdp) CH4 273 4509 13.53

Fe(bdp) CH4 273 4976 13.88

Fe(bdp) CH4 273 5508 14.26

Fe(bdp) CH4 273 6007 14.51

Fe(bdp) CH4 273 6490 14.81

Fe(bdp) CH4 273 6963 15.18

Fe(bdp) CH4 273 6057 14.51

Fe(bdp) CH4 273 5066 13.78

Fe(bdp) CH4 273 4064 12.98

Fe(bdp) CH4 273 3491 12.31

Fe(bdp) CH4 273 3041 11.19

Fe(bdp) CH4 273 2018 9.45

Fe(bdp) CH4 273 1769 8.89

Fe(bdp) CH4 273 1620 8.52

Fe(bdp) CH4 273 1424 8.04

Fe(bdp) CH4 273 1233 7.32

Fe(bdp) CH4 273 1117 6.20

Fe(bdp) CH4 273 1038 3.99

Fe(bdp) CH4 273 927 1.89

Fe(bdp) CH4 273 816 0.98

Fe(bdp) CH4 285 53 0.15

Fe(bdp) CH4 285 175 0.19

Fe(bdp) CH4 285 329 0.29

Fe(bdp) CH4 285 632 0.35

Fe(bdp) CH4 285 823 0.41

Fe(bdp) CH4 285 1025 0.47

Fe(bdp) CH4 285 1216 0.54

Fe(bdp) CH4 285 1418 0.64

Fe(bdp) CH4 285 1609 0.73

Fe(bdp) CH4 285 1747 0.78

Fe(bdp) CH4 285 1896 0.88

Fe(bdp) CH4 285 2018 1.02

Fe(bdp) CH4 285 2119 1.30

Fe(bdp) CH4 285 2199 2.36

Fe(bdp) CH4 285 2273 4.60

Fe(bdp) CH4 285 2353 7.10

Fe(bdp) CH4 285 2438 8.43

Fe(bdp) CH4 285 2533 9.01

Fe(bdp) CH4 285 2624 9.40

Fe(bdp) CH4 285 2725 9.72


Fe(bdp) CH4 285 2836 9.99

Fe(bdp) CH4 285 2942 10.14

Fe(bdp) CH4 285 3038 10.27

Fe(bdp) CH4 285 3219 10.57

Fe(bdp) CH4 285 3415 10.81

Fe(bdp) CH4 285 3617 11.05

Fe(bdp) CH4 285 3813 11.31

Fe(bdp) CH4 285 4015 11.95

Fe(bdp) CH4 285 4281 12.34

Fe(bdp) CH4 285 4514 12.61

Fe(bdp) CH4 285 4775 12.86

Fe(bdp) CH4 285 5014 13.10

Fe(bdp) CH4 285 5502 13.42

Fe(bdp) CH4 285 5980 13.77

Fe(bdp) CH4 285 6464 14.14

Fe(bdp) CH4 285 7011 14.50

Fe(bdp) CH4 285 6063 13.74

Fe(bdp) CH4 285 5071 12.96

Fe(bdp) CH4 285 4069 12.06

Fe(bdp) CH4 285 3523 10.99

Fe(bdp) CH4 285 2527 9.49

Fe(bdp) CH4 285 2029 8.55

Fe(bdp) CH4 285 1801 8.05

Fe(bdp) CH4 285 1419 6.87

Fe(bdp) CH4 285 1324 6.05

Fe(bdp) CH4 285 1229 4.36

Fe(bdp) CH4 285 1134 2.45

Fe(bdp) CH4 285 1018 1.40

Fe(bdp) CH4 285 917 0.92

Fe(bdp) CH4 298 53 0.15

Fe(bdp) CH4 298 175 0.19

Fe(bdp) CH4 298 292 0.20

Fe(bdp) CH4 298 409 0.24

Fe(bdp) CH4 298 611 0.29

Fe(bdp) CH4 298 829 0.26

Fe(bdp) CH4 298 1025 0.31

Fe(bdp) CH4 298 1222 0.31

Fe(bdp) CH4 298 1423 0.34

Fe(bdp) CH4 298 1625 0.36

Fe(bdp) CH4 298 1832 0.53

Fe(bdp) CH4 298 2013 0.61

Fe(bdp) CH4 298 2167 0.71

Fe(bdp) CH4 298 2273 0.83


Fe(bdp) CH4 298 2379 0.98

Fe(bdp) CH4 298 2459 1.24

Fe(bdp) CH4 298 2533 2.02

Fe(bdp) CH4 298 2597 3.57

Fe(bdp) CH4 298 2666 5.46

Fe(bdp) CH4 298 2762 7.22

Fe(bdp) CH4 298 2852 8.08

Fe(bdp) CH4 298 2958 8.61

Fe(bdp) CH4 298 3054 8.95

Fe(bdp) CH4 298 3171 9.23

Fe(bdp) CH4 298 3282 9.43

Fe(bdp) CH4 298 3383 9.67

Fe(bdp) CH4 298 3505 9.87

Fe(bdp) CH4 298 3622 10.04

Fe(bdp) CH4 298 3750 10.22

Fe(bdp) CH4 298 3877 10.39

Fe(bdp) CH4 298 4026 10.53

Fe(bdp) CH4 298 4275 10.83

Fe(bdp) CH4 298 4509 11.55

Fe(bdp) CH4 298 4769 11.75

Fe(bdp) CH4 298 5008 11.97

Fe(bdp) CH4 298 5513 12.28

Fe(bdp) CH4 298 5964 12.58

Fe(bdp) CH4 298 6517 12.94

Fe(bdp) CH4 298 7016 13.31

Fe(bdp) CH4 298 6047 12.52

Fe(bdp) CH4 298 5050 11.80

Fe(bdp) CH4 298 4064 10.57

Fe(bdp) CH4 298 3036 9.15

Fe(bdp) CH4 298 2516 8.38

Fe(bdp) CH4 298 2029 7.47

Fe(bdp) CH4 298 1801 6.90

Fe(bdp) CH4 298 1626 6.16

Fe(bdp) CH4 298 1447 3.86

Fe(bdp) CH4 298 1346 2.50

Fe(bdp) CH4 298 1230 1.45

Fe(bdp) CH4 311 53 0.15

Fe(bdp) CH4 311 175 0.19

Fe(bdp) CH4 311 292 0.20

Fe(bdp) CH4 311 409 0.24

Fe(bdp) CH4 311 611 0.29

Fe(bdp) CH4 311 829 0.26

Fe(bdp) CH4 311 1025 0.31


Fe(bdp) CH4 311 1222 0.31

Fe(bdp) CH4 311 1423 0.34

Fe(bdp) CH4 311 1625 0.36

Fe(bdp) CH4 311 1822 0.38

Fe(bdp) CH4 311 2024 0.40

Fe(bdp) CH4 311 2172 0.53

Fe(bdp) CH4 311 2300 0.55

Fe(bdp) CH4 311 2448 0.62

Fe(bdp) CH4 311 2581 0.68

Fe(bdp) CH4 311 2698 0.78

Fe(bdp) CH4 311 2820 1.10

Fe(bdp) CH4 311 2910 1.82

Fe(bdp) CH4 311 2995 3.38

Fe(bdp) CH4 311 3086 5.11

Fe(bdp) CH4 311 3176 6.53

Fe(bdp) CH4 311 3282 7.39

Fe(bdp) CH4 311 3378 7.96

Fe(bdp) CH4 311 3458 8.30

Fe(bdp) CH4 311 3564 8.61

Fe(bdp) CH4 311 3659 8.85

Fe(bdp) CH4 311 3760 9.03

Fe(bdp) CH4 311 3861 9.24

Fe(bdp) CH4 311 3941 9.42

Fe(bdp) CH4 311 4031 9.54

Fe(bdp) CH4 311 4281 9.94

Fe(bdp) CH4 311 4520 10.25

Fe(bdp) CH4 311 4780 10.56

Fe(bdp) CH4 311 5035 11.20

Fe(bdp) CH4 311 5508 11.64

Fe(bdp) CH4 311 6023 11.98

Fe(bdp) CH4 311 6458 12.25

Fe(bdp) CH4 311 7011 12.52

Fe(bdp) CH4 311 6047 11.89

Fe(bdp) CH4 311 5056 11.12

Fe(bdp) CH4 311 4065 9.82

Fe(bdp) CH4 311 3529 9.05

Fe(bdp) CH4 311 3020 8.37

Fe(bdp) CH4 311 2527 7.51

Fe(bdp) CH4 311 2024 6.58

Fe(bdp) CH4 311 1807 5.74

Fe(bdp) CH4 311 1643 4.21

Fe(bdp) CH4 311 1436 2.15

Fe(bdp) CH4 311 1325 1.55


Fe(bdp) CH4 311 1219 1.15

Fe(bdp) CH4 311 514 0.35

Fe(bdp) CH4 323 53 0.15

Fe(bdp) CH4 323 175 0.19

Fe(bdp) CH4 323 292 0.20

Fe(bdp) CH4 323 409 0.24

Fe(bdp) CH4 323 611 0.29

Fe(bdp) CH4 323 829 0.26

Fe(bdp) CH4 323 1025 0.31

Fe(bdp) CH4 323 1222 0.31

Fe(bdp) CH4 323 1423 0.34

Fe(bdp) CH4 323 1625 0.36

Fe(bdp) CH4 323 1822 0.38

Fe(bdp) CH4 323 2024 0.40

Fe(bdp) CH4 323 2231 0.43

Fe(bdp) CH4 323 2432 0.46

Fe(bdp) CH4 323 2618 0.52

Fe(bdp) CH4 323 2825 0.60

Fe(bdp) CH4 323 3017 0.77

Fe(bdp) CH4 323 3149 1.10

Fe(bdp) CH4 323 3250 1.96

Fe(bdp) CH4 323 3341 3.25

Fe(bdp) CH4 323 3431 4.66

Fe(bdp) CH4 323 3521 5.93

Fe(bdp) CH4 323 3606 6.74

Fe(bdp) CH4 323 3697 7.26

Fe(bdp) CH4 323 3792 7.64

Fe(bdp) CH4 323 3882 7.95

Fe(bdp) CH4 323 3978 8.18

Fe(bdp) CH4 323 4052 8.33

Fe(bdp) CH4 323 4153 8.57

Fe(bdp) CH4 323 4244 8.75

Fe(bdp) CH4 323 4329 8.94

Fe(bdp) CH4 323 4440 9.14

Fe(bdp) CH4 323 4525 9.24

Fe(bdp) CH4 323 4775 9.51

Fe(bdp) CH4 323 5030 9.76

Fe(bdp) CH4 323 5470 10.08

Fe(bdp) CH4 323 5970 10.35

Fe(bdp) CH4 323 6464 10.52

Fe(bdp) CH4 323 6979 10.71

Fe(bdp) CH4 323 6037 10.18

Fe(bdp) CH4 323 5051 9.48


Fe(bdp) CH4 323 4049 8.53

Fe(bdp) CH4 323 3530 7.90

Fe(bdp) CH4 323 3021 7.08

Fe(bdp) CH4 323 2522 6.26

Fe(bdp) CH4 323 2035 5.13

Fe(bdp) CH4 323 1823 3.77

Fe(bdp) CH4 323 1638 2.27

Fe(bdp) CH4 323 504 0.02


Table E3. Data points extracted graphically from Mason et al. [48]


Co(bdp) CH4 273 79 0.13

Co(bdp) CH4 273 214 0.18

Co(bdp) CH4 273 355 0.21

Co(bdp) CH4 273 512 0.26

Co(bdp) CH4 273 670 0.34

Co(bdp) CH4 273 827 0.40

Co(bdp) CH4 273 979 0.51

Co(bdp) CH4 273 1086 0.66

Co(bdp) CH4 273 1154 1.12

Co(bdp) CH4 273 1216 2.18

Co(bdp) CH4 273 1279 3.61

Co(bdp) CH4 273 1347 5.54

Co(bdp) CH4 273 1545 7.29

Co(bdp) CH4 273 1709 7.90

Co(bdp) CH4 273 1850 8.24

Co(bdp) CH4 273 1962 8.43

Co(bdp) CH4 273 2052 8.63

Co(bdp) CH4 273 2216 8.86

Co(bdp) CH4 273 2452 9.20

Co(bdp) CH4 273 2762 9.61

Co(bdp) CH4 273 3009 9.90

Co(bdp) CH4 273 3511 10.41

Co(bdp) CH4 273 3995 10.91

Co(bdp) CH4 273 4496 11.34

Co(bdp) CH4 273 4991 11.72

Co(bdp) CH4 273 5492 12.06

Co(bdp) CH4 273 5993 12.37

Co(bdp) CH4 273 6488 12.66

Co(bdp) CH4 273 6972 12.89


Co(bdp) CH4 273 6073 12.42

Co(bdp) CH4 273 5083 11.78

Co(bdp) CH4 273 4087 10.96

Co(bdp) CH4 273 3524 10.46

Co(bdp) CH4 273 2522 9.31

Co(bdp) CH4 273 1756 8.15

Co(bdp) CH4 273 1514 7.58

Co(bdp) CH4 273 1272 6.99

Co(bdp) CH4 273 1018 6.21

Co(bdp) CH4 273 882 5.72

Co(bdp) CH4 273 803 5.42

Co(bdp) CH4 273 753 5.19

Co(bdp) CH4 273 719 4.98

Co(bdp) CH4 273 690 4.74

Co(bdp) CH4 273 662 4.34

Co(bdp) CH4 273 621 3.22

Co(bdp) CH4 273 569 1.73

Co(bdp) CH4 273 512 0.72

Co(bdp) CH4 273 383 0.24

Co(bdp) CH4 285 101 0.12

Co(bdp) CH4 285 225 0.16

Co(bdp) CH4 285 349 0.20

Co(bdp) CH4 285 518 0.27

Co(bdp) CH4 285 653 0.34

Co(bdp) CH4 285 811 0.43

Co(bdp) CH4 285 957 0.48

Co(bdp) CH4 285 1120 0.55

Co(bdp) CH4 285 1278 0.65

Co(bdp) CH4 285 1357 0.94

Co(bdp) CH4 285 1424 1.73

Co(bdp) CH4 285 1481 2.96

Co(bdp) CH4 285 1555 4.35

Co(bdp) CH4 285 1680 6.02

Co(bdp) CH4 285 1815 6.78

Co(bdp) CH4 285 1928 7.20

Co(bdp) CH4 285 2035 7.47

Co(bdp) CH4 285 2221 7.93

Co(bdp) CH4 285 2407 8.26

Co(bdp) CH4 285 2609 8.54

Co(bdp) CH4 285 2818 8.81

Co(bdp) CH4 285 3009 9.03

Co(bdp) CH4 285 3257 9.33

Co(bdp) CH4 285 3521 9.58


Co(bdp) CH4 285 4000 10.05

Co(bdp) CH4 285 4501 10.45

Co(bdp) CH4 285 4996 10.84

Co(bdp) CH4 285 5492 11.23

Co(bdp) CH4 285 5993 11.57

Co(bdp) CH4 285 6494 11.84

Co(bdp) CH4 285 6983 12.16

Co(bdp) CH4 285 6049 11.52

Co(bdp) CH4 285 5070 10.85

Co(bdp) CH4 285 4063 10.04

Co(bdp) CH4 285 2532 8.40

Co(bdp) CH4 285 2031 7.64

Co(bdp) CH4 285 1778 7.17

Co(bdp) CH4 285 1513 6.61

Co(bdp) CH4 285 1321 6.13

Co(bdp) CH4 285 1169 5.71

Co(bdp) CH4 285 1028 5.26

Co(bdp) CH4 285 949 5.00

Co(bdp) CH4 285 898 4.79

Co(bdp) CH4 285 848 4.61

Co(bdp) CH4 285 819 4.39

Co(bdp) CH4 285 763 3.99

Co(bdp) CH4 285 723 3.09

Co(bdp) CH4 285 625 1.08

Co(bdp) CH4 285 405 0.26

Co(bdp) CH4 298 68 0.12

Co(bdp) CH4 298 180 0.13

Co(bdp) CH4 298 321 0.14

Co(bdp) CH4 298 613 0.18

Co(bdp) CH4 298 827 0.25

Co(bdp) CH4 298 940 0.29

Co(bdp) CH4 298 1131 0.32

Co(bdp) CH4 298 1300 0.35

Co(bdp) CH4 298 1452 0.43

Co(bdp) CH4 298 1542 0.58

Co(bdp) CH4 298 1593 0.77

Co(bdp) CH4 298 1661 1.41

Co(bdp) CH4 298 1734 2.51

Co(bdp) CH4 298 1819 3.73

Co(bdp) CH4 298 1899 4.78

Co(bdp) CH4 298 1989 5.50

Co(bdp) CH4 298 2079 5.97

Co(bdp) CH4 298 2164 6.28


Co(bdp) CH4 298 2248 6.54

Co(bdp) CH4 298 2327 6.75

Co(bdp) CH4 298 2423 7.01

Co(bdp) CH4 298 2524 7.21

Co(bdp) CH4 298 2654 7.51

Co(bdp) CH4 298 2812 7.76

Co(bdp) CH4 298 3020 8.11

Co(bdp) CH4 298 3262 8.41

Co(bdp) CH4 298 3515 8.71

Co(bdp) CH4 298 3999 9.14

Co(bdp) CH4 298 4506 9.54

Co(bdp) CH4 298 4996 9.92

Co(bdp) CH4 298 5502 10.28

Co(bdp) CH4 298 5998 10.62

Co(bdp) CH4 298 6493 10.92

Co(bdp) CH4 298 6994 11.22

Co(bdp) CH4 298 6054 10.65

Co(bdp) CH4 298 5058 10.00

Co(bdp) CH4 298 4062 9.23

Co(bdp) CH4 298 3528 8.70

Co(bdp) CH4 298 3027 8.17

Co(bdp) CH4 298 2526 7.54

Co(bdp) CH4 298 2030 6.73

Co(bdp) CH4 298 1766 6.25

Co(bdp) CH4 298 1512 5.71

Co(bdp) CH4 298 1270 5.11

Co(bdp) CH4 298 1112 4.67

Co(bdp) CH4 298 1016 4.38

Co(bdp) CH4 298 954 4.14

Co(bdp) CH4 298 920 3.95

Co(bdp) CH4 298 864 3.67

Co(bdp) CH4 298 818 3.24

Co(bdp) CH4 298 773 2.57

Co(bdp) CH4 298 715 1.51

Co(bdp) CH4 298 613 0.54

Co(bdp) CH4 311 90 0.08

Co(bdp) CH4 311 214 0.08

Co(bdp) CH4 311 338 0.10

Co(bdp) CH4 311 608 0.14

Co(bdp) CH4 311 822 0.21

Co(bdp) CH4 311 1013 0.22

Co(bdp) CH4 311 1204 0.29

Co(bdp) CH4 311 1407 0.30


Co(bdp) CH4 311 1525 0.39

Co(bdp) CH4 311 1666 0.45

Co(bdp) CH4 311 1795 0.73

Co(bdp) CH4 311 1858 1.24

Co(bdp) CH4 311 1931 2.04

Co(bdp) CH4 311 2005 2.92

Co(bdp) CH4 311 2089 3.80

Co(bdp) CH4 311 2169 4.56

Co(bdp) CH4 311 2264 5.18

Co(bdp) CH4 311 2360 5.59

Co(bdp) CH4 311 2462 5.92

Co(bdp) CH4 311 2558 6.18

Co(bdp) CH4 311 2653 6.39

Co(bdp) CH4 311 2755 6.59

Co(bdp) CH4 311 2834 6.76

Co(bdp) CH4 311 3014 7.14

Co(bdp) CH4 311 3267 7.53

Co(bdp) CH4 311 3515 7.87

Co(bdp) CH4 311 4005 8.39

Co(bdp) CH4 311 4500 8.82

Co(bdp) CH4 311 5001 9.19

Co(bdp) CH4 311 5491 9.56

Co(bdp) CH4 311 6003 9.87

Co(bdp) CH4 311 6498 10.16

Co(bdp) CH4 311 6994 10.41

Co(bdp) CH4 311 6048 9.86

Co(bdp) CH4 311 5046 9.19

Co(bdp) CH4 311 4039 8.35

Co(bdp) CH4 311 3026 7.24

Co(bdp) CH4 311 2525 6.57

Co(bdp) CH4 311 2030 5.80

Co(bdp) CH4 311 1770 5.31

Co(bdp) CH4 311 1517 4.79

Co(bdp) CH4 311 1320 4.35

Co(bdp) CH4 311 1168 3.96

Co(bdp) CH4 311 1077 3.71

Co(bdp) CH4 311 1021 3.49

Co(bdp) CH4 311 953 3.29

Co(bdp) CH4 311 919 3.09

Co(bdp) CH4 311 863 2.76

Co(bdp) CH4 311 823 2.25

Co(bdp) CH4 311 732 0.98

Co(bdp) CH4 323 73 0.08


Co(bdp) CH4 323 191 0.13

Co(bdp) CH4 323 338 0.18

Co(bdp) CH4 323 597 0.18

Co(bdp) CH4 323 799 0.18

Co(bdp) CH4 323 1013 0.24

Co(bdp) CH4 323 1199 0.30

Co(bdp) CH4 323 1413 0.32

Co(bdp) CH4 323 1621 0.34

Co(bdp) CH4 323 1767 0.41

Co(bdp) CH4 323 1874 0.47

Co(bdp) CH4 323 1959 0.69

Co(bdp) CH4 323 2032 1.17

Co(bdp) CH4 323 2111 1.81

Co(bdp) CH4 323 2184 2.52

Co(bdp) CH4 323 2269 3.38

Co(bdp) CH4 323 2371 4.18

Co(bdp) CH4 323 2456 4.73

Co(bdp) CH4 323 2563 5.08

Co(bdp) CH4 323 2653 5.41

Co(bdp) CH4 323 2743 5.66

Co(bdp) CH4 323 2839 5.83

Co(bdp) CH4 323 3013 6.27

Co(bdp) CH4 323 3272 6.71

Co(bdp) CH4 323 3526 7.13

Co(bdp) CH4 323 4016 7.67

Co(bdp) CH4 323 4505 8.14

Co(bdp) CH4 323 5001 8.51

Co(bdp) CH4 323 5496 8.88

Co(bdp) CH4 323 6003 9.18

Co(bdp) CH4 323 6498 9.45

Co(bdp) CH4 323 6999 9.71

Co(bdp) CH4 323 6036 9.17

Co(bdp) CH4 323 5046 8.52

Co(bdp) CH4 323 4050 7.68

Co(bdp) CH4 323 3020 6.57

Co(bdp) CH4 323 2519 5.89

Co(bdp) CH4 323 2018 5.09

Co(bdp) CH4 323 1764 4.65

Co(bdp) CH4 323 1511 4.13

Co(bdp) CH4 323 1330 3.74

Co(bdp) CH4 323 1167 3.41

Co(bdp) CH4 323 1066 3.10

Co(bdp) CH4 323 1015 2.96


Co(bdp) CH4 323 964 2.75

Co(bdp) CH4 323 919 2.50

Co(bdp) CH4 323 874 2.22

Co(bdp) CH4 323 823 1.81

Co(bdp) CH4 323 715 0.92

E4 – Data table for experimental data presented in Figure 2.9 and Figure 2.10

Table E4. Data table for experimental data presented in Figure 2.9 and Figure 2.10


ZIF‐7 CO2 233 0.09 0.020

ZIF‐7 CO2 233 0.46 0.084

ZIF‐7 CO2 233 1.13 0.147

ZIF‐7 CO2 233 1.87 0.333

ZIF‐7 CO2 233 3.96 1.646

ZIF‐7 CO2 233 26.09 1.879

ZIF‐7 CO2 233 19.21 1.868

ZIF‐7 CO2 233 9.54 1.808

ZIF‐7 CO2 233 5.12 1.735

ZIF‐7 CO2 233 3.87 1.696

ZIF‐7 CO2 233 3.12 1.663

ZIF‐7 CO2 233 1.91 1.573

ZIF‐7 CO2 233 1.04 0.591

ZIF‐7 CO2 233 0.49 0.445

ZIF‐7 CO2 238 0.10 0.017

ZIF‐7 CO2 238 0.48 0.064

ZIF‐7 CO2 238 1.03 0.098

ZIF‐7 CO2 238 2.00 0.178

ZIF‐7 CO2 238 3.00 0.729

ZIF‐7 CO2 238 4.30 1.551

ZIF‐7 CO2 238 16.12 1.812

ZIF‐7 CO2 238 20.96 1.838

ZIF‐7 CO2 238 30.62 1.872

ZIF‐7 CO2 238 39.76 1.893

ZIF‐7 CO2 238 28.92 1.869

ZIF‐7 CO2 238 19.34 1.835

ZIF‐7 CO2 238 9.87 1.767

ZIF‐7 CO2 238 5.18 1.679

ZIF‐7 CO2 238 3.99 1.637

ZIF‐7 CO2 238 3.10 1.592

ZIF‐7 CO2 238 2.08 1.463

ZIF‐7 CO2 238 1.02 0.457


ZIF‐7 CO2 238 0.51 0.390

ZIF‐7 CO2 244 0.56 0.055

ZIF‐7 CO2 244 1.23 0.163

ZIF‐7 CO2 244 3.32 0.483

ZIF‐7 CO2 244 5.30 1.356

ZIF‐7 CO2 244 39.25 1.853

ZIF‐7 CO2 244 42.61 1.862

ZIF‐7 CO2 244 44.75 1.867

ZIF‐7 CO2 244 44.95 1.868

ZIF‐7 CO2 244 40.80 1.859

ZIF‐7 CO2 244 38.21 1.853

ZIF‐7 CO2 244 35.72 1.847

ZIF‐7 CO2 244 34.97 1.845

ZIF‐7 CO2 244 32.53 1.839

ZIF‐7 CO2 244 29.97 1.831

ZIF‐7 CO2 244 27.49 1.822

ZIF‐7 CO2 244 25.01 1.812

ZIF‐7 CO2 244 22.49 1.800

ZIF‐7 CO2 244 20.02 1.787

ZIF‐7 CO2 244 17.51 1.771

ZIF‐7 CO2 244 15.05 1.752

ZIF‐7 CO2 244 12.10 1.722

ZIF‐7 CO2 244 9.14 1.679

ZIF‐7 CO2 244 6.89 1.632

ZIF‐7 CO2 244 4.80 1.564

ZIF‐7 CO2 244 3.92 1.484

ZIF‐7 CO2 244 2.79 1.134

ZIF‐7 CO2 247 4.97 0.454

ZIF‐7 CO2 247 10.82 1.519

ZIF‐7 CO2 247 24.38 1.750

ZIF‐7 CO2 247 27.34 1.766

ZIF‐7 CO2 247 30.38 1.780

ZIF‐7 CO2 247 31.88 1.787

ZIF‐7 CO2 247 35.13 1.800

ZIF‐7 CO2 247 38.00 1.809

ZIF‐7 CO2 247 40.93 1.818

ZIF‐7 CO2 247 44.04 1.827

ZIF‐7 CO2 247 46.95 1.834

ZIF‐7 CO2 247 50.01 1.841

ZIF‐7 CO2 247 53.06 1.847

ZIF‐7 CO2 247 56.07 1.853

ZIF‐7 CO2 247 60.04 1.861

ZIF‐7 CO2 247 65.08 1.869


ZIF‐7 CO2 247 69.97 1.876

ZIF‐7 CO2 247 75.05 1.883

ZIF‐7 CO2 247 80.02 1.890

ZIF‐7 CO2 247 85.04 1.896

ZIF‐7 CO2 247 90.00 1.902

ZIF‐7 CO2 247 95.05 1.908

ZIF‐7 CO2 247 100.04 1.913

ZIF‐7 CO2 247 93.24 1.908

ZIF‐7 CO2 247 88.23 1.904

ZIF‐7 CO2 247 83.23 1.899

ZIF‐7 CO2 247 80.03 1.896

ZIF‐7 CO2 247 75.07 1.890

ZIF‐7 CO2 247 70.08 1.884

ZIF‐7 CO2 247 65.11 1.878

ZIF‐7 CO2 247 59.96 1.871

ZIF‐7 CO2 247 55.97 1.865

ZIF‐7 CO2 247 52.93 1.860

ZIF‐7 CO2 247 50.10 1.855

ZIF‐7 CO2 247 47.13 1.849

ZIF‐7 CO2 247 43.96 1.842

ZIF‐7 CO2 247 40.96 1.835

ZIF‐7 CO2 247 37.97 1.828

ZIF‐7 CO2 247 34.98 1.819

ZIF‐7 CO2 247 31.95 1.810

ZIF‐7 CO2 247 29.01 1.800

ZIF‐7 CO2 247 26.01 1.787

ZIF‐7 CO2 247 23.01 1.773

ZIF‐7 CO2 247 20.02 1.756

ZIF‐7 CO2 247 17.06 1.735

ZIF‐7 CO2 247 14.04 1.709

ZIF‐7 CO2 247 11.09 1.674

ZIF‐7 CO2 247 8.18 1.625

ZIF‐7 CO2 247 4.77 1.524

ZIF‐7 CO2 253 4.82 0.378

ZIF‐7 CO2 253 14.43 1.639

ZIF‐7 CO2 253 44.29 1.805

ZIF‐7 CO2 253 48.60 1.817

ZIF‐7 CO2 253 51.65 1.824

ZIF‐7 CO2 253 53.01 1.828

ZIF‐7 CO2 253 56.13 1.834

ZIF‐7 CO2 253 60.06 1.841

ZIF‐7 CO2 253 65.00 1.850

ZIF‐7 CO2 253 70.05 1.858


ZIF‐7 CO2 253 75.05 1.865

ZIF‐7 CO2 253 80.07 1.872

ZIF‐7 CO2 253 85.05 1.878

ZIF‐7 CO2 253 90.03 1.884

ZIF‐7 CO2 253 95.05 1.889

ZIF‐7 CO2 253 100.03 1.895

ZIF‐7 CO2 253 93.26 1.889

ZIF‐7 CO2 253 88.28 1.884

ZIF‐7 CO2 253 83.27 1.879

ZIF‐7 CO2 253 80.03 1.876

ZIF‐7 CO2 253 75.07 1.870

ZIF‐7 CO2 253 69.95 1.863

ZIF‐7 CO2 253 65.05 1.856

ZIF‐7 CO2 253 60.12 1.848

ZIF‐7 CO2 253 55.95 1.841

ZIF‐7 CO2 253 52.94 1.835

ZIF‐7 CO2 253 49.97 1.829

ZIF‐7 CO2 253 46.96 1.822

ZIF‐7 CO2 253 43.97 1.815

ZIF‐7 CO2 253 40.95 1.807

ZIF‐7 CO2 253 37.98 1.798

ZIF‐7 CO2 253 34.95 1.789

ZIF‐7 CO2 253 32.01 1.778

ZIF‐7 CO2 253 28.99 1.765

ZIF‐7 CO2 253 26.03 1.751

ZIF‐7 CO2 253 23.01 1.734

ZIF‐7 CO2 253 20.01 1.714

ZIF‐7 CO2 253 17.05 1.690

ZIF‐7 CO2 253 14.06 1.658

ZIF‐7 CO2 253 11.11 1.618

ZIF‐7 CO2 253 8.21 1.560

ZIF‐7 CO2 253 5.23 1.273

ZIF‐7 CO2 253 2.58 0.323

ZIF‐7 CO2 253 1.28 0.222

ZIF‐7 CO2 273 2.62 0.067

ZIF‐7 CO2 273 7.22 0.217

ZIF‐7 CO2 273 13.79 0.345

ZIF‐7 CO2 273 15.43 0.383

ZIF‐7 CO2 273 19.72 0.604

ZIF‐7 CO2 273 24.83 1.121

ZIF‐7 CO2 273 43.80 1.561

ZIF‐7 CO2 273 46.42 1.574

ZIF‐7 CO2 273 51.19 1.593


ZIF‐7 CO2 273 54.96 1.606

ZIF‐7 CO2 273 60.00 1.622

ZIF‐7 CO2 273 65.04 1.636

ZIF‐7 CO2 273 70.07 1.648

ZIF‐7 CO2 273 75.04 1.659

ZIF‐7 CO2 273 80.05 1.669

ZIF‐7 CO2 273 85.04 1.679

ZIF‐7 CO2 273 90.05 1.687

ZIF‐7 CO2 273 95.03 1.696

ZIF‐7 CO2 273 100.01 1.703

ZIF‐7 CO2 273 93.42 1.695

ZIF‐7 CO2 273 88.33 1.688

ZIF‐7 CO2 273 83.37 1.680

ZIF‐7 CO2 273 79.94 1.674

ZIF‐7 CO2 273 75.12 1.665

ZIF‐7 CO2 273 70.05 1.655

ZIF‐7 CO2 273 65.14 1.643

ZIF‐7 CO2 273 60.01 1.630

ZIF‐7 CO2 273 55.16 1.617

ZIF‐7 CO2 273 49.98 1.601

ZIF‐7 CO2 273 44.97 1.583

ZIF‐7 CO2 273 39.99 1.562

ZIF‐7 CO2 273 35.00 1.537

ZIF‐7 CO2 273 30.05 1.508

ZIF‐7 CO2 273 25.07 1.472

ZIF‐7 CO2 273 20.21 1.423

ZIF‐7 CO2 273 14.81 1.153

ZIF‐7 CO2 273 9.72 0.363

ZIF‐7 CO2 273 4.00 0.208

ZIF‐7 CO2 273 2.58 0.163

ZIF‐7 CO2 273 0.92 0.105

ZIF‐7 CO2 283 4.89 0.119

ZIF‐7 CO2 283 8.59 0.173

ZIF‐7 CO2 283 11.26 0.206

ZIF‐7 CO2 283 13.82 0.239

ZIF‐7 CO2 283 16.98 0.282

ZIF‐7 CO2 283 19.87 0.328

ZIF‐7 CO2 283 22.85 0.386

ZIF‐7 CO2 283 25.60 0.470

ZIF‐7 CO2 283 29.11 0.699

ZIF‐7 CO2 283 32.38 0.999

ZIF‐7 CO2 283 38.55 1.243

ZIF‐7 CO2 283 40.83 1.296


ZIF‐7 CO2 283 44.60 1.373

ZIF‐7 CO2 283 46.93 1.410

ZIF‐7 CO2 283 50.02 1.447

ZIF‐7 CO2 283 53.03 1.471

ZIF‐7 CO2 283 55.99 1.488

ZIF‐7 CO2 283 59.99 1.504

ZIF‐7 CO2 283 80.03 1.558

ZIF‐7 CO2 283 99.52 1.595

ZIF‐7 CO2 283 79.64 1.559

ZIF‐7 CO2 283 60.04 1.508

ZIF‐7 CO2 283 54.70 1.491

ZIF‐7 CO2 283 52.83 1.484

ZIF‐7 CO2 283 50.11 1.474

ZIF‐7 CO2 283 46.96 1.461

ZIF‐7 CO2 283 43.97 1.448

ZIF‐7 CO2 283 40.97 1.434

ZIF‐7 CO2 283 37.98 1.418

ZIF‐7 CO2 283 34.99 1.401

ZIF‐7 CO2 283 32.01 1.381

ZIF‐7 CO2 283 29.03 1.358

ZIF‐7 CO2 283 26.13 1.326

ZIF‐7 CO2 283 22.67 1.228

ZIF‐7 CO2 283 19.51 0.940

ZIF‐7 CO2 283 17.40 0.535

ZIF‐7 CO2 283 14.58 0.351

ZIF‐7 CO2 283 11.03 0.258

ZIF‐7 CO2 283 7.86 0.195

ZIF‐7 CO2 283 4.87 0.137

ZIF‐7 CO2 293 5.08 0.113

ZIF‐7 CO2 293 7.85 0.145

ZIF‐7 CO2 293 11.80 0.182

ZIF‐7 CO2 293 13.78 0.199

ZIF‐7 CO2 293 16.96 0.225

ZIF‐7 CO2 293 19.96 0.249

ZIF‐7 CO2 293 22.93 0.274

ZIF‐7 CO2 293 25.96 0.300

ZIF‐7 CO2 293 28.94 0.329

ZIF‐7 CO2 293 31.97 0.361

ZIF‐7 CO2 293 34.95 0.396

ZIF‐7 CO2 293 37.90 0.438

ZIF‐7 CO2 293 40.89 0.493

ZIF‐7 CO2 293 43.80 0.568

ZIF‐7 CO2 293 46.58 0.689


ZIF‐7 CO2 293 49.77 0.876

ZIF‐7 CO2 293 54.35 1.084

ZIF‐7 CO2 293 56.54 1.151

ZIF‐7 CO2 293 59.39 1.223

ZIF‐7 CO2 293 82.69 1.519

ZIF‐7 CO2 293 99.40 1.582

ZIF‐7 CO2 293 79.81 1.540

ZIF‐7 CO2 293 60.47 1.480

ZIF‐7 CO2 293 54.91 1.456

ZIF‐7 CO2 293 52.79 1.446

ZIF‐7 CO2 293 50.01 1.430

ZIF‐7 CO2 293 47.03 1.409

ZIF‐7 CO2 293 44.08 1.379

ZIF‐7 CO2 293 41.07 1.337

ZIF‐7 CO2 293 38.15 1.279

ZIF‐7 CO2 293 35.32 1.189

ZIF‐7 CO2 293 32.59 1.039

ZIF‐7 CO2 293 28.61 0.675

ZIF‐7 CO2 293 23.25 0.387

ZIF‐7 CO2 293 21.97 0.364

ZIF‐7 CO2 293 19.69 0.324

ZIF‐7 CO2 293 16.91 0.279

ZIF‐7 CO2 293 14.00 0.238

ZIF‐7 CO2 293 11.03 0.201

ZIF‐7 CO2 293 8.12 0.166

ZIF‐7 CO2 293 5.15 0.126


Table E5. Data points extracted graphically from McDonald et al. [64]


mmen‐Fe2(dobpdc) CO2 298 0.00 0.04






































mmen‐Fe2(dobpdc) CO2 298 100.21 3.70

mmen‐Fe2(dobpdc) CO2 298 105.29 3.76

mmen‐Fe2(dobpdc) CO2 298 110.30 3.79






























mmen‐Fe2(dobpdc) CO2 313 100.20 3.22

mmen‐Fe2(dobpdc) CO2 313 105.28 3.26

mmen‐Fe2(dobpdc) CO2 313 110.07 3.28





























mmen‐Fe2(dobpdc) CO2 323 100.35 3.06

mmen‐Fe2(dobpdc) CO2 323 105.35 3.10

mmen‐Fe2(dobpdc) CO2 323 110.22 3.13

mmen‐CO2(dobpdc) CO2 298 0.29 0.06













































mmen‐CO2(dobpdc) CO2 298 102.38 3.77

mmen‐CO2(dobpdc) CO2 298 104.77 3.79

mmen‐CO2(dobpdc) CO2 298 109.68 3.83











































mmen‐CO2(dobpdc) CO2 313 102.31 3.30

mmen‐CO2(dobpdc) CO2 313 104.77 3.32

mmen‐CO2(dobpdc) CO2 313 107.30 3.34

mmen‐CO2(dobpdc) CO2 313 109.68 3.35











































mmen‐CO2(dobpdc) CO2 323 102.17 2.86

mmen‐CO2(dobpdc) CO2 323 104.91 2.92

mmen‐CO2(dobpdc) CO2 323 109.68 2.99

E6 – Data table for all experimental isotherm data presented in Chapter 3

Table E6. Data table for all experimental isotherm data presented in Chapter 3

Adsorbent Gas Temperature (°C) Pressure (kPa) Quantity Adsorbed (mmol/g)

TMAY ‐ CBV100 CH4 ‐15 1.41 0.034

TMAY ‐ CBV100 CH4 ‐15 2.67 0.056

TMAY ‐ CBV100 CH4 ‐15 4.88 0.094

TMAY ‐ CBV100 CH4 ‐15 9.74 0.178

TMAY ‐ CBV100 CH4 ‐15 19.64 0.339

TMAY ‐ CBV100 CH4 ‐15 29.98 0.490

TMAY ‐ CBV100 CH4 ‐15 40.19 0.623

TMAY ‐ CBV100 CH4 ‐15 50.16 0.741

TMAY ‐ CBV100 CH4 ‐15 60.01 0.850

TMAY ‐ CBV100 CH4 ‐15 70.18 0.954

TMAY ‐ CBV100 CH4 ‐15 80.12 1.045

TMAY ‐ CBV100 CH4 ‐15 89.97 1.129

TMAY ‐ CBV100 CH4 ‐15 99.98 1.227

TMAY ‐ CBV100 CH4 0 0.04 0.000

TMAY ‐ CBV100 CH4 0 0.12 0.001

TMAY ‐ CBV100 CH4 0 0.21 0.002

TMAY ‐ CBV100 CH4 0 0.29 0.003

TMAY ‐ CBV100 CH4 0 0.45 0.005

TMAY ‐ CBV100 CH4 0 0.68 0.008

TMAY ‐ CBV100 CH4 0 0.82 0.010

TMAY ‐ CBV100 CH4 0 1.02 0.012

TMAY ‐ CBV100 CH4 0 1.21 0.014

TMAY ‐ CBV100 CH4 0 1.50 0.018

TMAY ‐ CBV100 CH4 0 1.86 0.021

TMAY ‐ CBV100 CH4 0 2.20 0.026

TMAY ‐ CBV100 CH4 0 2.71 0.033


TMAY ‐ CBV100 CH4 0 3.32 0.039

TMAY ‐ CBV100 CH4 0 4.03 0.047

TMAY ‐ CBV100 CH4 0 4.96 0.059

TMAY ‐ CBV100 CH4 0 6.06 0.073

TMAY ‐ CBV100 CH4 0 7.37 0.089

TMAY ‐ CBV100 CH4 0 8.98 0.108

TMAY ‐ CBV100 CH4 0 10.96 0.131

TMAY ‐ CBV100 CH4 0 13.39 0.160

TMAY ‐ CBV100 CH4 0 16.31 0.193

TMAY ‐ CBV100 CH4 0 19.94 0.233

TMAY ‐ CBV100 CH4 0 24.29 0.281

TMAY ‐ CBV100 CH4 0 29.69 0.337

TMAY ‐ CBV100 CH4 0 36.18 0.403

TMAY ‐ CBV100 CH4 0 44.26 0.479

TMAY ‐ CBV100 CH4 0 54.14 0.560

TMAY ‐ CBV100 CH4 0 65.92 0.646

TMAY ‐ CBV100 CH4 0 80.65 0.745

TMAY ‐ CBV100 CH4 0 98.21 0.857

TMAY ‐ CBV100 CH4 30 5.43 0.039

TMAY ‐ CBV100 CH4 30 11.19 0.075

TMAY ‐ CBV100 CH4 30 19.93 0.129

TMAY ‐ CBV100 CH4 30 30.00 0.182

TMAY ‐ CBV100 CH4 30 39.87 0.234

TMAY ‐ CBV100 CH4 30 49.97 0.281

TMAY ‐ CBV100 CH4 30 60.12 0.326

TMAY ‐ CBV100 CH4 30 70.05 0.368

TMAY ‐ CBV100 CH4 30 80.13 0.411

TMAY ‐ CBV100 CH4 30 90.11 0.453

TMAY ‐ CBV100 CH4 30 100.04 0.499

TMAY ‐ CBV100 N2 ‐15 1.29 0.008

TMAY ‐ CBV100 N2 ‐15 2.80 0.012

TMAY ‐ CBV100 N2 ‐15 5.60 0.019

TMAY ‐ CBV100 N2 ‐15 9.99 0.031

TMAY ‐ CBV100 N2 ‐15 19.79 0.056

TMAY ‐ CBV100 N2 ‐15 29.94 0.082

TMAY ‐ CBV100 N2 ‐15 40.06 0.107

TMAY ‐ CBV100 N2 ‐15 50.04 0.132

TMAY ‐ CBV100 N2 ‐15 60.05 0.156

TMAY ‐ CBV100 N2 ‐15 70.03 0.179

TMAY ‐ CBV100 N2 ‐15 80.05 0.202

TMAY ‐ CBV100 N2 ‐15 89.99 0.225

TMAY ‐ CBV100 N2 ‐15 100.04 0.247

TMAY ‐ CBV100 N2 0 0.03 0.000


TMAY ‐ CBV100 N2 0 0.12 0.000

TMAY ‐ CBV100 N2 0 0.21 0.000

TMAY ‐ CBV100 N2 0 0.30 0.001

TMAY ‐ CBV100 N2 0 0.45 0.001

TMAY ‐ CBV100 N2 0 0.72 0.001

TMAY ‐ CBV100 N2 0 0.97 0.002

TMAY ‐ CBV100 N2 0 1.04 0.002

TMAY ‐ CBV100 N2 0 1.43 0.002

TMAY ‐ CBV100 N2 0 1.51 0.002

TMAY ‐ CBV100 N2 0 2.11 0.004

TMAY ‐ CBV100 N2 0 2.23 0.004

TMAY ‐ CBV100 N2 0 3.13 0.006

TMAY ‐ CBV100 N2 0 3.35 0.007

TMAY ‐ CBV100 N2 0 4.08 0.009

TMAY ‐ CBV100 N2 0 4.97 0.012

TMAY ‐ CBV100 N2 0 6.06 0.014

TMAY ‐ CBV100 N2 0 7.36 0.017

TMAY ‐ CBV100 N2 0 8.99 0.021

TMAY ‐ CBV100 N2 0 10.97 0.027

TMAY ‐ CBV100 N2 0 13.40 0.033

TMAY ‐ CBV100 N2 0 16.33 0.040

TMAY ‐ CBV100 N2 0 19.94 0.049

TMAY ‐ CBV100 N2 0 24.33 0.060

TMAY ‐ CBV100 N2 0 29.70 0.073

TMAY ‐ CBV100 N2 0 36.27 0.088

TMAY ‐ CBV100 N2 0 44.25 0.106

TMAY ‐ CBV100 N2 0 54.03 0.128

TMAY ‐ CBV100 N2 0 65.97 0.153

TMAY ‐ CBV100 N2 0 80.50 0.183

TMAY ‐ CBV100 N2 0 98.30 0.217

TMAY ‐ CBV100 N2 30 1.35 0.003

TMAY ‐ CBV100 N2 30 2.85 0.004

TMAY ‐ CBV100 N2 30 5.63 0.007

TMAY ‐ CBV100 N2 30 10.00 0.011

TMAY ‐ CBV100 N2 30 19.83 0.021

TMAY ‐ CBV100 N2 30 29.97 0.032

TMAY ‐ CBV100 N2 30 40.11 0.042

TMAY ‐ CBV100 N2 30 50.18 0.052

TMAY ‐ CBV100 N2 30 59.95 0.062

TMAY ‐ CBV100 N2 30 70.18 0.072

TMAY ‐ CBV100 N2 30 80.12 0.081

TMAY ‐ CBV100 N2 30 90.05 0.091

TMAY ‐ CBV100 N2 30 99.95 0.101


TMAY ‐ CBV720 CH4 ‐15 0.13 0.000

TMAY ‐ CBV720 CH4 ‐15 0.21 0.001

TMAY ‐ CBV720 CH4 ‐15 0.30 0.001

TMAY ‐ CBV720 CH4 ‐15 0.45 0.002

TMAY ‐ CBV720 CH4 ‐15 0.71 0.002

TMAY ‐ CBV720 CH4 ‐15 0.98 0.006

TMAY ‐ CBV720 CH4 ‐15 1.08 0.005

TMAY ‐ CBV720 CH4 ‐15 1.41 0.008

TMAY ‐ CBV720 CH4 ‐15 1.50 0.009

TMAY ‐ CBV720 CH4 ‐15 2.12 0.012

TMAY ‐ CBV720 CH4 ‐15 2.26 0.013

TMAY ‐ CBV720 CH4 ‐15 2.73 0.015

TMAY ‐ CBV720 CH4 ‐15 3.79 0.021

TMAY ‐ CBV720 CH4 ‐15 4.05 0.023

TMAY ‐ CBV720 CH4 ‐15 4.96 0.026

TMAY ‐ CBV720 CH4 ‐15 6.89 0.036

TMAY ‐ CBV720 CH4 ‐15 7.41 0.039

TMAY ‐ CBV720 CH4 ‐15 9.03 0.048

TMAY ‐ CBV720 CH4 ‐15 10.94 0.056

TMAY ‐ CBV720 CH4 ‐15 13.34 0.067

TMAY ‐ CBV720 CH4 ‐15 16.29 0.083

TMAY ‐ CBV720 CH4 ‐15 19.92 0.099

TMAY ‐ CBV720 CH4 ‐15 24.33 0.118

TMAY ‐ CBV720 CH4 ‐15 29.72 0.140

TMAY ‐ CBV720 CH4 ‐15 36.28 0.166

TMAY ‐ CBV720 CH4 ‐15 44.26 0.198

TMAY ‐ CBV720 CH4 ‐15 54.05 0.231

TMAY ‐ CBV720 CH4 ‐15 65.96 0.270

TMAY ‐ CBV720 CH4 ‐15 80.54 0.312

TMAY ‐ CBV720 CH4 ‐15 98.29 0.358

TMAY ‐ CBV720 CH4 0 0.13 0.000

TMAY ‐ CBV720 CH4 0 0.20 0.000

TMAY ‐ CBV720 CH4 0 0.30 0.001

TMAY ‐ CBV720 CH4 0 0.45 0.001

TMAY ‐ CBV720 CH4 0 0.69 0.002

TMAY ‐ CBV720 CH4 0 0.96 0.002

TMAY ‐ CBV720 CH4 0 1.16 0.003

TMAY ‐ CBV720 CH4 0 1.27 0.005

TMAY ‐ CBV720 CH4 0 1.51 0.006

TMAY ‐ CBV720 CH4 0 2.10 0.007

TMAY ‐ CBV720 CH4 0 2.55 0.010

TMAY ‐ CBV720 CH4 0 2.73 0.010

TMAY ‐ CBV720 CH4 0 3.80 0.014


TMAY ‐ CBV720 CH4 0 4.05 0.016

TMAY ‐ CBV720 CH4 0 4.95 0.018

TMAY ‐ CBV720 CH4 0 6.89 0.025

TMAY ‐ CBV720 CH4 0 7.35 0.028

TMAY ‐ CBV720 CH4 0 8.98 0.032

TMAY ‐ CBV720 CH4 0 12.53 0.045

TMAY ‐ CBV720 CH4 0 13.36 0.048

TMAY ‐ CBV720 CH4 0 16.33 0.058

TMAY ‐ CBV720 CH4 0 19.88 0.070

TMAY ‐ CBV720 CH4 0 24.27 0.082

TMAY ‐ CBV720 CH4 0 29.74 0.099

TMAY ‐ CBV720 CH4 0 36.20 0.117

TMAY ‐ CBV720 CH4 0 44.36 0.141

TMAY ‐ CBV720 CH4 0 54.18 0.167

TMAY ‐ CBV720 CH4 0 66.01 0.194

TMAY ‐ CBV720 CH4 0 80.52 0.223

TMAY ‐ CBV720 CH4 0 98.32 0.258

TMAY ‐ CBV720 CH4 30 0.69 0.001

TMAY ‐ CBV720 CH4 30 0.84 0.001

TMAY ‐ CBV720 CH4 30 1.16 0.003

TMAY ‐ CBV720 CH4 30 1.29 0.004

TMAY ‐ CBV720 CH4 30 1.73 0.002

TMAY ‐ CBV720 CH4 30 2.10 0.003

TMAY ‐ CBV720 CH4 30 2.24 0.005

TMAY ‐ CBV720 CH4 30 2.73 0.005

TMAY ‐ CBV720 CH4 30 3.81 0.006

TMAY ‐ CBV720 CH4 30 4.06 0.008

TMAY ‐ CBV720 CH4 30 4.94 0.007

TMAY ‐ CBV720 CH4 30 6.91 0.015

TMAY ‐ CBV720 CH4 30 7.37 0.015

TMAY ‐ CBV720 CH4 30 10.34 0.024

TMAY ‐ CBV720 CH4 30 10.96 0.026

TMAY ‐ CBV720 CH4 30 13.37 0.030

TMAY ‐ CBV720 CH4 30 18.24 0.044

TMAY ‐ CBV720 CH4 30 19.91 0.047

TMAY ‐ CBV720 CH4 30 24.35 0.057

TMAY ‐ CBV720 CH4 30 29.72 0.069

TMAY ‐ CBV720 CH4 30 36.31 0.085

TMAY ‐ CBV720 CH4 30 44.23 0.102

TMAY ‐ CBV720 CH4 30 54.04 0.121

TMAY ‐ CBV720 CH4 30 66.05 0.144

TMAY ‐ CBV720 CH4 30 80.56 0.167

TMAY ‐ CBV720 CH4 30 98.35 0.191


TMAY ‐ CBV720 N2 ‐15 2.25 0.001

TMAY ‐ CBV720 N2 ‐15 3.16 0.002

TMAY ‐ CBV720 N2 ‐15 3.32 0.003

TMAY ‐ CBV720 N2 ‐15 4.04 0.004

TMAY ‐ CBV720 N2 ‐15 5.66 0.006

TMAY ‐ CBV720 N2 ‐15 6.92 0.006

TMAY ‐ CBV720 N2 ‐15 8.42 0.008

TMAY ‐ CBV720 N2 ‐15 8.98 0.007

TMAY ‐ CBV720 N2 ‐15 12.56 0.013

TMAY ‐ CBV720 N2 ‐15 13.39 0.015

TMAY ‐ CBV720 N2 ‐15 16.33 0.018

TMAY ‐ CBV720 N2 ‐15 21.88 0.026

TMAY ‐ CBV720 N2 ‐15 24.30 0.031

TMAY ‐ CBV720 N2 ‐15 29.68 0.038

TMAY ‐ CBV720 N2 ‐15 36.27 0.048

TMAY ‐ CBV720 N2 ‐15 44.19 0.058

TMAY ‐ CBV720 N2 ‐15 54.04 0.073

TMAY ‐ CBV720 N2 ‐15 66.01 0.089

TMAY ‐ CBV720 N2 ‐15 80.56 0.106

TMAY ‐ CBV720 N2 ‐15 98.31 0.127

TMAY ‐ CBV720 N2 30 6.88 0.002

TMAY ‐ CBV720 N2 30 7.36 0.001

TMAY ‐ CBV720 N2 30 10.31 0.004

TMAY ‐ CBV720 N2 30 10.96 0.006

TMAY ‐ CBV720 N2 30 13.41 0.008

TMAY ‐ CBV720 N2 30 18.35 0.015

TMAY ‐ CBV720 N2 30 19.91 0.017

TMAY ‐ CBV720 N2 30 24.31 0.019

TMAY ‐ CBV720 N2 30 31.59 0.035

TMAY ‐ CBV720 N2 30 36.24 0.041

TMAY ‐ CBV720 N2 30 44.17 0.047

TMAY ‐ CBV720 N2 30 54.03 0.055

TMAY ‐ CBV720 N2 30 65.95 0.062

TMAY ‐ CBV720 N2 30 80.45 0.070

TMAY ‐ CBV720 N2 30 98.34 0.081

NaY CH4 0 1.28 0.013

NaY CH4 0 2.76 0.022

NaY CH4 0 4.90 0.035

NaY CH4 0 9.85 0.066

NaY CH4 0 19.75 0.126

NaY CH4 0 30.07 0.186

NaY CH4 0 40.07 0.243

NaY CH4 0 50.14 0.301


NaY CH4 0 60.07 0.355

NaY CH4 0 70.00 0.410

NaY CH4 0 80.14 0.465

NaY CH4 0 90.06 0.518

NaY CH4 0 100.06 0.571

NaY CH4 30 1.34 0.010

NaY CH4 30 2.82 0.015

NaY CH4 30 4.93 0.023

NaY CH4 30 9.91 0.042

NaY CH4 30 19.84 0.079

NaY CH4 30 29.95 0.117

NaY CH4 30 39.98 0.152

NaY CH4 30 49.93 0.186

NaY CH4 30 60.11 0.221

NaY CH4 30 70.19 0.256

NaY CH4 30 80.23 0.290

NaY CH4 30 90.16 0.322

NaY CH4 30 100.13 0.355

NaY N2 0 1.30 0.012

NaY N2 0 2.81 0.018

NaY N2 0 5.59 0.030

NaY N2 0 9.96 0.048

NaY N2 0 19.81 0.089

NaY N2 0 29.98 0.131

NaY N2 0 39.94 0.171

NaY N2 0 49.99 0.212

NaY N2 0 60.07 0.252

NaY N2 0 69.97 0.291

NaY N2 0 80.05 0.330

NaY N2 0 90.09 0.369

NaY N2 0 100.02 0.407

NaY N2 30 1.36 0.004

NaY N2 30 2.81 0.008

NaY N2 30 4.97 0.013

NaY N2 30 9.91 0.024

NaY N2 30 19.84 0.047

NaY N2 30 29.95 0.070

NaY N2 30 40.13 0.093

NaY N2 30 49.96 0.115

NaY N2 30 60.15 0.137

NaY N2 30 70.15 0.159

NaY N2 30 80.21 0.180

NaY N2 30 90.10 0.201


NaY N2 30 100.05 0.222

TMAY Not Degassed CH4 0 1.36 0.003


























TMA‐Mg‐Y CH4 0 1.42 0.016

TMA‐Mg‐Y CH4 0 2.69 0.026

TMA‐Mg‐Y CH4 0 4.84 0.044

TMA‐Mg‐Y CH4 0 9.75 0.083

TMA‐Mg‐Y CH4 0 19.67 0.160

TMA‐Mg‐Y CH4 0 29.97 0.237

TMA‐Mg‐Y CH4 0 40.12 0.312

TMA‐Mg‐Y CH4 0 50.13 0.382

TMA‐Mg‐Y CH4 0 60.18 0.450

TMA‐Mg‐Y CH4 0 70.08 0.517

TMA‐Mg‐Y CH4 0 80.19 0.583

TMA‐Mg‐Y CH4 0 89.98 0.645

TMA‐Mg‐Y CH4 0 100.07 0.707

TMA‐Mg‐Y CH4 30 1.32 0.007

TMA‐Mg‐Y CH4 30 2.77 0.013

TMA‐Mg‐Y CH4 30 4.93 0.022


TMA‐Mg‐Y CH4 30 9.87 0.041

TMA‐Mg‐Y CH4 30 19.75 0.080

TMA‐Mg‐Y CH4 30 30.07 0.120

TMA‐Mg‐Y CH4 30 40.13 0.158

TMA‐Mg‐Y CH4 30 50.10 0.195

TMA‐Mg‐Y CH4 30 60.13 0.232

TMA‐Mg‐Y CH4 30 70.08 0.267

TMA‐Mg‐Y CH4 30 79.99 0.302

TMA‐Mg‐Y CH4 30 90.10 0.337

TMA‐Mg‐Y CH4 30 100.09 0.372

TMA‐Mg‐Y N2 0 1.34 0.006

TMA‐Mg‐Y N2 0 2.84 0.010

TMA‐Mg‐Y N2 0 4.94 0.015

TMA‐Mg‐Y N2 0 9.92 0.028

TMA‐Mg‐Y N2 0 19.85 0.054

TMA‐Mg‐Y N2 0 29.96 0.079

TMA‐Mg‐Y N2 0 39.95 0.104

TMA‐Mg‐Y N2 0 49.95 0.128

TMA‐Mg‐Y N2 0 60.04 0.152

TMA‐Mg‐Y N2 0 70.10 0.176

TMA‐Mg‐Y N2 0 80.06 0.199

TMA‐Mg‐Y N2 0 90.04 0.222

TMA‐Mg‐Y N2 0 100.04 0.245

TMA‐Mg‐Y N2 30 1.38 0.004

TMA‐Mg‐Y N2 30 2.85 0.006

TMA‐Mg‐Y N2 30 5.67 0.010

TMA‐Mg‐Y N2 30 9.93 0.016

TMA‐Mg‐Y N2 30 19.89 0.030

TMA‐Mg‐Y N2 30 29.95 0.043

TMA‐Mg‐Y N2 30 39.95 0.057

TMA‐Mg‐Y N2 30 50.07 0.071

TMA‐Mg‐Y N2 30 60.07 0.084

TMA‐Mg‐Y N2 30 70.06 0.097

TMA‐Mg‐Y N2 30 80.07 0.110

TMA‐Mg‐Y N2 30 90.03 0.123

TMA‐Mg‐Y N2 30 100.01 0.137

TMA‐Base‐NaY CH4 0 1.36 0.020













TMA‐Base‐NaY CH4 0 100.07 0.857

TMA‐Base‐NaY N2 0 1.33 0.006

TMA‐Base‐NaY N2 0 2.80 0.010

TMA‐Base‐NaY N2 0 4.90 0.015

TMA‐Base‐NaY N2 0 9.90 0.028

TMA‐Base‐NaY N2 0 19.86 0.053

TMA‐Base‐NaY N2 0 29.99 0.078

TMA‐Base‐NaY N2 0 39.96 0.103

TMA‐Base‐NaY N2 0 50.12 0.127

TMA‐Base‐NaY N2 0 60.14 0.150

TMA‐Base‐NaY N2 0 70.02 0.173

TMA‐Base‐NaY N2 0 80.12 0.196

TMA‐Base‐NaY N2 0 90.10 0.217

TMA‐Base‐NaY N2 0 100.03 0.239

TMAOH‐NaY CH4 0 1.38 0.020

TMAOH‐NaY CH4 0 2.55 0.034

TMAOH‐NaY CH4 0 4.96 0.063

TMAOH‐NaY CH4 0 9.70 0.118

TMAOH‐NaY CH4 0 19.39 0.226

TMAOH‐NaY CH4 0 29.98 0.336

TMAOH‐NaY CH4 0 40.08 0.433

TMAOH‐NaY CH4 0 49.95 0.523

TMAOH‐NaY CH4 0 60.13 0.611

TMAOH‐NaY CH4 0 70.07 0.692

TMAOH‐NaY CH4 0 80.16 0.769

TMAOH‐NaY CH4 0 90.08 0.841

TMAOH‐NaY CH4 0 100.01 0.911

TMAOH‐NaY CH4 30 1.24 0.009

TMAOH‐NaY CH4 30 2.69 0.017

TMAOH‐NaY CH4 30 4.90 0.029

TMAOH‐NaY CH4 30 9.80 0.056

TMAOH‐NaY CH4 30 19.66 0.108

TMAOH‐NaY CH4 30 29.96 0.160

TMAOH‐NaY CH4 30 40.10 0.210

TMAOH‐NaY CH4 30 50.08 0.257

TMAOH‐NaY CH4 30 60.11 0.304

TMAOH‐NaY CH4 30 70.13 0.348

TMAOH‐NaY CH4 30 79.95 0.390


TMAOH‐NaY CH4 30 90.05 0.432

TMAOH‐NaY CH4 30 100.00 0.472

TMAOH‐NaY N2 0 1.34 0.005

TMAOH‐NaY N2 0 2.79 0.009

TMAOH‐NaY N2 0 4.88 0.014

TMAOH‐NaY N2 0 9.93 0.027

TMAOH‐NaY N2 0 19.86 0.052

TMAOH‐NaY N2 0 29.97 0.076

TMAOH‐NaY N2 0 39.93 0.099

TMAOH‐NaY N2 0 50.21 0.123

TMAOH‐NaY N2 0 60.11 0.145

TMAOH‐NaY N2 0 70.17 0.168

TMAOH‐NaY N2 0 80.21 0.190

TMAOH‐NaY N2 0 90.20 0.211

TMAOH‐NaY N2 0 100.10 0.233

TMAOH‐NaY N2 30 1.37 0.004

TMAOH‐NaY N2 30 2.84 0.006

TMAOH‐NaY N2 30 4.96 0.009

TMAOH‐NaY N2 30 9.94 0.015

TMAOH‐NaY N2 30 19.91 0.028

TMAOH‐NaY N2 30 29.97 0.041

TMAOH‐NaY N2 30 39.94 0.053

TMAOH‐NaY N2 30 50.08 0.066

TMAOH‐NaY N2 30 60.06 0.078

TMAOH‐NaY N2 30 70.07 0.090

TMAOH‐NaY N2 30 80.05 0.102

TMAOH‐NaY N2 30 90.04 0.113

TMAOH‐NaY N2 30 100.04 0.126

TMAOH‐HY600 CH4 0 1.26 0.010

TMAOH‐HY600 CH4 0 2.71 0.018

TMAOH‐HY600 CH4 0 4.90 0.033

TMAOH‐HY600 CH4 0 9.87 0.062

TMAOH‐HY600 CH4 0 19.71 0.116

TMAOH‐HY600 CH4 0 29.98 0.169

TMAOH‐HY600 CH4 0 40.05 0.219

TMAOH‐HY600 CH4 0 50.15 0.267

TMAOH‐HY600 CH4 0 60.16 0.312

TMAOH‐HY600 CH4 0 69.97 0.354

TMAOH‐HY600 CH4 0 80.07 0.396

TMAOH‐HY600 CH4 0 90.09 0.436

TMAOH‐HY600 CH4 0 100.05 0.474

TMAOH‐HY600 N2 0 1.38 0.004

TMAOH‐HY600 N2 0 2.86 0.005


TMAOH‐HY600 N2 0 5.68 0.009

TMAOH‐HY600 N2 0 10.04 0.015

TMAOH‐HY600 N2 0 19.89 0.027

TMAOH‐HY600 N2 0 29.97 0.040

TMAOH‐HY600 N2 0 40.04 0.052

TMAOH‐HY600 N2 0 50.04 0.064

TMAOH‐HY600 N2 0 60.07 0.076

TMAOH‐HY600 N2 0 70.06 0.088

TMAOH‐HY600 N2 0 80.07 0.100

TMAOH‐HY600 N2 0 90.11 0.111

TMAOH‐HY600 N2 0 100.04 0.123

TMAOH‐HY400 CH4 0 1.37 0.016

TMAOH‐HY400 CH4 0 2.63 0.027

TMAOH‐HY400 CH4 0 4.89 0.047

TMAOH‐HY400 CH4 0 9.72 0.089

TMAOH‐HY400 CH4 0 19.55 0.169

TMAOH‐HY400 CH4 0 30.10 0.250

TMAOH‐HY400 CH4 0 39.95 0.322

TMAOH‐HY400 CH4 0 50.13 0.391

TMAOH‐HY400 CH4 0 60.07 0.455

TMAOH‐HY400 CH4 0 70.13 0.516

TMAOH‐HY400 CH4 0 80.13 0.573

TMAOH‐HY400 CH4 0 89.91 0.627

TMAOH‐HY400 CH4 0 100.08 0.680

TMAOH‐HY400 N2 0 1.25 0.005

TMAOH‐HY400 N2 0 2.82 0.008

TMAOH‐HY400 N2 0 5.62 0.013

TMAOH‐HY400 N2 0 10.00 0.021

TMAOH‐HY400 N2 0 19.81 0.037

TMAOH‐HY400 N2 0 29.98 0.054

TMAOH‐HY400 N2 0 39.96 0.071

TMAOH‐HY400 N2 0 50.08 0.087

TMAOH‐HY400 N2 0 59.99 0.102

TMAOH‐HY400 N2 0 70.07 0.117

TMAOH‐HY400 N2 0 80.04 0.133

TMAOH‐HY400 N2 0 90.04 0.148

TMAOH‐HY400 N2 0 100.04 0.162

TMAOH‐HY400 CH4 30 1.24 0.008

TMAOH‐HY400 CH4 30 2.73 0.014

TMAOH‐HY400 CH4 30 4.85 0.023

TMAOH‐HY400 CH4 30 9.82 0.043

TMAOH‐HY400 CH4 30 19.71 0.082

TMAOH‐HY400 CH4 30 29.98 0.121


TMAOH‐HY400 CH4 30 39.94 0.157

TMAOH‐HY400 CH4 30 50.10 0.194

TMAOH‐HY400 CH4 30 60.11 0.228

TMAOH‐HY400 CH4 30 70.10 0.260

TMAOH‐HY400 CH4 30 80.14 0.293

TMAOH‐HY400 CH4 30 89.99 0.324

TMAOH‐HY400 CH4 30 100.14 0.355

TMAOH‐HY400 N2 30 1.34 0.004

TMAOH‐HY400 N2 30 2.86 0.005

TMAOH‐HY400 N2 30 5.69 0.007

TMAOH‐HY400 N2 30 10.00 0.011

TMAOH‐HY400 N2 30 19.91 0.020

TMAOH‐HY400 N2 30 29.96 0.029

TMAOH‐HY400 N2 30 39.95 0.038

TMAOH‐HY400 N2 30 49.94 0.046

TMAOH‐HY400 N2 30 60.14 0.055

TMAOH‐HY400 N2 30 70.06 0.064

TMAOH‐HY400 N2 30 80.00 0.072

TMAOH‐HY400 N2 30 90.10 0.080

TMAOH‐HY400 N2 30 100.11 0.088

Carbon (Norit RB3) CH4 0 1.23 0.045













Carbon (Norit RB3) N2 0 1.39 0.010













Carbon (Norit RB3) N2 0 100.02 0.454

Acid‐Carbon1 CH4 0 1.28 0.032












Acid‐Carbon1 CH4 0 100.11 1.009

Acid‐Carbon1 N2 0 1.27 0.008

Acid‐Carbon1 N2 0 2.75 0.015

Acid‐Carbon1 N2 0 4.94 0.024

Acid‐Carbon1 N2 0 9.86 0.045

Acid‐Carbon1 N2 0 19.80 0.086

Acid‐Carbon1 N2 0 29.97 0.127

Acid‐Carbon1 N2 0 39.96 0.164

Acid‐Carbon1 N2 0 50.10 0.201

Acid‐Carbon1 N2 0 60.08 0.236

Acid‐Carbon1 N2 0 70.07 0.269

Acid‐Carbon1 N2 0 80.07 0.301

Acid‐Carbon1 N2 0 90.10 0.333

Acid‐Carbon1 N2 0 100.07 0.364

TMA‐Carbon1 CH4 0 1.29 0.026

TMA‐Carbon1 CH4 0 2.43 0.045

TMA‐Carbon1 CH4 0 4.91 0.082

TMA‐Carbon1 CH4 0 9.76 0.147

TMA‐Carbon1 CH4 0 19.58 0.261

TMA‐Carbon1 CH4 0 30.08 0.364

TMA‐Carbon1 CH4 0 39.88 0.450

TMA‐Carbon1 CH4 0 49.98 0.531

TMA‐Carbon1 CH4 0 60.07 0.606

TMA‐Carbon1 CH4 0 70.07 0.676

TMA‐Carbon1 CH4 0 79.96 0.740

TMA‐Carbon1 CH4 0 90.08 0.803

TMA‐Carbon1 CH4 0 100.03 0.862

TMA‐Carbon1 N2 0 1.31 0.006


TMA‐Carbon1 N2 0 2.75 0.011

TMA‐Carbon1 N2 0 4.96 0.018

TMA‐Carbon1 N2 0 9.88 0.035

TMA‐Carbon1 N2 0 19.79 0.067

TMA‐Carbon1 N2 0 29.97 0.098

TMA‐Carbon1 N2 0 40.06 0.128

TMA‐Carbon1 N2 0 50.08 0.156

TMA‐Carbon1 N2 0 60.11 0.183

TMA‐Carbon1 N2 0 70.05 0.210

TMA‐Carbon1 N2 0 80.07 0.235

TMA‐Carbon1 N2 0 90.05 0.260

TMA‐Carbon1 N2 0 100.03 0.285

TMA‐Carbon2 CH4 0 1.21 0.011

TMA‐Carbon2 CH4 0 2.67 0.021

TMA‐Carbon2 CH4 0 4.92 0.035

TMA‐Carbon2 CH4 0 9.86 0.062

TMA‐Carbon2 CH4 0 19.85 0.108

TMA‐Carbon2 CH4 0 29.92 0.146

TMA‐Carbon2 CH4 0 40.12 0.181

TMA‐Carbon2 CH4 0 50.20 0.211

TMA‐Carbon2 CH4 0 60.07 0.239

TMA‐Carbon2 CH4 0 70.17 0.265

TMA‐Carbon2 CH4 0 80.20 0.290

TMA‐Carbon2 CH4 0 90.09 0.312

TMA‐Carbon2 CH4 0 100.13 0.334

TMA‐Carbon2 N2 0 1.36 0.004

TMA‐Carbon2 N2 0 2.83 0.006

TMA‐Carbon2 N2 0 5.68 0.010

TMA‐Carbon2 N2 0 10.00 0.015

TMA‐Carbon2 N2 0 19.88 0.028

TMA‐Carbon2 N2 0 29.96 0.039

TMA‐Carbon2 N2 0 39.95 0.051

TMA‐Carbon2 N2 0 49.94 0.062

TMA‐Carbon2 N2 0 60.06 0.072

TMA‐Carbon2 N2 0 70.03 0.082

TMA‐Carbon2 N2 0 80.04 0.092

TMA‐Carbon2 N2 0 90.05 0.102

TMA‐Carbon2 N2 0 100.01 0.111

TMA‐Carbon3 CH4 0 1.35 0.047

TMA‐Carbon3 CH4 0 2.65 0.085

TMA‐Carbon3 CH4 0 4.93 0.145

TMA‐Carbon3 CH4 0 9.62 0.251

TMA‐Carbon3 CH4 0 19.44 0.433


TMA‐Carbon3 CH4 0 29.82 0.592

TMA‐Carbon3 CH4 0 39.73 0.724

TMA‐Carbon3 CH4 0 50.05 0.847

TMA‐Carbon3 CH4 0 60.09 0.956

TMA‐Carbon3 CH4 0 70.06 1.055

TMA‐Carbon3 CH4 0 80.06 1.148

TMA‐Carbon3 CH4 0 89.95 1.235

TMA‐Carbon3 CH4 0 100.09 1.319

TMA‐Carbon3 N2 0 1.27 0.009

TMA‐Carbon3 N2 0 2.71 0.017

TMA‐Carbon3 N2 0 4.96 0.030

TMA‐Carbon3 N2 0 9.83 0.057

TMA‐Carbon3 N2 0 19.76 0.108

TMA‐Carbon3 N2 0 30.04 0.158

TMA‐Carbon3 N2 0 39.94 0.204

TMA‐Carbon3 N2 0 50.06 0.248

TMA‐Carbon3 N2 0 60.07 0.290

TMA‐Carbon3 N2 0 70.13 0.331

TMA‐Carbon3 N2 0 80.05 0.370

TMA‐Carbon3 N2 0 90.04 0.407

TMA‐Carbon3 N2 0 100.03 0.444

TMA‐Carbon4 CH4 0 1.34 0.038

TMA‐Carbon4 CH4 0 2.65 0.066

TMA‐Carbon4 CH4 0 4.79 0.110

TMA‐Carbon4 CH4 0 10.06 0.207

TMA‐Carbon4 CH4 0 19.82 0.358

TMA‐Carbon4 CH4 0 29.82 0.489

TMA‐Carbon4 CH4 0 39.94 0.606

TMA‐Carbon4 CH4 0 49.98 0.711

TMA‐Carbon4 CH4 0 60.08 0.808

TMA‐Carbon4 CH4 0 69.97 0.896

TMA‐Carbon4 CH4 0 80.10 0.981

TMA‐Carbon4 CH4 0 90.11 1.061

TMA‐Carbon4 CH4 0 100.15 1.136

TMA‐Carbon4 N2 0 1.24 0.007

TMA‐Carbon4 N2 0 2.71 0.013

TMA‐Carbon4 N2 0 4.84 0.023

TMA‐Carbon4 N2 0 9.78 0.045

TMA‐Carbon4 N2 0 19.75 0.088

TMA‐Carbon4 N2 0 30.08 0.130

TMA‐Carbon4 N2 0 40.03 0.169

TMA‐Carbon4 N2 0 50.08 0.206

TMA‐Carbon4 N2 0 60.05 0.242


TMA‐Carbon4 N2 0 70.03 0.277

TMA‐Carbon4 N2 0 80.05 0.311

TMA‐Carbon4 N2 0 90.00 0.343

TMA‐Carbon4 N2 0 100.04 0.376

TMA‐AcidBaseCarbon CH4 0 1.32 0.040













TMA‐AcidBaseCarbon N2 0 1.30 0.011













Carbon (Norit R2030) CH4 0 1.22 0.047

Carbon (Norit R2030) CH4 0 2.57 0.087

Carbon (Norit R2030) CH4 0 5.22 0.157

Carbon (Norit R2030) CH4 0 9.95 0.263

Carbon (Norit R2030) CH4 0 19.61 0.434

Carbon (Norit R2030) CH4 0 29.87 0.581

Carbon (Norit R2030) CH4 0 39.81 0.702

Carbon (Norit R2030) CH4 0 49.96 0.810

Carbon (Norit R2030) CH4 0 60.09 0.908

Carbon (Norit R2030) CH4 0 70.12 0.995

Carbon (Norit R2030) CH4 0 80.09 1.075

Carbon (Norit R2030) CH4 0 90.13 1.151

Carbon (Norit R2030) CH4 0 100.17 1.221


Carbon (Norit R2030) N2 0 1.39 0.010

Carbon (Norit R2030) N2 0 2.59 0.017

Carbon (Norit R2030) N2 0 4.87 0.029

Carbon (Norit R2030) N2 0 9.69 0.055

Carbon (Norit R2030) N2 0 19.49 0.104

Carbon (Norit R2030) N2 0 30.00 0.153

Carbon (Norit R2030) N2 0 40.12 0.196

Carbon (Norit R2030) N2 0 50.01 0.237

Carbon (Norit R2030) N2 0 60.07 0.276

Carbon (Norit R2030) N2 0 70.12 0.314

Carbon (Norit R2030) N2 0 80.04 0.349

Carbon (Norit R2030) N2 0 90.10 0.383

Carbon (Norit R2030) N2 0 100.11 0.416

Base‐Carbon CH4 0 1.31 0.050













Base‐Carbon N2 0 1.23 0.010




Base‐Carbon N2 0 19.74 0.110

Base‐Carbon N2 0 29.97 0.159

Base‐Carbon N2 0 40.03 0.203

Base‐Carbon N2 0 50.07 0.244

Base‐Carbon N2 0 60.15 0.284

Base‐Carbon N2 0 70.13 0.321

Base‐Carbon N2 0 80.08 0.357

Base‐Carbon N2 0 90.14 0.391

Base‐Carbon N2 0 100.11 0.424

APS‐Carbon CH4 0 1.30 0.024





APS‐Carbon CH4 0 19.65 0.241

APS‐Carbon CH4 0 30.05 0.333

APS‐Carbon CH4 0 39.98 0.410

APS‐Carbon CH4 0 49.97 0.479

APS‐Carbon CH4 0 59.93 0.543

APS‐Carbon CH4 0 70.10 0.603

APS‐Carbon CH4 0 79.95 0.657

APS‐Carbon CH4 0 90.17 0.709

APS‐Carbon CH4 0 100.06 0.759

APS‐Carbon N2 0 1.32 0.006

APS‐Carbon N2 0 2.80 0.011

APS‐Carbon N2 0 4.96 0.017

APS‐Carbon N2 0 9.88 0.033

APS‐Carbon N2 0 19.82 0.062

APS‐Carbon N2 0 29.94 0.090

APS‐Carbon N2 0 39.93 0.117

APS‐Carbon N2 0 50.09 0.143

APS‐Carbon N2 0 60.07 0.168

APS‐Carbon N2 0 70.10 0.192

APS‐Carbon N2 0 80.03 0.215

APS‐Carbon N2 0 90.09 0.237

APS‐Carbon N2 0 99.99 0.261

TMA‐APS‐Carbon CH4 0 1.28 0.033













TMA‐APS‐Carbon N2 0 1.36 0.007














Carbon‐E‐TMAY1 CH4 0 1.38 0.030












Carbon‐E‐TMAY1 CH4 0 100.04 1.029





Carbon‐E‐TMAY1 CH4 30 19.78 0.134

Carbon‐E‐TMAY1 CH4 30 30.10 0.196

Carbon‐E‐TMAY1 CH4 30 40.14 0.252

Carbon‐E‐TMAY1 CH4 30 50.12 0.305

Carbon‐E‐TMAY1 CH4 30 60.01 0.356

Carbon‐E‐TMAY1 CH4 30 70.13 0.405

Carbon‐E‐TMAY1 CH4 30 80.07 0.450

Carbon‐E‐TMAY1 CH4 30 90.08 0.494

Carbon‐E‐TMAY1 CH4 30 100.11 0.537

Carbon‐E‐TMAY1 N2 0 1.34 0.009



Carbon‐E‐TMAY1 N2 0 10.00 0.033

Carbon‐E‐TMAY1 N2 0 19.85 0.059

Carbon‐E‐TMAY1 N2 0 29.95 0.085

Carbon‐E‐TMAY1 N2 0 40.14 0.111

Carbon‐E‐TMAY1 N2 0 49.95 0.136

Carbon‐E‐TMAY1 N2 0 60.05 0.160

Carbon‐E‐TMAY1 N2 0 70.05 0.183

Carbon‐E‐TMAY1 N2 0 80.05 0.206

Carbon‐E‐TMAY1 N2 0 90.04 0.228


Carbon‐E‐TMAY1 N2 0 100.02 0.250

Carbon‐E‐TMAY1 N2 30 1.38 0.006

Carbon‐E‐TMAY1 N2 30 2.90 0.008

Carbon‐E‐TMAY1 N2 30 5.72 0.011

Carbon‐E‐TMAY1 N2 30 9.94 0.017

Carbon‐E‐TMAY1 N2 30 19.88 0.031

Carbon‐E‐TMAY1 N2 30 29.99 0.044

Carbon‐E‐TMAY1 N2 30 39.92 0.057

Carbon‐E‐TMAY1 N2 30 50.09 0.070

Carbon‐E‐TMAY1 N2 30 60.07 0.083

Carbon‐E‐TMAY1 N2 30 70.10 0.095

Carbon‐E‐TMAY1 N2 30 80.01 0.107

Carbon‐E‐TMAY1 N2 30 90.04 0.120

Carbon‐E‐TMAY1 N2 30 100.04 0.131

E7 – Data table for all experimental isotherm data presented in Chapter 4

Table E7. Data table for all experimental isotherm data presented in Chapter 4

Adsorbent Gas Temperature (°C) Pressure (bar) Quantity Adsorbed (mol/kg)

NaY N2 0 0.03 0.000

NaY N2 0 1.50 0.098

NaY N2 0 2.96 0.198

NaY N2 0 4.59 0.325

NaY N2 0 6.05 0.422

NaY N2 0 7.69 0.518

NaY N2 0 9.15 0.607

NaY N2 0 10.52 0.703

NaY N2 0 12.05 0.810

NaY N2 0 13.54 0.896

NaY N2 0 15.11 0.993

NaY N2 0 16.70 1.066

NaY N2 0 18.23 1.153

NaY N2 0 19.79 1.270

NaY N2 0 21.24 1.356

NaY N2 0 22.76 1.451

NaY N2 0 24.36 1.519

NaY N2 0 25.89 1.597

NaY N2 0 27.50 1.668

NaY N2 0 28.94 1.752

NaY N2 0 30.52 1.846

NaY N2 0 32.04 1.933

NaY N2 0 33.63 1.980


NaY N2 0 35.06 2.022

NaY N2 15 0.04 0.006

NaY N2 15 1.53 0.096

NaY N2 15 3.10 0.158

NaY N2 15 4.64 0.229

NaY N2 15 6.28 0.294

NaY N2 15 7.66 0.369

NaY N2 15 9.12 0.426

NaY N2 15 10.59 0.481

NaY N2 15 12.18 0.540

NaY N2 15 13.66 0.602

NaY N2 15 15.14 0.691

NaY N2 15 16.73 0.756

NaY N2 15 18.25 0.808

NaY N2 15 19.77 0.883

NaY N2 15 21.32 0.979

NaY N2 15 22.88 1.022

NaY N2 15 24.41 1.076

NaY N2 15 25.87 1.145

NaY N2 15 27.43 1.219

NaY N2 15 28.87 1.268

NaY N2 15 30.36 1.304

NaY N2 15 31.86 1.331

NaY N2 15 33.40 1.391

NaY N2 15 34.86 1.472

NaY N2 15 36.39 1.508

NaY N2 30 1.68 0.057

NaY N2 30 3.21 0.108

NaY N2 30 4.71 0.145

NaY N2 30 6.30 0.209

NaY N2 30 7.77 0.263

NaY N2 30 9.30 0.319

NaY N2 30 10.77 0.381

NaY N2 30 12.39 0.416

NaY N2 30 14.03 0.468

NaY N2 30 15.63 0.528

NaY N2 30 17.29 0.545

NaY N2 30 18.87 0.601

NaY N2 30 20.46 0.656

NaY N2 30 22.08 0.670

NaY N2 30 23.67 0.720

NaY N2 30 25.18 0.734

NaY N2 30 26.69 0.740


NaY N2 30 28.24 0.787

NaY N2 30 29.74 0.791

NaY N2 30 31.29 0.800

NaY N2 30 32.76 0.864

NaY N2 30 34.38 0.874

NaY N2 30 35.88 0.892

NaY N2 30 37.48 0.906

NaY N2 30 39.07 0.929

NaY N2 30 40.61 0.970

NaY N2 30 42.15 0.981

NaY N2 30 43.71 0.993

C‐R2030 N2 0 0.03 0.002

C‐R2030 N2 0 1.11 0.465

C‐R2030 N2 0 2.32 0.837

C‐R2030 N2 0 3.48 1.108

C‐R2030 N2 0 4.82 1.378

C‐R2030 N2 0 6.08 1.603

C‐R2030 N2 0 7.29 1.797

C‐R2030 N2 0 8.72 1.996

C‐R2030 N2 0 10.03 2.174

C‐R2030 N2 0 11.49 2.339

C‐R2030 N2 0 12.86 2.490

C‐R2030 N2 0 14.22 2.631

C‐R2030 N2 0 15.70 2.773

C‐R2030 N2 0 17.08 2.896

C‐R2030 N2 0 18.54 3.022

C‐R2030 N2 0 20.02 3.140

C‐R2030 N2 0 21.42 3.243

C‐R2030 N2 0 22.93 3.358

C‐R2030 N2 0 24.45 3.462

C‐R2030 N2 0 25.98 3.564

C‐R2030 N2 0 27.50 3.659

C‐R2030 N2 0 29.19 3.777

C‐R2030 N2 0 30.75 3.872

C‐R2030 N2 0 32.17 3.953

C‐R2030 N2 0 33.71 4.034

C‐R2030 N2 0 35.21 4.094

C‐R2030 N2 15 1.20 0.367

C‐R2030 N2 15 2.45 0.668

C‐R2030 N2 15 3.65 0.899

C‐R2030 N2 15 4.87 1.113

C‐R2030 N2 15 6.34 1.332

C‐R2030 N2 15 7.62 1.507


C‐R2030 N2 15 8.89 1.656

C‐R2030 N2 15 10.17 1.799

C‐R2030 N2 15 11.54 1.952

C‐R2030 N2 15 12.99 2.089

C‐R2030 N2 15 14.46 2.225

C‐R2030 N2 15 15.97 2.342

C‐R2030 N2 15 17.38 2.446

C‐R2030 N2 15 18.88 2.545

C‐R2030 N2 15 20.40 2.646

C‐R2030 N2 15 21.92 2.736

C‐R2030 N2 15 23.48 2.824

C‐R2030 N2 15 25.02 2.915

C‐R2030 N2 15 26.53 2.982

C‐R2030 N2 15 28.09 3.059

C‐R2030 N2 15 29.57 3.137

C‐R2030 N2 15 30.86 3.185

C‐R2030 N2 15 32.42 3.262

C‐R2030 N2 15 33.87 3.320

C‐R2030 N2 15 35.34 3.368

C‐R2030 N2 30 1.24 0.266

C‐R2030 N2 30 2.64 0.517

C‐R2030 N2 30 4.05 0.746

C‐R2030 N2 30 5.37 0.915

C‐R2030 N2 30 6.70 1.083

C‐R2030 N2 30 8.00 1.218

C‐R2030 N2 30 9.62 1.378

C‐R2030 N2 30 10.95 1.512

C‐R2030 N2 30 12.44 1.622

C‐R2030 N2 30 13.92 1.730

C‐R2030 N2 30 15.46 1.835

C‐R2030 N2 30 16.99 1.938

C‐R2030 N2 30 18.55 2.026

C‐R2030 N2 30 20.06 2.123

C‐R2030 N2 30 21.61 2.197

C‐R2030 N2 30 23.05 2.269

C‐R2030 N2 30 24.56 2.337

C‐R2030 N2 30 26.13 2.414

C‐R2030 N2 30 27.71 2.479

C‐R2030 N2 30 29.14 2.534

C‐R2030 N2 30 30.57 2.584

C‐R2030 N2 30 32.09 2.640

C‐R2030 N2 30 32.10 2.633

C‐R2030 N2 30 32.09 2.638


C‐R2030 N2 30 33.57 2.685

C‐R2030 N2 30 35.07 2.715

CEZ N2 0 1.21 0.556

CEZ N2 0 2.38 1.010

CEZ N2 0 3.61 1.413

CEZ N2 0 4.89 1.733

CEZ N2 0 6.14 2.053

CEZ N2 0 7.40 2.304

CEZ N2 0 8.68 2.551

CEZ N2 0 9.95 2.765

CEZ N2 0 11.52 3.011

CEZ N2 0 12.96 3.203

CEZ N2 0 14.36 3.387

CEZ N2 0 15.84 3.548

CEZ N2 0 17.26 3.728

CEZ N2 0 18.78 3.876

CEZ N2 0 20.30 4.017

CEZ N2 0 21.84 4.151

CEZ N2 0 23.31 4.296

CEZ N2 0 24.77 4.373

CEZ N2 0 26.27 4.498

CEZ N2 0 27.75 4.620

CEZ N2 0 29.25 4.709

CEZ N2 0 30.80 4.779

CEZ N2 0 32.29 4.901

CEZ N2 0 33.76 5.004

CEZ N2 0 35.26 5.086

CEZ N2 0 36.81 5.183

CEZ N2 0 38.30 5.276

CEZ N2 0 39.76 5.328

CEZ N2 0 41.29 5.385

CEZ N2 0 42.87 5.430

CEZ N2 0 43.74 5.342

CEZ N2 0 45.23 5.456

CEZ N2 0 46.10 5.364

CEZ N2 0 47.52 5.414

CEZ N2 0 48.32 5.351

CEZ N2 0 49.84 5.446

CEZ N2 0 50.63 5.479

CEZ N2 0 52.12 5.536

CEZ N2 0 52.84 5.580

CEZ N2 0 54.36 5.574

CEZ N2 0 55.08 5.599


CEZ N2 0 56.56 5.662

CEZ N2 0 57.33 5.645

CEZ N2 0 58.84 5.675

CEZ N2 0 59.63 5.634

CEZ N2 0 60.94 5.711

CEZ N2 0 61.65 5.659

CEZ N2 0 62.73 5.637

CEZ N2 0 63.24 5.623

CEZ N2 0 64.18 5.594

CEZ N2 15 1.27 0.426

CEZ N2 15 2.54 0.783

CEZ N2 15 3.80 1.111

CEZ N2 15 5.20 1.421

CEZ N2 15 6.56 1.680

CEZ N2 15 7.90 1.903

CEZ N2 15 9.22 2.107

CEZ N2 15 10.54 2.299

CEZ N2 15 12.06 2.485

CEZ N2 15 13.56 2.630

CEZ N2 15 14.96 2.776

CEZ N2 15 16.38 2.914

CEZ N2 15 17.90 3.048

CEZ N2 15 19.39 3.149

CEZ N2 15 20.90 3.247

CEZ N2 15 22.41 3.347

CEZ N2 15 23.94 3.445

CEZ N2 15 25.44 3.544

CEZ N2 15 26.86 3.614

CEZ N2 15 28.32 3.680

CEZ N2 15 29.74 3.790

CEZ N2 15 31.19 3.841

CEZ N2 15 32.78 3.903

CEZ N2 15 34.18 3.985

CEZ N2 15 35.70 4.042

CEZ N2 15 37.17 4.079

CEZ N2 15 38.75 4.060

CEZ N2 15 40.26 4.122

CEZ N2 15 41.77 4.098

CEZ N2 15 43.34 4.130

CEZ N2 15 44.19 4.085

CEZ N2 15 45.74 4.118

CEZ N2 15 46.50 4.141

CEZ N2 15 47.90 4.219


CEZ N2 15 48.62 4.216

CEZ N2 15 50.13 4.308

CEZ N2 15 50.95 4.257

CEZ N2 15 52.43 4.341

CEZ N2 15 53.20 4.343

CEZ N2 15 54.74 4.320

CEZ N2 15 55.51 4.268

CEZ N2 15 57.04 4.269

CEZ N2 15 57.82 4.232

CEZ N2 15 59.13 4.236

CEZ N2 15 59.80 4.186

CEZ N2 15 61.18 4.163

CEZ N2 15 61.91 4.067

CEZ N2 15 62.81 4.036

CEZ N2 15 63.24 3.985

CEZ N2 15 63.77 3.938

CEZ N2 15 64.03 3.845

CEZ N2 30 1.35 0.328

CEZ N2 30 2.69 0.606

CEZ N2 30 4.16 0.890

CEZ N2 30 5.50 1.113

CEZ N2 30 6.80 1.305

CEZ N2 30 8.10 1.495

CEZ N2 30 9.39 1.641

CEZ N2 30 10.78 1.803

CEZ N2 30 12.27 1.950

CEZ N2 30 13.79 2.093

CEZ N2 30 15.30 2.196

CEZ N2 30 16.82 2.331

CEZ N2 30 18.25 2.432

CEZ N2 30 19.73 2.527

CEZ N2 30 21.25 2.615

CEZ N2 30 22.74 2.710

CEZ N2 30 24.28 2.792

CEZ N2 30 25.84 2.869

CEZ N2 30 27.27 2.956

CEZ N2 30 28.84 3.027

CEZ N2 30 30.30 3.075

CEZ N2 30 31.86 3.129

CEZ N2 30 33.41 3.144

CEZ N2 30 34.95 3.212

CEZ N2 30 36.54 3.245

CEZ N2 30 38.12 3.296


CEZ N2 30 39.73 3.316

CEZ N2 30 41.32 3.347

CEZ N2 30 42.79 3.345

CEZ N2 30 44.31 3.373

CEZ N2 30 45.13 3.317

CEZ N2 30 46.70 3.358

CEZ N2 30 47.53 3.320

CEZ N2 30 49.01 3.280

CEZ N2 30 49.69 3.314

CEZ N2 30 51.26 3.279

CEZ N2 30 52.04 3.230

CEZ N2 30 53.57 3.268

CEZ N2 30 54.35 3.240

CEZ N2 30 55.95 3.182

CEZ N2 30 56.72 3.147

CEZ N2 30 58.09 3.127

CEZ N2 30 58.81 3.038

CEZ N2 30 60.24 2.999

CEZ N2 30 60.93 2.992

CEZ N2 30 62.13 2.921

CEZ N2 30 62.72 2.838

CEZ N2 30 63.22 2.828

CEZ N2 30 63.81 2.736

CEZ N2 30 64.08 2.635

NaY CH4 0 1.11 0.664

NaY CH4 0 2.17 1.251

NaY CH4 0 3.37 1.805

NaY CH4 0 4.63 2.260

NaY CH4 0 5.97 2.662

NaY CH4 0 7.26 2.963

NaY CH4 0 8.79 3.263

NaY CH4 0 10.15 3.502

NaY CH4 0 11.65 3.748

NaY CH4 0 13.16 3.960

NaY CH4 0 14.59 4.130

NaY CH4 0 16.09 4.339

NaY CH4 0 17.51 4.507

NaY CH4 0 18.94 4.643

NaY CH4 0 20.50 4.779

NaY CH4 0 22.06 4.933

NaY CH4 0 23.51 5.083

NaY CH4 0 24.92 5.239

NaY CH4 0 26.35 5.334


NaY CH4 0 27.80 5.461

NaY CH4 0 29.33 5.587

NaY CH4 0 30.70 5.696

NaY CH4 0 31.81 5.776

NaY CH4 0 32.77 5.872

NaY CH4 0 33.77 5.963

NaY CH4 0 34.65 6.001

NaY CH4 15 1.33 0.497

NaY CH4 15 2.54 0.905

NaY CH4 15 3.83 1.285

NaY CH4 15 5.21 1.636

NaY CH4 15 6.53 1.912

NaY CH4 15 7.91 2.160

NaY CH4 15 9.48 2.395

NaY CH4 15 10.86 2.593

NaY CH4 15 12.40 2.771

NaY CH4 15 13.96 2.948

NaY CH4 15 15.49 3.101

NaY CH4 15 16.93 3.248

NaY CH4 15 18.37 3.367

NaY CH4 15 19.96 3.479

NaY CH4 15 21.43 3.598

NaY CH4 15 22.99 3.719

NaY CH4 15 24.53 3.814

NaY CH4 15 26.11 3.908

NaY CH4 15 27.58 4.004

NaY CH4 15 29.14 4.103

NaY CH4 15 30.67 4.210

NaY CH4 15 31.86 4.281

NaY CH4 15 32.99 4.347

NaY CH4 15 33.95 4.382

NaY CH4 15 34.61 4.391

NaY CH4 30 1.38 0.259

NaY CH4 30 2.90 0.515

NaY CH4 30 4.43 0.708

NaY CH4 30 5.91 0.872

NaY CH4 30 7.30 1.012

NaY CH4 30 8.71 1.136

NaY CH4 30 10.04 1.246

NaY CH4 30 11.60 1.363

NaY CH4 30 13.12 1.487

NaY CH4 30 14.59 1.566

NaY CH4 30 16.07 1.660


NaY CH4 30 17.66 1.743

NaY CH4 30 19.24 1.855

NaY CH4 30 20.79 1.944

NaY CH4 30 22.27 2.014

NaY CH4 30 23.85 2.099

NaY CH4 30 25.41 2.175

NaY CH4 30 27.03 2.243

NaY CH4 30 28.60 2.318

NaY CH4 30 30.04 2.378

NaY CH4 30 31.45 2.444

NaY CH4 30 32.71 2.432

NaY CH4 30 33.74 2.439

NaY CH4 30 34.57 2.446

C‐R2030 CH4 0 0.65 0.839

C‐R2030 CH4 0 1.48 1.440

C‐R2030 CH4 0 2.45 1.917

C‐R2030 CH4 0 3.54 2.316

C‐R2030 CH4 0 4.79 2.683

C‐R2030 CH4 0 6.17 3.002

C‐R2030 CH4 0 7.60 3.293

C‐R2030 CH4 0 8.89 3.492

C‐R2030 CH4 0 10.19 3.692

C‐R2030 CH4 0 11.62 3.867

C‐R2030 CH4 0 12.97 4.040

C‐R2030 CH4 0 14.44 4.197

C‐R2030 CH4 0 15.96 4.345

C‐R2030 CH4 0 17.48 4.489

C‐R2030 CH4 0 18.91 4.600

C‐R2030 CH4 0 20.43 4.734

C‐R2030 CH4 0 21.97 4.853

C‐R2030 CH4 0 23.40 4.955

C‐R2030 CH4 0 24.82 5.049

C‐R2030 CH4 0 26.39 5.158

C‐R2030 CH4 0 27.83 5.253

C‐R2030 CH4 0 29.36 5.340

C‐R2030 CH4 0 30.65 5.421

C‐R2030 CH4 0 31.81 5.473

C‐R2030 CH4 0 32.95 5.527

C‐R2030 CH4 0 33.95 5.569

C‐R2030 CH4 0 34.67 5.577

C‐R2030 CH4 15 0.04 0.001

C‐R2030 CH4 15 0.75 0.735

C‐R2030 CH4 15 1.75 1.298


C‐R2030 CH4 15 2.77 1.700

C‐R2030 CH4 15 3.92 2.041

C‐R2030 CH4 15 5.09 2.339

C‐R2030 CH4 15 6.30 2.577

C‐R2030 CH4 15 7.55 2.802

C‐R2030 CH4 15 8.84 3.005

C‐R2030 CH4 15 10.07 3.166

C‐R2030 CH4 15 11.46 3.334

C‐R2030 CH4 15 12.95 3.505

C‐R2030 CH4 15 14.36 3.640

C‐R2030 CH4 15 15.87 3.784

C‐R2030 CH4 15 17.41 3.912

C‐R2030 CH4 15 18.75 4.019

C‐R2030 CH4 15 20.25 4.131

C‐R2030 CH4 15 21.68 4.219

C‐R2030 CH4 15 23.10 4.322

C‐R2030 CH4 15 24.68 4.400

C‐R2030 CH4 15 26.20 4.480

C‐R2030 CH4 15 27.65 4.558

C‐R2030 CH4 15 29.10 4.617

C‐R2030 CH4 15 30.40 4.673

C‐R2030 CH4 15 31.51 4.734

C‐R2030 CH4 15 32.67 4.759

C‐R2030 CH4 15 33.66 4.788

C‐R2030 CH4 15 34.20 4.788

C‐R2030 CH4 30 0.87 0.632

C‐R2030 CH4 30 1.91 1.100

C‐R2030 CH4 30 3.00 1.438

C‐R2030 CH4 30 4.25 1.775

C‐R2030 CH4 30 5.52 2.045

C‐R2030 CH4 30 6.89 2.283

C‐R2030 CH4 30 8.23 2.484

C‐R2030 CH4 30 9.56 2.659

C‐R2030 CH4 30 10.89 2.812

C‐R2030 CH4 30 12.32 2.972

C‐R2030 CH4 30 13.84 3.114

C‐R2030 CH4 30 15.41 3.235

C‐R2030 CH4 30 16.94 3.329

C‐R2030 CH4 30 18.39 3.437

C‐R2030 CH4 30 19.81 3.545

C‐R2030 CH4 30 21.40 3.618

C‐R2030 CH4 30 22.97 3.719

C‐R2030 CH4 30 24.53 3.799


C‐R2030 CH4 30 26.11 3.866

C‐R2030 CH4 30 27.69 3.923

C‐R2030 CH4 30 29.17 4.009

C‐R2030 CH4 30 30.50 4.026

C‐R2030 CH4 30 31.58 4.052

C‐R2030 CH4 30 32.58 4.108

C‐R2030 CH4 30 33.57 4.100

CEZ CH4 0 0.04 0.004

CEZ CH4 0 0.95 0.892

CEZ CH4 0 1.92 1.615

CEZ CH4 0 2.97 2.203

CEZ CH4 0 4.23 2.742

CEZ CH4 0 5.42 3.143

CEZ CH4 0 6.70 3.466

CEZ CH4 0 8.05 3.762

CEZ CH4 0 9.61 4.040

CEZ CH4 0 10.97 4.251

CEZ CH4 0 12.48 4.440

CEZ CH4 0 13.89 4.613

CEZ CH4 0 15.32 4.763

CEZ CH4 0 16.80 4.919

CEZ CH4 0 18.35 5.071

CEZ CH4 0 19.91 5.209

CEZ CH4 0 21.33 5.334

CEZ CH4 0 22.87 5.458

CEZ CH4 0 24.41 5.592

CEZ CH4 0 25.99 5.704

CEZ CH4 0 27.58 5.764

CEZ CH4 0 29.00 5.918

CEZ CH4 0 30.27 5.987

CEZ CH4 0 31.53 6.026

CEZ CH4 0 32.55 6.091

CEZ CH4 15 0.04 0.004

CEZ CH4 15 1.14 0.755

CEZ CH4 15 2.29 1.374

CEZ CH4 15 3.58 1.917

CEZ CH4 15 4.85 2.332

CEZ CH4 15 6.08 2.699

CEZ CH4 15 7.34 2.984

CEZ CH4 15 8.57 3.237

CEZ CH4 15 10.14 3.497

CEZ CH4 15 11.54 3.693

CEZ CH4 15 13.07 3.888


CEZ CH4 15 14.59 4.047

CEZ CH4 15 16.11 4.192

CEZ CH4 15 17.53 4.337

CEZ CH4 15 19.04 4.436

CEZ CH4 15 20.51 4.518

CEZ CH4 15 22.09 4.637

CEZ CH4 15 23.67 4.759

CEZ CH4 15 25.26 4.831

CEZ CH4 15 26.76 4.886

CEZ CH4 15 28.29 4.974

CEZ CH4 15 29.65 5.032

CEZ CH4 15 30.89 5.058

CEZ CH4 15 31.99 5.090

CEZ CH4 15 32.98 5.103

CEZ CH4 15 33.51 5.086

CEZ CH4 15 33.76 5.071

CEZ CH4 15 33.90 5.026

CEZ CH4 15 33.98 4.965

CEZ CH4 15 34.02 4.911

CEZ CH4 30 1.17 0.591

CEZ CH4 30 2.33 1.062

CEZ CH4 30 3.67 1.520

CEZ CH4 30 4.94 1.877

CEZ CH4 30 6.35 2.177

CEZ CH4 30 7.58 2.464

CEZ CH4 30 9.10 2.715

CEZ CH4 30 10.71 2.907

CEZ CH4 30 12.20 3.090

CEZ CH4 30 13.63 3.255

CEZ CH4 30 15.17 3.414

CEZ CH4 30 16.65 3.520

CEZ CH4 30 18.25 3.604

CEZ CH4 30 19.70 3.710

CEZ CH4 30 21.15 3.791

CEZ CH4 30 22.74 3.867

CEZ CH4 30 24.19 3.911

CEZ CH4 30 25.78 3.993

CEZ CH4 30 27.25 4.048

CEZ CH4 30 28.75 4.084

CEZ CH4 30 30.19 4.093

CEZ CH4 30 31.27 4.133

CEZ CH4 30 32.31 4.118

CEZ CH4 30 33.12 4.111


CEZ CH4 30 33.52 4.095

CEZ CH4 30 33.75 4.038

CEZ CH4 30 33.87 3.976

CEZ CH4 30 33.91 3.952

CEZ CH4 30 33.97 3.876

CEZ CH4 30 33.98 3.833

CEZ CH4 30 33.97 3.808

CEZ CH4 30 33.97 3.774

CEZ CH4 30 33.99 3.700

CEZ CH4 30 34.01 3.607


References

1. Dudley, B., BP statistical review of world energy, in BP Statistical Review, London, UK, accessed Aug. 2019. p. 2019.

2. Agency, I.E., World Energy Outlook 2019. 2019. p. Available online at www.iea.org/reports/world‐energy‐outlook‐2019/.

3. Administration, U.S.E.I., Annual Energy Outlook 2019. 2019, U.S. Energy Information Administration. p. Available online at https://www.eia.gov/outlooks/aeo/pdf/aeo2019.pdf.

4. Government of Australia ‐ Department of Industry, I.a.S., Resources and Energy Quarterly March 2019, I.a.S. Department of Industry, Editor. 2019, Government of Australia.

5. Government of Western Australia ‐ Department of Jobs, T., Science and Innovation, Western Australia Liquefied Natural Gas Profile October 2019, in WA LNG Industry Profile ‐ October 2019, T. Department of Jobs, Science and Innovation, Editor. 2019, Government of Western Australia.

6. (UNFCC), U.N.F.o.C.C. The Paris Agreement. 2019; Available online at https://unfccc.int/process‐and‐meetings/the‐paris‐agreement/the‐paris‐agreement].

7. Nachmany, M., S.S. Fankhauser, Joana, and A. Averchenkova, Global trends in climate change legislation and litigation: 2017 update, in London: Grantham Research Institute on Climate Change and the Environment and Centre for Climate Change Economics and Policy, London School of Economics and Political Science. 2017.

8. Setzer, J. and R. Byrnes, Global trends in climate change litigation: 2019 snapshot, in London: Grantham Research Institute on Climate Change and the Environment and Centre for Climate Change Economics and Policy, London School of Economics and Political Science. 2019.

9. (EEA), E.E.A. Atmospheric greenhouse gas concentrations. 2019 December 5, 2019; Available online at https://www.eea.europa.eu/data‐and‐maps/indicators/atmospheric‐greenhouse‐gas‐concentrations‐6/assessment‐1].

10. Administration, U.S.E.I. How much carbon dioxide is produced when different fuels are burned? 2019 June 4, 2019; Available online at https://www.eia.gov/tools/faqs/faq.php?id=73&t=11].

11. Jacobs, D., The Global Market for Liquefied Natural Gas. 2011, Reserve Bank of Australia (RBA). p. Available online at https://www.rba.gov.au/publications/bulletin/2011/sep/pdf/bu‐0911‐3.pdf.

12. Rufford, T.E., S. Smart, G.C. Watson, B. Graham, J. Boxall, J.D. Da Costa, and E. May, The removal of CO2 and N2 from natural gas: A review of conventional and emerging process technologies. Journal of Petroleum Science and Engineering, 2012. 94: p. 123‐154.

13. Kidnay, A.J., W.R. Parrish, and D.G. McCartney, Fundamentals of natural gas processing. 2011: CRC press.

14. Hubbard, B. New and Emerging Technologies (Petroskills workshop). in Gas Processors Association Convention. 2010. Austin, Texas: John M.Campbell & Co.

15. Bindwal, A.B., P.D. Vaidya, and E.Y. Kenig, Kinetics of carbon dioxide removal by aqueous diamines. Chemical engineering journal, 2011. 169(1‐3): p. 144‐150.

16. Wilkinson, D. and G. Johnson. Nitrogen Rejection Technology for Abu Dhabi. in GPA GCC 18th Annual Technical Conference, Oman. 2010.

17. FINN, A., Rejection strategies. Hydrocarbon engineering, 2007. 12(10): p. 49‐52. 18. Peters, L., A. Hussain, M. Follmann, T. Melin, and M.‐B. Hägg, CO2 removal from natural gas

by employing amine absorption and membrane technology—A technical and economical analysis. Chemical Engineering Journal, 2011. 172(2‐3): p. 952‐960.

19. Agrawal, R., D.M. Herron, H.C. Rowles, and G.E. Kinard, Cryogenic technology. Kirk‐Othmer Encyclopedia of Chemical Technology, 2000.


20. Maqsood, K., J. Pal, D. Turunawarasu, A.J. Pal, and S. Ganguly, Performance enhancement and energy reduction using hybrid cryogenic distillation networks for purification of natural gas with high CO 2 content. Korean Journal of Chemical Engineering, 2014. 31(7): p. 1120‐1135.

21. Tagliabue, M., D. Farrusseng, S. Valencia, S. Aguado, U. Ravon, C. Rizzo, A. Corma, and C. Mirodatos, Natural gas treating by selective adsorption: Material science and chemical engineering interplay. Chemical Engineering Journal, 2009. 155(3): p. 553‐566.

22. Alonso‐Vicario, A., Ochoa‐Gomez, JR Gil‐Rio, S. Gomez‐Jimerez‐Aberasturi, O. Ramirez‐Lopez, J. Torrecilla‐Soria, CA and Dominguez, A.:" Purification and upgradng of biogas by pressure swing adsorption on synthetic and natural zeolites. Microporous and Mesoporous Materials, 2010. 134: p. 100‐107.

23. Esteves, I.A., M.S. Lopes, P.M. Nunes, and J.P. Mota, Adsorption of natural gas and biogas components on activated carbon. Separation and Purification Technology, 2008. 62(2): p. 281‐296.

24. Warmuzinski, K., Harnessing methane emissions from coal mining. Process safety and environmental protection, 2008. 86(5): p. 315‐320.

25. Ruthven, D.M., Principles of adsorption and adsorption processes. 1984: John Wiley & Sons. 26. Seader, J.D., E.J. Henley, and D.K. Roper, Separation process principles. 1998. 27. Jensen, N.K., T.E. Rufford, G. Watson, D.K. Zhang, K.I. Chan, and E.F. May, Screening zeolites

for gas separation applications involving methane, nitrogen, and carbon dioxide. Journal of Chemical & Engineering Data, 2011. 57(1): p. 106‐113.

28. Rufford, T.E., G.C. Watson, T.L. Saleman, P.S. Hofman, N.K. Jensen, and E.F. May, Adsorption equilibria and kinetics of methane+ nitrogen mixtures on the activated carbon Norit RB3. Industrial & Engineering Chemistry Research, 2013. 52(39): p. 14270‐14281.

29. Watson, G., E. May, B. Graham, M. Trebble, R. Trengove, and K. Chan, Equilibrium adsorption measurements of pure nitrogen, carbon dioxide, and methane on a carbon molecular sieve at cryogenic temperatures and high pressures. Journal of Chemical & Engineering Data, 2009. 54(9): p. 2701‐2707.

30. Watson, G.C., N.K. Jensen, T.E. Rufford, K.I. Chan, and E.F. May, Volumetric adsorption measurements of N2, CO2, CH4, and a CO2+ CH4 mixture on a natural chabazite from (5 to 3000) kPa. Journal of Chemical & Engineering Data, 2011. 57(1): p. 93‐101.

31. Wei, Y., T.E. Parmentier, K.P. de Jong, and J. Zečević, Tailoring and visualizing the pore architecture of hierarchical zeolites. Chemical Society Reviews, 2015. 44(20): p. 7234‐7261.

32. Xu, X., X. Zhao, L. Sun, and X. Liu, Adsorption separation of carbon dioxide, methane and nitrogen on monoethanol amine modified β‐zeolite. Journal of Natural Gas Chemistry, 2009. 18(2): p. 167‐172.

33. Xu, X., X. Zhao, L. Sun, and X. Liu, Adsorption separation of carbon dioxide, methane, and nitrogen on Hβ and Na‐exchanged β‐zeolite. Journal of Natural Gas Chemistry, 2008. 17(4): p. 391‐396.

34. Xu, X., C. Song, B.G. Miller, and A.W. Scaroni, Adsorption separation of carbon dioxide from flue gas of natural gas‐fired boiler by a novel nanoporous “molecular basket” adsorbent. Fuel processing technology, 2005. 86(14‐15): p. 1457‐1472.

35. Herm, Z.R., R. Krishna, and J.R. Long, CO2/CH4, CH4/H2 and CO2/CH4/H2 separations at high pressures using Mg2 (dobdc). Microporous and Mesoporous Materials, 2012. 151: p. 481‐487.

36. Alhamami, M., H. Doan, and C.‐H. Cheng, A Review on Breathing Behaviors of Metal‐Organic‐Frameworks (MOFs) for Gas Adsorption. Materials, 2014. 7(4): p. 3198.

37. He, Y., J. Shang, Q. Gu, G. Li, J. Li, R. Singh, P. Xiao, and P.A. Webley, Converting 3D rigid metal–organic frameworks (MOFs) to 2D flexible networks via ligand exchange for enhanced CO 2/N 2 and CH 4/N 2 separation. Chemical Communications, 2015. 51(79): p. 14716‐14719.

38. Yilmaz, B., N. Trukhan, and U. Mueller, Industrial outlook on zeolites and metal organic frameworks. Chinese Journal of Catalysis, 2012. 33(1): p. 3‐10.


39. Demessence, A. and J.R. Long, Selective gas adsorption in the flexible metal–organic frameworks Cu (BDTri) L (L= dmf, def). Chemistry–A European Journal, 2010. 16(20): p. 5902‐5908.

40. Li, B., Z. Zhang, Y. Li, K. Yao, Y. Zhu, Z. Deng, F. Yang, X. Zhou, G. Li, and H. Wu, Enhanced binding affinity, remarkable selectivity, and high capacity of CO2 by dual functionalization of a

rht‐type metal–organic framework. Angewandte Chemie International Edition, 2012. 51(6): p. 1412‐1415.

41. Khan, N.A., Z. Hasan, and S.H. Jhung, Adsorptive removal of hazardous materials using metal‐organic frameworks (MOFs): a review. Journal of hazardous materials, 2013. 244: p. 444‐456.

42. Gu, C., D. Liu, W. Huang, J. Liu, and R. Yang, Synthesis of covalent triazine‐based frameworks with high CO 2 adsorption and selectivity. Polymer Chemistry, 2015. 6(42): p. 7410‐7417.

43. Xu, C. and N. Hedin, Synthesis of microporous organic polymers with high CO 2‐over‐N 2 selectivity and CO 2 adsorption. Journal of Materials Chemistry A, 2013. 1(10): p. 3406‐3414.

44. Langmuir, I., The constitution and fundamental properties of solids and liquids. II. Liquids. Journal of the American chemical society, 1917. 39(9): p. 1848‐1906.

45. Brunauer, S., L.S. Deming, W.E. Deming, and E. Teller, On a theory of the van der Waals adsorption of gases. Journal of the American Chemical society, 1940. 62(7): p. 1723‐1732.

46. Sips, R., Combined form of Langmuir and Freundlich equations. J. Chem. Phys, 1948. 16(429): p. 490‐495.

47. Farrusseng, D., Metal‐organic frameworks: applications from catalysis to gas storage. 2011: John Wiley & Sons.

48. Mason, J.A., J. Oktawiec, M.K. Taylor, M.R. Hudson, J. Rodriguez, J.E. Bachman, M.I. Gonzalez, A. Cervellino, A. Guagliardi, and C.M. Brown, Methane storage in flexible metal–organic frameworks with intrinsic thermal management. Nature, 2015. 527(7578): p. 357‐361.

49. Thallapally, P.K., J. Tian, M. Radha Kishan, C.A. Fernandez, S.J. Dalgarno, P.B. McGrail, J.E. Warren, and J.L. Atwood, Flexible (breathing) interpenetrated metal− organic frameworks for CO2 separation applications. Journal of the American Chemical Society, 2008. 130(50): p. 16842‐16843.

50. Li, Y. and R.T. Yang, Gas adsorption and storage in metal− organic framework MOF‐177. Langmuir, 2007. 23(26): p. 12937‐12944.

51. Ma, S. and H.‐C. Zhou, Gas storage in porous metal–organic frameworks for clean energy applications. Chemical Communications, 2010. 46(1): p. 44‐53.

52. Lee, J., O.K. Farha, J. Roberts, K.A. Scheidt, S.T. Nguyen, and J.T. Hupp, Metal–organic framework materials as catalysts. Chemical Society Reviews, 2009. 38(5): p. 1450‐1459.

53. Suh, M.P., H.J. Park, T.K. Prasad, and D.‐W. Lim, Hydrogen Storage in Metal–Organic Frameworks. Chemical Reviews, 2012. 112(2): p. 782‐835.

54. Stock, N. and S. Biswas, Synthesis of metal‐organic frameworks (MOFs): routes to various MOF topologies, morphologies, and composites. Chemical reviews, 2011. 112(2): p. 933‐969.

55. Farha, O.K., A.Ö. Yazaydın, I. Eryazici, C.D. Malliakas, B.G. Hauser, M.G. Kanatzidis, S.T. Nguyen, R.Q. Snurr, and J.T. Hupp, De novo synthesis of a metal–organic framework material featuring ultrahigh surface area and gas storage capacities. Nature chemistry, 2010. 2(11): p. 944‐948.

56. Férey, G. and C. Serre, Large breathing effects in three‐dimensional porous hybrid matter: facts, analyses, rules and consequences. Chemical Society Reviews, 2009. 38(5): p. 1380‐1399.

57. Schneemann, A., V. Bon, I. Schwedler, I. Senkovska, S. Kaskel, and R.A. Fischer, Flexible metal–organic frameworks. Chemical Society Reviews, 2014. 43(16): p. 6062‐6096.

58. Boutin, A., S. Couck, F.‐X. Coudert, P. Serra‐Crespo, J. Gascon, F. Kapteijn, A.H. Fuchs, and J.F. Denayer, Thermodynamic analysis of the breathing of amino‐functionalized MIL‐53 (Al) upon CO 2 adsorption. Microporous and Mesoporous Materials, 2011. 140(1): p. 108‐113.


59. Arami‐Niya, A., G. Birkett, Z. Zhu, and T.E. Rufford, Gate opening effect of zeolitic imidazolate framework ZIF‐7 for adsorption of CH 4 and CO 2 from N 2. Journal of Materials Chemistry A, 2017. 5(40): p. 21389‐21399.

60. Couck, S., J.F. Denayer, G.V. Baron, T. Rémy, J. Gascon, and F. Kapteijn, An amine‐functionalized MIL‐53 metal− organic framework with large separa on power for CO2 and CH4. Journal of the American Chemical Society, 2009. 131(18): p. 6326‐6327.

61. Li, G.K., J. Shang, Q. Gu, R.V. Awati, N. Jensen, A. Grant, X. Zhang, D.S. Sholl, J.Z. Liu, P.A. Webley, and E.F. May, Temperature‐regulated guest admission and release in microporous materials. Nature Communications, 2017. 8: p. 15777.

62. Coudert, F.‐X., M. Jeffroy, A.H. Fuchs, A. Boutin, and C. Mellot‐Draznieks, Thermodynamics of Guest‐Induced Structural Transitions in Hybrid Organic−Inorganic Frameworks. Journal of the American Chemical Society, 2008. 130(43): p. 14294‐14302.

63. Horike, S., S. Shimomura, and S. Kitagawa, Soft porous crystals. Nature chemistry, 2009. 1(9): p. 695‐704.

64. McDonald, T.M., J.A. Mason, X. Kong, E.D. Bloch, D. Gygi, A. Dani, V. Crocella, F. Giordanino, S.O. Odoh, and W.S. Drisdell, Cooperative insertion of CO2 in diamine‐appended metal‐organic frameworks. Nature, 2015. 519(7543): p. 303‐308.

65. Zhao, P., G.I. Lampronti, G.O. Lloyd, E. Suard, and S.A. Redfern, Direct visualisation of carbon dioxide adsorption in gate‐opening zeolitic imidazolate framework ZIF‐7. Journal of Materials Chemistry A, 2014. 2(3): p. 620‐623.

66. Du, Y., B. Wooler, M. Nines, P. Kortunov, C.S. Paur, J. Zengel, S.C. Weston, and P.I. Ravikovitch, New high‐and low‐temperature phase changes of ZIF‐7: elucidation and prediction of the thermodynamics of transitions. J. Am. Chem. Soc, 2015. 137(42): p. 13603‐13611.

67. Gao, S., C.G. Morris, Z. Lu, Y. Yan, H.G. Godfrey, C. Murray, C.C. Tang, K.M. Thomas, S. Yang, and M. Schröder, Selective Hysteretic Sorption of Light Hydrocarbons in a Flexible Metal–Organic Framework Material. Chemistry of Materials, 2016. 28(7): p. 2331‐2340.

68. Numaguchi, R., H. Tanaka, S. Watanabe, and M.T. Miyahara, Simulation study for adsorption‐induced structural transition in stacked‐layer porous coordination polymers: Equilibrium and hysteretic adsorption behaviors. The Journal of chemical physics, 2013. 138(5): p. 054708.

69. Coudert, F.‐X., M. Jeffroy, A.H. Fuchs, A. Boutin, and C. Mellot‐Draznieks, Thermodynamics of guest‐induced structural transitions in hybrid organic− inorganic frameworks. Journal of the American Chemical Society, 2008. 130(43): p. 14294‐14302.

70. Llewellyn, P.L., G. Maurin, T. Devic, S. Loera‐Serna, N. Rosenbach, C. Serre, S. Bourrelly, P. Horcajada, Y. Filinchuk, and G. Férey, Prediction of the conditions for breathing of metal organic framework materials using a combination of X‐ray powder diffraction, microcalorimetry, and molecular simulation. Journal of the American Chemical Society, 2008. 130(38): p. 12808‐12814.

71. Cai, W., T. Lee, M. Lee, W. Cho, D.‐Y. Han, N. Choi, A.C. Yip, and J. Choi, Thermal structural transitions and carbon dioxide adsorption properties of zeolitic imidazolate framework‐7 (ZIF‐7). Journal of the American Chemical Society, 2014. 136(22): p. 7961‐7971.

72. Gücüyener, C., J. van den Bergh, J. Gascon, and F. Kapteijn, Ethane/ethene separation turned on its head: selective ethane adsorption on the metal− organic Framework ZIF‐7 through a gate‐opening mechanism. Journal of the American Chemical Society, 2010. 132(50): p. 17704‐17706.

73. Morishige, K. and N. Tateishi, Adsorption hysteresis in ink‐bottle pore. The Journal of chemical physics, 2003. 119(4): p. 2301‐2306.

74. Neimark, A.V., P.I. Ravikovitch, and A. Vishnyakov, Adsorption hysteresis in nanopores. Physical Review E, 2000. 62(2): p. R1493.

75. Burgess, C.G., D.H. Everett, and S. Nuttall, Adsorption hysteresis in porous materials. Pure and Applied chemistry, 1989. 61(11): p. 1845‐1852.


76. Serre, C., S. Bourrelly, A. Vimont, N.A. Ramsahye, G. Maurin, P.L. Llewellyn, M. Daturi, Y. Filinchuk, O. Leynaud, and P. Barnes, An explanation for the very large breathing effect of a metal–organic framework during CO2 adsorption. Advanced materials, 2007. 19(17): p. 2246‐2251.

77. Neimark, A.V., F.‐X. Coudert, C. Triguero, A. Boutin, A.H. Fuchs, I. Beurroies, and R. Denoyel, Structural Transitions in MIL‐53 (Cr): View from Outside and Inside. Langmuir, 2011. 27(8): p. 4734‐4741.

78. Boutin, A., F.‐X. Coudert, M.‐A. Springuel‐Huet, A.V. Neimark, G. Férey, and A.H. Fuchs, The behavior of flexible MIL‐53 (Al) upon CH4 and CO2 adsorption. The Journal of Physical Chemistry C, 2010. 114(50): p. 22237‐22244.

79. Gomez, L.F., R. Zacharia, P. Bénard, and R. Chahine, Simulation of Binary CO2/CH4 Mixture Breakthrough Profiles in MIL‐53 (Al). Journal of Nanomaterials, 2015. 2015: p. 15.

80. Rodenas, T., M. van Dalen, E. García‐Pérez, P. Serra‐Crespo, B. Zornoza, F. Kapteijn, and J.

Gascon, Visualizing MOF Mixed Matrix Membranes at the Nanoscale: Towards Structure‐Performance Relationships in CO2/CH4 Separation Over NH2‐MIL‐53 (Al)@ PI. Advanced Functional Materials, 2014. 24(2): p. 249‐256.

81. Cavenati, S., C.A. Grande, and A.E. Rodrigues, Adsorption equilibrium of methane, carbon dioxide, and nitrogen on zeolite 13X at high pressures. Journal of Chemical & Engineering Data, 2004. 49(4): p. 1095‐1101.

82. Basu, S., M. Maes, A. Cano‐Odena, L. Alaerts, D.E. De Vos, and I.F.J. Vankelecom, Solvent resistant nanofiltration (SRNF) membranes based on metal‐organic frameworks. Journal of Membrane Science, 2009. 344(1–2): p. 190‐198.

83. Krause, S., V. Bon, I. Senkovska, U. Stoeck, D. Wallacher, D.M. Többens, S. Zander, R.S. Pillai, G. Maurin, and F.o.‐X. Coudert, A pressure‐amplifying framework material with negative gas adsorption transitions. Nature, 2016. 532(7599): p. 348‐352.

84. Yan, T., Y. Lan, D. Liu, Q. Yang, and C. Zhong, Large‐Scale Screening and Design of Metal–Organic Frameworks for CH4/N2 Separation. Chemistry–An Asian Journal, 2019. 14(20): p. 3688‐3693.

85. Wu, Y., D. Yuan, D. He, J. Xing, S. Zeng, S. Xu, Y. Xu, and Z. Liu, Decorated Traditional Zeolites with Subunits of Metal–Organic Frameworks for CH4/N2 Separation. Angewandte Chemie International Edition, 2019. 58(30): p. 10241‐10244.

86. Chen, Y., H. Wu, Y. Yuan, D. Lv, Z. Qiao, D. An, X. Wu, H. Liang, Z. Li, and Q. Xia, Highly rapid mechanochemical synthesis of a pillar‐layer metal‐organic framework for efficient CH4/N2 separation. Chemical Engineering Journal, 2020. 385: p. 123836.

87. Chen, L.‐H., X.‐Y. Li, J.C. Rooke, Y.‐H. Zhang, X.‐Y. Yang, Y. Tang, F.‐S. Xiao, and B.‐L. Su, Hierarchically structured zeolites: synthesis, mass transport properties and applications. Journal of Materials Chemistry, 2012. 22(34): p. 17381‐17403.

88. Deng, C., J. Zhang, L. Dong, M. Huang, B. Li, G. Jin, J. Gao, F. Zhang, M. Fan, and L. Zhang, The effect of positioning cations on acidity and stability of the framework structure of Y zeolite. Scientific reports, 2016. 6: p. 23382.

89. Yang, R.T., Adsorbents: fundamentals and applications. 2003: John Wiley & Sons. 90. Mahaffy, P.G., R. Bucat, R. Tasker, J.C. Kotz, P. Treichel, G.C. Weaver, and J. McMurry,

Chemistry: Human Activity, Chemical Reactivity. 2011: Nelson Education Toronto. 91. Sethia, G., R.S. Somani, and H.C. Bajaj, Adsorption of carbon monoxide, methane and nitrogen

on alkaline earth metal ion exchanged zeolite‐X: structure, cation position and adsorption relationship. RSC Advances, 2015. 5(17): p. 12773‐12781.

92. Hofman, P.S., T.E. Rufford, K.I. Chan, and E.F. May, A dynamic column breakthrough apparatus for adsorption capacity measurements with quantitative uncertainties. Adsorption, 2012. 18(3‐4): p. 251‐263.

93. LI, G., E.F. MAY, P.A. WEBLEY, S.H.‐W. HUANG, and K.I. CHAN, METHOD FOR GAS SEPARATION, U.S.P.a.T. Office, Editor. 2020.


94. Mills, B. Structural formula of tetramethylammonium chloride. 2015 23 April 2015 [cited 2020; Available from: https://commons.wikimedia.org/wiki/File:Tetramethylammonium‐chloride‐ion‐pair‐2D.png.

95. Li, G.K., J. Shang, P. Xiao, R. Singh, D. Danaci, P.A. Webley, and E.F. May, Separation of nitrogen/methane on organic cation zeolites for natural gas purification, in FOA11 – 11th International Conference on the Fundamentals of Adsorption. 2013: Baltimore, MD.

96. Saleman, T.L., G.C. Watson, T.E. Rufford, P.S. Hofman, K.I. Chan, and E.F. May, Capacity and kinetic measurements of methane and nitrogen adsorption on H+‐mordenite at 243–303 K and pressures to 900 kPa using a dynamic column breakthrough apparatus. Adsorption, 2013. 19(6): p. 1165‐1180.

97. Thu, M.S., Advanced Natural Gas Separation: Further Characterization of Ionic Liquidic Zeolites, in School of Mechanical & Chemical Engineering. 2015, University of Western Australia.

98. International, Z. Zeolite Y. Available from: https://www.zeolyst.com/our‐products/standard‐zeolite‐powders/zeolite‐y.html.

99. Talu, O., S.Y. Zhang, and D.T. Hayhurst, Effect of cations on methane adsorption by NaY, MgY, CaY, SrY, and BaY zeolites. The Journal of Physical Chemistry, 1993. 97(49): p. 12894‐12898.

100. Salem, M., P. Braeuer, M.v. Szombathely, M. Heuchel, P. Harting, K. Quitzsch, and M. Jaroniec, Thermodynamics of high‐pressure adsorption of argon, nitrogen, and methane on microporous adsorbents. Langmuir, 1998. 14(12): p. 3376‐3389.

101. Maurin, G., P. Llewellyn, T. Poyet, and B. Kuchta, Adsorption of argon and nitrogen in X‐faujasites: relationships for understanding the interactions with monovalent and divalent cations. Microporous and mesoporous materials, 2005. 79(1‐3): p. 53‐59.

102. Lee, K.‐M., Y.‐H. Lim, and Y.‐M. Jo, Evaluation of moisture effect on low‐level CO2 adsorption by ion‐exchanged zeolite. Environmental technology, 2012. 33(1): p. 77‐84.

103. Groen, J.C., J.C. Jansen, J.A. Moulijn, and J. Pérez‐Ramírez, Optimal aluminum‐assisted mesoporosity development in MFI zeolites by desilication. The Journal of Physical Chemistry B, 2004. 108(35): p. 13062‐13065.

104. Bhatnagar, A., W. Hogland, M. Marques, and M. Sillanpää, An overview of the modification methods of activated carbon for its water treatment applications. Chemical Engineering Journal, 2013. 219: p. 499‐511.

105. Chen, L., H. Xie, and W. Yu, Functionalization methods of carbon nanotubes and its applications. Carbon nanotubes applications on electron devices, 2011. 41(2): p. 215‐222.

106. Chingombe, P., B. Saha, and R. Wakeman, Surface modification and characterisation of a coal‐based activated carbon. Carbon, 2005. 43(15): p. 3132‐3143.

107. Chen, J.P. and S. Wu, Acid/base‐treated activated carbons: characterization of functional groups and metal adsorptive properties. Langmuir, 2004. 20(6): p. 2233‐2242.

108. Wu, Z., P.A. Webley, and D. Zhao, Comprehensive study of pore evolution, mesostructural stability, and simultaneous surface functionalization of ordered mesoporous carbon (FDU‐15) by wet oxidation as a promising adsorbent. Langmuir, 2010. 26(12): p. 10277‐10286.

109. Vinu, A., K.Z. Hossian, P. Srinivasu, M. Miyahara, S. Anandan, N. Gokulakrishnan, T. Mori, K. Ariga, and V.V. Balasubramanian, Carboxy‐mesoporous carbon and its excellent adsorption capability for proteins. Journal of Materials Chemistry, 2007. 17(18): p. 1819‐1825.

110. Moreno‐Castilla, C., F. Carrasco‐Marín, and A. Mueden, The creation of acid carbon surfaces by treatment with (NH4) 2S2O8. Carbon, 1997. 35(10‐11): p. 1619‐1626.

111. Moreno‐Castilla, C., M. Ferro‐Garcia, J. Joly, I. Bautista‐Toledo, F. Carrasco‐Marin, and J. Rivera‐Utrilla, Activated carbon surface modifications by nitric acid, hydrogen peroxide, and ammonium peroxydisulfate treatments. Langmuir, 1995. 11(11): p. 4386‐4392.

112. Zhang, D., W. Cheng, J. Ma, and R. Li, Influence of activated carbon in zeolite X/activated carbon composites on CH 4/N 2 adsorption separation ability. Adsorption, 2016. 22(8): p. 1129‐1135.


113. Miyake, M., Y. Kimura, T. Ohashi, and M. Matsuda, Preparation of activated carbon–zeolite composite materials from coal fly ash. Microporous and Mesoporous Materials, 2008. 112(1‐3): p. 170‐177.

114. Jha, V.K., M. Matsuda, and M. Miyake, Sorption properties of the activated carbon‐zeolite composite prepared from coal fly ash for Ni2+, Cu2+, Cd2+ and Pb2+. Journal of Hazardous Materials, 2008. 160(1): p. 148‐153.

115. Paskevicius, M., D.A. Sheppard, and C.E. Buckley, Thermodynamic changes in mechanochemically synthesized magnesium hydride nanoparticles. Journal of the American chemical Society, 2010. 132(14): p. 5077‐5083.

116. Purewal, J., Hydrogen adsorption by alkali metal graphite intercalation compounds. 2010, California Institute of Technology.

117. Wang, Y., B. Helvensteijn, N. Nizamidin, A.M. Erion, L.A. Steiner, L.M. Mulloth, B. Luna, and M.D. LeVan, High pressure excess isotherms for adsorption of oxygen and nitrogen in zeolites. Langmuir, 2011. 27(17): p. 10648‐10656.

118. Tekin, R., N. Bac, J. Warzywoda, and A. Sacco Jr, Encapsulation of a fragrance molecule in zeolite X. Microporous and Mesoporous Materials, 2015. 215: p. 51‐57.

119. Ahmed, I.A., S. Young, J. Mosselmans, N. Crout, and E. Bailey, Coordination of Cd2+ ions in the internal pore system of zeolite‐X: A combined EXAFS and isotopic exchange study. Geochimica et Cosmochimica Acta, 2009. 73(6): p. 1577‐1587.

120. Beaudoin, J.J. and J. Marchand, Pore structure. Handbook of Analytic Techniques in Concrete Science and Technology, Principles, Techniques, and Applications, Noyes Publication, New Jersey, 1999: p. 529‐628.

121. Ismail, A.F., K.C. Khulbe, and T. Matsuura, Gas separation membranes. Vol. 7. 2015: Springer. 122. Mao, Z. and S.B. Sinnott, Separation of organic molecular mixtures in carbon nanotubes and

bundles: molecular dynamics simulations. The Journal of Physical Chemistry B, 2001. 105(29): p. 6916‐6924.

123. Drahos, L. and K. Vékey, Determination of the thermal energy and its distribution in peptides. Journal of the American Society for Mass Spectrometry, 1999. 10(4): p. 323‐328.

124. Vékey, K., Internal energy effects in mass spectrometry. Journal of Mass Spectrometry, 1996. 31(5): p. 445‐463.

125. Murdock, C.R., B.C. Hughes, Z. Lu, and D.M. Jenkins, Approaches for synthesizing breathing MOFs by exploiting dimensional rigidity. Coordination Chemistry Reviews, 2014. 258: p. 119‐136.

126. Foo, K.Y. and B.H. Hameed, Insights into the modeling of adsorption isotherm systems. Chemical engineering journal, 2010. 156(1): p. 2‐10.

127. Malek, A. and S. Farooq, Comparison of isotherm models for hydrocarbon adsorption on activated carbon. AIChE Journal, 1996. 42(11): p. 3191‐3201.

128. Li, G., P. Xiao, D. Xu, and P.A. Webley, Dual mode roll‐up effect in multicomponent non‐isothermal adsorption processes with multilayered bed packing. Chemical engineering science, 2011. 66(9): p. 1825‐1834.

129. Ko, D., R. Siriwardane, and L.T. Biegler, Optimization of a pressure‐swing adsorption process using zeolite 13X for CO2 sequestration. Industrial & engineering chemistry research, 2003. 42(2): p. 339‐348.

130. Chue, K., J. Kim, Y. Yoo, S. Cho, and R. Yang, Comparison of activated carbon and zeolite 13X for CO2 recovery from flue gas by pressure swing adsorption. Industrial & Engineering Chemistry Research, 1995. 34(2): p. 591‐598.

131. Xiao, G., Z. Li, T.L. Saleman, and E.F. May, Adsorption equilibria and kinetics of CH4 and N2 on commercial zeolites and carbons. Adsorption, 2017. 23(1): p. 131‐147.

132. Saha, D., Z. Bao, F. Jia, and S. Deng, Adsorption of CO2, CH4, N2O, and N2 on MOF‐5, MOF‐177, and zeolite 5A. Environmental science & technology, 2010. 44(5): p. 1820‐1826.


133. Shang, J., G. Li, R. Singh, Q. Gu, K.M. Nairn, T.J. Bastow, N. Medhekar, C.M. Doherty, A.J. Hill, and J.Z. Liu, Discriminative separation of gases by a “molecular trapdoor” mechanism in chabazite zeolites. Journal of the American Chemical Society, 2012. 134(46): p. 19246‐19253.

134. Li, G.K., J. Shang, Q. Gu, R.V. Awati, N. Jensen, A. Grant, X. Zhang, D.S. Sholl, J.Z. Liu, and P.A. Webley, Temperature‐regulated guest admission and release in microporous materials. Nature communications, 2017. 8: p. 15777.

135. Sumida, K., D.L. Rogow, J.A. Mason, T.M. McDonald, E.D. Bloch, Z.R. Herm, T.‐H. Bae, and J.R. Long, Carbon dioxide capture in metal–organic frameworks. Chemical reviews, 2011. 112(2): p. 724‐781.

136. Dinca, M. and J.R. Long, Strong H2 binding and selective gas adsorption within the microporous coordination solid Mg3 (O2C‐C10H6‐CO2) 3. Journal of the American Chemical Society, 2005. 127(26): p. 9376‐9377.

137. Yuan, W., Y. Lin, and W. Yang, Molecular sieving MFI‐type zeolite membranes for pervaporation separation of xylene isomers. Journal of the American Chemical Society, 2004. 126(15): p. 4776‐4777.

138. Shang, J., G. Li, R. Singh, P. Xiao, J.Z. Liu, and P.A. Webley, Potassium chabazite: a potential nanocontainer for gas encapsulation. The Journal of Physical Chemistry C, 2010. 114(50): p. 22025‐22031.

139. Kuznicki, S.M., V.A. Bell, S. Nair, H.W. Hillhouse, R.M. Jacubinas, C.M. Braunbarth, B.H. Toby, and M. Tsapatsis, A titanosilicate molecular sieve with adjustable pores for size‐selective adsorption of molecules. Nature, 2001. 412(6848): p. 720.

140. Webley, P.A., Seventh Pacific Basin Conference on Adsorption Science and Technology, 24–27 September 2015, Xiamen, China. 2016, Springer.

141. Li, G., P. Xiao, J. Zhang, P.A. Webley, and D. Xu, The role of water on postcombustion CO2 capture by vacuum swing adsorption: Bed layering and purge to feed ratio. AIChE Journal, 2014. 60(2): p. 673‐689.

142. Li, G., P. Xiao, P. Webley, J. Zhang, R. Singh, and M. Marshall, Capture of CO 2 from high humidity flue gas by vacuum swing adsorption with zeolite 13X. Adsorption, 2008. 14(2‐3): p. 415‐422.

143. Göltl, F. and J. Hafner, Alkane adsorption in Na‐exchanged chabazite: The influence of dispersion forces. The Journal of chemical physics, 2011. 134(6): p. 064102.

144. Thrush, K.A. and S.M. Kuznicki, Characterization of chabazite and chabazite‐like zeolites of unusual composition. Journal of the Chemical Society, Faraday Transactions, 1991. 87(7): p. 1031‐1035.

145. Liu, L., R. Singh, G. Li, P. Xiao, P. Webley, and Y. Zhai, Synthesis and adsorption properties of titanosilicates ETS‐4 and ETS‐10 from fly ash. Journal of hazardous materials, 2011. 195: p. 340‐345.

146. Bourgogne, M., J.‐L. Guth, and R. Wey, Process for the preparation of synthetic zeolites, and zeolites obtained by said process. 1985, Google Patents.

147. Liu, B., Y. Zheng, N. Hu, T. Gui, Y. Li, F. Zhang, R. Zhou, X. Chen, and H. Kita, Synthesis of low‐silica CHA zeolite chabazite in fluoride media without organic structural directing agents and zeolites. Microporous and Mesoporous Materials, 2014. 196: p. 270‐276.

148. Van Heest, T., S.L. Teich‐McGoldrick, J.A. Greathouse, M.D. Allendorf, and D.S. Sholl, Identification of metal–organic framework materials for adsorption separation of rare gases: Applicability of ideal adsorbed solution theory (IAST) and effects of inaccessible framework regions. The Journal of Physical Chemistry C, 2012. 116(24): p. 13183‐13195.

149. Kresse, G. and J. Furthmüller, Efficient iterative schemes for ab initio total‐energy calculations using a plane‐wave basis set. Physical Review B, 1996. 54(16): p. 11169‐11186.

150. Kresse, G. and D. Joubert, From ultrasoft pseudopotentials to the projector augmented‐wave method. Physical review b, 1999. 59(3): p. 1758.


151. Jeong, G.‐Y., A.K. Singh, M.‐G. Kim, K.‐W. Gyak, U. Ryu, K.M. Choi, and D.‐P. Kim, Metal‐organic framework patterns and membranes with heterogeneous pores for flow‐assisted switchable separations. Nature communications, 2018. 9(1): p. 1‐9.

152. Sadakiyo, M., T. Yamada, K. Kato, M. Takata, and H. Kitagawa, A significant change in selective adsorption behaviour for ethanol by flexibility control through the type of central metals in a metal–organic framework. Chemical science, 2016. 7(2): p. 1349‐1356.

153. Carrington, E.J., C.A. McAnally, A.J. Fletcher, S.P. Thompson, M. Warren, and L. Brammer, Solvent‐switchable continuous‐breathing behaviour in a diamondoid metal–organic framework and its influence on CO 2 versus CH 4 selectivity. Nature Chemistry, 2017. 9(9): p. 882‐889.

154. Wannapaiboon, S., A. Schneemann, I. Hante, M. Tu, K. Epp, A.L. Semrau, C. Sternemann, M. Paulus, S.J. Baxter, G. Kieslich, and R.A. Fischer, Control of structural flexibility of layered‐pillared metal‐organic frameworks anchored at surfaces. Nature Communications, 2019. 10(1): p. 346.

155. Henke, S., M.T. Wharmby, G. Kieslich, I. Hante, A. Schneemann, Y. Wu, D. Daisenberger, and A.K. Cheetham, Pore closure in zeolitic imidazolate frameworks under mechanical pressure. Chemical science, 2018. 9(6): p. 1654‐1660.

156. Neimark, A.V., F.‐X. Coudert, A. Boutin, and A.H. Fuchs, Stress‐based model for the breathing of metal− organic frameworks. The journal of physical chemistry letters, 2009. 1(1): p. 445‐449.

157. Hazra, A., D.P. van Heerden, S. Sanyal, P. Lama, C. Esterhuysen, and L.J. Barbour, CO 2‐induced single‐crystal to single‐crystal transformations of an interpenetrated flexible MOF explained by in situ crystallographic analysis and molecular modeling. Chemical science, 2019. 10(43): p. 10018‐10024.

158. Evans, J.D., S. Krause, S. Kaskel, M.B. Sweatman, and L. Sarkisov, Exploring the thermodynamic criteria for responsive adsorption processes. Chemical science, 2019. 10(19): p. 5011‐5017.

159. Li, G., J. Shang, Q. Gu, R.V. Awati, N. Jensen, A. Grant, X. Zhang, D.S. Sholl, J.Z. Liu, P.A. Webley, and E.F. May, Temperature‐regulated guest admission and release in microporous materials. 2017. 8: p. 15777.

160. Neimark, A.V., P.I. Ravikovitch, M. Grün, F. Schüth, and K.K. Unger, Pore size analysis of MCM‐41 type adsorbents by means of nitrogen and argon adsorption. Journal of Colloid and Interface Science, 1998. 207(1): p. 159‐169.

161. Wedler, C., K. Lotz, A. Arami‐Niya, G. Xiao, R. Span, M. Muhler, E.F. May, and M. Richter, Influence of Mineral Composition of Chars Derived by Hydrothermal Carbonization on Sorption Behavior of CO2, CH4, and O2. ACS omega, 2020.

162. Couck, S., J.F.M. Denayer, G.V. Baron, T. Rémy, J. Gascon, and F. Kapteijn, An Amine‐Functionalized MIL‐53 Metal−Organic Framework with Large Separa on Power for CO2 and CH4. Journal of the American Chemical Society, 2009. 131(18): p. 6326‐6327.

163. Arami‐Niya, A., G. Birkett, Z. Zhu, and T.E. Rufford, Gate opening effect of zeolitic imidazolate framework ZIF‐7 for adsorption of CH4 and CO2 from N2. Journal of Materials Chemistry A, 2017. 5(40): p. 21389‐21399.

164. Yang, X., A. Arami‐Niya, G. Xiao, and E. May, Flexible Adsorbents at High Pressure: Observations and Correlation of ZIF‐7 Sorption Isotherms for Argon, Nitrogen and Other Gases. Journal of Material Chemistry A 2020. Submitted.

Advanced Adsorbents with High Selectivity and Enhanced ...

Documents