Metal-Organic Frameworks of Pyrazolate Derivatives: Synthesis and Applications for Chemical Sensing, Gas Separation, and Catalysis

Home , Europium hydride

Charlie Eric William Kivi

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Chemistry University of Toronto

Metal-Organic Frameworks of Pyrazolate Derivatives: Synthesis and Applications for Chemical Sensing, Gas Separation, and Catalysis

Charlie Eric William Kivi

Doctor of Philosophy

Department of Chemistry University of Toronto

2017 Abstract

In this work we examine the structure and applications of metal-organic frameworks (MOFs)

synthesized from pyrazolate derivatives. In total, six pyrazolate ligands were examined: 4-(4-

(3,5-dimethyl-1H-pyrazol-4-yl)phenyl)pyridine (HL), 1,1’-methylenebis(3,5-dimethyl-1H-

pyrazolyl-4-carboxylic acid) (H2BPM), triethyl-1,1',1''-methanetriyltris(3,5-dimethyl-1H-

pyrazole-4-carboxylate) (TPM-1), triethyl-4,4',4''-(methanetriyltris[3,5-dimethyl-1H-pyrazole-

1,4-diyl])tribenzoate (TPM-2), 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5-dimethyl-1H- pyrazole-4-carboxylic acid) (H4TPP), tetraethyl-1,1',1'',1'''-(1,4-phenylenebis[methanetriyl])

tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylate)pyridinephenylpyrazole (TPX). These ligands

were used to synthesize a variety of MOFs that were subsequently investigated for luminescent

sensing, gas storage, and catalytic performance.

In Chapter 2, three MOFs synthesized from the reaction of HL and CuX (X = Br, I) were

investigated. These materials were found to form MOFs which contained the luminescent

trinuclear Cu(I) pyrazolate unit. Only CuBr and HL alone resulted in a MOF of sufficient purity

and stability to permit evaluation as a luminescent sensor. Experiments with various volatile

organic compounds revealed turn-on (luminescence enhancement) behaviour for ethyl acetate,

pentane, and benzene and turn-off (luminescence quenching) behaviour for diethyl ether and chloroform for this material.

In Chapters 3 and 4, ten MOFs are described. These were synthesized from H2BPM and metal- acetate salts. Chapter 3 encompassed two MOFs: [Ni(BPM)]n∙xDMSO and

[Cd(BPM)]n∙xDMSO. These MOFs are isostructural, microporous materials. Evacuation of the

pores led to collapse of the [Cd(BPM)]n∙xDMSO material but [Ni(BPM)]n∙xDMSO proved to be

permanently porous. Further investigations revealed [Ni(BPM)]n∙xDMSO selectively adsorbed

methane over nitrogen indicating it may serve as a porous material for coal mine methane

capture. Chapter 4 investigated the many other structures H2BPM can assume in the presence of manganese, cobalt, iron, copper, and europium. Cobalt in particular proved highly tunable with 4 separate structures being synthesized such as [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O). This

was achieved through varying the synthetic conditions.

[Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O)and [Eu(BPM)(OAc)]n were further investigated for

the oxidation of olefins and luminescence respectively.

In Chapter 5, metal complexes Mo(TPM-1)(CO)3, Mo(TPM-2)(CO)3, Pd2(H4TPP)Cl4, and

Pd2(TPX)Cl4 were synthesized. The catalytic performance of Mo(TPM-1)(CO)3 was evaluated

and found to be effective for the oxidation of olefins. MOF synthesis was attempted with all four

compounds but was ultimately unsuccessful. Although one MOF was produced from Mo(TPM-

1)(CO)3, large scale synthesis was not possible preventing full investigation of its properties.

iii

Acknowledgments

I would like to thank my doctoral supervisor Professor Datong Song for his guidance throughout the last few years. Also thanks to my internal committee members Professor Robert Morris and Professor Ulrich Fekl who provided invaluable direction and assistance. Lastly I would like to acknowledge my examination committee members, Professor Bernie Kraatz, Professor Jik Chin, and external examiner Professor Suning Wang, who provided insightful feedback for the final version of this thesis.

Several people assisted with the work contained within this thesis. Thanks to Andrew Proppe of the Ted Sargent Group for quantum yield and luminescent lifetime measurements of

[Cu9L6Br2][CuBr2]. My collaborators at the University of Calgary, Professor George Shimizu

and Benjamin Gelfand, assisted with gas sorption measurements of [Ni(BPM)]n∙xDMSO. Our collaborators at the University of Ottawa, Professor Tom Woo and Hana Dureckova, contributed

to computational modeling of [Ni(BPM)]n∙xDMSO. I also had the privilege of working with a stellar undergraduate student, Cindy Ma. After some initial guidance, Cindy seemed to make every metal acetate in our lab form a coordination polymer or MOF. Cindy was also able to deduce the correct procedure to synthesize [Ni(BPM)]n∙xDMSO which required an unintuitive trisolvent mixture. Following Cindy’s procedures Sheree Zhang and Michaela Deng helped make bulk samples of [Ni(BPM)]n∙xDMSO for characterization and gas sorption testing. Riley Choi aided further investigation into iron coordination polymers and Jingning Zhou cracked the

synthesis for [Eu(BPM)(OAc)]n. Bulk synthesis of coordination polymers and MOFs synthesized from H2BPM (compounds 7-13) for characterization was conducted by Yujin Yamamotto and, again, Michaela Deng. Lastly, I would like to thank Quiming (Walter) Liang

who worked tirelessly on synthesizing derivatives of Mo(TPM-1)CO3 (14) in an effort to make

an isoreticular TPM MOF series. Walter’s work allowed for the synthesis of Mo(TPM-2)CO3 (15) and investigation of its catalytic properties.

A special thanks to all the Song Group members past and present: Runyu Tan, Yu Li, Tao Bai, Shaolong Gong, Xiaofei Li, Kim Osten, Rhys Batcup, Tara Cho, Celia Gendron-Herndon, Adam Pantaleo, Yanxin Yang, Trevor Janes, Fred Chiu, Ellen Yan, and Daniel Dalessandro. You all livened up the lab. An extra special thanks to my undergraduate students, Cindy Ma and Walter Liang, and my other volunteer students who were wonderful to work with. iv

No PhD can be successful without support from outside of the lab. Therefore, I would like to thank Anna Liza Villavelez (grad office), Ken Greaves (chem stores), John Ford (machine shop supervisor), Jack O’Donnell (glass shop), Ahmed Bobat (machine shop – creator of our MOF heating blocks), Darcy Burns (NMR facilities manager), Dmitry Pichugin (NMR facility), Rose Balazs (EA technician), Jack Jackiewicz (electronics shop), and Alan Lough (x-ray).

It also takes a lot of effort to get into the grad program to begin with. To that end, a big thank you to Professor Tom Baker for allowing me the opportunity to work in your lab as an inexperienced second year undergrad at the University of Ottawa. Thank you to Dr. Daniel Harrison for being an amazing post-doc when I came back to Dr. Baker’s lab as a more experienced fourth year student. And thanks to Professor George Shimizu and Dr. Simon Iremonger who mentored me as a third year exchange student at the University of Calgary working in the wonderful world of MOFs. You all sparked my interest in MOFs and inspired me to make the leap to grad school.

I also wouldn’t have been able to complete this doctorate without the support of my friends (both inside and outside the lab) and my family. I’m particularly proud, and thankful, that my papa, Edwin Kivi, will be around to see the first “Doctor Kivi” in the family. I will be able to answer “Yes!” when he asks if I’ve graduated yet! Thank you to my sister, Michelle, and brother in law, Richard Jagielowicz, for being there for me through thick and thin. Thank you to my parents for their love and support especially my mom, Catherine, for encouraging my interest in the natural world, and my father, Eric, for showing me how to think outside the box.

Lastly, thanks to my special someone, Anneliese Neumann, for the long walks, companionship, and interesting lab book entries.

Table of Contents

Acknowledgments...... iv

Table of Contents ...... vi

List of Tables ...... x

List of Schemes ...... xii

List of Equations ...... xiii

List of Figures ...... xiv

List of Appendices ...... xxviii

List of Abbreviations and Symbols...... xxix

Chapter 1 ...... 1

1 Introduction ...... 1

1.1 Introduction to Metal-Organic Frameworks ...... 1

1.1.1 Definition of a Metal-Organic Framework ...... 1

1.1.2 Early Examples of MOFs...... 2

1.1.3 Applications of MOFs...... 3

1.2 Chemical Sensing in MOFs ...... 4

1.2.1 Introduction to Chemical Sensing ...... 4

1.2.2 VOC sensing in luminescent MOFs ...... 5

1.3 Gas Storage and Selective Adsorption in MOFs ...... 6

1.3.1 Gas Storage in MOFs ...... 6

1.3.2 Selective Gas Adsorption in MOFs ...... 9

1.4 Catalysis in MOFs...... 11

1.4.1 Introduction to Catalytic MOFs ...... 11

1.4.2 Metalloligands for Catalytic MOFs ...... 12

1.4.3 Reactions at Metal Nodes ...... 13

1.5 Thesis Scope and Objectives ...... 14

Chapter 2 ...... 17

2 Trinuclear Cu(I) Pyrazolate-based MOFs for Volatile Organic Compound Sensing ...... 17

2.1 Abstract ...... 17

2.2 Introduction ...... 17

2.2.1 Applications for VOC Sensing ...... 17

2.2.2 Luminescent Trinuclear Cu(I) Pyrazolate MOFs ...... 18

2.3 Results and Discussion ...... 18

2.3.1 Preparation of HL ...... 18

2.3.2 Luminescent Cationic Metal–Organic Framework featuring [Cu–Pyrazolate]3 Units for Volatile Organic Compound Sensing ...... 19

2.3.3 β-CuBr MOF ...... 30

2.3.4 α-CuI MOF ...... 34

2.3.5 Independent Synthesis of Small Molecule Cu3L3 Unit...... 37

2.3.6 Conclusions ...... 39

2.4 Experimental ...... 39

2.4.1 Synthesis ...... 39

2.4.2 Experimental Properties and Supporting Information ...... 46

2.4.3 X-Ray Diffractometry ...... 56

2.4.4 Luminescent Sensing ...... 59

Chapter 3 ...... 70

3 A New MOF constructed from Carboxylate Functionalized Bispyrazolylmethane for Coal Mine Methane Capture ...... 70

3.1 Abstract ...... 70

3.2 Introduction ...... 70

3.3 Results and Discussion ...... 72

3.3.1 Synthesis and Properties of H2BPM Ligand ...... 72 vii

3.3.2 Synthesis and Crystal Structure Description...... 72

3.3.3 Gas Sorption Studies of 5 ...... 77

3.3.4 Conclusions ...... 82

3.4 Experimental ...... 83

3.4.1 General Considerations ...... 83

3.4.2 Ligand Synthesis ...... 83

3.4.3 Synthesis of MOFs ...... 87

3.4.4 Thermal Gravimetric Analysis, Powder X-Ray, and IR ...... 88

3.4.5 Gas Sorption Studies ...... 90

3.4.6 Computational Studies ...... 97

3.4.7 X-Ray Diffractometry ...... 101

Chapter 4 ...... 104

4 Structural Investigation of Bispyrazolylpropane MOFs ...... 104

4.1 Abstract ...... 104

4.2 Introduction ...... 104

4.3 Results and Discussion ...... 105

4.3.1 Methods of Assessing BPM Ligand Conformation ...... 105

4.3.2 Cobalt Coordination Polymers ...... 106

4.3.3 BPM MOFs of Manganese, Iron, Copper, Zinc, and Europium ...... 119

4.3.4 Investigated Properties ...... 124

4.3.5 Conclusions ...... 126

4.4 Experimental ...... 127

4.4.1 General Considerations ...... 127

4.4.2 Synthesis of MOFs ...... 127

4.4.3 Powder X-Ray and TGA Characterization ...... 129

4.4.4 X-Ray Diffractometry ...... 133 viii

Chapter 5 ...... 138

5 Towards Catalytic MOFs Utilizing Tris- and Tetrapyrazolyl Ligands ...... 138

5.1 Abstract ...... 138

5.2 Introduction ...... 138

5.2.1 Known Trispyrazolylmethane and Tetrapyrazolyl Compounds ...... 138

5.3 Results and Discussion ...... 141

5.3.1 Functionalized Molybdenum Trispyrazolylmethane Tricarbonyl Complexes ....141

5.3.2 Tetrapyrazolyl Complexes as Ligands for MOF Synthesis ...... 147

5.3.3 Conclusions ...... 150

5.4 Experimental ...... 151

5.4.1 General Considerations ...... 151

5.4.2 Ligand Synthesis ...... 151

5.4.3 Catalytic Conditions and Methodology ...... 163

5.4.4 MOF Synthesis...... 163

5.4.5 X-Ray Diffractometry ...... 164

Chapter 6 ...... 167

6 Conclusions, Outlook, and Future Work ...... 167

6.1 General Overview ...... 167

6.2 Chapter 2 ...... 167

6.3 Chapter 3 ...... 168

6.4 Chapter 4 ...... 168

6.5 Chapter 5 ...... 169

Chapter 7 References ...... 171

About the Author ...... 181

Appendices ...... 183

List of Tables

Table 2.1: Examples of MOF synthesis test conditions utilizing 4 as a starting material...... 38

Table 2.2: Results of Quantum Yield Measurements for 1...... 48

Table 2.3: Fitting parameters for the luminescent lifetimes of 1 with different solvent vapour and observation times...... 51

Table 2.4: Luminescent lifetimes for 1 monitored for various periods under N2 (as-synthesized), ethyl acetate vapour and water vapour. All lifetimes are in nanoseconds...... 52

Table 2.5: Crystallographic properties of MOFs 1 to 3...... 57

Table 2.6: Crystallographic properties of Cu3L3 (4)...... 58

Table 2.7: Solvochromic effects on 1’s luminescence over sequential experiments...... 65

Table 3.1: Monte Carlo simulated coal mine methane working capacities and selectivities for 123,124 reported MOFs in comparison to fully activated [Ni(BPM)]n...... 82

Table 3.2: Correlation coefficient for isotherms fits and calculating isosteric heat of adsorption

using the Clausius-Clapeyron equation for [Ni(BPM)]n∙xDMSO (5)...... 91

Table 3.3: Post adsorption elemental analysis of 5...... 96

Table 3.4: N2-NIMF Force Field parameters: bl is the distance of the atom to the molecular mass centre. Lorentz-Berthelot mixing rules were used to determine parameters between atoms of different types...... 100

Table 3.5: Crystallographic properties of 5 and 6...... 102

2 Table 4.1: Metal-metal distances and μ -OH2 bonding angles for dinuclear cobalt building units of 7, 8 and 9 in comparison to analogous literature compounds...... 119

Table 4.2: Crystallographic Properties of Cobalt Coordination Polymers 7-10...... 134

Table 4.3: Crystallographic properties of other BPM MOFs 11-13 and

[Fe2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (FeBPM Ladder)...... 136

Table 4.4 Selected properties of the BPM ligand for compounds 7 through 13...... 137

Table 5.1: Crystallographic properties of tetrapyrazolyl ligands...... 164

Table 5.2 Crystallographic properties of metallo-ligand complexes 14, 16, and 17...... 165

Supplemental Table

Supplemental Table A.1: Metal-Carbonyl stretching frequencies for 14, 14-(COOH)3, and 15 demonstrating the similarity of the molybdenum coordination environment...... 190

List of Schemes

Scheme 2.1: Synthesis of HL...... 19

Scheme 2.2: Synthesis of the α-CuBr MOF (1) under nitrogen atmosphere...... 20

Scheme 2.3: Synthesis of the β-CuBr MOF (2)...... 31

Scheme 2.4: Synthesis of the α-CuI MOF (3)...... 35

Scheme 3.1: Synthesis of ligand H2BPM...... 72

Scheme 5.1: Synthesis of TPM-1 and TPM-2...... 141

Scheme 5.2: Synthesis of metalloligands 14 and 15...... 142

Scheme 5.3: Synthesis of 14-(COOH)3 from 14 and subsequent MOF synthesis...... 144

Scheme 5.4: Synthesis of H4TPP in two steps...... 147

Scheme 5.5: Synthesis of palladium complex 16 from H4TPP...... 148

Scheme 5.6: Synthesis of TPX and palladium complex 17...... 149

xii

List of Equations

Equation 2.1: Percent difference versus water of MOF luminescence intensity...... 24

Equation 2.2: Fit for luminescent lifetimes of 1 where I0 is the fitting offset, k1 and k2 are the

decay rates, A1 and A2 are the exponential coefficients, σ is the width of the excitation pulse, and

t0 is the time-zero offset. Erf is the error function...... 51

Equation 2.3: Equations for the determination of crystallographic confidence factors R1 and wR2...... 58

Equation 3.1: Fitting parameter for gas sorption isotherms...... 91

Equation 3.2: The bulk phase chemical potential of an ideal gas...... 98

Equation 3.3: Probability of a molecular translation/rotation within the pore...... 99

Equation 3.4: Probability of addition of a new gas molecule to the pore...... 99

Equation 3.5: Probability of removal of a gas molecule from the pore...... 99

Equation 3.6 Selectivity of gas 1 versus gas 2 where μ = gas uptake and p = partial pressure. . 101

xiii

List of Figures

Figure 1.1: Some possible structures accessible using organic linkers and metal nodes...... 1

Figure 1.2: a) SBU of MOF-5 showing 4 zinc atoms, 1 μ4-O, and 6 carboxylate groups from 1,4- benzenedicarboxylate. Hydrogen and other ligand carboxylate groups are not shown. b) MOF-5 viewed down the a-axis showing porous cavities...... 2

Figure 1.3: a) SBU of HKUST-1 showing 4 copper atoms, 4 carboxylate groups from 1,3,5- benzenetricarboxylate, and two apically coordinated solvent water. Hydrogen and other ligand carboxylate groups are not shown. b) HKUST-1 viewed down the a-axis showing porous cavities...... 3

Figure 1.4: a) Jablonski diagram showing typical processes that result in fluorescence and phosphorescence. b) A scheme for the various energy and emission pathways that would result in luminescence within a MOF...... 6

Figure 1.5: Potential energy diagram for physisorption versus chemisorption...... 7

2+ Figure 1.6: a) Asymmetric unit of CO2 adsorbing MOF showing 2 Zn , 2 CO2, oxalate, and 2 3-

amino-1,2,4-triazole ligands. b) Crystal structure of MOF viewed down the b-axis showing CO2 interacting with amino groups...... 8

Figure 1.7: Grams of CO2 produced per MJ of combustion energy (ΔHc) for various n-alkyl compounds...... 9

Figure 1.8: Scheme for selective adsorption of CO2 in a MOF (post-combustion carbon capture) resulting in an N2 enriched gas stream...... 10

Figure 1.9: a) Locations of possible active sites within a MOF. b) General scheme for MOF catalysis where a substrate (green) enters the pore, undergoes a reaction, and exits as product (red, blue, yellow)...... 11

Figure 1.10: a) 5,10,15,20-tetrakis(4-(pyrazolate-4-yl)-phenyl-porphyrin with coordinated nickel. b) MOF pore showing location of Ni-porphyrin. Other faces and vertices coloured yellow for clarity. c) Catalytic C-H bond halogenation with Mn-porphyrin MOF...... 13 xiv

Figure 1.11: MOF catalyzed decomposition of chemical warfare agents...... 14

Figure 1.12: Ligands synthesized and employed for the formation of new MOFs...... 15

Figure 2.1: Binding modes of HL showing copper and ancillary coordination sites...... 19

+ Figure 2.2: a) [Cu3L3] building block of 1; [Cu3Br3] cluster and pore anion omitted for clarity. + b) [Cu3Br2] building unit of 1; [Cu3L3] cluster, protons, and pore anion omitted for clarity...... 21

Figure 2.3: The ‘sandwich’ repeating units of 1 with the top and bottom [Cu3L3] units colour coded in purple and red, respectively. The left image shows the side view of the unit (middle ligands truncated to just the pyridine nitrogen), and right shows the top view of the ‘sandwich’ repeating unit (down the a+c direction)...... 21

Figure 2.4: Trigonal prismatic channels in 1 along the a+c direction...... 22

Figure 2.5: Average fluorescence excitation and emission spectra of a bulk sample of vacuum dried 1 after PXRD structure-activity experiments (see section 2.4.4.1) monitored at a

wavelength of 582 nm and 468 nm respectively under N2...... 23

Figure 2.6: Average emission intensity of 1 after VOC treatment (EtOA – red) relative to average pre-exposure water vapour emission (blue) maximum using an excitation wavelength of 469 nm as determined by Equation 2.1. Note significant increase of emission intensity...... 24

Figure 2.7: VOC sensing results measured as percent difference compared to water-treated 1 (Equation 2.1). Excitation wavelength is 469 nm and response is monitored at the emission maxima of each solvent (Table 2.7). Errors are +/- one standard deviation...... 25

Figure 2.8: PXRD pattern comparison of freshly synthesized 1 (blue) sequentially treated with water vapour (red), ethyl acetate vapour (purple), water vapour (green), and ethyl acetate vapour (orange). Note loss of major peaks at 6.11°, 6.29° and the emergence of a new peak at 6.90° upon water exposure and restoration of the original structure on exposure to ethyl acetate...... 27

Figure 2.9: Sequential exposure of 1 to dry and wet THF following standard procedures. Sample illuminated at 469 nm and monitored at the peak emission of THF (559 nm)...... 28

Figure 2.10: PXRD comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with chloroform vapour (red)...... 29

Figure 2.11: PXRD comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with diethyl ether vapour (pink)...... 30

I Figure 2.12: a) [Cu3L3] unit of 2; disordered atoms, bridging [Cu2Br] groups, and Cu Br2 anion + not shown. b) [Cu2Py3Br] unit of 2; coordinating pyridine truncated to their nitrogen donors (blue). All other atoms omitted...... 31

I + Figure 2.13: “Weave” of [Cu 2Py6Br] units showing the lower (yellow) ligand arm reaching up between the ligands of the purple unit...... 32

Figure 2.14: a) Parallelepiped channels down the b direction of 2; b) all other directions are closed packed (a direction shown)...... 33

Figure 2.15: Luminescent excitation and emission spectra of a bulk sample of vacuum dried 2 monitored at a wavelength of 620 nm and 410/469 nm respectively...... 34

Figure 2.16: The ‘sandwich’ repeating unit with the top and bottom [Cu3L3] units colour coded in green and red, respectively. The middle ligands truncated to just the pyridine nitrogen, copper is yellow, and iodide is purple. Left image depicts a shows side view of unit and the right image shows a top down view of the ‘sandwich’ repeating unit along the a+c direction...... 36

Figure 2.17: Trigonal prismatic channels in 3 shown along the a+c direction...... 36

Figure 2.18: Crystal structure of 4 showing small molecule [Cu3L3] unit. Co-crystallized solvent

(CHCl3) not shown...... 37

1 Figure 2.19: H NMR spectrum in CDCl3 of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole. .... 41

13 Figure 2.20: C NMR spectrum in CDCl3 of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole. ... 41

1 Figure 2.21: H NMR spectrum of HL in DMSO-d6...... 43

13 Figure 2.22: C NMR spectrum of HL in DMSO-d6...... 43

xvi

1 Figure 2.23: H NMR spectrum of 4 in CDCl3. Note the coordination of CDCl3/CHCl3 to 4

(representing 0.8 CHCl3 per molecule) as seen by the shoulders at 8.62 and 7.71 ppm. N-H peak not visible due to water present in solvent...... 45

13 Figure 2.24: C NMR spectrum of 4 in CDCl3...... 46

Figure 2.25: Quantum yield of as-synthesized 1 with excitation at 375 nm...... 47

Figure 2.26: Quantum yield of water vapour treated 1 with excitation at 375 nm...... 47

Figure 2.27: Quantum yield of ethyl acetate vapour treated 1 with excitation at 375 nm...... 48

Figure 2.28: Observed luminescence decay for ethyl acetate treated 1 over 26 μs with a biexponential fit and an R2 value of 0.9708...... 49

Figure 2.29: Observed luminescence decay for as-synthesized 1 over 26 μs with a biexponential fit and an R2 value of 0.6953...... 49

Figure 2.30: Observed luminescence decay for water treated 1 over 26 μs with a biexponential fit and an R2 value of 0.2695...... 50

Figure 2.31: Theoretical versus observed PXRD pattern for 1...... 53

Figure 2.32: Theoretical versus observed PXRD pattern for 2...... 53

Figure 2.33: Mass change upon heating 1 after water vapour treatment for 24 hours. There is a 3.5% mass loss (assigned as pore/coordinating water) between 40 and 200 °C with the greatest mass loss occurring between 100 and 140 °C. Structure decomposition begins at ~325 °C...... 54

Figure 2.34: TGA spectrum of β-CuBr MOF (2)...... 55

Figure 2.35: TGA spectrum of α-CuI MOF (3)...... 55

Figure 2.36: Infrared spectrum of neat sample of 1...... 56

Figure 2.37: Cuvette and VOC exposure setup for luminescent gas sensing experiments...... 59

xvii

Figure 2.38: Configuration of luminescence spectrometer for luminescent gas sensing experiments...... 60

Figure 2.39: Averaged emission intensity of 1 after VOC treatment (pentane – green) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm...... 61

Figure 2.40: Averaged emission intensity of 1 after VOC treatment (benzene – orange) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm...... 62

Figure 2.41: Averaged emission intensity of 1 after VOC treatment (acetone – red) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm...... 62

Figure 2.42: Averaged emission intensity of 1 after VOC treatment (wet THF – orange, dry THF - green) relative to averaged pre-exposure water vapour emission (blue for wet THF, purple for dry THF) using an excitation wavelength of 470 nm...... 63

Figure 2.43: Averaged emission intensity of 1 after VOC treatment (MeCN – red) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm...... 63

Figure 2.44: Averaged emission intensity of 1 after VOC treatment (MeOH – purple) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm...... 64

Figure 2.45: Averaged emission intensity of 1 after VOC treatment (Et2O – green) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm...... 64

Figure 2.46: Averaged emission intensity of 1 after VOC treatment (CHCl3 - orange) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm...... 65

xviii

Figure 2.47: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with ethyl acetate vapour (purple)...... 67

Figure 2.48: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with pentane vapour (red)...... 67

Figure 2.49: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with acetone vapour (black)...... 68

Figure 2.50: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with acetonitrile vapour (orange)...... 68

Figure 2.51: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with tetrahydrofuran vapour (purple)...... 69

Figure 2.52: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with methanol vapour (grey)...... 69

Figure 3.1: Diagram of H2BPM with carboxylate (orange), flexible methylene linker (pink), and potential pyrazolyl chelate site (blue) indicated...... 72

Figure 3.2: a) Building unit of [Ni(BPM)]n∙xDMSO, 5, showing the ligands coordinating Δ to the nickel centre. BPM ligands are truncated with solvent and hydrogen atoms omitted for clarity. b) 128.5° bend of BPM chelated to Ni; purple = carboxylate bound BPM. c) 117° bend of carboxylates bound to Ni; yellow = chelating BPM...... 73

Figure 3.3: a) Smallest assembly of metal nodes within 5 containing 4 nickel centres and 4 BPM ligands; BPM ligands truncated for clarity. b) Assembly of rings (colour coded) to form large diameter pores within 5; BPM truncated for clarity, nickel centres shown as large colour coded spheres...... 74

Figure 3.4: a) 5 viewed down the b direction showing large (20.5 Å x 4.1 Å) and small (3.6 Å x 3.6 Å) pores. b) 5 viewed down the a direction showing similar large and small pores. c) 5 viewed down the c direction showing exclusively small pores. All rotation, rotoinversion, and screw axis in 5 follow the c direction. Note in all cases pore solvent is omitted...... 75

xix

Figure 3.5: a) Building unit of [Cd(BPM)]n∙xDMSO, 6; BPM ligands truncated for clarity. b) 6 viewed down the b direction showing large and small pores...... 76

Figure 3.6: PXRD pattern of [Cd(BPM)]n∙xDMSO (6) showing rapid structural change upon desolvation of the material. Depicted are the predicted pattern (yellow), a fresh sample of 6 still wet with DMSO (blue), and partially (orange) and fully (grey) air dried solids...... 77

Figure 3.7: Selected gas uptake plots for methane adsorbing onto 5. The full figure is shown in section 3.4...... 78

Figure 3.8: IAST determined selectivity for a 20% CO2/80% N2 gas stream at 298 K adsorbing onto 5...... 79

Figure 3.9: IAST determined selectivity for a 1:1 mixture of CH4 /N2 at 298 K adsorbing onto 5...... 79

Figure 3.10: Comparison of IAST derived dual component gas uptake to Monte Carlo predicted gas uptake for 5 with no residual pore solvent...... 80

Figure 3.11: Monte Carlo simulated coal mine methane selectivities for reported MOFs123,124

versus fully activated [Ni(BPM)]n...... 81

1 Figure 3.12: H NMR spectrum of II in CDCl3...... 84

13 Figure 3.13: C NMR spectrum of II in CDCl3...... 85

1 Figure 3.14: H NMR spectrum of H2BPM in DMSO-d6...... 86

13 Figure 3.15: C NMR spectrum of H2BPM in DMSO-d6...... 86

Figure 3.16: TGA spectrum of partially dried [Ni(BPM)]n∙xDMSO (5)...... 88

Figure 3.17: TGA spectrum of partially dried [Cd(BPM)]n∙xDMSO (6)...... 88

Figure 3.18: IR Spectrum of neat [Ni(BPM)]n∙xDMSO MOF (5)...... 89

Figure 3.19: IR Spectrum of air dried [Cd(BPM)]n∙xDMSO (6) submitted for EA. Note changes in relative peak intensity due to collapsed structure of 6 in comparison to 5...... 89

Figure 3.20: PXRD pattern of as-synthesized [Ni(BPM)]n∙xDMSO (5)...... 90

Figure 3.21: BET_des surface area for N2 at 77 K within 5...... 92

Figure 3.22: BET_des surface area for CO2 at 195 K within 5...... 92

Figure 3.23: mmol of CO2 adsorbed per gram of 5 versus pressure...... 93

Figure 3.24: mmol of CH4 adsorbed per gram of 5 versus pressure...... 93

Figure 3.25: mmol of N2 adsorbed per gram of 5 versus pressure...... 94

Figure 3.26: Heat of absorption for CO2 within 5...... 94

Figure 3.27: Heat of Absorption for CH4 within 5...... 95

Figure 3.28: Heat of absorption for N2 within 5...... 95

Figure 3.29: Post adsorption TGA of 5 showing 15.80% mass loss at 250°C: 11.84% from DMSO, 3.96% from surface moisture (sample left in air for ~2 weeks before analysis)...... 96

Figure 4.1: a) Scheme for the assessment of angle (α) within the BPM moiety measured between the carboxylate carbon atoms and methylene linker. b) Scheme for the assessment of pyrazolyl torsion angle (β) measured from the four nitrogen atoms of the pyrazolyl rings...... 106

Figure 4.2: a) The dinuclear building unit of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O), 7; methyl and methylene groups on pyrazolyl rings and symmetry related elements omitted for clarity. b) Coordination network “ladder” viewed down the a+b+c direction, c) “ladder” viewed down the b+c direction; yellow = BPM, dinuclear cobalt node, blue = Bpy...... 107

Figure 4.3: a) Four adjacent “ladders” of 7 viewed down the a-axis showing staggered conformation. b) Four adjacent ladders viewed down the b+c direction; non-coordinating solvent omitted for clarity. c) Adjacent “ladders” viewed down the a+b+c direction showing role of coordinated solvent (hydrogen omitted for clarity) in formation of hydrogen bonded network between 1D ladders...... 108

Figure 4.4 TGA analysis of air dried 7 showing stepwise loss of solvent...... 109

xxi

Figure 4.5: a) The dinuclear building unit of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O), 8; methyl and methylene groups on pyrazolyl rings and symmetry related elements omitted for clarity. b) Adjacent “ladders” viewed down the a+b+c direction showing role of coordinated solvent (hydrogen omitted for clarity) in formation of hydrogen bonded network between 1D ladders. Note that all possible positions of the disordered DMSO molecules are shown...... 110

Figure 4.6: TGA analysis of air dried 8 showing stepwise loss of solvent. Note that 1.4 molecules of solvent water have already evaporated upon air drying...... 110

Figure 4.7: a) Dinuclear building unit of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9); methyl and methylene groups on pyrazolyl rings omitted for clarity. b) 1D chain of dinuclear-BPM “diamonds” viewed down the b direction and c) “diamonds” viewed down the a+c direction. 112

Figure 4.8: a) Hydrogen bonding (purple) linking two adjacent dinuclear cobalt nodes of 9 together; methyl and methylene groups on pyrazolyl rings and DMF not involved in hydrogen bonding omitted for clarity. b) 4 adjacent 1D chains viewed down the a+c direction. Note that chains hydrogen bond to form 2D sheets along the b direction. c) “Diamond” units of 1D chains viewed down the b direction with non-coordinating solvent shown...... 113

Figure 4.9: TGA analysis of air dried 9 showing stepwise loss of solvent. Note that the mass contribution of all coordinating water is equal to one DMF (3.56%). As such, water may be a component of the assigned mass losses...... 114

Figure 4.10: a) Building unit of [Co(BPM)(H2O)(DMSO)]n (10); methyl and methylene groups on pyrazolyl rings and hydrogen omitted for clarity. b) Parallel 1D chains of 10 propagating along the c direction with c) limited hydrogen bonding (purple) between adjacent 1D chains (shown as red and green in b)...... 115

Figure 4.11: TGA analysis of air dried 10 showing initial stepwise loss of solvent. After ~200 °C observed mass changes no longer correlated well to expected mass changes based on known composition of 10 derived from other characterization methods...... 116

Figure 4.12: a) Dinuclear building unit of the Fonari MOF; benzene dicarboxylate truncated for clarity. b) Coordination network “diamonds” of the Fonari MOF viewed down the a+b direction; yellow = benzene dicarboxylate, green = cobalt...... 118

xxii

Figure 4.13: a) 9 viewed down the b direction; adjacent 1D chains in orange and blue, cobalt centre shown in purple. b) Coordination network “diamonds” of Fonari MOF viewed down the a+b direction; yellow = benzene dicarboxylate, green = cobalt...... 118

Figure 4.14: a) Dinuclear manganese node of [Mn(BPM)(H2O)2]n MOF (11); methyl and methylene groups on pyrazolyl rings and bridging BPM-carboxylate omitted for clarity. b) Schematic representation of dinuclear manganese node. c) Pyrazolyl-carboxylate bridge between adjacent nodes along the a direction; methyl and methylene groups on pyrazolyl rings omitted for clarity. c) Staggered confirmation of BPM viewed down the c direction; hydrogen omitted for clarity. d) 3D MOF structure of 11 showing close packed nature of material (no voids present)...... 120

Figure 4.15: a) Copper paddlewheel structure of [Cu(BPM)(DMSO)]n MOF (12); methyl and methylene groups on pyrazolyl rings omitted for clarity. b) A single 2D sheet viewed down the a+b+c direction with c) interpenetration of 12; parallel 2D sheets (blue, purple) are orthogonally interpenetrated (green, yellow) leaving no accessible cavities. d) View of interpenetrating 2D sheets viewed down the a+b direction showing paddlewheel unit located inside cavities of the interpenetrated 2D sheet (green)...... 121

Figure 4.16 a) Europium nodes of [Eu(BPM)2(OAc)]n MOF (13); methyl and methylene groups on pyrazolyl rings omitted for clarity. Note that due to orientation of chains the second pyrazolyl-carboxylate is hidden behind the first as shown by b) a schematic representation. c) Linkages between 1D europium metal chains via BPM on both the top (red) and bottom (green) of 2D sheet c) resulting 2D sheet on the bc plane...... 122

Figure 4.17 TGA spectrum of [Eu(BPM)2(OAc)]n MOF (13)...... 123

Figure 4.18: a) Dinuclear building unit of [Fe2(BPM)2(H2O)4(Bpy)0.5]n; methyl and methylene groups on pyrazolyl rings and symmetry related Bpy unit omitted for clarity. b) Four adjacent “ladders” viewed down the a direction showing staggered confirmation; non-coordinating solvent not shown...... 124

Figure 4.19: Excitation and emission spectra for 13 monitored at 615 nm and irradiated at 393 nm respectively showing predominantly europium-centred excitations...... 126

xxiii

Figure 4.20: TGA spectrum of [Mn(BPM)(H2O)2]n MOF (11). Note slow loss of coordinated water between 100 and 200 °C...... 129

Figure 4.21: TGA spectrum of [Cu(BPM)(DMSO)]n MOF (12). The absence of mass loss prior to decomposition indicates the non-porous, interpenetrated nature of the MOF...... 130

Figure 4.22: PXRD pattern of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (7)...... 130

Figure 4.23: PXRD pattern of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O) (8)...... 131

Figure 4.24: PXRD pattern of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9)...... 131

Figure 4.25: PXRD pattern of [Co(BPM)(H2O)(DMSO)]n (10)...... 132

Figure 4.26: PXRD pattern of [Mn(BPM)(H2O)2]n (11)...... 132

Figure 4.27: PXRD pattern of [Cu(BPM)(DMSO)]n (12)...... 133

Figure 4.28: PXRD pattern of [Eu(BPM)2(OAc)]n (13)...... 133

Figure 5.1: Possible sites for tuning/derivatization of the TPM ligand; R = different functional groups for metal node coordination...... 139

Figure 5.2: Similarities and differences of H4TPP and TPX ligands with desired locations for metal coordination sites indicated...... 140

Figure 5.3: a) Mo(TPM-1)(CO)3 complex (14) side view and b) top down view showing triangular arrangement of the disordered ester groups...... 142

Figure 5.4: Observed catalytic performance of 14; blue line is a visual aid only. The inset shows the catalytic reaction conditions (1.5 eq of tertbutylhydrogen peroxide [5.5 M in decane]; n-butyl ether was used as an internal standard) and observed products...... 143

Figure 5.5: Catalytic performance of 14 versus 15...... 144

Figure 5.6: a) Dinuclear building unit of MOF synthesized from Mo(TPM-1)(CO)3 in two steps;

ligand truncated for clarity. b) MOF asymmetric unit showing intact Mo(TPM-1)(CO)3 and coordinated zinc. c) Model of the MOF showing pores propagating down the b direction and d)

xxiv

space filling model of MOF showing small pores down b direction. Solvent molecules omitted for clarity...... 146

Figure 5.7: Crystal structures of a) H4TPP ligand and b) Pd2(H4TPP)Cl4 (16) complex; solvent molecules not shown...... 148

Figure 5.8: Crystal structure of a) TPX ligand and b) Pd2(TPX)Cl4 (17) palladium complex; solvent not shown...... 150

1 Figure 5.9: H NMR spectrum of TPM-1 in CD2Cl2...... 152

13 Figure 5.10: C NMR spectrum of TPM-1 in CD2Cl2...... 152

1 Figure 5.11: H NMR spectrum of 14 in CD2Cl2...... 153

13 Figure 5.12: C NMR spectrum of 14 in CD2Cl2...... 154

1 Figure 5.13: H NMR spectrum of B in CDCl3...... 156

13 Figure 5.14: C NMR spectrum of B in CDCl3...... 157

1 Figure 5.15: H NMR spectrum for H4TPP in DMSO-d6...... 158

13 Figure 5.16: C NMR spectrum for H4TPP in DMSO-d6...... 158

1 Figure 5.17: H NMR spectrum of 16 in DMSO-d6. Excess water present in the solvent suppresses the 12.33 ppm acid peak signal...... 159

13 Figure 5.18: C NMR spectrum of 16 in DMSO-d6...... 160

1 Figure 5.19: H NMR spectrum of TPX in CDCl3...... 161

13 Figure 5.20: C NMR spectrum of TPX in CDCl3...... 161

1 Figure 5.21: H NMR spectrum of 17 in CDCl3...... 162

13 Figure 5.22: C NMR spectrum of 17 in CDCl3...... 163

xxv

Supplemental Figures

Supplemental Figure A.1: a) Samples of 1 exposed to 365 nm light after exposure to water vapour (orange, left) and ethyl acetate (yellow, right). Ethyl acetate (yellow, left) and water vapour (orange, right) exposed samples of 1 side by side under b) normal light and c) 365 nm UV light...... 183

Supplemental Figure A.2: CIE colour representation of ethyl acetate vapour treated 1 with an excitation wavelength of 469 nm...... 184

Supplemental Figure A.3: a) Prepared air free flask for the synthesis of 1 followed by b) multiday heating protected from light on a thermocouple controlled hot plate with c) resulting crystals at the end of the synthesis of 1. Crystal size, yield, and purity is dependent on stirring (this sample was not stirred)...... 184

Supplemental Figure A.4: Typical MOF reaction setup with 2 dram vials in an aluminum heating block. A similar setup was used for larger, 20 mL vials for larger scale synthesis...... 185

Supplemental Figure A.5: Schematic representation of coal mine methane (CMM) extraction using pressure swing adsorption...... 185

Supplemental Figure A.6: IR spectrum of [Co2(BPM)2(H2O)4(Bipy)0.5]n∙2DMF∙2(H2O) (7). ... 186

Supplemental Figure A.7: IR spectrum of [Co2(BPM)2(H2O)4(Bipy)0.5]n∙2DMSO∙2(H2O) (8). 186

Supplemental Figure A.8: IR spectrum of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9)...... 187

Supplemental Figure A.9: IR spectrum of [Co(BPM)(H2O)(DMSO)]n (10)...... 187

Supplemental Figure A.10: IR spectrum of [Mn(BPM)(H2O)2]n (11)...... 188

Supplemental Figure A.11 IR spectrum of [Cu(BPM)(DMSO)]n (12)...... 188

Supplemental Figure A.12: IR spectrum of [Eu(BPM)2(OAc)]n (13)...... 189

Supplemental Figure A.13: CIE colour representation of as synthesized 13 with an excitation wavelength of 393 nm...... 189

xxvi

Supplemental Figure B.1: The Song Group Spring 2012...... 190

Supplemental Figure B.2: The Song Group Fall 2012...... 191

Supplemental Figure B.3: The Song Group Fall 2013...... 191

Supplemental Figure B.4: The Song Group Spring 2015...... 192

Supplemental Figure B.5: The Song Group Fall 2015...... 192

Supplemental Figure B.6: The Song Group Spring 2017...... 193

Supplemental Figure B.7: The Song Group (plus alumnus Vince) Summer 2017...... 193

xxvii

List of Appendices

Appendices ...... 183

A Supplemental Information ...... 183

A.1 Chapter 2 ...... 183

A.2 Chapter 3 ...... 185

A.3 Chapter 4 ...... 186

A.4 Chapter 5 ...... 189

B The Song Group: 2011 through 2017 ...... 190

xxviii

List of Abbreviations and Symbols

HL 4-(4-(3,5-dimethyl-1H-pyrazol-4-yl)phenyl)pyridine H2BPM 1,1’-methylenebis(3,5-dimethyl-1H-pyrazolyl-4-carboxylic acid) TPM-1 Triethyl-1,1',1''-methanetriyltris(3,5-dimethyl-1H-pyrazole-4-carboxylate) TPM-2 Triethyl-4,4',4''-(methanetriyltris(3,5-dimethyl-1H-pyrazole-1,4- diyl))tribenzoate H4TPP 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4- carboxylic acid) TPX Tetraethyl-1,1',1'',1'''-(1,4-phenylenebis(methanetriyl))tetrakis(3,5-dimethyl-1H- pyrazole-4-carboxylate) Bpy 4,4’-bipyridine SBU Secondary building unit Bpy-N-oxide 4,4’-bipyridine-N-oxide OAc Acetate Å Angstrom, 10-10 m Anal. Calcd Calculated elemental analysis based on a formula δ Chemical shift ° Degree DFT Density functional theory DCM Dichloromethane Et2O Diethyl ether DMSO Dimethylsulfoxide DMF N,N-dimethylformamide THF Tetrahydrofuran eq Equivalent ν Frequency, stretching frequency (IR) κ Kappa, prefix for a coordinating heteroatom μ Mu, prefix for a bridging ligand 3[MMCT] Metal-to-ligand charge transfer 3[MLCT], Metal-to-ligand charge transfer 3[LMCT], Ligand-to-metal charge transfer 3[LLCT], Ligand-to-ligand charge transfer 3[XLCT] Halide-to-ligand charge transfer Pz Pyrazolate Py Pyridine [Cu3Pz3] Trinuclear Copper (I) Pyrazolate Chromophore + [Cu3Br2] Cationic Tricopper Dibromide metal node + [Cu2Br] Cationic Dicopper Bromide metal node NMR nuclear magnetic resonance J nuclear spin-spin coupling constant s/d/t/m singlet/doublet/triplet/multiplet (NMR) ppm parts per million Ph phenyl π pi orbital tBu tert-butyl UV ultraviolet xxix

Chapter 1 1 Introduction 1.1 Introduction to Metal-Organic Frameworks 1.1.1 Definition of a Metal-Organic Framework

Metal-organic frameworks (MOFs) are materials comprised of metal nodes joined together by organic linkers. MOFs may be either two or three dimensional materials. Two dimensional (2D) MOFs are composed of repeating metal-ligand units that propagate along two directions to form sheets of material. 2D MOFs typically interpenetrate to form a pseudo three dimensional (3D) structure. Interpenetration occurs when one framework grows within another. MOFs are a subset of porous coordination polymers (PCPs) as the MOF is held together via coordinate bonds. Note that while all MOFs are coordination polymers, not all coordination polymers are MOFs. This is the case with simple 1D polymers (Figure 1.1).1 Additionally, if the metal node’s geometry and composition has been demonstrated in a variety of MOFs, it is known as a secondary building unit (SBU). Once identified, SBUs’ allow for the prediction of MOF structures from the known geometry of the metal nodes.2,3

Figure 1.1: Some possible structures accessible using organic linkers and metal nodes.

1 2

1.1.2 Early Examples of MOFs

Although there are examples of what are now recognized as metal-organic frameworks very early in the literature (e.g. a 1959 report of a Bis(adiponitrilo)copper(I) Nitrate MOF 4) the field of MOF chemistry truly began in the mid to late 1990’s with the seminal work done by Yaghi (MOF-5 and others)5,6 and Williams (HKUST-1).7 These early MOFs were highly influential as they demonstrated the first SBUs which spawned many successor materials. Shown below is the zinc-oxo cluster SBU of MOF-5 (Figure 1.2) and the copper-paddlewheel structure of HKUST-1 (Figure 1.3).

The porous nature of MOFs allows for diverse chemistry to be performed within them. Both MOF-5 and HKUST-1 hinted at this future due to their large pores and high surface areas. One of the earliest demonstrated properties of MOFs were the accessibility of their pores to gases and solvents allowing for storage and seperation.5–7 Since these early reports, the scope of chemistry possible within MOFs has exploded to encompass many potential applications. These include luminescent sensing8 and catalysis9 which will be discussed in detail in subsequent sections.

1.1.3 Applications of MOFs

As a consequence of the well-defined architecture of MOFs, which results in large surface areas, large pore volumes, and predictable composition, MOFs lend themselves to many potential uses.8–10 The focus of this thesis will be three main applications for MOFs. They are as follows:

1) Chemical sensing via incorporation of luminescent functionality into the MOF (see Section 1.2 and Chapter 2).8

2) Gas storage and separation through selective physisorption of gas molecules into the porous MOF structure (see Section 1.3 and Chapter 3).10

3) Catalysis by incorporation of catalytically active sites within the framework (see Section 1.4 and Chapter 5).9

Additional MOF applications include their use as proton carriers for fuel cells,11 atmospheric water harvesting,12 certain medical devices (e.g. enzyme encapsulation13,14 and biocomposites15), in molecular magnetism,16 as switchable mechanically interlocked materials,17 and as electronic devices.18 These applications are beyond the scope of this thesis. For further reading, a number of excellent review articles have been published in a special issue of Chemical Society Reviews entitled “Metal-organic frameworks and porous polymers – current and future challenges”.8–10,19

1.2 Chemical Sensing in MOFs 1.2.1 Introduction to Chemical Sensing

Chemical sensing occurs when a target analyte induces a measurable change in the material or device being used as a sensor (e.g. a MOF). While the detection of liquid reagents and dissolved ions are possible, the most relevant form of sensing for MOF applications is the detection of volatile organic compounds (VOCs). VOCs are present at industrial sites and leaks pose a risk to both the environment and worker’s health. Therefore, inexpensive, reliable detection of low concentration VOCs is desirable.

At the moment, the state of the art for VOC detection is via Photo Ionization Detection (PID).20,21 PID works by using high energy photons (typically ultraviolet) to ionize analyte gases which then produce a measurable current within the device. Although other solid state VOC detectors are also used,22 PID is the quickest and simplest method for monitoring toxic vapours in industrial settings. However, certain VOCs, such as halogenated compounds, are difficult to detect due to the high ionization energy required.

Conversely, luminescent detection of VOCs does not require ionization. Instead, the VOC interacts with the luminescent sites of the MOF to cause observable luminescence emission changes.8,23 How the VOC interacts with both the chromophore and MOF structure would dictate the range of detectable VOCs as opposed to relying on ionization. Therefore, having a low

energy, luminescent detector would be far more versatile as it could detect a wider array of VOCs.

1.2.2 VOC sensing in luminescent MOFs

Our group was interested in the ability of MOFs to perform as luminescent chemical sensors for VOCs.8,23–25 While there are many recent examples, most known luminescent MOF sensors for VOCs rely on ligand-based26,27 or lanthanide-based28,29 emissions.23,25,30,31

Luminescent detection requires the target analyte (in our case a VOC) to change the intrinsic luminescence of the material (in this case a MOF). This can occur in three general ways: luminescence enhancement, luminescence quenching, and/or red/blue shifting the emission maxima. Of these three mechanisms, luminescence enhancement (classified as turn-on sensing) is the most desirable. It is the least likely to result in a false positive and is the easiest to detect. As described by Mortellaro and Nocera: “the obvious benefit of [turn-on sensing] is that the analyte triggers a light signal against a dark background to achieve higher sensitivities”.32

The general mechanism by which luminescence can be adjusted is by alteration of the excited state of the material by the VOC. For example, luminescence quenching may occur via interactions between an analyte (e.g. chloroform) and an accessible chromophore (e.g. anthracene). This facilitates a charge transfer and non-radiative excitation decay.33 Luminescence enhancement may occur via analyte interactions (e.g. [H+]) altering the charge transfer states of a complex to disfavour non-radiative transfers (e.g. 3LLCT – triplet ligand-ligand charge transfer) and promote emissive pathways (e.g. 3MLCT – triplet metal-ligand charge transfer).34 Our group previously reported a Europium MOF as a turn-on sensor for dimethyl formamide (DMF).28 This MOF was constructed from europium and 2′,5′-bis(methoxymethyl)-[1,1′:4′,1′′-terphenyl]-4,4′′- dicarboxylate. Adsorption of DMF vapour into the MOF’s pores locked the confirmation of the terphenyl rings. Locking of the rings leads to a more efficient ligand to metal exciton transfer resulting in luminescence enhancement.28 For a full description of the many effects that can alter MOF luminescence see Figure 1.4 which is adapted from the literature.8

1.3 Gas Storage and Selective Adsorption in MOFs 1.3.1 Gas Storage in MOFs

MOFs’ permanent porosity led to gas storage and separation as one of their earliest investigated applications.6 Since then, there have been ongoing efforts to improve the volumetric and gravimetric uptake of gases in MOFs for storage purposes. To this end, large pore MOFs with greater internal surface areas are frequently being developed as requirements for carbon capture 35,36 37,38 (CO2) and US Department of Energy Targets for energy storage (CH4 and H2) become more aggressive.

The key mechanism allowing MOFs to be applied for gas storage is physisorption of the target gas (Figure 1.5). Physisorption is adsorption governed by non-bonding physical interactions between the gas and the MOF pore as opposed to the formation of formal chemical bonds. For example, a reported MOF constructed from Zn2(Atz)2(ox) (Atz = 3-amino-1,2,4-triazole, ox = 39 oxalate) (Figure 1.6a) displayed excellent adsorption characteristics for carbon dioxide (CO2).

Within the pores of Zn2(Atz)2(ox) amino nitrogen lone pairs from Atz interacted with the partial + positive (δ ) carbon CO2 in order to favour adsorption of CO2 at low pressures and ambient temperature.39 Physisorption in general is a low energy process as no bonds need to be formed or broken as reliance is on interactions between the pores of the MOF and the target substrate.10 This is in sharp contrast to chemisorption where energy is expended for bond formation and

cleavage between the substrate and MOF. One such example is the reversible capture and release

of bromine and chlorine from a MOF constructed from Co2Cl2BTDD (BTDD = bis(1H-1,2,3- triazolo[4,5-b],[4,5-i])dibenzo[1,4]dioxin). Chlorine and bromine can reversibly coordinate to the CoII nodes of the MOF via a redox process. However, regeneration of the MOF requires reduction of the nodes back to CoII under forcing conditions (temperatures in excess of 150 °C while under vacuum).40 In another example of chemisorption within a MOF, bromine and iodine halogenated the unsaturated bonds of extended 4,4’-[1,4-phenylenebis(ethyne-2,1-diyl)]- dibenzoate ligands used to construct a zirconium MOF.41 However, this reaction was irreversible preventing the release of dihalide.41

Distance from Surface

Physisorption Chemisorption Potential energyPotential to scale) (not

Figure 1.5: Potential energy diagram for physisorption versus chemisorption.

2+ Figure 1.6: a) Asymmetric unit of CO2 adsorbing MOF showing 2 Zn , 2 CO2, oxalate, and 2 3-

amino-1,2,4-triazole ligands. b) Crystal structure of MOF viewed down the b-axis showing CO2 interacting with amino groups.

Although MOFs have been applied for the storage of a number of gases, carbon dioxide and methane garner the most attention. Methane and CO2 are both potent greenhouse gases and are produced as a consequence of industrial processes and energy production. For example, global release of carbon is approaching 10 000 million metric tons per year.42 Storage of these gases

within MOFs serves two main goals: sequestration of CO2 for carbon capture and increasing the safety and efficacy of methane storage for use in transportation.

At present, carbon capture is optimized through the use of chemisorption. A CO2 gas stream is passed through aqueous amine solvents.43,44 However, there is a large “parasitic” energy cost 45 associated with regeneration of the amine solvents. Since MOFs sequester CO2 via physisorption the “parasitic” energy requirements for regeneration are greatly reduced. Typically, reduced pressure and mild heating are sufficient for regeneration making MOFs ideal for carbon capture.46

Additionally, lack of technology to store methane means that methane produced at remote sites or under low pressure (e.g. coal mines) is usually vented to the atmosphere or burned as a waste gas. However, per unit carbon, methane is the cleanest hydrocarbon fuel available on Earth

(Figure 1.7)47. Safe, cost efficient means to store methane could significantly remove one of the impediments to its use as a fuel.48

/MJ) /MJ)

2 55

(g CO 50 produced per unit per energy produced 2

CO 45 0 2 4 6 8 Number of n-Alkyl Carbons

Figure 1.7: Grams of CO2 produced per MJ of combustion energy (ΔHc) for various n-alkyl compounds.

1.3.2 Selective Gas Adsorption in MOFs

In tandem with the storage of gases in MOFs, selective capture of gases is frequently investigated.49 A MOF that is a high performing storage material often favours one gas over another. This is a result of the enthalpy of adsorption, uptake, polarizability, and kinetic diameter

of gases not being identical. For example, N2, CO2, and CH4 have kinetic diameters of 3.64 Å, 3.30 Å, and 3.80 Å respectively.50 Therefore, a MOF with pore apertures of only 3.5 Å will be

able to adsorb CO2 but not N2 or CH4. This effect is known as molecular sieving and is even

more pronounced when comparing gases with larger differences (e.g. water vs NH3 or benzene vs toluene).10 As a consequence, MOFs may be designed to favour one gas over another. In this manner binary gas mixtures may be separated to produce pure gases with very low energy requirements.

At present, the most common way gas separation is accomplished industrially is through Pressure Swing Adsorption (PSA).51,52 For PSA, a column of adsorbent (i.e. MOF) is first pressurized with a gas mixture. As the gas mixture diffuses through the column, one component

is selectively adsorbed (e.g. CO2 for post-combustion carbon capture). Once the column becomes saturated, the gas steam is switched to a parallel adsorbent column to maintain continual flow. The pressure is reduced in the saturated column to desorb and recover the captured gas and regenerate the column. In this manner purity of the gas stream (mobile phase) or captured gas may be increased to the desired level using the MOF (Figure 1.8). Molecular sieving (size 39 selective separation), favourable dipole interactions (e.g. Zn2(Atz)2(ox)), etc. all impact the selectivity of a MOF.10 For some specific examples, as mentioned above, amine functionalized 39 MOFs may selectively capture CO2 through lone pair amine interactions with CO2. The polarizability of small molecule alkanes may also allow for a greater affinity for the MOF framework for certain molecules, with larger alkanes such as propane being favoured over ethane or methane.53 This affinity is a result of polarizability of the molecular orbitals of alkanes to favour interactions within the MOF with a) the metal nodes, b) the ligands, and c) other gases within the pore.54 Lastly, pore aperture size (molecular sieving) may be optimized to allow for a specific target gas to enter the MOF preferentially.55

Figure 1.8: Scheme for selective adsorption of CO2 in a MOF (post-combustion carbon capture)

resulting in an N2 enriched gas stream.

1.4 Catalysis in MOFs 1.4.1 Introduction to Catalytic MOFs

As MOFs incorporate both ligands and metal sites they are also attractive materials for catalysis.9,56 Reactions may occur at the metal nodes themselves through either direct substrate coordination or via an outer sphere reaction or at reactive ligands (Figure 1.9).9 MOFs also allow for the heterogenizing of homogeneous catalysts: whereby a catalytically active metal complex is incorporated into the MOF’s framework (e.g. MnIII-porphyrin).57 Metallation may occur either pre- or post-synthetically to produce a reactive metalloligand. Use of these homogeneous catalysts as metalloligands allows MOFs to exploit the advantages of being a heterogeneous material (e.g. ease of catalyst recovery, high catalytic site density) potentially without compromising the efficiency of the catalysts themselves. Incorporation of reactive complexes with chiral active sites allows for enantioselectivity to be imparted to the substrate. Furthermore, due to their nature as porous materials, any such catalysis is inherently selective: only materials that can enter the pores may be acted upon. For instance, small narrow pores will restrict the substrate size.

1.4.2 Metalloligands for Catalytic MOFs

Porphyrin, salen (N,N′-ethylenebis(salicylimine)), and binol (1,1'-bi-2-naphthol) are three ligands that have seen great success when used as metalloligands incorporated into catalytically active MOFs.56 The binol and salen family of ligands in particular were capable of achieving enantioselective catalysis.56,58 The first example of a MOF incorporating a metallo-salen ligand was synthesized by Hupp et. al., who used a pyridine functionalized manganese-salen complex to crosslink 2D sheets constructed from Zn2+ and 4,4’-biphenyldicarboxylate.59 Although this MOF was interpenetrated, it still possessed large pore volumes and was capable of the asymmetric epoxidation of 2,2-dimethyl-2H-chromene with greater than 80 % enantiomeric excess.59 The pioneering work for binol functionalized MOFs was performed by Lin et. al.60,61 Although their original research focus was on hydrogen storage,60 Lin et. al. quickly realized the catalytic potential of their MOF system. (R)-6,6‘-dichloro-2,2‘-dihydroxy-1,1‘-binaphthyl-4,4‘- 61 bipyridine, L, may be reacted with CdCl2 to form a MOF with a composition Cd3Cl6L3. The binol ligand can then be post-synthetically metallated with tetrakis(iso-propoxy)titanium (IV) to i 61 form a catalytically active site composed of TiL(O Pr)2. Once metallated, the binol-MOF becomes a highly active asymmetric catalyst for the conversion of aromatic aldehydes to chiral secondary alcohols.61

The first microporous, catalytically active MOF to incorporate porphyrin was also engineered by Hupp et. al.62 Since then, research into catalytically active MOFs featuring one or more porphyrin moieties has continued to see rapid advancement. Nagaraja et. al. reported a Ni(II)- porphyrin MOF made of 2D sheets stacking together through π-π interactions. This supramolecular framework was capable of photocatalytically reducing various nitroaromatic compounds to their corresponding amines using visible light at room temperature.63 More recently, another new nickel porphyrin MOF was reported which proved to be highly stable under basic aqueous conditions. This is notable as many MOFs are not stable towards aqueous basic or acidic conditions. Exchange of porphyrin coordinated Ni2+ for Mn3+ allowed the MOF to effectively catalyze the halogenation of the C-H bonds in cyclopentane and cyclohexane with high yields (Figure 1.10).57

1.4.3 Reactions at Metal Nodes

Another main mechanism by which MOFs may be used as catalysts is through catalysis at the building units (metal nodes) that comprise the MOF’s structure.19 A recent development in this area is the utilization of MOFs for the decomposition of chemical warfare agents.19 As noted by Farha et. al. and Navarro et. al. the zinc oxo and zirconium oxo-hydroxo secondary building units of certain MOFs mimics the Zn-OH-Zn active site of phosphotriesterase enzymes.64,65 It was therefore observed that the zinc and zirconium oxo-hydroxo metal nodes in certain MOFs serve as effective catalysts for the hydrolytic decomposition of Ethyl ({2-[bis(propan-2- yl)amino]ethyl}sulfanyl)(methyl)phosphinate (VX) and sarin gas analogues (Figure 1.11).64 It was noted that for sarin gas a MOF catalyzed partial oxidation is also possible although over oxidation is a concern as the doubly oxidized compound remains a potent toxin.64 These MOFs

are just two examples of materials that can effectively neutralize stockpiles of chemical weapons 19 and other toxic compounds such as SOx and NOx.

O O hydrolysis P Me N P Me N + HO S SH O O

VX nerve agent

hydrolysis HO OH + HCl Cl Cl S S Sulphur mustard oxidation Cl Cl S O

Figure 1.11: MOF catalyzed decomposition of chemical warfare agents.

It should be noted that decomposition of chemical warfare agents is not the limit of building unit centered catalysis in MOFs. Many other examples have been realized such as catalytic cyanosilylation of benzaldehyde in activated HKUST-1.56 Other applications are continually under investigation as new materials are discovered.66

1.5 Thesis Scope and Objectives

Pyrazole is a very versatile ligand and can play many different roles in chemical systems. However, it remains relatively underexplored in MOFs. Most reports of MOFs incorporating pyrazole use it as a terminal LX type ligand (such as the catalytic porphyrin example cited above57) ignoring the rich, versatile chemistry of pyrazole ligands.67–69 The aim of this work was to synthesize new MOFs by exploiting the versatility of pyrazole-based ligand systems.

We investigated the incorporation of pyrazole-based ligands into MOFs in four main areas: 1) synthesis of luminescent materials for volatile organic compound sensing based on trinuclear CuI MOFs utilizing a monopyrazolate ligand (Chapter 2); 2) synthesis of gas storage materials based on MOFs incorporating a bispyrazolylmethane ligand (Chapter 3); 3) an in-depth exploration of the coordination chemistry of a bispyrazolylmethane ligand in a variety of MOFs (Chapter 4); and 4) synthesis of catalytically active MOFs incorporating trispyrazolylmethane, tetrapyrazolyl-

p-xylene, and 1,1’,3,3’-tetrapyrazolylpropane ligands (Chapter 5). In all cases, neutral pyrazole- alkyl species were targeted (as opposed to pyrazolyl-borates) for ease of synthesis, characterization, and purification of the ligand intermediates. In order to form bonds with metal nodes, and thus synthesize a MOF, carboxylates and pyridyl groups were incorporated into the pyrazolyl ligands due to their simplistic structure and bonding, ease of synthesis, and well- studied characteristics. In total, six pyrazolate ligands were examined: 4-(4-(3,5-dimethyl-1H- pyrazol-4-yl)phenyl)pyridine (HL), 1,1’-methylenebis(3,5-dimethyl-1H-pyrazolyl-4-carboxylic acid) (H2BPM), triethyl-1,1',1''-methanetriyltris(3,5-dimethyl-1H-pyrazole-4-carboxylate) (TPM-1), triethyl-4,4',4''-(methanetriyltris[3,5-dimethyl-1H-pyrazole-1,4-diyl])tribenzoate (TPM-2), 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylic acid) (H4TPP), Tetraethyl-1,1',1'',1'''-(1,4-phenylenebis[methanetriyl])tetrakis(3,5-dimethyl-1H- pyrazole-4-carboxylate)pyridinephenylpyrazole (TPX) (Figure 1.12).

H HO O O OH O O O O O NH O N O N N N N N N N O O O N N N N N N N N (HL) (TPM-1) H N N N N N N N N O O O O N O N N N N N OH N N O HO N N O O HO O O OH O O O O (H2BPM) (TPM-2) (H4TPP) (TPX)

Figure 1.12: Ligands synthesized and employed for the formation of new MOFs.

Chapter 6 provides conclusions and future directions for the four research chapters.

The experimental work herein was performed by the author including collection of all x-ray crystallographic data. Single crystal x-ray structures were solved by the author and Professor Datong Song. In Chapter 2, quantum yield and luminescent lifetime measurements of

[Cu9L6Br2][CuBr2] were conducted by Andrew Proppe of the Ted Sargent Group at the

University of Toronto. MOF trials for CuL3 were assisted by undergraduate students Cindy Ma

and Stefan Jevtić. In Chapter 3, gas sorption measurements for [Ni(BPM)]n∙xDMSO and

[Cd(BPM)]n∙xDMSO were performed by Benjamin Gelfand under supervision of Professor George Shimizu at the University of Calgary. Computer simulations for the gas uptake of

[Ni(BPM)]n∙xDMSO were performed by Hana Dureckova under the supervision of Professor

Tom Woo at the University of Ottawa. Synthesis of [Ni(BPM)]n∙xDMSO was initially performed

by Cindy Ma under the author’s supervision with later assistance by high school volunteer Sheree Zhang and undergraduate volunteer Michaela Deng. For Chapter 4 compounds 7 through 12 were initially synthesized by Cindy Ma under the author’s supervision with later assistance by undergraduate volunteers Riley Cho, Michaela Deng, and Yujin Yamamotto. Synthesis of

[Eu(BPM)(OAc)]n was initially performed by Jingning Zhou under the author’s supervision. In

Chapter 5 compound TPM-2 and Mo(TPM-2)(CO)3 were initially synthesized and characterized by undergraduate student Quiming (Walter) Liang under the author’s supervision.

Parts of Chapter 2 have been published and Chapters 3 through 5 of this thesis are in preparation for publication:

Chapter 2: Charlie E. Kivi, Datong Song, “A luminescent cationic metal–organic framework featuring [Cu–pyrazolate]3 units for volatile organic compound sensing” Dalton Transactions, 2016, 45, 17087-17090. This paper details the luminescence properties and VOC sensing performance of 1.

Chapter 3: Charlie E. Kivi, Benjamin Gelfand, Hana Dureckova, Cindy Ma, Datong Song, George K. H. Shimizu, Tom Woo, “Coal Mine Methane Capture with a New Nickel MOF” In Preparation. This paper details the selective gas sorption properties of 5 with regards to coal mine methane capture. This work is being prepared for publication pending completion of computational studies.

Chapter 4: Charlie E. Kivi, Cindy Ma, Datong Song, “Bis(3,5-dimethyl-4-carboylic acid)pyrazolylmethe as a Versatile Ligand for MOF and Coordination Polymer Synthesis” In Preparation. This paper details the new structures and properties of BPM coordination polymers and MOFs. Submission is being withheld pending publication of our more impactful work regarding coal mine methane capture with 5.

Chapter 5: Charlie E. Kivi, Datong Song, “Three New Metal Complexes for Catalytic Transformations”, In Preparation. This work is being prepared for publication pending further catalytic testing of the metalloligands.

Chapter 2 2 Trinuclear Cu(I) Pyrazolate-based MOFs for Volatile Organic Compound Sensing 2.1 Abstract

A new luminescent MOF, [Cu9L6Br2][CuBr2] (1), incorporating trinuclear Cu(I) pyrazolate

[Cu3Pz3] units shows promise as a sensor for hydrophobic VOCs. 1 was synthesized via the reaction of pyridinephenylpyrazole (HL) and CuBr in DMF under air free conditions. Exposure of 1 to ethyl acetate, benzene, and pentane resulted in luminescence enhancement whereas chloroform and ethyl ether had a quenching effect as measured versus water treated 1. Hydrophilic solvents had no significant effect on the observed luminescence of 1. Based on literature precedence, the sensing mechanism should be a combination of metal-to-metal charge transfer 3[MMCT], metal-to-ligand charge transfer 3[MLCT], and halide-to-ligand charge transfer 3[XLCT] excited states. Two other MOFs were synthesized using HL: a CuBr MOF possessing a different structure (2) and an isostructural CuI MOF (3). However, these materials were not evaluated for VOC sensing.

2.2 Introduction 2.2.1 Applications for VOC Sensing

As discussed at length in Chapter 1, MOFs have garnered recent attention due to their many applications. Of particular interest to us is the use of MOFs as chemical sensors for volatile organic compounds (VOCs).8,23–25 While there are many recent examples, most known luminescent MOF sensors for VOCs rely on ligand-based26,27 or lanthanide-based28,29 emissions.23,25,30,31

In contrast, luminescence involving the transition metal centres (e.g. metal-ligand charge transfer)70,71 remains relatively underexplored in MOFs research. Some known examples include Lin’s RuII-bipyridine MOF72 and Hanuza’s CrIII MOF.73 As such, we were very interested in how we could expand this field through the use of pyrazolates.

2.2.2 Luminescent Trinuclear Cu(I) Pyrazolate MOFs

I The trinuclear Cu –pyrazolate (Cu3Pz3) chromophore caught our attention due to its proven sensing ability in homogeneous solution74 and the active role of the CuI centres in 75–77 I 77,78 79–81 luminescence. Simple luminescent Cu 3Pz3 complexes consisting of prisms , cages and oligomers82 have all been reported. To our knowledge, there are only six published MOF I 83–90 types that utilize Cu 3Pz3 building blocks. In the first instance, using commercially available bipyrazole, a neutral interpenetrated MOF was synthesized that exhibited luminescence83 and gas sorption ability.84 Interestingly, this MOF demonstrated that simple aromatic molecules had a templating effect for MOF synthesis.84 A classical pillared-paddlewheel MOF synthesized from 4-(4-carboxylatophenyl)-3,5-dimethyl pyrazolate was reported but this MOF displays no 85 luminescence properties. In situ reduction of trinuclear Cu(II) pyrazolates also allows Cu3Pz3 MOFs to be accessed.86 The reduced MOF is exceptionally water stable and remains porous but luminescence activity was not reported in the paper.86

The most relevant Cu3Pz3 containing MOFs for VOC sensing are those with a pyridyl (Py) functional group on the ligand.87–90 By using pyridine, secondary CuI clusters may be constructed to produce MOFs with ancillary luminescent sites.87–90 4-(pyrid-4’-yl)-3,5-dimethylpyrazole and 3,5-diethyl-4-(4-pyridyl)pyrazole have been used to produce two luminescent MOFs which rely on Cu(CN) and CuX (X = halide) or Cu2I2 clusters respectively as their secondary building units (SBUs).87,88 3-(4-pyridyl)-5-p-tolyl-pyrazolate) and its derivatives have also been used to 89,90 produce luminescent MOFs using tetranuclear Cu4I4 clusters as the SBU. However, in all four cases, VOC sensing was not demonstrated. Indeed, only thermochromism87,89 and a “chemopalette effect” (doping the MOF with different coordinating solvents during synthesis)90 I were reported. All the known Cu 3Pz3-containing MOFs have charge neutral skeletons with no reported influence of the species within the MOF’s pores on luminescent behaviour.83–90

2.3 Results and Discussion

2.3.1 Preparation of HL

To construct the Cu3Pz3 unit we desired in our MOFs, a new ligand, 4-(4-(3,5-dimethyl-1H- pyrazol-4-yl)phenyl)pyridine (HL), was synthesized in a two-step process. Using literature procedures91 3-(4-bromophenyl)pentane-2,4-dione was first prepared. This dione was then

reacted with hydrazine monohydrate to produce 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole. Lastly, a Suzuki coupling reaction was performed to afford HL (Scheme 2.1). This ligand, once deprotonated, features three nitrogen donors: one neutral and one anionic on the pyrazolate group I and a distant pyridine. In this fashion, the pyrzaole ring may bind to Cu to form Cu3Pz3 while the L-donor pyridine remains available for coordination to additional metal nodes (Figure 2.1).

N Br 1eq K PO O 3 4 1.5eq PyB(OH)2 N2H4 0.5 mol% Pd(PPh3)4 O HN MeOH dioxane/water HN N 110C 36h N Br HL 84.9 % Yield ~40 % Average Yield

Scheme 2.1: Synthesis of HL.

Figure 2.1: Binding modes of HL showing copper and ancillary coordination sites.

2.3.2 Luminescent Cationic Metal–Organic Framework featuring [Cu– Pyrazolate]3 Units for Volatile Organic Compound Sensing 2.3.2.1 Synthesis and Characterization

Yellow crystals of [Cu9L6Br2][CuBr2] (α-CuBr MOF, 1) can be obtained in moderate yields by heating CuBr and HL in dry, degassed DMF under a nitrogen atmosphere at 135 °C for four I days (Scheme 2.2). This MOF features the Cu 3Pz3 chromophore and a cationic skeleton with I − [Cu Br2] counterions in the pores and crystallizes in the Monoclinic C2/c space group.

Br Cu Br

N N Cu Cu N N N N Cu N

Br N N CuBr Cu Cu Cu DMF N N 135°C Br

N NH N Cu N N N Cu Cu N N

Scheme 2.2: Synthesis of the α-CuBr MOF (1) under nitrogen atmosphere.

Two building blocks can be found in the skeleton of 1: (a) trigonal [Cu3L3] units (Figure 2.2a), + similar to known [Cu3Pz3] compounds, and (b) trigonal bipyramidal shaped [Cu3Br2] clusters with the two bromides occupying the apical positions (Figure 2.2b). The repeating unit of the cationic skeleton of 1 can then be described as a ‘sandwich’ with the trigonal bipyramidal shaped I + I [Cu 3Br2] clusters sandwiched between two staggered trigonal [Cu 3L3] layers; each apical I + I I bromide of the [Cu 3Br2] cluster interacting with all three Cu centres of one [Cu 3L3] unit I (Figure 2.3). It is important to note that the other apical site of the [Cu 3L3] unit remains open.

+ Figure 2.2: a) [Cu3L3] building block of 1; [Cu3Br3] cluster and pore anion omitted for clarity. + b) [Cu3Br2] building unit of 1; [Cu3L3] cluster, protons, and pore anion omitted for clarity.

Each repeating unit is linked to six adjacent units through the coordination of the pyridine I nitrogen donor from one [Cu3Pz3] unit (see blue pyridine nitrogens in Figure 2.3) to the Cu + centre in the middle of the [Cu3Br2] unit, resulting in the formation of a 3D framework. There

are trigonal prismatic channels along the a+c direction of the crystal lattice, where the disordered counterions and solvent molecules reside (Figure 2.4).

Figure 2.4: Trigonal prismatic channels in 1 along the a+c direction.

2.3.2.2 Luminescent Properties of α-CuBr MOF 1

In order to verify the luminescent properties of the [Cu3Pz3] unit within MOF 1, the identity of the MOF was first verified via powder x-ray diffraction (PXRD). MOF samples were then degassed under high vacuum over 48 hours then analyzed via luminescent spectroscopy. The luminescence measurements were conducted inside a quartz cuvette equipped with a septum top under pseudo-air free conditions via nitrogen purging of the cuvette. Measurements were recorded on a QuantaMaster spectrofluorometer. These samples show broadband excitation and emission spectra in the solid state with maxima at 469 and 560 nm, respectively (Figure 2.5) and a quantum yield of 8.6% (Figure 2.25).

0.8

0.6 Excitation Emission

Intensity (a.u.) 0.4

0.2

0 300 400 500 600 700 800 Wavelength (nm)

Figure 2.5: Average fluorescence excitation and emission spectra of a bulk sample of vacuum dried 1 after PXRD structure-activity experiments (see section 2.4.4.1) monitored at a wavelength of 582 nm and 468 nm respectively under N2.

Treatment of 1 with water vapour to eliminate any residual channel organic solvents results in a decrease of quantum yield to 1.8% (Figure 2.26). Based on the observed quenching of the luminescence of the luminescence of 1 by water we opted to use the luminescence intensity of water treated 1 as our zero point for subsequent VOC sensing experiments.

2.3.2.3 VOC Sensing Experiments with MOF

2.3.2.3.1 Influence of VOCs on the Luminescence changes of 1

In order to demonstrate that 1 has the potential to act as a luminescent VOC sensor, water treated 1 was exposed to ethyl acetate vapours until the sample was saturated and the luminescence was recorded (see section 2.4.4.1). During VOC exposure, samples were measured every ten minutes. A sample was determined to be saturated after three consecutive measurements showed no luminescence change. In Figure 2.6, ethyl acetate produces a “turn-on” behaviour versus water.

1.75

1.5

1.25

1 Ethyl Acetate 0.75 Water Vapour

0.5

0.25

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

After this initial confirmation of sensing potential, representatives of each common solvent type were measured and compared as the percent difference of VOC saturated 1 versus water (Figure 2.7). Benzene and pentane (in addition to ethyl acetate) all show notable “turn-on” behaviour with a greater than 60 % increase in luminescence intensity versus the water baseline indicating that 1 may be a potential sensor for esters, alkanes and aromatics. Acetone shows a slight increase (20%), while acetonitrile, tetrahydrofuran (THF) and methanol all have negligible impact on luminescence intensity. Diethyl ether and chloroform both cause notable “turn-off” behaviour with a luminescence intensity decrease of more than 25%. Lastly, some solvochromism is observed but the overall effect is minimal, i.e., all luminescence maxima are within the range of 555 to 585 nm.

% = 𝐼𝐼𝑉𝑉𝑉𝑉𝑉𝑉 − 𝐼𝐼𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝐼𝐼𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 Equation 2.1: Percent difference versus water of MOF luminescence intensity.

100.00%

75.00% Ethyl Acetate Pentane Benzene 50.00% Acetone Dry THF Wet THF 25.00% Acetonitrile Methanol

PercentDiffernece vsWater Diethyl Ether 0.00% Chloroform

-25.00%

-50.00%

2.3.2.3.2 Luminescence Lifetime Measurements

Ethyl acetate vapour causes the most significant luminescence intensity enhancement resulting in a quantum yield of 24.3 % (Figure 2.27). Such a high quantum yield allows us to measure the luminescence lifetime accurately. Multiple lifetimes were observed (Table 2.3 and Table 2.4) I with the longest being 2.6 μs which is comparable to other MOFs containing Cu 3Pz3 units and is indicative of metal-based phosphorescence.92–94

2.3.2.3.3 MOF Structure-Activity Relationship: Cycling of Water and Ethyl Acetate Vapour

In order to verify that the observed VOC sensing was a result of migration of solvent molecules into the MOF and not just an effect due to outside surface adsorption, a structure-activity

relationship was investigated. If solvent was being adsorbed by the MOF to cause the observed luminescence changes there should be observable structural changes via powder x-ray diffraction (PXRD). These structural changes would manifest as loss or gain of peaks in the PXRD pattern and/or significant shifts of the peaks from the as-synthesized sample. We postulated that these changes should be reversible upon using water vapour to restore the structure. In this manner, solvent switching should produce a notable, and reversible, change in the PXRD pattern to mirror the observed luminescence changes.

As VOC with the largest observable luminescence enhancement, ethyl acetate was examined first. A sample of 1 was prepared on a silicon zero background PXRD sample holder (i.e. the sample holder produces no PXRD signals) via dropcasting from an ethyl acetate solution. The sample holder was placed in a degassed, sealed 250 mL flask fitted with a septum and protected from light. Solvent vapour, using nitrogen as a carrier gas, was diffused into the flask for the listed time. Times were optimized over several experiments. If the previous solvent was water, additional diffusion time was necessary (e.g. EtOAc after 42 hours still showed trace water present [Figure 2.8]).

6000

5000

4000

96 h EtOAc vapour 3000 23 h water vapour 42 h EtOAc vapour

Relative Intensity Relative Water vapour 2000 Fresh (1)

1000

0 5 10 15 20 25 30 2θ (°)

As a result of this experiment a clear trend emerged. Treatment of 1 with water vapour resulted in the loss of major peaks at 6.11° and 6.29° which are present in the freshly prepared sample (Figure 2.8, red trace). A new peak emerged at 6.90°. Treatment with ethyl acetate reversed these changes, restoring the original pattern (Figure 2.8, purple trace). These changes could be reproduced within the same sample via sequential exposure to water and ethyl acetate vapour. This proves that a) water causes an observable structural change to 1 and b) that restoration of the original structure, via ethyl acetate treatment, is possible. However, the exact location and coordination mode of water remains unknown as this information is not discernable from the PXRD pattern.

2.3.2.3.4 Role of Water and Proposed Sensing Mechanism

In order to further probe the role of water in the observed luminescence changes, PXRD studies of all of the tested VOCs were conducted to demonstrate the structure–activity (Figure 2.47 to Figure 2.52). All other solvents (except diethyl ether) cause a similar structural change as ethyl acetate. Based on the studies of similar compounds in the literature,92 the luminescence of 1 likely originates from the metal-to-metal charge transfer 3[MMCT], metal-to-ligand charge transfer 3[MLCT], and halide-to-ligand charge transfer 3[XLCT] excited states. The 3[MMCT] excited state is likely the main contributor, evidenced by the broad and featureless emission spectrum of 1. All three types of excited states are well-known contributors for related luminescent Cu(I) complexes.92,95–97 Coordinating solvents (such as water, acetonitrile, alcohols, and THF) are known to quench the luminescence of Cu(I) compounds via exciplex formation.98,99 An exciplex is “an electronically excited complex of definite stoichiometry, ‘non- 100 bonding’ in the ground state.” For 1, the excited complex is the [Cu3Pz3] unit. Further luminescent studies contrasting wet (from reagent bottle) versus dry (distilled from sodium/benzophenone) THF revealed no statistically significant changes originating from the dryness of the VOC (Figure 2.9).

10.00%

8.00%

6.00%

4.00% Wet THF Vapour 2.00% Dry THF Vapour 0.00%

PercentDiffernece vsWater -2.00%

-4.00%

-6.00%

Figure 2.9: Sequential exposure of 1 to dry and wet THF following standard procedures. Sample illuminated at 469 nm and monitored at the peak emission of THF (559 nm).

Presumably, the more substantial turn-on behaviour caused by water-immiscible solvents (i.e. ethyl acetate, pentane, benzene) is due to their non-coordinating nature and ability to displace the quencher water from the pores. Water-miscible solvents (i.e. acetonitrile, THF, acetone, and methanol) may also displace water molecules in the pore but can also quench luminescence via exciplex formation due to their coordinating nature. Therefore, most water-miscible VOCs cause no significant luminescence intensity change overall. Only the less coordinating acetone exhibits a 25% luminescent turn-on behaviour compared to water-treated 1.

Since chloroform is able to displace water (Figure 2.10) and halocarbons are well known to quench the transition metal-based luminescence through electron transfer,101 the interactions between chloroform and the luminescent sites within the pores likely account for the observed luminescence intensity decrease. Diethyl ether causes no structural change to the water-treated 1 (Figure 2.11) implying that the luminescence quenching caused by ether is probably a surface effect.

4000

3500

3000

2500

2000 19 h chloroform vapour 23 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.10: PXRD comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with chloroform vapour (red).

4000

3500

3000

2500

2000 18 h diethyl ether vapour 21 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.11: PXRD comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with diethyl ether vapour (pink).

2.3.3 β-CuBr MOF

2.3.3.1 Synthesis and Crystal Structure Description

Yellow crystals of [Cu8L6Br][CuBr2] (β-CuBr MOF, 2) can be obtained by heating CuBr and

HL in dry, degassed DMF under an N2 atmosphere at 135 °C and increasing the temperature to 150 °C over five days in the presence of 0.7 equivalents of 2,2’-bipyridine. Similar to 1, this I I − MOF also features the Cu 3Pz3 chromophore and cationic skeleton with [Cu Br2] counterions in the pores (Scheme 2.3).

Cu Cu N Br N

N N

N N Cu N N Cu N N N N N N N CuBr Cu Cu Cu Cu N N Br Cu Br N N DMF 135°C

N NH

N N

Scheme 2.3: Synthesis of the β-CuBr MOF (2).

Unlike 1, 2 crystallizes in the Monoclinic C2 space group. Two building blocks can be found in I + the skeleton of 1: (a) trigonal [Cu3Pz3] units (Figure 2.12a) and (b) [Cu 2Py6Br] clusters with

bromide residing on the C2 axis bridging two tetrahedral CuPy3 groups (Figure 2.12b).

Unlike 1, the bridging bromides in 2 are covalently bonded to copper with a bond length of 2.509 Å. There are no observable electrostatic interactions between the bridging bromide and the I I + [Cu 3L3] cluster. Instead, the pyridine groups of [Cu 2Py6Br] cluster effectively shield the

I bromide. The apical sites of the [Cu 3L3] unit instead either remain open or are involved with I adjacent [Cu 3L3] units via π-stacking. As a consequence of these differing building units the MOF is less symmetrical than 1. The repeating unit of the cationic skeleton of 2 is the I + I [Cu 2Py6Br] unit and can then be described as a weave. The ligands attached to Cu weave I + through one another eventually linking distant [Cu 2Py6Br] units together (Figure 2.13).

I + Figure 2.13: “Weave” of [Cu 2Py6Br] units showing the lower (yellow) ligand arm reaching up between the ligands of the purple unit.

This weave allows 2 to form a 3D structure. Parallelepiped channels form along the b direction of the crystal lattice, where the disordered counterions and solvent molecules reside (Figure 2.14a). However, the weave occludes all other directions (Figure 2.14b). As a consequence, there are void spaces to allow for potential VOC sensing within the MOF. Furthermore, as there is

nothing directly coordinating to the [Cu3Pz3] units, 2 should be more sensitive to analyte gases.

Figure 2.14: a) Parallelepiped channels down the b direction of 2; b) all other directions are closed packed (a direction shown).

2.3.3.2 Luminescent Properties of β-CuBr MOF 2

Freshly synthesized 2 was vacuum dried and tested for luminescence activity. A peak emission of 620 nm was observed. This emission peak remained the same whether the sample was excited at 410 or 469 nm (Figure 2.15). The higher excitation indicates that 2 requires more energy to excite than 1.

80000

70000

60000

50000 Emission of 2 with 410 nm Excitation 40000 Emission of 2 with 469 nm 30000 Excitation Excitation of 2 with 620 nm (a.u.) Intensity Relative 20000 Emission 10000

0 350 450 550 650 750 850 Wavelength (nm)

Figure 2.15: Luminescent excitation and emission spectra of a bulk sample of vacuum dried 2 monitored at a wavelength of 620 nm and 410/469 nm respectively.

Attempts to examine the efficacy of 2 towards VOC sensing were unsuccessful. The open nature

of the [Cu3Pz3] units proved to be far more sensitive towards oxidation. 2 is stable as a dry material allowing for luminescence and PXRD measurements. However, attempts to expose 2 to VOCs using our testing apparatus (Figure 2.37) decomposed the MOF. The decomposition is likely a result of adventitious oxygen present in the testing apparatus. It is important to note that + 1 was stable under these testing conditions likely due to the interactions between the [Cu3Br2]

and [Cu3Pz3] units.

Although initial experiments were performed to evaluate the luminescence changes in 2 for N2

and CO2 versus vacuum, results discouraged further study.

2.3.4 α-CuI MOF

2.3.4.1 Synthesis and Crystal Structure and Description of [Cu3I3]

Yellow crystals of [Cu9L6I2][CuI2] (3) can be obtained in poor yield by heating CuI and HL in

dry, degassed DMF under an N2 atmosphere at 135 °C for four days (Scheme 2.4). This MOF I I − features the Cu 3Pz3 chromophore and cationic skeleton with [Cu I2] counterions in the pores

and crystallizes in the Monoclinic C2/c space group.

I Cu I

N N Cu Cu N N N N Cu N

I N N CuI Cu Cu Cu DMF N N 135°C I

N NH N Cu N N N Cu Cu N N

Scheme 2.4: Synthesis of the α-CuI MOF (3).

+ 3 is isostructural to 1 in that it is composed of the same [Cu3Pz3] and [Cu3I2] units with iodide replacing bromide. The same “sandwich” building unit is formed (Figure 2.16) with trigonal prismatic channels along the a+c direction of the crystal lattice where the disordered counterions and solvent molecules reside (Figure 2.17). Due to the increased ionic radius of iodide, the unit cell parameters of 3 are similar, but not identical, to 1.

Figure 2.17: Trigonal prismatic channels in 3 shown along the a+c direction.

However, 3 is much harder to synthesize than 1 due to possible I2 contamination of our CuI starting material. This is a known problem with commercially available CuI. Purification of the

starting material does improve the yield of the resulting MOF but contaminates are still present. As a result, large scale synthesis required to make enough material to conduct VOC sensing measurements, was not possible due to required mechanical separation of 3 from unreacted ligand and oxidized copper.

2.3.5 Independent Synthesis of Small Molecule Cu3L3 Unit 2.3.5.1 Synthesis and Crystal Structure Description

Clear crystals of Cu3L3 (4) can be obtained in poor yield by heating Cu2O and HL in absolute

ethanol under an N2 atmosphere at 110 °C for 60 hours (Scheme 2.4). This material features the I Cu 3Pz3 unit and crystallizes in the Triclinic P-1 space group. It is important to note that unlike 1, 2, and 3, 4 is a small molecule: the pyridine groups remain available for coordination (Figure 2.18). 4 is stable as a dry solid but oxidizes in solution when exposed to air.

Figure 2.18: Crystal structure of 4 showing small molecule [Cu3L3] unit. Co-crystallized solvent

(CHCl3) not shown.

Synthesis of [Ag3L3] was attempted. However, the yields for this material was considerably lower than 4. Due to the difficulty with making sufficient quantities for investigation, its

tendency to decompose on exposure to light, and challenges with purification, [Ag3L3] was not

pursued beyond preliminary investigations. [Au3L3] was considered as a target molecule but not investigated due to starting material costs.

2.3.5.2 MOF Synthesis Attempts

With compound 4 in hand, synthesis of MOFs was attempted. The primary aim was to replicate the synthesis of 1, 2, and 3 to elucidate the synthetic route. A secondary goal was to produce 10 bimetalic MOFs. Reactions of [Cu3Pz3] with other simple d metal salts (e.g. zinc or cadmium) would not quench luminescence and may allow for the discovery of a new, diverse family of

MOFs. Unfortunately, no materials were produced which maintained the [Cu3Pz3] unit (Table

2.1). Of the conditions tested, only ZnCl2 yielded a crystalline product determined to be

[ZnLCl]n indicating decomposition of 4.

Table 2.1: Examples of MOF synthesis test conditions utilizing 4 as a starting material.

Metal Mass of Solvent Conditions Result Salt 4 (mg)

CHCl :DCM:DMSO No crystals or fluorescence, green 2 eq CuCl 15 3 70°C, 4 days (2:1:3) solution

CHCl :DCM:DMSO No crystals or fluorescence, green 2 eq CuBr 15 3 70°C, 4 days (2:1:3) solution

CHCl :DCM:DMSO No crystals or fluorescence, green 2 eq CuI 15 3 70°C, 4 days (2:1:3) solution

2 eq CHCl :DCM:DMF Green solution, with luminescent 15 3 70°C, 7 days ZnCl2 (2:1:3) crystals identified as [ZnLCl]n

2 eq CHCl :DCM:DMF No crystals, fluorescence Zn(NO ) ∙ 15 3 70°C, 7 days 3 2 (2:1:3) observed, green solution 6H2O

The synthetic difficulties were a consequence of air sensitivity for copper and the iterative nature of MOF synthesis. Also, as mentioned when discussing 1 there are multiple reactions that occur

during the synthesis. These intermediate species may serve to either template the eventual MOF + + formation or act as key intermediates for the [M3X2] or [M2X] building unit formation.

2.3.6 Conclusions

In summary, we have prepared a new luminescent Cu(I)–pyrazolate MOF, 1, featuring the

[Cu3Pz3] chromophore with secondary Cu(I) clusters as building units. The exposure of water- treated 1 to the vapours of water-immiscible VOCs causes luminescence signal modulations, indicating that 1 may prove effective for the detection of VOCs. Water miscible VOCs did not have a significant impact on the observed luminescence of 1. Preliminary studies show that most VOCs cause structural changes to the water-treated 1 except for diethyl ether. These structural changes indicate that there is a structure-activity relationship proving that there are interactions between the VOC and the MOF, not just the VOC and chromophore.

Two additional MOFs containing the [Cu3Pz3] unit were also synthesized. However, both proved unsuitable for VOC detection due to low stability (2) and low synthetic yield/purity (3). The small molecule analogue of 1, 4, was successfully isolated. However, attempts to make a bimetalic MOF proved unfruitful.

2.4 Experimental 2.4.1 Synthesis

2.4.1.1 General Considerations

Elemental analyses were performed at the ALANEST Facility in our Chemistry Department on a on a Thermo Flash 2000 CHN analyzer. Thermogravimetric analyses (TGA) were performed on a TA Instruments SDT Q600 instrument under a dinitrogen atmosphere with a heating rate of 5 °C per minute. NMR spectra were recorded on a Bruker Avance 400 spectrometer. Both 1H and 13C NMR spectra were referenced and reported relative to the solvent's residual signals. Photoluminescence spectra were measured using either a QunataMaster spectrofluorimeter (Photon Technology International, Edison, New Jersey) (1, all VOC sensing) or a SPEX Fluorolog-3 spectrofluorometer (Jobin Yvon/SPEX, Edison, New Jersey) (2) with a Xenon lamp for steady-state measurements. Steady-state photoluminescence spectra for quantum yield experiments were performed with a Xenon lamp excitation source and detected in on a Fluorolog-3 (Horiba) equipped with a Quanta-Phi quantum yield measurement accessory

(Horiba). Photoluminescence lifetimes were measured using a Horiba Time-Correlated Single Photon Counting setup with a 375 nm picosecond laser. (DeltaDiode Horiba). Sample emission was detected at 575 nm after passing through a spectrometer (iHR320, Horiba) and detected on a single photon detector (PPD, Horiba). Powder X-Ray Diffraction (PXRD) experiments were performed on a Rigaku MiniFlex 600 diffractometer equipped with a Cu-Kα source operating at 40 kV/15 mA at the Walter Curlook Materials Characterization & Processing Laboratory at the University of Toronto Department of Materials Science and Engineering. A step scan mode was used for data acquisition with a step size of 0.02° 2θ. All the PXRD samples were prepared by dropcasting onto a silicon zero background sample holder. The infrared spectrum of 1 was recorded using a neat sample on a Bruker Alpha FT-IR spectrometer equipped with a Platinum ATR sampling unit in air. Unless otherwise stated, all manipulations were performed under dinitrogen. All reagents were purchased from commercial sources and used without further purification. DMF for the self-assembly of MOFs was dried and degassed on a PureSolv-MD (Innovative Technology, Inc., Newburyport, Maryland) Solvent Purification System containing activated alumina. 4-(4-bromophenyl)-2,4-pentadione was prepared following literature 91 procedures. Pd(PPh3)4 was synthesized following modified standard procedures from 102 Pd(OAc)2.

2.4.1.2 Synthesis and Characterizations of Ligands

2.4.1.2.1 Synthesis of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole (Done in air):

To a suspension of 3-(4-bromophenyl)pentane-2,4-dione (25.0 g, 980 mmol) in methanol (350 mL) was added 12 N HCl aq. (1 mL). The resulting mixture was then cooled in an ice bath and hydrazine monohydrate was added dropwise (12 mL, 245 mmol). The reaction mixture was further stirred for 1 h at 0°C before being left to react overnight at room temperature. The solution was concentrated under reduced pressure until the volume was halved and the remaining solution was cooled in a freezer at –20°C. White crystals of 1 were collected via vacuum filtration, washed with cold methanol, and dried under vacuum (21 g, 85 % yield). A second and 1 third crop of crystals may be collected from the supernatant. H NMR (400 MHz, CDCl3) δ: 11.61 (s, N-H), 7.53 (d, J = 8.4 Hz, 2H, Ph), 7.14 (d, J = 8.4 Hz, 2H, Ph), 2.32 (s, 6H, Me). 13C

NMR (100 MHz, CDCl3) δ: 141.90, 132.82, 131.72, 131.00, 120.35, 117.52, 11.64. Anal. Calc.

for C11H10N2Br: C 52.61, H 4.42, N 11.16; found C 52.68, H 4.74, N 11.38.

170000

160000

150000

140000

130000 2.32 7.13 7.16 7.26 Chloroform-d 7.52 7.54 11.61

120000

110000

100000

90000

80000

70000 Intensity

60000

50000

40000

30000

20000

10000

-10000 6.00 2.00 1.93 1.06

2.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 f1 (ppm)

1 Figure 2.19: H NMR spectrum in CDCl3 of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole.

18000

17000

16000

15000

14000 11.64 77.16 Chloroform-d 117.52 120.35 131.00 131.72 132.82 141.90 13000

12000

11000

10000

9000

8000 Intensity 7000

6000

5000

4000

3000

2000

1000

-1000

140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 2.20: C NMR spectrum in CDCl3 of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole.

2.4.1.2.2 Synthesis of 4-(4-(3,5-dimethyl-1H-pyrazol-4-yl)phenyl)pyridine (HL)

To a degassed mixture of 4-(4-bromophenyl)-3,5-dimethyl-1H-pyrazole (7.50 g, 29.9 mmol),

K3PO4 (6.34 g, 29.9 mmol), 4-pyridine boronic acid (4.22 g, 34.3 mmol), and Pd(PPh3)4 (172.6 mg, 0.149 mmol) was added 360 mL of degassed dioxane/water (1:1) mixed solvent. The resulting reaction mixture was then refluxed and the reaction progress was monitored via TLC. Upon completion (2–3 days) the solution was made alkaline with 4 g of KOH in 150 mL of water and extracted with ethyl acetate (4 x 100 mL). The organic extracts were combined,

washed with brine, dried over MgSO4, filtered, and concentrated to dryness under reduced pressure to yield 7 g of crude material which was purified via column chromatography (eluant begins as EtOAc and switches to 5% MeOH in EtOAc) to produce 5.0 g (67 % yield) of the pure 1 product. Note that HL reacts with chloroform. H NMR (400 MHz, DMSO-d6) δ 12.38 (s, 1H, NH), 8.63 (d, J = 6.1 Hz, 2H, Py), 7.85 (d, J = 8.4 Hz, 2H, Ph), 7.74 (d, J = 6.1 Hz, 2H, Py), 7.45 13 (d, J = 8.4 Hz, 2H, Ph), 2.24 (s, 6H, Pz-Me). C NMR (100 MHz, DMSO-d6) δ 150.21, 146.69, 145.13. 136.08 135.24, 134.20, 129.27, 126.84, 120.94, 116,14, 12.99, 10.16. Anal. Calc. for

C16H15N3: C 77.08, H 6.06, N 16.85; found C 77.13, H 5.89, N 16.64.

14000

13000

12000

11000

10000

9000 2.24 2.50 Dimethyl Sulfoxide-d6 3.33 12.38 8.64 8.63 7.86 7.84 7.75 7.74 7.46 7.44

8000

7000

6000 Intensity

5000

4000

3000

2000

1000

0 6.05 2.00 1.99 2.00 0.95 2.00 -1000

12.5 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 f1 (ppm)

1 Figure 2.21: H NMR spectrum of HL in DMSO-d6.

18000

17000

16000

15000

14000

13000

12000

150.21 146.69 145.13 136.08 135.24 134.20 129.27 126.84 120.94 116.14 39.52 Dimethyl Sulfoxide-d6 12.99 10.16 11000

10000

9000

8000 Intensity

7000

6000

5000

4000

3000

2000

1000

-1000

150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 2.22: C NMR spectrum of HL in DMSO-d6.

2.4.1.3 Synthesis of [Cu3L3] containing MOFs

2.4.1.3.1 Synthesis of α-CuBr MOF (1)

Approximately 100 mL of DMF was transferred to a dry, degassed mixture of HL (182 mg, 0.731 mmol) and CuBr (129 mg, 0.901 mmol) in a 350 mL Schlenk bomb. The orange mixture was subjected to three consecutive freeze-pump-thaw cycles and backfilled with nitrogen gas to remove trace oxygen. The Schlenk bomb was sealed, covered with aluminum foil, and heated with gentle stirring at 135°C for 5 d. The resulting yellow crystals were collected via vacuum filtration and washed with acetone and hexanes, dried under vacuum, and stored under N2. 116 mg (53 % yield) of yellow micro-crystals collected. X-ray quality crystals may be collected by not stirring the reaction mixture at the cost of reduced yield. Anal. Calc. for vacuum dried

C48H42N9Br2Cu5: C 47.16, H 3.46, N 10.31; found C 46.54, H 2.93, N 10.20.

2.4.1.3.2 Synthesis of β-CuBr MOF (2)

Approximately 100 mL of DMF was transferred to a dry, degassed mixture of HL (248.3 mg, 0.9959 mmol), 2,2’bipyridine (111 mg, 0.708 mmol) and CuBr (179 mg, 1.25 mmol) in a 350 mL Schlenk bomb. The resulting red mixture was subjected to three consecutive freeze-pump- thaw cycles and backfilled with nitrogen gas to remove any trace oxygen. The system was sealed, protected from light and left to react at 135 °C. After crystal nucleation was observed, the temperature was increased to 150 °C. Total reaction time was five days. The resulting yellow crystals were collected via vacuum filtration and washed with acetone and ether. 65 mg (20 %

yield) of pure yellow crystals collected. Anal. Calc. for vacuum dried C48H42N9Br1.5Cu4.5: C 50.10, H 3.86, N 10.95; found C 49.86, H 3.41, N 11.13.

2.4.1.3.3 Synthesis of α-CuI MOF (3)

10 mL of DMF was equally proportioned atop dry mixtures of HL (15.0 mg, 0.0602 mmol) and CuI (13.8 mg, 0.0722 mmol) in 20 mL glass sleeves inside two stainless steel high pressure reactors. The resulting mixture was purged 20 times with 200 PSI nitrogen gas to remove trace oxygen prior to 30 min of sonication and a further 10 purges. The system was then sealed under 1 atm of nitrogen, protected from light, and left to react at 105 °C for 1 week. After heating, the reactors were cooled and the solids were collected on a Hirsch funnel and washed with acetone. The resulting yellow crystals were mechanically separated from unreacted ligand and dried under

45 vacuum for 48 hours. 5.7 mg (15 % yield) of pure yellow crystals collected. Anal. Calc. for vacuum dried C96H84I4Cu10N18: C 43.79, H 3.22, N 9.58; found C 43.72, H 3.34, N 9.44.

2.4.1.4 Synthesis of Small Molecule Cu3L3 (4)

75 mL of absolute ethanol was charged into a flask containing HL (500 mg, 2.01 mmol) and

Cu2O (861 mg, 6.02 mmol). The system was deoxygenated via four freeze-pump-thaw cycles and left to reflux on a Schlenk line at 110 °C for 60 hours. After cooling, crude solids were collected and the filtrate discarded. The solids were suspended in a 2:1 mixture of DCM:CHCl3, sonicated for 10 minutes and filtered. This was repeated twice. The collected filtrate concentrated to dryness under reduced pressure. 496 mg (78 % yield) of white solids collected. A 1 sample was recrystallized from CHCl3 for single crystal, NMR and EA analysis. H NMR (400

MHz, CDCl3) δ 8.68 (d, J = 6.0 Hz, 2H, Py), 7.715 (d, J = 8.3 Hz, 2H, Ph), 7.56 (d, J = 6.0 Hz, 13 2H, Py), 7.56 (d, J = 8.3 Hz, 2H, Ph), 2.44 (s, 6H, Pz-Me). C NMR (100 MHz, CDCl3) δ

150.50, 129.97, 127.16, 121.52, 13.41.Anal. Calc. for air dried C48H42Cu3N9•0.8(CHCl3): C 56.85, H 4.18, N 12.23; found C 56.74, H 4.14, N 12.51.

7500 8.69 8.67 8.63 8.62 7.72 7.70 7.57 7.55 7.43 7.41 7.26 2.44 1.56 7000

6500

6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

-500 1.01 0.29 0.25 0.97 0.95 1.07 2.74

9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 f1 (ppm)

1 Figure 2.23: H NMR spectrum of 4 in CDCl3. Note the coordination of CDCl3/CHCl3 to 4

(representing 0.8 CHCl3 per molecule) as seen by the shoulders at 8.62 and 7.71 ppm. N-H peak not visible due to water present in solvent.

7000

6500 150.50 129.97 127.16 121.52 77.16 13.41

6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

-500

150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 2.24: C NMR spectrum of 4 in CDCl3.

2.4.2 Experimental Properties and Supporting Information

2.4.2.1 Quantum Yield and Luminescent Lifetime Measurements for 1

Quantum yield measurements (Figure 2.25 to Figure 2.27 and Table 2.2) and luminescent lifetimes (Figure 2.28 to Figure 2.30, Table 2.3, and Table 2.4) of VOC treated samples of 1 were prepared as follows. First, freshly prepared 1 was treated with the appropriate vapour (water or ethyl acetate) in a sealed vial using the same apparatus as that used for VOC luminescent sensing (see section 2.4.4.1 for full description). Next, the background (“blank”) properties of a clean empty quartz cuvette was measured. Once the value was known, vapour treated 1 was quickly transferred from the vial to the quartz cuvette, sealed, and the measurement collected. Use of an air free quartz cuvette was not possible due to sample size restrictions within the instrument. Luminescent lifetimes were determined using Equation 2.2.

2000000 7000 1800000 6000 1600000

1400000 5000 1200000 4000 Blank Corrected 1000000 Sample Corrected 3000 800000 Blank EM Corrected Emission(counts) (counts) Excitation 600000 2000 Sample FL corrected 400000 1000 200000 0 0 350 400 450 500 550 600 650 700 Wavelength (nm)

Figure 2.25: Quantum yield of as-synthesized 1 with excitation at 375 nm.

2000000 1600

1800000 1400 1600000 1200 1400000 1200000 1000 Blank Corrected 1000000 800 Sample Corrected 800000 600 Blank EM Corrected Emission(counts) (counts) Excitation 600000 400 Sample FL corrected 400000 200000 200 0 0 350 400 450 500 550 600 650 700 Wavelength (nm)

Figure 2.26: Quantum yield of water vapour treated 1 with excitation at 375 nm.

2000000 10000 1800000 9000 1600000 8000

7000 1400000 6000 1200000 5000 Blank Corrected 1000000 4000 Sample Corrected 800000 3000 Blank EM Corrected Emission(counts) (counts) Excitation 600000 2000 Sample FL corrected 400000 1000 200000 0 0 -1000 350 400 450 500 550 600 650 700 wavelength (nm)

Figure 2.27: Quantum yield of ethyl acetate vapour treated 1 with excitation at 375 nm.

Table 2.2: Results of Quantum Yield Measurements for 1.

Quantum Yield (%)

As-synthesized (N2) 8.6

Water Vapour 1.8

Ethyl Acetate Vapour 24.3

1.2

0.8

0.6 Observed Biexponential Fit (a.u.) Intensitry 0.4

0.2

0 0 5000 10000 15000 20000 25000 time (ns)

Figure 2.28: Observed luminescence decay for ethyl acetate treated 1 over 26 μs with a biexponential fit and an R2 value of 0.9708.

1.2

0.8

0.6 Observed Biexponential Fit (a.u.) Intensity 0.4

0.2

0 0 5000 10000 15000 20000 25000 Time (ns)

Figure 2.29: Observed luminescence decay for as-synthesized 1 over 26 μs with a biexponential fit and an R2 value of 0.6953.

1.2

0.8

0.6 Observed Biexponential Fit (a.u.) Intensity 0.4

0.2

0 0 5000 10000 15000 20000 25000 Time (ns)

Figure 2.30: Observed luminescence decay for water treated 1 over 26 μs with a biexponential fit and an R2 value of 0.2695.

Table 2.3: Fitting parameters for the luminescent lifetimes of 1 with different solvent vapour and observation times.

Gas N2 EtOAc Water

Obs. 100 ns 26 μs 26 μs 100 ns 26 μs 26 μs 100 ns 26 μs 26 μs Period

A1 1.713 0.4927 0.4685 1.777 0.7709 0.4329 - 0.1985 0.9954

A2 0.1424 - 0.2155 0.2128 - 0.3967 - - 0.1499

6.283 x 5.776 x I 0.08023 0.1845 0.1705 0.1073 - 0.1214 0.1206 0 10-2 10-2

6.934 x 2.474 x 5.948 x 1.114 x 1.79 x k 8.125 10.8 - 0.421 1 10-4 10-3 10-4 10-3 10-3

2.557 x 1.544 x 3.872 x 1.201 x k 0.01835 - - - - 2 10-4 10-2 10-4 10-3

0.0996 0.0619 3.80 x σ (ns) 0.0639 0.7786 0.5191 1.614 - 0.3709 2 7 10-6 t0 (ns) 17.96 20.9 20.42 17.93 24.36 23.18 - 19.05 24.76

R2 0.7798 0.7374 0.6953 0.7629 0.9744 0.9708 - 0.3524 0.2695

σ σ [{ } ] [{ } ] 0 2 0 ( ) = + 2 1 + erf + 2 1 + erf 2 σ �𝑘𝑘2 ∗ − 𝑥𝑥−𝑡𝑡0 ∗𝑘𝑘2 � σ �𝑘𝑘1 ∗ − 𝑥𝑥−𝑡𝑡0 ∗𝑘𝑘1 � 2 2 σ 𝑥𝑥2− 𝑡𝑡 σ 𝑥𝑥2− 𝑡𝑡 𝑓𝑓 𝑦𝑦 𝐼𝐼0 �𝐴𝐴1� 𝑒𝑒 ∗ � � 2 �� 𝐴𝐴2� ∗ � � 2 �� 2 ∗ 𝑘𝑘1 2 ∗ 𝑘𝑘2 ∗ √ − ∗ √ − √ √ Equation 2.2: Fit for luminescent lifetimes of 1 where I0 is the fitting offset, k1 and k2 are the decay rates, A1 and A2 are the exponential coefficients, σ is the width of the excitation pulse, and t0 is the time-zero offset. Erf is the error function.

Table 2.4: Luminescent lifetimes for 1 monitored for various periods under N2 (as-synthesized), ethyl acetate vapour and water vapour. All lifetimes are in nanoseconds.

Gas N2 EtOAc water

Obs. 100 ns 26 μs 26 μs 100 ns 26 μs 26 μs 26 μs 26 μs Period

τ1 (ns) 0.12 1442.17 404.20 0.09 1681.24 897.67 558.66 2.38

τ2 (ns) 54.50 - 3910.83 64.77 - 2582.64 - 832.64

2.4.2.2 Powder X-Ray Characterization

PXRD was used to determine bulk sample purity for all analyzed MOFs. The observed PXRD pattern of dropcast samples of finely powdered MOF on a silicon zero background holder were compared to the PXRD pattern predicted from the single crystal structure for each MOF. In the case of 1, the observed pattern is a very close match to the predicted pattern with few discrepancies (Figure 2.31). These minor discrepancies (i.e. peak shape, size, and location) are a consequence of predicting the pattern from an idealized single crystal. Under experimental conditions numerous variables can alter the observed pattern. These include but are not limited to: thermal expansion (single crystals are collected at low temperature), unaccounted for solvent (e.g. disordered solvent removed from the single crystal solution), and MOF structural defects (e.g. a missing ligand in the structure). The absence of any unexplained peaks (e.g. not present in the predicted pattern) and close correlation between the predicted and observed pattern indicate a pure sample of 1. This analysis is further corroborated by thermal gravimetric analysis and elemental analysis of the bulk sample.

3500

3000

2500

2000

Theoretical

Intensity 1500 Observed 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.31: Theoretical versus observed PXRD pattern for 1.

3500

3000

2500

2000

Observed 1500 Predicted Relative Intensity Relative 1000

500

0 4.5 9.5 14.5 19.5 24.5 29.5 2θ (°)

Figure 2.32: Theoretical versus observed PXRD pattern for 2.

2.4.2.3 Thermal Gravimetric Analysis

Thermal gravimetric analysis (TGA) was used to determine the solvent loading and thermal stability of compounds 1-3 (Figure 2.33 to Figure 2.35). As samples are heated, solvent is liberated and is observed as a mass change. This mass change can then be used to determine the quantity of solvent present within the MOF to assess the accuracy of subsequent elemental analysis results.

100.00%

80.00%

60.00%

PercentMass 40.00%

20.00%

0.00% 0 200 400 600 800 1000 Temperature (°C)

100%

80%

60%

40% MassPercent

20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 2.34: TGA spectrum of β-CuBr MOF (2).

100%

80%

60%

40% MassPercent

20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 2.35: TGA spectrum of α-CuI MOF (3).

2.4.2.4 Infrared Characterization

Infrared (IR) spectra were collected to verify the presence of organic components (L in the case of 1) in bulk MOF samples.

Wavenumber (cm-1) 3500 3000 2500 2000 1500 1000 500 100%

90%

80%

70%

60%

50%

Percent Transmission (%) 40%

30%

Figure 2.36: Infrared spectrum of neat sample of 1.

2.4.3 X-Ray Diffractometry

X-ray quality single crystals of 1 to 4 were obtained as described in the synthesis above. The crystals were mounted on the tip of a MiTeGen MicroMount. The single-crystal X-ray

diffraction data were collected on a Bruker Kappa Apex II CCD diffractometer with Mo Kα radiation (λ = 0.71073 Å) operating at 50 kV and 30 mA, at 150 K controlled by an Oxford Cryostream 700 series low temperature system. The data integration and absorption correction were performed with the Bruker Apex 2 software package.103 The structure was solved by direct methods and refined using SHELXTL V2016/4.104,105 Disordered solvent molecules and residual electron density were removed using the Platon Squeeze method.106 All non-hydrogen atoms except for the atoms involved in the disordered portions were refined anisotropically. The positions of the hydrogen atoms were calculated using the riding model.

Table 2.5: Crystallographic properties of MOFs 1 to 3.

Name α-CuBr MOF (1) β-CuBr MOF (2) α-CuI MOF (3)

2(C96H84BrCu8N18), Formula C96H84Br4Cu10N18 C96H84I4Cu10N18 2(CuBr2)

Formula Weight 2444.85 4602.91 2632.81 [g/mol]

Crystal System Monoclinic Monoclinic Monoclinic

Space Group C2/c (No. 15) C2 (No. 5) C2/c (No. 15)

a [Å] 19.6076(13) 42.358(6) 19.5049(15)

b [Å] 27.6940(16) 7.0699(9) 27.695(2)

c [Å] 21.551(2) 17.436(3) 22.206(2)

β [°] 107.054(2) 95.362(7) 106.533(4)

Volume [Å3] 11187.9(14) 5198.7(13) 11499.5(16)

Z 4 1 4

D(calc) [g/cm3] 1.452 1.470 1.521

μ(MoKα) [ /mm ] 3.336 3.006 2.930

F(000) 4864 2304 5152

Crystal Size 0.02 x 0.06 x 0.08 0.01 x 0.01 x 0.13 0.04 x 0.06 x 0.10 (estimated) [mm]

Temperature (K) 150 150 150

Radiation [Å] 0.71073 0.71073 0.71073

θ Min-Max [°] 1.8, 27.5 1.0, 27.7 1.3, 27.5

-25: 25 ; -35: 35 ; -27: -25: 22; -35: 35; -28: Dataset -54: 54; -9: 9; -20: 22 28 28

Tot., Uniq. Data, 61644, 12861, 0.070 23393, 11834, 0.084 50097, 13173, 0.115 R(int)

Observed data [I 8030 6632 7175 > 2.0 σ(I)]

Nref, Npar 12861, 606 11834, 459 13173, 588

R1, wR2, GooF 0.0666, 0.1899, 1.04 0.0854, 0.2305, 0.94 0.0820, 0.2749, 1.04

Max. and Av. 0.00, 0.00 0.00, 0.00 0.00, 0.00 Shift/Error

Min. and Max. Resd. Dens. -1.20, 1.94 -1.33, 1.34 -4.49, 2.39 [e/Å3]

( ) = = ( ) = 2 2 2 ∑ 𝐹𝐹𝑜𝑜−𝐹𝐹𝑐𝑐 2 ∑ 𝑤𝑤�𝐹𝐹𝑜𝑜 −𝐹𝐹𝑐𝑐 � 1 2 𝑤𝑤 2 2 𝑅𝑅 ∑ 𝐹𝐹𝑜𝑜 𝑤𝑤𝑅𝑅 𝑅𝑅 𝐹𝐹 � ∑ 𝑤𝑤�𝐹𝐹𝑜𝑜 �

Equation 2.3: Equations for the determination of crystallographic confidence factors R1 and wR2.

Table 2.6: Crystallographic properties of Cu3L3 (4).

Name Cu3L3 (4)

Formula C48H42Cu3N9, CHCl3

Formula Weight [g/mol] 1054.89

Crystal System Triclinic

Space Group P-1 (No. 1)

a [Å] 11.3660(8)

b [Å] 11.7261(9)

c [Å] 19.2941(14)

α [°] 76.251(4)

β [°] 77.251(4)

γ [°] 73.642(3)

Volume [Å3] 2363.9(3)

Z 2

D(calc) [g/cm3] 1.482

μ (MoKα) [ /mm ] 1.552

F(000) 1076

Crystal Size (estimated) [mm] 0.02 x 0.02 x 0.10

Temperature (K) 150

Radiation [Å] 0.71073

θ Min-Max [°] 1.1, 27.6

Dataset -14: 14; -15: 15; -25: 24

Tot., Uniq. Data, R(int) 35686, 10765, 0.065

Observed data [I > 2.0 σ(I)] 6348

Nref, Npar 10765, 583

R1, wR2, GooF 0.0556, 0.1614, 1.01

Max. and Av. Shift/Error 0.00, 0.00

Min. and Max. Resd. Dens. [e/Å3] -0.75, 0.60

2.4.4 Luminescent Sensing

2.4.4.1 Apparatus for the Evaluation of Bulk Powders with VOCs

Figure 2.37: Cuvette and VOC exposure setup for luminescent gas sensing experiments.

Figure 2.38: Configuration of luminescence spectrometer for luminescent gas sensing experiments.

1 was loaded into a 4 mm wide quartz cuvette until the viewing window was covered by free- flowing material (~150 mg) and fitted with a screw-top cap containing a plastic septum. A water vapour baseline was established by bubbling nitrogen through distilled water at a rate of 60 mL/min and introducing it to the cuvette via needles (Figure 2.37). The needles were removed and the luminescence spectrum measured (Figure 2.38). The cuvette was then agitated to display a fresh sample surface and the measurements were repeated a total of seven times. In this manner, the average baseline response of the material is recorded.

1 was then exposed to VOC vapour following the same procedure. The MOF is periodically monitored until there is no intensity change between measurements (sample is saturated with solvent vapour). Saturation time was highly dependent on the vapour pressure of the VOC. Once saturated, seven measurements were recorded to determine the average luminescence of 1. The cuvette was agitated between measurements ensure a fresh sample surface. Measurements are shown in the discussion in addition to Figure 2.39 through Figure 2.46.

After every VOC solvent gas, 1 was regenerated by extended vacuum (≥48 hours). If shorter durations or other methods are used VOC may remain and skew subsequent measurements.

Regeneration is judged effective if the water baseline closely matches that of the previous sample.

2.4.4.2 VOC Effect on Luminescence for Water Treated 1

1.75

1.5

1.25

Pentane 0.75 Water Vapour 0.5

0.25

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.39: Averaged emission intensity of 1 after VOC treatment (pentane – green) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm.

1.75

1.5

1.25

Benzene 0.75 Water Vapour 0.5

0.25

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.40: Averaged emission intensity of 1 after VOC treatment (benzene – orange) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm.

1.5

1.25

0.75 Acetone Water Vapour 0.5

0.25

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.41: Averaged emission intensity of 1 after VOC treatment (acetone – red) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm.

1.2

0.8

Wet THF 0.6 Dry THF Water Vapour (Wet THF) 0.4 Water Vapour (Dry THF) 0.2 Average IntensityRelative to Wter

0 500 520 540 560 580 600 620 Wavelength (nm)

1.2

0.8

0.6 Acetonitrile Water Vapour 0.4

0.2 Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.43: Averaged emission intensity of 1 after VOC treatment (MeCN – red) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 469 nm.

1.2

0.8

0.6 Methanol Water Vapour 0.4

0.2

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.44: Averaged emission intensity of 1 after VOC treatment (MeOH – purple) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm.

1.2

0.8

0.6 Diethyl Ether Water Vapour 0.4

0.2 Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.45: Averaged emission intensity of 1 after VOC treatment (Et2O – green) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm.

1.2

0.8

0.6 Chloroform Water Vapour 0.4

0.2

Average Intensity Relative to Relative (a.u.) Water Intensity Average 0 500 520 540 560 580 600 620 Wavelength (nm)

Figure 2.46: Averaged emission intensity of 1 after VOC treatment (CHCl3 - orange) relative to averaged pre-exposure water vapour emission (blue) using an excitation wavelength of 470 nm.

Table 2.7: Solvochromic effects on 1’s luminescence over sequential experiments.

Max (nm) Standard Deviation

water 574.9 1.3

EtOAc 560.3 1.1

water 575.8 1.3

Benzene 566.7 1.0

water 573.6 2.4

Pentane 562.2 0.8

water 573.4 2.4

MeCN 579.3 1.4

water 577.6 2.4

THF 558.7 1.6

water 571.6 2.0

Acetone 572.9 1.2

water 582.2 2.5

MeOH 579.1 1.9

water 564.1 1.3

Et2O 561.0 1.0

water 569.0 1.4

CHCl3 554.9 1.1

2.4.4.3 PXRD Investigation of Structure-Activity Relationship: Structural Changes of Water vs Solvent Vapour:

A bulk sample of 1 was exposed to 19+ hours of water vapour before being drop cast with acetone on a zero background sample holder to establish a water baseline. Bulk water treated 1 was then divided into vials, placed into degassed, septum sealed 16 mm x 150 mm test tubes and protected from light. Solvent vapour, using nitrogen as a carrier gas, was then diffused into the various test tubes at a rapid rate for 18+ hours. After solvent exposure, the samples of 1 were drop cast on a zero background sample holder using the analyte solvent for comparison with the water baseline. Results of the structure-activity relationship are shown in Figure 2.47 through Figure 2.52.

4000

3500

3000

2500

2000 19 h ethyl acetate vapour 23 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.47: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with ethyl acetate vapour (purple).

4000

3500

3000

2500

2000 18 h pentane vapour 21 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.48: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with pentane vapour (red).

4000

3500

3000

2500

2000 19 h acetone vapour 23 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.49: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with acetone vapour (black).

4000

3500

3000

2500

2000 18h acetonitrile vapour 21 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.50: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with acetonitrile vapour (orange).

4000

3500

3000

2500

2000 18 h tetrahydrofuran vapour 21 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.51: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with tetrahydrofuran vapour (purple).

4000

3500

3000

2500

2000 18 hours methanol vapour 21 h water vapour 1500

Relative Intensity Relative Fresh (1) 1000

500

0 5 10 15 20 25 30 2θ (°)

Figure 2.52: PXRD pattern comparison of fresh 1 (blue), water vapour treated 1 (green), and subsequent treatment with methanol vapour (grey).

Chapter 3 3 A New MOF constructed from Carboxylate Functionalized Bispyrazolylmethane for Coal Mine Methane Capture 3.1 Abstract

Two new MOFs were synthesized using M(OAc2) [M = Ni, Cd] and 1,1’-methylenebis(3,5-

dimethyl-1H-pyrazolyl-4-carboxylic acid) (H2BPM). [Ni(BPM)]n∙xDMSO (5) proved to be a microporous, high surface area MOF, with voids occupying ~50% of the MOF’s volume as observed via single crystal x-ray diffraction and nitrogen gas sorption measurements. Variable temperature adsorption measurements indicated 5 had enthalpy of adsorption characteristics favourable for selective gas sorption. Ideal adsorbed solution theory indicated 5 should behave as a selective adsorbent for coal mine methane capture. Computational studies partially confirm this result. [Cd(BPM)]n∙xDMSO (6) was isostructural to 5 but was not stable towards pore evacuation leading to decomposition of the material.

3.2 Introduction

While a lot of work with MOFs has focused on post combustion CO2 capture (removal of CO2 50,107–109 110–113 from N2), natural gas upgrading (removal of CO2 from CH4 at high pressures), and 114,115 biogas upgrading (removal of CO2 and N2 from CH4 at low pressures) comparatively little has been done to investigate coal mine methane capture. Methane is a potent greenhouse gas 116 with a warming factor 25 times greater than that of CO2. At present, coal mines account for ~9% of the total US methane emissions116 with the US being the second largest contributor to coal mine methane atmospheric release after China.117 This is both an environmental problem and an economic problem as methane itself is a useful fuel for energy production. At the moment, coal mine methane is vented into the atmosphere or flared (burned) off. For select coal mines, the methane is of high enough purity for direct use on site (> 50% CH4) or for sale to the national pipeline system (> 90% CH4). However, coal mine methane is typically in the range of 118–120 30% - 60% with the balance being dry N2 which makes it unsuitable for direct use.

Recognising this problem, industry currently uses pressure swing adsorption (PSA) at some mining sites to recover methane.117–119 Currently, state of the art technology uses activated carbon as the absorbent at ambient temperatures in the pressure range of roughly 0.25 bar to 3 bar. Silica gel, alumina, and zeolites have also been used for PSA coal mine methane capture but zeolites have only been tested in laboratory settings.118,119 However, empirical and theoretical studies have shown MOFs greatly outperform activated carbon with regards to methane uptake and selectivity.121

Despite an extensive literature search only a handful of MOFs have been investigated for coal mine methane capture. Recent reports focus on computational studies122,123 with only a few containing actual gas breakthrough experiments performed on MOF samples.121,124 To the best of

our knowledge, the best MOF tested to date for coal mine methane capture is Cu(INA)2 [INA =

isonicotinic acid] with reported values of 7.0-7.6 dependent on system pressure. Cu(INA)2 has a predicted maximum selectivity of 8.34 for coal mine methane capture (recall: separation of 1:1 124 CH4/N2). For due diligence, two of the best performing general purpose MOFs for CH4 125–128 capture and storage – CPO-27-Co ([Co2(DOTP )]n, DOTP = 2,5-dioxidoterephthalate) and III MIL-47-V ([V (OH)(TP)]n, TP = terephthalate) – were also considered in our evaluations of any synthesized coal mine methane capture MOFs.129–131

Of particular interest to us was bispyrazolylmethane (BPM) as a potential ligand backbone. Its semi-flexible nature and proclivity to reduce π-stacking were envisioned forming ideal pore sizes for gas uptake when incorporated into a MOF. While many BPM MOFs and coordination polymers have been reported, most feature simple BPM moieties with little or no gas sorption data.132–138 Our goal was to use the simplest variant: a 4-carboxylate substituted BPM to access appropriately sized pores for CO2 and CH4 capture by targeting the restricted geometries of six coordinate metal centres. Previous work by Carrano, Doonan, and Sumby showed how carboxylate substituents at the 4 position on pyrazolates produced porous MOFs.139–142 A Cu-

BPM MOF even demonstrated CO2 capture potential via molecular sieving though low surface areas (< 200 m2/g) and pore diameters (<3.5 Å) were measured.139 Further iterations by Doonan et. al. were able to increase the surface area and pore diameters at the cost of selectivity due to loss of the sieving effect.140 We therefore designed a new ligand, 3,5-dimethyl-1H-pyrazole-4- carboxylic acid (H2BPM) to explore these structures.

3.3 Results and Discussion

3.3.1 Synthesis and Properties of H2BPM Ligand

O O O OH

4.5 KOH O O eq 4.5 eq K2CO3 10 KOH 8 mol% [TBA][HSO4] N N N eq N N N DCM N 9:1 N O HO 65 °C, 2 days MeOH:H2O HN N 90 °C, overnight O O I II H2BPM

Scheme 3.1: Synthesis of ligand H2BPM.

As shown in Scheme 3.1, H2BPM is synthesized in two steps from the literature compound 3,5- dimethylpyrazole-4-carboxylic acid ethyl ester (I).143,144 A subsequent substitution reaction in dichloromethane followed by ester hydrolysis afforded the final ligand in good yield. The

benefits of using H2BPM for MOF synthesis can be seen in Figure 3.1. This ligand can act as a typical carboxylate linear ligand (orange) while maintaining a degree of flexibility via the

methylene linker (pink). In addition, pre- or postsynthetic metalation of H2BPM is also possible by coordinating a metal to the pyrazolyl nitrogen chelate site (blue).

O O N N HO N N OH

Figure 3.1: Diagram of H2BPM with carboxylate (orange), flexible methylene linker (pink), and potential pyrazolyl chelate site (blue) indicated.

3.3.2 Synthesis and Crystal Structure Description

3.3.2.1 [Ni(BPM)]n∙xDMSO (5)

Reacting Ni(OAc)2 tetrahydrate with 1 eq of H2BPM in a 10:4:1 mixture of DMSO, DMF, and methanol affords tetragonal green crystals of MOF (5). The nickel center is distorted octahedral and bonds to three BPM ligands. Two ligands bond κ2 via their carboxylate groups. The other BPM chelates to the nickel centre via the pyrazolyl nitrogen lone pairs (Figure 3.2a). Due to the distorted octahedral geometry the three BPM ligands assume various angles relative to the nickel centre. The chelating BPM is bent to a dihedral angle of 128.5° between the pyrazolyl rings

(Figure 3.2b). The carboxylate bound BPM are separated by 117° relative to one another as measured from the central carbon of the pyrazolyl ring to the nickel centre (Figure 3.2c). Two

unique nickel environments, Δ-[Ni(BPM)]n∙xDMSO and Λ-[Ni(BPM)]n∙xDMSO, are possible and are equally distributed within the MOF.

Due to these ligand distortions, close packing of the building units and/or π-stacking of the pyrazolyl rings are not possible. Instead, the MOF is assembled from a series of small, concentric

rings (Figure 3.3a).The units within the concentric rings are related by S4 and C2 symmetry. The rotation axis follows the c direction. From these rings, the pores of the MOF are constructed. Whereas small pore openings consist of only one ring assembly, multiple rings work in concert (under geometric constraints) to assemble larger pores (Figure 3.3b). This series of large and

small pores form the [Ni(BPM)]n∙xDMSO MOF (Figure 3.4). The MOF has alternating large (20.5Å x 4.1Å) and small (3.6Å x 3.6Å) channels running down the a and b direction of the material (Figure 3.4a, b). Along the c direction of 5, only small channels are observed (Figure 3.4c). Due to these channels, 5 is roughly 50% empty space, perfect for gas sorption studies.

3.3.2.2 [Cd(BPM)]n∙xDMSO (6)

Reacting Cd(OAc)2•x-hydrate with 1 eq of H2BPM affords clear tetragonal crystals of MOF (6). This MOF is isostructural to 5. The most notable difference is an elongation of the unit cell dimensions (Table 3.5) to account for the larger diameter of the Cd2+ cation. This means that 6 is also constructed from a distorted octahedral metal centre coordinating to three BPM ligands (Figure 3.5a). These then assemble into rings resulting in large and small pores (Figure 3.5b).

Figure 3.5: a) Building unit of [Cd(BPM)]n∙xDMSO, 6; BPM ligands truncated for clarity. b) 6 viewed down the b direction showing large and small pores.

Unfortunately, 6 was not a permanently porous structure likely due to the weaker bonds between BPM and the d10 cadmium (II) metal centre. This is a consequence of the lack of ligand-field stabilization for cadmium (II) complexes.145 Under high vacuum (necessary to activate the MOF) 6 collapsed to form an amorphous material. This collapse occurred in less than 5 minutes. A dropcast MOF sample was observed via PXRD spectroscopy to change into a different material on route to an amorphous compound (Figure 3.6). As a consequence, it was neither possible to collect a high quality single crystal X-ray structure nor study the gas adsorption properties of 6.

1400

1200

1000

800 Fully Air Dried 600 ~5 min Air Drying Damp Sample 400 Predicted Relative Intensity (a.u.) Intensity Relative 200

0 5 10 15 20 2θ (°)

3.3.3 Gas Sorption Studies of 5 3.3.3.1 Surface Area, Uptake and Heat of Adsorption

To confirm the porosity, 5 was activated via solvent exchange with chloroform followed by heating under reduced pressure. Despite numerous solvent exchanges, it was determined that 0.6

molecules of DMSO per nickel centre remained within the MOF’s pores. N2 isotherms were collected at 77K and showed a BET surface area of 1168(8) m2/g with a void volume of 0.807cm3/g. This moderate surface area and permanent porosity is a notable improvement over MOFs constructed from similar ligands.132–138

Given the large pore volume, we investigated the gas sorption properties of 1 towards CO2, CH4,

and N2. At 298K and 1 bar 5 showed uptakes of 3.71 mmol/g, 1.76 mmol/g, and 0.47 mmol/g for

CO2, CH4, and N2, respectively. Adsorption measurements for CO2 and CH4 at various temperatures (e.g. Figure 3.7) allowed for the determination of the enthalpy of adsorption. At

maximum loading, ΔHad was determined to be 19.8(4) kJ/mol, 31.6(4) kJ/mol, and 27.8(7)

kJ/mol for N2, CO2, and CH4, respectively, which are comparable to other MOFs. However,

there were difficulties determining the ΔHad of N2.

2.5

2 Measured CH at 278K

Calculated CH₄ at 278K 1.5 Measured CH ₄ at 288K Calculated CH₄ at 288K 1 Measured CH ₄ at 298K Calculated CH₄ at 298K 0.5 ₄ Amount Adsorbed Adsorbed Amount (mmol/g) 0 0 200 400 600 800 1000 1200 Pressure (mbar)

Figure 3.7: Selected gas uptake plots for methane adsorbing onto 5. The full figure is shown in section 3.4.

3.3.3.2 Ideal Adsorbed Solution Theory (IAST) Selectivity

Given that the enthalpies of adsorption for 5 were within the range of other selectively adsorbing

MOFs, 5 was investigated for post-combustion CO2 capture and coal mine methane capture.

Using single component gas sorption isotherms, binary isotherms for post-combustion CO2

capture (CO2 vs N2) and coal mine methane purification (CH4 and N2) can be calculated using ideal adsorbed solution theory (IAST).146,147 IAST approximates the adsorption equilibria of a gaseous mixture by using classical surface thermodynamics to derive an expression similar to Raoult’s law.

Using this approximation allowed for the determination that at 1.2 bar and 298 K 5 would have a predicted IAST selectivity of 77.5 (Figure 3.8). While substantial, this selectivity is not

competitive with the best performing MOFs for post combustion CO2 capture. Materials with selectivities in excess of 200 have been reported.148 Conversely, the IAST selectivity of 5 for a

1:1 mixture of CH4/N2 at 298K and 2 bar was 8.0 (Figure 3.9). Not only is this comparable to

Cu(INA)2 (which did not use IAST to determine maximum selectivity) but it is in excess of the 124 highest observed methane selectivity for Cu(INA)2.

2.5 80 2

60 Selectivity 1.5

40 1

(mmol/g) Loading 0.5 20

0 0 0 300 600 900 1200 Total Pressure (mbar) Loading CO Loading N selectivity

₂ ₂ Figure 3.8: IAST determined selectivity for a 20% CO2/80% N2 gas stream at 298 K adsorbing onto 5.

1.8 10

1.5 8 1.2 Selectivity 6 0.9 4 0.6

(mmol/g) Loading 0.3 2 0 0 0 400 800 1200 1600 2000 Total Pressure (mbar) Loading CH Loading N selectivity

₄ ₂ Figure 3.9: IAST determined selectivity for a 1:1 mixture of CH4 /N2 at 298 K adsorbing onto 5.

3.3.3.3 Computational Investigations

However, as noted, IAST relies on extrapolation of data and the use of assumptions to predict dual component isotherms. Coal mine methane capture typically has an operating pressure between 2 to 5 bar. However, our instruments could only measure the single component adsorptions to 1.2 bar. As the data is further extrapolated, IAST becomes less reliable. In order to overcome these difficulties, Monte Carlo simulations were conducted to compare 5 with the best known demonstrated and predicted MOFs for coal mine methane capture.123,124

In order to validate the gas sorption model, dual component isotherms were simulated and compared to IAST calculations within their range of accuracy. In order to simplify the calculations, gas uptake in 5 was modelled for a fully activated MOF (i.e. no residual pore solvent). The model was not a perfect match to the experimental system (Figure 3.10). Multiple simulations determined that the residual DMSO present in the pore (9.6 molecules per unit cell) was enhancing the selectivity of the MOF for methane. This has been observed in other systems where residual pore solvent acts to enhance the observed gas uptake and selectivity.149

2.0 1.8 1.6

1.4 1.2 Experimental CH Uptake 1.0 Experimental N Uptake₄ 0.8 Simulated CH Uptake₂ 0.6 Simulated N Uptake

Gas Adsorbed(mmol/g) ₄ 0.4 ₂ 0.2 0.0 0.00 0.50 1.00 1.50 2.00 Pressure (bar)

Figure 3.10: Comparison of IAST derived dual component gas uptake to Monte Carlo predicted gas uptake for 5 with no residual pore solvent.

In order to overcome these challenges a Monte Carlo simulation allowing for optimization of both the adsorbed gas and residual solvent is required. However, such a simulation is non-trivial to perform and very compute intensive due to the large number of molecules and variables required for the model. Such calculations are ongoing as of this writing. However, insight into the performance of 5 for coal mine methane capture may still be approximated with the simple, fully evacuated model. Comparing this model to other reported MOFs reveals that 5 has the highest selectivity of all known coal mine methane MOFs at 1 bar (Table 3.1, Figure 3.11).

Below 1 bar, 5 meets or exceeds the best proven MOF for coal mine methane, Cu(INA)2. Furthermore, the predicted selectivity for 5 at 1 bar is 11.74 which is in excess of both the 124 observed (7.0-7.6) and predicted (8.36) selectivity for Cu(INA)2. Lastly, it must be noted that these calculations are very sensitive to the choice of force field used. Monte Carlo originally over

predicted uptake in Cu(INA)2 threefold by using a N2 force field applicable for microporous

MOFs but that did not model the extremely small pore sizes and surface areas of Cu(INA)2 effectively.

14.00

12.00

) 2 10.00 /N 4 8.00

6.00

Selectivity (CH Selectivity 4.00

2.00

0.00 0 20 40 60 80 100 Pressure (kPa)

Ni-MOF (no DMSO) Cu(INA)2 CPO-27-Co MIL-47 ZIF-8 ZIF-7

Figure 3.11: Monte Carlo simulated coal mine methane selectivities for reported MOFs123,124

versus fully activated [Ni(BPM)]n.

Monte Carlo simulations were also used to estimate the working capacities of the MOFs reported for coal mine methane capture (Table 3.1). Comparing this data indicates that 5 should again

have comparable performance to Cu(INA)2 and is a significant improvement over other MOFs that have been tested for coal mine methane capture. It should be further noted that with proper corrections for the residual DMSO present in 5, these performances should further improve.

Table 3.1: Monte Carlo simulated coal mine methane working capacities and selectivities for 123,124 reported MOFs in comparison to fully activated [Ni(BPM)]n.

CH Working Capacity Selectivity MOF 4 (mmol/g) (1 bar)

[Ni(BPM)]n 1.44 11.7

Cu(INA)2 1.58 11.4

MIL-47 1.06 6.90

CPO-27-Co 0.66 5.86

ZIF-8 0.51 5.20

ZIF-7 0.48 4.04

3.3.4 Conclusions

We have produced two new MOFs, 5 and 6, constructed from a simple dicarboxylate

functionalized bispyrazolylmethane ligand, H2BPM. 5 exhibits moderate CO2 absorption

selectivity versus N2. 5 is also an excellent selective absorbent for CH4 versus N2 comparable to

the best known MOF, Cu(INA)2, as reported in the literature. Therefore, 5 has excellent potential as an adsorbent for coal mine methane capture. The isostructural MOF 6 was not stable towards gas sorption due to the weaker bonds between the ligand and cadmium,. Ongoing work includes

investigating how to extend this type of system towards other CO2 capture MOFs and exploring other possible MOFs using metals with different preferred coordination numbers and geometries.

3.4 Experimental 3.4.1 General Considerations

Elemental analyses were performed on a Thermo Flash 2000 CHN analyzer. Thermogravimetric analyses (TGA) were performed on a TA Instruments SDT Q600 instrument under a nitrogen atmosphere with a heating rate of 5 °C per minute in an alumina sample pan with an empty pan as a reference. NMR spectra were recorded on a Bruker Avance 400 spectrometer. Both 1H and 13C NMR spectra were referenced and reported relative to the solvent's residual signals. Powder X-Ray Diffraction (PXRD) experiments were performed on a Rigaku MiniFlex 600 diffractometer equipped with a Cu-Kα source operating at 40 kV/15 mA at the Walter Curlook Materials Characterization & Processing Laboratory at the University of Toronto Department of Materials Science and Engineering. A step scan mode was used for data acquisition with a step size of 0.02° 2θ. All the PXRD samples were prepared by dropcasting with acetone powdered sample onto a silicon zero background sample holder. The infrared spectra of 5 and 6 were recorded using a neat sample on a Bruker Alpha FT-IR spectrometer equipped with a Platinum ATR sampling unit in air. The adsorption isotherms were conducted using an Accelerated Surface Area & Porosimetry System (ASAP) 2020 supplied by Micromeritics Instruments Inc. All reagents were purchased from commercial sources and used without further purification. Ethyl diacetylacetate was prepared following literature procedures.143

3.4.2 Ligand Synthesis

3.4.2.1 Synthesis of 3,5-dimethylpyrazole-4-carboxylic acid ethyl ester (I)

In a modification of literature procedures,144 ethyl diacetylacetate (25.0 g, 145 mmol) was dissolved in 290 mL of methanol in a 0 °C ice bath. 12 M hydrochloric acid (1.5 mL, 18 mmol) was added to the solution and stirred for 20 min. Hydrazine monohydrate (17.5 mL, 359 mmol) was added dropwise to the solution after which the flask was allowed to slowly warm to room temperature overnight. After stirring, the methanol was removed under reduced pressure and 100 mL of water was introduced to induce precipitation white solids. The solids were collected on a Brukner funnel, washed with water (5 x 20 mL), and air dried. The crude product was dissolved in dichloromethane (DCM) and filtered through silica gel. After removal of DCM under reduced pressure 20 g of white solids (81 % yield) were collected. Spectroscopic data matches literature data.

3.4.2.2 Synthesis of diethyl 1,1’-methylenebis(3,5-dimethyl-1H-pyrazole-4- carboxylate) (II)

Following literature procedures,150 I (10 g, 59.45 mmol) was added to freshly powdered potassium hydroxide (15 g, 267.55 mmol), potassium carbonate (37 g, 267.55 mmol) and tetra-n- butylammonium hydrogen sulfate (1.62 g, 4.76 mmol) in DCM (400 mL). The solution was refluxed at 65 °C until it assumed a deep orange colour (normally 2-3 days). The solution was rinsed once with water (~400 mL) and four times with brine (~200 mL). The collected organic phase was dried over magnesium sulfate, filtered and dried to yield 12 g (85 % yield) of white solids. Crystals suitable for EA and x-ray analysis may be grown via the slow diffusion of 1 pentane into chloroform. H NMR (400 MHz, CDCl3) δ 6.10 (s, 2H), 4.26 (q, J = 7.1 Hz, 4H), 13 2.73 (s, 6H), 2.37 (s, 6H), 1.33 (t, J = 7.1 Hz, 6H). C NMR (101 MHz, CDCl3) δ 164.36,

151.50, 145.89, 111.06, 60.09, 59.91, 14.51, 14.44, 11.56. Anal. Calc. for C17H24N4O4: C 58.61, H 6.94, N 16.08; found C 58.50, H 6.92, N 16.14.

50000

45000

40000 1.33 2.37 2.73 4.27 4.26 6.10 7.26 Chloroform-d

35000

30000

25000 Intensity 20000

15000

10000

5000

0 5.86 6.02 6.00 3.87 2.01

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 f1 (ppm)

1 Figure 3.12: H NMR spectrum of II in CDCl3.

24000

22000

20000

18000 11.56 14.44 14.51 59.91 60.09 77.16 Chloroform-d 111.06 145.89 151.50 164.36

16000

14000

12000

10000 Intensity

8000

6000

4000

2000

-2000

160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 3.13: C NMR spectrum of II in CDCl3.

3.4.2.3 Synthesis of 1,1’-methylenebis(3,5-dimethyl-1H-pyrazolyl-4- carboxylic acid) (H2BPM)

II (1.04 g, 2.87 mmol) was combined with freshly powdered potassium hydroxide (1.61 g, 28.7 mmol) in a 9:1 methanol (45 mL) water (5 mL) solution. The solids were suspended in solution and refluxed overnight at 90 °C. The clear yellow solution was quenched with 12 M hydrochloric acid (2.4 mL, 28.7 mmol) resulting in the precipitation of white solids. The solvent was removed under vacuum and the white solids were suspended in water, collected via vacuum filtration and washed four times with water to yield 2.1 g of crude material. Pure crystals suitable for elemental analysis and MOF synthesis may be grown as needed via the slow diffusion of water into a saturated solution of DMSO at 65 °C. Yield of the first crop was 240 mg (43% 1 yield). H NMR (400 MHz, DMSO-d6) δ 12.32 (s, 2H), 6.23 (s, 2H), 2.67 (s, 6H), 2.24 (s, 6H). 13 C NMR (101 MHz, DMSO-d6) δ 165.00 , 150.22 , 145.26 , 110.11 , 58.85 , 14.04 , 10.88.

Anal. Calc. for C13H16N4O4: C 53.42, H 5.52, N 19.17; found C 53.25, H 5.56, N 19.29.

120000

110000

100000

90000

80000 2.24 2.50 Dimethyl Sulfoxide-d6 2.67 6.23 12.32

70000

60000 Intensity 50000

40000

30000

20000

10000

0 3.00 3.01 0.99 0.98 -10000

12.5 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 f1 (ppm)

1 Figure 3.14: H NMR spectrum of H2BPM in DMSO-d6.

35000

30000

25000 10.88 14.04 39.52 Dimethyl Sulfoxide-d6 58.85 110.11 145.26 150.22 165.00 20000

15000 Intensity

10000

5000

170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 3.15: C NMR spectrum of H2BPM in DMSO-d6.

3.4.3 Synthesis of MOFs

3.4.3.1 Synthesis of [Ni(BPM)]n∙xDMSO (5)

Solutions of H2BPM (340 mg, 1.163 mmol) in 17 mL of DMSO and Ni(OAc)2•4H2O (289 mg,

1.163 mmol) in DMF (6.8 mL) and methanol (1.7 mL) were prepared. 1 mL of H2BPM solution was added to 2 dram scintillation vials and 0.5 mL of nickel solution was syringed on top. The vials were sealed and left to heat at 110 °C for five days in an aluminum block. After heating, 332 mg (56 % yield) of x-ray quality electric lime green crystals were collected from solution and washed with DMSO. Elemental analysis of vacuum dried sample showed 1.35 DMSO

molecules remaining (Fig S5). Anal. Calc. for NiC13H14N4O4•(C2H6SO)1.35: C 41.49, H 4.90, N 12.33; found C 41.13, H 4.40, N 11.88.

3.4.3.2 Synthesis of [Cd(BPM)]n∙xDMSO (6)

Solutions of 30 mg (0.103 mmol) H2BPM in 6 mL of DMSO (A) and 25.2 mg (0.103 mmol) of

Cd(OAc)2•xH2O (assumed Cd(OAc)2•H2O) in 6 mL of distilled water (B) were prepared. The solutions were equally proportioned into 2 x 20 scintillation vials, sealed, and left to react at 90 °C for 18 days. After cooling the white solids were collected on a Hirsch funnel, washed three times with the mother liquor and once with DMSO. 14.0 mg (61.5 % yield) of solids were

collected. Anal. Calc. for C13H14N4O4Cd•2(C2H6SO): C 36.53, H 4.69, N 10.02; found C 36.22, H 4.74, N 9.62.

3.4.4 Thermal Gravimetric Analysis, Powder X-Ray, and IR

100.00%

80.00%

60.00%

40.00% MassPercent (%) 20.00%

0.00% 25 125 225 325 425 525 Temperature (°C)

Figure 3.16: TGA spectrum of partially dried [Ni(BPM)]n∙xDMSO (5).

100%

80%

60%

40% MassPercent

20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 3.17: TGA spectrum of partially dried [Cd(BPM)]n∙xDMSO (6).

Wavenumber (cm-1) 3900 3400 2900 2400 1900 1400 900 400 100%

90%

80%

70%

60%

Percent Transmission (%) Transmission Percent 50%

40%

Figure 3.18: IR Spectrum of neat [Ni(BPM)]n∙xDMSO MOF (5).

Wavenumber (cm-1) 3900 3400 2900 2400 1900 1400 900 400 100%

90%

80%

70%

60% Percent Transmission Percent 50%

Figure 3.19: IR Spectrum of air dried [Cd(BPM)]n∙xDMSO (6) submitted for EA. Note changes in relative peak intensity due to collapsed structure of 6 in comparison to 5.

1600

1400

1200

1000

800 Observed 600 Predicted 400 Relative Intensity (a. (a. u.) Intensity Relative 200

0 5 10 15 20 25 30 35 40 2θ (°)

Figure 3.20: PXRD pattern of as-synthesized [Ni(BPM)]n∙xDMSO (5).

3.4.5 Gas Sorption Studies

3.4.5.1 Adsorption Analysis

5 was prepared for gas sorption analyses by exchanging the solvent with methanol (3x, ~15 mL). Fresh methanol was used each time and allowed to sit for three hours. The methanol was then replaced with chloroform (3x, ~15 mL), using fresh chloroform each time. This was allowed to sit overnight. Finally, the solvent was exchanged with fresh CHCl3 (3x, ~15 mL) using fresh chloroform each time. The sample was then loaded into an analysis tube in a suspended solution of chloroform with excess solvent removed manually. The sample was then slowly evacuated on a Schlenk Line at approximately 10-3 mbar for ninety minutes before being backfilled with argon. Finally, the sample was activated on the ASAP2020 instrument for two hours at 60⁰C followed by 16 hours of activation at 100⁰C at which point no outgassing was observed. The sample was then backfilled with N2 before being transferred to the analysis port where it was evacuated for a further 120 minutes before the analysis was started.

3.4.5.2 Isotherm Fitting

Adsorption isotherms were fit to a Dual Site Langmuir model with fits used to achieve a value as close as possible to R2 = 1. For these isotherms:

( ) ( ) = 1 = 1 ( )2 ( ) 2 2 ∑ 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 ∑ 𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 − 𝑛𝑛𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑅𝑅 − 2 − 2 ∑ 𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 ∑ 𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 Equation 3.1: Fitting parameter for gas sorption isotherms.

3.4.5.3 Isosteric Heat of Adsorption Calculation

From the isotherm fitting, a modified Clausius-Clapeyron equation was used with all six 151 temperatures to determine the heat of adsorption (ΔHads) for CH4, CO2, and N2.

Table 3.2: Correlation coefficient for isotherms fits and calculating isosteric heat of adsorption

using the Clausius-Clapeyron equation for [Ni(BPM)]n∙xDMSO (5).

CH4 CO2 N2

R2 for isotherm fitting 0.999927 0.996277 0.99999 (average ± st. dev.) ± 0.000065 ± 0.005984 ± 0.000001

Min. R2 for isotherm 0.999826 0.984333 0.999997 fitting

R2 for ΔH ads 0.998521 0.99924 0.998734 calculation ± 0.000578 ± 0.000283 (average ± st. dev.) ± 0.001228

Min. R2 for ΔH ads 0.994160 0.997917 0.998086 calculation

0.0008

0.0007

0.0006

1)] 0.0005 y = 0.0032x + 5E-06 R² = 0.9993 0.0004

0.0003 1 / [n (Po/P[n - 1 /

0.0002

0.0001

0 0 0.05 0.1 0.15 0.2 0.25 Relative Pressure (P/Po)

Figure 3.21: BET_des surface area for N2 at 77 K within 5.

0.0014

0.0012

0.001 y = 0.0044x + 2E-05

1)] R² = 0.9957

- 0.0008

0.0006 1 / [n(Po/P 0.0004

0.0002

0 0 0.05 0.1 0.15 0.2 0.25 0.3 Relative Pressure (P/Po)

Figure 3.22: BET_des surface area for CO2 at 195 K within 5.

6 Measured CO at 273K Calculated CO at 273K 5 ₂ Measured CO at 278K ₂ Calculated CO at 278K 4 ₂ Measured CO at 283K ₂ Calculated CO at 283K 3 ₂ Measured CO at 288K ₂ Calculated CO at 288K 2 ₂ Measured CO at 293K ₂ Calculated CO at 293K ₂

Amount Adsorbed Adsorbed Amount (mmol/g) 1 Measured CO at 298K ₂ Calculated CO at 298K 0 ₂ 0 200 400 600 800 1000 1200 ₂ Pressure (mbar)

Figure 3.23: mmol of CO2 adsorbed per gram of 5 versus pressure.

3 Measured CH at 273K Calculated CH at 273K 2.5 ₄ Measured CH at 278K ₄ Calculated CH at 278K 2 ₄ Measured CH at 283K ₄ Calculated CH at 283K 1.5 ₄ Measured CH at 288K ₄ Calculated CH at 288K 1 ₄ Measured CH at 293K ₄ Calculated CH at 293K ₄

Amount Adsorbed Adsorbed Amount (mmol/g) 0.5 Measured CH at 298K ₄ Calculated CH at 298K 0 ₄ 0 500 1000 ₄ Pressure (mbar)

Figure 3.24: mmol of CH4 adsorbed per gram of 5 versus pressure.

1 Measured N at 273K Calculated N at 273K ₂ 0.8 Measured N at 278K ₂ Calculated N at 278K ₂ Measured N at 283K 0.6 ₂ Calculated N at 283K ₂ Measured N at 288K 0.4 ₂ Calculated N at 288K ₂ Measured N at 293K ₂ Calculated N at 293K 0.2 ₂ Measured N at 298K ₂ Amount Adsorbed Adsorbed Amount (mmol/g) Calculated N at 298K 0 ₂ 0 500 1000 ₂ Pressure (mbar)

Figure 3.25: mmol of N2 adsorbed per gram of 5 versus pressure.

ΔH ads 20 (kJ/mol)

0 0 1 2 3 4 5 Ammount Adsorbed (mmol/g)

Figure 3.26: Heat of absorption for CO2 within 5.

ΔH ads 15 (kJ/mol)

0 0 0.5 1 1.5 2 Ammount Adsorbed (mmol/g)

Figure 3.27: Heat of Absorption for CH4 within 5.

ΔHads (kJ/mol) 10

0 0 0.05 0.1 0.15 Ammount Adsorbed (mmol/g)

Figure 3.28: Heat of absorption for N2 within 5.

3.4.5.4 Post Adsorption Residual Solvent Analysis via Thermal Gravimetric and Elemental Analysis

Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) was performed using a NETZSCH STA 409PC Luxx. The sample was loaded into the aluminum sample carrier, with an empty aluminum sample carrier as a reference. The sample was then heated to 450 °C at 2°C min-1.

100

40 MassPercent (%) 30

0 0 50 100 150 200 250 300 350 400 450 500 Temperature (°C)

Figure 3.29: Post adsorption TGA of 5 showing 15.80% mass loss at 250°C: 11.84% from DMSO, 3.96% from surface moisture (sample left in air for ~2 weeks before analysis).

Table 3.3: Post adsorption elemental analysis of 5.

Run 1 Run 2 Average C13H14N4O4Ni-(C2H6SO)0.6

%C 43.35 43.40 43.38 43.08

%H 4.79 4.85 4.82 4.48

%N 13.75 13.81 13.78 14.15

3.4.6 Computational Studies

3.4.6.1 Overview

Using the Grand Canonical Monte Carlo (GCMC) algorithm, we can accurately predict the gas adsorption properties of MOFs from single crystal structures. GCMC relates the microscopic energy interactions between guest molecules and the MOF framework to macroscopic properties such as total gas uptake, selectivity, and heat of adsorption. The GCMC algorithm accomplishes this by sampling a statistical mechanical ensemble.

3.4.6.2 The Monte Carlo Algorithm

An ensemble refers to a large number of replicas of a system. These replicas have the same thermodynamic variables such as temperature, pressure, chemical potential, and volume, but the atoms in each replica have different positions and momenta. Together, the replicas describe all possible atomic configurations in a system at a given set of thermodynamic variables. As an example of an ensemble, consider a set of guest molecules residing in all possible locations in the pores of a MOF. Each configuration of the guest molecules in the pores corresponds to a single state of the ensemble. The specific positions of the guests in each configuration allows for determination of the total potential energy of the system. A generation of all possible configurations of guest positions in the pores is referred to as ensemble. The sampling of guest positions is generated to satisfy the statistical mechanical distribution function corresponding to the specific ensemble. The type of ensemble being sampled depends on which thermodynamic variables are kept constant. In a canonical ensemble the temperature, volume and number of molecules are fixed. In order to obtain the macroscopic properties of a system, the ensemble must be sufficiently sampled over an extended number of configurations.

Statistical ensembles are commonly sampled using a Monte Carlo algorithm that randomly perturbs (translates/rotates/inserts/deletes) the molecules in the pores. A configuration with a high potential energy is less likely to be observed and contributes less to the average macroscopic properties. In order to prevent these disfavoured states from being sampled frequently, every perturbation is evaluated based on a set of rules called the Metropolis Monte Carlo algorithm. In the case of a canonical ensemble, the rules are as follows:

1) A trial perturbation (translation or rotation) is accepted if the energy of the newer state is lower than the previous state.

2) If the energy of the new state is higher, it is accepted based on a probability proportional

to exp(-∆E/kBT), which is a Boltzmann weighting (canonical ensemble distribution

function) of the difference in energy between the two states, where kBT is the thermal energy of the system.

3) If the random perturbation is rejected, then the previous state is counted again towards the statistical average ensuring the favourable sampling of lower energy states.

When the number of guest molecules is fixed, sampling the ensemble yields a spatial 3D probability distribution of the guest positions in the MOF pores.

3.4.6.3 The Grand Canonical Monte Carlo Simulation for Calculating Gas Isotherms

Determination of the amount of gas that a MOF will adsorb given an external pressure of bulk phase gas at a specific temperature is a frequent computational goal. This requires modelling an open system that allows for the transfer of particles between the MOF and the gas phase. The chemical potential, system volume, and temperature of the sampled grand canonical ensemble must therefore be fixed but the number of particles is allowed to fluctuate. Equilibrium occurs when the chemical potential of the guests in the adsorbed phase equals the guests in the gas phase. The bulk gas phase can be approximated as an ideal gas to determine the chemical potential where P0 is the standard pressure of the system, is the standard state chemical 0 potential of the gas at P0, and R, T, and P are the gas constant,𝜇𝜇𝑖𝑖𝑖𝑖 temperature and pressure of the gas, respectively (Equation 3.2).

( , ) = + 𝑃𝑃 𝜇𝜇𝑖𝑖𝑖𝑖 𝑝𝑝 𝑇𝑇 𝜇𝜇𝑖𝑖𝑖𝑖 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 � � 𝑃𝑃0 Equation 3.2: The bulk phase chemical potential of an ideal gas.

Metropolis Grand Canonical Monte Carlo sampling includes two new moves in addition to translation and rotation: the insertion and removal of particles.152 In order to efficiently sample the Boltzmann distribution of states which in the grand canonical ensemble becomes the grand

canonical probability distribution, the Monte Carlo moves are accepted or rejected based on the following probabilities: translating/rotating a molecule already in the pore (Equation 3.3), inserting a new gas molecule into the pore (Equation 3.4), and removal of a gas molecule from the pore (Equation 3.5). Where N is the number of gas molecules in the pores of the MOF, E is the internal energy of the states (described by subscripts), V is the total system volume, and P represents the pressure of the molecules in the gas phase.152,153

( ) ( ) = min 1, exp

− 𝐸𝐸𝑛𝑛𝑛𝑛𝑛𝑛 − 𝐸𝐸𝑜𝑜𝑜𝑜𝑜𝑜 𝑃𝑃 𝑜𝑜𝑜𝑜𝑜𝑜 → 𝑛𝑛𝑛𝑛𝑛𝑛 � � �� 𝑘𝑘𝐵𝐵𝑇𝑇 Equation 3.3: Probability of a molecular translation/rotation within the pore.

( ) ( + 1) = min 1, exp ( + 1) 𝑉𝑉𝑉𝑉 − 𝐸𝐸𝑁𝑁+1 − 𝐸𝐸𝑁𝑁 𝑃𝑃 𝑁𝑁 → 𝑁𝑁 � � �� 𝑘𝑘𝐵𝐵𝑇𝑇 𝑁𝑁 𝑘𝑘𝐵𝐵𝑇𝑇 Equation 3.4: Probability of addition of a new gas molecule to the pore.

( ) ( 1) = min 1, exp ( ) 𝑁𝑁 − 𝐸𝐸𝑁𝑁−1 − 𝐸𝐸𝑁𝑁 𝑃𝑃 𝑁𝑁 → 𝑁𝑁 − � � �� 𝑘𝑘𝐵𝐵𝑇𝑇 𝑉𝑉𝑉𝑉 𝑘𝑘𝐵𝐵𝑇𝑇 Equation 3.5: Probability of removal of a gas molecule from the pore.

In order to obtain accurate thermodynamic averages, approximately 107 Monte Carlo moves consistent with the three conditions discussed above need to be performed. If an ab initio energy calculation was performed at each step, this would take approximately 4 months. For this reason we perform classical force field GCMC calculations. In order to make the classical GCMC simulations feasible, two approximations are made. First, the guest and framework structures are considered rigid. Second, the intermolecular interactions (van der Waals and electrostatic interactions between guest-guest and guest-framework atoms) are divided into short-range steric and dispersion interactions, and long-range Coulombic interactions. These approximations allow for thermodynamic properties such as gas uptake in the MOF to be accurately calculated except for MOFs with flexible structures or open metal sites.

100

When simulating the uptake of gas mixtures, we need to apply acceptance conditions to each gas

in the mixture with its special partial pressure, Pi. Gas adsorption properties for 5 were calculated using an in-house developed grand canonical Monte Carlo (GCMC) code written based on the DL_POLY 2.0 molecular dynamics package at the University of Ottawa.154

3.4.6.4 Representations of Intermolecular Forces

The Lennard-Jones (LJ) potentials were used to model the van der Waals non-bonding dispersion interactions. Ewald summations were used to represent the electrostatic interactions. The LJ parameters for the frameworks were taken from the Dreiding force field155 supplemented with the universal force field (UFF).156 Partial atomic charges were calculated using the REPEAT method.157 The REPEAT method uses density functional theory (DFT) gauge-modified electrostatic potential calculated using VASP158,159 with the PBE160 functional with a planewave

cut-off a 520 eV. The LJ potential parameters for CO2 and CH4 guest molecules were taken from 161 the TraPPE force field and the LJ potential parameters for N2 guest molecules were taken from 162 the NIMF force field. Cu(INA)2 required the a simplified N2-TraPPE force field calculation.

The N2-NIMF force field was developed in house at the University of Ottawa. The parameters used for the N2-NIMF force field may be found in Table 3.4. GCMC simulations were run for 20,000 equilibration cycles and 20,000 production cycles. A cycle consists of a certain fixed number of distinct GCMC.

Species Atom bl (Å) q (e) (ϵ/kb) (K) σ (Å)

N 0.5500 -0.4820 39.966 2.4549

COM 0.0 0.9640 0 0

Variable (~0.05 - H - 22.142 2.5711 N2 0.2)

Variable (~0.2 - B - 90.581 3.6375 0.4)

N - Variable (~ -0.4) 34.722 3.2607

101

3.4.6.5 Comparison of Simulation Results to Experimental Results

GCMC simulations can calculate gas uptake for gas mixtures from which selectivity for one component over another can be obtained. Experimentally, for a gas mixture, only the sum of the uptake of the two gases can be measured. In order to compare simulated selectivities to experimental values, experimental single component gas isotherms were converted to binary gas isotherms using pyIAST with the isotherms fit at the temperatures and mixtures indicated in the figure captions.146,147 From the binary gas isotherms, selectivity for the gas of interest (gas 1) over the other gas (gas 2) was calculated using Equation 3.6.

= 𝑔𝑔𝑔𝑔𝑔𝑔 1 𝜇𝜇 � 𝑔𝑔𝑔𝑔𝑔𝑔 1� 𝑝𝑝 𝑆𝑆 𝜇𝜇𝑔𝑔𝑔𝑔𝑔𝑔 2 �𝑝𝑝𝑔𝑔𝑔𝑔𝑔𝑔 2� Equation 3.6 Selectivity of gas 1 versus gas 2 where μ = gas uptake and p = partial pressure.

3.4.7 X-Ray Diffractometry

Å) for 5 and a Bruker Kappa Apex II CCD diffractometer with Mo Kα radiation (λ = 0.71073 Å) for 6 operating at 50 kV and 30 mA at low temperature controlled by an Oxford Cryostream 700 series system. The data integration and absorption correction were performed with the Bruker Apex 2 software package.103 The structure was solved by direct methods and refined using SHELXTL V2016/4.104,105 Disordered solvent molecules and residual electron density were removed using the Platon Squeeze method.106 All non-hydrogen atoms except for the solvent atoms were refined anisotropically. The positions of the hydrogen atoms were calculated using the riding model.

The microporous nature of 5 and 6 necessitated special collection methods. In order to collect a complete data set it was necessary to utilize Cu Kα radiation for 5. While Mo Kα radiation did yield a solvable structure, the R1 and wR2 values were outside of acceptable limits. Due to the rapid structural collapse of 6 upon desolvation it was not possible to collect a complete data set.

102

As a consequence, Mo Kα radiation was used on the best picked crystal to determine the approximate structure of 6 for comparison to 5. This structure should not be considered “solved”.

Table 3.5: Crystallographic properties of 5 and 6.

Name [Ni(BPM)]n∙xDMSO (5) [Cd(BPM)]n∙xDMSO (6)

Formula NiC13H14N4O4 + solvent CdC13H14N4O4 + solvent

Formula Weight [g/mol] 902.20 418.69

Crystal System Tetragonal Tetragonal

Space Group I41/a (No. 88) I41/a (No. 88)

a [Å] 19.5477(7) 20.4525(13)

b [Å] 19.5477(7) 20.4525(13)

c [Å] 29.8873(11) 31.291(4)

Volume [Å3] 11420.3(9) 13089(2)

Z 16 16

D(calc) [g/cm3] 1.049 0.850

μ(CuKα) [ /mm ] 1.928 -

μ(MoKα) [ /mm] - 0.682

F(000) 3744 3328

Crystal Size (estimated) [mm] 0.17 x 0.17 x 0.20 0.10 x 0.16 x 0.17

Temperature (K) 147 150

Radiation [Å] CuKa: 1.54178 MoKa: 0.71073

θ Min-Max [°] 2.7, 67.5 1.2, 22.2

Dataset -22: 23 ; -23: 22 ; -35: 35 -21: 21; -21: 21; -32: 33

Tot., Uniq. Data, R(int) 120306, 5106, 0.084 20218, 4034, 0.100

Observed data [I > 2.0 σ(I)] 4263 1785

Nref, Npar 5106, 276 4034, 138

103

R1, wR2, GooF 0.1125, 0.3108, 2.56 0.3056, 0.6229, 3.16

Max. and Av. Shift/Error 0.00, 0.00 0.37, 0.06

Min. and Max. Resd. Dens. -0.64, 0.79 -0.87, 2.34 [e/Å3]

104

Chapter 4 4 Structural Investigation of Bispyrazolylpropane MOFs 4.1 Abstract

n+ A series of new MOFs and coordination polymers were synthesized using M (OAc)n [M = Mn2+, Fe2+, Co2+, Cu2+, and Eu3+] and 1,1’-methylenebis(3,5-dimethyl-1H-pyrazolyl-4-

carboxylic acid) (H2BPM). These materials all proved to have unique structural characteristics. Co-BPM formed four unique coordination polymers (7-10) constructed from a dinuclear cobalt metal node, ancillary ligand, and BPM. Coordination polymer formation may be controlled by

altering the synthetic condition. [Mn(BPM)(H2O)2]n (11) and [Cu(BPM)(DMSO)]n (12) formed

close packed, non-porous MOF materials. [Eu(BPM)2(OAc)]n (13) formed 2D sheets. Although

a [Fe2(BPM)2(H2O)4(Bpy)0.5]n compound was also realized, purification, isolation, and characterization were not possible. These wide varieties of compounds, with the potential for future applications, highlight the importance of structural investigations and the versatility of the

H2BPM ligand.

4.2 Introduction

The synthesis of MOFs is iterative and very sensitive to synthetic conditions. Consequently, small changes in synthetic conditions (e.g. use of a different metal salt or solvent system) can lead to radically different structures. An early example was 1,3,5-bezene tricarboxylate. In the presence of cobalt(II) nitrate and pyridine, it produces a simple 2D sheet.6 The metal-carboxylate groups form 2D layers while pyridine, coordinating to the cobalt centres, facilitates stacking of the sheets.6 However, the use of copper (II) nitrate forms HKUST-1, a 3D porous MOF.7 As discussed at length in Chapter 3, we have successfully synthesized two microporous MOFs using

nickel (II) and cadmium (II) acetate that incorporates H2BPM (compounds 5 and 6).

[Ni(BPM)]n∙xDMSO (5) proved to be permanently porous and potentially useful for coal mine

methane purification. Given that H2BPM had already successfully formed a MOF with cadmium

(6), we set out to investigate the reactivity of H2BPM towards a variety of different metal salts and in the presence of ancillary ligands (4,4’-bipyridine [Bpy] and 4,4’-bipyridine-N-oxide [Bpy- oxide]). By tuning the synthetic conditions a wide variety of MOFs should be accessible using

the same H2BPM ligand.

105

The purpose of extending this chemistry to a variety of different metals and ancillary ligands was to a) determine whether synthesis of coordination polymers was possible and b) to investigate how the preferred geometries and coordination numbers of these metals would impact the resulting structures in comparison to 5 and 6. Our initial work focused on other first row transition metals (manganese, iron, cobalt, copper, zinc) in addition to europium. Manganese, cobalt, copper, and zinc were chosen because they are well known to serve as nodes in other MOF systems and are relatively inexpensive. Iron was selected in an attempt to construct stable FeII species that may be investigated for catalysis at the metal nodes such as the recent report by Guo et. al.163 A series of MIl-53(Fe)-type materials (MIL-53(Fe): node = FeIII, ligand = 1,4- benzenedicarboxylate) were prepared with nodes containing different ratios of FeII and FeIII. As the quantity of FeII increased, the MOF was found to be effective for the catalytic oxidation of phenol via the formation of hydroxyl radicals.163 Europium was investigated in an effort to produce a luminescent MOF and investigate the behaviour of BPM towards metals with preferred coordination numbers above 6.

4.3 Results and Discussion

4.3.1 Methods of Assessing BPM Ligand Conformation

As noted in chapter 3, the BPM ligand is flexible and has two potential binding sites for metals (the carboxylate and pyrazolyl sites). To assess the geometry of the BPM moiety in its structures, two key angles can be measured. Namely the angle formed by the carboxylate groups with the methylene linker (α, Figure 4.1:a) and the torsion angle of the pyrazolyl rings (β, Figure 4.1:b). The degree of bending along BPM indicates how close the ligand is to linearity. For example, an angle of 0° would indicate BPM is acting as a linear linker similar to 4,4’- bipheylcarboxylate. Conversely, torsion angles measure whether the pyrazolyl nitrogen lone pairs of BPM have the correct geometry to chelate to a metal centre (i.e. torsion angle of < 10°). Both metrics provided geometric explanations for how the structure of any new MOFs arose.

106

4.3.2 Cobalt Coordination Polymers

4.3.2.1 [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (7)

Reacting Co(OAc)2 tetrahydrate with 1 eq of H2BPM and 0.5 eq of Bpy in a DMF/H2O solution affords triclinic pink crystals that contain a dinuclear cobalt building unit (Figure 4.2:2a). Both cobalt centres (Co1 and Co2) are octahedral and bind to 6 atoms. Co1 binds to three terminal aqua ligands, two oxygen atoms from carboxylate groups belonging to two different BPM ligands, and a bridging aqua ligand. Co2 bonds to 4 carboxylate oxygen atoms originating from four different BPM ligands and one nitrogen atom from Bpy which coordinates trans to the μ- 2 H2O. Two carboxylate groups bind κ to the dinuclear cobalt unit and 2 carboxylate groups bind κ1 to the dinuclear node. The non-metal-coordinating oxygen atoms from the κ1 carboxylates engage in hydrogen bonding with the bridging aqua ligand. When combined, each metal node consists of two cobalt atoms, one bridging water, three coordinating water, 4,4’-bipyridine, and four pyrazolyl-carboxylate groups (Figure 4.2:a).

107

The dinuclear cobalt node assembles into a coordination network via a structure resembling a ladder. Each dinuclear node is linked via the BPM’s methylene bridge or through Bpy. Looking down the a+b+c direction, diamond shaped units are observed which are formed by two BPM ligands bent at α = 105° and 110° and two metal nodes (Figure 4.2:b). These diamond units propagate along the a direction to make 1D chains (the ladder’s sides). It is important to note that the pyrazolyl nitrogens are non-coordinating as they are bent away from one another (β = 159° and 166°). It is also important to note that the carboxylate bonding mode remains the same for each BPM ligand in the diamond unit. In other words, either both carboxylates bind to the adjacent dinuclear cobalt node in a κ2 fashion, or both will bind κ1 to the node and hydrogen bind to the bridging aqua. 4,4’-bipyridine joins two of these chains together by linking adjacent cobalt nodes (the ladder’s rungs, Figure 4.2:c). However, due to the water coordinating to Co1, no 2D sheets or 3D structures are possible through direct coordination to the dinuclear cobalt nodes. Instead, hydrogen bonding between coordinated water and residual solvent allow for the formation of a hydrogen bonded network with pseudo 3D structure (Figure 4.3:).

108

Observing the hydrogen bonded network down the a direction (Figure 4.3:a), it is clear that the “ladders” pack with each other in a staggered conformation. This staggering is also visible down the b+c direction (Figure 4.3:b). These staggered conformations allow the aqua groups of each dinuclear cobalt node to hydrogen bond to both adjacent nodes and the non-coordinating pyrazolyl nitrogen lone pairs. When viewing the structure down the a+b+c direction it is clear that residual solvent resides within the diamond units of the ladder’s sides while only water occupies space between the 1D chains (Figure 4.3:c). These water molecules serve to further link the “ladders” together. Evidence for this hydrogen bonded network can be observed via TGA (Figure 4.4), which indicates that stepwise loss of the solvent occurs upon heating, as opposed to

109 uncontrolled dehydration. The assigned species are the best match for the observed mass loss based on the known composition of 7 from the crystal structure. A secondary analytical technique (e.g. GCMS) would be required for conclusive identification of the labialized species.

Figure 4.4 TGA analysis of air dried 7 showing stepwise loss of solvent.

4.3.2.2 [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O) (8)

Substitution of DMSO for DMF in the synthesis of 7 results in the formation of 8. 8 contains the same dinuclear cobalt unit as 7 (Figure 4.5:a) and forms a near identical structure. The notable differences are a slight increase in the α-angle of the diamond shaped ladder sides from α = 106.5° and 111.5° to α =105° and 110°(Figure 4.5:b), further bending of the pyrazolyl rings away from one another (Table 4.4), and disordered DMSO as opposed to DMF in the structure. Akin to 7, 8 forms a hydrogen bonded 3D network (Figure 4.5:b) which exhibits stepwise loss of solvent (Figure 4.6:) although due to partial drying not all of solvent water can be identified.

110

Figure 4.6: TGA analysis of air dried 8 showing stepwise loss of solvent. Note that 1.4 molecules of solvent water have already evaporated upon air drying.

111

4.3.2.3 [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9)

In an effort to access a 2D or 3D coordination network, Bpy was replaced with 4,4’-bipyridine- N-oxide (Bpy-N-oxide) for the synthesis of 7. Since cobalt (II) is very oxophilic, we hypothesized that Bpy-N-oxide would be able to compete with the aqua ligands for coordination to Co1. This was not observed. Instead, the Bpy-N-oxide ancillary ligand was unable to compete with solvent for coordination to either Co1 or Co2 resulting in the formation of 9.

In the crystal structure of 9 cobalt centres Co1 and Co2 each bind to 6 atoms and adopt an octahedral geometry. However, the coordination sphere of Co1 has subtlety changed. While Co1 still binds to two oxygen atoms from carboxylate groups belonging to two different BPM ligands and a bridging aqua ligand, the two terminal aqua ligands present in 7 and 8 have been exchanged for two DMF solvent molecules. The new description of the Co1 bonding site is thus

μ-H2O, terminal aqua, two DMF and two carboxylate oxygen. As before, Co2 binds to 4 carboxylate oxygen atoms originating from four different BPM ligands but the coordination site trans to the μ-H2O is occupied by DMF, as opposed to the anticipated Bpy-oxide. The non- metal-coordinating oxygen atoms from the κ1 carboxylate still hydrogen bind to the bridging aqua ligand. Combined, each metal node now consists of two cobalt centres, one bridging and one coordinating water, three DMF solvent molecules, and four pyrazolyl-carboxylate groups (Figure 4.7:a).

112

The dinuclear cobalt node assembles into a coordination polymer via a structure resembling a chain of diamonds when viewed down the b direction. Each node is linked to an adjacent node through the BPM’s methylene bridge. Unlike 7 and 8, there are four unique BPM ligands per asymmetric unit in the crystal structure resulting in α-angles between 119.4° and 122.2° (Figure 4.7:b). The pyrazolyl nitrogens remain non-coordinating (Table 4.4). The diamond units propagate along the a+c direction to make 1D chains (Figure 4.7:c). One of the unique features of 9 is the alternation of carboxylate bonding modes of the BPM ligands: one carboxylate arm will bond κ2 to the dinuclear cobalt node whereas the other will bond κ1 to the node and hydrogen bond to the bridging aqua. As a consequence, adjacent dinuclear cobalt units on the 1D diamond chain will be rotated 180° relative to one another (Figure 4.8:a).

113

Due to the solvent coordinating to both Co1 and Co2, no 2D sheets or 3D structures are possible through direct coordination to the dinuclear cobalt nodes. Instead, hydrogen bonding between the single coordinated aqua of Co1 and a coordinated DMF on an adjacent (along the b direction) Co1 allow for the formation of a hydrogen bonded network with pseudo 2D structure (Figure 4.8:a). The 1D “diamonds” are arrayed as 2D sheets that propagate along the b direction to form a high order hydrogen bonded network (Figure 4.8:b). Viewing the 2D sheets along the b direction reveals that non coordinating DMF resides within the diamond units (Figure 4.8:c). Rapid loss of this DMF is observed via TGA (Figure 4.9:). The lack of other H-donors (such as solvent water for 7 and 8) prevents hydrogen bonding between adjacent sheets.

114

4.3.2.4 [Co(BPM)(H2O)(DMSO)]n (10)

Employing the same strategy of substituting Bpy for Bpy-N-oxide resulted in a radically

different structure, 10. Combining Co(OAc)2•4H2O, H2BPM, and 0.5 eq Bpy-N-oxide in a

DMSO/H2O solution did not result in the expected dinuclear cobalt nodes and resultant coordination polymers. This is likely a result of the interactions between Bpy-N-oxide and cobalt being less favoured than cobalt and DMSO. 10 is instead constructed from a single cobalt atom with distorted octahedral geometry. Coordinating to the cobalt centre are two DMSO, an aqua trans to the DMSO, and two carboxylate groups one of which binds κ2 to the cobalt and the other of which binds κ1 (Figure 4.10:a). The κ1 bonded carboxylate hydrogen bonds with the coordinated aqua ligand. Presumably, steric hindrance by the coordinating DMSO prevents the formation of the dinuclear cobalt nodes observed for 7, 8, and 9. As such, only simple 1D chains, propagating along the c-axis are formed (Figure 4.10:b). These chains have a zig-zag appearance due to an α angle of 118° for the BPM unit. Close packing of the 1D chains and a lack of further hydrogen bonding results in a β-angle of only 134.5°, the lowest observed in this series of cobalt coordination polymers (Table 4.4).

115

There is still limited hydrogen bonding observed in 10 due to the aqua ligand. As mentioned, the aqua ligand forms a hydrogen bond with the κ1 carboxylate. It also forms a hydrogen bond to the κ2 carboxylate of an adjacent 1D Co-BPM chain (Figure 4.10:c). This limited hydrogen bonding results in two 1D chains of 10 pairing with one another through the aqua-carboxylate hydrogen bond. It is important to note that this pairing does not change the dimensionality of the structure. Hydrogen bonding is always between adjacent pairs of 1D chains to create discrete sets. Therefore there is no assembly of a 2D hydrogen bonded network as observed in 9 nor a 3D hydrogen bonded network as observed in 7 and 8. Further evidence for this lack of an extensive hydrogen bonding network is observed via TGA where there is significant early mass loss (Figure 4.11:).

116

4.3.2.5 Similarities of the Dinuclear Cobalt Nodes to other MOFs

Compounds 7, 8, and 9 all assemble coordination networks through dinuclear cobalt nodes and hydrogen bonding. Formations of such nodes have been reported previously in the literature for other cobalt MOFs/coordination polymers.164–169 These prior reports all share certain commonalities with our reported dinuclear cobalt nodes: use of carboxylate ligands and aqueous conditions to produce a dinuclear cobalt node with octahedral cobalt metal centres, a bridging aqua ligand, and carboxylate groups binding κ1 and κ2 to a multinuclear cobalt node.

The first reported MOF possessing dinuclear cobalt nodes was by Zheng et. al. who, upon

reacting 1,3,5-benzenetricarboxylic acid (H3btc) with CoCl2 in DMF, isolated a MOF which partially collapses upon desolvation. This MOF possessed two separate dinuclear cobalt nodes, one with two octahedral cobalt centres and another with one octahedral and one trigonal 164 bipyramidal cobalt centre. Switching to a more complex solvent system, Co(NO3)2∙6H2O, and

using 2,4,6-tris(4-pyridyl)-1,3,5-triazine (tpt) as an ancillary ligand in addition to H3btc yielded a

MOF where H3btc and Co formed small cubohemioctahedra supramolecular building blocks. Whereas in 7 and 8 Bpy coordinates to the Co2 apical site, in this MOF tpt fills that role and is thus able to make a 3D MOF.165 Wu et. al., Lu et. al., and Sumby et. al. were also successful in

117

forming MOFs166,170 and coordination polymers,169 respectively with multinuclear cobalt nodes by employing moderately large, semi-rigid ligands. Lastly, Tasiopoulos et. al. reported a MOF

constructed from Co(NO3)2•6H2O and isonicotinic acid (HINA) which consists of three unique

metal nodes: an octahedral Co(INA)6 and two Co2(μ2-H2O)(INA)x units which have slightly different Co-Co separations across the dinuclear unit.167

The closest analogue to compounds 7, 8, and 9 was reported by Fonari et. al. Assembled from

benzene dicarboxylate (bdc), thionictinamide, and Co(OAc)2•4H2O, this MOF is essentially identical to 9: Co1 binds to one aqua ligand, one DMF solvent molecule, thionictinamide, two oxygen atoms from carboxylate groups belonging to two different bdc ligands, and a bridging aqua ligand. Co2 binds to 4 carboxylate oxygen atoms originating from four different bdc

ligands and the coordination site trans to the μ-H2O is occupied by thionictinamide. The non- metal-coordinating oxygen atoms from the κ1 carboxylate hydrogen bond to the bridging aqua ligand. Combined, each metal node consists of two cobalt centres, one bridging and one coordinating water, one DMF solvent molecule, two thionictinamide, and four benzene- carboxylate groups (Figure 4.12:a).168 The nodes allow for the formation of 2D sheets. Viewing these sheets down the a+b direction (Figure 4.12:b) reveals a diamond like pattern similar to that observed in 9 (Figure 4.7:b). By using the small, rigid bdc ligand, the metal nodes (shown in green) are able to assemble into an extended framework. In contrast, for 9, the methylene groups of the BPM ligands prevent such coordination. As shown in Figure 4.13: the methylene groups of 9 (Figure 4.13:a) block direct coordination of the 1D chains. Without the methylene groups Fonari’s MOF is able to coordinate easily to the metal nodes (Figure 4.13:b). As a consequence, 9 is unable to form an extended 2D structure through direct metal-ligand coordination unlike the MOF reported by Fonari et. al.

118

Overall, 7, 8, and 9 share a similar structural node to these literature compounds with Co-OH2- Co angles and Co-Co distances being similar to those reported in the literature (Table 4.1). As such, the dinuclear cobalt building unit may be classified as a secondary building unit. The extensive use of tripodal ligands164–166 and pyridyl groups165,167 in assembling reported MOFs with this building unit indicates functionalization of BPM’s methylene linker with either a carboxylate or pyridyl group may allow for compounds analogous to 7, 8, and 9 to be synthesized that are true MOFs. Furthermore, the identification of the dinuclear cobalt node as

119

the seventh iteration of an underexplored secondary building unit highlights the versatility and structural control possible with Co-BPM structure. This should allow for general predictions regarding the geometries and expected structures for future Co-BPM compounds. For instance, replacement of Bpy with tpt may add further dimensionality to 7 and 8 allowing for the potential construction of 1D tripodal columns.

2 Table 4.1: Metal-metal distances and μ -OH2 bonding angles for dinuclear cobalt building units of 7, 8 and 9 in comparison to analogous literature compounds.

Compound Co-Co distance (Å) Co-OH2-Co angle (°)

7 3.5753 (5) 114.66 (6)

8 3.594 (1) 114.9 (2)

9 3.5275 (6) 114.6 (1)

Ref 166 3.577 (1), 3.6017 (9) 112.1 (1), 116.4 (1)

Ref 167 3.579 (5) 113.30 (7)

4.3.3 BPM MOFs of Manganese, Iron, Copper, Zinc, and Europium

4.3.3.1 [Mn(BPM)(H2O)2]n (11)

Manganese acetate, H2BPM, and Bpy reacted to yield a new MOF, 11. Unlike the cobalt coordination polymers (7-10), 11 is a true MOF with a 3D structure. The building unit consists of a dinuclear node formed from two octahedral manganese metal centres that are coordinated to three carboxylate and four aqua ligands. Each carboxylate bonds κ2 to the unit and originates from a different BPM ligand (Figure 4.14:a,b). The dinuclear nodes are linked together to form 1D chains of manganese along the a direction via a bridging BPM-carboxylate (Figure 4.14:c). This bridging carboxylate also serves to charge balance the dinuclear node. In order to facilitate this arrangement BPM is bent at a near right angle and twisted with a β-angle of 177° or 167° (Table 4.4). A staggered confirmation is observed when viewed from the side (Figure 4.14:d). As a consequence, each dinuclear manganese node is vertically offset from its neighbours. This arrangement prevents the formation of any voids due to hydrogen bonding between the pyrazole

120

nitrogen and the coordinated aqua groups. 11 closely packs around itself forming 1D manganese chains resulting in a non-porous 3D MOF (Figure 4.14:e).

Although Bpy is not incorporated into the structure in any way, 11 will not form in its absence.

Therefore, Bpy is likely acting as a weak base to deprotonate the H2BPM ligand.

121

4.3.3.2 [Cu(BPM)(DMSO)]n (12)

Copper acetate and H2BPM react to form a traditional Cu2(RCOO)4 paddlewheel building unit (i.e. the same unit as HKUST-17). The dinuclear copper paddlewheel of 12 is composed of two square pyramidal copper centres coordinating to four BPM-carboxylate groups. The apical sites of the copper centres are occupied by coordinated DMSO (Figure 4.15:a). While the BPM ligand is bent far more than observed in 7-11 (132.47°), the pyrazolyl rings are twisted significantly less: only 119.8(3)° (Table 4.4). As a consequence, 6 is able to form well-ordered 2D sheets with well-defined rectangular cavities (Figure 4.15:b). However, 12 is fully interpenetrated with 2D sheets growing orthogonally through their respective cavities (Figure 4.15:c). Due to the paddlewheel unit residing within the rectangular cavities of the interpenetrating sheets, 12 is not a porous material (Figure 4.15:d).

122

4.3.3.3 [Eu(BPM)2(OAc)]n (13)

Europium (III) acetate and H2BPM combined in a heated DMF/H2O solution to form another new MOF, 13. This demonstrated that BPM can effectively bind to lanthanides in addition to transition metals. The building unit for 13 is rather complex as there are no discrete units. Instead, the core of the MOF is formed by a 1D chain that propagates down the b direction. This chain contains distorted pentagonal bipyramidal europium centres that are joined by 2 bridging BPM-carboxylates and an acetate that bonds κ2 to Eu1 and also bridges between Eu1 and Eu2 (Figure 4.16a). Each Europium centre is seven coordinate (Figure 4.16b). These 1D chains are then linked together via the methylene groups of the BPM ligand (Figure 4.16c). The BPM is bent at 120.1° and twisted (Table 4.4) to create an over-under weave of the europium chains. This allows for the formation of 2D sheets (Figure 4.16d) as opposed to the close packed, 3D MOF 11 which has a far more acute angle and greater torsion of the BPM ligand. The 2D sheets form in the bc plane and stack along the a direction. Careful observation of the 2D sheets reveals that the europium centres are completely occluded from further substrate coordination by the BPM and acetate methyl groups although the pyrazolyl lone pairs remain available.

123

Linkages between 1D europium metal chains via BPM on both the top (red) and bottom (green) of 2D sheet c) resulting 2D sheet on the bc plane.

Another interesting property of 13 is its high thermal stability (Figure 4.17). When heated under

N2, 13 showed no mass change until ~400 °C. In comparison, most europium coordination compounds decompose between 200 – 300 °C.171–173 This is another excellent indicator that 13 has the potential to be a very robust luminescent sensing material provided a mechanism of luminescence enhancement/quenching can be incorporated into the MOF.

100%

80%

60%

40% MassPercent

20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 4.17 TGA spectrum of [Eu(BPM)2(OAc)]n MOF (13).

4.3.3.4 Attempts to Synthesize Zinc and [Fe2(BPM)2(H2O)4(Bpy)0.5]n MOFs

Attempts were also made to react H2BPM with both zinc and iron with mixed success. Although reactions with zinc always provided crystalline materials, it was not possible to isolate pure species for analysis. Combining Fe(OAc)2, H2BPM, and Bpy in DMF/H2O did result in a coordination polymer identical to 7, only the dinuclear node is comprised of two iron metal centres, not two cobalt metal centres (Figure 4.18:a). The [Fe2(BPM)2(H2O)4(Bpy)0.5]n material again formed a hydrogen bonded 3D network (Figure 4.18:b). Unfortunately, the material proved to be both difficult to isolate and very air sensitive. Synthesis had to be conducted under air free conditions and always resulted in a mixture of [Fe2(BPM)2(H2O)4(Bpy)0.5]n and unidentified amorphous materials. This may be a consequence of disproportionation of the FeII starting

124

material during synthesis. Mechanical separations of the crystals were not possible as any exposure to air or removal from the mother liquor resulted in rapid degradation of the material: rich purple crystals immediately turned brown. Attempts to analyze the brown material showed it to be amorphous. While the [Fe2(BPM)2(H2O)4(Bpy)0.5]n material remains interesting, a more rigorous synthetic approach must be devised in order to produce enough polymer for full characterization.

4.3.4 Investigated Properties

With compounds 7-13, in hand preliminary experiments were undertaken to determine the reactive properties of the materials. Similar cobalt MOFs in the literature with related building groups have reported Co-MOFs as being active for olefin oxidations.174,175 7 was chosen to evaluate the catalytic potential of the CoBPM coordination polymers as it was the easiest to synthesize in bulk. Taking the experimental conditions of Khang et. al. as a baseline174 multiple conditions were attempted using 7 as a catalyst for cyclooctene oxidation using tertbutylhydrogen peroxide as the oxidant. While catalysis was observed, there was not a significant

125

improvement over the homogeneous catalyst Co(OAc)2. Furthermore, we were unable to replicate the control experiments described in the paper.174 It is possible that under experimental conditions the hydrogen bonded network of 7 collapses and the cobalt redissolves into the reaction solution. As such, future evaluation of 7-10 for catalytic performance must demonstrate stability of the coordination network in the reaction mixture first.

Manganese and copper were not evaluated for any chemical reactivity as the frameworks were non-porous. However, the europium MOF 13 was successfully demonstrated as a luminescent material. The MOF has two strong excitation maxima with the global maximum occurring at 393 nm. A typical emission spectrum for europium was observed with a maximum at 615 nm (Figure 4.19:). 13 was also evaluated as a luminescent sensor for VOCs via interactions with the BPM pyrazolyl nitrogen lone pairs since, as noted in Figure 4.16c, the europium centres are completely occluded. Testing with VOCs did not have a statistically significant impact on the observed luminescent intensity of 13. This is likely a consequence of BPM serving as a poor “antenna” for the europium metal centre.176,177 Lack of ligand sensitization has been observed in other Europium MOFs.176 For 13 to function as a VOC sensor it would first be necessary to improve the LMCT process through the modification of BPM.

126

1200

1000

800

600 Excitation Emission 400

Relative Intensity (a.u.) Intensity Relative 200

0 375 450 525 600 675 750 Wavelength (nm)

Figure 4.19: Excitation and emission spectra for 13 monitored at 615 nm and irradiated at 393 nm respectively showing predominantly europium-centred excitations.

4.3.5 Conclusions

In order to better understand the utility of BPM towards the formation of coordination polymers and MOFs, seven new compounds were synthesized, characterized, and studied. Co-BPM demonstrated that a wide variety of structures were possible (7-10) using simple synthetic controls. Furthermore, although [Mn(BPM)(H2O)2]n (11) and [Cu(BPM)(DMSO)]n (12) were non porous, both demonstrated how BPM can behave as a typical di-carboxylate linker to form typical MOF structures. Lastly, [Eu(BPM)2(OAc)]n (13) demonstrated that BPM is not limited to transition metal nodes and can form lanthanide complexes. These lanthanide complexes are also very thermally stable indicating the 13 may hold promise as a luminescent sensing material.

127

4.4 Experimental 4.4.1 General Considerations

The ligand synthesis is the same as that outlined in chapter 3.4. Reagents were used as received. All reactions were performed under air unless otherwise noted. Instruments used for sample analysis as per chapter 3.4.

4.4.2 Synthesis of MOFs

4.4.2.1 Synthesis of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (7)

A 18 mL aqueous solution 76.5 mg (0.307 mmol) of Co(OAc)2•4H2O was added to a 18 mL

DMSO solution containing 90 mg (0.308 mmol) H2BPM and 24.5 mg (0.157 mmol) of 4,4’- bipyridine in round bottom flask. The pink solution was stirred at 90 °C for two days to afford pink crystals. The crystals were collected on a Hirsch funnel and washed with DMF/water. 67 mg (42 % yield) of crystals collected. A second crop of crystals may be recovered by heating the filtrate for a further week (6.9 mg, total yield 47 %). Anal. Calc. for C37H58N11O16Co2: C 43.11, H 5.67, N 14.95; found C 43.37, H 5.42, N 14.93.

4.4.2.2 Synthesis of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O) (8)

A 2 mL aqueous solution 8.5 mg (0.034 mmol) of Co(OAc)2•4H2O was added to a 2 mL DMSO

solution containing 10 mg (0.034 mmol) H2BPM and 3.3 mg (0.021 mmol) of 4,4’-bipyridine in a 2 dram vial. The sealed vial was heated at 110 °C for four days to afford pink crystals. The crystals were collected on a Hirsch funnel and washed with DMSO. 14 mg (68 % yield) of crystals collected. The crystals rapidly lost coordinated water when left in air. Anal. Calc. for

C31H40N9O12Co2-2(C2H6SO)0.6(H2O): C 41.39, H 5.28, N 12.41; found C 41.52, H 5.21, N 12.27.

4.4.2.3 Synthesis of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9)

A 2 mL aqueous solution containing 8.5 mg (0.034 mmol) of Co(OAc)2•4H2O was added to a 2

mL DMF solution containing 10 mg (0.034 mmol) H2BPM and 3.2 mg (0.017 mmol) of 4,4’- bipyridine-N-oxide in a 2 dram vial. The sealed vial was heated at 90 °C overnight to afford a clear, pink solution with some amorphous purple solids. The vial was left to sit for 18 months during which purple crystals formed from a clear greenish solution. The crystals were collected

128

on a Hirsch funnel and washed with DMF. 14 mg (80 % yield) of crystals collected. Anal. Calc.

for C76H120N24O28Co4: C 44.45, H 5.89, N 16.37; found C 44.43, H 5.94, N 16.13.

4.4.2.4 Synthesis of [Co(BPM)(H2O)(DMSO)]n (10)

A 2 mL aqueous solution containing 8.5 mg (0.034 mmol) of Co(OAc)2•4H2O was added to a 2

mL DMSO solution containing 10 mg (0.034 mmol) H2BPM and 3.2 mg (0.017 mmol) of 4,4’- bipyridine-N-oxide in a 2 dram vial. The sealed vial was heated at 90 °C overnight to afford a clear, pink solution. The vial was left to sit for 18 months during which purple crystals formed in a clear jelly-like substrate. The crystals were collected on a Hirsch funnel and the jelly was removed via washing with DMSO. 13 mg (71 % yield) of crystals collected. Anal. Calc. for

C17H28N4O7S2Co: C 39.01, H 5.39, N 10.70; found C 39.26, H 5.28, N 10.43.

4.4.2.5 Synthesis of [Mn(BPM)(H2O)2]n MOF (11)

Solutions of 30 mg (0.10 mmol) H2BPM and 8.7 mg (0.056 mmol) of 4,4’-bipyridine in 6 mL of

DMSO and 12.9 mg (0.0746 mmol) of Mn(OAc)2 in 6 mL of distilled water were prepared. The solutions were equally proportioned into 3 x 2 dram vials, sealed, and left to react at 90 °C for 5 days. After cooling the white solids were collected on a Hirsch funnel, washed three times with DMSO and once with water. 18 mg (61 % yield) of solids were collected. Anal. Calc. for

C26H36N8O12Mn2: C 40.96, H 4.76, N 14.70; found C 41.15, H 4.68, N 14.26.

4.4.2.6 Synthesis of [Cu(BPM)(DMSO)]n MOF (12)

A 2 mL aqueous solution containing 6.2 mg (0.034 mmol) of Cu(OAc)2•xH2O (assumed

Cu(OAc)2) was added to a 2 mL DMF solution containing 10 mg (0.034 mmol) H2BPM in a 2 dram vial. The sealed vial was heated at 110 °C for four days to afford rod shaped blue crystals. The crystals were collected on a Hirsch funnel and washed with DMSO. 10 mg (68 % yield) of

crystals collected. Anal. Calc. for C15H20N4O5CuS-0.25(C2H6SO): C 41.24, H 4.80, N 12.41; found C 41.42, H 4.59, N 12.27.

4.4.2.7 Synthesis of [Eu(BPM)2(OAc)]n (13)

Solutions of 88.6 mg (0.303 mmol) H2BPM in 18 mL of DMF and 79.2 mg (0.241 mmol) of

Eu(OAc)3•xH2O (assumed Eu(OAc)3) in 18 mL of distilled water were prepared. The solutions were equally proportioned into 9 x 20 mL scintillation vials and further diluted with 3 mL of

129

DMF and 3 mL of water per vial. The sealed vials were left to react at 90 °C for 2 days followed by 2 days at 110 °C. After cooling the white solids were collected on a Hirsch funnel, washed three times with DMSO and once with water. 78 mg (51 % yield) of solids were collected. X-ray quality crystals may be grown by heating for a longer duration (~1 week) at a lower temperature

(70 °C). Anal. Calc. for C15H17N4O6Eu: C 35.94, H 3.42, N 11.18; found C 36.32, H 3.44, N 11.27.

4.4.3 Powder X-Ray and TGA Characterization

4.4.3.1 TGA of compounds 11 and 12

100%

80%

60%

40% Mass Percent Mass 20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 4.20: TGA spectrum of [Mn(BPM)(H2O)2]n MOF (11). Note slow loss of coordinated water between 100 and 200 °C.

130

100%

80%

60%

40% MassPercent

20%

0% 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 4.21: TGA spectrum of [Cu(BPM)(DMSO)]n MOF (12). The absence of mass loss prior to decomposition indicates the non-porous, interpenetrated nature of the MOF.

4.4.3.2 Powder X-Ray Diffraction of 7-13

500 450 400 350 300 250 Observed 200 Predicted 150

(a.u.) Intensity Relative 100 50 0 5 10 15 20 25 30 35 2 Theta (°)

Figure 4.22: PXRD pattern of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (7).

131

500 450 400 350 300 250 Observed 200 Predicted 150

(a.u.) Intensity Relative 100 50 0 5 10 15 20 25 30 35 40 2 Theta (°)

Figure 4.23: PXRD pattern of [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O) (8).

300

250

200

150 Observed

100 Predicted (a.u.) Intensity Relative 50

0 5 10 15 20 25 30 35 40 2 Theta (°)

Figure 4.24: PXRD pattern of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9).

132

1800 1600 1400 1200 1000 800 Observed 600 Predicted

(a.u.) Intensity Relative 400 200 0 5 10 15 20 25 30 35 40 2 Theta (°)

Figure 4.25: PXRD pattern of [Co(BPM)(H2O)(DMSO)]n (10).

1000

800

600

Observed 400 Predicted

(a.u.) Intensity Relative 200

0 5 15 25 35 45 2 Theta (°)

Figure 4.26: PXRD pattern of [Mn(BPM)(H2O)2]n (11).

133

250

200

150

Observed 100 Predicted

(a.u.) Intensity Relative 50

0 5 15 25 35 2 Theta (°)

Figure 4.27: PXRD pattern of [Cu(BPM)(DMSO)]n (12).

600

500

400

300 Observed

200 Predicted (a.u.) Intensity Relative 100

0 5 10 15 20 25 30 35 40 2 Theta (°)

Figure 4.28: PXRD pattern of [Eu(BPM)2(OAc)]n (13).

4.4.4 X-Ray Diffractometry

X-ray quality single crystals of 7-13 were obtained as described in the synthesis above. The crystals were mounted on the tip of a MiTeGen MicroMount. The single-crystal X-ray

diffraction data were collected on a Bruker Kappa Apex II CCD diffractometer with Mo Kα radiation (λ = 0.71073 Å) operating at 50 kV and 30 mA, at 150 K (unless otherwise noted)

134 controlled by an Oxford Cryostream 700 series low temperature system. The data integration and absorption correction were performed with the Bruker Apex 2 software package.103 The structure was solved by direct methods and refined using SHELXTL V2016/4.104,105 All non-hydrogen atoms except for the atoms involved in the disordered portions were refined anisotropically. The positions of the hydrogen atoms were calculated using the riding model.

Note that for [Co2(BPM)2(H2O)4(Bpy)0.5]n∙2DMSO∙2(H2O) (8) the data quality was low as a consequence of the crystals being thin.

Table 4.2: Crystallographic Properties of Cobalt Coordination Polymers 7-10.

Name (7) (8) (9) (10)

C31H40N9O12Co2- C70H106N22O26Co4- Formula C37H58N11O16Co2 C17H28N4O7S2Co 2(C2H6SO)0.6(H2O) 2(C3H7NO)

Formula Weight 1030.80 1040.86 2053.67 523.48 [g/mol]

Crystal Triclinic Triclinic Triclinic Triclinic System

Space Group P-1 (No. 2) P-1 (No. 2) P-1 (No. 2) P-1 (No. 2)

a [Å] 12.0921(12) 12.1781 (17) 11.0570(5) 8.0447(11)

b [Å] 14.3464(14) 13.8645 (18) 19.9279(8) 11.5152(15)

c [Å] 15.3102(16) 14.483 (2) 21.6288(9) 13.2015(19)

α [°] 77.840(6) 81.911 (7) 94.714(2) 73.081(6)

β [°] 70.808(6) 76.867 (7) 101.202(2) 85.823(7)

γ [°] 67.955(5) 73.032 (7) 99.387(2) 80.360(7)

Volume [Å3] 2313.4(4) 2270.5 (5) 4580.3(3) 1153.1(3)

Z 2 2 2 2

D(calc) 1.480 1.523 1.489 1.508 [g/cm3]

135

μ(MoK ) α 0.797 0.900 0.802 0.970 [ /mm ]

F(000) 1078 1086 2152 546

Crystal Size 0.03 x 0.04 x 0.02 x 0.02 x (estimated) 0.01 x 0.03 x 0.06 0.04 x 0.04 x 0.16 0.05 0.02 [mm]

Temperature 160 149 150 150 (K)

Radiation [Å] 0.71073 0.71073 0.71073 0.71073

θ Min-Max 1.9, 27.6 1.8, 27.5 1.9, 27.6 2.1, 27.6 [°]

-15: 15; -18: 18; -15: 15; -17: 17; -14: 14; -25: 25; -10: 10; -14: 14; Dataset -16: 19 -18: 17 -28:26 -17: 17

Tot., Uniq. 35793, 10387, 31619, 10184, 75772, 20467, 23521, 5285, Data, R(int) 0.032 0.120 0.073 0.104

Observed data [I > 2.0 8316 5259 11615 3081 σ(I)]

Nref, Npar 10387, 607 10184, 590 20467, 1221 5285, 296

R1, wR2, 0.0344, 0.0915, 0.1075, 0.2757, 0.0558, 0.1230, 0.0671, 0.1515, GooF 1.02 1.03 1.01 1.03

Max. and Av. 0.00, 0.00 0.00, 0.00 0.00, 0.00 0.00, 0.00 Shift/Error

Min. and Max. Resd. -0.34, 0.42 -1.19, 1.12 -0.70, 0.56 -0.60, 0.686 Dens. [e/Å3]

136

Table 4.3: Crystallographic properties of other BPM MOFs 11-13 and

[Fe2(BPM)2(H2O)4(Bpy)0.5]n∙2DMF∙2(H2O) (FeBPM Ladder).

Name (11) (12) (13) FeBPM Ladder

Formula C26H36N8O12Mn2 C15H20N4O5CuS C15H17N4O6Eu C37H58N11O16Fe2

Formula Weight 762.51 431.95 501.28 1024.64 [g/mol]

Crystal System Monoclinic Orthorhombic Orthorhombic Triclinic

Space Group P21/c (No. 14) Pbcn (No. 60) Pbcm (No. 57) P-1 (No. 2)

a [Å] 9.1531(9) 17.7266(7) 8.6328(16) 12.116 (5)

b [Å] 16.534(2) 11.6874(5) 8.3314(15) 14.476 (6)

c [Å] 21.537(3) 18.1453(7) 25.778(6) 15.380 (6)

α [°] 90 90 90 77.520 (13)

β [°] 101.922(6) 90 90 71.034 (10)

γ [°] 90 90 90 67.943 (10)

Volume [Å3] 3189.1(7) 3759.3(3) 1854.0(6) 2350.7 (17)

Z 4 8 4 2

D(calc) [g/cm3] 1.588 1.526 1.796 1.448

μ(MoKα) [ /mm ] 0.866 1.306 3.421 0.696

F(000) 1576 1784 984 1074

Crystal Size 0.01 x 0.02 x 0.01 x 0.01 x 0.02 x 0.02 x 0.005 x 0.02 x (estimated) [mm] 0.05 0.03 0.18 0.03

Temperature (K) 150 150 150 150

Radiation [Å] 0.71073 0.71073 0.71073 0.71073

θ Min-Max [°] 1.6, 27.5 2.1, 27.5 2.4, 27.5 1.9, 25.2

-11: 11; -21: 21; -18: 23; -15: 15; -11: 11; -10: 9; - -14: 14; -17: 16; Dataset -27: 27 -23: 23 32: 33 -18, 18

137

Tot., Uniq. Data, 51991, 7307, 57894, 4312, 10204, 2170, 28014, 8184, R(int) 0.094 0.064 0.090 0.200

Observed data 4590 3338 1547 3422 [I > 2.0 σ(I)]

Nref, Npar 7307, 441 4312, 244 2170, 129 8184, 607

0.0430, 0.1051, 0.0431, 0.1171, 0.0579, 0.1517, 0.0840, 0.2081, R , wR , GooF 1 2 1.01 1.07 0.99 1.00

Max. and Av. 0.00, 0.00 0.00, 0.00 0.00, 0.00 0.00, 0.00 Shift/Error

Min. and Max. Resd. Dens. -0.47, 0.44 -0.61, 1.55 -2.51, 3.86 -0.54, 0.61 [e/Å3]

Table 4.4 Selected properties of the BPM ligand for compounds 7 through 13.

α Angle (°) β Angle (°)

7 104.97(5), 110.21(5) 165.8(2), 159.2(2)

8 111.6(2), 106.5(2) 161.2(6), 172.0(6)

9 120.40(9), 119.45(9), 121.21(9), 122.21(9) 149.8(3), 147.3(3), 149.6(3), 146.3(3)

10 118.3(1) 134.5(4)

11 97.24(7), 100.13(7) 177.3(3), 167.4(3)

12 132.47(9) 119.8(3)

13 120.1 131.7(6)

138

Chapter 5 5 Towards Catalytic MOFs Utilizing Tris- and Tetrapyrazolyl Ligands 5.1 Abstract

A series of new ligands, triethyl-1,1',1''-methanetriyltris(3,5-dimethyl-1H-pyrazole-4- carboxylate) (TPM-1), 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4- carboxylic acid) (H4TPP), and tetraethyl-1,1',1'',1'''-(1,4-phenylenebis(methanetriyl))tetrakis(3,5- dimethyl-1H-pyrazole-4-carboxylate) (TPX), were synthesized. These ligands were, in turn, coordinated to Mo(CO)3 (TPM-1) and PdCl2 (H4TPP and TPX) to investigate their ability to act as MOF metalloligands. A second variant of TPM-1, containing a phenyl spacer was also synthesized: triethyl-4,4',4''-(methanetriyltris(3,5-dimethyl-1H-pyrazole-1,4-diyl))tribenzoate

(TPM-2), in an effort to produce an isoreticular MOF series. Evaluation of Mo(TPM-1)(CO)3

and Mo(TPM-2)(CO)3 indicated that they were effective pre-catalysts for the oxidation of cyclooctene to cyclooctene oxide. However, only one MOF was realized, synthesized from the

acid form of Mo(TPM-1)(CO)3 and Zn(NO3)2 in very low yield and purity due to decomposition

of the Mo(TPM-1)(CO)3 coordination compound. As such, while these ligands and metal complexes hold great promise, we have been unable to synthesize and fully characterize any MOFs thus far.

5.2 Introduction 5.2.1 Known Trispyrazolylmethane and Tetrapyrazolyl Compounds

As discussed in Chapter 1, a common goal for MOFs is to add catalytic functionality to the frameworks in order to access difficult to synthesize chemicals. The best proven method for accomplishing this is the heterogenization of homogeneous catalysts. In this manner, the benefits of a homogenous catalyst, namely well-defined active sites and understood mechanisms, may be combined with the robustness of heterogeneous catalysts. Although many examples of catalytic MOFs are known, to our knowledge no one has successfully synthesized a MOF containing a trispyrazolemethane active site.

139

Use of bis- and trispyrazolyl compounds as ligands was first described by Trofimenko in the 1970’s who synthesized a whole family of pyrazolyl borate and alkyl ligands for metal coordination.67 Due to their well-known and extensive chemistry, trispyrazolylmethane (TPM) in particular makes for an attractive target for a MOF metalloligand. TPM ligands, also called “scorpionates” are tridentate ligands that bind to a metal centre in a fac manner via the pyrazolyl lone pairs.178,179 These ligands offer many possible tunable sites: the backbone may be charged or neutral, binding site sterics are easy to alter, choice of functional group allows for coordination to different metal nodes, and coordination to different metals imparts different catalytic activities (Figure 5.1:). Rather than synthesize every possible ligand derivative, we chose various restraints in order to simplify our ligand design. Although trispyrazolylborate ligands (TP) were developed first,67 we elected to pursue TPM. Unlike TP, TPM is charge neutral allowing for silica gel column purification and reaction monitoring via NMR during ligand design and optimization. Specifically, tris(3,5-dimethylpyrazolyl)methane was chosen due to relative ease of synthesis and its well documented coordination chemistry.179 Lastly, carboxylate groups were targeted to functionalize the TPM core. This allows the ligand to coordinate to ancillary metals and form a MOF.

Figure 5.1: Possible sites for tuning/derivatization of the TPM ligand; R = different functional groups for metal node coordination.

140

Of particular interest to us was the use of trispyrazolylmethane molybdenum tricarbonyl as a pre- 180 catalyst for olefin oxidation. The Mo(TPM)(CO)3 pre-catalyst is a neutral, stable species that is relatively easy to synthesize from TPM and molybdenum hexacarbonyl.181 Under the catalytic conditions the oxidant of choice (i.e. tert-butylhydrogen peroxide) oxidizes Mo0 to MoIV or MoVI oxo. The olefin may then coordinate to the reactive oxo groups in and outer sphere reaction in order to produce an epoxide.182

Given our successes with MOFs and coordination polymers synthesized from H2BPM we also set out to investigate tetrapyrazolyl compounds for MOF synthesis. Although some are known such as Sumby et. al.’s hinged ligand for gas sorption,183 other recent examples are scarce. Tetrapyrazolyl complexes in general offer a unique opportunity for synthesis of large pore MOFs with chelation sites that may be metallated. Two potential targets are tetraethyl-1,1',1'',1'''-(1,4- phenylenebis(methanetriyl))tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylate) (TPX)184–187 and 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylic acid) 188–191 (H4TPP). Most importantly, these materials have successfully been used for the synthesis of supramolecular structures184,185 and coordination polymers.186,188,189 Homogenous examples also proved to be effective catalysts when coordinated to palladium dichloride.187 By further functionalizing these materials with carboxylate groups we hope to expand this chemistry to encompass true 3D MOFs as stable platforms for catalysis (Figure 5.2:).

Figure 5.2: Similarities and differences of H4TPP and TPX ligands with desired locations for metal coordination sites indicated.

141

5.3 Results and Discussion 5.3.1 Functionalized Molybdenum Trispyrazolylmethane Tricarbonyl Complexes

5.3.1.1 Isostructural Trispyrazolylmethane Ligand Series

O O

O O N O NaHCO 3 N N N [n-Bu4N]Br O water : CHCl3 (2:1) N N NH reflux 72 hours N

O TPM-1

O O O O Br

NaHCO 2.5 mol% Pd(OAc)2 3 5 mol% Xantphos [n-Bu4N]Br

10 atm CO water : CHCl3 (2:1) EtOH/NEt3 reflux 72 hours

HN N HN N N N

3 CH A TPM-2

Scheme 5.1: Synthesis of TPM-1 and TPM-2.

In an effort to successfully synthesize a MOF containing Mo(TPM)(CO)3 an isoreticular series of trispyrazolyl ligands was investigated. Isoreticular series are common in MOF chemistry. The term originates from “iso” meaning the same and “reticular” meaning “of, relating to, or forming a network”. Typically, this means the MOFs’ secondary building units remain the same but the framework linkers are enlarged by adding phenyl groups. The classic example of this was the expansion of MOF-5 by Yaghi et. al. to make progressively larger MOFs with more intricate pores.192

As shown in Scheme 5.1, two derivatives TPM-1 and TPM-2, were synthesized from the literature compound 3,5-dimethylpyrazole-4-carboxylic acid ethyl ester143,144 in one step and from the hydrazine treated literature compound 4-bromo-phenyl-2,4-pentanedione193 in two steps. Subsequent substitution reaction in chloroform yields the desired ligand precursors.194

Further treatment of the ligands with Mo(CO)6 in refluxing mesitylene or 1,4-dioxane yields the desired molybdenum coordination complexes (Scheme 5.2).181

142

O O H N O O O N O N N N N N mesitylene N + Mo(CO)6 N O N O N 140 °C O Mo O 1.5 hours N OCOC CO O

O TPM-1 14

O O

O N dioxane N N R + Mo(CO)6 N R 110 °C N O N 1 hour Mo OC CO N N OC R 3 CH TPM-2 15

Scheme 5.2: Synthesis of metalloligands 14 and 15.

Although recrystallization of 15 was unsuccessful, we were able to isolate crystals of 14. The crystal structure clearly shows the anticipated fac coordination of 14 and the arrangement of the carboxylate groups in a triangular fashion desired for successful MOF formation (Figure 5.3:).

Figure 5.3: a) Mo(TPM-1)(CO)3 complex (14) side view and b) top down view showing triangular arrangement of the disordered ester groups.

143

5.3.1.2 Performance of Molybdenum Trispyrazolylmethane Tricarbonyl for the Catalytic Oxidation of Olefins

Following the protocols of Gonçalves et. al.180 14 and 15 were evaluated for the catalytic oxidation of cyclooctene. 14 was found to quickly and cleanly synthesize cyclooctene oxide with complete conversion in 1 hour under mild conditions (Figure 5.4:). Monitoring the reaction via gas chromatography/mass spectrometry revealed the major product (>90%) was cyclooctene oxide with trace amounts of cyclooctanone. 15 was also found to be catalytically active however the reaction products were not identifiable indicating possible over oxidation of substrate. Furthermore, although 15 had a much higher initial reaction rate (80% substrate consumption within 5 min) the reaction did not go to completion with 10% substrate remaining after one hour indicating possible deactivation or decomposition of 15 (Figure 5.5:).

144

100%

75%

50% 14 15

PercentConversion 25%

0% 0 15 30 45 60 Time (min)

Figure 5.5: Catalytic performance of 14 versus 15.

5.3.1.3 MOF Synthesis Attempts and MOF Properties

O O O OH O O HO O

N N NN N N N N .6 CO NaOH CO Zn(NO3)2 H2O H Mo H Mo MOF Crystals! CO THF:water CO DMF:H2O ° N N CO 70C N N CO 85 C 3 days 5 Days - EtOH

O O O OH 14 14-(COOH)3

Scheme 5.3: Synthesis of 14-(COOH)3 from 14 and subsequent MOF synthesis.

To use 14 or 15 as metalloligands for MOF synthesis, it is necessary to convert the ester groups to carboxylates. Repeated ester hydrolysis attempts were unsuccessful for 15. The compound was too sensitive and decomposed upon reaction with base. For 14 it was possible to synthesize a tricarboxylic acid form of the ligand, 14-(COOH)3 (Scheme 5.3). However, NMR analysis of the isolated compound always showed >10% free ligand. The likely reason for this is TPM ligands are fluxional by nature. It is not uncommon for one or more pyrazolyl arms to dissociate from the metal centre in solution. Once dissociated, the basic pyrazolyl nitrogen is susceptible to

145

protonation resulting in loss of molybdenum from the complex. Furthermore, adventitious oxygen may oxidize the complex resulting in a mixture of species, further degrading the yield and purity of the isolated product. The isolated yellow solids are not air stable. Decomposition of - 14-(COO )3 resulting in free ligand can be observed via the increasing ligand to complex ratio as measured by 1H NMR spectroscopy.

Further reacting 14-(COOH)3 with zinc nitrate yielded MOF crystals. However, the MOF synthesis proved to be very sensitive to temperature, concentration, and solvent composition. Furthermore, the yield was too low to allow for effective characterization of the MOF. A typical 10 mg reaction would yield, at most, half a dozen crystals that had to be mechanically separated from white amorphous solids. These solids are likely a mixture of free ligand and oxidized molybdenum due, again, to adventitious oxygen.

The resulting MOF consists of a dinuclear node formed from two tetrahedral zinc metal centres that are joined by two bridging pyrazolyl-carboxylate groups. Each zinc centre further coordinated to two pyrazolyl-carboxylates bonding in a κ1 fashion (Figure 5.6:a). In order to charge balance the dinuclear node either two carboxylate groups remain protonated or a disordered counterion resides in the pores. It is important to note that the molybdenum tricarbonyl unit remains intact and, as predicted, three zinc bind to the pyrazolyl-carboxylate groups (Figure 5.6:b).

146

Figure 5.6: a) Dinuclear building unit of MOF synthesized from Mo(TPM-1)(CO)3 in two steps;

ligand truncated for clarity. b) MOF asymmetric unit showing intact Mo(TPM-1)(CO)3 and coordinated zinc. c) Model of the MOF showing pores propagating down the b direction and d) space filling model of MOF showing small pores down b direction. Solvent molecules omitted for clarity.

The tripodal units of the MOF constructed from Mo(TPM-1)(CO)3 are closely packed due to the short distance between carboxylate groups. As a consequence, the only visible channels in the

MOF run along the b direction restricting access to larger pores lined with Mo(CO)3 groups (Figure 5.6:c). Examining a space filling model of the MOF reveals that the pore openings are restricted to less than 1 nm once the van der Waals radius of hydrogen is accounted for. Therefore, while the material has the potential to be catalytic, only the smallest substrates would

147

be capable of diffusing into the MOF to access the Mo(CO)3 pre-catalyst. Potentially, if 15-

(COOH)3 could be successfully synthesized, a MOF constructed from the larger complex would relieve the pore size constraints by providing more space between metal nodes, thus expanding the pore openings.

5.3.2 Tetrapyrazolyl Complexes as Ligands for MOF Synthesis

5.3.2.1 Tetrapyrazolylpropane

O O HO O

O O O O O N N O O O S N N H2N OH O OH 4 mol% N 15eq KOH N N N N N 9:1 MeOH:H O O O HEAT O N 2 N N N 90C overnight O N N

O OH

N NH O O O HO 4.5 eq

B H4TPP

Scheme 5.4: Synthesis of H4TPP in two steps.

195 Initially, H4TPP proved very hard to isolate. Communications with Smith, et. al. indicated that temperature and solvent removal (reaction proceeds by eliminating methanol) were crucial for high yields and purity for their tetrapyrazolylpropane ligands. Further optimization also indicated that the choice of catalytic acid impacted the yield of the final product (Scheme 5.4).

Although synthesis of H4TPP was now possible, the overall reaction yield remained low at 20% complicating scale up and subsequent screening trials.

H4TPP itself was tested under a variety of conditions to make a MOF. However, no crystalline

products were formed. Considering the geometry of the H4TPP ligand (Figure 5.7a) it is likely that the propyl linker is too flexible. Combined with rotation of the pyrazolyl rings TPP likely favours amorphous coordination polymers due to its lack of defined geometry.

148

HO O

HO O O OH O N N OH N N N N N N PdCl2 Cl Cl Pd Pd N N N Cl Cl O N MeCN N N N N 90 °C OH

O HO HO O O OH 16 TPP

Scheme 5.5: Synthesis of palladium complex 16 from H4TPP.

In order to add rigidity TPP was reacted with palladium dichloride to produce a metalloligand, 16 (Scheme 5.5). Coordination of palladium to the pyrazolyl chelate site effectively locked the confirmation of the pyrazolyl rings producing a semi-rigid metalloligand (Figure 5.7b). Although free rotation around the propyl group was still possible the defined geometry should facilitate MOF formation. Furthermore, the production of a PdII complex also allows for the future investigation of the catalytic properties of 16.187

Figure 5.7: Crystal structures of a) H4TPP ligand and b) Pd2(H4TPP)Cl4 (16) complex; solvent molecules not shown.

With 16 in hand we set out to perform approximately 100 MOF synthesis attempts evaluating different metal salts to form the connective nodes, different solvent systems, and different

149

temperatures. To date there have been no successes. The challenge lies in the inherently iterative nature of MOF synthesis. While 16 now has the correct geometry and characteristics to make a MOF there is still a high degree of flexibility and potential steric effects that must be overcome. Until such time as a crystalline MOF is formed, 16 may be successful at making a bimetallic species by coordinating an oxophilic metal to the available carboxylate groups. In this fashion, we may evaluate the reactivity of a mixed-metal TPP species towards a one pot, multi-step catalytic process. Hydrogen bonding may also be taken advantage of to produce hydrogen bonded networks of 16 (as seen with compounds 7, 8, and 9) for catalytic evaluation.

5.3.2.2 Tetrapyrazolyl-p-xylene

O O O O O O O N N 1eq NaH O O 1) (95%) 0.25eq Br N O 2) N N N Cl Br N N Pd Br N Cl Cl PdCl2 Pd N N N Br Cl HN O N N N MeCN N DMSO N O 80 °C 80 °C O O O O O O

TPX 17

Scheme 5.6: Synthesis of TPX and palladium complex 17.

After myriad optimizations, a reliable synthesis for TPX was realized. The main challenge was caused by the competitive side reaction of tetrabromoparaxylene with base, quenching the reagent. In order to prevent this, pyrazole was first fully deprotonated using sodium hydride. To this basic mixture, tetrabromoparaxylene was added via slow addition with a syringe pump. The slower the addition, the higher the yield and purity of TPX (Scheme 5.6, Figure 5.8a). With the ligand in hand, desterification was attempted in order to form TPX-(COOH)4. Unfortunately, we were not successful.

Mild conditions such as ester hydrogenation were unreactive and did not yield the desired carboxylic acid. Standard ester hydrolysis with base instead results in deprotonation of the benzylic carbon and subsequent decomposition of the ligand. Attempts were made to synthesize the benzyl ester form of the TPX ligand. Unfortunately, the benzyl ester behaved the same as the

150

ethyl ester variant. As a consequence, although the ligand precursor could be synthesized, the

desired H4TPX carboxylate could not. TPX may still reacted with PdCl2 to synthesize 17 in order to demonstrate the potential of TPX as a metalloligand and allow for future catalytic investigations (Figure 5.8b).187 Additionally, removal of the benzylic proton for a non-reactive group may stabilize TPX enough to allow for desterification and subsequent investigations of the ligand for MOF formation.

Figure 5.8: Crystal structure of a) TPX ligand and b) Pd2(TPX)Cl4 (17) palladium complex; solvent not shown.

5.3.3 Conclusions

Four new ligands, TPM-1, TPM-2, H4TPP, and TPX were successfully synthesized and

characterized although the synthesis of H4TPP and TPX proved quite challenging. Subsequent

coordination to Mo(CO)6 and Pd2Cl2, respectively, led to complexes 14-17. These complexes all have the potential to act as catalytic metalloligands for MOF formation and 14 and 15 were both demonstrated as effective pre-catalysts for the oxidation of cylcooctene. However, no MOFs were realized with 15-17. 14 could not produce a MOF of sufficient yield and purity to allow for catalytic testing. This was due to a combination of factors such as the fluxional nature of TPM

ligands, an overly flexible propyl backbone in H4TPP, and the fact that TPX was not stable/reactive towards desterification. That being said, these materials hold great potential provided these synthetic hurdles can be overcome.

151

5.4 Experimental 5.4.1 General Considerations

Elemental analyses were performed on a Thermo Flash 2000 CHN analyzer. NMR spectra were recorded on a Bruker Avance 400 spectrometer. Both 1H and 13C NMR spectra were referenced and reported relative to the solvent's residual signals. Catalytic trials were monitored using an Agilent Technologies 7890A Gas Chromatography System with a silica column coupled to an Agilent Technologies 5975C inert XL Mass Spectrometer with a triple axis detector. n-butylether and mesitylene were used as internal standards for the investigation of 14 and 15 respectively. All reagents were purchased from commercial sources and used without further purification

unless otherwise noted. DMSO was distilled from a refluxing CaH2 mixture and stored over activated molecular sieves. Ethyl diacetylacetate and 4-bromo-phenyl-2,4-pentanedionewere prepared following literature procedures.143, 193 Synthesis of trispyrazolylmethyl compounds and coordination of molybdenum were adapted from literature procedures as detailed below.181,196,197 Synthesis of tetrapyrazolyl-p-xylene and tetrapyrazolylpropane adapted from modified literature procedures.194,198

5.4.2 Ligand Synthesis

5.4.2.1 Synthesis of Triethyl-1,1',1''-methanetriyltris(3,5-dimethyl-1H- pyrazole-4-carboxylate) (TPM-1)

Ethyl ester pyrazolyl (5.1 g, 30 mmol), tetra-n-butylammonium bromide (791 mg, 2.4 mmol), and sodium bicarbonate (18.1 g, 215 mmol) were combined with chloroform (14 mL) and water

(29 mL) and allowed to reflux for 72 hours at 70 °C under a N2 atmosphere for five days. After cooling, the organic layer was extracted with ether (30 mL) and washed with water (20mL), dried over magnesium sulfate, filtered and dried under reduced pressure to yield 5 g of crude brown oil. The crude product was purified by flash column chromatography on silica gel (80% / 20% hexane / ethyl acetate) and dried under reduced pressure to produce a solid. 1.0 g (20 % yield) of product collected with 60% of starting material recovered after column. Note NMR 1 may also be performed with CDCl3. CD2Cl2 used to directly compare TPM-1 and 14. H NMR

(400 MHz, CD2Cl2) δ 8.14 (s, 1H), 4.26 (q, J = 7.1 Hz, 6H), 2.36 (s, 9H), 2.32 (s, 9H), 1.33 (t, J 13 = 7.1 Hz, 9H). C NMR (101 MHz, CD2Cl2) δ 164.37, 151.77, 146.54, 112.79, 81.03, 60.52,

152

14.83, 14.68, 11.25. Anal. Calc. for C25H34N6O6: C 58.35, H 6.66, N 16.33; found C 58.53, H 6.94, N 16.07.

16000 2.36 2.32 1.33 4.26 8.14 4.27

15000

14000

13000

12000

11000

10000

9000

8000

7000

6000

5000

4000

3000

2000

1000

-1000 0.95 5.95 8.65 8.60 9.00

8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 f1 (ppm)

1 Figure 5.9: H NMR spectrum of TPM-1 in CD2Cl2.

2600 164.37 151.77 146.54 112.79 81.03 60.52 14.83 14.68 11.25 2400

2200

2000

1800

1600

1400

1200

1000

800

600

400

200

-200

-400

-600

-800

170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 5.10: C NMR spectrum of TPM-1 in CD2Cl2.

153

5.4.2.2 Synthesis of Molybdenum Triethyl-1,1',1''-methanetriyltris(3,5- dimethyl-1H-pyrazole-4-carboxylate) Tricarbonyl (14)

TPM-1 (2.29 g, 4.45 mmol) was combined with molybdenum hexacarbonyl (1.41g, 5.33 mmol) in mesitylene (42 mL) and allowed to heat to 150 °C for 1.5 hours under nitrogen. While heating the solution turned yellow and a yellow precipitate formed. The yellow powder was washed with toluene, ethanol, and ether before drying. 2.7 g (86 % yield) of yellow product collected. 1H

NMR (400 MHz, CD2Cl2) δ 8.09 (s, 1H), 4.30 (q, J = 7.1 Hz, 6H), 2.83 (s, 9H), 2.82 (s, 9H), 13 1.34 (t, J = 7.1 Hz, 9H). C NMR (101 MHz, CD2Cl2) δ 163.04, 158.13, 144.87, 113.26, 67.93,

61.45, 16.27, 14.58, 11.42. Anal. Calc. for MoC28H34N6O9: C 48.42, H 4.93, N 13.82; found C 48.53, H 5.09, N 12.04. 4.29 4.31 2.83 2.82 1.34 8.09 6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

-500 0.92 6.05 17.81 9.00

8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 f1 (ppm)

1 Figure 5.11: H NMR spectrum of 14 in CD2Cl2.

154

1600 11.42 14.58 16.27 61.45 67.93 113.26 144.87 158.13 163.04 1500

1400

1300

1200

1100

1000

900

800

700

600

500

400

300

200

100

-100

160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 5.12: C NMR spectrum of 14 in CD2Cl2.

5.4.2.3 Synthesis of 1,1',1''-methanetriyltris(3,5-dimethyl-1H-pyrazole-4- carboxylic acid) (14-[COOH]3)

14 (500 mg, 0.642 mmol) and sodium hydroxide (6.60 g, 194 mmol) were combined in a flask with water (25 mL) and uninhibited tetrahydrofuran (25 mL). The system was purged with nitrogen for three minutes and left to reflux under nitrogen at 70 °C for 5 days. The reaction was deemed complete when the THF layer was colourless. The THF was removed under vacuum and the reaction was quenched with dilute acid to induce precipitation of a yellow solid. This solid was collected on a filter frit and washed with water. The solids were redissolved, filtered again, and dried under vacuum to yield 270 mg of crude yellow solids (~ 70 % pure via 1H NMR, ~37 % yield). The isolated crude yellow solids were not stable and were used immediately for MOF synthesis. 1H NMR (400 MHz, DMSO-d6) δ 12.33 (s, 4H), 6.80 (t, J = 6.4 Hz, 2H), 3.68 (t, J = 6.4 Hz, 2H), 2.43 (s, 12H), 2.24 (s, 12H)

5.4.2.4 Synthesis of Ethyl-4-(3,5-dimethyl-1H-pyrazol-4-yl)benzoate (A)

4-bromo-phenyl-2,4-pentanedione (125 mg, 0.50 mmol), xantphos (14.4 mg, 0.025 mmol), and palladium acetate (2.79 mg, 0.013 mmol) were combined with ethanol (2 mL, 34.25 mmol) and triethylamine (2 ml) in the glass sleeve in a high pressure reactor vessel. The vessel was purged

155

with CO and left to react under 10 atm CO(g) for 24 hours at 70 °C. The reaction mixture was allowed to cool to room temperature, diluted with ethyl acetate (10mL), and dried under reduced pressure. The crude solids were dissolved in dichloromethane (~20mL) and washed with brine (~10mL) before collecting the organic layer. The crude product was dry loaded onto silica gel and purified by flash column chromatography on silica gel (1:1 dichloromethane / ethyl acetate) to produce yellow oil. Trituration with hexanes resulted in a brown solid which was dried under reduced pressure. 560 mg (93 % yield) of solids were collected from five combined reactions. m.p.: 114-118 °C; 1H NMR (400 MHz, CDCl3): 8.12 (d, J = 7.9 Hz, 2H), 7.29 (d, J = 8.0 Hz, 2H), 4.40 (q, J = 7.1 Hz, 2H), 2.70 (d, 6H), 1.41 (t, J = 7.1Hz, 3H).

5.4.2.5 Synthesis of Triethyl-4,4',4''-(methanetriyltris[3,5-dimethyl-1H- pyrazole-1,4-diyl])tribenzoate (TPM-2)

A (500 mg, 2.05 mmol), tetra-n-butylammonium bromide (52.78 mg, 0.16 mmol), and sodium bicarbonate (2.58 g, 30.70 mmol) were combined with chloroform (12.5 mL) and water (27.5

mL) and allowed to reflux for 72 hours at 70 °C under a N2 atmosphere. After cooling the organic layer was separated and washed with water (20mL) and brine (10mL). The organic fraction was treated with magnesium sulfate, filtered, and dried under reduced pressure. The crude product was purified by flash column chromatography on silica gel (60% / 40% hexane / ethyl acetate) and dried under reduced pressure to provide 340 mg (68 % yield) of pale yellow 1 solids. H NMR (400 MHz, CDCl3): 8.29 (s, 1H), 8.07 (d, J = 8.3 Hz, 6H), 7.33 (d, J = 8.2 Hz, 6H), 4.38 (q, J = 7.1 Hz, 6H), 2.23 (s, 9H), 2.15 (s, 9H), 1.39 (t, J = 7.1Hz, 9H).

5.4.2.6 Synthesis of Molybdenum Triethyl-4,4',4''-(methanetriyltris[3,5- dimethyl-1H-pyrazole-1,4-diyl])tribenzoate Tricarbonyl (15)

TPM-2 (100 mg, 0.13 mmol) and molybdenum hexacarbonyl (89 mg, 0.34 mmol) were stirred together for 1 hour in dioxane (4 mL) at 110 °C. During this time the color of the mixture changed from yellow to dark brown. The reaction mixture was allowed to cool to room temperature and pentane (5mL) was added. The precipitate was collected by vacuum filtration, washed with pentane (2 × 10mL) and diethyl ether (2 × 10mL) and dried under reduced pressure 1 1 for 30 min. 40 mg (32 % yield) of pale yellow solids were collected. H NMR ( H; CDCl3, 400 MHz): 8.12 (d, J = 7.4 Hz, 6H), 7.29 (d, J = 7.7 Hz, 6H), 4.41 (q, J = 7.1 Hz, 6H), 2.76 (s, 18H), 1.41 (t, J = 7.0Hz, 9H). IR (KBr, υ/cm-1): 1909, Mo(CO) stretch.

156

5.4.2.7 Synthesis of Tetraethyl-1,1',1'',1'''-(propane-1,1,3,3- tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylate) (B)

Ethyl ester pyrazolyl (1.70 g, 10.1 mmol) and taurine (12.7 mg, 0.101 mmol) were placed into a dry two neck flask equipped with a reflux condenser and degassed under vacuum at 80 °C. Against a vigorous flow of nitrogen 1,1’,3,3’-tetramethoxy propane (0.366 mL, 2.22 mmol) was added to the flask. The system was kept under nitrogen and heated at 150 °C overnight before raising the temperature to 210 °C for 17 hours. After cooling, the resulting black tar was purified via flash column chromatography on silica gel (ethyl acetate / hexanes) after which 415 mg (27 1 % yield) of white solids were collected. H NMR (400 MHz, CDCl3) δ 6.38 (t, J = 7.2 Hz, 2H), 4.27 (q, J = 7.1 Hz, 8H), 3.79 (t, J = 7.2 Hz, 2H), 2.41 (s, 12H), 2.40 (s, 12H), 1.34 (t, J = 7.1 Hz, 13 12H). C NMR (101 MHz, CDCl3) δ 165.00 , 150.22 , 145.26 , 110.11 , 58.85 , 14.04 , 10.88.

Anal. Calc. for C35H48N8O8: C 59.31, H 6.83, N 15.81; found C 59.40, H 6.67, N 15.33.

17000

16000 6.38 4.28 4.26 3.79 2.41 2.40 1.34

15000

14000

13000

12000

11000

10000

9000

8000

7000

6000

5000

4000

3000

2000

1000

-1000 1.97 8.04 2.00 23.54 12.28

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 f1 (ppm)

1 Figure 5.13: H NMR spectrum of B in CDCl3.

157

24000 164.27 151.56 144.79 111.27 66.79 60.02 36.74 14.69 14.49 10.87 22000

20000

18000

16000

14000

12000

10000

8000

6000

4000

2000

-2000

170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

13 Figure 5.14: C NMR spectrum of B in CDCl3.

5.4.2.8 Synthesis of 1,1',1'',1'''-(propane-1,1,3,3-tetrayl)tetrakis(3,5- dimethyl-1H-pyrazole-4-carboxylic acid) (H4TPP)

B (79.7 mg, 0.112 mmol), potassium hydroxide (110 mg, 1.96 mmol), 0.5 mL of water and 4.5 mL of methanol were combined in a flask. The clear yellow solution was refluxed at 90 °C overnight. The solution was concentrated under vacuum and quenched with 12 M hydrochloric acid (0.15 mL, 1.8 mmol). 5 mL of water was added to the resulting slurry after which the solids were collected via gravity filtration and washed 4 times with water. After vacuum drying 41 mg 1 1 (62 % yield) of white solids were collected. H NMR (400 MHz, DMSO-d6) δ H NMR (400

MHz, DMSO-d6) δ 12.30 (s, 4H), 6.78 (t, J = 6.3 Hz, 2H), 3.66 (t, J = 6.3 Hz, 2H), 2.41 (s, 12H), 13 2.23 (s, 12H). C NMR (101 MHz, DMSO-d6) δ 164.91, 149.82, 144.24, 110.43, 66.53, 14.09,

10.47. Anal. Calc. for C27H32N8O8-0.5(H2O): C 53.62, H 5.49, N 18.50; found C 53.62, H 5.46, N 18.33.

158

5000 3.66 6.78 12.30

4500

4000

3500

3000

2500

2000

1500

1000

500

0 12.00 12.12 2.25 2.11 3.78

12.5 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 f1 (ppm)

1 Figure 5.15: H NMR spectrum for H4TPP in DMSO-d6.

1500

1400 10.47 14.09 149.82 144.24 110.43 66.53 164.91

1300

1200

1100

1000

900

800

700

600

500

400

300

200

100

-100

170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 5.16: C NMR spectrum for H4TPP in DMSO-d6.

159

5.4.2.9 Synthesis of Dipalladium 1,1',1'',1'''-(propane-1,1,3,3- tetrayl)tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylic acid) Tetrachloride Complex (16)

H4TPP (50.0 mg, 0.0838 mmol), palladium dichloride (40.2 mg, 0.189 mmol), and 15 mL of acetonitrile were charged into a Schlenk flask. The red solution was degassed via three freeze- pump-thaw cycles and left to heat at 90 °C overnight. Orange crystals precipitated from solution. The crystals were collected on a Hirsch funnel and washed with the mother liquor and acetonitrile. 58 mg (70 % yield) of orange crystals were collected. 1H NMR (400 MHz, DMSO-

d6) δ 12.33 (s, 4H), 6.80 (t, J = 6.4 Hz, 2H), 3.68 (t, J = 6.4 Hz, 2H), 2.43 (s, 12H), 2.24 (s, 12H). 13 C NMR (101 MHz, DMSO-d6) δ 164.91, 149.83, 144.26, 110.40, 66.54, 38.89, 14.09, 10.48.

Anal. Calc. for C27H32N8O8Pd2Cl4: C 34.09, H 3.39, N 11.78; found C 33.97, H 3.81, N 11.54.

3500 12.33 6.80 3.68 2.43 2.24

3000

2500

2000

1500

1000

500

0 1.29 2.00 2.45 11.71 11.73

13.0 12.5 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 f1 (ppm)

1 Figure 5.17: H NMR spectrum of 16 in DMSO-d6. Excess water present in the solvent suppresses the 12.33 ppm acid peak signal.

160

500 164.91 149.83 144.26 110.40 66.54 38.89 14.09 10.48

450

400

350

300

250

200

150

100

-50

-100

-150

210 200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 -10 f1 (ppm)

13 Figure 5.18: C NMR spectrum of 16 in DMSO-d6.

5.4.2.10 Synthesis of Tetraethyl-1,1',1'',1'''-(1,4-phenylenebis[methanetriyl]) tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylate) (TPX)

95% Sodium Hydride (78.0 mg, 3.25 mmol) and ethyl ester pyrazole (547 mg, 3.25 mmol) were placed in a Schlenk flask and degassed over ~30 min. 5 mL of dry DMSO was added and the flask was heated at 80 °C for one hour to form a clear yellow solution. Tetrabromo-p-xylene (343 mg, 0.813 mmol) was dissolved in 5.7 mL of DMSO in a vial and slowly added to the reaction flask via syringe pump at a rate of 50 μL/min. After addition, the orange solution was left to stir at 80 °C overnight under nitrogen. After cooling, the reaction mixture was decanted into a clean flask and quenched with 50 mL of distilled water. The solids were collected on a Bruckner funnel, washed three times with water and dried to yield 323 mg (52 % yield) of pale yellow solids. X-ray quality crystals may be grown via slow diffusion of pentane into chloroform. 1H

NMR (400 MHz, CDCl3) δ 7.77 (s, 2H), 6.94 (s, 4H), 4.28 (q, J = 7.1 Hz, 9H), 2.51 (s, 12H), 13 2.40 (s, 12H), 1.35 (t, J = 7.1 Hz, 12H). C NMR (101 MHz, CDCl3) δ 127.39, 60.03, 14.70,

14.52, 12.13. Anal. Calc. for C40H50N8O8: C 62.32, H 6.54, N 14.54; found C 62.45, H 6.44, N 14.29.

161

8000 7.77 6.94 4.29 4.28 2.51 2.40 1.35 7500

7000

6500

6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

-500 2.10 3.93 9.06 11.74 11.71 12.00

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 f1 (ppm)

1 Figure 5.19: H NMR spectrum of TPX in CDCl3.

1400

127.39 60.03 14.70 14.52 12.13 1300

1200

1100

1000

900

800

700

600

500

400

300

200

100

-100

170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

13 Figure 5.20: C NMR spectrum of TPX in CDCl3.

162

5.4.2.11 Synthesis of Dipalladium Tetraethyl-1,1',1'',1'''-(1,4-phenylenebis [methanetriyl])tetrakis(3,5-dimethyl-1H-pyrazole-4-carboxylate) Tetrachloride Complex (17)

TPX (300 mg, 0.389 mmol) and palladium dichloride (187 mg, 0.876 mmol) were placed in a Schlenk flask to which 70 mL of acetonitrile was added. The resulting solution was degassed via three freeze-pump-thaw cycles after which it was sealed under nitrogen and left to heat at 90 °C. After three days of heating, the solution was cooled down to room temperature and concentrated under vacuum to ~15 mL at which point solids precipitated from the yellow solution. The solids were collected via gravity filtration, washed twice with the filtrate and once with clean acetonitrile. After drying 184 mg (42 % yield) of orange solids were collected. X-ray quality 1 crystals may be grown via slow diffusion of ether into acetonitrile. H NMR (400 MHz, CDCl3) δ 8.07 (s, 2H), 7.18 (s, 4H), 4.33 (q, J = 7.1 Hz, 9H), 2.83 (s, 12H), 2.79 (s, 12H), 1.36 (t, J = 7.1 13 Hz, 12H). C NMR (101 MHz, CDCl3) δ 162.04, 157.88, 147.79, 135.80, 128.78, 113.01, 69.65,

61.69, 16.21, 14.41, 11.58. Anal. Calc. for C40H50N8O8-3(H2O): C 40.73, H 4.79, N 9.50; found C 40.74, H 4.77, N 9.32.

24000

22000 8.07 7.26 CDCl3 7.18 4.34 4.32 2.83 2.79 1.68 1.36 20000

18000

16000

14000

12000

10000

8000

6000

4000

2000

-2000 1.93 3.87 8.00 23.30 6.02 11.83

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 f1 (ppm)

1 Figure 5.21: H NMR spectrum of 17 in CDCl3.

163

3000

2800 162.04 157.88 147.79 135.80 128.78 113.01 69.65 61.19 16.21 14.41 11.58

2600

2400

2200

2000

1800

1600

1400

1200

1000

800

600

400

200

-200

-400

160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 f1 (ppm)

13 Figure 5.22: C NMR spectrum of 17 in CDCl3.

5.4.3 Catalytic Conditions and Methodology

In a two neck flask, 14 (30 mg, 0.038 mmol) was suspended in cyclooctene (1.0 mL, 7.7 mmol), sparged with nitrogen for 10 minutes, and heated to 55 °C producing a yellow solution. To this solution, 1.0 mL of n-butyl ether was added as an internal standard. Once at the reaction temperature, a 0.05 mL aliquot of solution was taken, quenched with cold DCM containing a

catalytic amount of MnO2, diluted, and analyzed via GCMS to determine the ratio of starting material to internal standard at t = 0 min. Tertbutylhydrogen peroxide (5.5 M in hexanes, 2.1 mL, 11.5 mmol) was then introduced and aliquots of solution were taken every 15 min and analyzed.

A similar procedure was followed for the catalytic investigation of 15.

5.4.4 MOF Synthesis

14-(COOH)3 (5 mg, 6.2 μmol) was dissolved in 1.75 mL of DMF. To this solution was added zinc nitrate (2.8 mg, 9.4 μmol) in 0.5 mL of DMF. The reaction mixture was then diluted with 1.75 mL of water, sealed in a vial, and left to react at 85 °C for one week. After heating, trace amounts of yellow MOF crystals were observed (3-5 crystals per vial).

164

5.4.5 X-Ray Diffractometry

X-ray quality single crystals were obtained as described in the synthesis above. The crystals were mounted on the tip of a MiTeGen MicroMount. The single-crystal X-ray diffraction data were collected on a Bruker Kappa Apex II CCD diffractometer with Mo Kα radiation (λ = 0.71073 Å) operating at 50 kV and 30 mA, at 147 K controlled by an Oxford Cryostream 700 series low temperature system. The data integration and absorption correction were performed with the Bruker Apex 2 software package.103 The structure was solved by direct methods and refined using SHELXTL V2016/4.104,105 Disordered solvent molecules and residual electron density were removed using the Platon Squeeze method.106 All non-hydrogen atoms except for the solvent atoms were refined anisotropically. The positions of the hydrogen atoms were calculated using the riding model.

Table 5.1: Crystallographic properties of tetrapyrazolyl ligands.

Name H4TPP TPX

Formula C27H32N8O8, (H2O)0.5 C40H50N8O8

Formula Weight [g/mol] 605.61 770.88

Crystal System Triclinic Triclinic

Space Group P-1 (No. 2) P-1 (No. 2)

a [Å] 8.6362 (14) 8.5466 (5)

b [Å] 12.797 (2) 11.1455 (6)

c [Å] 14.917 (3) 13.3595 (9)

α [°] 72.097 (8) 104.119 (3)

β [°] 89.876 (8) 97.414 (3)

γ [°] 80.964 (8) 96.988 (3)

Volume [Å3] 1547.4 (5) 1208.20 (13)

Z 2 1

D(calc) [g/cm3] 1.300 1.059

165

μ(MoKα) [ /mm ] 0.099 0.075

F(000) 638 410

Crystal Size (estimated) 0.08 x 0.11 x 0.40 0.20 x 0.35 x 0.47 [mm]

Temperature (K) 150 150

Radiation [Angstrom] 0.71073 0.71073

θ Min-Max [°] 1.4, 27.5 1.9, 27.5

Dataset -11: 10; -16: 16; -19: 19 -11: 10; -14: 14; -17: 17

Tot., Uniq. Data, R(int) 25143, 7051, 0.041 25220, 5551, 0.032

Observed data [I > 2.0 σ(I)] 4255 4270

Nref, Npar 7051, 417 5551, 260

R1, wR2, GooF 0.0612, 0.1880, 1.03 0.0459, 0.13545, 1.06

Max. and Av. Shift/Error 0.00, 0.00 0.01, 0.00

Min. and Max. Resd. Dens. -0.55, 0.77 -0.30, 0.33 [e/Å3]

Table 5.2 Crystallographic properties of metallo-ligand complexes 14, 16, and 17.

Name Mo(TPM-1)(CO) (14) Pd2(H4TPP)Cl4 (16) Pd2(TPX)Cl4 (17)

Formula C28H34N6O9Mo C29H37N9O9Pd2Cl4 C44.8H57.2N10.4O8Pd2Cl4

Formula Weight 694.55 1010.27 1224.01 [g/mol]

Crystal Orthorhombic Monoclinic Monoclinic System

Space Group Pbca (No. 61) P 21/n (No. 14) P 21/c (No. 14)

a [Å] 16.6962(6) 13.8299 (16) 9.9865 (4)

b [Å] 18.9469(8) 19.690 (2) 20.0055 (8)

166

c [Å] 19.4915(7) 15.5199 (16) 14.2511 (6)

β [°] 90 104.227 (6) 105.501 (2)

Volume [Å3] 6166.0(4) 4096.6 (8) 2743.6 (2)

Z 8 4 2

D(calc) 1.496 1.638 1.482 [g/cm3]

μ(MoK ) α 0.487 1.197 0.907 [/mm ]

F(000) 2864 2024 1246

Crystal Size (estimated) 0.08 x 0.35 x 0.52 0.02 x 0.03 x 0.07 0.06 x 0.06 x 0.07 [mm]

Temperature 150 150 150 (K)

Radiation [Å] 0.71073 0.71073 0.71073

θ Min-Max [°] 1.9, 31.3 1.7, 25.0 2.0, 27.5

Dataset -24: 23; -20: 20; -18: 25 -16: 15; -23: 23; -18: 18 -12: 9; -25: 25; -18: 18

Tot., Uniq. 38033, 7084, 0.082 34109, 7225, 0.165 45646, 6266, 0.065 Data, R(int)

Observed data 4603 4013 4456 [I > 2.0 σ(I)]

Nref, Npar 7084, 403 7225, 480 6266, 331

R1, wR2, GooF 0.0523, 0.1424, 1.03 0.0867, 0.2371, 1.03 0.0469, 0.1428, 1.06

Max. and Av. 0.00, 0.00 0.00, 0.00 0.00, 0.00 Shift/Error

Min. and Max. Resd. Dens. -0.74, 0.78 -1.23, 2.51 -0.98, 1.70 [e/Å3]

167

Chapter 6 6 Conclusions, Outlook, and Future Work 6.1 General Overview

This thesis is concerned with the exciting chemistry of metal-organic frameworks derived from pyrazolate derivatives. In total, five novel pyrazolate ligands were synthesized. Their reactivities and MOF structures were explored. 4-(4-(3,5-dimethyl-1H-pyrazol-4-yl)phenyl)pyridine (HL) was investigated in Chapter 2 and found to successfully produce a luminescent MOF, 1, that was effective at sensing VOCs. This project was originally started due to a serendipitous occurrence: decomposition of tris(pyridylphenylpyrazolyl)methane during a MOF synthesis attempt yielded stable luminescent crystals. These stable compounds were the first hint of a new investigative direction.

1,1’-methylenebis(3,5-dimethyl-1H-pyrazolyl-4-carboxylic acid) (H2BPM) formed the basis of chapters 3 and 4 wherein a MOF effective for the selective gas adsorption of methane, 5, was synthesized in addition to a host of other new materials (compounds 6-13). With assistance from some talented volunteers and collaborators we were able to fully explore the structures of these compounds to discover how BPM interacts with different metal centres and allow for predictions to be made regarding potential future materials.

My final thesis chapter, 5, concerns the attempted synthesis of catalytically active MOFs. Several new materials were synthesized (14-17), two of which were demonstrated as effective catalysts for olefin oxidation (14 and 15). However, only one MOF was synthesized and not in sufficient quantities to fully characterize.

6.2 Chapter 2

In chapter two, we describe three new MOFs synthesized from CuX (X = Cl, Br) and HL that

incorporated the luminescent [Cu3Pz3] chromophore into their structure. It was demonstrated that

two possible MOFs could be formed from HL and CuBr, [Cu9L6Br2][CuBr2] (α-CuBr MOF, 1)

and [Cu8L6Br][CuBr2] (β-CuBr MOF, 2), depending on whether a competing ligand is present in solution. Furthermore, the luminescence intensity of 1 was observed to change in the presence of hydrophobic VOCs. Therefore, 1 may act as a luminescence sensor for hydrophobic VOCs such

168

as ethyl ester, pentane, and benzene (turn-on response in 1) or chloroform and ethyl ether (turn- off response in 1). However, 2 and 3 were not amicable to investigation of VOC sensing.

An interesting avenue for future investigations of these compounds would be to synthesize the silver and gold derivatives. As mentioned in the Chapter 2 discussion, reactions with silver were attempted but did not prove fruitful when investigated by Cindy Ma, second year undergraduate student, and Stefan Jevtić, summer volunteer undergraduate. Additionally, a simple trinuclear CuI ethylesterpyrazolate compound, 4, was successfully synthesized. Although 4 was able to catalyze azide-alkyne cycloadditions it did not serve as an effective ligand for MOF synthesis.

6.3 Chapter 3

Employment of H2BPM allowed for the formation of two microporous MOFs, 5 and 6. Although

[Cd(BPM)]n∙xDMSO (6) proved to decompose upon desolvation, [Ni(BPM)]n∙xDMSO (5) remained porous upon solvent exchange and material activation. Once activated, 5 was able to

selectively adsorb both CO2 and CH4. Although 5’s CO2 capture performance was not competitive with regards to better known literature materials, 5 was competitive as a MOF for coal mine methane capture. Indeed, preliminary experiments show 5 should have performance

equal to or greater than the best performing MOF presently reported, Cu(INA)2.

Future directions for this project are twofold. Firstly, computational simulations are adequate for identifying probable applications for a MOF but actual experiments need to be conducted to verify the simulations. Therefore, gas breakthrough experiments must be conducted on 5 either in-house or with the assistance of a collaborator. Secondly, 5 offers many potential sites for derivatization, pore engineering, and the development of an isoreticular series. Exploring the role of coordinated solvent, removing one or more pyrazolyl methyl group, and using a larger BPM ligand (e.g. bis(carboxylphenyl)pyrazolylmethane) would all provide insight into the optimization of 5 for coal mine methane capture. Further optimization and tuning may allow this microporous material to be competitive for other selective gas sorption applications such as flue

gas CO2 capture or biogas upgrading.

6.4 Chapter 4

In chapter 4, seven new coordination polymers and MOFs were described. These materials all

had interesting structures and presented insight into the bonding modes of H2BPM. Of particular

169

interest is the identification of the dinuclear cobalt secondary building unit observed in compounds 7-9. Identification of this secondary building unit allows for future predictions to be

made regarding the coordination of H2BPM to cobalt allowing for limited predictive design. Although the manganese and copper MOFs were non-porous, the remaining materials present interesting opportunities for further explorations of BPM MOFs.

We conducted preliminary reactivity studies but the potential applications of 7-13 remain relatively underexplored. The most promising material for future investigation is 13, the 2D

sheets of [Eu(BPM)2(OAc)]n. These europium sheets are very thermally robust. Future directions for this material may include solution state sensing whereby the pyrazole lone pairs coordinate to an analyte (e.g. dissolved ions) to modulate the emission spectrum. Another possibility is 2 introducing a new ligand to compete with the κ ,μ2-acetate. If these acetate groups could be replaced with, for instance, benzene dicarboxylate a true 3D MOF should be possible. If the thermal stability is retained this hypothetical [Eu(BPM)bdc]n material should have very interesting luminescent properties. It may be an effective platform for sensing VOCs with mild

thermal treatment regenerating the pristine MOF. Additionally, [Cu(BPM)(DMSO)]n (12) may be exploited in the future if interpenetration is avoided and/or a pillaring ligand is coordinated to the paddlewheel, as opposed to DMSO, to make a true 3D material.

Lastly, use of tris- or tetrapyridyl ligands could replace Bpy when synthesizing cobalt complexes 7-10. This subtle change should allow for greater crosslinking between the dinuclear cobalt secondary building units. This in turn may allow for the formation of a true 2D or 3D MOF.

6.5 Chapter 5

Of all the work presented in this thesis, chapter 5 contains the lowest lows and highest highs. The initial focus of or project was to synthesize catalytic MOFs. Spanning 2011-2013, and 2016, much work was invested towards realizing this goal. While four new, never reported metalloligands were synthesized (14-17) we were unable to make large quantities of MOF (in the case of 14, the trispyrazolylmethane complex), nor find the correct synthetic conditions to make a MOF (15, 16), and were stymied by the properties of the ligand itself (17).

Despite setbacks, we now have a range of new metalloligands that, given the right combination of skill and luck, should all make very interesting MOFs with catalytic properties. 14 and 16

170

especially seem to consist of one solvent/metal salt combination away from repeated production of MOFs for catalytic testing. Compound 14 already yields a MOF as a minor species under the synthetic conditions attempted. 16 with its inherent flexibility and multiple carboxylate groups may lend itself towards formation of catalytic aerogels. Only two other tetrapyrazolylpropane complexes are known in the literature. 14 and 16 are new complexes which suggest a wide range of further studies both within and without the scope of materials chemistry.

A potential remains to explore tetrapyrazolyl ligands for the synthesis of coordination polymers via the metal chelate site as has been demonstrated in the literature. The carboxylate groups may then serve as a secondary site for post-synthetic metalation or to increase the solubility of the resulting polymers. Additionally, the trispyrazolylborate derivatives of the TPM ligands may be synthesized and studied. Although preliminary reactions were performed to demonstrate that synthesis of these compounds is possible, difficulties isolating and purifying the resulting ligands dissuaded us from pursuing it further.

171

Chapter 7 References

1 S. R. Batten, N. R. Champness, X.-M. Chen, J. Garcia-Martinez, S. Kitagawa, L. Öhrström, M. O’Keeffe, M. Paik Suh and J. Reedijk, Pure Appl. Chem., 2013, 85, 1715– 1724. 2 M. Eddaoudi, D. B. Moler, H. Li, B. Chen, T. M. Reineke, M. O’Keeffe and O. M. Yaghi, Acc. Chem. Res., 2001, 34, 319–330. 3 D. J. Tranchemontagne, J. L. Mendoza-Cortés, M. O’Keeffe and O. M. Yaghi, Chem. Soc. Rev., 2009, 38, 1257–1283. 4 Y. Kinoshita, I. Matsubara, T. Higuchi and Y. Saito, Bull. Chem. Soc. Jpn., 1959, 32, 1221–1226. 5 H. Li, M. Eddaoudi, M. O’Keeffe and O. M. Yaghi, Nature, 1999, 402, 276–279. 6 O. M. Yaghi, G. Li and H. Li, Nature, 1995, 378, 703–706. 7 S. S.-Y. Chui, S. M.-F. Lo, J. P. H. Charmant, A. G. Orpen and I. D. Williams, Science, 1999, 283, 1148–1150. 8 W. P. Lustig, S. Mukherjee, N. D. Rudd, A. V. Desai, J. Li and S. K. Ghosh, Chem. Soc. Rev., 2017, 46, 3242–3285. 9 S. M. J. Rogge, A. Bavykina, J. Hajek, H. Garcia, A. I. Olivos-Suarez, A. Sepúlveda- Escribano, A. Vimont, G. Clet, P. Bazin, F. Kapteijn, M. Daturi, E. V. Ramos-Fernandez, F. X. L. i Xamena, V. V. Speybroeck and J. Gascon, Chem. Soc. Rev., 2017, 46, 3134– 3184. 10 K. Adil, Y. Belmabkhout, R. S. Pillai, A. Cadiau, P. M. Bhatt, A. H. Assen, G. Maurin and M. Eddaoudi, Chem. Soc. Rev., 2017, 46, 3402–3430. 11 B. Joarder, J.-B. Lin, Z. Romero and G. K. H. Shimizu, J. Am. Chem. Soc., 2017, 139, 7176–7179. 12 H. Kim, S. Yang, S. R. Rao, S. Narayanan, E. A. Kapustin, H. Furukawa, A. S. Umans, O. M. Yaghi and E. N. Wang, Science, 2017, 356, 430–434. 13 P. Li, J. A. Modica, A. J. Howarth, E. Vargas L., P. Z. Moghadam, R. Q. Snurr, M. Mrksich, J. T. Hupp and O. K. Farha, Chem, 2016, 1, 154–169. 14 P. Li, S.-Y. Moon, M. A. Guelta, S. P. Harvey, J. T. Hupp and O. K. Farha, J. Am. Chem. Soc., 2016, 138, 8052–8055. 15 C. Doonan, R. Riccò, K. Liang, D. Bradshaw and P. Falcaro, Acc. Chem. Res., 2017, 50, 1423–1432. 16 F. Liu, H. Huang, W. Gao, X. Zhang and J. P. Liu, CrystEngComm, 2017, 19, 3660– 3665. 17 J. E. M. Lewis, P. D. Beer, S. J. Loeb and S. M. Goldup, Chem. Soc. Rev., 2017, 46, 2577–2591.

172

18 I. Stassen, N. Burtch, A. Talin, P. Falcaro, M. Allendorf and R. Ameloot, Chem. Soc. Rev., 2017, 46, 3185–3241. 19 N. S. Bobbitt, M. L. Mendonca, A. J. Howarth, T. Islamoglu, J. T. Hupp, O. K. Farha and R. Q. Snurr, Chem. Soc. Rev., 2017, 46, 3357–3385. 20 P. C. Hsi (Rae Systems, Inc.), WO1994027141A1, 1994; P. C. Hsi (Rae Systems, Inc.), US Pat, 5393979, 1995; P. C. Hsi (Rae Systems, Inc.), US Pat, 5561344A, 1996. 21 J. D. Coy, P. L. Bigelow, R. M. Buchan, J. D. Tessari and J. O. Parnell, AIHAJ, 2000, 61, 268–274. 22 E. Comini, G. Faglia and G. Sberveglieri, Eds., Solid state gas sensing, Springer, New York, N.Y, 2009. 23 J. J. Perry, C. A. Bauer and M. D. Allendorf, in Metal-Organic Frameworks, ed. D. Farrusseng, Wiley-VCH Verlag GmbH & Co. KGaA, 2011, pp. 267–308. 24 L. E. Kreno, K. Leong, O. K. Farha, M. Allendorf, R. P. Van Duyne and J. T. Hupp, Chem. Rev., 2012, 112, 1105–1125. 25 Z. Hu, B. J. Deibert and J. Li, Chem. Soc. Rev., 2014, 43, 5815–5840. 26 M. M. Wanderley, C. Wang, C.-D. Wu and W. Lin, J. Am. Chem. Soc., 2012, 134, 9050– 9053. 27 M. Zhang, G. Feng, Z. Song, Y.-P. Zhou, H.-Y. Chao, D. Yuan, T. T. Y. Tan, Z. Guo, Z. Hu, B. Z. Tang, B. Liu and D. Zhao, J. Am. Chem. Soc., 2014, 136, 7241–7244. 28 Y. Li, S. Zhang and D. Song, Angew. Chem. Int. Ed., 2013, 52, 710–713. 29 C. Zhan, S. Ou, C. Zou, M. Zhao and C.-D. Wu, Anal. Chem., 2014, 86, 6648–6653. 30 L. Zhang, Z. Kang, X. Xin and D. Sun, CrystEngComm, 2016, 18, 193–206. 31 D. Zhao, Y. Cui, Y. Yang and G. Qian, CrystEngComm, 2016, 18, 3746–3759. 32 M. A. Mortellaro and D. G. Nocera, ChemTech, 1996, 26, 17–23. 33 K. Nakabayashi, S. Inoue, Y. Abiko and H. Mori, Macromolecules, 2013, 46, 4790– 4798. 34 K. M.-C. Wong, W.-S. Tang, X.-X. Lu, N. Zhu and V. W.-W. Yam, Inorg. Chem., 2005, 44, 1492–1498. 35 K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, T.-H. Bae and J. R. Long, Chem. Rev., 2012, 112, 724–781. 36 B. Li, M. Chrzanowski, Y. Zhang and S. Ma, Coord. Chem. Rev., 2016, 307, Part 2, 106–129. 37 S. Ma and H.-C. Zhou, Chem. Commun., 2010, 46, 44–53. 38 S.-J. Lee and Y.-S. Bae, J. Phys. Chem. C, 2014, 118, 19833–19841. 39 R. Vaidhyanathan, S. S. Iremonger, G. K. H. Shimizu, P. G. Boyd, S. Alavi and T. K. Woo, Science, 2010, 330, 650–653.

173

40 Y. Tulchinsky, C. H. Hendon, K. A. Lomachenko, E. Borfecchia, B. C. Melot, M. R. Hudson, J. D. Tarver, M. D. Korzyński, A. W. Stubbs, J. J. Kagan, C. Lamberti, C. M. Brown and M. Dincă, J. Am. Chem. Soc., 2017, 139, 5992–5997. 41 R. J. Marshall, S. L. Griffin, C. Wilson and R. S. Forgan, Chem. – Eur. J., 2016, 22, 4870–4877. 42 O. US EPA, Global Greenhouse Gas Emissions Data, https://www.epa.gov/ghgemissions/global-greenhouse-gas-emissions-data, (accessed 18 June 2017). 43 I. Sreedhar, T. Nahar, A. Venugopal and B. Srinivas, Renew. Sustain. Energy Rev., 2017, 76, 1080–1107. 44 X. Yang, R. J. Rees, W. Conway, G. Puxty, Q. Yang and D. A. Winkler, Chem. Rev., 2017, 117, 9524–9593. 45 Q. Yang, G. Puxty, S. James, M. Bown, P. Feron and W. Conway, Energy Fuels, 2016, 30, 7503–7510. 46 Y.-S. Bae and R. Q. Snurr, Angew. Chem. Int. Ed., 2011, 50, 11586–11596. 47 W. M. Haynes and D. R. Lide, Eds., CRC handbook of chemistry and physics: a ready- reference book of chemical and physical data, CRC Press, Boca Raton, Fla., 96. ed., 2015-2016., 2015. 48 Y. He, W. Zhou, G. Qian and B. Chen, Chem. Soc. Rev., 2014, 43, 5657–5678. 49 Z. Kang, L. Fan and D. Sun, J. Mater. Chem. A, 2017, 5, 10073–10091. 50 Z. Zhang, Z.-Z. Yao, S. Xiang and B. Chen, Energy Environ. Sci., 2014, 7, 2868–2899. 51 N. Casas, J. Schell, L. Joss and M. Mazzotti, Sep. Purif. Technol., 2013, 104, 183–192. 52 A. Sinha, L. A. Darunte, C. W. Jones, M. J. Realff and Y. Kawajiri, Ind. Eng. Chem. Res., 2017, 56, 750–764. 53 J. Li, X. Luo, N. Zhao, L. Zhang, Q. Huo and Y. Liu, Inorg. Chem., 2017, 56, 4141– 4147. 54 H. Kim, J. Park and Y. Jung, Phys. Chem. Chem. Phys., 2013, 15, 19644–19650. 55 Z. Kang, L. Fan and D. Sun, J. Mater. Chem. A, 2017, 5, 10073–10091. 56 C.-D. Wu and M. Zhao, Adv. Mater., 2017, 29, n/a-n/a. 57 X.-L. Lv, K. Wang, B. Wang, J. Su, X. Zou, Y. Xie, J.-R. Li and H.-C. Zhou, J. Am. Chem. Soc., 2017, 139, 211–217. 58 J. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. T. Nguyen and J. T. Hupp, Chem. Soc. Rev., 2009, 38, 1450–1459. 59 S.-H. Cho, B. Ma, S. T. Nguyen, J. T. Hupp and T. E. Albrecht-Schmitt, Chem. Commun., 2006, 2563–2565. 60 B. Kesanli, Y. Cui, M. R. Smith, E. W. Bittner, B. C. Bockrath and W. Lin, Angew. Chem. Int. Ed., 2005, 44, 72–75. 61 C.-D. Wu, A. Hu, L. Zhang and W. Lin, J. Am. Chem. Soc., 2005, 127, 8940–8941.

174

62 A. M. Shultz, O. K. Farha, J. T. Hupp and S. T. Nguyen, J. Am. Chem. Soc., 2009, 131, 4204–4205. 63 M. S. Deenadayalan, N. Sharma, P. K. Verma and C. M. Nagaraja, Inorg. Chem., 2016, 55, 5320–5327. 64 Y. Liu, S.-Y. Moon, J. T. Hupp and O. K. Farha, ACS Nano, 2015, 9, 12358–12364. 65 E. López-Maya, C. Montoro, L. M. Rodríguez-Albelo, S. D. Aznar Cervantes, A. A. Lozano-Pérez, J. L. Cenís, E. Barea and J. A. R. Navarro, Angew. Chem. Int. Ed., 2015, 54, 6790–6794. 66 A. W. Peters, Z. Li and O. K. Farha, ACS Cent. Sci., 2017, 3, 367–368. 67 S. Trofimenko, Chem. Rev., 1972, 72, 497–509. 68 H. V. R. Dias, H. V. K. Diyabalanage, M. G. Eldabaja, O. Elbjeirami, M. A. Rawashdeh- Omary and M. A. Omary, J. Am. Chem. Soc., 2005, 127, 7489–7501. 69 G. A. Santillan and C. J. Carrano, Inorg. Chem., 2008, 47, 930–939. 70 V. W.-W. Yam, V. K.-M. Au and S. Y.-L. Leung, Chem. Rev., 2015, 115, 7589–7728. 71 A. Vogler and H. Kunkely, in Transition Metal and Rare Earth Compounds, ed. H. Yersin, Springer Berlin Heidelberg, 2001, pp. 143–182. 72 C. A. Kent, D. Liu, T. J. Meyer and W. Lin, J. Am. Chem. Soc., 2012, 134, 3991–3994. 73 M. Mączka, B. Bondzior, P. Dereń, A. Sieradzki, J. Trzmiel, A. Pietraszko and J. Hanuza, Dalton Trans., 2015, 44, 6871–6879. 74 H. V. R. Dias, H. V. K. Diyabalanage, M. G. Eldabaja, O. Elbjeirami, M. A. Rawashdeh- Omary and M. A. Omary, J. Am. Chem. Soc., 2005, 127, 7489–7501. 75 M. A. Omary, M. A. Rawashdeh-Omary, M. W. A. Gonser, O. Elbjeirami, T. Grimes, T. R. Cundari, H. V. K. Diyabalanage, C. S. P. Gamage and H. V. R. Dias, Inorg. Chem., 2005, 44, 8200–8210. 76 B. Hu, G. Gahungu and J. Zhang, J. Phys. Chem. A, 2007, 111, 4965–4973. 77 G.-F. Gao, M. Li, S.-Z. Zhan, Z. Lv, G. Chen and D. Li, Chem. Eur. J., 2011, 17, 4113– 4117. 78 M. Veronelli, S. Dechert, S. Demeshko and F. Meyer, Inorg. Chem., 2015, 54, 6917– 6927. 79 P.-C. Duan, Z.-Y. Wang, J.-H. Chen, G. Yang and R. G. Raptis, Dalton Trans., 2013, 42, 14951–14954. 80 M. Grzywa, B. Bredenkötter, D. Denysenko, S. Spirkl, W. Nitek and D. Volkmer, Z. Für Anorg. Allg. Chem., 2013, 639, 1461–1471. 81 T. Jozak, Y. Sun, Y. Schmitt, S. Lebedkin, M. Kappes, M. Gerhards and W. R. Thiel, Chem. Eur. J., 2011, 17, 3384–3389. 82 J.-H. Wang, M. Li, J. Zheng, X.-C. Huang and D. Li, Chem. Commun., 2014, 50, 9115– 9118. 83 J. He, Y.-G. Yin, T. Wu, D. Li and X.-C. Huang, Chem. Commun., 2006, 2845–2847.

175

84 J.-P. Zhang and S. Kitagawa, J. Am. Chem. Soc., 2008, 130, 907–917. 85 Z. Wei, D. Yuan, X. Zhao, D. Sun and H.-C. Zhou, Sci. China Chem., 2013, 56, 418–422. 86 J.-H. Wang, M. Li and D. Li, Chem. Eur. J., 2014, 20, 12004–12008. 87 J.-X. Zhang, J. He, Y.-G. Yin, M.-H. Hu, D. Li and X.-C. Huang, Inorg. Chem., 2008, 47, 3471–3473. 88 L. Hou, W.-J. Shi, Y.-Y. Wang, H.-H. Wang, L. Cui, P.-X. Chen and Q.-Z. Shi, Inorg. Chem., 2011, 50, 261–270. 89 S.-Z. Zhan, M. Li, X.-P. Zhou, J.-H. Wang, J.-R. Yang and D. Li, Chem. Commun., 2011, 47, 12441–12443. 90 S.-Z. Zhan, M. Li, S. W. Ng and D. Li, Chem. Eur. J., 2013, 19, 10217–10225. 91 C. B. Aakeröy, A. S. Sinha, P. D. Chopade and J. Desper, Dalton Trans., 2011, 40, 12160–12168. 92 L. Hou, W.-J. Shi, Y.-Y. Wang, H.-H. Wang, L. Cui, P.-X. Chen and Q.-Z. Shi, Inorg. Chem., 2011, 50, 261–270. 93 S.-Z. Zhan, M. Li, X.-P. Zhou, J.-H. Wang, J.-R. Yang and D. Li, Chem. Commun., 2011, 47, 12441–12443. 94 S.-Z. Zhan, M. Li, S. W. Ng and D. Li, Chem. Eur. J., 2013, 19, 10217–10225. 95 A. Vogler and H. Kunkely, in Transition Metal and Rare Earth Compounds, ed. H. Yersin, Springer Berlin Heidelberg, 2001, pp. 143–182. 96 B. Hu, G. Gahungu and J. Zhang, J. Phys. Chem. A, 2007, 111, 4965–4973. 97 P. C. Ford, E. Cariati and J. Bourassa, Chem. Rev., 1999, 99, 3625–3648. 98 D. R. McMillin, J. R. Kirchhoff and K. V. Goodwin, Coord. Chem. Rev., 1985, 64, 83– 92. 99 J. V. Lockard, S. Kabehie, J. I. Zink, G. Smolentsev, A. Soldatov and L. X. Chen, J. Phys. Chem. B, 2010, 114, 14521–14527. 100 I. U. of P. and A. Chemistry, IUPAC Gold Book - exciplex, https://goldbook.iupac.org/html/E/E02246.html, (accessed 6 September 2017). 101 H.-K. Yip, H.-M. Lin, K.-K. Cheung, C.-M. Che and Y. Wang, Inorg. Chem., 1994, 33, 1644–1651. 102 D. R. Coulson, L. C. Satek and S. O. Grim, in Inorganic Syntheses, ed. R. J. Angelici, John Wiley & Sons, Inc., 1990, pp. 107–109. 103 3. Apex 2 Software Package, Bruker AXS Inc., Madison, WI, 2013. 104 G. M. Sheldrick, Acta Crystallogr. A, 2008, 64, 112–122. 105 The SHELX homepage, http://shelx.uni-ac.gwdg.de/SHELX/index.php, (accessed 17 May 2017). 106 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.

176

107 J.-R. Li, J. Yu, W. Lu, L.-B. Sun, J. Sculley, P. B. Balbuena and H.-C. Zhou, Nat. Commun., 2013, 4, 1538. 108 D. J. Darensbourg, W.-C. Chung, K. Wang and H.-C. Zhou, ACS Catal., 2014, 4, 1511– 1515. 109 S. Kumar, M. Y. Wani, C. T. Arranja, J. de A. e Silva, B. Avula and A. J. F. N. Sobral, J. Mater. Chem. A, 2015, 3, 19615–19637. 110 H. Huang, W. Zhang, D. Liu and C. Zhong, Ind. Eng. Chem. Res., 2012, 51, 10031– 10038. 111 B. Yuan, D. Ma, X. Wang, Z. Li, Y. Li, H. Liu and D. He, Chem. Commun., 2012, 48, 1135–1137. 112 O. Shekhah, Y. Belmabkhout, K. Adil, P. M. Bhatt, A. J. Cairns and M. Eddaoudi, Chem. Commun., 2015, 51, 13595–13598. 113 Z. Qiao, C. Peng, J. Zhou and J. Jiang, J. Mater. Chem. A, 2016, 4, 15904–15912. 114 S. Chaemchuen, N. A. Kabir, K. Zhou and F. Verpoort, Chem. Soc. Rev., 2013, 42, 9304– 9332. 115 R. Muñoz, L. Meier, I. Diaz and D. Jeison, Rev. Environ. Sci. Biotechnol., 2015, 14, 727– 759. 116 Environmental Protection Agency., Overview of Greenhouse Gases, https://www.epa.gov/ghgemissions/overview-greenhouse-gases#methane, (accessed 12 May 2017). 117 C. Ö. Karacan, F. A. Ruiz, M. Cotè and S. Phipps, Int. J. Coal Geol., 2011, 86, 121–156. 118 Upgrading Drained Coal Mine Methane to Pipeline Quality, EPA, 2008. 119 Best Practice Guidance for Effective Methane Drainage and Use in Coal Mines, UN, New York and Geneva, 2010. 120 F. Zhou, T. Xia, X. Wang, Y. Zhang, Y. Sun and J. Liu, J. Nat. Gas Sci. Eng., 2016, 31, 437–458. 121 J. Hu, T. Sun, X. Liu, S. Zhao and S. Wang, Microporous Mesoporous Mater., 2016, 225, 456–464. 122 M. Fischer, Microporous Mesoporous Mater., 2016, 219, 249–257. 123 Q. Li, M. Ruan, B. Lin, M. Zhao, Y. Zheng and K. Wang, J. Porous Mater., 2016, 23, 107–122. 124 J. Hu, T. Sun, X. Liu, Y. Guo and S. Wang, RSC Adv., 2016, 6, 64039–64046. 125 E. J. García, J. P. S. Mowat, P. A. Wright, J. Pérez-Pellitero, C. Jallut and G. D. Pirngruber, J. Phys. Chem. C, 2012, 116, 26636–26648. 126 H. Wu, W. Zhou and T. Yildirim, J. Am. Chem. Soc., 2009, 131, 4995–5000. 127 D. Yu, A. O. Yazaydin, J. R. Lane, P. D. C. Dietzel and R. Q. Snurr, Chem. Sci., 2013, 4, 3544.

177

128 B. Pato-Doldán, M. H. Rosnes and P. D. C. Dietzel, ChemSusChem, 2017, 10, 1710– 1719. 129 Q. Li, M. Ruan, B. Lin, M. Zhao, Y. Zheng and K. Wang, J. Porous Mater., 2016, 23, 107–122. 130 K. Barthelet, J. Marrot, D. Riou and G. Férey, Angew. Chem. Int. Ed., 2002, 41, 281–284. 131 N. Rosenbach, H. Jobic, A. Ghoufi, F. Salles, G. Maurin, S. Bourrelly, P. L. Llewellyn, T. Devic, C. Serre and G. Férey, Angew. Chem. Int. Ed., 2008, 47, 6611–6615. 132 G. Durá, M. C. Carrión, F. A. Jalón, B. R. Manzano and A. M. Rodríguez, Eur. J. Inorg. Chem., 2013, 2013, 5943–5957. 133 G. Durá, M. C. Carrión, F. A. Jalón, A. M. Rodríguez and B. R. Manzano, Cryst. Growth Des., 2013, 13, 3275–3282. 134 J. R. Gardinier, H. M. Tatlock, J. S. Hewage and S. V. Lindeman, Cryst. Growth Des., 2013, 13, 3864–3877. 135 F.-L. Li, Q. Chen, H.-B. Song, B. Dai and L.-F. Tang, Polyhedron, 2014, 83, 102–107. 136 D. L. Reger, A. E. Pascui, M. D. Smith, J. Jezierska and A. Ozarowski, Inorg. Chem., 2015, 54, 1487–1500. 137 E. Y. Semitut, V. Y. Komarov, E. Y. Filatov, A. S. Kuznetsova, A. I. Khlebnikov and A. S. Potapov, Inorg. Chem. Commun., 2016, 64, 23–26. 138 G. A. Santillan and C. J. Carrano, Inorg. Chem., 2008, 47, 930–939. 139 W. M. Bloch, R. Babarao, M. R. Hill, C. J. Doonan and C. J. Sumby, J. Am. Chem. Soc., 2013, 135, 10441–10448. 140 C. J. Coghlan, C. J. Sumby and C. J. Doonan, CrystEngComm, 2014, 16, 6364–6371. 141 W. M. Bloch, A. Burgun, C. J. Coghlan, R. Lee, M. L. Coote, C. J. Doonan and C. J. Sumby, Nat. Chem., 2014, 6, 906–912. 142 W. M. Bloch, A. Burgun, C. J. Doonan and C. J. Sumby, Chem. Commun., 2015, 51, 5486–5489. 143 M. W. Rathke and P. J. Cowan, J. Org. Chem., 1985, 50, 2622–2624. 144 J.-C. Jung, E. B. Watkins and M. A. Avery, Tetrahedron, 2002, 58, 3639–3646. 145 S. J. Ashcroft, J. Chem. Soc. Inorg. Phys. Theor., 1970, 0, 1020–1024. 146 C. M. Simon, B. Smit and M. Haranczyk, Comput. Phys. Commun., 2016, 200, 364–380. 147 A. L. Myers and J. M. Prausnitz, AIChE J., 1965, 11, 121–127. 148 Y.-P. Zhao, Y. Li, C.-Y. Cui, Y. Xiao, R. Li, S.-H. Wang, F.-K. Zheng and G.-C. Guo, Inorg. Chem., 2016, 55, 7335–7340. 149 E. Sánchez-González, E. González-Zamora, D. Martínez-Otero, V. Jancik and I. A. Ibarra, Inorg. Chem., 2017, 56, 5863–5872. 150 L. D. Field, B. A. Messerle, M. Rehr, L. P. Soler and T. W. Hambley, Organometallics, 2003, 22, 2387–2395.

178

151 J.-S. Lee, J.-H. Kim, J.-T. Kim, J.-K. Suh, J.-M. Lee and C.-H. Lee, J. Chem. Eng. Data, 2002, 47, 1237–1242. 152 J. Kärger, D. M. Ruthven and D. N. Theodorou, in Diffusion in Nanoporous Materials, Wiley-VCH Verlag GmbH & Co. KGaA, 2012, pp. 191–226. 153 F.-X. Coudert, in Metal-Organic Frameworks, ed. D. Farrusseng, Wiley-VCH Verlag GmbH & Co. KGaA, 2011, pp. 49–68. 154 W. Smith and T. R. Forester, J. Mol. Graph., 1996, 14, 136–141. 155 S. L. Mayo, B. D. Olafson and W. A. Goddard, J. Phys. Chem., 1990, 94, 8897–8909. 156 A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard and W. M. Skiff, J. Am. Chem. Soc., 1992, 114, 10024–10035. 157 C. Campañá, B. Mussard and T. K. Woo, J. Chem. Theory Comput., 2009, 5, 2866–2878. 158 G. Kresse and J. Hafner, Phys. Rev. B, 1993, 47, 558–561. 159 G. Kresse and J. Hafner, Phys. Rev. B, 1994, 49, 14251–14269. 160 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868. 161 M. G. Martin and J. I. Siepmann, J. Phys. Chem. B, 1998, 102, 2569–2577. 162 J. J. Potoff and J. I. Siepmann, AIChE J., 2001, 47, 1676–1682. 163 Q. Sun, M. Liu, K. Li, Y. Han, Y. Zuo, J. Wang, C. Song, G. Zhang and X. Guo, Dalton Trans., 2016, 45, 7952–7959. 164 F. Luo, Y. Che and J. Zheng, Cryst. Growth Des., 2008, 8, 176–178. 165 Z. Chang, D.-S. Zhang, T.-L. Hu and X.-H. Bu, Cryst. Growth Des., 2011, 11, 2050– 2053. 166 C. Shen, T. Sheng, Q. Zhu, S. Hu and X. Wu, CrystEngComm, 2012, 14, 3189–3198. 167 E. E. Moushi, A. Kourtellaris, I. Spanopoulos, M. J. Manos, G. S. Papaefstathiou, P. N. Trikalitis and A. J. Tasiopoulos, Cryst. Growth Des., 2015, 15, 185–193. 168 D. Chisca, L. Croitor, E. B. Coropceanu, O. Petuhov, S. G. Baca, K. Krämer, S.-X. Liu, S. Decurtins, H. J. Rivera-Jacquez, A. E. Masunov and M. S. Fonari, CrystEngComm, 2016, 18, 384–389. 169 P. K. Capon, A. Burgun, C. J. Coghlan, R. S. Crees, C. J. Doonan and C. J. Sumby, CrystEngComm, 2016, 18, 7003–7010. 170 Q.-H. Tan, Y.-Q. Wang, H.-T. Liu and Z.-L. Liu, Inorg. Chem. Commun., 2015, 58, 67– 70. 171 Y. Wang, H. Li, L. Gu, Q. Gan, Y. Li and G. Calzaferri, Microporous Mesoporous Mater., 2009, 121, 1–6. 172 K. Sheng, B. Yan, H.-F. Lu and L. Guo, Eur. J. Inorg. Chem., 2010, 2010, 3498–3505. 173 P. He, H. H. Wang, H. G. Yan, W. Hu, J. X. Shi and M. L. Gong, Dalton Trans., 2010, 39, 8919–8924.

179

174 G. Zhang, H. Li, F. Zhao, H. Hu, H. Huang, H. Li, X. Han, R. Liu, H. Dong, Y. Liu and Z. Kang, Dalton Trans., 2013, 42, 9423–9427. 175 S. Sha, H. Yang, J. Li, C. Zhuang, S. Gao and S. Liu, Catal. Commun., 2014, 43, 146– 150. 176 R. Decadt, K. Van Hecke, D. Depla, K. Leus, D. Weinberger, I. Van Driessche, P. Van Der Voort and R. Van Deun, Inorg. Chem., 2012, 51, 11623–11634. 177 K. Binnemans, Chem. Rev., 2009, 109, 4283–4374. 178 C. Pettinari and R. Pettinari, Coord. Chem. Rev., 2005, 249, 525–543. 179 H. R. Bigmore, S. C. Lawrence, P. Mountford and C. S. Tredget, Dalton Trans., 2005, 635–651. 180 A. C. Gomes, P. Neves, S. Figueiredo, J. A. Fernandes, A. A. Valente, F. A. Almeida Paz, M. Pillinger, A. D. Lopes and I. S. Gonçalves, J. Mol. Catal. Chem., 2013, 370, 64– 74. 181 P. Chaudhuri, K. Weighardt, Y. H. Tsai and C. Krueger, Inorg. Chem., 1984, 23, 427– 432. 182 A. Hroch, G. Gemmecker and W. R. Thiel, Eur. J. Inorg. Chem., 2000, 2000, 1107–1114. 183 C. J. Coghlan, C. J. Sumby and C. J. Doonan, CrystEngComm, 2014, 16, 6364–6371. 184 D. L. Reger, R. P. Watson, M. D. Smith and P. J. Pellechia, Organometallics, 2005, 24, 1544–1555. 185 D. L. Reger, R. P. Watson and M. D. Smith, Inorg. Chem., 2006, 45, 10077–10087. 186 A. S. Potapov, E. A. Nudnova, A. I. Khlebnikov, V. D. Ogorodnikov and T. V. Petrenko, Inorg. Chem. Commun., 2015, 53, 72–75. 187 N. Dehury, N. Maity, S. K. Tripathy, J.-M. Basset and S. Patra, ACS Catal., 2016, 6, 5535–5540. 188 D. L. Reger, R. P. Watson, M. D. Smith and P. J. Pellechia, Cryst. Growth Des., 2007, 7, 1163–1170. 189 L. Zhang, Z.-G. Ren, H.-X. Li and J.-P. Lang, CrystEngComm, 2011, 13, 1400–1405. 190 A. Beheshti, A. Lalegani, A. Noshadian, G. Bruno and H. A. Rudbari, Polyhedron, 2015, 85, 690–696. 191 A. Beheshti, F. Safaeiyan, F. Hashemi, H. Motamedi, P. Mayer, G. Bruno and H. A. Rudbari, J. Mol. Struct., 2016, 1123, 225–237. 192 M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O’Keeffe and O. M. Yaghi, Science, 2002, 295, 469–472. 193 C. B. Aakeröy, A. S. Sinha, P. D. Chopade and J. Desper, Dalton Trans., 2011, 40, 12160–12168. 194 L. J. Charbonnière and R. Ziessel, Tetrahedron Lett., 2003, 44, 6305–6307. 195 D. L. Reger, J. R. Gardinier, R. F. Semeniuc and M. D. Smith, Dalton Trans., 2003, 1712–1718.

180

196 D. L. Reger, T. C. Grattan, K. J. Brown, C. A. Little, J. J. S. Lamba, A. L. Rheingold and R. D. Sommer, J. Organomet. Chem., 2000, 607, 120–128. 197 L. J. Charbonnière and R. Ziessel, Tetrahedron Lett., 2003, 44, 6305–6307. 198 J.-X. Wang, Z.-R. Zhu, F.-Y. Bai, X.-Y. Wang, X.-X. Zhang and Y.-H. Xing, Polyhedron, 2015, 99, 59–70.

181

About the Author

Biography:

Charlie Kivi was born in Thunder Bay, Ontario. He attended the University of Ottawa graduating magna cum laude, Bachelor of Science with Honours, Specialization in Chemistry, in 2011. Charlie’s undergraduate thesis work concerned the investigation of iron and tungsten complexes for the catalytic synthesis of fluoroalkanes under the supervision of Professor R. Tom Baker. While an undergraduate student, Charlie had the opportunity to study MOFs under the supervision of Professor George Shimizu at the University of Calgary as part of the 2010 Inorganic Chemistry Exchange Program. This ignited his interest in these versatile compounds.

Charlie obtained his PhD in 2017 advised by Professor Datong Song at the University of Toronto. While at U of T, Charlie had the opportunity to learn x-ray crystallography. This fueled a consuming, and at times confounding, interest in crystal structures with high degrees of structural disorder. Additionally, Charlie was able to hone his air free synthetic techniques and develop his own apparatus for testing VOCs. When not in the lab, Charlie can be found fishing, on the hiking trail, stalking birds with a camera, or playing board games with his friends.

Publications:

1. Charlie E. Kivi, Datong Song, “Three New Metal Complexes for Catalytic Transformations”, In Preparation. 2. Charlie E. Kivi, Cindy Ma, Datong Song, “Bis(3,5-dimethyl-4-carboylic acid)pyrazolylmethe as a Versatile Ligand for MOF and Coordination Polymer Synthesis” In Preparation. 3. Charlie E. Kivi, Benjamin Gelfand, Hana Dureckova, Cindy Ma, Datong Song, George K. H. Shimizu, Tom Woo, “Coal Mine Methane Capture with a New Nickel MOF” In Preparation. 4. Charlie E. Kivi, Datong Song, “A luminescent cationic metal–organic framework

featuring [Cu–pyrazolate]3 units for volatile organic compound sensing” Dalton Transactions, 2016, 45, 17087-17090.

182

Conference Presentations:

1. Charlie E. Kivi, Benjamin Gelfand, Hana Dureckova, Cindy Ma, Datong Song, George K. H. Shimizu, Tom Woo, “Coal Mine Methane Recovery using MOFs”, Poster Presentation, 100th Canadian Chemistry Conference and Exhibition (CSC), Toronto, Ontario, May 28-June 1, 2017 2. Charlie E. Kivi, Datong Song, “Metal-Organic Frameworks of Pyrazolate Derivatives”, Poster Presentation, 100th Canadian Chemistry Conference and Exhibition (CSC), Toronto, Ontario, May 28-June 1, 2017 3. Charlie E. Kivi, Datong Song, “Metal-Organic Frameworks of Pyrazole Derivatives”, Poster Presentation, 98th Canadian Chemistry Conference and Exhibition (CSC), Ottawa, Ontario, June 13-17, 2015 4. Charlie E. Kivi, “Chemistry of Good Communication”, Oral Presentation, 23rd IUPAC International Conference on Chemical Education, Toronto, Ontario, July 13-18, 2014 5. Charlie E. Kivi, Datong Song, “MOF Catalysis using Scorpionate Metalloligands”, Poster Presentation, 45th Inorganic Discussion Weekend, Ottawa, Ontario, November 2-4, 2012.

Appendices A Supplemental Information A.1 Chapter 2

183 184

Supplemental Figure A.2: CIE colour representation of ethyl acetate vapour treated 1 with an excitation wavelength of 469 nm.

185

A.2 Chapter 3

Supplemental Figure A.4: Typical MOF reaction setup with 2 dram vials in an aluminum heating block. A similar setup was used for larger, 20 mL vials for larger scale synthesis.

Supplemental Figure A.5: Schematic representation of coal mine methane (CMM) extraction using pressure swing adsorption.

186

A.3 Chapter 4

The infrared spectra of 7 to 13 were recorded using a neat sample on a Bruker Alpha FT-IR spectrometer equipped with a Platinum ATR sampling unit in air.

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.6: IR spectrum of [Co2(BPM)2(H2O)4(Bipy)0.5]n∙2DMF∙2(H2O) (7).

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.7: IR spectrum of [Co2(BPM)2(H2O)4(Bipy)0.5]n∙2DMSO∙2(H2O) (8).

187

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.8: IR spectrum of [Co2(BPM)2(H2O)4(DMF)3]n∙DMF (9).

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.9: IR spectrum of [Co(BPM)(H2O)(DMSO)]n (10).

188

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.10: IR spectrum of [Mn(BPM)(H2O)2]n (11).

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.11 IR spectrum of [Cu(BPM)(DMSO)]n (12).

189

Wave Number (cm-1)

4000 3600 3200 2800 2400 2000 1600 1200 800 400 100%

90%

80%

70%

PercentTransmission 60%

50%

Supplemental Figure A.12: IR spectrum of [Eu(BPM)2(OAc)]n (13).

Supplemental Figure A.13: CIE colour representation of as synthesized 13 with an excitation wavelength of 393 nm.

A.4 Chapter 5

Metal-carbonyl stretching frequencies were collected on a Perkin Elmer Spectrum One FT-IR spectrometer. Samples were prepared as pressed KBr disks.

190

Supplemental Table A.1: Metal-Carbonyl stretching frequencies for 14, 14-(COOH)3, and 15 demonstrating the similarity of the molybdenum coordination environment.

-1 Species Metal-CO stretching frequencies (cm )

Free CO 2149

14 1912

14-(COOH)3 1910

15 1909

B The Song Group: 2011 through 2017

Supplemental Figure B.1: The Song Group Spring 2012.

191

Supplemental Figure B.2: The Song Group Fall 2012.

Supplemental Figure B.3: The Song Group Fall 2013.

192

Supplemental Figure B.4: The Song Group Spring 2015.

Supplemental Figure B.5: The Song Group Fall 2015.

193

Supplemental Figure B.6: The Song Group Spring 2017.

Supplemental Figure B.7: The Song Group (plus alumnus Vince) Summer 2017.