<%BANNER%>

Development of Nonadentate Ligands for the Selective Separation of Lanthanides and Actinides

Permanent Link: http://ufdc.ufl.edu/UFE0042507/00001

Material Information

Title: Development of Nonadentate Ligands for the Selective Separation of Lanthanides and Actinides A Computational and Synthetic Investigation
Physical Description: 1 online resource (106 p.)
Language: english
Creator: Hyatt, Ivan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: actinides, dft, effective, lanthanides, molecular, semi, sparkle, triphenoxymethane
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: To meet the energy demands of the future, nuclear power provides one of the best options for providing electricity. The encumbrance that nuclear power bears is that the radioactive waste must be stored for many thousands of years. To counteract the problems associated with radioactive waste, the actinides can be separated from the waste stream in order to decrease waste volume and, after transmutation, decrease storage time of the hazardous materials. Transmutation of actinides is hindered by the presence of lanthanides lowering the efficiency of the process. The following work presents ligands that can separate these two groups of metals from each other. The design of the ligands is based on radioactive stability, hard-soft acid-base principles, and the use of environmentally friendly elements. Tridentate ligands with nitrogen donor groups can then be attached to a triphenoxymethane scaffold to make a rigid structure that bonds metals in a nonadentate fashion. Several computational methods were also used to investigate the geometric and electronic structure of the metal complexes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ivan Hyatt.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Scott, Michael J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042507:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042507/00001

Material Information

Title: Development of Nonadentate Ligands for the Selective Separation of Lanthanides and Actinides A Computational and Synthetic Investigation
Physical Description: 1 online resource (106 p.)
Language: english
Creator: Hyatt, Ivan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: actinides, dft, effective, lanthanides, molecular, semi, sparkle, triphenoxymethane
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: To meet the energy demands of the future, nuclear power provides one of the best options for providing electricity. The encumbrance that nuclear power bears is that the radioactive waste must be stored for many thousands of years. To counteract the problems associated with radioactive waste, the actinides can be separated from the waste stream in order to decrease waste volume and, after transmutation, decrease storage time of the hazardous materials. Transmutation of actinides is hindered by the presence of lanthanides lowering the efficiency of the process. The following work presents ligands that can separate these two groups of metals from each other. The design of the ligands is based on radioactive stability, hard-soft acid-base principles, and the use of environmentally friendly elements. Tridentate ligands with nitrogen donor groups can then be attached to a triphenoxymethane scaffold to make a rigid structure that bonds metals in a nonadentate fashion. Several computational methods were also used to investigate the geometric and electronic structure of the metal complexes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ivan Hyatt.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Scott, Michael J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042507:00001


This item has the following downloads:


Full Text

PAGE 1

1 DEVELOPMENT OF NONAD ENTATE LIGANDS FOR T HE SELECTIVE SEPARATION OF LANTHANIDES AND A CTINIDES: A COMPUTAT IONAL AND SYNTHETIC INVESTIGATION By IVAN FABE DEMPSEY HYATT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Ivan Fabe Dempsey Hyatt

PAGE 3

3 To Elisa and my family

PAGE 4

4 ACKNOWLEDGMENTS I would first like to thank my par ents and family for the support they have given me in my scholastic journey I would not be at this level of academics without their help and most importantly their faith that I could accomplish my goals. I thank my mother, Elaine Hyatt, for motivating me to strive for the best and teaching me how to focus my mind on academics like only a teacher can. I am very thankful for my father, Andrew Hyatt, for instilling the drive to take pride in my work and how to excel through competition. I would also like t o thank my extended family, especially James and Alexander Hyatt, for always giv ing me something to smile about. I could have never succeeded in academics without the support of my significant other (p < 0.0001) Elisa Livengood. She has contributed more to my success than any one person has. Her academic competitiveness has propelled us both and our mutual love for science is nearly as great as our love for each other. Eternal appreciation is given to my high school calculus and physics teacher, Mr. Greg Hardin, who taught me the level of commitment that is needed to excel in school and to my high school chemistry teacher, Mr Elton Cavi ness, who first inspired me to pursue the infinite possibilities of a career in chemistry. I thank my undergraduate advis er, Dr. Andrew Sargent, for encouraging me to go to graduate school, and teaching me how to undertake computational chemistry research. The skill s he taught have been invaluable to my research. I also thank Dr. Tammy Davidson for always giving me the oppo rtunity to teach organic chemistry lab and Dr. Lisa McElwee White for letting me do computational work with her. I thank Dr. Ben Smith for helping me transition to a new research group.

PAGE 5

5 The last five years have been challenging but my friends are always th ere to help. I thank John Deaton, Rusty Coco, Mark Moseley, Benjamin Baldwin and Mike Hoose for letting me escape chemistry even if it was brief venture back to North Carolina I also thank my fellow Scott group members, Patrick Hillesheim, Gary Guillet, A nna Sberega e va, Nathan Strutt and Candace Z i elenuik, for teaching me proper synthetic techniques and always listening to my off the wall ideas. I also have much appreciation and respect for my friends in the inorganic division especially Matt Jeletic, Soumya Sarkar and Shreya Mukherjee. In closing, I wish to express gratitude to my adviser, Dr. Mike Scott. I will always remember his unique way of teaching and try to use what I have learned from him in my own career. I am most thankful tha t he accepte d me into the group at a time of great need. I appreciate the independence and trust he gave me in exploring my own computational research and for contributing to my knowledge of chemistry as a whole.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 13 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 1.1 The Current Status of Nuclear Power ................................ ............................... 16 1.2 Nuclear Waste Reprocessing ................................ ................................ ........... 16 1.2.1 Plutonium Uranium Recovery by Extraction Method ............................... 18 1.2.2 Transuranium Extraction Method ................................ ............................ 19 1.2.3 Diamide Extraction Method ................................ ................................ ..... 20 1.2.4 Selective Actinide Extraction Method ................................ ...................... 21 1.3 Separation of Lanthanides and Actinides ................................ .......................... 22 1.4 Benefits of Preorganization on a Molecular Platf orm ................................ ........ 23 1.5 Experimental Research Objectives ................................ ................................ ... 26 1.6 Computation Research Objectives ................................ ................................ ... 27 1.6.1 Semi Empirical Calculations ................................ ................................ .... 28 1.6.2 Density Functional Theory Calculations ................................ .................. 29 2 STRUCTURAL ANALYSIS OF NONADENTATE LANTHANIDE COMPLEXES USING SEMI EMPIRICAL METHODS ................................ ................................ ... 31 2.1 Introduction ................................ ................................ ................................ ....... 31 2.2 Computed Structures ................................ ................................ ........................ 32 2.3 Structural Evaluation ................................ ................................ ......................... 36 2. 4 Lanthanide Contraction ................................ ................................ ..................... 38 2.4 Conclusions ................................ ................................ ................................ ...... 40 2.5 Computational Details ................................ ................................ ....................... 41 3 SYNTHESIS OF BENZIMIDAZOLE PRIDYL FUNCTIONALIZED NONADENTATE LIGANDS ................................ ................................ .................... 42 3.1 Introduction ................................ ................................ ................................ ....... 42 3.2 Synthetic Difficulties of Ether Linkage Ligands ................................ ................. 42 3.3 Synthesis of Amide Linkage Ligands ................................ ................................ 45 3.3.1 Carboxylic Acid Functionalization of the Ligand Arm ............................... 46

PAGE 7

7 3.3.2 Incorporation of the Amide Linkage as a Donor Group for Metal Binding ................................ ................................ ................................ .......... 48 3.4 Synthesis of Triazole Linkage Ligands ................................ .............................. 49 3.5 Conclusions ................................ ................................ ................................ ...... 50 3.6 Experimental Section ................................ ................................ ........................ 52 methyl 2 (6 (diethylcarbamoyl)pyridin 2 yl) 1 ethyl 1H benzo[d]imidazole 5 carboxylate (3 21): ................................ ................................ ........................ 52 2 (6 (diethylcarbamoyl)pyridin 2 yl) 1 ethyl 1H benzo[d]imidazole 5 carboxylic acid (3 20): ................................ ................................ ................... 53 methyl 6 (ethyl(4 (methoxymethyl) 2 nitrophenyl)carbamoyl)picolinate (3 23): ................................ ................................ ................................ ................ 54 6 (1 ethyl 5 (methoxymethyl) 1H benzo[d]imidazol 2 yl)picolinic acid (3 25): .. 54 2 (6 bromopyridin 2 yl) 1 methyl 1H benzo[d]imidaz ole (3 30): ....................... 55 1 methyl 2 (6 ((trimethylsilyl)ethynyl)pyridin 2 yl) 1H benzo[d]imidazole (3 31a): ................................ ................................ ................................ .............. 56 2 (6 ethynylpyridin 2 yl) 1 methyl 1H benzo[d]imidazole (3 31b): .................... 56 Tris(2 ethylacetoxy 3 methyl 5 tert pentylphenyl)methane (3 14c): ................. 57 Tris[2 (2 hydroxylethoxy) 3 methyl 5 tert pentylphenyl]methane (3 15c): ........ 57 Tris[2 (2 toluenesulfonlyethoxy) 3 methyl 5 tert pent ylphenyl]methane (3 16c): ................................ ................................ ................................ .............. 58 Tris (2 (2 azidoethoxy) 3,5 di tert pentylphenyl)methane (3 32): ...................... 59 tris(2 (2 (4 (6 (1 methyl 1H benzo[d]imidazol 2 yl)pyridin 2 yl) 1H 1,2,3 triazol 1 yl)ethoxy) 3,5 di tert pentylphenyl)methane (3 33): ........................ 59 General Procedure for the Synthesis of 3 22 and 3 26 ................................ .... 60 4 METAL COMPLEXATION EXPERIMENTS OF LANTHANIDES USING BENZIMIDAZOLE PRIDYL FUNCTIONALIZED NONADENTATE LIGANDS ........ 62 4.1 Introduction ................................ ................................ ................................ ....... 62 4.2 Extracti on Results ................................ ................................ ............................. 62 4.3 Crystal Structure ................................ ................................ ............................... 64 4.4 Conclusions ................................ ................................ ................................ ...... 67 4.5 Experimental Section ................................ ................................ ........................ 69 4.5.1 Extraction Experiment ................................ ................................ ............. 69 4.5.2 Crystal Structure of Yb(3 26) 3+ ................................ ................................ 69 4.5.2 Crystal Structure of Yb(3 33) 3+ ................................ ................................ 70 5 ELECTRONIC STRUCTURE OF EUROPIUM, AMERICIUM, AND CURIUM COMPLEXES USING DENSITY FUNCTIONAL THEORY ................................ ..... 73 5.1 Introduction ................................ ................................ ................................ ....... 73 5.2 Geometry Optimizations ................................ ................................ ................... 74 5.3 Electronic Structure Analysis ................................ ................................ ............ 77 5.3.1 Mulliken and Natural Charge Analysis ................................ ..................... 77 5.3.2 Molecular Orbital Analysis ................................ ................................ ....... 78 5.4 Conclusions ................................ ................................ ................................ ...... 88 5.5 Computational Details ................................ ................................ ....................... 90

PAGE 8

8 6 SUMMARY ................................ ................................ ................................ ............. 92 APPENDIX : NUCLEAR MAGNETIC RESONANCE SPECTRA OF SYNTHESIZED COMPOUNDS ................................ ................................ ................................ ........ 95 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 106

PAGE 9

9 LIST OF TABLES Table page 2 1 Average Metal Bond Lengths () and Twist Angles ( ) of Amide Structures ...... 36 2 2 Average Metal Bond Lengths () and Twist Angles ( ) of Triazole Structures .... 37 4 1 Extraction Efficiency ................................ ................................ ........................... 63 4 2 Comparison of bond lengths () and twist angles ( ) between the crystal structure, SE calculation, and crystal SE calculation of Yb(3 26) ....................... 67 4 3 Comparison of bond lengths () and twist angles ( ) between the crystal structure, SE calculation, and crystal SE calculation of Yb(3 33) ....................... 68 4 4 X ray data for crystal structures Yb complexed by 3 22 and 3 33 using triflate counter ions grown from the diffusion of ether into methano l ............................. 72 5 1 Calculated M N bond distances for structure 5 1 ................................ ................ 75 5 2 Calculated M N bond distances for structure 5 2 ................................ ................ 76 5 3 Calculated M N bond distances for structure 5 3 ................................ ................ 76 5 4 Comparison of Mulliken Charges for ML 3 3+ ................................ ........................ 77 5 5 Orbital compositions (%) of bonding interactions ................................ ................ 89

PAGE 10

10 LIST OF FIGURES Figure page 1 1 The nuclear fuel cycle. Filled boxes represent the currently used open fuel cycle and unfilled boxes represent the proposed closed fuel cycle. ................... 17 1 3 Structure of tributylphosphate (TBP), an extractant used in the PUREX process ................................ ................................ ................................ ............... 19 1 4 Structure of N,N d iisobutylcarbamoyl methyl octylphenyl phosphineoxide (CM PO), an extractant used in the TRUEX process ................................ ........... 20 1 5 Structure of DMDOHEMA and DMDBTDMA, extractants used in the DIAMEX process ................................ ................................ ................................ ............... 20 1 6 Structure of Cyanex 301 and BTP, extractants used in the TRUEX process. .... 22 1 7 Tricapped trigonal prismatic (TTP) geometry ................................ ...................... 23 1 8 A simple calixarene. Functionalities can be added to bolster metal extraction. .. 24 1 9 CMPO moieties attached to a calixarene platform ................................ .............. 24 1 10 Triphenoxymethane molecular platform ................................ ............................. 26 1 11 DGA attached to triphenoxymethane ................................ ................................ .. 26 1 12 Common moiety used in all s refer to the various linkages, donor groups and solubility modifiers. .................. 27 2 1 Semi empirical calculation structures ................................ ................................ 33 2 2 Semi empirical geometry optimization of structure 2 7 ................................ ....... 34 2 3 Semi empirical geometry optimization of structure 2 8 ................................ ....... 35 2 4 The graph shows the metal to donor atom bond lengths for structure 2 4. ......... 39 2 5 The graph shows the metal to donor atom bond lengths for structure 2 7. ......... 39 3 1 Tripodal helicate ligand developed by Piguet et al. ................................ ............. 43 3 2 Synthesis of liga nd arm 3 10, which contains the chloride functionality needed to attach triphenoxymethane via Williamson ether synthesis. ................ 43 3 3 Synthesis of two carbon linker triphenoxymethane. ................................ ............ 45 3 4 Synthesis of the three carbon linker amine triphenoxymethane, 3 18 ................ 46

PAGE 11

11 3 5 Synthesis of carboxylic acid functionalized ligand arm, 3 20 .............................. 46 3 6 Synthesis of the nonadentate ligand, 3 22 ................................ ......................... 47 3 7 Synthesis of the carboxylic acid functionalized benzimidazole pridyl ligand arm, 3 25 ................................ ................................ ................................ ............ 48 3 8 Synthesis of the nonadentate ligand, 3 26 ................................ ......................... 48 3 9 Synthesis of t he alkyne ligand arm, 3 31 ................................ ............................ 49 3 10 Synthesis of the nonadentate triazole ligand, 3 33 ................................ ............. 50 4 1 Crystal structure of Yb complexed by ligand 3 26 with ellipsoids at 50% probability ................................ ................................ ................................ ........... 65 4 2 Crystal structure of Yb complexed by ligand 3 33 with ellipsoids at 50% probability ................................ ................................ ................................ ........... 66 5 1 Simplified ligands computed by DFT. Each ligand was computed with Eu, Am, and Cm as the metal. ................................ ................................ .................. 75 5 2 Representation of benzimidazole centered HOMO of Cm(5 3) 3 3+ ...................... 79 5 3 Representation of ligand to metal donation in the HOMO 9 of Am (5 1) 3 3+ ......... 80 5 4 The HOMO 1 of Am(5 1) 3 3+ The black lines signify anti bonding interactions .. 81 5 5 The HOMO 12 of Am(5 1)) 3 3+ does not display any overlap of orbitals since the overlap population was computed as zero. ................................ .................. 82 5 6 MO for Eu(5 1) 3 3+ Am(5 1) 3 3+ and Cm(5 1) 3 3+ Metal energy levels shifted by approximately 18 eV and ligand orbitals by 8 eV. ................................ .............. 83 5 7 Degenerate orbitals HOMO 43 and HOMO 44 of Cm(5 1) 3 3+ at 17.803 eV ...... 84 5 8 MO for Eu(5 2) 3 3+ Am(5 2) 3 3+ and Cm(5 2) 3 3+ Metal energy levels shifted by approximately 18 eV and ligand orbitals by 8 eV. ................................ .............. 85 5 9 Representation of ligand to metal donation in the HOMO 36 of Cm(5 2) 3 3+ ....... 86 5 10 MO for Am(5 3) 3 3+ and Cm(5 3) 3 3+ Metal energy levels shifted by approximately 18 eV and ligand orbitals by 8 eV. ................................ .............. 87 A 1 NMR of 3 21 ................................ ................................ ................................ ....... 95 A 2 NMR of 3 20 ................................ ................................ ................................ ....... 96 A 3 NMR of 3 23 ................................ ................................ ................................ ....... 96

PAGE 12

12 A 4 NMR of 3 25 ................................ ................................ ................................ ....... 97 A 5 NMR of 3 30 ................................ ................................ ................................ ....... 97 A 6 NMR of 3 31b ................................ ................................ ................................ ..... 98 A 7 NMR of 3 14c ................................ ................................ ................................ ..... 98 A 8 NMR of 3 15c ................................ ................................ ................................ ..... 99 A 9 NMR of 3 33 ................................ ................................ ................................ ....... 99 A 10 NMR of 3 22 ................................ ................................ ................................ ..... 100 A 11 NMR of 3 26 ................................ ................................ ................................ ..... 100

PAGE 13

13 LIST OF ABBREVIATION S 6 31G* Gaussian basis set with diffuse functions on atoms other than hydrogen AM1 Austin model 1, semi empirical method Am(III) Trivalent americium An(III) Trivalent actinides B3LYP Becke 3 parameter, Lee, Yang, Parr functional BTP bis triazinyl pyridine CMPO N,N d iisobutylcarbamoyl methyl octy lphenyl phosphineoxide Cyanex 301 (bis(2,4,4 trimeth ylpentyl)dithiophosphinic acid) DCM Dichloromethane DIAMEX Diamide Extraction DFT Density functional theory DGA D igycolamides DMDBTDMA D imethyl dibutyl tetradecyl malonamide DMDOHEMA D imethyl dioctyl hexy lethoxy malonamide DMF N,N dimethyl formamide %E Extraction efficiency ECP Effective core potential ESI MS Electrospray mass spectrometry Eu(III) Trivalent europium HOFO Highest occupied fragment orbital HOMO Highest occupied molecular orbital L Ligand LAN L2DZ Los Alamos national lab double zeta density functional basis set Ln(III) Trivalent lanthanides

PAGE 14

14 LUFO Lowest unoccupied fragment orbital LUMO Lowest unoccupied molecular orbital M Metal MO Molecular orbital MOPAC Molecular Orbital Package NMR Nuclear magnetic resonance spectroscopy PM3 Parameterized model number 3, semi empirical method PM6 Parameterized model number 6, semi empirical method PUREX Plutonium uranium extraction PyBOP benzotriazol 1 yl oxytripyrrolidinophosphonium hexafluorophosphate RECP Relativistic effective core potential SANEX Selective actinide extraction SCF Self consistent field SE Semi empirical S N 2 bimolecular substitution reaction SPARKLE A semi empirical method for treating trivalent lanthnides. The + o xidation state of lanthanides TBP Tributyl Phosphate THF Tetrahydrofuran Tf Trifluoromethanesulfonate aka Triflate TRUEX Trans uranium extraction TTP Tricapped trigonal prismatic geometry

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the U niversity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DEVELOPMENT OF NONAD ENTATE LIGANDS FOR T HE SELECTIVE SEPARAT ION OF LANTHANIDES AND A CTINIDES: A COMPUTAT IONAL AND SYNTHETIC INVESTIGATION By Ivan Fabe Demps ey Hyatt December 2010 Chair: Mike Scott Major: C hemistry To meet the energy demands of the future nuclear power provides one of the best options for providing electri city. The encumbrance that nuclear power bears is that the radioactive waste must be stored for many thousands of years. To counteract the problems associated with radioactive waste the actinides can be separ ated from the waste stream in order to decrease waste volume and, after transmutation, decrease s torage time of the hazardous materials Transmutation of actinides is hindered by the presence of lanthanides lowering the efficiency of the process. The following work presents ligands that can separate these two groups of metals from each other. The desi gn of the ligands is based on radioactive stability, hard soft acid base principles, and the use of environmentally friendly eleme nts. Tridentate ligands with nitrogen donor groups can then be attached to a triphenoxymethane scaffold to make a rigid struct ure that bonds metals in a nonadentate fashion Several computational methods were also used to investigate the geometric and electronic structure of the metal complexes.

PAGE 16

16 CHAPTER 1 INTRODUCTION 1.1 The Current Status of Nuclear Power about 441 nuclear reactors. In the United States 103 of these reactors provide us with 20% of our electricity. 1 By the year 2030 there will be a 57% increase in worldwide energy consumption with a significant fraction coming from developing nations. 2 Curren tly, the wo rld does not have enough power plants available to generate the required electricity and with the rising cost of oil and concern about carbon dioxide emissions, the burning of fossil fuels may not be the best method to supply the energy demand. In contrast there is an ample supply of uranium to fuel nuclear power and the entire process from mining to electricity distribution, generates less greenhouse gas than burning fossil fuels. The main detriment to nuclear power is the generation of radioactive waste which must be disposed of and stored in order to prevent contamination. 3 1.2 Nuclear Waste Reprocessing Since 1968 the United States as produced over 47,000 metric tonnes of spent fuel. 4 Without a national repository the radioactive waste is usually stored onsite at the reactor facility fr om which it was generated. In countries such as France nuclear power provides around 80% of the electricity and produc es approximately 1,135 tonnes of waste each year, but they have developed reprocessing methods that can recover reusable materials thus lowering the volume of waste. The commercial reprocessing plant in La Hague can reprocess up to 1,700 tonnes of spent fu el each year. 5 A simplified closed fuel cycle is shown in Figure 1 1. 6 All current methods being employed

PAGE 17

17 do not separate the radiotoxic minor actinides (Np, Am, Cm) from the waste stream. These minor actinides account for the most dangerous waste source after removal of U and Pu. 7 Figure 1 1. The nuclear fuel cycle. Filled boxes represent the currently used open fuel cycle and unfilled box es represent the proposed closed fuel cycle. The main waste product in a nuclear reactor is the spent fuel rod. Nuclear fuel rods are made from various forms of uranium ore which is processed and enriched to contain 3 4% radioactive 235 U. After three years of use the rod must be replaced. The spent fuel rod contains 95.5% uranium, 3.1% other stable isotopes and 1.4% harmful radioactive elements including the minor actinides, 99 Tc, 129 I, 90 Sr, and 135 Cs. 8 The spent

PAGE 18

18 fuel rod is vitrified in 1 3M nitric acid and from here the separation of all waste products from the remaining uranium can begin. The most radiotoxic elements in the waste are plutonium and the minor actinides. The radiotoxicity of these elements can be significantly reduced by transmutation. During transmutation the elements are bombarded with neutrons, which cause them to undergo fission reactions. Once the fission process is complete the daughter products that are generated from the process are either quickly eliminated by nuclear decay or are stable non radioactive materials Transmutation is severely impeded if there is a mixture of lanthanides [Ln(III)] and actinides [An(III)] because the Ln(III) have a higher neutron capture cross section. When bomb arded with neutrons the Ln(III) ions more readily absorb the neutrons thus the An(III) are not transmuted efficiently. 9 10 If these waste products could be effectively sep arated and transmuted there would be significant reduction in waste volume and the r adioactive components leftover would require much less storage time of hundreds rather than thousands of years. With a shorter amount of time required for safe storage, the re is much less chance of damaging an underground repository from a geological change. 11 1.2.1 Plutonium Uranium Recovery by Ex traction Method The Plutonium Uranium Recovery by Ex traction ( PUREX ) method is used to ext ract the plutonium and uranium for the waste stream. PUREX is a liquid liquid extraction technique that uses tributylphosphate (TBP) shown in Figure 1 3, in an inert diluent (kerosene or dodecane) as the organic phase and nitric acid as the aqueous phase. First the fuel rod assembly is dismantled and the fuel pellets are dissolved in hot nitric acid. The U and Pu are then extracted from the acidic solution and then separated from each other to be either used as fuel or disposed 12 The key issue in PUREX is the

PAGE 19

19 recovery of solvent. The solvent is contaminated by degradation and radiolysis of TBP which in turn causes emulsions that inhibit extraction effectiveness. The emulsion is counter acted by adding alkaline, carbonate, and water washes to remove the degradation products. 13 Figure 1 3 Structure of tributylphosphate (TBP), an extractant used in the PUREX process 1.2.2 Transuranium Ex traction Method Another method that has shown to be capable at removing the tri tetra and hexavalent actinides and lanthani des is called the Transuranium Ex traction (TRUEX ) process. The TRUEX process is similar to PUREX in that it also works by a solvation mechanism b ut it uses a combination of N ,N d iisobutylcarbamoyl methyl octylphenyl phosphineoxide (CMPO) shown in Figure 1 4, and TBP in an organic solvent. In acidic solutions the TRUEX process still works well and eliminates the need for pH adjustments thus the vol ume of waste generated is decreased. Although it is selective in extracting these products it cannot separate the trivalent lanthanides from the trivalent actinides. Another problem with the TRUEX method also contains phosphorus since hazardous secondary w aste is produced upon incineration the process is far from ideal 11

PAGE 20

20 Figure 1 4 Structure of N,N d iisobutylcarbamoyl methyl octylphenyl phosphineoxide (CMPO), an extractant used in the TRUEX process 1.2.3 Diamide Ext raction Method In congruence with the TRUEX process the Diamide Ext raction (DIAMEX ) process was conceived. 14 Both methods have the same purpose but DIAMEX follows the CHON principle (ligands only contain c arbon, hydrogen, oxygen and nitrogen) thus it can be incinerated with less hazardous byproducts. Two ligands used for the DIAMEX process are dimethyl dioctyl hexylethoxy malonamide (DMDOHEMA) and dimethyl dibutyl tetradecyl malonamide (DMDBTDMA) and are shown in Figure 1 5 Fi gure 1 5 Structure of DMDOHEMA and DMDBTDMA, extractant s used in the DIAMEX process DMDOHEMA bonds bidentate and the efficie ncy of these ligands are greatly influenced on the position of the N substituents due to the fact that they can affect the

PAGE 21

21 basicity of the carbonyl oxygens. 15, 16 The hard oxygen donors in DIAMEX ligands do not disti nguish the actinides from the lanthanides thus another process is n eeded to separate the Ln(III) fro m the minor actinides in the raffinate. 1.2.4 Selective Actinide Ex traction Method The Selective Actinide Ex traction ( SANEX ) process tries to separate the Ln(III) from the minor actinides by using ligands that contain acidic sulfur or heterocyclic nitrogen donors. There are several methods for measuring the ability of extractants for Ln(III)/An(III) separation in the literature. 11 To quantify the ability of an extractant the distribution coefficient is used. The distribution coefficient is a measure of the metal concentration in the organic phase divided by the metal concentration in the aqueous phase of an extraction. A nother common term used in measuring the ability of an extractant to selectively extract actinides over lanthanides is the separation factor. The separation factor is defined as the distribution coefficient of the actinide divided by the distribution coeff icient of the lanthanide. The most common metals used to show separation factors are the comparison between europium and americium. Dithiophosphonic acids such as Cyanex 301 (bis(2,4,4 trimeth ylpentyl)dithiophosphinic acid), shown in Figure 1 6, exhibits h igh selectivity for Am over the lanthanides. Unfortunately, the ligand generates secondary waste due to the presence of sulfur and phosphorus and it is susceptible to radiation damage Additionally, the pH of the waste stream must be adjusted to ~3, requir ing significant diliution in order to extract efficiently 17, 18 R ecently interest has focused on ligands with multidentate heterocyclic nitrogen donor atoms, using Eu(III) to represent the lanthanide series in test scenarios 11 It has been theorized that the electron density and the basicity of the nitrogen donor atom influences how the ligand interacts with the metal cation. 19

PAGE 22

22 The most thoroughly investigated and highest Am(III)/Eu(III) separation factors (SF) ligands with heterocyclic nitrogen donor atoms has been 2,6 bis(1,2,4 triazin 3 yl)pyridines (BTP) and is shown in Figure 1 6 The scrutinization of BTP also reve aled that it is susceptible to hydrolysis and radiolysis which makes them a poor choice for an industrial process. 20 The work described herein uses multi coordinate benzimidazole bearing ligands because it is known that they do not suffer from hydrolysis, are stable at high temperature, and have been shown to bind lanthanides and actinides although limited data on their extraction efficiency is available. 21, 22 Fi gure 1 6 Structure of Cyanex 301 and BTP extractants used in the TRUEX process. 1.3 Separation of Lanthanides and Actinides The similarities between An(III) and Ln(III) make separation an arduous task. An(III) and Ln(III) cations are classified as hard acids by the HSAB principle, hence their bonding is primarily ionic and mainly governed by charge density. 23 Separation by exploiting their common bonding principles also proves difficult due to the similar ionic radii from the f element contractions an d the identical oxidation state One difference that can be exploited is that the An(III) favor covalent bonds more than Ln(III). Therefore, many successful ligands employ bonding of atoms so fter than oxygen. The

PAGE 23

23 advantageous covalency of the An(III) is due to the 5f orbitals being more spatially extended than the 4f orbitals. The extension leads to better overlap if the f orbitals are involved in bonding. Whether or not the 5f orbitals are in volved in bonding has been disputed with some implying that the s orbital is the contributor due to the many coordination geometries and others attributing high separation factors to the difference in covalency. 24, 25 T o facilitate b onding multicoordinate ligands that have preorganized chelating groups have been synthesized to selectively separate lanthanides and actinides Since trivalent f elements have high coordination numbers of eight or g reater and prefer a tricapped trigonal pr ismatic (TTP) geometry shown in Figure 1 7 If the binding ligands are preorganized to facilitate the TTP geometry the result would be better extractions and possibly higher metal selectivities The preorganized pla tforms provide opportuni ties for different stoichiometric ligand amounts for complete complexation and different steric requirements. Fi gure 1 7 Tricapped trigonal prismatic (TTP) geometry 1.4 Benefits of Preorganization on a Molecular Platform Many groups have developed ligands based on using molecular platforms. One of the first synthetic accomplishments involved attaching multiple CMPO moieties to a

PAGE 24

24 variety of calixarenes (Figure 1 8 ). 26 30 The calixarene platform can be functionalized on the narrow lower rim and at the wide upper rim of the platform. The functionalization can also be used to tune solubility and atta ch different c helating groups (Figure 1 9 ). Fi gure 1 8. A simple calixarene. Functionalities can be added to bolster metal extraction. Calixarenes also facilitate the extraction of metal cations by inhibiting the approach of water molecules to the metal coordination site. 31 Calixarenes have been extensively studied for the preorganization of chelating groups and can range from a three phenyl to an eight membered macrocycle. 32 Some problems with calixarenes include solubility, not being easily modified and inability to provide C 3 symmetry due to steric strain. Fi gure 1 9 CMPO moieties attached to a calixarene platform Another class of molecules that is used as a molecular platform is the C 3 symmetric triphenoxymethane scaffold shown in Figure 1 10 Using three tridentate

PAGE 25

25 ligands attached to the platform will build a nonadentate ligand capable of bonding in a tricapped trigonal prismatic geometry. Upon the addition of the multi coordinate ligand to the extraction medium any water or nitrates that are bonded to the inner coordination sphere of th e metal are displaced. The removal of the nine water or ni trate molecules by one nonadentate ligand leads to an increase in the spontaneity of reaction by increasing entropy A thermodynamic study has shown that for the heavier lanthanides the complexation reaction is mainly driven by entropy thus the predicted entropic gain of a preorganized molecular platform correlates with their data. 33 34 An important feature of t riphenoxymethane is that it can be conformationally locked so that three hydroxyl grou ps point in the same direction orientation of the hydroxyl groups facilitate preorganization of the final nonadentate ligand in that the phenyl rings do not rotate around the central carbon of triphenoxymethane The substituents on the phenyl rings can be changed to tune solubility, cavity size and influence other steric properties T he solubility of the ligand can be set early in the synthesis by altering groups at the 2 and 4 positions of the phenols. With the ability to change these groups, the interaction of the molecule at the extraction phase boundary can be tuned to facilitate the transfer of the metal from the aqueous to organic phase. The entropic gain, preorganization and tunable solubility engender the triphenoxymethane platform with useful attributes for t he develop ment of extraction agents Many variations using triphenoxymethane have been experimented on by changing the substituents and trying different chelating groups. 35, 36 37

PAGE 26

26 Fi gure 1 10 Triphenoxymethane molecular platform Kornelia Matolka, a previous member of the Scott group, prepared a triphenoxymethane based ligand using three trident ate chelating groups. The chelating group us ed different moieties of digycolamides (DGA) as shown in Figure 1 11 Fi gure 1 11 DGA attached to triphenoxymethane The ligand chelating groups used oxygen do nors to bind the metal and showed tolerance to acidic solutions. The ligand showed increased affi nity for the later lanthanides and maintained good extraction ability even w ith a 1:1 ligand:Ln(III) ratio, where no rmally a 10:1 ratio is required to achieve such results 38 Unfortunately, the hard donors in DGA hin der ed it from binding actinides selectively. In an attempt to decrease the hardness of th e donors, the oxygen atoms were replaced by sulfur atoms. Inclusion

PAGE 27

27 of softer sulfur atoms lead to poor extraction efficiency. The Scott group has since put considerab le effort into exploring ligands that incorporate nitrogen donor atoms. The nitrogen donor atoms should still have the relative softness when compared the DGA oxygen donors but still be hard enough to bond selectively to the actinides. 1.5 Experimental Research Objectives Based on the knowledge gained from earlier work in the Scott group a new class of ligands with pyridyl benzimidazole donors was designed to selectively separate the lanthani des and actinides. The choice of donor atoms was guided by com putational chemistry results in order to correlate a predictive analysis to experimental data. When applicable biphasic extraction experiments were performed using the synthesized ligand and lanthanide elements in a nitric acid solution in order to simula te the industrial separation process. The common moiety used for eac h ligand is shown in Figure 1 12 and relies on the synthetic procedures developed by Piguet et al. but has never been attached to any variation of triphenoxymethane. 39 Fi gure 1 12 Common moiety used in all synt hesized ligands refer to the various linkages, donor groups and solubility modifiers 1.6 Computation Research Objectives C omputatio nal chemistry was used for this project to establish a predictive method for screening possible ligands and to investigate what electronic properties facilitate the

PAGE 28

28 selective separation of lanthanides and actinides. The screening process used semi empirical calculations to evaluate the structu re and symmetry of several ligands and based on this information synthetic targets were chosen DFT calculations were n ot performed on the full nonadentate structures because of the intensive computation time and resources needed Instead, DFT calculations were utilized t o investigate the electronic structure of a truncate d version of the ligands The accuracy of DFT became necessary for calculating the energy level matching and HOMO LUMO ga p energies of the metal ligand complexes The elect ronic structure properties that facilitate the selective separati on of lanthanides and actinides were investigated and correlated to experimental data when applicable. 1.6.1 Semi Empirical Calculations The laborious multi step procedure needed to synthesiz e a nonadentate ligand capable of fu lly encapsulating a lanthanide or actinide often requires many moiety var iations to optimize the bonding geometry By performing quantu m chemical calculations on the ligands and metal complexes one can predict whether t he synthetic effort will be fruitful. In order to accurately predict whether a ligand will be a good extractant some chemical property must be evaluated that correlates to it. In previous work completed by the Scott group, any deviation f rom TTP geometry and concomitant increase in twist angle correlated with decrease in extraction efficiency. The twist angle is defined as the angle to which the two triangles in the TTP are deviated from their eclipsed geometry. Thus in theor y, the TTP geometry can be modeled from an optimized geometry calculation With this calculation, one can predict the extraction efficiency.

PAGE 29

29 Recently, the SPARKLE/AM1 model was developed and it showed the model could give accurate results for Ln(III) complex es. SPARKLE/AM1 is a semi empirical method that is exclusively implemented in the free software package, MOPAC2009. 40 SPARKLE refers to the parameterization of the trivalent l anthanides and AM1 is the actually level of theory being used in the calculation. Several papers show results using the AM1 method and the more recently developed PM6 method are comparable to ab intio DFT calculations and to crystallographic data from the Cambridge Structural Database. 41 Because the method is semi empirical, it uses a simple Hamiltonian as well as parameters obtained from experimental results or ab initio calculations to quickly calculate molecular properties including geome try. The only feasible means of optimizing the geometry of large (250 atoms or more) lanthanide complexes is to use the SPARKLE/AM1 method. In this study many structures are computed to assess the twist angle and the semi empirica l method is evaluated across lanthanide series to its accuracy with our specific compounds. 1.6.2 Density Functional Theory Calculations Due to the radiotoxicity and expense of working with actinides many groups in this field have begun to use Density Fun ctional Theory (DFT) in order to investigate how to better design ligands for Ln(III)/An(III) separation. Calculations of Ln(III) and An(III) complexes is not as perspicuous as other structures because the large number of electrons at the metal requires re lativistic effects must be taken into consideration. The average radius of an atom is proportional to the Bohr radius a 0 and a 0 is inversely proportional to the electron mass. Therefore, the relativistic increase of mass with velocity decreases the radiu s of the inner s orbitals in a heavy atom. To maintain orthogonality with the inner s orbitals, the outer s orbitals must also decrease in radius.

PAGE 30

30 The retraction of s orbitals causes all of the p orbitals to shrink as well but to a lesser degree. As the s and p orbitals get smaller they screen the nucleus more effectively than nonrelativistic atoms thus causing an expansion of the d and f orbitals. All electron calculations of relativistic molecules are time consuming so to increase the speed the calculatio ns the use a relativistic effective core potential (RECP) is employed. RECP freezes the core electrons based on one all electron iteration and then only the valence electrons are treated explicitly. 42 Accounting f or relativistic effects with an effective core potential has been shown to give accurate data while minimizing refinement time. 43 The most prominent calculation methods to aid in the development of extraction complexes include using either Gaussian or the Amsterdam Density Functional (ADF) program package. In order to run a calculation on Ln(III) complexes a generated basis set must be used to include relativistic effects. In addition, the common practice is to use the generated basis set on just the metal and much lower level (i.e. 6 31G*) basis set on the rest of the atoms. 44, 45 The lower level of theory is used for all other atoms in order to decrease the calculation time. This method gives accurate data for both packages but the calculation time is quite long. In this study, DFT methods were utilized t o investigate the properties of several heterocyclic nitrogen donor ligands bound to either Eu, Am, or Cm. The calculations performed used three terdentate ligands in order to simulate the environment provided by the nonadentate triphenoxymethane ligand. T he triphenoxymethane platform was truncated in order to decrease the computational time The molecules underwent geometry optimizations and their molecular orbitals evaluated.

PAGE 31

31 CHAPTER 2 STRUCTURAL ANALYSIS OF NONADENTATE LANTH ANIDE COMPLEXES USIN G SEMI EMPIRICAL METHODS 2.1 Introduction Semi empirical (SE) calculations have long been used to evaluate chemical systems that contain too many electrons to calculate by normal Hatree Fock methods or DFT. SE methods use an approximated Hamiltonian operator that estimates integrals using experimental and ab initio calculation data in order to avoid cumbersome all electron calculations. Recently, the semi empirical program MOPAC incorporated the SPARKLE model which allows for trivalent lanthanide complexes to be calculated with impressive accuracy. 40, 41 46 In an independent review the SPARKLE model was said to 47 due to the accuracy in which it can estimate bond lengths. These reviews made MOPAC a logical choice for full geometry optimization of the structures before synthetic efforts were attempted The rapid speed of the calculations also provides immediate structural information to help modulate ligand design instead of exploring time consuming and expensive synthesis. There ar e many benefits to using SE methods for ligand design of which the most significant is that the full nonadentate ligand using the triphenoxymethane scaffold can be incorpora ted where as with any other method the triphenoxymethane scaffold must be truncated and the structure evaluated as three separated tridentate ligands. One of the largest drawbacks to using SE methods is the known inaccurac y for calculating amide bonds geometries 48 This is problematic for this project since the most successful synthetic method of joining the metal binding units to triphenoxymethane is through an amide linkage. I nspection of amide bond rotation was closely monitored for deviations in

PAGE 32

32 planarity. Another issue is that the actinides are not able to be calculated with SE methods because the lack of data on them and complex electronic structure of their diffuse 5f orbitals. In this chapter, SE c alculations are used t o develop a means of visualizing the structure of non adentate ligands formed by triphenoxymethane derivatives as molecular scaffol ds. The twist angle for each calculated structure is evaluated in order to obtain information on ideal bonding geometries Finally, the method is tested to assess how precisely it predicts the la nthanide contraction. The SE optimiz ed structures also offer a starting geometry for further DFT optimization. The main goal of this project is to model and evaluate the structures to predict the most viable ligand design that facilitates bonding. The stru ctures that exhibit the best results were then deemed worthy of synthetic efforts. 2.2 Computed Structures The following structures in Figure 2 1 were computed in MOPAC using SPARKLE/PM6 method and showed the most promising results when assessing their st ructural properties. The ligand design followed the strict criteria of only using C, H, O and N atoms, triphenoxymethane to form a nonadentate structure upon complexation and containing synthons that have been previously published. The benzimidazole contai ning ligand arm was based on the benzimidazole lanthanide complex work performed by Piguet et al. and the triphenoxymethane work from the Scott group. 49, 50 Structures 2 1, 2 3, 2 4, 2 6 and 2 7 used linkages previously explored by the Scott group to attach the ligand arm to triphenoxymethane. The synthetic linkages that have not been attempted before include structures 2 2, 2 5, 2 8 and 2 9

PAGE 33

33 Figure 2 1. Semi empirical calculation structures

PAGE 34

34 Figure 2 2 Semi empirical geometry optimization of structure 2 7

PAGE 35

35 Figure 2 3 Semi empirical geometry optimization of structure 2 8

PAGE 36

36 The length of the arm linking the donor groups to the platform was also varied with either two or three carbon atoms. The extra CH 2 group was included to test whether the increase in flexibility would influence the binding of the three arms to the metal. Different alkyl groups were used to ascertain if bulky substituents would affect the bonding environment. Since the alkyl groups on triphenoxymethane are the main factor influencing solubility, knowledge of their steric effects on metal binding is important. Europium was used for each structure due to its importance in lanthanide actinide separation but als o because it represents the average size and charge of the lanthanide series. The calculated structures of 2 7 and 2 8 are shown in Figure 2 2 and Figure 2 3, respectively. 2.3 Structural Evaluation The metal to ligand bond distances, twist angles and the standard deviation (STD) of the twist angles are shown in Table 2 1 and 2 2. The values listed in the tables represent the main parameters on which ligand design was evaluated with SE calculations. Visual inspection of the geometry optimization a lso was useful but not easily quantifiable due to the many subtle differences between each structur e. Table 2 1 Average Metal Bond Leng ths () and Twist Angle s ( ) of Amide Structures O N py N bzd Twist Angle STD Twist Angle 2 1 2.43 2.56 2.47 12.94 1.15 2 2 2.43 2.56 2.47 11.33 0.30 2 3 2.43 2.56 2.47 13.27 0.80 2 4 2.43 2.56 2.47 12.48 0.51 2 5 2.45 2.53 2.46 15.70 1.29 2 6 2.43 2.56 2.47 14.09 1.77 2 7 2.45 2.53 2.46 16.02 4.81 The oxygen donor of the amide, nitrogen donor of the pyridine and nitrogen donor of the benzimidazole are notated as O, N py and N bzd respectively.

PAGE 37

37 Table 2 2 Average Metal Bond Lengths () and Twist Angle s ( ) of Triazole Structures N tz N py N bzd Twist Angle STD Twist Angle 2 8 2.51 2.52 2.47 10.36 0.58 2 9 2.50 2.52 2.47 10.36 0.58 The nitrogen donor of the triazole, nitrogen donor of the pyridine and nitrogen donor of the benzimidazole are notated as N tz N py and N bzd respectively. The twist angle for each complex was obtained by calculating the centroid of each triangle of the TTP geometry and averaging the three dihedral angles of donor atom (top), centroid (top), centroid (bottom), donor atom (bottom). All of the amide type ligand s have similar bond l engths with an exception of structures 2 5 and 2 7. It is likely that the small deviation is due to the steric strain with the short acyl linkage in 2 5 and the amide linkage containin g the donor oxygen in 2 7. The triazole type ligand s have approximately the same bond length to the benzimidazole as the amide type ligands but the triazole moiety accounts for a longer bond as opposed to the oxygen donating atom of the amide type ligands. The twist angle for each complex fell within a na rrow range of 10.36 to 16.02 degrees and additional analysis were performed to extract more discerning data from each structure The standard deviation for each of the three twist angles in the structures was calculated in order to evaluate the symmetry of each molecule. A higher standard deviation b etween twist angles for any given structure implies more asymmetry in the optimized geometry and a lower theorized extraction efficiency Using the twist angle and the standard deviation as criteria the SE calculations predicted that 2 2, 2 4, 2 8, and 2 9 would be possible candidates for synthetic efforts. Structures 2 2 and 2 4 both use d a thre e carbon linker validating that the additional flexibility gives a more sym metric structure than their two carbon linking counterparts, 2 1 and 2 3. In contrast, s tructures 2 8 and 2 9 had the exact same twist angle values despite one

PAGE 38

38 being two carbon linker and the other a three carbon linker. It is possible that something speci fic to the planarity and direct donation of the triazole moiety or a combination of many effec ts caused 2 8 and 2 9 to be different from the rest of the calculations Structure 2 7 had the largest twist angle but also a significantly higher standard deviat ion compared to the other structures. The inability of SE calculations to accurately define the conjugated planarity of amide bonds is most directly seen in structure 2 7 thus it was chosen as a possible candidate for synthesis because of the inconclusive evidence provided by the calculation. SE calculations were completed using different alkyl groups on triphenoxymethane but there was no change in metal ligand bond lengths or twist angles thus the wide variety of groups at these positions in structure 2 1 through 2 9 were deemed negligible. It is unfortunate that the alkyl groups of methyl, tert butyl, tert pentyl were ineffective in changing the bonding environment of the metal because it represents a possible parameter to alter selectivity but experiment ally it is a positive result with regards to modifying solubility without possible fluctuations to the metal ligand bond. Other possible alkyl groups were not calculat ed because at the time no synthetic procedures were established for their synthesis. 2. 4 Lanthanide Contraction A self consi stent process of analyzing the accuracy of the method is to use the same ligand but change the lanthanide for each calculation Across the lanthanide series there should be a decrease in bond length of the metal ligan d bond due to the fact that the ionic radius gets smaller. W hen increasing electrons on the metal there is an increase in shielding and decreases ionic radius following the trend of the lanthanide

PAGE 39

39 contraction. From Figure 2 4 and Figure 2 5 i t is shown that ligand 2 4 and 2 7 both have a decrease in bond length across the lanthanide series, respectively. Figure 2 4 The graph shows the metal to donor atom bond lengths for structure 2 4. Figure 2 5 The graph shows the metal to donor ato m bond lengths for structure 2 7

PAGE 40

40 In Figure 2 4 and 2 5 the oxygen donor of the amide, nitrogen donor of the pyridine and nitrogen donor of the benzimidazole are notated as O, N py and N bzd respectively. The degree of error becomes apparent when comparing lantha nides that are adjacent to each other in the periodic table suggesting the method cannot accurately reproduce such small variations in size. Interestingly, the ligand used for structure 2 7 represented the lanthanide contraction even though the large fluct uations in twist angle persisted. One could speculate that the SE method used has a much higher accuracy for the ligand lanthanide bonds than it does with other atomic interaction in the molecule. Despite the minor deviations with adjacent metals the met hod accurately represent ed the lanthanide contraction and affirmed its ability to accurately measure trends for the nonadentate complexes calculated in this study. 2.4 Conclusions In this study semi empirical calculations were used to assist in practical ligand design. The main objective was to eliminate possible ligand structures using computational data and symmetry arguments. Out of nine possible ligands five were chosen to be wo rthy of synthetic efforts. The decision was based on the analysis of the twist angle of each calculated structure. Varying the alkyl groups on triphenoxymethane did not have an effect on the twist angle and deemed to have negligible effect on the metal bon ding environment. The lanthanide contraction was reproduced fo r two representative ligands thus giving validity to the method for the uncommon structures computed in this study.

PAGE 41

41 2.5 Computational Details All semi empirical calculations were performed with in MOPAC on PC with a Windows operating system. The level of theory used was PM6 and trivalent lanthanides used the SPARKLE model.

PAGE 42

42 CHAPTER 3 SYNTHESIS OF BENZIMI DAZOLE PRIDYL FUNCTIONALIZE D NONADENTATE LIGANDS 3.1 Introduction In this study the syn thesis of several benzimidazole pyridyl functionalized ligands was attempted. The basic procedure was to first synthesis the ligand arm which usually contained two or three possible chelating groups then attaching three of the ligand arms to triphenoxymeth ane in order to form a nonadentate ligand. Several synthetic procedures developed by Piguet et al. were employed and improved upon to achieve the benzimidazole pyridyl functionality that they use for luminescence experiments. Although they have not attache d their ligands to triphenoxymethane they have been able to attach the ligand arms to another type of tripodal platform shown in Figure 3 1 The li gand in Figure 3 1 l ack ed rigidity and had limited solubility modifier s thus limiting its use in An( III)/Ln(III ) extractions. However, when three of the arm s are attached to triphenox ymethane the close proximity lead s to a more preorganized structure for the binding of the metal. The goal for the synthetic project was to obtain several nonadentate ligand s containing benzimidazole pridyl moieties in as high yield and lowest number of steps as possible in order to be applicable to an industrial process The synthetic targets were chosen based on semi empirical calculations 3.2 S ynthetic Difficulties of Et her Linkage Ligands With the knowledge that the ligands synthesized by Piguet et al. could be attached to a tripodal platform through an ether linkage and the information gathered by the predictive semi empirical calculations discussed in the previous chap ter attempts were made to substitute in triphenoxymethane as the molecu lar platform via ether linkage. The first steps taken involved the synthesis of ligand arm as shown in Figure 3 2.

PAGE 43

43 Figure 3 1 Tripodal helicate ligand developed by Piguet et al. Figure 3 2. Synthesis of ligand arm 3 10, which contains the chloride functionality needed to attach triphenoxymethane via Williamson ether synthesis.

PAGE 44

44 The synthesis began with commercially available starting materials and proceeded smoothly except for the formation of product 3 8. It was found that the benzimidazole nitrogen was susceptible to nitrone formation during the procedure used to remove the excess iron from the reaction. EDTA was used to remove the iron by varying the pH and adding hydrogen per oxi de. The hydrogen peroxide was assumed to be the culprit for the nitrone formation and the placement of the nitrone on the benzimidazole was determined by obtaining a n X ray crystal structure of the unwanted byproduct. Other than the nitrone formation the synthesis does not deviate much from the literature procedure with the exception that the yield on 3 3 was increased from 85% to 100% by reacting for two weeks at room temperature instead one hour in a reaction vessel and that the yield on 3 9 was nev er reproduced in the claimed 94% yield. The discovery that 3 3 could be synthesized from a reaction time of two weeks was made by analyzing the reaction by TLC. Once the ligand arm (3 10) was synthesized the two carbon l inker triphenoxymethane (3 15a) was synthesized by the procedure in Figure 3 3. Even though the three carbon ether linkage displayed better twist angle parameters from semi empirical calculations the two carbon ether linkage to triphenoxymethane was used because of the previously establishe d synthetic procedure developed by Matloka. To attach 3 10 to 3 15a NaH was used to deprotonate the hydroxyl group on 3 15a but the subsequent S N 2 reaction on 3 10 was unsuccessful using a variety of solvents and reaction conditions. Another attempt was m ade by adding a tosyl leaving group on 3 15a to make 3 16a and reacting it with 3 9 in the presence of base but this reaction failed to attach the two reactants and form the desired product. After

PAGE 45

45 exhaustive reaction conditions were explored for the ether linkage it became apparent that some other means of tethering the ligand arm to triphenoxymethane would be requir ed. Figure 3 3. Synthesis of two carbon linker triphenoxymethane. 3.3 Synthesis of Amide Linkage Ligands Another method that had been establi shed to attach a ligand arm to triphenoxymethane was to use an amide linkage. The Scott group had already explored how to synthesi ze a two and three carbon amine functionalized triphenoxymethane and subsequently attach a ligand arm with a carboxylic acid functionality by using the reagent benzotriazol 1 yl oxytripyrrolidinophosphonium hexafluorophosphate ( PyBOP) The three carbon amine triphenoxymethane was chosen for the synthesis due to the favorable twist angle data obtained from the semi empir ica l calculations comparing 2 3 and 2 4. The synthesis for the three carbon linker amine 3 18 shown in Figure 3 4 was published by Matloka. The synthesis also incorporated di t butyl groups on

PAGE 46

46 triphenoxymethane due to solubility issues involved with the me thyl t butyl triphenoxymethane (3 13a) Figure 3 4. Synthesis of the three carbon linker amine triphenoxymethane 3 18 3.3.1 Carboxylic Acid Functionalization of the Ligand Arm The ligand arm was modified to a novel carboxylic acid moiety using two separ ate procedures shown in Figure 3 5 All reactions using chromium reagents, hydrogen peroxide, and sodium hypochlorite to oxidize 3 8 or 3 9 directly to the carboxylic acid failed. The acid was formed after preparation of aldehyde, 3 19, followed by oxidati on oxidation Figure 3 5. Synthesis of carboxylic acid functionalized ligand arm, 3 20

PAGE 47

47 Due to the low yield to form 3 9 and 3 20 efforts to improve the protocol were undertaken to increase the yield of the overall synthesis. The elimination of the methyl deprotection step in the synthesis of 3 9 and direct oxidation of the e ther (3 8) to a n ester (3 21) increased the overall yield to th e carboxylic acid from 3.3% to 22%. The overall yield for the synthesis of the nonadentate ligand 3 22 was 9.4 % in 16 steps Figure 3 6. Synthesis of the nonadentate ligand, 3 22 The low yield for the amide linkage and benzimidazole formation of 3 23 and 3 24 was due to side reactions that occur red from using the weak methy l ester protection group In addition, 3 24 resulted in a mixture of methyl and ethyl esters due to the use of ethanol as a solve nt but it is inconsequential since the ester is hydrolyzed in the subsequent step. With a carboxylic acid directly attached to the ort ho position of the pyridine the coupling reaction with PyBOP forms the link to triphenoxymethane and the amide metal donor group in one synthetic step (Figure 3 8). The overall yield for the final nonadentate ligand 3 26, is 2.2 % with a total of eleven st eps.

PAGE 48

48 3.3.2 Incorporation of the Amide Linkage as a Donor Group for Metal Binding To improve upon the final structure the amide linkage was included as a donor group to the metal. The synthetic target was chosen to be the inconclusive semi empirical calculated structure, 2 7. To obtain the targeted ligand 3 18 was used as the molecular platform but a new ligand arm was needed. The synthesis of the new amide ligand arm is shown in Figure 3 7. Figure 3 7. Synthesis of the carboxylic acid functionalize d benzimidazole pridyl ligand arm, 3 25 Figure 3 8. Synthesis of the nonadentate ligand, 3 26

PAGE 49

49 3.4 Synthesis of Triazole Linkage Ligands Since t he previous ligand arm synthetic protocols were complicated by low yield s and a large number of steps a comple t ely new synthetic scheme was developed. In comparison to the other methods Figure 3 9 shows a route that uses a different reagent to form the benzimidazole ring, offers highly combinatorial options through halogen exchange and cross coupling reactions and ultimately creates a softer all nitrogen donating environment for metal complexation. Figure 3 9 Synthesis of the alkyne ligand arm, 3 31 The synthesi s of 3 28 is from the literature procedure developed by Chuang et al. 51 The commercially available reactant, 3 29, was was substituted for 3 1 in order to limit the number of steps and increase the overall yield of the nonadentate ligand. The benzimida zole formation using Fe/HCl for 3 8 and 3 24 proved too harsh to retain the bromine of 3 28 Using Na 2 S 2 O 4 as the benzimidazole ring closing reagent not only formed the targeted product but offered a n easier purification procedure. The Sonogashira reaction introduced the alkyne to 3 30 by adding trimethylsilyl acetylene in the presence of base and CuI. The overall yield for the ligand arm was 22% in five steps. It should be noted the compound 3 31 is likely photosensitive as it began to darken when stored in the presence of light.

PAGE 50

50 The azide was substituted for the tosyl group of 3 16c as shown in Figure 3 10. Since semi empirical calculations did not show a difference in twist angle between structures 2 8 and 2 9 and because there was no established procedure for a three carbon linker the two carbon linker was again utilized. The alkyl substituents on the triphenoxymethane rings were changed to di tert pentyl in order to attempt to alleviate solubility problems. The final step in the synthesis relies on the 1, 3 dipolar cycloaddition of the azide on triphenoxyme thane and the alky ne on the ligand arm to produce 3 33. With nine total steps the overall yield for the nonadentate ligand 3 33, was 8%. Figure 3 10. Synthesis of the nonadentate triazole ligand, 3 33 3.5 Conclusions In this chapter the synthesis of four types of nonadentate ligands were presented of which three were successful ly completed The first synthetic target was the structure

PAGE 51

51 2 1 but the ether linkage failed to react and the product was not obtained even after leaving group changes and many a ttempts using various conditions. The second synthetic target, 3 22, was produced in the highest yield of all the synthesized ligands but also the highest number of steps The structure of the ligand used an amide linkage to connect the triphenoymethane p latform to the ligand arm but did not incorporate the amide linkage into metal binding. The third synthetic target, 3 26, includ ed the linking amide as a donor group for metal binding and eliminated five synthetic steps compared to 3 22 but due to the we ak methyl ester protecting group and harsh reaction conditions the yield was decreased The fourth synthetic target, 3 33, c ontained all heterocyclic nitrogen donors of benzimidazole, pyridine, and triazole. The ligand arm synthesis was completely changed in order to increase the overall yield eliminate steps and to form an alkyne for the triazole linkage. The final nonadentate ligand, 3 33, was obtained in the highest yield and with the least number of steps than the other ligands. Each consecuti ve ligand had an improvement in yield or synthetic steps but the overall yield is still questionable in terms of an industrial process It appears that further optimization of the reactions or alternate synthetic pathways would be needed for large scale pr oduction. F or an industrial extraction process they have the advantage of encompassing the metal at a 1:1 ligand to metal ratio whereas three tridentate ligands would need a 3:1 ratio. When comparing nonadentate ligands to three tridentate ligands the reac tion yield needs to be three times as high for the three tridentate ligands to fit the appropriate stoichiometry. For instance the best yi eld obtained was for 3 22 at 9 % so any tridentate sy nthesis would need at least 28 % overall yield to be competitive.

PAGE 52

52 The objective of the synthetic project was to synthesize nonadentate ligands using a triphenoxymethane platform and to expand upon previously developed methods for obtaining benzimidazol e pridyl containing ligand arms. Three novel nonadentate ligands were synthesized and their procedures were optimized to achieve the highest yields. 3.6 Experimental Section Starting materials ( 3 1, 3 4, 3 11a,b,c, 3 propylbromo phthalimide, N methyl 2 nitroaniline 2,6 dibromopyridine ) were purchased through Sigma Aldrich. The synthetic methodology for the preparation of compounds 3 2, 3 5, 3 6, 3 7, 3 8, 3 9, and 3 10, were adapted from literature procedures. 52 49 Synthesis of 3 27 used the procedure of Chuang et al. 51 Methods for the synthesis of triphenoxymethane followed literature procedures. 34, 35, 53, 54 All of the 1 H and 13 C NMR spectra were recorded on a Varian VXR 300 or Mercury 300 spectrometer at 299.95 MHz. Each mass spec sample was dissolved in a ppropriate solvent and underwent direct injection through an autosampler, followed by ESI or APCI analysis with methanol (with or without 0.2% FA) as mobile phase. Solvent is used only when necessary for DART. The ions were detected with the Agile nt 6210 TOF MS while the data was process ed with the MassHunterTMsoftware. methyl 2 (6 (diethylcarbamoyl)pyridin 2 yl) 1 ethyl 1H ben zo [d]imidazole 5 carboxylate (3 21 ): N,N diethyl 6 (1 ethyl 5 (methoxymethyl) 1H benzo[d]imidazol 2 yl)picolinamide ( 3 8 ) (1.08 g, 2.95 mmol), KMnO 4 (4.66 g, 29.5 mmol), and triethylbenzylammonium chloride (9.50 g, 29.5 mmol) were refluxed in dichloromethane (100 mL) for 24 hours. The excess KMnO 4 was removed by the addition of a concentrated solution of sodium bisulfite. The organic phase was separated and the aque ous layer was extracted with

PAGE 53

53 dichloromethane. The organic phases were combined, dried with sodium sulfate, filtered and concentrated until crystalization occurred to afford 0.906 g (81%) of 3 21 1 H NMR (CDCl 3 ): = 1.07 (t, 3H, J=7.0 Hz, CH3), 1.28 (t, 3H, J=1.0 Hz), 1.48 (t, 3H, J=6.9 Hz), 3.34 (q, 2H, J=1.0 Hz), 3.62 (q, 2H, J=6.9 Hz), 3.96 (s, 3H), 4.78 (q, 2H, J=7.1 Hz), 7.48 (d, 1H, J=8.7 Hz), 7.58 (d, 1H, J=7.8 Hz), 7.97 (t, 1H, J=7.8 Hz), 8.07 (d, 1H, J=8.7 Hz), 8.41 (d, 1H, J=7.9 Hz), 8.55 (s, 1H) ppm. 13 C NMR (CDCl 3 ): = 12.60, 14.10, 15.13, 39.32, 40.67, 42.59, 51.88, 109.59, 122.37, 124.67, 124.93, 137.91, 139.20, 142.03, 148.72, 150.78, 154.37, 167.25, 168.07 ppm. ESI MS m/z : 381.1914 [M + H] + 403.175 3 [M + Na] + 783.3591 [2M + Na] + 2 (6 (diethylcarbamoyl)pyridin 2 yl) 1 ethyl 1H benzo[d ]i midazole 5 carboxylic acid (3 20 ): Compound 3 21 (0.50 g, 1.3 mmol) and NaOH (0.1 g, 2.5 mmol) were refluxed in methanol/water (20 mL/20 mL) for 30 mins. The mixture was then concentrated under vacuum, precipitated with concentrated hydrochloric acid (35%), and filtered. The white solid was dried under vacuum to give 0.48 g (99 % yield) of 3 22 1 H NMR (CDCl 3 ): = 1.10 (t, 3H, J=7.2 Hz), 1.30 (t, 3H, J=7.1 Hz), 1.52 (t, 3H, J=7.0 Hz), 3.37 (d, 2H, J=7.0 Hz), 3.64 (d, 2H, J=7.2 Hz), 4.81 (d, 2H, J=7.3 Hz), 7.54 (d, 1H, J=8.6 Hz), 7.63 (dd, 1H, J=7.7, 0.9 Hz), 8.03 (t, 1H, J=7.9 Hz), 8.17 (dd, 1H, J=8.6, 1.5 Hz), 8 .50 (dd, 1H, J=8.0, 0.8 Hz), 8.71 (d, 1H, J=1.3 Hz), 9.89 10.92 (bs, 1H) ppm. 13 C NMR (CDCl 3 ): = 13.03, 14.53, 15.54, 39.84, 41.28, 43.10, 110.21, 123.24, 123.38, 125.11, 125.80, 125.88, 138.58, 139.62, 141.59, 148.49, 151.10, 154.78, 168.55, 171.36. E SI MS m/z : 367.1787 [M + H] + 389.1584 [M + Na] + 733.3445 [2M + H] + 755.3262 [2M + Na] +

PAGE 54

54 methyl 6 (ethyl(4 (methoxymethyl) 2 nit rophenyl)carbamoyl)picolinate (3 23 ): A mixture of 3 5 (13.00 g, 71.8 mmol), CH 2 Cl 2 (120 mL), thionyl chloride (20.8 mL, 285 m mol), and DMF (0.1 mL) was refluxed for 1.5 h under a nitrogen atmosphere and evaporated to dryness. The white residue was dried under vacuum for 30 min, dissolved in CH 2 Cl 2 (60 mL), a nd cooled to 0 C. A mixture of 3 3 (15.34 g, 73.0 mmol), triethylamine (44 mL), and CH 2 Cl 2 (100 mL) was added dropwise. The resulting solution was stirred for 10 min at 0C, refluxed for 2 h, and evaporated to dryness. The residual brown oil was dissolved in CH 2 Cl 2 /aqueous half saturated NH 4 Cl, the organic layer was separated and the aqueous layer was extracted with CH 2 Cl 2 The combined organic phases were washed with deionized water, dried over MgSO 4 filtered, and evaporated to dryness. The resulting crude compound was purified by column chromatography using 100% DCM to 96 % DCM/ 4% MeOH gradient to afford 3 23 as a br own oil (11.4 g, 24.6 mmol, 42 %). 1 H NMR (CDCl 3 ): = 1.25 (t, 3H, J=7.2 Hz), 3.41 (s, 3H), 3.72 (dq, 1H, J=13.9, 7.2 Hz), 3.80 (s, 3H), 4.23 (dq, 1H, J=13.9, 7.1 Hz), 4.47 (s, 2H), 7.35 (d, 1H, J=8.2 Hz), 7.51 (dd, 1H, J=8.4, 1.8 Hz), 7.82 (t, 1H, J=1.0 Hz), 7.89 7.99 (m, 2H), 8.06 (dd, 1H, J=7.8, 1.2 Hz) ppm. 13 C NMR (CDCl 3 ): = 12.66, 31.11, 46.40, 52.52, 58.73, 72.95, 124.31, 125.94, 127.89, 132.10, 136.43, 138.05, 139.45, 145.73, 145.96, 152.56, 164.94, 165.91 ppm. ESI MS m/z : 374.1342 [M + H] + 396.1184 [M + Na] + 769.2426 [2M + Na] + 6 (1 ethyl 5 (methoxymethyl) 1H benzo[d ]imidazol 2 yl)picolinic acid (3 25 ): A mixture of 3 23 (9.19 g, 24.6 mmol), ethanol (750 mL), water (200 mL), powdered iron (11.0 g, 197 mmol), and concentrated hydrochloric acid (37%, 50 mL) was refluxed for 18 h under a nitrogen atmosphere, filtered, and concentrated under vacuum. The residual aqueous layer was poured into a mixture of CH 2 Cl 2 (200 mL),

PAGE 55

55 water (370 mL) and Na 2 H 2 EDTA2H 2 O (132 g, 393 mmol). The pH was adjusted to 7 with a 25% aqueous ammonia solution then 30% hydrogen peroxide solution (7 mL, 68.8 mmol) was very slowly added and the mixture was stirred for 15 min. The pH was adjusted to 8.5 with a 25% aqueous ammonia solution, the aqueous layer was extracted with CH 2 Cl 2 and the combined organic layers were washed with deionized water until neutral, dried over MgSO 4 filtered, and evaporated to dryness to afford compound 2 24 The mixture of methyl and ethyl esters was then hydrolyzed by KOH (2.76 g, 49.2 mmol) in 75 mL of water at reflux for 1 hour. Concentrated HCl was added until pH=2 and a white precipitate formed which was collected by filt ration to give 3 25 (2.60 g, 8.36 mmol, 34% yield.) 1H NMR (CDCl 3 ): = 1.52 (t, 3H, J=6.9 Hz), 3.33 (s, 3H), 4.59 (s, 2H), 4.95 (q, 2H, J=1.0 Hz), 7.53 (d, 1H, J=8.6 Hz), 7.81 (s, 1H), 7.98 (d, 1H, J=8.6 Hz), 8.23 8.42 (m, 2H), 8.73 (d, 1H, J=7.7 Hz) p pm. 13C NMR (CDCl 3 ): = 14.88, 41.92, 57.65, 73.18, 112.53, 114.28, 125.56, 126.70, 128.29, 132.89, 133.45, 136.78, 139.55, 144.49, 145.72, 148.51, 165.34 ppm. DART MS m/z: 312.1350 [M + H]+, 623.2613 [2M+ H]+ 2 (6 bromopyridin 2 yl) 1 m ethyl 1H benzo[d]i midazole (3 30 ): Under a nitrogen atmosphere 6 bromopicolinaldehyde (3.2 g) and N methyl 2 nitroaniline (2.6 g) were stirred in a degassed mixture of ethanol (256 mL) and water (64 mL) at 100C until all solids completely dissolved. To this solution, soli d Na 2 S 2 O 4 (9.6 g) was added in one portion and the mixture was stirred at 100C for 24 hr to give a yellow solution. The solution was concentrated under vacuum and extracted with ethyl acetate/water. The organic phases were combined, dried with sodium sulf ate, and evaporated. Purification was achieved by column chromatography using a 100% DCM to 95% DCM/ 5% MeOH gradient and gave 2.0 g of product (41%). 1 H NMR (CDCl 3 ):

PAGE 56

56 = 4.13 (s, 3H), 7.18 7.36 (m, 3H), 7.39 (d, 1H, J=7.9 Hz), 7.55 (t, 1H, J=7.8 Hz), 7.68 7.77 (m, 1H), 8.26 (d, 1H, J=7.7 Hz) ppm. 13 C NMR (CDCl 3 ): = 32.56, 109.85, 119.85, 122.67, 122.95, 123.55, 127.74, 137.00, 138.85, 140.30, 142.04, 148.07, 150.77 ppm. ESI MS m/z : 288.0140 [M + H] + 1 methyl 2 (6 ((trimethylsilyl)ethynyl)pyridin 2 yl) 1H benzo[d]imidazole ( 3 31 a ): The following Sonogashira coupling reaction took place under a nitrogen atmosphere. Compound 3 30 (1.59 g, 8.57 mmol), Pd(PPh 2 )Cl 2 (0.30 1 g, 5 mol %), trimethylsilylacetylene (1.4 mL, 9.43 mmol) and CuI (0.082 g, 5 mol %), were added to a solution of anhydrous DMF. To this solution dry, degassed triethylamine (5 mL, 35.6 mmol) was added and allowed to stir for 2 hours at room temperature. After the reaction was complete the mixture was filtered over Celite and the filter pad was rinsed with a small amount of THF. The filtrate was then acidified with 1M HCl and extracted with DCM. The organic phase was washed with water, dried, and evaporate d. Purification was achieved by column chromatography using 100% DCM to afford 0.83 g of product (70%). 1 H NMR (CDCl 3 ): = 0.30 (s, 9H), 4.29 (s, 3H), 7.29 7.38 (m, 2H), 7.40 7.46 (m, 1H), 7.51 (d, 1H, J=1.0 Hz), 7.74 7.86 (m, 2H, J=7.9 Hz), 8.32 (d 1H, J=1.0 Hz) ppm. 13 C NMR (CDCl 3 ): 0.00, 32.97, 95.22, 103.84, 110.20, 120.31, 122.88, 123.69, 124.30, 127.67, 137.17, 137.52, 142.14, 142.70, 149.84, 151.03 ppm. C 18 H 19 N 3 Si calculated: C, 70.78; H, 6.27; N, 13.76. found: C, 70.40; H, 6.21; N, 13.5 0 2 (6 ethynylpyridin 2 yl) 1 methyl 1H benzo[d]imidazole (3 31 b ): A solution of 3 31a (0.734 g, 3.62 mmol), CsF (0.365 g, 4.34 mmol), MeOH (25 mL), and THF (25 mL) was stirred for 4 hours. After the reaction was complete the mixture was evaporated and then a 1M HCl solution was added. The solution was extracted with DCM, the organic phases combined and washed with water. The organic

PAGE 57

57 phase was then dried with sodium sulfate, filtered, and evaporated. The crude product was then recrystallized from minimal diethyl ether to afford 0.515 g product (92%). 1 H NMR (CDCl 3 ): = 3.18 (s, 1H), 4.26 (s, 3H), 7.27 7.36 (m, 2H) 7.37 7.44 (m, 1H), 7.49 (d, 1H, J=7.6 Hz), 7.73 7.85 (m, 2H), 8.36 (d, 1H, J=7.9 Hz) ppm. 13 C NMR (CDCl 3 ): = 33.03, 77.38, 82.90, 110.24, 120.29, 122.96, 123.79, 124.63, 127.67, 137.31, 137.50, 141.30, 142.57, 149.52, 151.09 ppm. ESI MS m/z : 234.10 30 [M + H] + 467.1976 [2M + H] + Tris(2 ethylacetoxy 3 methyl 5 tert pentylphenyl)methane (3 14c ): Under a nitrogen atmosphere 0.50 g of 3 13c was dissolved in dry acetone and 0.40 g of ethyl bromoacetate and 0.91 g of Cs 2 CO 3 was added. The mixture was ref luxed for 12 15 h and then cooled to room temperature. The acetone was removed under vacuum and the solids dissolved in diethyl ether. Solid MgSO 4 was added and the insoluble salts and drying agent were filtered off. The ether was removed from the filtrat e and to give the crude product. The white solid was recrystallized in ethanol to give 0.49 g (71%) of product 3 14c 1 H NMR (CDCl 3 ): = 0.46 0.63 (m, 18H), 1.14 (s, 18H), 1.27 (t, 9H, J=7.2 Hz), 1.32 (s, 18H), 1.48 (q, 6H, J=7.3 Hz), 1.68 (q, 6H, J=7.3 Hz), 4.16 (s, 4H), 4.24 (q, 6H, J=7.2 Hz), 6.31 (s, 1H), 6.99 (s, 2H), 7.07 (s, 3H) ppm. 13C NMR (CDCl 3 ): = 9.34, 9.71, 14.33, 28.72 29.16, 34.62, 37.11, 37.86, 39.26, 60.85, 68.38, 69.61, 125.12, 127.42, 136.98, 140.51, 143.39, 152.54, 169.10 ppm. Yield = 71%, ESI MS m/z: 971.6936 [M + H]+, 989.7055 [M + H 3 O]+ Tris[2 (2 hydroxylethoxy) 3 methyl 5 tert pentylphenyl]methane (3 15c ): A dry diethyl ether solution of 3 14c was added over 1 2h with an addition funnel to a slurry of LiAlH 4 (0.49 g, 0.503 mmol) in 100 mL dry diethyl ether cooled to 0 mixture was then warmed to room temperature and stirred 12 15 h. The excess

PAGE 58

58 reductant was destroyed with 1 M HCl (100 mL). The ether layer was separated and further extracted with 1 M HCl (2 x 100mL) and brine (100 mL). The ether was then dried with MgSO 4 After filtration of the drying agent, the ether was removed to give 0.38 g (90%) of white crystalline material. 1 H NMR (CDCl 3 ): = 0.53 (m, 18H), 1.18 (s, 18H), 1.32 (s, 18H), 1.50 (s, 6H), 1.70 (m, 6H), 3.50 (m, 2H), 3.74 (m, 4H), 3.98 (m, 4H), 4.23 (m, 2H), 6.43 (s, 1H), 7.06 (s, 3H), 7.21 (s, 3H) ppm. 13 C NMR (CDCl 3 ): = 9.09, 9.48, 34.78, 36.94, 37.64, 39.10, 62.07, 73.12, 124.83, 127.72, 137.31, 139.92, 142.74, 152.12 ppm. Yield = 90%. ESI MS m/z : 845.6654 [M + H] + 862.6916 [M + NH 4 ] + 1690.3219 [2M] + 1707.3464 [2M+NH 4 ] + Tris[2 (2 toluenesulfonlyethoxy) 3 methyl 5 tert pent ylphenyl]methane (3 16c ): In a dry flask 0.283 g (0.334 mmol) of 3 15c was dissolved in 100 mL of dry sulfonyl chloride was added and the reaction mixture was stirred for for 12 15 h at room temperature. The pyridine was removed under vacuum and the solid material dissolved in 100 mL methylene chloride and the extracted with 1 M HCl (2 x 100 mL). The organic phase was the dried with MgSO 4 filtered, an d the solvent removed. Methanol was added and the white solid product filtered off and washed with more dry methanol to afford 0.41 g (93%) of 3 16c. 1 H NMR (CDCl 3 ): = 0.36 0.60 (m, 18H), 1.12 (s., 18H), 1.23 (s, 18H), 1.38 1.51 (m, 6H), 1.54 1.77 (m, 6H), 2.43 (s, 9H), 3.13 3.46 (m, 2H), 3.51 3.80 (m, 2H), 3.83 4.20 (m, 4H), 4.36 4.83 (m, 4H), 6.19 (s, 1H), 7.02 (s, 3H), 7.09 (s, 3H), 7.37 (d, 6H, J=8.2 Hz), 7.85 (d, 6H, J=7.9 Hz) ppm. 13 C NMR (CDCl 3 ): = 9.03, 9.32, 21.57, 29.26, 29.67, 3 4.83, 36.81, 37.62, 39.12, 69.46, 70.00, 125.10, 127.54, 127.95, 129.91, 132.99, 137.01, 140.32, 143.31, 144.63,

PAGE 59

59 152.14 ppm. C 76 H 106 O 12 S 3 calculated: C, 69.80; H, 8.17; found: C, 69.32; H, 8.39 Yield = 93%. Tris(2 (2 azidoethoxy) 3,5 di tert pentylphenyl)methane (3 32 ): 3 16c (15.31g, 11.7mmol) was combined with sodium azide (6.85g, 105mmol) in 325mL of DMF in a dry 500mL Schlenk flask fitted with a reflux condenser. The reaction mixture was heated to reflux overnight. The solution was all owed to cool to room temperature when 250mL of water was added and the solution extracted with dichloromethane (3 x 100mL). The combined organics were washed with water (3 x 100mL) and brine (1 x 100mL). The organics were then dried with sodium sulfate a nd the solvent removed under reduced pressure yielding pure 3 32 (10.82g, 99 %). 1 H NMR ( CDCl 3 ): = 0.52 (t, 9H, J=7.5 Hz), 0.58 (t, 9H, J=7.5 Hz), 1.15 (s, 18H), 1.39 (s, 18H), 1.44 1.57 (m, 6H), 1.64 1.90 (m, 6H), 3.10 4.06 (m, 9H), 6.40 (s, 1H), 7 .08 (s, 3H), 7.12 (s, 3H) ppm. 13 C NMR (CDCl 3 ) = 9.06, 9.51, 29.42, 35.18, 36.83, 37.66, 38.57, 39.17, 51.80, 70.52, 125.12, 127.80, 137.44, 140.25, 143.20, 152.59 ppm. C 55 H 85 N 9 O 3 calculated: C, 71.78; H, 9.31; N, 13.70; found: C, 71.87; H, 9.62; N, 13.5 0 tris(2 (2 (4 (6 (1 methyl 1H benzo[d]imidazol 2 yl)pyridin 2 yl) 1H 1,2,3 triazol 1 yl)ethoxy) 3,5 di tert pentylphenyl)methane ( 3 33 ): A solution of tris(2 (2 azidoethoxy) 3,5 di tert pentylphenyl)methane (3 32) (0.458 g, 0.497 mmol), 3 31b (0.406 g 1.74 mmol), CuI (0.01 g, 10 mol%), and Et 3 N (0.55 mL, 3.98 mmol) were combined in THF (30 mL) under a nitrogen atmosphere. The solution was stirred for 24 hours and turned dark brown over time. The resulting mixture was poured into a half saturated ammoni um chloride solution and then extracted with DCM and subsequently evaporated. Diethyl ether was poured into the dark mixture and filtered. The filtrate was then evaporated and purified by column chromatography (100%

PAGE 60

60 DCM to 4% MeOH : 96% DCM gradient) to af ford 0.492 g of product (61%). 1 H NMR (CDCl 3 ): = 0.52 (m, 18H), 1.00 1.29 (m, 36H), 1.41 (d, 6H, J=7.3 Hz), 1.52 (d, 6H, J=7.0 Hz), 3.75 3.92 (m, 6H), 3.96 (s, 9H), 4.52 (m, 3H), 4.95 (m, 1H), 6.55 (s, 1H), 7.03 7.24 (m, 12H), 7.31 (br. s., 3H), 7. 67 (td, 3H, J=7.9, 1.5 Hz), 7.74 (d, 3H, J=7.9 Hz), 8.11 (d, 3H, J=7.9 Hz), 8.20 (d, 3H, J=7.9 Hz), 8.69 (s, 3H) ppm. 13 C NMR (CDCl 3 ): = 9.01, 9.40, 28.00, 28.81, 28.93, 29.05, 32.10, 34.32, 36.71, 37.71, 38.95, 51.05, 53.34, 69.96, 109.65, 119.79, 119.9 6, 122.45, 123.17, 123.61, 124.35, 125.24, 128.14, 136.89, 136.96, 137.54, 140.37, 142.21, 143.92, 148.03, 148.88, 149.73, 150.02, 151.94 ppm. APCI MS m/z : 1619.9719 [M + H] + General Procedure for the Synthesis of 3 22 and 3 26 A DMF solution of 1.0 equiv of carboxylic acid, 1.1 equiv PyBOP, and 2 equiv of diethyl isopropylamine was stirred together for 30 minutes at room temperature. To this solution was added 0.3 equiv of the triphenoxymethane amine and was stirred for 24 hours. After the reaction was com plete 10% HCl solution was added and then extracted with either dichloromethane or diethyl ether. The organic phases were combined and underwent several water washes to removed residual DMF. The organic phase were then dried with sodium sulfate, filtered. and evaporated to affored the product. Analysis of Compound 3 22 : 1 H NMR (CDCl 3 ): = 1.23 (s, 27H), 1.37 (s, 36H), 1.90 2.08 (m, 3H), 2.10 2.36 (m, 9H), 3.36 (s, 9H), 3.52 3.70 (m, 6H), 3.82 4.01 (m, 6H), 4.50 (s, 6H), 4.53 4.75 (m, 6H), 6.54 (s, 1H), 7.17 7.26 (m, 6H), 7.28 7.37 (m, 6H), 7.57 (t, 3H, J=7.9 Hz), 7.68 (s, 3 H), 7.73 (dd, 3H, J=7.6, 1.1 Hz), 8.08 8.17 (m, 3H), 8.21 (dd, 3H, J=7.8, 1.0 Hz) ppm. 13 C NMR (CDCl 3 ): = 15.38, 30.69, 31.38, 31.48, 34.50, 35.49, 37.61, 40.24, 57.77, 70.15, 74.93, 109.75, 119.65, 121.98, 122.35, 123.83, 126.85, 127.23, 132.94,

PAGE 61

61 135.4 9, 137.58, 137.98, 141.77, 142.45, 144.60 ppm. ESI MS m/z : 840.5151 [M + 2H] 2+ 851.5058 [M + H + Na] 2+ 1701.9972 [M + Na] + Yield: 85% Analysis of Compound 3 26 : 1 H NMR (CDCl 3 ): = 0.98 (t, 9H, J=7.2 Hz), 1.06 1.54 (m, 72H, M13), 1.94 2.49 (m, 6H), 3.23 (q, 6H, J=7.1 Hz), 3.51 (q, 12H, J=7.4 Hz), 3.67 3.91 (m, 6H), 4.30 4.74 (m, 6H), 6.51 (s, 1H), 7.12 (s, 6H), 7.17 7.32 (m, 3H), 7.41 (d, 3H, J=7.4 Hz), 7.62 7.82 (m, 6H), 8.01 (d, 3H, J=7.7 Hz), 8.21 (s, 3H) ppm. 13 C NMR (CDCl 3 ): = 13.00, 14.50, 15.30, 31.06, 31.73, 31.80, 34.72, 35.75, 38.88, 39.70, 41.28, 42.98, 70.87, 110.68, 118.54, 122.52, 123.29, 123.90, 125.19, 127.40, 130.90, 137.47, 137.73, 138.47, 140.10, 142.06, 144.63, 147.55, 149.91, 153.74, 154.91, 168.14, 168.42 ppm. APCI MS m/z : 1845.1415 [M + H] + 1867.1237 [M + Na] + Yield: 80%

PAGE 62

62 CHAPTER 4 METAL COMPLEXATION EXPERIMENTS OF LANTHANIDES USING BENZIMIDAZOLE PRIDYL FUNCTIONALIZED NONAD ENTATE LIGANDS 4.1 Introduction In this chapter the extraction results and crystal structure of the ligands synthesized in chapter three are discussed. The objective of the experim entation was to analyze each ligands extr action efficiency for lanthanides and to evaluate t he crystal structure of the ligands after metal comple xation The solvent extraction study used ligands 3 22 3 26 and 3 33 to examine how relatively soft nitrogen donor atoms play ed a role in lanthanide extraction. In each study dichloromethane containing a 10 3 M solution of ligand was used as the organic phase and 1 M nitric acid containing 10 4 M Ln(III) nitrate was used as the aqueous phase. Once the extraction event was performed UV Vis spectrophotometry was used to measure the ability of each ligand to extract the lanthanides. The extr action efficiency of each trivalent lanthanide was calculated as %E = 100/(A 1 A)/ (A 1 A 0 ), where A is the absorbance of the ex tracted aqueous phase A 1 is the absorbance of the aqueous phase befo re extraction and A 0 is the absorbance of metal free 1 M nitric acid solution. The crystal structures of 3 26 and 3 33 both with ytterbium were obtained. Both crystals were formed from ytterbium (III) triflate. The triflate anion balance s the trivalent cation of the complex. Structural parameters are reported and the crystal structure is compared to computational data. 4.2 Extraction Results The data in Table 4 1 shows that ligands 3 22 and 3 26 ga ve the highest %E while 3 33 gives %E values within the error limits of the analysis method.

PAGE 63

63 Table 4 1. Extraction Efficien cy Ligand 3 22 3 26 3 33 C ation %E %E %E La(III) 5.7 5.1 0.9 Ce(III) 3.9 4.3 n.d. Pr(III) 2.7 4.2 n.d. Nd(III) 5.7 4.5 0.1 Eu(III) 5.0 3.0 0.1 Gd(III) 4.2 4.0 n.d. Tb(III) 4.6 4.0 n.d. Dy(III) 5.1 5.8 0.2 Er(III) 5.3 5.4 n.d. Tm(III) 10.8 12.3 n.d. Yb(III) 6.2 7.2 0.7 It is not surprising that the hard oxygen donors of the amide functional group in ligands 3 22 and 3 26 had a greater affinity for lanthanides in comparison to the softer triazole donor of 3 33 since the lanthanides themselves are classified as hard acids. T he rigidity of the two carbon linker arm in 3 33 could hamper to metal complexation in comparison to the three carbon linker in 3 22 and 3 26 although semi empirical calculations seem to sugg est otherwise for this particular ligand The lowered extraction efficiency of rigid two carbon linkers that connect the three ligand arms to the triphenoxymethane platform verses three carbon linkers was previously evaluated experimentally by the Scott gr oup. 53 With the exception of Tm and Yb all of the lanthanides had approximately 5% extraction efficiency As previously stated, t he increase in %E for the late la nthanides are mainly governed by entropic effects rather than enthalpy of complexation. It is also possible that the binding pocket of the ligands 3 22 and 3 26 are more favorable for smaller lanthanides. The similarity in %E between 3 22 and 3 26 demonstrates that the extra amide group in 3 26 does not drastically affect the bonding environment despite the increased opportunity for hydrogen bonding.

PAGE 64

64 Since the %E for ligands 3 22 and 3 26 are still all less than 13% they may still prove viable to give high separation factors if actinide extracti on studi es were performed. The actinide extraction studies for this set of ligands were not performed because of the expense, radiotoxicity and government regulation of actinides. The most structurally similar molecule to 3 33 that has undergone solvent ex traction experiments consists of a bis benzimidazole pridyl skeleton and did not use a molecular scaffold. Drew et al. reported that their ligands containing this structure also showed comparably low distribution coefficients for Eu(III) although it should be noted that their extraction conditions are different than what is used in this study. 19 They also reported an increase in distribution coefficients when using Am(III) leading to separation factors as high as 9 for Am(III)/Eu(III). Although this value is dwarfed by the BTP maximum separation factor of 150 it does show proof of concept that benzimidazole containing ligands may offer acceptable separation between lanthanides and actinides if conditions are opti mized. 4.3 Crystal Structure The crystals presented in this study were grown by the slow diffusion of methyl tert butyl ether or diethyl ether in to methanol. All efforts to crystallize l anthanide nitrate salts resulted in precipitation of a fine powder D espite numerous attempts crystals of 3 22 could not be obtained In a typical lanthanid e complexation procedure equal molar portions of a lanthanide salt were reacted with ligand in ethyl acetate for several hours. In the case of nitrate salts the produc t immediately precipitated from the solution The precipitate was filtered to obtain the metal ligand complex. Triflate salt reactions were evaporated to dryness to obtain the complex. Figure 4 1 and Figure 4 2 show the crystal structure of ytterbium complexed by 3 26 and 3 33, respectively.

PAGE 65

65 Figure 4 1. Crystal structure of Yb complexed by ligand 3 26 with ellipsoids at 50% probability

PAGE 66

66 Figure 4 2. Crystal structure of Yb complexed by ligand 3 33 with ellipsoids a t 50% probability

PAGE 67

67 4.4 Conclusions The ext raction experiments using ligands 3 22, 3 26, and 3 33 were performed to test their ability to extract lanthanide ions from an acidic solution If the extraction efficiency is too high for the lanthanides the select ive separation of actinides would not be likely. The extraction efficiency for ligands 3 22, and 3 26 were approximately 5% and increased for the heavier lanthanides. The low extraction efficiency with lanthanides could be a positive result in that the sel ectivity of the lanthanides is low but it could also mean that the ligand is not bonding the metal or the solubility of the ligand in solution is poor The crystal structure s of Yb( 3 26 ) and Yb( 3 33 ) demonstrate d that the ligand can complex the metal. The crystal structures were evaluate d for their bond lengths, twist angle and standard deviation ( STD ) of twist angle. The solid state structure was also compared to SE calculations to evaluate accuracy of the theoretical method. To analyze the agreement between the theoretical SE calculations and the experimentally obtained crystal structure a new set of calculations were performed that used the solid state geometry as the initial guess for the SE geometry optimization. The results are reported in Table 4 2 and 4 3. Table 4 2. Comparison of bond lengths () and twist angles ( ) between the crystal structure, SE calculation and crystal SE calculation of Yb(3 26) Crystal Structure SE calculation SE calculation of crystal O 2.31 (7) 2.38 2.36 N py 2.4 6(9) 2.43 2.44 N bzd 2.55 (9) 2.36 2.34 Twist Angle 9.08 8.57 4.04 STD Twist Angle 1.24 4.58 0.05 The oxygen donor of the amide, nitrogen donor of the pyridine and nitrogen donor of the benzimidazole are notated as O, N py and N bzd respectively.

PAGE 68

68 Table 4 3. Comparison of bond lengths () and twist angles ( ) between the crystal structure, SE calculation and crystal SE calculation of Yb(3 33 ) Crystal Structure SE calculation SE calculation of crystal N triazole 2.47 (6) 2.42 2.43 N py 2.51 (7) 2.43 2.44 N bzd 2.47 (6) 2.38 2.37 Twist Angle 8.92 5.53 5.18 STD Twist Angle 1.06 0.27 0.77 The nitrogen donor of the triazole, nitrogen donor of the pyridine and nitrogen donor of the benzimidazole are notated as N triazole N py and N bzd respectively. T he bond lengths in each structure were not accurately represented at the benzimidazole donor group and it is concluded that an alteration in the way that benzimidazole is treated in the SE calculations is needed T he SE calculation of Yb(3 26) was found to be in agreemen t with the crystal structure twist angle but this could be coincidental since the STD of the twist angles do not agree. When the crystal structure geometry of Yb(3 26) was used as a starting point for the SE geometry optimization the structure was taken to a more symmetric minimum as shown by the decrease in STD The SE calculations of Yb( 3 33 ) were similar to those of Yb( 3 26 ) in that the benzimidazole donor was not accurately represente d. The twist angles of Yb(3 33) did not ach ieve the accuracy of Yb( 3 26 ) but the STD value was in better agreement. When the crystal structure geometry of Yb(3 33) was used as a starting point for the SE geometry optimization, the structure was not significantly changed. Overall, the SE calculation s are validated by comparison with the crystal structure with whatev er starting geometry is chosen.

PAGE 69

69 4.5 Experimental Section 4.5.1 Extraction Experiment The lanthanide and actinide salts, La(NO 3 ) 3 6H 2 O (Alpha Aesar), Ce(NO 3 ) 3 6H 2 O, Pr(NO 3 ) 3 6H 2 O, Nd(NO 3 ) 3 6H 2 O, Eu(NO 3 ) 3 5H 2 O, Gd(NO 3 ) 3 6H 2 O, Tb(NO 3 ) 3 6H 2 O, Dy(NO 3 ) 3 5H 2 O, Er(NO 3 ) 3 5H 2 O, Tm(NO 3 ) 3 5H 2 O, Yb(NO 3 ) 3 5H 2 O Yb(Tf ) 3 5H 2 O deionized water, TraceMetal grade HNO 3 (Fisher Scientific), and HPLC grade organic solvents. The extraction efficiency was measure by utilizing t he Arsenazo(III) assay and was performed on a Varian Cary 50 UV vis spectrophotometer The metal extraction experiments followed previously reported procedure. 35, 55, 56 Metal stock solutions were standardized by colorimetric titration with Na 2 H 2 EDTA and xylenol orange All extraction experiments used 10 3 M solutions of ligand in dichloromethane and a 10 4 M solutions of Ln(III) nitrate in 1 M nitric acid. The extraction took place in a 10 mL sealed vial using 4 mL of organic solution and 4 mL of aqueous solution and agitated on an oscillating table for 24 hours. The errors, based on the precision of the spectrophotometer and the standard deviation from the mean of at least three measurements, were in most cases no higher than 1.5%. The distribution coefficients ( defined as the ratio of concentrations of metal in the organic and aqueous layers) were obtained and reported. 4.5.2 Crystal Structure of Yb(3 26) 3+ X Ray i ntensity data were collected at 100 K on a Bruker DUO diffractometer using MoK radiation ( = 0.710 73 ) and an APEXII CCD area detector. Raw data frames were read by program SAINT and integrated using 3D profiling algorithms. 57 The resulting data were reduced to produce hkl reflections and their intensities and estimated standard deviations. The data were corrected for Lorentz and polarization

PAGE 70

70 effects and numerical absorption corrections were applied based on indexed and measured faces. The structure was solved and refined in SHELXTL6.1, using full matrix l east squares refinement. The non H atoms were refined with anisotropic thermal parameters and all of the H atoms were calculated in idealized positions and refine d riding on their parent atoms. 58 In the final cycle of refinement, 17922 reflections (of which 12086 are observed with I > 2 (I)) were used to refine 1661 parameters and the resulting R 1 wR 2 and S (goodness of fit) were 5.77%, 14.43% and 0.932, respectively. The refinement was carried out by minimizing the wR 2 function using F 2 rather than F values. R 1 is calculated to provide a reference to the conventional R value but its function is not minimized. Further par ameters are reported in Table 4 4 Several solvent molecules were disordered and could not be modeled properly, thus program SQUEEZE, a part of the PLATON package of crystallographic software, was used to calculate the solvent disorder area and remove its contribution to the overall intensi ty data. 59 4.5.2 Crystal Structure of Yb(3 33 ) 3+ X Ray Intensity data were collected at 100 K on a Bruker DUO diffractometer using MoK radiation ( = 0.71073 ) and an APEXII CCD area detector. Raw data frames were read by program SAINT and integrated using 3D profiling algorithms. 57 The resulting data were reduced to produce hkl reflections and their intensities and estimated standard deviations. The data wer e corrected for Lorentz and polarization effects and numerical absorption corrections were applied based on indexed and measured faces. The structure was solved and refined in SHELXTL6.1, using full m atrix least squares refinement. 58 The non H atoms were refined with anisotropic thermal

PAGE 71

71 parameters and all of the H atoms were calculated in idealized positions and refined riding on their parent atoms. In the final cycle of refinement, 19530 reflections (of which 1518 8 are observed with I > 2 (I)) were used to refine 1260 parameters and the resulting R 1 wR 2 and S (goodness of fit) were 7.96%, 20.13% and 1.027, respectively. The refinement was carried out by minimizing the wR 2 function using F 2 rather than F values. R 1 is calculated to provide a reference to the conventional R value but its function is not minimized. Further par ameters are reported in Table 4 4 Several solvent molecules were disordered and could not be modeled properly, thus program SQUEEZE, a part of the PLATON package of crystallographic software, was used to calculate the solvent disorder area and remove its contribution to the overall intensi ty data. 59 The largest residual electron density peak could not be properly accounted for chemically and considering the amount of disorder in this structure, it is very likely that it is the result of a slight l ower quality data than expected.

PAGE 72

72 Table 4 4 X ray data for crystal structures Yb complexed by 3 22 and 3 33 using triflate counter ions grown from the diffusion of ether into methanol Yb(3 33 ) 3+ (CF 3 SO 3 ) 3 Yb(3 26 ) 3+ (CF 3 SO 3 ) 3 Total Reflections 50337 17922 Unique Reflections Reflections I (I) 19530/15188 17922/12086 1.49 < < 25.00 1.54 < < 27.50 Formula C 103 H 118 F 9 N 18 O 12 S 3 Yb C 106 H 130 F 9 N 12 O 18 S 3 Yb M r 2240.37 2300.44 Crystal System Triclinic Rhombohedral Space Group P R3c a () 13.9187(16) 42.262(3) b () 14.9433(18) c () 29.236(4) 45.351(3) 90.593(2) 91.506(2) 114.085(2) V c ( 3 ) 5548.0(11) 70148(8) D c (g cm 1 ) 1.341 1.307 Z 2 24 F( 000 ) 2314 28632 [Mo K ] (mm 1 ) 0.979 0.933 R 1 [ I (I)data] 0.0796 0.0577 w R 2 [ I (I)data] 0.2013 0.1443 GoF 1.027 1.089 Largest Peak, deepest trough (e 3 ) +4.588, 1.802 +1.081, 0.817 a ) Obtained with monochromatic Mo K radiation ( = 0.71073 ) b ) R 1 = F o F c / F o c ) w R 2 = { [ w ( F o 2 F c 2 ) 2 / [ w ( F o 2 ) 2 ]} 1/2

PAGE 73

73 CHAPTER 5 ELECTRONIC STRUCTURE OF EUROPIUM, AMERICI UM, AND CURIUM COMPLEXES USING DENS ITY FUNCTIONAL THEOR Y 5.1 Introduction The complex bonding and radiotoxicity of actinides has spurred many theoretical investigations in recent years. The advent of cheap and available computing power has made many electron calculations involving heavy metals more feasible. Theoreticians have a lso invented effective core potentials (ECP) for heavy metals in which co re electrons are calculated as part of the nuclei and treated as a rigid non polarizable p otential while the valence electrons are computed with a separate valence wav e function. 42 The approximation drastically decreases the amount of computational resources needed to calculate actinide complexes. In particular, the effective core potentials used in this study are relativistic effective core po tentials (RECP) because they also account for the relativistic energy changes that occur as the innermost electrons start to approach the speed of light. For an accurate description of heavy metals, such as actinides, accounting for relativistic effects ha s been shown to be critical for computation. 43 The methods chosen for the calculations were based on computational resources, program availability, and literature results. All calculations were performed using the Gau ssian 03 program package on the University of Florida High Performance Computing network. Each geometry optimization and single point calculation of a metal ligand complex used a maximum of four processors and eight gigabytes of memory. The hybrid B3LYP fu nctional was chosen over other DFT functionals in order to facilitate self consistent field ( SCF ) convergence. Americium, Europium and Curium were chosen to be computed for their importance to lanthanide actinide separation

PAGE 74

74 methods. A small core relativist ic effective core potential was used for the metals. The double zeta basis set, 6 31G*, was chosen for all other atoms as a triple zeta basis set proved to be too costly in computational resources. In accordance with literature data high spin configuratio n was used for all computed complexes which corresponds to the following spin multiplicities for each metal: Eu = 7, Am = 7, Cm = 8. In this study calculations of the molecular orbital (MO) data produced large files of text and to facilitate interpretati on the data was plotted graphically. The data for the MO diagrams in this chapter were processed in the AOMix program package. In order to decrease the complexity of the diagrams only MO contributions that are above 4% were included The metal and ligand e nergies in the diagrams have been to highlight the frontier orbitals but the metal ligand complex energies have not been altered. Because of the many approximations used in the calculations the results for the different elements can be compared but the e nergy values are not accurate. The objective of the DFT study was to evaluate the geometry and elect ronics of the metallated complexes of the synthesized compounds. In order to decrease calculation time the ligands were truncated to three independent ter dentate ligands as opposed to one nonadentate ligand. For comparison the BTP ligand was computed in the same manner since the actinide selectivity of it is often attributed to the orbital interactions. Since the structure has never been modeled with the B3LYP functional, Am(BTP) 3 3+ was also included in the calculation in view of its relevance to the area of study. 5.2 Geometry Optimizations Geometry optimizations were carried out using metals Eu 3+ Am 3+ and Cm 3+ in combination with structures 5 1 5 2 and 5 3 (Figure 5 1 ).

PAGE 75

75 Figure 5 1. Simplified l igands computed by DFT. Each ligand was computed with Eu, Am, and Cm as the metal. The optimizations for each structure were D 3 or C 3 symmetric with less t han 2 kcal/mol difference in energy between the full y optimized and the symmetr y restricted structure. Ligand 5 1 was computed for comparison and the bond distance s were similar to previously published reports using a similar method. 60 Structure 5 2 represents a truncated version of 3 22 and 3 26 with the commonality of amide, pyridine, and benzimidazole donors. The triphenoxymethane and other atoms irrelevant to the electronic structure were removed. Structure 5 3 resembles the 3 33 donor atoms but the linkage to triphenoxymethane was replaced by a methyl group. In addition, the alkyl group of the benzimidazole ni trogen was been replaced with hydrogen in order to facilitate SCF convergence The removal of the e lectron donating methyl group may cause small electronic differences to the experimental comparison. The bond distances for the M L bonds are reported in Table 5 1, 5 2, and 5 3 where the pyridine, triazine, benzimidazole, amide, and triazole donor atom s are labeled as N py N triazine N bzd O amide and N triazole respectively. Table 5 1. Calculated M N bond distances for structure 5 1 d (M N py ) d (M N triazine ) [Eu(5 1) 3 ] 3+ 2.61 2.60 [Am(5 1) 3 ] 3+ 2.64 2.62 [Cm(5 1) 3 ] 3+ 2.64 2.63

PAGE 76

76 Table 5 2. Calculated M N bond distances for structure 5 2 d (M N py ) d (M N bzd ) d (M O amide ) [Eu(5 2) 3 ] 3+ 2.69 2.68 2.45 [Am(5 2) 3 ] 3+ 2.69 2.69 2.47 [Cm(5 2) 3 ] 3+ 2.69 2.67 2.47 Table 5 3. Calculated M N bond distances for structure 5 3 d (M N py ) d (M N bzd ) d (M N triazole ) [Eu(5 3) 3 ] 3+ n.d. n.d. n.d. [Am(5 3) 3 ] 3+ 2.68 2.63 2.63 [Cm(5 3) 3 ] 3+ 2.70 2.64 2.64 The re are only small difference s in bon d distances between Am and Cm complexes with each ligand. In each case the Cm distances were slightly longer but likely within the error of the calculation. The Eu complexes exhibit shorter bond s compared to the Am and Cm complexes emphasizing that separation based on size of the cation would not be efficient. The [Eu(5 3 ) 3 ] 3+ structure failed to converge and hence no data is reported. All N heterocyclic nitrogen donors have very similar bond lengths which are divergent from the oxygen donor bond lengths The increased hardness of the oxygen donors makes for a stronger bond and is reflected in the shorter bond distance. The solid state structure of Eu(5 1) 3 3+ and Cm(5 1) 3 3+ has been determined and the authors reported average nitrogen bond distances of 2.55 and 2.57 respectively 61 The geometry optimizations had longer bond lengths but should not significantly affect the molecular orbital analysis.

PAGE 77

77 5.3 Electronic Structure Analysis 5.3.1 Mulliken and Natural Charge Analysis Mulliken charges can be used as a computational assessment of the degree of covalency between a metal ligand bond. Generally the increased covalency between a neutral ligand and metal suggests a stronger bond. The charge on the metal should be 3 + for each computed molecule and any deviation from that value equat es to a change in electron density around the metal. A stronger electron density donation will lower the Mulliken charge. Mulliken charges are less intuitive if back bonding is occurring in the metal ligand interaction because when the metal donates electr on density back to the ligand the Mulliken charge is increased. The Mulliken charges for each calculation are shown in Table 5 4 The [Eu(5 3 ) 3 ] 3+ structure failed to converge and hence no data is reported. Table 5 4. Comparison of Mulliken Charges for ML 3 3+ 5 1 5 2 5 3 Eu 1.75 1.65 n.d. Am 1.61 1.33 1.52 Cm 1.44 1.28 1.40 The low value can be attributed to the fact that Mulliken charges generally suggest excessively covalent interactions, although global trends are accurate. 62 From the MO analysis presented later it was determined that the structures do not display any substantial back bonding thus the value shown is pure donation of the donor atom electrons in to the metal. Higher values of Eu complexes imply a more covalent bond occurs with the actinides than the lanthanides. 63, 64 The increased covalency for actinides is in agreement with the fact that the diffuse 5f orbitals can cont ribute to bonding in actinides, whereas, the energetically buried 4f orbitals of the lanthanides do

PAGE 78

78 not because the energy discrepancy between bonding orbitals is too high The calculations also show increased covalency for ligand 5 2 which can be explaine d by the hard oxygen donor of the methyl amide functional group. Since lanthanides and actinides are both considered hard acids it is not surprising that the largest charge donation occurs at the oxygen atom. The all nitrogen donor ligands, 5 1 and 5 3, sh ow similar results with the more electronegative triazole moiety of 5 3 decreasing the Mulliken charge more than the triazinyl moiety of 5 1. Overall the Mulliken charges were effective at determining which metal has more covalency with the ligand set and which ligand donates the most electron density to the metal. 5.3.2 Molecular Orbital Analysis The possible covalency of a ligand can be measured by absolute hardness. Since absolute hardness is proportional to the HOMO LUMO gap the difference in energy between the HOMO and LUMO of a ligand can give insight into how the ligand will interact with a metal 65 The HOMO LUMO gap for structures ( 5 1 ) 3 ( 5 2 ) 3 ( 5 3 ) 3 in their computed equilibrium geometry for binding is 88.2, 98.7, and 99.9 kcal/mol, respectively. This shows the expected softness of 5 1 in comparison to the other ligands. It is likely that the dominance of the benzimidazole moiety on the relative HOMO LUMO energies is the reason why the gap energies f or structure 5 2 and 5 3 are so similar. In fact, the MO for the AnL 3 3+ complexes of 5 2 and 5 3 are mainly comprised of ligand centered orbitals in the frontier orbital region. The electron density of the HOMO in structures 5 2 and 5 3 is not entirely com prised of the benzene ring but also of the nitrogen donor as seen in Figure 5 2. Since some of the HOMO is comprised of the nitrogen donor atom it is reasonable to assume that it could be used for bonding if the energetics and orbital symmetries of the met al and ligand were more congruent. To

PAGE 79

79 alter energetic matching i t might be possible to modify benzimidazole to a lower energy by replacing substituents The likely areas of modification include N alkylation or aromatic substitution on the benzene ring. Ano ther modification could be to add more conjugation to the structure in order to lower the HOMO LUMO gap. Figure 5 2. Representation of benzimidazole centered HOMO of Cm( 5 3 ) 3 3+ An example of good energetic matching occurs in the HOMO 9 of Am(5 1) 3 3+ which consists of 39.2% f orbital inter acting with 22.7% of the ligand where the remaining percentage of the MO is not involved with the ligand to metal donation. The MO representation of HOMO 9 is shown in Figure 5 3. The bonding interaction is tracked i n the calculation as having a positive overlap population which is clearly seen in Figure 5 3 where the orbitals on one of the three ligands has the triazinyl nitrogens directly facing the f xyz orbital on Am.

PAGE 80

80 Figure 5 3. Representation of ligand to metal donation in the HOMO 9 of Am(5 1) 3 3+ A n example of proper energetic matching is shown in Figure 5 4. T he HOMO 1 composition of Am( 5 1 ) 3 3+ is 42.8 % ligand and 57 .2 % meta l and the interaction occur s in the frontier orbitals. E ven though there is a good energetic matching for the HOMO 1 there is no ove rlap of the orbitals. Figure 5 4 depicts a negative overlap population where the orbital on the triazinyl nitrogen is not directly facing the metal orbital in the sigma bonding arrangement one would expect T he multiple contributions that make up many of the molecular orbitals can create a complex bonding environment For instance, the HOMO 1 consists of 52.1% f orbital anti bonding to 31.6% of the ligand, 2.6% f orbital in an ionic interaction with the ligan d, and 2.5% d orbital overlapping with

PAGE 81

8 1 the ligand in a ligand to metal donation. Since the dominating interaction of the MO is anti bonding the representation of HOMO 1 shown in Figure 5 4 is mainly anti bonding with the f orbital slightly distorted from t he d orbital contribution. Figure 5 4 The HOMO 1 of Am(5 1) 3 3+ T he black lines signify anti bonding interactions Another instance that can be mistaken for a bonding interaction is when there is zero overlap population for the interaction. An example of zero overlap population occurs in the HOMO 12 of Am(5 1 ) 3 3+ and is shown in Figure 5 5. The zero overlap population can be viewed as an ionic bond where the electrostatic interaction between

PAGE 82

82 the negatively charged nitrogen and the positively charged met al are attracted to each other but no orbital overlap occurs. These non covalent type interactions are similar in lanthanides and actinides thus are not tracked since they should not be an important factor in selectivity. Figure 5 5 The HOMO 1 2 of Am(5 1) ) 3 3+ does not display any overlap of orbitals since the overlap population was computed as zero. The MO interaction diagrams of Eu(5 1) 3 3+ Am(5 1) 3 3+ and Cm(5 1) 3 3+ are shown in Figure 5 6. With the exception of Am(5 1) 3 3+ most frontier orbitals have 96 % ligand composition When viewing the MO interaction diagrams it is important to note that although there appears to be a bond forming it is also possible for the interaction to be anti bonding or ionic bonding interaction where there is no overlap betw een the metal

PAGE 83

83 and ligand orbitals. For clarity, all MO diagrams that have contributions below 4% do not have a line drawn as an interaction. Figure 5 6 MO for Eu(5 1) 3 3+ Am(5 1) 3 3+ and Cm(5 1) 3 3+ Metal energy levels shifted by approximately 18 eV an d ligand orbitals by 8 eV. In particular, Eu( 5 1 ) 3 3+ shows poor energetic matching between the metal and ligand because of the energetically buried 4f orbital s of europium. The first positive overlap populated orbital of Eu( 5 1 ) 3 3+ occurs in the HOMO 35 with a composition of 3.4% d orbital and 53.1% ligand. The small composition of the d orbital in HOMO 35 implies that the energetics of the interaction is poor. There are no f orbital contributions

PAGE 84

84 found until the HOMO 120 at 33.414 eV with only a 1.1% composition which is deemed negligible in terms of covalency. In comparison Am(5 1) 3 3+ has the first f orbital positive overlap population at HOMO 9 with 39.2% f orbital interacting with 22.7% of the ligand. For the Cm(5 1) 3 3+ calcula tion the first f orbital positive overlap population occurred at the HOMO 33 with 29.8% metal and 22.5% ligand. Even though the HOMO 32 is lower in energy it is a consequence of Cm being lower in energy than Am. The Cm(5 1) 3 3+ calculation also has several pairs of degenerate orbitals in the electronic structure (Figure 5 7). As previously mentioned, the re is no significant back bonding in the structures therefore there are no lines drawn from the HOMO to the LUFO of the ligand in the MO diagram (Figure 5 6). Figure 5 7 Degenerate orbitals HOMO 43 and HOMO 44 of Cm(5 1) 3 3+ at 17.803 eV The MO interaction diagram for metals Eu, Am, and Cm complexed by ligand 5 2 is shown in Figure 5 8. The first few ligand centered MOs of Eu(5 2) 3 3+ Am(5 2) 3 3+ and Cm(5 2) 3 3+ are comprised of benzimidazole character and look similar to the HOMO of

PAGE 85

85 Cm(5 3) 3 3+ shown in Figure 5 2. The Eu(5 2) 3 3+ calculation is similar to Eu(5 1) 3 3+ in that there are no f orbital positive overlap populations in the frontier orbitals. Figure 5 8 MO for Eu(5 2) 3 3+ Am(5 2) 3 3+ and Cm(5 2) 3 3+ Metal energy levels shifted by approximately 18 eV and ligand orbitals by 8 eV. When comparing the MO interaction diagram of Am(5 2) 3 3+ with that of Am(5 1) 3 3+ there are less interactions between Am and ligand (5 2) 3 than with ligand (5 1) 3 When the overlap population was analyzed it showed no f orbital overlap in the frontier orbitals and only small bonding contributions from the d orbitals. The lack of covalent interactions with Am(5 2) 3 3+ is a sign that selectivity between Am and Eu might not be

PAGE 86

86 achieved using ligand 5 2. Unlike Am(5 2) 3 3 + the calculation of Cm(5 2) 3 3 + does show positive overlap populations with the f orbitals. An example of the ligand to metal donation for the HOMO 36 of Cm(5 2) 3 3 + is shown in Figure 5 9. The composition of HOMO 36 is 36.7% f orbital and 19.7% ligand The similar compositions mean the energetics are favorable for the interaction and that ligand 5 2 may be selective for Cm over Eu and possibly other lathanides. As with ligand 5 1, there is no substantial back bonding in ligand 5 2. Figure 5 9 Representation of ligand to metal donati on in the HOMO 36 of C m(5 2) 3 3+ The MO of ligand 5 3 shows many similarities to ligand 5 2 because of the common benzimidazole but the triazole moiety displayed unique properties as well. The MO diagram s for Am and Cm complexed by l igand 5 3 are shown in Figure 5 10

PAGE 87

87 Figure 5 10 MO for Am(5 3 ) 3 3+ and Cm(5 3 ) 3 3+ Metal energy levels shifted by approximately 18 eV and ligand orbitals by 8 eV.

PAGE 88

88 As discussed previously the HOMO is dominated by the benzimidazole in both structures. When m oving down in energy by approximately 1 eV from the HOMO interactions begin to appear in the Am complex and likewise in the Cm structure at appr oximately 2 eV from the HOMO. The calculation of Am(5 3) 3 3+ showed very similar results to Am(5 2) 3 3 + in that there was no positive overlap population with the f orbitals. Likewise, the calculation of Cm(5 3) 3 3 + showed positive f orbital overlap similar to the calculation of Cm(5 2) 3 3 + thus giving evidence for ligand 5 3 to be selective for Cm over Eu and possibly other lanthanides. One of the main differences of Cm(5 3) 3 3 + from the other calculations is that it was the only calculation of have an absence o f d orbital contribution in the frontier orbitals. All of the calculations showed at least small percentages of positive overlap populations of the d orbitals. It is not likely that the absence of d orbitals would cause changes in selectivity since the d o rbital contributions are so low. 5.4 Conclusions The DFT study presented in this chapter calculated three ligand structures with Eu, Am, and Cm. Because of convergence problems two of the Eu complexes were not presented. The geometry optimization results were compared amongst the calculations and experimental structural data for Cm(5 1) 3 3+ The Mulliken charg es of each computed structure were calculated. The Mulliken charges show that harder donors like oxygen donate more electron density to the metal than the nitrogen donor groups. The Mulliken charges on Eu were higher than Am and Cm because the energetic matching of orbitals did not facilitate donation. According to the MO analysis t he Cm charges were slightly lower than Am despite Am complexes having mo re ligand to metal donations. The discrepancy between charges was increased when using the softer

PAGE 89

89 ligan d 5 1 and 5 3 as compared to 5 2 with harder donor s Since the softer ligands cause a larger difference between Am and Cm which implies back donation fro m Am might be used to explain the difference but the population analysis did not uncover any large degree of back donation. The other option to explain the difference in Mulliken charges is that the ECP for Cm uses 36 valance electrons while Eu and Am use an ECP with 35 electrons, therefore, the different basis set causes the minor discrepancy The population analysis investigated the orbital environment to try and evaluate what aspect of ligand design facilitates selective separation of lanthanides from a ctinides. The bonding in each structure is mainly ionic with covalency generated by ligand metal donation. The ligand to populating empty metal ns, (n 1)d and (n 2)f orbitals where n=7 for acti nides and n=6 for lanthanides. To help summarize the results of the population analysis Table 5 5 shows the average contributions of all the positive overlap populations in the frontier orbitals for each calculation. Table 5 5. Orbital compositions (%) of bonding interactions Eu Am Cm (5 1) (5 2) (5 3) (5 1) (5 2) (5 3) (5 1) (5 2) (5 3) f orbital nd 25.0 17.0 33.7 32.1 Ligand (f) nd 33.6 19.0 10.5 6.5 d orbital 3.3 2.4 nd 1.9 1.5 2.4 1.3 2.6 Ligand (d) 48.6 28.7 nd 29.7 13.6 19.4 32.5 20.6 Table 5 5 has a dashed line for contributions that did not appear in the calculation but may be pr esent in less than 1% contributions. The overall bonding for the frontier orbitals of Eu with any ligand is that it only has minimal d orbital composition with large differences in composition of the metal and the ligand. The difference in composition implies that the energetics of the orbitals do not match and that the covalency is less

PAGE 90

90 significant. The calculations show the Eu complexes are mainly comprised of ionic bonds. For the Am calculations ligand 5 1 was the only ligand that showed positive overlap populations for the f orbitals. All of the Am calculations had small amounts of d orbital bonding but since they are small i n composition the covalency of the d orbitals is n ot significant. The ligand analogs 5 2 and 5 3 did not show positive overlap populati ons with the f orbitals, hence they are likely not to be selective for Am over Eu. The Cm calculations all have positive overlap populations with the f orbitals. The MO c ompositions of Cm(5 1) 3 3+ are not as high or evenly matched between metal and ligand as the Am(5 1) 3 3+ calculation but they are likely to be significant for covalent interactions The Cm(5 2) 3 3+ and Cm(5 3) 3 3+ calculations have a n average ligand composition lower than the average metal compositions (Table 5 5) which might be altered with modification s that effect the energetics of the ligands. The Cm(5 3) 3 3+ complex was the only calculation to not have d orbital contributions in the frontier orbitals. The triazole moiety is the likely culprit but the low contributions of d orbitals in the other structures indicate that it will not significantly alter potentia l selectivity. 5.5 Computational Details All density functional theory (DFT) calculations were performed with the Gaussian 66 Spin unrestricted multiplicities were used to account for the formal f n configuration s for lanthanides and actinides. Each complex had its geometry optimized by using the B3LYP functional, the 6 31G* basis set for C, H, N, and O atoms and a small core relativistic effective core potential (RECP) for Eu, Am, and Cm. The RECP data was taken from the EMSL basis set library which used work from the Stuttgart and Dresden groups. 67, 68 Subsequent single point calculations were performed with the optimized geometries in order t o format them for input into the AOMix program package.

PAGE 91

91 Using the AOMix program package the molecular orbitals, natural charges and populations were analyzed. 69, 70 Spin contamination was closely monitored and was found to be close to ideal values.

PAGE 92

92 CHAPTER 6 SUMMARY In this proj ect ligands that facilitate the selective partitioning of lantha nides and actinides were designed, synthesized, and analyzed The objectives of the project were to develop a rational means of ligand design through semi emp irical calculations, synthesize the desired ligand s collect experimental data from the synthesized ligands, and use DFT to model variations of th e ligand s and evaluate any features that appear to contribute to the selectivity of binding actinides over lanthanides. The penultimate goal was to produce ligands that had experimental and theoretical evidence supporting the selective separation of actinides from the nuclear waste stream. The semi empirical (SE) calculation project tested several metal ligand complexes for their twist angles in order to pre dict which complex would be closer to the ideal TTP geometry. When comparing the calculated structure s to the expe rimental data obtained from the crystal structure s the calculations only showed small difference s in twist angle The method was unable to reproduce accurate bond lengths for structure 2 7 with the major deviation coming from the benzimidazole nitrogen do nor. For accurate bond lengths DFT methods would be preferred in place of SE calculations but DFT becomes impractical with a large selection of ligands due to the increased computational demand In this project the SE calculations outperform ab initio met hods in predictive ligand design when deciding between multiple candidates. The synthetic project investigated several synthetic pathways for producing ligands containing benzimidazole p y ridyl moieties attached to a triphenoxymethane molecular platform. T he ligands synthesized were chosen based on SE calculation data obtained from a library of possible ligands. The synthesis of the ligands initially relied on

PAGE 93

93 previously synthesiz ed molecules as intermediates for the novel nonadentate final products. Overal l three ligands were successfully synthesized of which the t wo amide linkage ligands, 3 22 and 3 26, showed extraction efficiency of approximately 5% for the lanthanides. The triazole functionalized ligand, 3 33, showed less than 1% extraction efficiency for the lanthanide s Since the goal of the project is the selective binding of actinides low extraction efficiency of lanthanides is actually desirable Theoretically, if the ligand can covalently bind a lanthanide it can likely bind actinides due to their extended 5f orbital s All three ligands synthesized show promising initial results for selectively binding actinides over lanthanides although actinide experimentation is needed for confirmation. In the DFT co mputational project the donor groups of the synthesized molecules are compared to structure 5 1. The selectivity exhibite d by 5 1 toward trivalent lanthanides and actinides is complex, but through theoretical investigations trends are starting to appear that may help fine tune l igand design for optimal performance. Since actinides contain more diffuse f orbitals than lanthanides covalency c an be exploited to separate Am and Cm from Eu The highest degree of covalency would occur with ligands that facilitate back bonding in additi on to ligand to metal donation. Back bonding was found to be inadequate for the ligands used in this study mainly because the hardness of the metals. S ince the Am(5 1) 3 3+ calculation did not show back bonding even though ligand 5 1 demonstrates selectivity it appears that back bonding is only an ideality and not necessary for selectivity Utility of using benzimidazole type ligands seem applicable if the metal ligand energetics can be matched up thro ugh modification of the ligand. By tracking the positiv e overlap populations in the

PAGE 94

94 calculations it was deter mined that ligands 5 2 and 5 3 c ould be selective for Cm. The calculations presented show a promising extrapolation to the synthetic nonadentate analogs and gives some validity to what experimental data would be expected with acti nides. The objectives for this project included the guided design, synthesis, analysis, and theoretical evaluation of a series of ligands for the selective separation of actinides over lanthanides for use in nuclear waste reproc essing. The work presented herein was able to complete all objectives except experimentation with the actinides The SE calculations proved to be a wor thy predictor for ligand design. S ynthesis and subsequent experimentation of ligands was performed while DFT calculations provided a systematic theoretical approach for predicting selectivity between the lanthanides and actinides

PAGE 95

95 APPENDIX A NUCLEAR MAGNETIC RES ONANCE SPECTRA OF SY NTHESIZED COMPOUNDS The following Nuclear Magne tic Resonance (NMR) spectra ar e provided as an a ssessment of purity for the synthesized c ompounds discussed in Chapter 3. Only the NMR spectra of compounds that did not include elemental analysis are shown. Figure A 1. NMR of 3 21

PAGE 96

96 Figure A 2. NMR of 3 20 Figure A 3. NMR of 3 23

PAGE 97

97 Figure A 4. NMR of 3 25 Figure A 5. NMR of 3 30

PAGE 98

98 Figure A 6. NMR of 3 31b Figure A 7. NMR of 3 14c

PAGE 99

99 Figure A 8 NMR of 3 15c Figure A 9. NMR of 3 33

PAGE 100

100 Figure A 10. NMR of 3 22 Figure A 11 NMR of 3 26

PAGE 101

101 References (1) John, Moens; U.S. Nuclear Generation of Electricity ;EIA Survey Form 906; http://www.eia.doe.gov/cneaf/nuclear/page/nuc_generation/gensum.html (2) International Energy Ou tlook 2007. DOE/EIA 0484(2007). http://www.eia.doe.gov/oiaf/ieo/index.html (3) Madic, C.; Bourges, J.; J.F., D. In In International Conference on Accelerator Driven Transmutation Tech nology and Applications; (4) Total U.S. Commercial Spent Nuclear Fuel Discharges, 1968 2002. Energy Information Administration ; Form RW 859 http://www.eia.doe. gov/cneaf/nuclear/spent_fuel/ussnfdata.html (5) Country Profile France Nuclear Energy & Radioactive Waste Management. http://www.andra.fr/IMG/pdf/country profile.pdf (6) Baisde n, P. A.; Choppin, G. R. In Radiochemistry and Nuclear Chemistry; Encyclopedia of Life Support Systems ; Eolss Publishers: Oxford, UK, 2007; (7) Sasahara, A.; Matsumura, T.; Nicolaou, G.; Papaioannou D. J. Nuc. Sci. Tech 2004 41 448 456. (8) Madic, C .; Lecomte, M.; Baron, P.; Boullis, B. C. R. Physique 2002 3 797 811. (9) Mughabghab, S. F.; Divadeenam, M.; Holden, N. E. Academic Press 1981 (10) Mathur, J.; Murali, M.; Nash, K. Solvent Extraction & Ion Exchange 2001 19 357. (11) Dam, H. H.; Reinhoudt, D. N.; Verboom, W. Chem. Soc. Rev. 2007 36 367 377. (12) Recycle and reuse of materials and components from waste streams of nuclear fuel cycle facilities IAEA WMDB ST 1 International Atomic Energy Agency, 2000 (13) Prade l, P. In In The organic waste treatment in UP3 La Hague; Proc. 3rd Int. Conf. on Nuclear Fuel Reprocessing and Waste Management; pp 1101 1106. (14) Cuillerdier, C.; Musikas, C. Sep. Sci. Technol., 1991 26 1229. (15) Chan, G. Y. S.; Drew, M. G. B.; Hudson, M. J.; Iveson, P. B.; Liljenzin, J. .; Skalberg, M.; Spjuth, L.; Madic, C. J. Chem. Soc. Dalton Trans. 1997 649. (16) Spjuth, L.; Liljenzin, J. .; Hudson, M. J.; Drew, M. G. B.; Iveson, P. B.; Madic, C. Solvent Extr. Ion Exch. 2000 18 1.

PAGE 102

102 (17) Zhu, Y.; Chen, J.; Jiao, R. Sep. Sci. Technol. 1996 31 2723 2731. (18) Zhu, Y.; Chen, J.; Jiao, R. Solvent Extr. Ion Exch. 1996 14 61 68. (19) Drew, M. G. B.; Hill, C.; Hudson, M. J.; Iveson, P. B.; Madic, C.; Vaillant, L.; Youngs, T. G. A. N ew J. Chem. 2004 28 462 470. (20) C. Madic, M. J. Hudson; J. O. Liljenzin; J. P. Glatz; R. Nannicini; A. Facchini; Z. Kolarik; R. Odoj, In Fourth International Conference of Nuclear Fuel Reprocessing and Waste Management, London, UK, 1994 (21) Addison, A. W.; Burke, P. J. J. Heterocycl. Chem. 1981 18 803 805. (22) Jensen, T. B.; Scopelliti, R.; Bunzli, J. G. Inorg. Chem. 2006 45 7806 7814. (23) Nash, K. L. Solvent Extr. Ion Exch. 1993 11 729. (24) Choppin, G. R. J. Alloys Compds 2002 344 55 5 9. (25) Guoxin, T.; Yaning, Z.; Jing, Z. Inorg. Chem. 2003 43 735 741. (26) Arnaud Neu, F.; Bahmer, V.; Dozol, J. F.; Grattner, C.; Jakobi, R. A.; Kraft, D.; Mauprivez, O.; Rouquette, H.; Schwing Weill, M.; Simon, N.; Vogt, W. J. Chem. Soc. Perkin Trans. 2 1996 1175. (27) Matthews, S. E.; Saadioui, M.; Bahmer, V.; Barboso, S.; Arnaud Neu, F.; Schwing Weill, M.; Carrera, A. G.; Dozol, J. F. J. Prakt. Chem. 1999 341 264. (28) Barboso, S.; Carrera, A. G.; Matthews, S. E.; Ar naud Neu, F.; Bohmer, V.; Dozol, J. F.; Rouquette, H.; Schwing Weill, M. J. J. Chem. Soc. Perkin Trans 1999 719 723. (29) Delmau, L. H.; Simon, N.; Schwing Weill, M. J.; Arnaud Neu, F.; Dozol, J. F.; Eymard, S.; Tournois, B.; Gruttner, C.; Musigmann, C .; Tunayar, A.; Bohmer, V. Sep. Sci. and Technol. 1999 6 836 876. (30) Arduini, A.; Bohmer, V.; Delmau, L.; Desreux, J. F.; Dozol, J. F.; Carrera, M. A. G.; Lambert, B.; Musigmann, C.; Pochini, A.; Shivanyuk, A.; Ugozzoli, F. Chem. Eur. J. 2000 6 2135. (31) Gutsche, C. D. In Calixarenes revisited; The Royal Society of Chemistry: Cambridge, 1998 ; (32) Arnaud Neu, F.; Schwing Weill, M. .; Dozol, J. In Calixarenes 2001; Kluwer Academic Publishers: Dordrecht, 2001 ;

PAGE 103

103 (33) Miguirditchian, M.; Guillaneux, D.; Guillaumont, D.; Moisy, P.; Madic, C.; Jensen, M. P.; Nash, K. L. Inorg. Chem. 2005 44 1404 1412. (34) Peters, M. W.; Werner, E. J.; Scott, M. J. Inorg. Chem. 2002 41 1707 1716. (35) Dinger, M. B.; Scott, M. J. European Journal of Or ganic Chemistry 2000 2467 2478. (36) Matloka, K.; Gelis, A.; Regalbuto, M.; Vandegrift, G.; Scott, M. J. J. Chem. Soc. Dalton Trans. 2005 3719. (37) Rudzevich, V.; Schollmeyer, D.; Braekers, D.; Desreux, J. F.; Diss, R.; Wipff, G.; Bahmer, V. J. Org. Chem. 2005 70 6027. (38) Matloka, K. SYNTHETIC DEVELOPMENT OF C 3 SYMMETRIC TRIPHENOXYMETHANE BASED REAGENTS FOR THE SELECTIVE RECOGNITION AND SEQUESTRATION OF LANTHANIDES AND ACTINIDES, 2006 (39) Piguet, C.; Bernardinelli, G.; Bocquet, B.; Quattropan i, A.; Williams, A. F. J. Am. Chem. Soc. 1992 114 7440 7451. (40) MOPAC2009, James J. P. Stewart, Stewart Computational Chemistry, Colorado Springs, CO, USA, HTTP://OpenMOPAC.net ( 2009 ) (41) Rocha, G. B.; Freire, R. O.; da Costa, N. B.; de Sa, G. F.; Simas, A. M. Inorg. Chem. 2004 43 2346 2354. (42) Levine, I. N., Quantum Chemistry 5 th Ed; Prentice Hall: 2000 ; pp 602 603. (43) Pyykko, P. Chem. Rev. 1988 88 563. (44) Guillaumont, D. Journal of Molecular Structure: THEOCHEM 2006 771 105 110. (45) Guillaumont, D. The Journal of Physical Chemistry A 2004 108 6893 6900. (46) Freire, R.; Rocha, G.; Simas, A. Journal of Molecular Modeling 2006 12 373 389. (47) Seitz, M.; Alzakhem, N. Journal of Chemical Information and Modeling 2010 50 217 220. (48) Ludwig, O.; Schinke, H.; Brandt, W. Journal of Molecular Modeling 1996 2 341 350. (49) Elhabiri, M.; Scopelliti, R.; Bunzli, J. G.; Piguet, C. J. Am. Chem. Soc. 1999 121 10747 10 762.

PAGE 104

104 (50) Matloka, K.; Gelis, A.; Regalbuto, M.; Vandegrift, G.; Scott, M. J. Dalton Trans. 2005 3719 3721. (51) Canard, G.; Koeller, S.; Bernardinelli, G.; Piguet, C. J. Am. Chem. Soc. 2008 130 1025 1040. (52) Chuang, C.; dos Santos, O.; Xu, X.; Can ary, J. W. Inorg. Chem. 1997 36 1967 1972. (53) Matloka, K.; Sah, A. K.; Peters, M. W.; Srinivasan, P.; Gelis, A. V.; Regalbuto, M.; Scott, M. J. Inorg. Chem. 2007 46 10549 10563. (54) Mitra, R.; Peters, M. W.; Scott, M. J. Dalton Trans. 2007 3924 3935. (55) R. Malevann, E.N. Tsvetkov, M. Kabachni, Zh. Obshch. Khim. 1971 41 1426. (56) Z. Marczenko, in: E. Horwood (Ed.) In Separation and Spectrophotometric Determination of Elements; Halsted Press: New York, 1986 ; (57) van der Sluis, P.; Spek, A. L. Acta Crystallographica Section A 1990 46 194 201. (58) SHELXTL6 ( 2000 ). Bruker AXS, Madison, Wisconsin, USA. (59) Spek, A. L. Acta Crystallographica Section D 2009 65 148 155. (60) Petit, L.; Adamo, C.; Maldivi, P. Inorg. Chem. 2006 45 8517 8522. (61) Maldivi, P.; Petit, L.; Adamo, C.; Vetere, V. Comptes Rendus Chimie 2007 10 888 896. (62) Mulliken, R. S. J. Chem. Phys. 1955 23 1833 1840. (63) Petit, L.; Joubert, L.; Maldivi, P.; Adamo, C. J. Am. Chem. Soc. 2006 128 2190 2191. (64) De Proft, F.; Martin, J. M. L.; Geerlings, P. Chemical Physics Letters 1996 250 393 401. (65) Pearson, R. G. Proc. Natl. Acad. Sci. U. S. A. 1986 83 8440 8441.

PAGE 105

105 (66) Gaussian 03, Revision E .02 J. A.; Vreven, T.; Kudin, K. N.; B urant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O. ; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. ; Gaussian, Inc., Wa llingford CT, 2004 (67) Feller, D. Journal of Computational Chemistry 1996 17 1571 1586. (68) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. Journal of Chemical Information and Modeling 20 07 47 1045 1052. (69) Gorelsky, S. I.; Lever, A. B. P. Journal of Organometallic Chemistry 2001 635 187 196. (70) S. I. Gorelsky, AOMix: Program for Molecular Orbital Analysis http://www.sg chem.net/, University of Ottawa, version 6.4, 2010

PAGE 106

106 BIOGRAPHICAL SKETCH Ivan Fabe Dempsey Hyatt was bor n in Greensboro, North Carolina in the year of 1982 For the next eighteen years h e lived in the small rural town of Staley, North Carolina and attended Eastern Randolph High School. In 2001, he began his undergraduate studies at East Carolina University under the supervision of Professor Andrew Sargent. His research consisted of computationally modeling the reaction mechanism of rhodium catalyzed hydroacylation. In 2005, he graduated with a B achelor of S ci ence in chemistry and a B achelor in A rts in mathematics and enrolled into the University of Florida as a Grinter Fellow. He initially joined the physical chemistry division but after his first year in graduate school he switched to the inorganic chemistry division under the supervision of Dr. Michael J. Scott. He completed the requirements for the degree of Doctor in Philosophy in December 2010.