SYNTHESIS, POLYMERIZATION, AND METAL BINDING STUDIES OF NEW PHOSPHAALKENES BEARING DONOR SUBSTITUENTS by Amber Michelle Juilfs B. Sc., Illinois Institute of Technology, 2009 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Chemistry) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2012 © Amber Michelle Juilfs, 2012 ii Abstract A new P-Mes (Mes = 2,4,6-trimethylphenyl) phosphaalkene bearing a chelating C-substitutent has been prepared: MesP=C(Ph)(o-DMA) (o-DMA = ortho- dimethylaniline) (1a). Phosphaalkene 1a was prepared by the standard phospha-Peterson reaction. Radical polymerization of 1a afforded [MesP-C(Ph)(o-DMA)]n (2a). Lithiation followed by protonation of 1a and previously synthesized phosphaalkene MesP=C(Ph)(2- Pyd) (2-Pyd = 2-pyridine) (2a) afforded model compounds (Mes)(Me)P-CH(Ph)(o- DMA) (3a) and (Mes)(Me)P-CH(Ph)(2-Pyd) (3b), respectively. Model compounds 3a and 3b were complexed to Pd(II) resulting in metal complexes (Mes)(Me)P-CH(Ph)(o-DMA)PdCl2 (4a), [(Mes)(Me)P-CH(Ph)(o- DMA)]2PdCl2 (4b), and (Mes)(Me)P-CH(Ph)(2-Pyd)PdCl2 (4c), respectively. The metal binding properties of the metal complexes were investigated by NMR spectroscopy and X-ray crystallography. The results showed that metal complex 4b coordinated Pd(II) in a two to one fashion by analysis of X-ray crystallography whereas complex 4c coordinated Pd(II) in a bidentate fashion through the phosphorus and nitrogen atoms. Crystals were not obtained for complex 4a, however NMR spectroscopy shows the complex differs from the free ligand 3a and the two to one complex 4b. iii Preface Some of the work in this thesis was achieved in collaboration with other researchers. All crystallographic data were obtained by Paul Siu and Spencer Serin with some assistance from Brian Patrick using the departmental X-ray diffractometer. The solution and refinement of the molecular structure of MesP=C(Ph)(o-DMA) (o-DMA = ortho-dimethylaniline) was completed by Paul Siu. The solution and refinement of the molecular structure of (Mes)(Me)P-C(Ph)(2-Py)PdCl2 (2-Py = 2-pyridine) was completed by Spencer Serin. The solution and refinement of the molecular structure of ((Mes)(Me)P-C(Ph)(o-DMA))2PdCl2 was done by Brian Patrick. All of the work presented in Chapters 2 and 3 was completed by myself. iv Table of Contents Abstract ............................................................................................................................... ii Preface ................................................................................................................................ iii Table of Contents ............................................................................................................... iv List of Tables .................................................................................................................... vii List of Figures .................................................................................................................. viii List of Abbreviations ...........................................................................................................x Acknowledgements ........................................................................................................... xii Dedication ........................................................................................................................ xiii Chapter One Introduction ....................................................................................1 1.1 Metal Scavenging ..............................................................................................1 1.1.1 Metal Scavenging via Distillation ....................................................2 1.1.2 Metal Scavenging via Extraction .....................................................4 1.1.3 Metal Scavenging via Crystallization ..............................................7 1.1.4 Metal Scavenging via Adsorption ..................................................10 1.1.4.1 Trimercaptotriazine ...........................................................11 1.1.4.2 Polystyrene-bound TMT ...................................................13 1.1.4.3 Polystyrene-bound Ethylenediamine ................................15 1.1.4.4 Silica-bound Metal Scavengers .........................................18 1.2 Phosphorus-Carbon Polymers .........................................................................22 1.3 Research Objectives ........................................................................................24 1.4 Contributions by Other Researchers to This Work .........................................26 v Chapter Two Synthesis and Polymerization Studies of New Phosphaalkenes Bearing Donor Substituents as Potential Chelating Ligands ...27 2.1 Introduction .....................................................................................................27 2.1.1 2-Pyridyl-Substituted Phosphaalkene ............................................27 2.2 Results and Discussion ...................................................................................30 2.2.1 Synthesis of Phosphaalkene 2.3 .....................................................32 2.2.2 X-Ray Crystallographic Analysis of Phosphaalkene 2.3 ...............39 2.2.3 Synthesis of Phosphaalkene 2.5 .....................................................43 2.2.4 Polymerization of Phosphaalkene 2.3 ............................................46 2.3 Summary .........................................................................................................51 2.4 Experimental ...................................................................................................52 2.4.1 Preparation of MesP=C(Ph)(o-DMA) (2.3) ...................................53 2.4.2 Preparation of [MesP-C(Ph)(o-DMA)]n (2.17) ..............................54 Chapter Three Synthesis and Metal Binding Studies of Phosphaalkene Model Compounds and Complexes Bearing Donor Substituents as Potential Chelating Ligands ........................................................56 3.1 Introduction ......................................................................................................56 3.2 Results and Discussion ...................................................................................57 3.2.1 Synthesis of Model Compound 3.3 ................................................57 3.2.2 Synthesis of Model Compound 3.4 ................................................62 3.2.3 Synthesis of Model Complex 3.5 ...................................................65 3.2.4 X-Ray Analysis of Model Complex 3.5 ........................................69 3.2.5 Synthesis of Model Complex 3.8 ...................................................74 vi 3.2.6 Synthesis of Model Complex 3.9 ...................................................76 3.2.7 X-Ray Analysis of Model Complex 3.9 ........................................79 3.3 Summary .........................................................................................................83 3.4 Experimental ...................................................................................................84 3.4.1 Preparation of (Mes)(Me)P-CH(Ph)(2-Py) (3.3) ...........................85 3.4.2 Preparation of (Mes)(Me)P-CH(Ph)(o-DMA) (3.4) ......................85 3.4.3 Preparation of (Mes)(Me)P-CH(Ph)(2-Py)PdCl2 (3.5) ..................86 3.4.4 Preparation of (Mes)(Me)P-CH(Ph)(o-DMA)PdCl2 (3.8) .............87 3.4.5 Preparation of [(Mes)(Me)P-CH(Ph)(o-DMA)]2PdCl2 (3.9) .........87 Chapter Four Overall Conclusions and Future Work ......................................88 4.1 Summary of Thesis Work ................................................................................88 4.2 Future Work ....................................................................................................93 4.2.1 New Bidentate and Tridentate Phosphaalkenes .............................93 4.2.2 Copolymerization of Phosphaalkenes ............................................93 4.2.3 Metal-Scavenging Tests .................................................................94 4.3 Concluding Remarks .......................................................................................94 Bibliography .....................................................................................................................95 Appendices ........................................................................................................................99 Appendix A: 2D NMR Experiments .....................................................................99 vii List of Tables Table 1.1 Removal of Palladium using Polystyrene-bound TMT .............................14 Table 1.2 Removal of Palladium from 1.36 using Silica-bound Metal Scavengers ..20 Table 1.3 Selected Examples of Metal Scavenging Techniques ...............................21 Table 2.1 Summary of NMR Resonances for Phosphaalkene 2.3 .............................39 Table 2.2 X-ray Crystallographic Data for 2.3 ..........................................................40 Table 2.3 Important Metrical Parameters for 2.1 and 2.3 ..........................................43 Table 3.1 X-ray Crystallographic Data for 3.5 ..........................................................71 Table 3.2 Important Metrical Parameters for P,N Bidentate Ligands bound to Palladium (II) .............................................................................................73 Table 3.3 X-ray Crystallographic Data for 3.9 ..........................................................81 Table 3.4 Important Metrical Parameters for 3.9 and 3.5 ..........................................82 viii List of Figures Figure 1.1 Several polystyrene-bound amines useful for metal scavenging ...............15 Figure 1.2 Several silica-bound chelating atoms useful for metal scavenging ...........18 Figure 2.1 Targeted functionalized phosphaalkenes ...................................................30 Figure 2.2 31P{1H} NMR spectrum (CDCl3) of 2.3 ....................................................36 Figure 2.3 1H NMR spectrum (CDCl3) of 2.3 .............................................................37 Figure 2.4 13C NMR spectrum (CDCl3) of 2.3 ............................................................38 Figure 2.5 Molecular structure of MesP=C(Ph)(o-DMA) (2.3) ..................................41 Figure 2.6 31P{1H} NMR spectrum (CDCl3) of polymer 2.17 ....................................49 Figure 2.7 1H NMR spectrum (CDCl3) of polymer 2.17 .............................................50 Figure 2.8 GPC (THF) trace of polymer 2.17 .............................................................51 Figure 3.1 31P{1H} NMR spectrum (CDCl3) of 3.3 ....................................................59 Figure 3.2 1H NMR spectrum (CDCl3) of 3.3 .............................................................60 Figure 3.3 13C NMR spectrum (CDCl3) of 3.3 ............................................................61 Figure 3.4 31P{1H} NMR spectrum (CDCl3) of 3.4 ....................................................63 Figure 3.5 1H NMR spectrum (CDCl3) of 3.4 .............................................................64 Figure 3.6 13C NMR spectrum (CDCl3) of 3.4 ............................................................65 Figure 3.7 31P{1H} NMR spectrum (CDCl3) of 3.5 ....................................................67 Figure 3.8 Comparison of 1H NMR spectrum (CDCl3) of 3.5 against the 1H NMR spectrum (CDCl3) of 3.3 ............................................................................68 Figure 3.9 Molecular structure of (Mes)(Me)PC(Ph)(2-Pyd)PdCl2 (3.5) ...................70 Figure 3.10 31P{1H} NMR spectrum (CDCl3) of 3.8 ....................................................75 ix Figure 3.11 Comparison of 1H NMR spectrum (CDCl3) of 3.8 against the 1H NMR spectrum (CDCl3) of 3.5 ............................................................................76 Figure 3.12 31P{1H} NMR spectrum (THF) of 3.9 .......................................................78 Figure 3.13 Molecular structure of ((Mes)(Me)PC(Ph)(o-DMA))2PdCl2 (3.9) ............80 x List of Abbreviations Anal. analysis API active pharmaceutical ingredient aq. aqueous Ar aryl br broad (spectra) Bu butyl n-Bu n-butyl t-Bu tert-butyl C Celsius Calcd calculated cod 1,5-cyclooctadiene COSY correlation spectroscopy d doublet (spectra) d day deg degree DMA dimethylaniline DMF dimethylformamide DMSO dimethylsulfoxide dppf 1,1’-Bis(diphenylphosphino)ferrocene E engegen (configuration) e.g. example xi EI electron impact equiv equivalent(s) FW formula weight GPC gel permeation chromatography HMBC heteronuclear multiple-bond correlation spectroscopy h/hr hour HSQC heteronuclear single-quantum correlation spectroscopy Hz hertz m multiplet (spectra) M+ molecular ion Me methyl Mes mesityl; 2,4,6-trimethylphenyl Mes* super mesityl; 2,4,6-tri-tert-butylphenyl MHz megahertz min minute mmol millimole MS mass spectrometry m/z mass-to-charge ratio NMR nuclear magnetic resonance o ortho O olefin p para PA phosphaalkene xii Ph phenyl PMP poly(methylenephosphine) PMP-r-PO poly(methylenephosphine)-random-polyolefin PO-b-PMP polyolefin-block-poly(methylenephosphine) ppm parts per million i-Pr iso-propyl 2-Py 2-pyridyl RT room temperature s singlet (spectra) t triplet (spectra) THF tetrahydrofuran TMEDA N,N,N’,N’-tetramethylethylenediamine TMSCl chlorotrimethylsilane TMT trimercaptotriazine VAZO 1,1’-azobis(cyclohexanecarbonitrile) Z zusammen (configuration) xiii Acknowledgements Professor Derek P. Gates For advice, patience, encouragement, and reference letters Dr. Julien Dugal-Tessier, Dr. Josh Bates, Dr. Eamonn Conrad, Dr. Ivo Krummenacher, Dr. Thom Hey, Paul Siu, Benjamin Rawe, Spencer Serin, Andrew Priegert and former and present undergrads For help in the lab and advice Staff in Department of Chemistry (UBC) For their assistance, and technical and administrative support NSERC, UBC, and Department of Chemistry (UBC) For financial support Angela Crane, Ashlee Howarth, Lyndsey Earl and Kim Osten For fun social shenanigans, theater goings, and girls nights My Parents For eternal belief in me and support Dawn and Jeff Kohrt For inspiration and support Brittany Juilfs For being an amazing little sister xiv Dedication To my family for their unconditional belief and unfailing support 1 Chapter 1 Introduction 1.1 Metal Scavenging The last few decades have seen a drastic increase in the usage of transition metals as catalysts in the synthesis of fine chemicals, pharmaceutical intermediates, and active pharmaceutical ingredients (APIs). The use of these catalysts has increased the efficiency of the synthesis of complex chemical compounds by reducing the synthetic steps and reducing waste.1 Recent advances have improved catalyst design and have encouraged the use of catalysis in the production of fine chemicals and active pharmaceutical ingredients.2 The most prominent transition metal in use for the synthesis of active pharmaceutical ingredients is palladium, however other late transition metals such as platinum3, ruthenium4, and rhodium5 are also important in the pharmaceutical industry. Unfortunately, the use of these late transition metals has an undesired effect of contaminating the product with transition metal-containing impurities after the catalytic reaction has been completed.6 This is especially important in active pharmaceutical ingredients, where a small amount of transition metal can prove toxic and stringent limits7 (e.g. < 10 ppm Pd in a 10g daily dose) are placed on metal contamination in drug products. Late transition metals are also of low abundance and their use in large-scale production required in the pharmaceutical industry can be extremely expensive. With these unfortunate side-effects that occur from the usage of late transition metals, it is necessary that a method of removing the metal catalysts from the chemical product is developed. 2 The development of inexpensive and efficient methods to remove transition metal catalysts from active pharmaceutical ingredients once the product has been synthesized is a current goal of the pharmaceutical industry.8 The general term for the removal of transition metal catalysts from chemical products after the products have been synthesized is known as metal-scavenging. Metal scavenging techniques fall into four general categories: distillation, extraction, crystallization, and adsorption.6 Each technique has advantages and disadvantages depending on the chemical product being purified and the specific reaction conditions utilized for each product. 1.1.1 Metal Scavenging Via Distillation Distillation is the process where two chemicals are separated from one another based on the differences in their volatility. Palladium and other heavy metals are not significantly volatile at “normal” laboratory distillation temperatures (<250ºC) and pressures. It is sometimes possible to distill the API or desired chemical product out of the reaction mixture, leaving the heavy metal behind in the residual waste. An example of this approach was used by Beutler9 et. al. in the preparation of a key intermediate of terbinafine hydrochloride (1.4; the active ingredient in Lamisil®). A Pd-coupling reaction was used to couple compounds 1.1 and 1.2 to synthesize product 1.3 (Scheme 1.1). Synthetic product 1.3 was purified by an aqueous extraction. The palladium catalyst remained in the crude product (177 ppm Pd) along with a small bit of copper (19 ppm Cu). Distillation of product 1.3 resulted in removal of 176 ppm of the palladium and nearly all the copper, leaving only 1 ppm of palladium and <1 ppm of copper in the 3 product. API 1.4 was synthesized from the product 1.3 resulting in an active ingredient with heavy metal traces well below the required specifications. N CH3 Cl H3C CH3 CH3 H N CH3 H3C CH3 CH3 1.1 1.2 1.3 1.4 Scheme 1.1 i) Pd(PPh3)2Cl2, Cu-I, n-C4H9NH2 ii) distillation i), ii) N CH3 H3C CH3 CH3 -HCl 4 1.1.2 Metal Scavenging Via Extraction Extraction is the process where two chemicals are separated from one another based on the differences in their solubility. For an extraction of palladium or another heavy metal, the palladium and API must have a significant difference in solubility. In other words, one should be mostly soluble in water and the other mostly soluble in the organic solvent. This results in extraction of the palladium from the solvent in which the active pharmaceutical ingredient is soluble. An example of this approach was used by Prashad and co-workers10 in the synthesis of compound 1.7. Compound 1.7 is an intermediate in the synthesis of fungicidal derivatives and its synthesis was achieved through α-arylation of pinacolone (1.5) with 1-bromo4-chlorobenzene (1.6) using sodium t-butoxide, palladium acetate, and toluene under ligand-free conditions (Scheme 1.2). In their first attempt to purify compound 1.7, residual palladium was removed using activated carbon (see section 1.1.4). Filtration of the activated carbon from the API resulted in the product containing 32 ppm Pd, which was not sufficient for the required specifications. After further study, it was found that the palladium could be removed by extraction of the toluene layer with an aqueous solution of L-cysteine at 85-90ºC, followed by washing with a solution of L-cysteine and sodium thiosulfate at 78-82ºC. This technique resulted in the product containing <3 ppm Pd, which was sufficient for the required specifications. 5 Another example of the use of extraction to remove palladium was done by O’Shea and co-workers11 in the preparation of compound 1.10. Compound 1.10 has potential use in bone regrowth. A Suzuki coupling of 1.8 and 1.9 resulted in the synthesis of compound 1.10 (Scheme 1.3). They found the crude product contained ~8000 ppm of palladium and iron. The initial techniques attempted to remove the palladium (chromatography and the use of n-Bu3P to solubilize the Pd) were not effective at removing the metals to the required specifications. With further research they discovered that compound 1.10 tends to form a water-soluble salt when reacted with lactic acid. To remove the palladium from the API, the API was dissolved in ethyl acetate and treated with n-Bu3P to complex and solubilize the Pd in the organic phase. The mixture was then extracted with aqueous lactic acid, resulting in the formation of the salt of 1.10 and transferring the salt to the aqueous phase. After the organic layer, which contained the Pd-phosphine complex, was removed a significantly purer API was obtained. The product 1.10 was regenerated from the salt by treatment of the aqueous layer with sodium carbonate and extraction into ethyl acetate. Further treatment of API Br Cl 1.5 O CH3 H3C H3C CH3 1.6 Cl O H3C H3C CH3 1.7 i) Scheme 1.2 i) Pd(OAc)2, t-BuONa, toluene 6 1.10 with activated carbon (see section 1.1.4) and crystallization from toluene resulted in an API with <50 ppm of Pd and Fe. From the two examples listed above, it is apparent that extraction of palladium from active pharmaceutical ingredients only works in specific cases where the product or metal has a change in its solubility when interacting with another compound in an aqueous phase (e.g. L-cysteine and lactic acid). The change in solubility is critical for the technique of extraction to be effective in removing palladium from the API. 1.8 1.10 Scheme 1.3 i) PdCl2(dppf)CH2Cl2, K2CO3, DMF/toluene, 2h, 80oC Br H N O N CH3 H3C B(OH)2 N HN 1.9 i) N HN H N O N CH3 H3C 7 1.1.3 Metal Scavenging Via Crystallization Crystallization is the process where two chemicals are separated from one another by solidifying the desired product out of solution while the palladium and other waste remain in solution. An example of this approach was used by Prasad and co-workers12 in the synthesis of compound 1.14. Compound 1.14 is an antimitotic agent used for skin disorders. A Sonogashira coupling of 1.11 and 1.12 resulted in the synthesis of compound 1.14 (Scheme 1.4). They found the crude product to contain ~400 ppm of residual palladium and initial techniques attempted to remove the palladium (activated carbon) resulted in the palladium levels reduced to 70 – 150 ppm, which was not sufficient for required specifications. Further attempts resulted in using an additive to complex with the palladium and solubilize the complex in the crystallization solution. The additive used was N-acetylcysteine and addition of this to the crystallization solution of compound 1.13 resulted in a reduction of palladium from 400 ppm to <30 ppm, which was not sufficient for required specifications. However, further addition of N- acetylcysteine for the crystallization of compound 1.14 resulted in reduction of the palladium to 1-2 ppm. 8 1.11 1.13 Scheme 1.4 i) Pd(0), Cu-I, Et3N, DMF ii) H2, Pd/i-PrOH 1.12 i) OCH3 OCH3 H N N Br CH3 OCH3 OCH3 N N H3C ii) OCH3 OCH3 N N H3C 1.14 9 Another example of the use of crystallization to remove palladium was done by Manley, Acemoglu et. al.13 in the preparation of compound 1.17. Compound 1.17 has potential use in the treatment of asthma. A Negishi coupling of 1.15 and 1.16 resulted in the synthesis of compound 1.17 (Scheme 1.5). A large study into crystalline salts was done to determine the most efficient method of removing palladium from the desired API. It was determined that a salt derived from maleic acid and product 1.17 was the most successful. Crystallization of the hemi-maleate salt reduced the palladium content from 100-800 ppm Pd of the crude product to 10-50 ppm. The salt was converted back to API 1.17 with aqueous sodium carbonate in methyl acetate. The organic phase containing the drug product was then treated with activated carbon (see section 1.1.4), filtered, and recrystallized to reduce the palladium content to <0.5 ppm Pd. 10 1.1.4 Metal Scavenging Via Adsorption Adsorption is the process where the transition metal is separated from the active pharmaceutical ingredient by chemically binding or adsorbing to another chemical compound or macromolecule then filtered from the API solution. This is the largest field of metal-scavenging and many examples of this type of purification exist.6 In the crystallization examples above, this technique was also being used, however the 1.15 Scheme 1.5 i) Pd(PPh3)4, THF, pentane, hexane, 0oC (15 min), 22oC (1h) 1.16 i) N N O NZnCl H3CO Br N H3CO N O N 1.17 11 compounds that bound to the palladium were single molecules and were not filtered from solution, but allowed the API to be crystallized while the palladium remained in the waste solution. Activated carbon is a common material for adsorption because of its low cost and ease of use.14 In the examples listed above in the previous sections, activated carbon was commonly used before or after the other methods of purification. Several chemical suppliers provide pre-made filtration systems with activated carbon. These pre-made cartridges remove the need to add carbon powder to a solution, which often results in carbon contamination of the product. 1.1.4.1 Trimercaptotriazine An example of the use of adsorption to remove palladium was done by Bristol- Myers Squibb15 in the preparation of compound 1.21. The synthesis of 1.21 was achieved by a palladium-catalyzed indolization of compound 1.18 and 1.19 (Scheme 1.6). The most effective method found to remove palladium was treatment of intermediate 1.20 with a slurry of trimercaptotriazine (1.22; TMT), charcoal, and diatomaceous earth. The slurry was cooled to 0-5ºC and filtered. The treatment resulted in a reduction of palladium content to 1-4 ppm Pd. This technique was effectively scaled to 12 kg with palladium content reduced to the same extent as in the laboratory process. 12 1.18 Scheme 1.6 i) Pd(OAc)2, PPh3, LiCl, Na2CO3, CH3CN, H2O, reflux ii) 1.22, Darco G-60, HyFlo filter aid, 0-5oC (1h) 1.19 i), ii) I 1.20 SO2NHCH3 NH2 N N OCH3 N N SiEt3 N H N N N N SO2NHCH3 H3CO SiEt3 N H N N N N SO2NHCH3 H3CO H 1.21 1.22 N N N SH SHHS 13 1.1.4.2 Polystyrene-bound TMT Trimercaptotriazine is an effective treatment for removal of palladium from active pharmaceutical ingredients. However, if TMT is bound to polystyrene by reacting Merrifield resin (1.23) with TMT (Scheme 1.7), the metal-scavenging capabilities increase. The low-solubility of the polystyrene portion of the metal-scavenging agent increases the ease of filtration after the palladium has been complexed. Ishihara and co- workers16 performed a study on the effectiveness of polystyrene-bound TMT in purifying a number of solutions containing Pd(II)Cl2. After the solution was stirred in the presence of 1.24 (3.8/11.5 equivs. of resin per Pd(II)) at room temperature for 1 day, the slurries were filtered and the palladium content of the filtrate was determined. Table 1.1 shows the results of their tests. The resin was effective at reducing the palladium content at low pH, with the effectiveness reduced as the pH increased. The resin was also effective at removing Pd(OAc)2 from an organic solution as well as in an aqueous solution. 14 Table 1.1 – Removal of Palladium Using Polystyrene-Bound TMT Entry Solvent Solution pH Residual Pd (ppm) Adsorption (%) 1a 1 M HCl (aq.) 0.1 0.061b 99.9 2a 1 M CH3CO2H (aq.) 2.2 0.110b 99.8 3a 5 wt% NH4Cl (aq.) 3.8 19.0b 68.3 4a 5 wt% NH4OAc (aq.) 6.9 33.0b 45.0 5c 5 wt% NH4Cl (aq.) 3.8 0.180b 99.7 6c 5 wt% NH4OAc (aq.) 6.9 2.0b 96.7 7a 1 M CH3CO2H in THF --- 0.045d 99.91 8a THF --- 0.002e 99.99 a 3.8 equivs. of 1.24/Pd(II) was used. b Each solution contained ~60.0 ppm Pd originally. c 11.5 equivs. of 1.24/Pd(II) was used. d The solution originally contained 47.4 ppm Pd(II). e The solution originally contained 23.7 ppm Pd(II). 1.22 Scheme 1.7 i) Et3N, KI, MeOH, reflux, 1d ii) NaH, DMF, reflux, 1d 1.23 i) or ii) N N N HS 1.24 SH SH Polystyrene Cl Polystyrene S N N N RS SR R = H or Polystyrene CH2 15 1.1.4.3 Polystyrene-bound Ethylenediamine As seen above, polystyrene-bound TMT is an effective metal scavenger for palladium due to the chelating ability of trimercaptotriazine combined with the ease of filtering the polystyrene from the solutions of active pharmaceutical ingredients. Due to the success of this compound, others have bound chelating molecules to polystyrene in attempts to mimic the success of polystyrene-bound TMT. One such attempt was done with polystyrene-bound ethylenediamine.17 Polystyrene-bound ethylenediamine was used in the synthesis of API 1.31. The synthesis of intermediate 1.30 was achieved via the Suzuki-Miyaura coupling reaction. It 1.25 1.26 Polystyrene 1.27 Polystyrene NH N NH2 H2N HN NH H2N Polystyrene HN NH2 Figure 1.1 – Several polystyrene-bound amines useful for metal scavenging. 16 was initially found that addition of ethylenediamine to the reaction solution effectively complexed to the palladium, however the resulting oil was difficult to separate from the product. To improve separation, polystyrene-bound ethylenediamine (Figure 1.1) was used to scavenge the palladium from the API. After the Suzuki-Miyaura coupling reaction, the solution was washed with water and polystyrene-bound ethylenediamine was added to the organic layer and the resulting slurry was stirred at 60ºC for 17 hours. Filtration of the polystyrene-bound ethylenediamine resulted in the palladium content being reduced from 2000-3000 ppm to 100-300 ppm. Further purification was done by treatment with (+)-di-p-toluoyl-D-tartaric acid and water. Crystallization, filtration, and washing resulted in the product salt, which contained 2-35 ppm Pd. The product was then carried onto the final API (1.31), which contained palladium levels sufficient for required specifications. 17 1.28 Scheme 1.8 i) PdCl2(PPh3)2, K3PO4-n H2O, toluene, reflux 1.29 i) 1.30 N Cl Br CH3 NH B N O O F 1.31 N Cl CH3 NH F N N Cl CH3 N F N 18 1.1.4.4 Silica-bound Metal Scavengers Silica provides an additional support for metal scavengers, like polystyrene has been mentioned as a support for TMT and ethylenediamine. Silica-based scavengers have a few advantages over polymer-based resins. Silica particle size can be controlled more easily than polymers and the resins do not swell in solution. Like the polystyrene- bound metal scavengers, silica can also provide support for a wide variety of chelating molecules like ethylenediamine and thiols (Figure 1.2). Allmendinger and co-workers18 conducted a study of a variety of silica-bound functionalized resins capable of metal scavenging. They chose to study the hydrogenation of an azide-precursor to Aliskiren base (Scheme 1.9). Very few purification techniques are effective for this reaction. Therefore they studied a wide variety of metal scavengers to determine which was most effective in purifying the Si O OO S S SH Si O OO S Si O O O N N N SH SH N HN HN NH 1.32 1.33 1.34 Figure 1.2 – Several silica-bound chelating atoms useful for metal scavenging. 19 product. Typically, the synthesis of 1.36 may be contaminated with up to 20 ppm Pd. 4 mL of the metal-contaminated solutions (10% 1.36 in ethanol) were treated with functionalized resins (100-300 mg), and the mixtures were shaken in screw cap vials for 20 h at room temperature. The samples were filtered and the residues analyzed for the metals. Table 1.2 shows a selection of silica-bound resins (for a more thorough analysis see reference 15). Scheme 1.9 1.35 R2 R1 OH N3 H2, Pd/C EtOH O OO NH2 OH O H N O NH2 1.36 20 Table 1.2 – Removal of Palladium from 1.36 using Silica-Bound Metal Scavengers Pd level (ppm) of 1.36 after treatment with Resina Loading (mmol/g) 100 mg 300 mg Silica 1.00 14 14 1.32 0.04 9 11 1.33 0.03 11 13 1.34 0.26 12 9 a Refer to Figure 1.2 21 Table 1.3 - Selected Examples of Metal Scavenging Techniques Pd Source Scavenging Technique Pd (ppm) after treatment Reference Pd(PPh3)2Cl2 Distillation 1 9 Pd(OAc)2 Extraction with aqueous L-cysteine and toluene < 3 10 PdCl2(dppf)CH2Cl2 Extraction with n-Bu3P in organic phase and aqueous lactic acid (forms salt of 1.10) Followed by adsorption with activated carbon < 50 11 Pd(0) Crystallization with N-acetylcysteine to solubilize palladium 1 - 2 12 Pd(PPh3)4 Crystallization with maleic acid to form salt of organic product Followed by adsorption with activated carbon < 0.5 13 Pd(OAc)2 Adsorption with TMT (1.22) 1 - 4 15 Pd(PPh3)2Cl2 Adsorption with polystyrene-bound ethylenediamine (1.27) 2 - 35 17 HO O NH2 SH HO O OH H3C O SH O OH O OH O OH 22 1.2 Phosphorus-Carbon Polymers In the examples of metal-scavenging materials in section 1.1, several were supported by a macromolecule. Often polystyrene is used as a support for the metal- scavenging material because it is easily synthesized and can be easily modified after polymerization to contain the functional group capable of chelating transition metals. Several phosphorus polymers of this type have been previously made, where the phosphine moiety is incorporated into a side-chain of a polymer after the polymer has been synthesized.19 However, incorporating phosphorus into the main chain of the polymer is difficult and is relatively rare.20-23 Recently, a polymer containing a series of alternating phosphorus-carbon bonds in the main chain was synthesized using common polymerization techniques typically utilized for the polymerization of organic polymers.24 This was accomplished because of the similarities between the chemistry of phosphorus-carbon double bonds (P=C) and the chemistry of carbon-carbon double bonds (C=C).25-27 Treating a phosphaalkene (1.37) with a radical or anionic initiator results in poly(methylenephosphine), a polymer with a phosphine at every second atom of the polymer backbone (1.38) (Scheme 1.10). The simplicity of the synthesis allows the polymerization of a wide variety of phosphaalkenes resulting in tunable and functional macromolecules that have high potential for use as a ligand for metal-scavenging.28 23 It has also been recently discovered that phosphaalkenes can be co-polymerized with an olefin (e.g. styrene, isoprene) to result in random and block co-polymers.29-31 In some cases the polymerization is living, which results in a highly tunable and controllable polymer. Previous work by Tsang, Yam et al.32 of the Gates group has also shown that poly(methylenephosphine) and poly(methylenephosphine)-styrene copolymers are capable of supporting palladium catalysis. The study was performed on the Suzuki cross- coupling reaction with 1.38 and 1.41 in solution. For comparison, the reaction was also performed with triphenylphosphine. The polymers gave similar results to the phosphine system, which indicates that the polymers were involved in the catalytic reaction in a similar way to triphenylphosphine. From the study, it can be assumed that the polymer is known to chelating palladium in solution as a catalytic supporter. If the phosphorus- carbon polymer can support catalysis then it is safe to assume that it can also scavenge metals after palladium has been used in a catalytic reaction. Scheme 1.10 P C Mes Ph Ph P C Mes Ph Ph n Initiator 1.37 1.38 P C C C Initiator P C x C C y n 1.39 1.40 1.41 24 This polymer has high potential as a metal scavenger. The phosphorus-carbon polymer has several advantages. First, the phosphorus in the backbone of the polymer results in a polymer that is ready to scavenge metals without any further need to add phosphines on a side-chain of the polymer. The polystyrene examples of metal scavengers shown earlier need further work after polymerization to attach a metal- binding molecule to the polymer, this polymer already has a metal-binding atom in the backbone, overall reducing the work and energy needed. Second, the polymer can be widely varied by changing the phosphaalkene polymerized as well as by co- polymerization with styrene. Variability may result in a library of phosphorus polymers with metal-scavenging capabilities. However there is currently one disadvantage in that the examples of metal scavenging materials shown in the first section have at least two functional groups capable of chelating the transition metals in the metal-scavenging portion of the material (e.g. amines, thiols). So far, these polymers only have the phosphorus in the backbone of the polymer available for metal binding. The addition of a second or third binding group would increase the potential of synthesizing a material capable of removing metal impurities in active pharmaceutical ingredients below the stringent requirements for drug products. 1.3 Research Objectives It is clear that a great deal of research has been done in the field of metal scavengers. It is also clear that the pharmaceutical industry is in need of metal scavengers capable of removing transition metals below stringent requirements. The research shown above often utilizes two or more materials in sequence or in tandem to 25 reach those strict requirements. A large number of metal scavengers are built upon supports like polystyrene or silica with one or more functional group capable of binding metals. The polymeric support results in easier removal once the metal has bound to the scavenging material and the functional groups provide a place for the metal to bind. Previous work in our group has synthesized a polymer with a phosphorus-carbon backbone capable of binding palladium as a support for catalysis. The same phosphaalkene monomer can copolymerize with styrene to form both random and block copolymers. The usage of such a material as a metal scavenger to remove a transition metal after it has been used as a catalyst is highly possible. However, in the examples of metal scavengers given previously, a second or third functional group capable of binding transition metals would provide a stronger material for metal scavenging. The first objective of this thesis is to synthesize a new phosphaalkene with a second functional group substituted on the carbon atom of the phosphaalkene that is capable of binding palladium and other transition metals. This phosphaalkene must also be polymerizable, which may result in a new metal scavenging material. The synthesis and polymerization of this phosphaalkene is discussed in Chapter 2. In order to study the metal binding properties of the polymeric material, model compounds and complexes were synthesized and examined by NMR spectroscopy and X-ray crystallography. The synthesis and analysis of the model compounds and complexes are discussed in Chapter 3. The long-term goal behind the work in Chapters 2 and 3 is to develop a metal scavenging material based on poly(methylenephosphine). This would require studies, such as metal adsorption and metal binding studies, and further development of new 26 phosphaalkenes. Thus, the results included in Chapters 2 and 3 represent the ground work towards the long-term goal of poly(methylenephosphine) based metal scavengers. 1.4 Contributions by Other Researchers to This Work Some of the work in this thesis was achieved in collaboration with other researchers. All crystallographic data were obtained by Paul Siu and Spencer Serin with some assistance from Brian Patrick using the departmental X-ray diffractometer. The solution and refinement of the molecular structure of MesP=C(Ph)(o-DMA) (o-DMA = ortho-dimethylaniline) was completed by Paul Siu. The solution and refinement of the molecular structure of (Mes)(Me)P-C(Ph)(2-Py)PdCl2 (2-Py = 2-pyridine) was completed by Spencer Serin. The solution and refinement of the molecular structure of ((Mes)(Me)P-C(Ph)(o-DMA))2PdCl2 was done by Brian Patrick. All of the work presented in Chapters 2 and 3 was completed by myself. 27 Chapter 2 Synthesis and Polymerization Studies of New Phosphaalkenes Bearing Donor Substituents as Potential Chelating Ligands 2.1 Introduction As mentioned previously in the introduction, the overall goal of this project was to synthesize a new phosphaalkene bearing donor substituents with the potential to chelate transition metals for use as a metal scavenger. The polymerization of this new phosphaalkene would, in theory, result in a material that could act as a metal scavenger of transition metals for use in the pharmaceutical industry. 2.1.1 2-Pyridyl-Substituted Phosphaalkene Previously a bidentate phosphaalkene containing a 2-pyridine substituted in place of a phenyl on the carbon has been synthesized.33 This phosphaalkene (2.1) contained a nitrogen in the ortho position on the pyridine functional group, resulting in the nitrogen being a few bonds away from the phosphorus. This phosphaalkene was unique in that other reported phosphaalkenes bearing a pyridyl-substitution have employed the bulkier supermesityl substituent at phosphorus (e.g., Mes*P=CH(2-py),34 Mes*P=CH(2,6- py)HC=PMes*,34 and Mes*P=C(R)(2-Py),35,36 with R=H, SiMe3, and tBu; and Mes* = 2,4,6-tri-tert-butylphenyl) whereas phosphaalkene 2.1 had the less bulky mesityl substituent at phosphorus. This pyridine-substituted phosphaalkene was synthesized 28 using the phospha-Peterson reaction from the ketone 2-benzoylpyridine and lithiated MesP(SiMe3)2. Previous work has shown that the pyridine-substituted phosphaalkene 2.1 was capable of chelating both palladium and platinum using (cod)PdCl2 and (cod)PtCl2 as the palladium and platinum sources (Scheme 2.1).37 The compound was bound to the metal from both the phosphorus and the nitrogen atoms to form a bidentate system. The success of this phosphaalkene in binding transition metals led to the goal of synthesizing other phosphaalkenes with donor substituents for the potential to chelate transition metals, particularly Pd and Pt. P C Mes Ph N 2.1 29 Figure 2.1 shows examples of phosphaalkenes with donor substitutents (nitrogen, oxygen, and sulfur) with the potential to chelate transition metals. These phosphaalkenes may also be polymerizable and eventually used as a metal scavenging material. Only phosphaalkenes 2.1 and 2.3 have been synthesized and preliminary work has been done to synthesize phosphaalkene 2.5. 2.1 Scheme 2.1 2.2a-b M Cl N P Cl CMes Ph P C Mes Ph N (cod)MCl2 CH2Cl2 M = Pd, Pt 30 2.2 Results and Discussion As mentioned in the previous section, bidentate phosphaalkenes have been synthesized previously and shown that they are capable of chelating transition metals.33,37 This section discusses the synthesis and analysis of a new P(sp2)-N(sp3) phosphaalkene with the potential to chelate metals for metal sequestration applications. This phosphaalkene was synthesized utilizing the oft-mentioned phospha-Peterson reaction P C Mes Ph N P C Mes Ph N P C Mes Ph S P C Mes CMe2 Ph S P C Mes CMe2 Ph N P C Mes N N 2.1 2.3 2.4 2.5 2.6 2.7 Figure 2.1 – Targeted functionalized phosphaalkenes, tuned to serve as chelating ligands in anticipation of polymerization and usage as metal scavengers. 31 resulting in a phosphaalkene bearing a dimethylaniline substitution as well as modest bulk to promote stabilization.33 The general reaction used to synthesize phosphaalkenes is the phospha-Peterson reaction (Scheme 2.2). The phospha-Peterson reaction requires a ketone and MesP(SiMe3)2. Unfortunately, the ketone required for the synthesis of phosphaalkene 2.3 is unavailable via purchase from a chemical supplier, which means that both the ketone and MesP(SiMe3)2 must be synthesized prior to the synthesis of the phosphaalkene itself. Scheme 2.2 P C R R' R'' R P SiMe3 SiMe3 O R'R'' MeLi R P Li SiMe3 32 2.2.1 Synthesis of Phosphaalkene 2.3 The ketone 2.9 was synthesized following the route previously obtained by Horaguchi’s research group.38 To a solution of 2-aminobenzophenone in dimethyl sulfoxide was added methyl iodide (2.5 equivalent). The reaction mixture was heated at reflux. After 2 hours, methyl iodide was added again and the reaction mixture was heated at reflux for another few hours. This process was repeated until the amine signal in the 1H NMR spectrum indicated all of the primary amine was converted to the dimethylamine (2.9, δ = 2.7 ppm), in total an average of 6 equivalents of methyl iodide was needed to fully methylate the amine. The ketone was purified via several washes with hexanes and column chromatography and then dried in vacuo before the phospha- Peterson reaction was attempted. The synthesis of MesP(SiMe3)2 was completed following established literature procedures (Scheme 2.4).39 To a solution of magnesium in THF was added mesityl bromide slowly in increments. Once all of the mesityl bromide was added, the solution was heated at reflux for approximately one hour resulting in the Grignard reagent mesityl 2.92.8 O NH2 MeI DMSO O NMe2 Scheme 2.3 33 magnesium bromide (MesMgBr). In a separate flask a solution of phosphorus trichloride in THF was cooled to -78°C. MesMgBr was added drop wise into the stirred solution of phosphorus trichloride.. Once all of the mesityl magnesium bromide was added, the solution was slowly warmed to room temperature. The THF was then dried in vacuo and the product extracted in hexanes. The reaction mixture was monitored via 31P NMR spectroscopy to observe the three singlet resonances (167.2 ppm, 161.0 ppm, 152.6 ppm) resulting from the mono-substituted product (MesPCl2, MesPBr2, & MesPClBr). The reaction mixture was dried in vacuo and redissolved in a solution of diethyl ether. The mesityl dichlorophosphine solution was added to a solution of lithium aluminum hydride in diethyl ether at -78°C. The resulting solution was quenched with degassed deionized water and the ether layer extracted and dried over MgSO4. The reaction mixture was then dried in vacuo and distilled to isolate the mono-substituted phosphine (MesPH2). An analysis of the 31P NMR spectrum was consistent with the formulation of the mono- substituted phosphine (-155.0 ppm). A solution of the mesityl phosphine in THF was cooled to -78°C and methyllithium was added until the solution was a bright yellow. The solution was warmed to room temperature and stirred until the solution was a clear yellow. After 12 hours chlorotrimethylsilane was added to the solution and stirred for 15 minutes until the solution was colorless and cloudy. The reaction mixture was monitored by 31P NMR spectroscopy and the process of lithiation and addition of chlorotrimethylsilane repeated (approximately 2-3x) until the desired product, MesP(SiMe3)2 (δ = -161.9 ppm), was completely converted from MesPH2 (δ = -155.0 ppm). The product was purified via distillation and resulted in a clear oil (Overall Yield: 30%). 34 With both reagents synthesized, the phospha-Peterson reaction was attempted. The ortho-dimethylaniline-substituted phosphaalkene 2.3 was synthesized via the phospha-Peterson reaction for the P=C bond forming step (Scheme 2.5). A solution of ketone 2.9 in THF was added drop wise to a cooled (-78°C) solution of MesP(Li)SiMe3 in THF. The reaction occurred quickly and an aliquot was removed for analysis by 31P NMR spectroscopy. Most importantly, the signal assigned to MesP(Li)SiMe3 (δ = -187 ppm) was replaced by two new singlet resonances at 239 ppm and 230 ppm, which is in the range consistent with that expected for a phosphaalkene (MesP=CPh2: δ = 233 ppm). The presence of two signals in 31P NMR spectrum of the phosphaalkene solution (CDCl3) suggests that the compound 2.3 forms as mixtures of E and Z isomers. This is comparable to the previously synthesized phosphaalkene, 2.1, which also shows two signals in the same low-region of the 31P NMR spectrum at 260 ppm and 242 ppm.33 As observed in multiple reactions, the ratio of the E isomer to the Z isomer is one to one. X- ray structure experiments (discussed below) for single crystals of 2.3 reveal that the mixture crystallizes as the E isomer. Scheme 2.4 MesMgBr PCl3MesBr + Mg MesPCl2 LiAlH4 MesPH2 MeLi TMSCl Mes P SiMe3 SiMe3 MesPH2 2.12 2.10 2.11 2.11 35 By the analysis of several 2D NMR spectroscopy experiments (1H-1H COSY, 1H- 13C HSQC, 1H-13C HMBC, 1H-31P HMBC), several of the resonances can be identified into the two separate isomers (see Appendix A for 2D spectra). Unable to determine which signals correspond to the E or to the Z isomer, the signals for the isomers are simply labeled as isomer A and isomer B with isomer A having the lower field signal in the 31P NMR spectrum (δ = 239 ppm) and isomer B having the higher field signal in the 31P NMR spectrum (δ = 230 ppm). Scheme 2.5 2.12 P C Mes Ph Mes P SiMe3 SiMe3 MeLi Mes P Li SiMe3 N O NMe2 2.9 2.3 36 In the high field of the 1H NMR spectrum, the resonances for the amine protons can be resolved and identified as 2.41 ppm for isomer A and 2.55 ppm for isomer B. The resonances corresponding to the para and ortho methyl groups on the mesityl substituent on the phosphorus atom are indistinguishable between the two isomers and the signals fall at 2.28 ppm for the ortho-methyl groups and at 2.18 ppm for the para-methyl group. In the phenyl region of the 1H NMR spectrum, there are multiple signals between 7.72 ppm and 6.65 ppm that are difficult to resolve between the two isomers. For comparison, the 1H NMR spectroscopy for the related phosphaalkene 2.1 was completely resolved and identified separately as the E and Z isomers. For phosphaalkene 2.1, the signals for the phenyl region fall between 8.67 ppm for the protons adjacent to the nitrogen in the pyridyl substituent to 6.67 ppm for the protons in the meta position on the mesityl Figure 2.2 – 31P{1H} NMR spectrum of phosphaalkene 2.3 in CDCl3 37 substituent. For phosphaalkene 2.1, the only signals at high field correspond to the methyls on the mesityl group which fall at 2.28 ppm and 2.26 ppm for the ortho groups and 2.17 and 2.16 for the para groups, which match well with the assignments for phosphaalkene 2.3 (o-mesityl 2.28 ppm, p-mesityl 2.18 ppm). The similarities of the signals in the NMR spectroscopy between the pyridyl-substituted phosphaalkene 2.1 and the ortho-dimethylaniline-substituted phosphaalkene 2.3 indicate that the two phosphaalkenes are similar in structure. Figure 2.3 – 1H NMR spectrum of phosphaalkene 2.3 in CDCl3. Square, circle, and star icons indicate signals corresponding to specific protons shown on the chemical structure. 38 In the 13C NMR spectrum, the high field signals of the amine methyl groups can be resolved between the two isomers. The signal at 43.5 ppm corresponds to isomer A and the signal at 42.8 ppm corresponds to isomer B. The resonances for the mesityl methyls are indistinguishable between the two isomers and fall at 22.2 ppm and 21.9 ppm for the ortho-methyl groups and at 21.3 ppm for the para-methyl group. The ortho groups have a C-H J-coupling constant of 10 Hz and the para group has a J-coupling constant of 12 Hz. The resonances of the 13C NMR spectroscopy in the aryl region (δ = 151.4 – 117.6 ppm) are indistinguishable between the two isomers. In the lower region of the 13C NMR spectrum, the faint signals from the phosphorus-carbon double bond (P=C) are observed at 194.8 ppm and 190.6 ppm with a PC coupling constant of 123 Hz each. A summary of the resolved resonances for each isomer can be found in Table 2.1. Figure 2.4 – 13C NMR spectrum of phosphaalkene 2.3 in CDCl3 39 Table 2.1 – Summary of NMR Resonances for Phosphaalkene 2.3 Isomer A Isomer B 31P NMR 239 ppm 230 ppm 1H NMR amine (Ha) 2.41 ppm 2.55 ppm o-methyls (Hb) 2.28 ppm p-methyl (Hc) 2.18 ppm aryl (Hd) 7.72 – 6.65 ppm 13C NMR amine (Ca) 43.5 ppm 42.8 ppm o-methyls (Cb) 22.2 ppm & 21.9 ppm p-methyl (Cc) 21.3 ppm aryl (Cd) 151.4 – 117.6 ppm P=C (Ce) 194.8 ppm & 190.6 ppm *Refer to image of 2.3 below for labeling. 2.2.2 X-Ray Crystallographic Analysis of Phosphaalkene 2.3 To further investigate the structural features and bonding in phosphaalkene 2.3, the compound has been analyzed by X-ray crystallography. A summary of cell constants P C N 2.3 b b c d a a e 40 and data collection parameters for 2.3 is included in Table 2.2. The molecular structure of phosphaalkene 2.3 is shown in Figure 2.5. Important metrical parameters for 2.3 are tabulated in Table 2.3 and, for comparison, the metrical parameters are also provided for the closely related phosphaalkene 2.1.33 Table 2.2 - X-Ray Crystallographic Data for 2.3 2.3 formula C48H52N2P2 fw 718.86 cryst syst triclinic space group P1̄ color yellow a (Å) 8.7100(14) b (Å) 14.886(2) c (Å) 16.860(3) α (deg) 109.285(3) β (deg) 98.199(4) γ (deg) 99.850(3) V (Å3) 1985.8(6) T (K) 183(2) Z 2 µ (Mo Kα) (cm-1) 14.5 cryst size (mm3) 0.05 x 0.10 x 0.25 calcd density (Mg m-3) 1.202 2θ (max) (deg) 27.86 No. of reflns 34 062 No. of unique data 9394 R (int) 0.0240 Refln/param ratio 19.6 R1 0.035 wR2 (all data) 0.0904 The ortho-dimethylaniline-substituted E-2.3 has an average P=C bond length (1.6933(12) Å) compared to the typical range found for C-substituted phosphaalkenes 41 (1.61 – 1.71 Å)40, albeit the bond length for E-2.3 falls at the long end of the range. For comparison, the P=C bond length for phosphaalkene 2.1 (1.7082(13)) also falls at the C8 C4 C5 C10 C14 C18 C22 C3 C9 C13C17 C21 C2 C7 C1 C6 C11 C12 C15 C16 C19C20 P1 N1 C23 C24 Figure 2.5 – Molecular structure of MesP=C(Ph)(o-DMA) (2.3). Ellipsoids are drawn at the 50% probability level; Atoms are labeled and hydrogens have been omitted for clarity. Selected bond lengths (Å) and angles (deg): P(1)-C(10) 1.6933(12), P(1)-C(1) 1.8376(12), C(10)-C(11) 1.4948(16), C(10)-C(19) 1.4824(16); C(1)-P(1)-C(10) 106.66(6), P(1)-C(10)-C(11) 116.87(11), P1-C(10)-C(19) 128.24(9), C(11)-C(10)-C(19) 115.13(10). 42 long end of the typical phosphaalkene range. Interestingly, the P-CMes bond length of phosphaalkene 2.3 (1.8376(12) Å) is short compared with the typical P-C single bond range (1.85 – 1.90 Å)41, but is comparable to the bond length of the previously obtained E-2.1 (1.8378(16) Å). The slight elongation of the P-C bond and shortening of the P-CMes bond may indicate some π-conjugation between the Mes group and the P-C bonds. The bond angle at phosphorus, ∠C-P=C, in phosphaalkene 2.3 (106.66(6)°) is smaller than the average 108° for a phosphaalkene and is smaller than the angle of phosphaalkene 2.1 (107.80(7)°) by a full degree. The 108° bond angle at phosphorus is consistent with a high degree of s-character in the lone pair on phosphorus, with the σ- bonds being higher in p-character. The geometry at C10 is essentially planar in phosphaalkene 2.3 (sum of angles 360(1)°), which is typical of a phosphaalkene and is comparable to the geometry of phosphaalkene 2.1, which also has a planar geometry around the double-bonded carbon. It is interesting to note that in Arcis (∠P=C-Ccis ≈ 128°) bends further away from the P-C bond than Artrans (∠P=C-Ctrans ≈ 116°). This fact reflects the greater steric congestion between the Mes and Ph groups. This congestion is also observed for the bond angles of phosphaalkene 2.1 with the bond angle for Arcis (∠P=C-Ccis ≈ 125°) and the bond angle for Artrans (∠P=C-Ctrans ≈ 116°). 43 Table 2.3 – Important Metrical Parameters for 2.1 and 2.3 E – 2.1a E – 2.3b Bond Lengths (Å) P = C 1.7082(13) 1.6933(12) 1.6953(13) P – CMes 1.8378(16) 1.8376(12) 1.8345(12) C – Ctrans 1.496(2) 1.4948(16) 1.4906(16) C – Ccis 1.498(2) 1.4824(16) 1.4828(17) Bond Angles (deg) ∠ CMes – P = C 107.80(7) 106.66(6) 105.79(6) ∠P = C – Ctrans 116.87(11) 116.59(8) 116.54(9) ∠P = C – Ccis 125.41(11) 128.24(9) 126.78(9) ∠ Ccis – C – Ctrans 117.48(13) 115.13(10) 116.63(10) a Reference 33. b Two independent molecules are present in the asymmetric unit. Data on the top line are for molecule 1; the bottom line, for molecule 2. 2.2.3 Synthesis of Phosphaalkene 2.5 In order to broaden the donor-substituted phosphaalkene library, attempts were made to synthesize other new phosphaalkenes. One in particular that consumed a great deal of time with limited success was the phosphaalkene 2.5. This phosphaalkene is similar to the ortho-dimethylaniline-substituted phosphaalkene 2.3, however this phosphaalkene has a double substitution of ortho-dimethylaniline at the carbon in comparison to the single substitution of phosphaalkene 2.3. 44 Like the synthesis of phosphaalkene 2.3, this phosphaalkene requires the relevant ketone and MesP(SiMe3)2 for the phospha-Peterson reaction. Also like phosphaalkene 2.3, the ketone for this compound is unavailable via a chemical supplier and must be synthesized prior to it being utilized in the phospha-Peterson reaction. The general idea for the synthesis of this ketone (2.3d) was to ortho-lithiate dimethylaniline using directed ortho metalation42 and react the lithiated dimethylaniline43 with a carbonyl source to produce bi(dimethylaniline) benzophenone (Scheme 2.6). P C N 2.5 N Mes Scheme 2.6 O NN 2.15 2.14a, 2.14b N Li N n-BuLi, TMEDA reflux 2.14a = 2.14b = Cl OEt O N N O N N 2.13 45 The first step of the reaction, the lithiation of dimethylaniline, was fairly simple. A solution of dimethylaniline in hexanes was reacted with n-butyllithium in the presence of TMEDA. The solution was heated at reflux for several hours and afterwards the reaction mixture was cooled to room temperature. A solution of the carbonyl source in hexanes was added to the lithiated dimethylaniline. The reaction solution was stirred for several hours then quenched with degassed deionized water. The resulting lithium salt was removed via filtration and the product isolated by extraction of the organic phase from the aqueous phase. The first carbonyl source to be attempted was ethyl chloroformate (2.14a). In the analysis of the 1H NMR spectroscopy and mass spectroscopy, it was determined that the major product obtained was not, in fact, the desired bi(dimethylaniline) benzophenone. Instead the addition of the lithiated dimethylaniline resulted in triple addition resulting in the alcohol (2.16). The 1H NMR spectroscopy indicated a clean singlet at 3.03 ppm and the mass spectroscopy showed the m/z at 389. OH N N 2.16 N 46 It was possible that the choice of carbonyl source resulted in the first addition of the lithiated dimethylaniline to be more reactive than the original source, which resulted in the production of the triply substituted product. In order to bypass this problem the reaction was attempted with another carbonyl source, carbonyldiimidazole (2.14b). Unfortunately analysis with 1H NMR spectroscopy indicated that this reaction also resulted in the triply substituted product. It might be possible to obtain the desired ketone by oxidation of the alcohol, but unfortunately the attempt was never made. A literature search into other ways of synthesizing bi(dimethylaniline) benzophenone resulted in the discovery of a possible route in which a Grignard reagent (o-dimethylaminophenylmagnesium bromide) was reacted with carbon dioxide as the carbonyl source.44 Unfortunately the experiment was never attempted due to lack of time and successes in other areas. 2.2.4 Polymerization of Phosphaalkene 2.3 With the success of synthesizing a new phosphaalkene (2.3) with the potential to chelate metals, the polymerization of the phosphaalkene would indicate that it also has the potential to be made into a material capable of extracting metals from active pharmaceutical ingredients. As mentioned in the introductory chapter, inorganic polymers with an alternating phosphorus-carbon backbone are now easily synthesized using known radical and anionic polymerization techniques used for common organic polymers.24,32 The polymers can be synthesized by radical polymerization using the common radical initiator, VAZO and the phosphaalkene monomers can also be 47 polymerized by anionic polymerization using MeLi as an initiator, resulting in poly(methylenephosphine) polymers (Scheme 2.7). Polymerization of phosphaalkene 2.3 followed the same general procedure used previously to polymerize phosphaalkenes (Scheme 2.7). In a pyrex vacuum tube phosphaalkene 2.3 was added along with the initiator, VAZO. The tube was flame sealed in vacuo and heated in an oven with a rocking tray at 200°C for 48 hours. After the allotted time, the mixture became highly viscous and the tube was cooled to room temperature before being opened under nitrogen. The polymer was purified by dissolving in dichloromethane and precipitated with hexanes. The resulting polymer (2.17) was yellow-brown in appearance. P C P C n PA PMP Init = R. or R- [I] Scheme 2.7 48 A broad signal in the phosphine region of the 31P NMR spectrum indicates the formation of the phosphorus-carbon polymer (δ = 18.9 ppm). The resonances for both the phenyl region (δ = 7.76 – 6.7 ppm) and the alkane region (δ = 2.6 – 2.2 ppm) in the 1H NMR spectrum also appear broad, indicating the formation of a large molecule. For comparison, the polymer 2.18 shows a broad signal (δ = -10 ppm) in the phosphine region of the 31P NMR spectrum24 and the 1H NMR spectrum shows a broad signal in the phenyl region (δ = 7.8 – 6.6 ppm) as well as in the alkane region (δ = 2.5 – 1.9 ppm). The large difference in the NMR signals (-10 ppm vs. 18 ppm) could be a result of some oxidation of the polymer from exposure to the atmosphere. P C Mes Ph N P C Mes Ph n N VAZO 2.3 2.17 200oC 48 hr Scheme 2.8 P C Mes Ph n Ph 2.18 49 Figure 2.6 – 31P{1H} NMR spectrum of phosphorus – carbon polymer 2.17 in CDCl3 50 The polymer was also examined via GPC (THF) analysis. Unfortunately, the polymer does not readily dissolve in THF, which resulted in unreliable measurement of the molecular weight calculations. However, the light scattering data of the GPC shows clearly and cleanly the presence of polymer (Figure 2.8). Figure 2.7 – 1H NMR spectrum of phosphorus – carbon polymer 2.17 in CDCl3 51 2.3 Summary This chapter has examined the synthesis of a new phosphaalkene (2.3) bearing a donor substituent as a potential chelating ligand and the new inorganic phosphorus- carbon polymer (2.17) derived from the donor-substituted phosphaalkene. The new ortho-dimethylaniline-substituted phosphaalkene is similar to a previously made pyridine-substituted phosphaalkene that has already shown the ability to chelate metals. The phosphaalkene contains a dimethylaniline group on the carbon substituent resulting in a phosphaalkene with two donating atoms capable of binding to a metal (P, N). The synthesis of this phosphaalkene may pave the way for the synthesis of other phosphaalkenes with two or even three different functional groups capable of binding Figure 2.8 – GPC (THF) trace of phosphorus-carbon polymer 2.17 52 metals. The synthesis of other donor-substituted phosphaalkenes may result in a broad library of bidentate and possibly tridentate phosphaalkenes with the potential as chelating ligands. The polymerization of the new phosphaalkene resulted in a new inorganic phosphorus-carbon polymer and shows the potential of the phosphaalkene to be engineered into a material capable of removing metals from active pharmaceutical ingredients. Although untested in actual solutions, this polymer has high potential to remove transition metals from solutions. The synthesis of this new polymer may show the potential of phosphorus-carbon polymers in metal-sequestration resulting in the development of a range of new inorganic polymers with the purpose of removing metal impurities from active pharmaceutical ingredients. 2.4 Experimental General Procedures. All manipulations of air- and/or water-sensitive compounds were performed under a nitrogen atmosphere using standard Schlenk or glovebox techniques. Hexanes and dichloromethane were deoxygenated with nitrogen and dried by passing through a column containing activated alumina. THF was freshly distilled from sodium/benzophenone ketyl. Deionized water was degassed prior to use. CDCl3 was distilled from P2O5 and degassed. VAZO was recrystallized in methanol and dried in vacuo. 2-aminobenzophenone, methyl iodide, and methyllithium were purchased from 53 Aldrich and used as received. MesP(SiMe3)210 was prepared following literature procedures. Equipment. 1H, 31P, and 13C NMR spectra were recorded at room temperature on Bruker Avance 300 or 400 MHz spectrometers. Chemical shifts are reported relative to residual CHCl3 (δ = 7.24 for 1H), 85% H3PO4 as an external standard (δ = 0.0 for 31P), and CDCl3 (δ = 77.0 for 13C). Assignments of NMR spectra were made with the aid of 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, 1H-31P HMBC experiments. Elemental analyses were performed in the University of British Columbia Chemistry Microanalysis Facility. Mass spectra were recorded on a Kratos MS 50 instrument in EI mode (70 eV). X-ray crystallography was performed on a Bruker X8 APEX diffractometer with graphite-monochromated Mo Kα radiation. Gel permeation chromatography (GPC) was performed using a Waters liquid chromatograph equipped with a Waters 515 HPLC pump, Waters 717 plus autosampler, Waters Styragel, columns (4.6x300mm) HR2, HR4, and HR5E and a Waters 2410 differential refractometer (refractive index detector). A flow rate of 0.3 ml/min was used and samples were dissolved in THF (ca. 1mg/ml) and filtered before injection. 2.4.1 Preparation of MesP=C(Ph)(o-DMA) (2.3) To a stirred solution at 55°C of MesP(SiMe3)2 (8.3g, 28.1 mmol) dissolved in THF (40 mL) was added MeLi (1.6M, 17.5 mL, 28.1 mmol). The reaction was monitored by 31P NMR (δ = -185) and after 2 hours the reaction was cooled to -78°C. 54 The reaction mixture was added to a solution of 2-dimethylaminobenzophenone (6.34 g, 28.1 mmol) in THF (30 mL). After approximately 15 minutes the reaction mixture was quenched with excess chlorotrimethylsilane. The solvent was removed in vacuo leaving a dark yellow oil. To the oil was added minimal hexanes and cooled to give yellow crystals. Yield = 6.38g (87%). Isomer A - 31P NMR (CDCl3): δ 239.03 (s). 1H NMR (CDCl3): δ 7.72-6.65 (m, 11H, Ar), 2.41 (s, 6H, NCH3), 2.28 (s, 6H, o-CH3), 2.18 (s, 3H, p-CH3). 13C (CDCl3): δ151.4 – 117.6 (m, Ar), 43.48 (s, N(CH3)2), 22.18 (d, J = 10 Hz, o-CH3), 21.90 (d, J = 10 Hz, o- CH3), 21.28 (d, J = 12 Hz, p-CH3). Isomer B - 31P NMR (CDCl3): δ 230.64 (s). 1H NMR (CDCl3): δ 7.72-6.65 (m, 11H, Ar), 2.55 (s, 6H, NCH3), 2.28 (s, 6H, o-CH3), 2.18 (s, 3H, p-CH3). 13C (CDCl3): δ151.4 – 117.6 (m, Ar), 42.80 (s, N(CH3)2), 22.18 (d, J = 10 Hz, o-CH3), 21.90 (d, J = 10 Hz, o- CH3), 21.28 (d, J = 12 Hz, p-CH3). Unresolved Signals: 13C (CDCl3) (E/Z mixture): δ 194.8 (d, JPC=123.6 Hz, P=C), 190.64 (d, JPC=123.3 Hz, P=C). MS (EI): 359 [M+], 344 [M+ - Me], 208 [M+ - MesP + H]. Anal. Calcd for C24H26PN: C, 80.19; H, 7.29; N, 3.90. Found: C, 76.05; H, 7.05; N, 3.77. 2.4.2 Preparation of [MesP-C(o-DMA)(Ph)]n (2.17) To a pyrex tube was added phosphaalkene 2.3 (2.0 g, 5.57 mmol) and VAZO (0.05g, 0.204 mmol). The tube was flame sealed in vacuo. After being heated at 200°C in an oven equipped with a rocking tray for 48 hours the polymerization mixture became very viscous. The tube was broken under nitrogen and dissolved in minimal CH2Cl2 (3 55 mL) and precipitated into hexanes (90 mL). Precipitation was repeated 3 times. A brownish-yellow powder was obtained and dried. 31P NMR (CDCl3): δ 18.9 (br). 1H NMR (CDCl3): δ 7.76 – 6.7 (aryl-H, br), δ 4.70 (br), δ 2.6 – 2.3 (CH3 & NCH3, br). 56 Chapter 3 Synthesis and Metal Binding Studies of Phosphaalkene Model Compounds and Complexes Bearing Donor Substituents as Potential Chelating Ligands 3.1 Introduction In the previous chapter (Chapter 2) the synthesis of a new phosphaalkene and its derivative polymer were discussed. The purpose of synthesizing the new phosphaalkene and the polymer was in the goal of developing a material that is capable of removing transition metals from active pharmaceutical ingredients. In order to show that the polymers are capable of binding metals, model compounds must be synthesized and bound to transition metals, primarily Pd, to show exactly how the monomer binds metals. Because the binding of the macromolecular polymer is difficult to analyze, it is presumed that the individual monomers within the polymer bind similarly to the model compound. This chapter discusses the synthesis of the model compounds of the nitrogen containing phosphaalkenes mentioned in the previous chapter: 2-pyridine-substituted phosphaalkene (3.1) and ortho-dimethylaniline-substituted phosphaalkene (3.2). This chapter also discusses how the model compounds bind with palladium (II) to form model complexes. These model complexes indicate how the polymers will bind to transition 57 metals once they are used to remove heavy metals from active pharmaceutical ingredients. 3.2 Results and Discussion This section discusses the model compounds and model complexes synthesized from the 2-pyridine substituted phosphaalkene 3.1 and the ortho-dimethylaniline- substituted phosphaalkene 3.2. Previous work with phosphaalkene 3.1 has shown that the phosphaalkene binds both palladium (II) and platinum (II) through both the phosphorus and the nitrogen chelating atoms.37 Presumably, the phosphine model compound of the phosphaalkene will bind in the same manner. The following sections discuss the synthesis of the two model compounds and their binding capabilities. 3.2.1 Synthesis of Model Compound 3.3 The synthesis of the 2-pyridyl substituted model compound was a fairly simple reaction (Scheme 3.1). To a solution of MesP=C(Ph)(2-Py) (3.1) in THF was added methyllithium. The reaction was monitored by 31P NMR spectroscopy (δ = -47 ppm). After approximately twenty minutes, the reaction solution was quenched with methanol P C Mes Ph N 3.1 P C Mes Ph N 3.2 58 to form the model compound (3.3) (δ = -22 ppm). The THF was removed in vacuo resulting in a pale yellow solid. The solid was purified by washing with hexanes and filtered to remove the lithium salt byproduct. The hexanes was removed in vacuo to produce the pure, pale yellow product. Analysis of the 1H, 31P, and 13C NMR spectra is consistent with the proposed formulation of the model compound. The most obvious change is in the 31P NMR spectrum (Figure 3.1) where the resonance of the model compound resides at -22 ppm. The parent compound, phosphaalkene 3.1, displays two signals at 260.1 ppm and 242.1 ppm, which is the typical range for a phosphaalkene (MesP=CPh2: δ = 233 ppm). The two signals represent the E and Z isomers of the parent compound, the reduction of those two signals into one is consistent with the removal of the double bond character in the parent compound to the single bond in the model compound. In comparison to similar phosphines with a nitrogen atom within 4-5 bonds away (δ = -11.2345, -3.9745, -24.5846), the phosphorus signal is typical. P C Mes Ph N P CH Mes Ph Me N MeLi MeOH 3.1 3.3 Scheme 3.1 59 The 1H NMR spectrum (Figure 3.2) displays a resonance at 1.35 ppm, which corresponds to the newly added methyl group on the phosphorus atom. It also shows the formation of the resonance at 4.99 ppm, which is assigned to the newly added proton on the carbon atom. The other resonances at 2.43 ppm and 2.20 ppm match the resonances for the methyls on the mesityl that can also be seen in the 1H NMR spectroscopy of the parent compound, 2-pyridyl substituted phosphaalkene 3.1.33 Figure 3.1 – 31P{1H} NMR spectrum of model compound 3.3 in CDCl3 60 In the 31C NMR spectrum (Figure 3.3) the appearance of the resonance at 10.3 ppm (JPC = 18 Hz) is assigned to the new methyl carbon bonded to the phosphorus atom and the resonance at 53.7 ppm shows the change in the phosphorus-carbon double bond to a single bond. A typical phosphorus-carbon double bond resonates in the low field of the spectrum, around 200 – 190 ppm. The shift from the low field resonance of the parent compound 3.1 (190 ppm) to the high field resonance of the model compound 3.3 Figure 3.2 – 1H NMR spectrum of model compound 3.3 in CDCl3. Diamond, circle, star and square icons indicate signals corresponding to specific protons shown on the chemical structure. 61 (53.7 ppm) indicates a distinct change in the compound’s bonding character from a phosphorus-carbon double bond to a single bond. Figure 3.3 – 13C NMR spectrum of model compound 3.3 in CDCl3 62 3.2.2 Synthesis of Model Compound 3.4 The synthesis of the ortho-dimethylaniline substituted model compound is a fairly simple reaction (Scheme 3.2). To a solution of MesP=C(Ph)(o-DMA) (3.1) in THF was added methyllithium. The reaction mixture was monitored by 31P NMR spectroscopy to observe the lithiation of the phosphaalkene (δ = -47 ppm). After approximately twenty minutes, the reaction solution was quenched with methanol to form the model compound (3.4) (δ = -20.9 ppm). The THF was removed in vacuo, resulting in a pale yellow solid. The solid was washed with hexanes and filtered to remove the lithium salt byproduct. The hexanes was removed in vacuo to produce the pale yellow product. Analysis of the 1H, 31P, and 13C NMR spectra is consistent with the proposed formulation of the model compound. The most obvious change is in the 31P NMR spectroscopy (Figure 3.4) where the resonance of the model compound resides at -20.9 ppm. The parent compound, phosphaalkene 3.2, displays two signals at 239 ppm and 230 ppm, which is the typical range for a phosphaalkene (MesP=CPh2: δ = 233 ppm). The two signals are assigned to the E and Z isomers of the parent compound, the reduction of Scheme 3.2 MeLi MeOH 3.2 3.4 P C Mes Ph N P CH Mes Ph Me N 63 those two signals into one is consistent with the removal of the double bond character in the parent compound to the single bond in the model compound. In comparison to similar phosphines with a nitrogen atom within 4-5 bonds away (δ = -11.2345, -3.9745, - 24.5846) the phosphorus signal is typical. The 1H NMR spectrum (Figure 3.5) displays a resonance at 1.33 ppm, which corresponds to the newly added methyl group on the phosphorus atom. It also displays a formation of the resonance at 5.82 ppm, which is assigned to the newly added proton on the carbon atom. The other resonances at 2.37 ppm and 2.17 ppm match the resonances for the methyls on the mesityl that can also be seen in the 1H NMR spectroscopy of the Figure 3.4 – 31P{1H} NMR spectrum of model compound 3.4 in CDCl3 64 parent compound, ortho-dimethylaniline substituted phosphaalkene 3.2. The resonance at 2.52 ppm is assigned to the amine methyls and is consistent with the spectra for the parent compound. In the 31C NMR spectrum (Figure 3.6), the appearance of the resonance at 10.3 ppm (JPC = 18 Hz) is assigned to the new methyl carbon bonded to the phosphorus atom and the resonance at 45.9 ppm shows the change in the phosphorus-carbon bond from a double bond to a single bond. A typical phosphorus-carbon double bond resonates in the Figure 3.5 – 1H NMR spectrum of model compound 3.4 in CDCl3. Diamond, circle, star, square and triangle icons indicate signals corresponding to specific protons shown on the chemical structure. 65 low field of the spectrum, around 200 – 190 ppm. The shift from the low field resonance of the parent compound 3.2 (194.8 ppm) to the high field resonance of the model compound 3.4 (45.9 ppm) indicates a distinct change in the compound’s bonding character from a phosphorus-carbon double bond to a single bond. 3.2.3 Synthesis of Model Complex 3.5 In order to observe the binding characteristics of the phosphine model compound in the presence of transition metals, the model compound was reacted with palladium (II). Addition of palladium (II) resulted in the formation of a model complex that represents Figure 3.6 – 13C NMR spectrum of model compound 3.4 in CDCl3 66 the related polymer. By observing the binding characteristics of the model complex, assumptions about the binding characteristics of the polymeric material can be made. To a solution of 2-pyridine-substituted model compound 3.3 was added one equivalent of (cod)PdCl2. The mixture was stirred and the reaction mixture was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.5 was formed (δ = 45.5). The solution was dried in vacuo and residual 1,5-cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was dried in vacuo and recrystallized via slow addition of hexanes to a solution of the model complex in dichloromethane. The large shift of the 31P NMR spectrum of approximately 60 ppm [-22 ppm (free phosphine) to 45 ppm (bound phosphine)] is consistent with the change indicative of a phosphorus ligand bound to a metal.46 The large shift verifies the coordination of the phosphorus atom to the palladium source. A small amount of secondary product (δ = 41.2) can be observed in the spectrum. The secondary product is most probably the same model complex with a difference in chirality. Since the model compound is synthesized from both the E and Z isomers of the phosphaalkene, the chirality around both the 3.3 3.5 P CH Mes Ph Me N Pd Cl Cl(cod)PdCl2 P N Me CH Ph Mes Scheme 3.3 67 phosphorus and the carbon atoms may differ slightly, resulting in a small difference in the chemical shifts of the two compounds. Figure 3.8 shows a comparison of the 1H NMR spectrum of the model complex 3.5 with the 1H NMR spectrum of the model compound 3.3. In the 31P NMR spectroscopy, a small amount of secondary product (δ = 41.2) can be observed, which accounts for some of the additional signals in the 1H NMR spectroscopy. However, overall the spectra matches up and a few large shifts can be seen in the spectrum of the model complex compared to the spectrum of the free phosphine. A prominent shift can be observed in the resonance that occurs near 10 ppm in the model complex 3.5, which is assigned to the proton adjacent to the nitrogen on the pyridyl ring. When the compound is bound to palladium (II), as opposed to the free ligand, the resonance can be observed at Figure 3.7 – 31P{1H} NMR spectrum of model complex 3.5 in CDCl3 68 a higher frequency at approximately 8 ppm. The other phenyl resonances remain mostly unchanged. The resonance at approximately 5 ppm, which is assigned to the proton on the carbon adjacent to phosphorus, also shifts to a slightly lower position in response to the chelation of palladium (II) to the phosphorus. The signals of the methyls on the mesityl group remain mostly unchanged in the chelation of the ligand to palladium (II). Figure 3.8 – Comparison of the 1H NMR spectrum of model complex 3.5 in CDCl3 (top) against the 1H NMR spectrum of model compound 3.3 in CDCl3 (bottom). 69 3.2.4 X-Ray Analysis of Model Complex 3.5 To further investigate the structural features and bonding in model complex 3.5, the compound has been analyzed by X-ray crystallography. A summary of cell constants and data collection parameters for 3.5 is included in Table 3.1. The molecular structure of model complex 3.5 is shown in Figure 3.9. Important metrical parameters for 3.5 are tabulated in Table 3.2 and, for comparison, the metrical parameters are also provided for the closely related model complexes.45,46 70 Figure 3.9 – Molecular structure of (Mes)(Me)PC(Ph)(2-Pyd)PdCl2 (3.5). Ellipsoids are drawn at the 50% probability level; Atoms are labeled and hydrogens have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Pd(1) – Cl(1) 2.2960(8), Pd(1) – Cl(2) 2.3733(7), Pd(1) – P(1) 2.2083(7), Pd(1) – N(1) 2.074(2), P(1) – C(7) 1.861(3), N(1) – C(8) 1.359(3), C(7) – C(8) 1.514(4), Cl(2)-Pd(1)-P(1) 175.02(3), Cl(1)-Pd(1)-N(1) 173.25(7), Cl(1)-Pd(1)-P(1) 89.04(3), Cl(2)-Pd(1)-N(1) 95.07(7), P(1)-Pd(1)-N(1) 84.59(7), Pd(1)-P(1)-C(7) 103.73(9), Pd(1)-N(1)-C(8) 121.05(19), P(1)-C(7)-C(8) 109.74, N(1)-C(8)-C(7) 119.3(3). 71 Table 3.1 - X-Ray Crystallographic Data for 3.5 3.5 formula C22H24PNPdCl2 fw 510.69 cryst syst orthorhombic space group Fdd2 color yellow a (Å) 63.471(3) b (Å) 10.7688(4) c (Å) 12.1580(5) α (deg) 90 β (deg) 90 γ (deg) 90 V (Å3) 8310.1(6) T (K) 183(2) Z 16 µ (Mo Kα) (cm-1) 72.90 cryst size (mm3) 0.13 x 0.15 x 0.09 calcd density (Mg m-3) 1.633 2θ (max) (deg) 60.1 No. of reflns 31416 No. of unique data 5516 R (int) 0.038 Refln/param ratio 22.24 R1 0.030 wR2 (all data) 0.059 2-pyridine-substituted model complex 3.5 has a slightly distorted square planar geometry at palladium. The two Pd-Cl lengths are comparable with that trans to P, 2.3733(7) Å, somewhat longer than that trans to N, 2.2960(8) Å, due to the larger trans influence of the phosphine relative to that of the pyridine.47 The Pd-P distance, 2.2083(7) Å, is relatively short48, however it is reasonable for P trans to Cl. The Pd-N distance, 2.074(2) Å is typical for a pyridine-Pd bond.48,49 The Cl-Pd-Cl angle, 91.08(3)°, is 72 comparable to that of complex 3.7 (91.4(5)°)46 and reveals only minor distortions from the ideal value of a square planar molecule. The cis Cl-Pd-N angle (95.07(7)°), the trans P-Pd-Cl (175.02(3)°), and N-Pd-Cl (173.25(7)°) also have only minor distortions. The chelate angle P-Pd-N (84.59(7)°) is comparable to those of complexes 3.645 (85.98(18)°) and 3.746 (84.4(8)°) and is normal for such a five-membered ring and is presumably related to the slightly larger than 90° angles noted above. Complexes 3.6 and 3.7 are both bidentate ligands, which chelate through a phosphorus and a nitrogen. Both also have a five-membered ring once chelated to palladium (II), which makes them ideal for comparison to complex 3.5, which also has a five-membered ring once complexed. Complex 3.6 has a rigid backbone whereas complex 3.7 has a flexible backbone, by comparing complex 3.5 to both varieties of ligand, it can be observed how similar complex 3.5 is to each and how it differs. The biggest difference between the three complexes is that complex 3.5 has the nitrogen atom in a pyridyl ring whereas complexes 3.6 and 3.7 are a tertiary amine and a secondary amine, respectively. 73 Table 3.2 Important Metrical Parameters for P,N Bidentate Ligands bound to Palladium (II) 3.5 3.6a 3.7b Bond Lengths (Å) Pd – Cltrans N 2.2960(8) 2.378(2) 2.31(1) Pd – Cltrans P 2.3733(7) 2.291(2) 2.36(1) Pd – P 2.2083(7) 2.187(2) 2.19(1) Pd – N 2.074(2) 2.135(6) 2.16(3) P – Cbackbone 1.861(3) 1.798(7) 1.83(4) N – Cbackbone 1.359(3) 1.471(10) 1.50(4) Bond Angles (deg) ∠ Cl – Pd – P 175.02(3) 174.44(8) 177.6(5) ∠ Cl – Pd – N 173.25(7) 175.16(17) 174.0(8) ∠ Cl – Pd – Pcis 89.04(3) 89.24(8) 90.1(4) ∠ Cl – Pd – Ncis 95.07(7) 93.72(18) 93.7(8) ∠ P – Pd – N 84.59(7) 85.98(18) 84.4(8) ∠ Cl – Pd – Cl 91.08(3) 91.4(5) aReference 45. bReference 46. Pd P Cl N Cl Pd P Cl NH Cl O 3.6 3.7 74 3.2.5 Synthesis of Model Complex 3.8 In order to observe the binding characteristics of the phosphine model compound in the presence of transition metals, the model compound was reacted with palladium (II). Addition of palladium (II) resulted in the formation of a model complex that represents the related polymer. By observing the binding characteristics of the model complex, assumptions about the binding characteristics of the polymeric material can be made. To a solution of ortho-dimethylaniline-substituted model compound 3.4 was added one equivalent of (cod)PdCl2. The mixture was stirred and the reaction mixture was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.8 was formed (δ = 62.0). The solution was dried in vacuo and residual 1,5-cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was dried in vacuo and recrystallized via slow addition of hexanes to a solution of the model complex in dichloromethane. Like model complex 3.5, the 31P NMR spectroscopic shift for model complex 3.8 was also significant, from -20 ppm for the free phosphine to 62 ppm for the metal-bound Scheme 3.4 P CH Mes Ph Me N Pd Cl Cl 3.4 (cod)PdCl2 P N H CMes Ph Me 3.8 75 phosphine, if slightly larger than the shift for complex 3.5. The large shift verifies the coordination of the phosphorus atom to the palladium. Figure 3.11 shows a comparison of the 1H NMR spectrum of the model complex 3.8 to the 1H NMR spectrum of the model compound 3.4. The spectrum of both compounds match fairly well. The spectrum of the model complex 3.8 shows a slight shift to lower field for some of the signals in the phenyl region, however unlike in the spectrum for model complex 3.5, there are fewer phenyl protons directly affected by the bonding of the compound to the palladium atom. An overall broadening of the spectra as well as the low field shift of the amine methyls (2.5 ppm - 2.7 ppm) indicate the coordination of the nitrogen to the palladium. Figure 3.10 – 31P{1H} NMR spectrum of model complex 3.8 in CDCl3. 76 Unfortunately, a crystal of complex 3.8 was not grown that was sufficient for X- ray crystallography to be performed in order to examine the compound more thoroughly. Since a crystal was unobtainable, another complex of the ortho-dimethylaniline- substituted model compound was reacted with palladium (II) in a two to one fashion. 3.2.6 Synthesis of Model Complex 3.9 The complex was synthesized following the same synthetic procedure of complex 3.8, however two equivalents of the model compound 3.4 was added to one equivalent of Figure 3.11 – Comparison of the 1H NMR spectrum of model complex 3.8 in CDCl3 (bottom, 1) against the 1H NMR spectrum of model compound 3.4 in CDCl3 (top, 2). 77 (cod)PdCl2 in dichloromethane. The mixture was stirred and the reaction mixture was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.9 was formed (δ = 23.1). The solution was dried in vacuo and residual 1,5-cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was dried in vacuo and recrystallized via slow addition of hexanes to a solution of the model complex in dichloromethane. Like model complex 3.5 and 3.8, the 31P NMR spectroscopic shift for model complex 3.9 was also significant, from -20 ppm for the free phosphine to 23 ppm for the metal-bound phosphine, if not quite as large as either. The shift indicates coordination of the soft phosphorus atom to the palladium compound, but the differences in the shift (∆ 43 ppm) compared to the much larger shift of model complex 3.8 (∆ 82 ppm), indicate that the coordination of the model compound to palladium(II) is clearly different. Scheme 3.5 P CH Mes Ph Me N 3.4 0.5 eq(cod)PdCl2 3.9 P H C Mes Pd Me Ph NMe2 Cl Cl PC Mes Me PhMe2N H 78 The 1H NMR spectrum was unobtainable due to lack of solubility. Model compound 3.9 was insoluble in CDCl3 resulting in the inability to run an NMR spectroscopy experiment. The lack of the 1H NMR spectrum regrettably limits the analysis of the compound and therefore cannot be compared to the 1H NMR spectrum of the free compound 3.4 and the 1:1 compound 3.8. However, unlike the model complex 3.8, crystallization of model complex 3.9 resulted in a crystal sufficient for X-ray crystallography analysis, which gives a clear image of the binding characteristics of the model complex. Figure 3.12 – 31P{1H} NMR spectrum of model complex 3.9 in THF. 79 3.2.7 X-Ray Analysis of Model Complex 3.9 To further investigate the structural features and bonding in model complex 3.9, the compound has been analyzed by X-ray crystallography. A summary of cell constants and data collection parameters for 3.9 is included in Table 3.3. The molecular structure of model complex 3.9 is shown in Figure 3.13. Important metrical parameters for 3.9 are tabulated in Table 3.4 and, for comparison, the metrical parameters are also provided for the 2-pyridyl-substituted model complex 3.5. 80 Figure 3.13 – Molecular structure of ((Mes)(Me)PC(Ph)(o-DMA))2PdCl2 (3.9). Ellipsoids are drawn at the 50% probability level; Atoms are labeled and hydrogens have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Pd(1) – P(1) 2.3202(11), Pd(1) – P(2) 2.3166(10), Pd(1) – Cl(1) 2.3100(11), Pd(1) – Cl(2) 2.3066(11), P(1) – C(2) 1.872(4), P(2) – C(27) 1.873(4), Cl(1)-Pd(1)-Cl(2) 168.37(4), P(1)-Pd(1)-P(2)172.95(4), Cl(1)-Pd(1)-P(2) 87.69(4), Cl(1)-Pd(1)-P(1) 92.30(4), Cl(2)- Pd(1)-P(2) 92.03(4), Cl(2)-Pd(1)-P(1) 86.56(4), C(2)-P(1)-Pd(1) 115.26(13), C(27)- P(2)-Pd(1) 115.85(13). 81 Table 3.3 - X-Ray Crystallographic Data for 3.9 3.9 formula C50H60P2N2PdCl2 fw 928.24 cryst syst monoclinic space group P 21/c color yellow a (Å) 13.0239(9) b (Å) 10.006(1) c (Å) 40.248(4) α (deg) 90 β (deg) 95.007(3) γ (deg) 90 V (Å3) 5586.2(2) T (K) 183(2) Z 4 µ (Mo Kα) (cm-1) 5.15 cryst size (mm3) 0.03 x 0.05 x 0.16 calcd density (Mg m-3) 1.104 2θ (max) (deg) 45 No. of reflns 58860 No. of unique data 7249 R (int) 0.123 Refln/param ratio 13.91 R1 0.079 wR2 (all data) 0.088 Ortho-dimethylaniline-substituted model complex 3.9 has a slightly distorted square planar geometry at palladium. The Pd-P lengths (2.3202(11) Å and 2.3166(10) Å) are longer than those for the model complex 3.5 (2.2083(7) Å) due to the fact that the phosphine groups are trans to each other. The Cl-Pd-Cl and P-Pd-P angles are slightly less than 180° (168.37(4)° and 172.95(4)° respectively) showing the slight distortion of the square planar geometry at palladium. The four P-Pd-Cl angles are all roughly 90° 82 with two slightly larger angles (92.30(4)°, 92.02(4)°) and two with slightly smaller angles (87.69(4)°, 86.56(4)°). Table 3.4 Important Metrical Parameters for 3.9 and 3.5 3.9a 3.5 Bond Lengths (Å) Pd – P 2.3202(11) 2.3166(10) 2.2083(7) Pd – Cl 2.3100(11) 2.3066(11) 2.3733(7)b 2.2960(8)c P – C 1.872(4) 1.873(4) 1.861(3) Bond Angles (deg) ∠ Cl – Pd – P 87.69(4) 86.56(4) 92.30(4) 92.03(4) ∠ P – Pd – P 172.95(4) ∠ Cl – Pd – Cl 168.37(4) ∠ Pd – P – C 115.26(13) 115.85(13) aCompound contains two molecules bound to one palladium, parameters for both molecules are listed. bDistance listed for Pd-Cl trans to P. cDistance listed for Pd-Cl cis to P The chelating characteristics of compound 3.9 give an interesting insight into how a polymeric material of the phosphaalkene (3.2) would bind transition metals. In a polymeric material, the number of atoms capable of binding metals is ideally higher than the number of transition metal atoms to be removed from solvent. If that is the case, then the softer phosphorus atoms are more likely to bind to the transition metals than the harder nitrogen atoms. In the case of model complex 3.9, the polymeric material would bind with two of the phosphorus atoms, resulting in the cross-linking of the polymer and 83 more securely attaching to the transition metal atom by surrounding the transition metal atom with the organic portions of the polymer. The isolation of the transition metal from the solvent solution would increase the effectiveness of a material in scavenging transition metals from solution. 3.3 Summary In this section, several new compounds were synthesized and examined. In the first subsection, the synthesis of new model compounds was discussed and their properties examined. In the second subsection the model compounds were chelated to palladium (II) in order to examine their binding properties. The pyridine-substituted model compound formed a new model complex with both the phosphorus and nitrogen of the pyridine bound to the palladium (II) forming a 5-sided chelate ring. Addition of palladium (II) to the ortho-dimethylaniline-substituted model compound resulted in two new model complexes, one with a one-to-one ratio of model compound to palladium (II) and the other with a two-to-one ratio of model compound to palladium (II). The first was characterized via NMR spectroscopy and was shown to be capable of binding through both the phosphorus and the nitrogen atoms to the palladium. The second was crystallized and proven to have two model compounds bound to a palladium (II) atom through the phosphorus of each atom with the nitrogen atoms not involved in the binding. This two-to-one binding shows an interesting property in the potential ability of the polymer to bind transition metals by cross-linking of the polymer. Polymer cross-linking could result in isolating the desired transition metals from solution by surrounding the 84 atoms with the organic portions of the cross-linked polymer, effectively removing the metals from the solvent. The work done in this section shows the metal scavenging potential of the polymers derived from bidentate phosphaalkenes by examining the binding properties of their related model complexes. 3.4 Experimental General Procedures. All manipulations of air- and/or water-sensitive compounds were performed under a nitrogen atmosphere using standard Schlenk or glovebox techniques. Hexanes and dichloromethane were deoxygenated with nitrogen and dried by passing through a column containing activated alumina. THF was freshly distilled from sodium/benzophenone ketyl. Distilled water was degassed prior to use. CDCl3 was distilled from P2O5 and degassed. The complex (cod)PdCl2 were prepared according to literature procedures.50 Methyllithium was purchased from Aldrich and used as received. MesP(SiMe3)249 was prepared following literature procedures. Equipment. 1H, 31P, and 13C NMR spectra were recorded at room temperature on Bruker Avance 300 or 400 MHz spectrometers. Chemical shifts are reported relative to residual CHCl3 (δ = 7.24 for 1H), 85% H3PO4 as an external standard (δ = 0.0 for 31P), and CDCl3 (δ = 77.0 for 13C). Assignments of NMR spectra were made with the aid of 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, 1H-31P HMBC experiments. Elemental analyses were performed in the University of British Columbia Chemistry Microanalysis 85 Facility. Mass spectra were recorded on a Kratos MS 50 instrument in EI mode (70 eV). X-ray crystallography was performed on a Bruker X8 APEX diffractometer with graphite-monochromated Mo Kα radiation. 3.4.1 Preparation of (Mes)(Me)P-CH(Ph)(2-Py) (3.3) To a stirred solution at RT of MesP=C(Ph)(2-Pyd) (0.50g, 1.57mmol) dissolved in THF (5 mL) was added rapidly MeLi (1.6M, 1.2mL, 1.88mmol). The reaction was monitored by 31P NMR spectroscopy(δ = -47) and after twenty minutes the solution was quenched with MeOH to form the product (δ = -22). The solvent was removed in vacuo leaving a pale yellow solid. To the solid was added hexanes (3 x 10 mL) and the suspension was filtered and the solvent removed in vacuo. The product was purified by crystallization in hexanes to give a pale yellow solid. Yield = 0.50g (96%). 31P NMR (CDCl3): δ -22.04. 1H NMR (CDCl3): δ 8.66-6.74 (m, 11H, Ar), 4.99 (s, 1H, CH), 2.43 (s, 6H, o-CH3), 2.20 (s, 3H, p-CH3), 1.35 (s, 3H, P-CH3). 13C{1H} (CDCl3): δ 149.5 – 121.5 (m, 17C, Ar), 53.7 (s, 1C, P-CH), 23.6, 23.4, 21.0 (s, 3C, Ph- CH3), 10.3 (s, 1C, P-CH3). MS (EI): 333, 334 [M++H], 318 [M+ - CH3], 168 [M+ - P(Me)(Mes)]. Anal. Calcd for C22H24PN: C, 79.2; H, 7.2; N, 4.2. Found: C, 78.08; H, 7.18; N, 4.55. 3.4.2 Preparation of (Mes)(Me)P-CH(Ph)(o-DMA) (3.4) To a stirred solution at RT of MesP=C(Ph)(o-DMA) (0.28g, 0.78mmol) dissolved in THF (5 mL) was added rapidly MeLi (1.6M, 0.6mL, 0.93mmol). The reaction was monitored by 31P NMR spectroscopy(δ = -47.7 ) and after twenty minutes the solution 86 was quenched with MeOH to form the product (δ = -20.9). The solvent was removed in vacuo leaving a pale yellow solid. To the solid was added hexanes (3 x 10 mL) and the suspension was filtered and the solvent removed in vacuo. The product was purified by crystallization in hexanes to give a pale yellow solid. Yield = 0.27g (94%). 31P NMR (CDCl3): δ -20.9. 1H NMR (CDCl3): δ 7.95-6.70 (m, 11H, Ar), 5.82 (s, 1H, CH), 2.52 (s, 6H, N(CH3)2), 2.37 (s, 6H, o-CH3), 2.17 (s, 3H, p-CH3), 1.33 (s, 3H, P- CH3). 13C{1H} (CDCl3): δ 152.9 – 120.6 (m, 18C, Ar), 45.9 (s, 1C, P-CH), 42.9 (s, 2C, N(CH3)2), 23.2, 21.0 (s, 3C, Ph-CH3), 10.3 (s, 1C, PCH3). MS (EI): 375 [M+], 210 [M+ - P(Me)(Mes)]. Anal. Calcd for C25H30PN: C, 79.95; H, 8.07; N, 3.73. Found: C, 79.78; H, 8.00; N, 3.49. 3.4.3 Preparation of (Mes)(Me)P-CH(Ph)(2-Py)PdCl2 (3.5) To a solution of 2-pyridine-substituted model compound 3.3 (0.05 g, 0.15 mmol) was added (cod)PdCl2 (0.042g, 0.15mmol). The mixture was stirred and the reaction progress was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.5 was formed (δ = 45.54). The solution was vacuum dried and residual 1,5-cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was vacuum dried and recrystallized via slow addition of hexanes to a solution of product and dichloromethane. 31P NMR (CDCl3): δ 45.54. 1H NMR (CDCl3): δ 9.9 (s, Ar), 7.8 (m, Ar), 7.3 (m, Ar), 6.9 (m, Ar), 5.3 (d, CH), 3.3 (s), 3.0 (s), 2.6 (s, Ar-CH3), 2.5 (s, Ar-CH3), 2.2 (s, Ar-CH3), 1.8 (s), 1.7 (s), 1.4 (s, PCH3). Anal. Calcd for C22H24PNPdCl2: C, 51.85; H, 4.71; N, 2.74. Found: C, 52.63; H, 4.88; N, 2.64. 87 3.4.4 Preparation of (Mes)(Me)P-CH(Ph)(o-DMA)PdCl2 (3.8) To a solution of ortho-dimethylaniline-substituted model compound 3.4 (0.09 g, 0.24 mmol) was added (cod)PdCl2 (0.066g, 0.24mmol). The mixture was stirred and the reaction progress was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.8 was formed (δ = 62.07). The solution was vacuum dried and residual 1,5- cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was vacuum dried and recrystallized via slow addition of hexanes to a solution of product and dichloromethane. 31P NMR (CDCl3): δ 62.07 1H NMR (CDCl3): 8.2 (br, m, Ar), 7.7 – 7.2 (br, m, Ar), 6.6 (br, s, Ar), 2.7 (br, s, NCH3), 2.2 (br, s, Ar-CH3), 1.3 (br, s, PCH3). Anal. Calcd for C25H30PNPdCl2: C, 54.43; H, 5.43; N, 2.53. Found: C, 54.65; H, 5.64; N, 2.37. 3.4.5 Preparation of [(Mes)(Me)P-CH(Ph)(o-DMA)]2PdCl2 (3.9) To a solution of ortho-dimethylaniline-substituted model compound 3.2b (0.1 g, 0.27 mmol) was added (cod)PdCl2 (0.037g, 0.13mmol). The mixture was stirred and the reaction progress was monitored by 31P NMR spectroscopy. After 1 hour, the complex 3.3e was formed (δ = 23.18). The solution was vacuum dried and residual 1,5- cyclooctadiene was removed by a cold hexane wash (3x). The pale yellow product was vacuum dried and recrystallized via slow addition of hexanes to a solution of product and dichloromethane. 31P NMR (CDCl3): δ 23.18, 20.21 88 Chapter 4 Overall Conclusions and Future Work 4.1 Summary of Thesis Work Previously, a phosphaalkene (4.1) bearing a pyridine substituent was synthesized via the standard phospha-Peterson reaction.33 This phosphaalkene was shown to chelate to the late transition metals palladium and platinum (Scheme 4.1), indicating that the phosphaalkene could fill the role of a ligand.37 The ability of the phosphaalkene to bind late transition metals such as palladium and platinum is critical in achieving one goal of this thesis, which is to discover a new bidentate phosphaalkene capable of binding transition metals like palladium. This previous work shows that the potential exists for other bidentate phosphaalkenes to be synthesized with the ability to bind transition metals in a similar manner. It has also been shown that phosphaalkenes can be polymerized using simple polymerization techniques usually reserved for olefins.24 Phosphaalkenes (PA) can be P C Mes Ph N (cod)MCl2 CH2Cl2 M = Pd, Pt4.1 4.2a-b M N P Cl CMes Ph Cl Scheme 4.1 89 polymerized via radical or anionic initiation resulting in poly(methylenephosphine) (PMP) (Scheme 4.2). More work in this area has also shown that phosphaalkenes are capable of copolymerization with olefins (O) such as styrene and isoprene to afford random (PMP-r-PO) and block-copolymers (PO-b-PMP) (Scheme 4.3).32 The living polymerization allows control over the composition and molecular weight of the polymers. With such control, one can control the number of bidentate molecules within the polymer as a whole. P C P C n PA PMP Init = R. or R- [I] Scheme 4.2 PA P C C C P C n C C x y R. PMP-r-PO C Cn R- C C P C n mPA P Cm PO-b-PMP Scheme 4.3 O O 90 The combination of these two ideas led to the overall goal of this thesis: to create a new material capable of chelating late transition metals, like palladium and platinum. The new material would be efficient as a metal scavenging material for the removal of palladium and other metals from active pharmaceutical ingredients. We have now successfully prepared a new bidentate phosphaalkene MesP=C(o- DMA)(Ph) (o-DMA = ortho-dimethylaniline) (4.4). The synthesis of 4.4 was achieved by the standard phospha-Peterson reaction (Scheme 4.4). Crystals suitable for X-ray diffraction analysis were obtained for 4.4. Radical polymerization of 4.4 afforded a new phosphorus-carbon polymer 4.5 (Scheme 4.5). Polymer 4.5 was analyzed by GPC to show clearly the success of the polymerization reaction. All of the synthetic work described above is presented in Chapter 2. P C Mes Ph Mes P SiMe3 SiMe3 MeLi Mes P Li SiMe3 N O NMe2 4.3 4.4 60oC 1-2 hrs -78oC Scheme 4.4 91 To further understand the metal binding capabilities of the bidentate phosphaalkenes and their derivative polymers, model compounds were synthesized from phosphaalkenes 4.1 and 4.4. This was achieved by lithiation of the phosphaalkenes followed by addition of methanol. This reaction afforded the phosphines 4.6 and 4.7 (Scheme 4.6). P C Mes Ph N P C Mes Ph n N VAZO 4.4 4.5 200oC 48 hr Scheme 4.5 Scheme 4.6 P C Mes Ph N P CH Mes Ph Me N MeLi MeOH 4.1 4.6 P C Mes Ph N P CH Mes Ph Me N MeLi MeOH 4.4 4.7 92 To show in detail how the phosphorus polymers bind transition metals, the phosphines 4.6 and 4.7 were chelated to palladium (II) (Scheme 4.7). Each new compound was analyzed by NMR spectroscopy and crystals suitable for x-ray diffraction analysis were obtained for complexes 4.8 and 4.10. The analysis of the data gives a representation of how the bidentate polymers of phosphaalkenes 4.1 and 4.4 will bind to late transition metals. 4.6 4.8 P CH Mes Ph Me N Pd Cl Cl(cod)PdCl2 P N Me CH Ph Mes Scheme 4.7 P CH Mes Ph Me N Pd Cl Cl 4.7 (cod)PdCl2 P N H CMes Ph Me 4.9 P CH Mes Ph Me N 4.7 0.5 eq(cod)PdCl2 4.10 P H C Mes Pd Me Ph NMe2 Cl Cl PC Mes Me PhMe2N H 93 4.2 Future Work 4.2.1 New Bidentate and Tridentate Phosphaalkenes While our studies have provided spectroscopic evidence that shows the synthesis of bidentate phosphaalkenes is possible, further work is needed to expand the library of bidentate phosphaalkenes. With further work it is possible to synthesis a series of bidentate, or even tridentate, phosphaalkenes capable of chelating late transition metals like palladium and platinum. The inclusion of different chelating atoms, such as oxygen and sulfur, in the synthesis of new bidentate phosphaalkenes would diversify the binding strengths of the resultant polymers. This would perhaps allow a wider variety of transition metals to be chelated, resulting in polymers with the potential to scavenge a wider range of transition metals. Tridentate phosphaalkenes are also a potential area for future work. Phosphaalkenes with two chelating atoms in addition to the phosphorus could also increase the strength of the bonds to transition metals. 4.2.2 Copolymerization of Phosphaalkenes The studies discussed in this thesis that pertain to polymerization show that the phosphaalkene 4.4 is polymerizable. While a homopolymer of a bidentate phosphaalkene is a great achievement, copolymerization of the bidentate phosphaalkene with styrene or another olefin would lead to a wider variety of metal-chelating polymers. With a wider range of phosphorus content in the main chain of the polymer, the effectiveness of the polymer at metal scavenging can be tuned and perfected. Copolymerization with styrene would result in a library of polymers with different metal-scavenging capabilities and allow for the choice of material best suited to a specific situation. 94 4.2.3 Metal-Scavenging Tests Work in this thesis has shown that the model compounds 4.6 and 4.7, which represent the macromolecular polymer, are capable of binding palladium. This indicates that the polymer 4.5 is also capable of binding palladium. However, the work discussed in this thesis does not show exactly how capable polymer 4.5 is at scavenging metals from solution. A series of experiments need to be completed to show how much palladium a certain amount of polymer can remove from solution. These experiments would demonstrate the metal-scavenging capabilities of the bidentate polymer. The model complexes 4.9 and 4.10 indicate that polymer 4.5 will be an excellent metal- scavenging material, however the experiments need to be run to generate exact numbers to show how effective it really is. 4.3 Concluding Remarks The synthesis of a new bidentate phosphaalkene and its derivative polymer was accomplished. An investigation into the metal-binding capabilities of the phosphaalkenes and their derivative polymers was accomplished by the synthesis of model compounds. The model compounds were chelated to palladium and x-ray analysis shows the binding characteristics of the model complexes. The results obtained by these studies represent the foundation towards metal-scavenging materials derived from bidentate phosphaalkenes. Future development of new bidentate phosphaalkenes and the polymers and copolymers derived from them is an exciting and expanding area of research, and thus the story of bidentate phosphaalkenes and their derivatives is to be continued. 95 Bibliography 1) Carey, J.S.; Laffan, D.; Thomson, C.; Williams, M. T. Org. Biomol. Chem. 2006, 4, 2337. 2) Zapf, A.; Beller, M. Top. Catal. 2002, 19, 101. 3) Baiker, A. J. Mol. Cat. A: Chem. 1997, 115, 473 4) Ohta, T.; Takaya, H.; Kitamura, M.; Nagai, K.; Noyori, R. J. Org. Chem. 1987, 52, 3174 5) Appleby, I.; Boulton, L. T.; Cobley, C.J.; Hill, C.; Hughes, M.L.; Koning, P.D.; Lennon, I.C.; Praguin, C.; Ramsden, J.A.; Samuel, H.J.; Willis, N. Org Lett. 2005, 7, 1931 6) Garrett, C. E.; Prasad, K. Adv. Synth. Catal. 2004, 346, 889. 7) Abernethy, D. R.; DeStefano, A. J.; Cecil, T. L.; Zaidi, K.; Williams, R. L.; USP Metal Impurities Advisory Panel. Pharmaceutical Research 2010, 27, 750. 8) Thayer, A. Chem. Eng. News. 2005, 83, 55. 9) Beutler, U.; Schneitter, R.; Schorter, P. (Novartis), personal communication. 10) Prashad, M.; Liu, Y.; Repic, O. Adv. Synth. Catal. 2003, 345, 533. 11) Chen, C.-Y.; Dagneau, P.; Grabowski, E. J. J.; Oballa, R.; O’Shea, R.; Prasit, P.; Robichaud, J.; Tillyer, R.; Wang, X. J. Org. Chem. 2003, 68, 2633. 12) Königsberger, K.; Chen, G.-P.; Wu, R.; Girgis, M. J.; Prasad, K.; Repic, O.; Blacklock, T. J. Org. Proc. Res. Dev. 2003, 7, 733. 13) Manley, P. W.; Acemoglu, M.; Marterer, W.; Pachinger, W. Org. Proc. Res. Dev. 2003, 7, 436. 96 14) Prashad, M.; Bin, H.; Har, D.; Repic, O.; Blacklock, T. J. Adv. Synth. Catal. 2001, 343, 461. 15) Rosso, V. W.; Lust, D. A.; Bernot, P. J.; Grosso, J. A.; Modi, S. P.; Rsowicz, A.; Sedergran, T. C.; Simpson, J. H.; Srivastava, S. K.; Humora, M. J.; Anderson, N. G. Org. Proc. Res. Dev. 1997, 1, 311. 16) Ishihara, K.; Nakayama, M.; Kurihara, H.; Itoh, A.; Haraguchi, H. Chem. Lett. 2000, 1218. 17) Urawa, Y.; Miyazawa, M.; Ozeki, N.; Ogura, K. Org. Proc. Res. Dev. 2003, 7, 191. 18) Barbaras, D.; Borzio, J.; Johannsen, I.; Allmendinger, T. Org. Proc. Res. Dev. 2009, 13, 1068. 19) Fellmann, J. D.; Garrou, H. P.; Withers, H. P.; Seyferth, D.; Traficante, D. Organometallics 1983, 2, 818. 20) Baumgartner, T.; Reau, R. Chem. Rev. 2006, 106, 4681. 21) Peckham, T. J.; Massey, J. A.; Honeyman, C. H.; Manners, I. Macromolecules, 1999, 32, 2830. 22) Naka, K.; Umeyama, T.; Nakahashi, A.; Chujo, Y. Macromolecules, 2007, 40, 4854. 23) Ouchi, Y.; Morisaki, Y.; Ogoshi, T.; Chujo, Y. Chem. Asian. J. 2007, 2, 397. 24) Tsang, C.-W.; Yam, M.; Gates, D. P. J. Am. Chem. Soc. 2003, 125, 1480. 25) Regitz, M.; Scherer, O. J. Multiple bond and low coordination in phosporus chemistry, Thieme, Stuttgart, 1990. 97 26) Dillon, K. B.; Mathey, F.; Nixon, J. F. Phosphorus: the carbon copy, Wiley, 1998. 27) Mathey, F. Angew. Chem. Int. Ed. 2003, 42, 1578. 28) Gillon, B. H.; Patrick, B. O.; Gates, D. P. Chem. Commun. 2008, 2161. 29) Noonan, K. J. T.; Gates, D. P. Angew. Chem. Int. Ed. 2006, 45, 7271. 30) Noonan, K. J. T.; Gates, D. P. Macromolecules 2008, 41, 1961. 31) Noonan, K. J. T.; Gillon, B. H.; Cappello, V.; Gates, D. P. J. Am. Chem. Soc. 2008, 130, 12876. 32) Tsang, C.-W.; Baharloo, B.; Riendl, D.; Yam, M.; Gates, D. P. Angew. Chem. Int. Ed. 2004, 43, 5682. 33) Yam, M.; Chong J. H.; Tsang, C.; Patrick, B. O.; Lam, A. E.; Gates, D. P. Inorg. Chem. 2006, 45, 5225. 34) Jouaiti, A.; Geoffroy, M.; Bernardenelli, G. Tetrahedron Lett. 1992, 33, 5071. 35) Ionkin, A. S.; Marshall, W. J. Heteroat. Chem. 2002, 13, 662 36) Van der Sluis, M.; Beverwijk, V.; Termaten, A.; Gavrilova, E.; Bickelhaupt, F.; Kooijman, H.; Velman, N.; Spek, A. L. Organometallics 1997, 16, 1144. 37) Dugal-Tessier, J.; Dake, G. R.; Gates, D. P. Organometallics 2007, 26, 6481. 38) Creencia, E. C.; Takahashi, A.; Horaguchi, T. Heterocycles 2009, 78 (6), 1549. 39) Becker, G.; Uhl, W.; Wessely, H. J. Z. Anorg. Allg. Chem. 1981, 479, 41. 40) Appel, R. In Multiple Bonds and Low Coordination in Phosphorus Chemistry; Regitz, M., Scherer, O. J., Eds.; Thieme: Stuttgart Germany, 1990. 41) Characteristic bond lengths in free molecules. CRC Handbook of Chemistry and Physics, 84th ed.; CRC Press: Boca Raton, FL, 2003. 98 42) Alessi, M.; Larkin, A. L.; Ogilvie, K. A.; Green, L. A.; Lai, S.; Lopez, S.; Snieckus, V. J. Org. Chem. 2007, 72, 1588. 43) Slocum, D. W.; Reece, T.; White, C. B.; Sandlin, R. D.; Lalonde, K.; Bolton, J.; Warnell, S.; Bragg, S. A.; Whitley, P. E. Abstracts, 235th National Meeting of the American Chemical Society, New Orleans, LA, April 2008, section ORGN, paper 594. 44) Holmberg, G. A. Acta Chem. Scand. 1955, 9, 555. 45) Albinati, A.; Lianza, F.; Berger, H.; Pregosin, P. S.; Rüegger, H.; Kunz, R. W. Inorg. Chem. 1993, 32, 478. 46) Albinati, A.; Lianza, F.; Berger, H.; Arz, C.; Pregosin, P.S. Inorg. Chim, Acta 1992, 198-200, 771. 47) Orpen, A. G.; Brammer, L.; Allen, F. H.; Kennard, O.; Watson, D. G.; Taylor, R. J. Chem. Soc. Dalton Trans. 1989, S1. 48) Palenik, G. J.; Giordano, T. J. J. Chem. Soc. Dalton Trans. 1987, 1175. 49) McDermott, J.X.; White, J.F.; Whitesides, G.M. J. Am. Chem. Soc. 1976, 98, 6521. 50) Dai, Y.; Feng, X.; Wang, B.; He, R.; Bao, M. J. Organometallic Chem. 2012, 696, 4309. 99 Appendix A: 2D NMR Experiments 1H-1H COSY NMR: Zoom 7.0 ppm – 6.25 ppm/ 2.0 ppm – 2.5 ppm 100 1H-1H COSY NMR: Zoom 7.9 ppm – 6.25 ppm/ 7.9 ppm – 6.25 ppm 101 1H-31P HMBC NMR: Zoom 8.0 ppm – 0.0 ppm/ 246 ppm – 223 ppm 102 1H-13C HMBC NMR: Zoom 7.1 ppm – 6.25 ppm/ 27 ppm – 10 ppm 103 1H-13C HMBC NMR: Zoom 7.9 ppm – 6.7 ppm/ 155 ppm – 115 ppm 104 1H-13C HMBC NMR: Zoom 2.8 ppm – 1.9 ppm/ 155 ppm – 115 ppm 105 1H-13C HSQC NMR: Zoom 2.8 ppm – 2.0 ppm/ 50 ppm – 19 ppm 106 1H-13C HSQC NMR: Zoom 8.0 ppm – 6.6 ppm/ 138 ppm – 112 ppm