Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Aqueous solution studies of multidentate ligands with trivalent metal ions Caravan, Peter 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-195597.pdf [ 9.06MB ]
Metadata
JSON: 831-1.0059510.json
JSON-LD: 831-1.0059510-ld.json
RDF/XML (Pretty): 831-1.0059510-rdf.xml
RDF/JSON: 831-1.0059510-rdf.json
Turtle: 831-1.0059510-turtle.txt
N-Triples: 831-1.0059510-rdf-ntriples.txt
Original Record: 831-1.0059510-source.json
Full Text
831-1.0059510-fulltext.txt
Citation
831-1.0059510.ris

Full Text

AQUEOUS SOLUTION STUDIES OF MULTIDENTATE LIGANDS WITH TRIVALENT METAL IONS by Peter Caravan B.Sc. (Hons.), Acadia University, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 1996 © Peter Caravan In presenting this thesis. in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholariy purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) Abstract Four water-soluble amine phenols have been prepared: tris(((2-hydroxy-5-sulfobenzyl)amino)ethyl)amine (H6TRNS), l,l,l-tris(((2-hydroxy-5-sulfobenzyl)amino)methyl)ethane (H^TAMS), l,2,3-tris(((2-hydroxy-5-sulfobenzyl)amino)propane (H6TAPS), cis, ds-l,3,5-tris(((2-hydroxy-5-sulfobenzyl)amino)cyclohexane (HgTACS). Complex formation constants have been measured at 25 °C (\i = 0.16 M NaCl): log K [M(TRNS)]3' ([M(HTRNS)]2") Ga 28.55 (36.90), In 29.3 (34.9); log K [M(H3TRNS)] ([M(H3TRNS)2]3") La 5.65, Nd 6.41 (6.34), Gd 6.67 (7.69), Ho 7.67 ( 8.75), Yb 8.53 (9.73); log K [M(TAMS)]3" ([M(HTAMS)]2') Al 22.5 (29.3), Ga 31.83, In 28!49, La 9.17, Nd 11.19, Gd 11.86 (18.41), Ho 12.71 (19.40), Yb 13.78 (20.11); log K [M(TAPS)]3- ([M(HTAPS)]2") Al 22.8 (29.0), Ga 31.54 (35.15), In 27.56 (31.93), La 11.33 (18.47), Nd 13.59 (20.13), Gd 14.50 (20.88), Ho 14.71 (21.15), Yb 15.15 (21.54). The complexes were characterized in situ by multinuclear ( 1 H, 1 3 C , 1 7 0 , 2 7 A1, 7 1 G a , 1 1 5In) NMR and UV spectroscopy. A new preparation of N,N'-bis(2-methylpyridyl)ethylenediamine-N,N'-diacetic acid (H2bped) is reported and some trivalent metal ion complexes have been prepared. Complex formation constants have been measured at 25 °C (\i = 0.16 M NaCl): log K [M(bped)]+ ([M(bped)(OH)]) Al 10.85 (6.37), Ga 19.89 (15.62), In 22.6 (15.44), La 10.81 (10.75), Nd 11.99 (10.45), Gd 12.37 (10.25), Ho 12.31 (9.31), Yb 13.42 (8.98). The complexes have been characterized in the solid state as PF6" or C104" salts by IR, +LSIMS, elemental analysis, and by X-ray crystallography (Co[bped]PF6); and in solution by multinuclear NMR and UV spectroscopy. ii TABLE OF CONTENTS Abstract ii Table of Contents iii List of Figures v List of Tables x List of Abbreviations xii Acknowledgments xviii Dedication xix Chapter 1 General Introduction 1 References 16 Chapter 2 Tripodal Aminophenolate Ligand Complexes 19 of Al(III), Ga(III), and In(III) in Water 2.1 Introduction 19 2.2 Experimental 23 2.3 Results and Discussion 38 2.4 Conclusions 71 2.5 References 72 Chapter 3 Aqueous Solution Studies of Tripodal 76 Aminophenolate Ligands with the Lanthanides 3.1 Introduction 76 3.2 Experimental 80 3.3 Results 87 3.4 Discussion 99 3.5 Conclusions 113 3.6 References 114 iii Chapter 4 On the Effect of Pyridyl Donors in the Chelation 118 of Al(III), Ga(III), and In(III) 4.1 Introduction 118 4.2 Experimental 121 4.3 Results 128 4.4 Discussion 141 4.5 Conclusions 153 4.6 References 154 Chapter 5 Cationic Lanthanide Complexes of N,N'-Bis(2- 158 methylpyridyl)ethylenediamine-N,N'-diacetic acid 5.1 Introduction 158 5.2 Experimental 160 5.3 Results 164 5.4 Discussion 173 5.5 Conclusions 179 5.6 References 179 Chapter 6 Conclusions and Further Thoughts 183 6.1 General Conclusions 183 6.2 Suggestions for Future Work 185 6.3 References 187 Appendix I X-Ray Crystallographic Data for 188 [Co(bped)][PF 6]»CH 3CN»H 20. iv List of Figures Figure 1.1. log Keff vs pH for FeL (top), GaL (middle), and InL 15 (bottom). L = edta ; L = HBET ; L = HBED Figure 2.1. Variable p H UV Spectra of H 6 TAPS (~2 mM). 40 Uncorrected for Dilution. Figure 2.2. Plot of Molar Extinction Coefficient (e, M'^cm"1) at 41 256 nm vs. pH for H 6 TRNS, H 6 TAPS, H 6 TACS, and H 6 T A M S . Figure 2.3. Variable pH 1 H NMR Titration of H 6 TAPS. 43 Figure 2.4. Aliphatic and Benzylic Portions of 1 H NMR Spectrum 45 (200 MHz) of H 6 TAPS at Various pD. Figure 2.5. Aliphatic and Benzylic Portions of 1 H NMR Spectrum 47 (300 MHz) of H 6 TAPS at pD = 4.3; Bottom (Simulated), Middle (25 °C), Top (85 °C). Figure 2.6. Speciation Diagram of 2 mM M(III) : 2 mM H 6 TAMS. 52 Dashed Lines Indicate the Region Where Precipitation Occurs. Charges Omitted for Simplicity. Figure 2.7. Speciation Diagram of 2 mM M(III) : 2 mM H 6 TAPS. 54 Dashed Lines Indicate the Region Where Precipitation Occurs. Charges Omitted for Simplicity. Figure 2.8. Experimental Titration Curves for 2 mM M(III) : 2 mM 55 H 6 L . Dashed Lines Represent H 6 TAPS Titrations, Solid Lines Represent H 6 T A M S . Figure 2.9. Partial Speciation Diagram ([Ga(OH) x]n + Species Omitted 57 for Clarity) for 1 Ga(III): 1 H 6 TAPS (30 uM) and the Molar Extinction Coefficient (e, M^cm" 1) at 249 nm vs. pH. [Ga(HTAPS)]2"- - ^ [Ga(TAPS)]3" / ([Ga(HTAPS)]2" + [Ga(TAPS)]3") / e (249 nm) • Figure 2.10. Plot of Change in % Chemical Shift (Left Axis) 59 for H ortho to Hydroxyl Group (H(a) A , H(b) o , H(c) • as in Scheme 2.6) and Plot of Molar Extinction Coefficient (e, M'^cm"1) at 250 nm (Right Axis) both vs. p H (Filled Symbols) for [Ga(TRNS)]3\ Figure 2.11. Speciation Diagram of 2 mM M(III) : 2 mM H 6 TRNS. 61 Charges Omitted for Simplicity. Figure 2.12. 7 1 G a NMR Spectra (91.5 MHz) of [Ga(TAPS)]3" (Top, 64 1000 Transients), [Ga(TACS)]3" (Middle, 1000 Transients), and [Ga(TAMS)]3" (Bottom, 10000 Transients); [Ga] = 30 mM, pD = 9.3,20 °C. Figure 3.1. Top: Experimental Lanthanide Titration Curves at 89 2 mM H 6 TRNS : 1 mM Ln(III). Bottom: Experimental Plots of n vs. log [H3TRNS 3-] (Symbols) and Curves Generated from the Calculated Stability Constants (Solid Lines). Figure 3.2. Experimental Lanthanide Titration Curves at 2 mM 91 H 6 T A M S : 2 mM Ln(III), Top, and 2 mM H 6 TAPS : 2 mM Ln(III), Bottom. Figure 3.3. Experimental (Symbols) Plots of n vs. log [TAMS6"] 92 (Top) and h vs. log [TAPS6"] (Bottom). The Solid Lines are Generated from the Calculated Stability Constants, K L n(XAMS) a n ^ KLn(TAPS)-Figure 3.4. X H NMR Spectrum (200 MHz) of the Benzylic Region 94 for Various H 3 T R N S 3 " : L u 3 + Ratios. vi Figure 3.5. Plot of Gd.I.S vs. pH for 4mM TRNS : 2mM Gd(III) 97 (Top), and the Relevant Speciation Diagram Calculated from the Equilibrium Constants Shown in Table 3.1 (Bottom). Figure 3.6. Plot of Dy.I.S. vs. [Dy(III)] (mM) for D y ( a q ) 3 + , A , 98 [Dy(TAMS)]3", • , [Dy(TAPS)]3", • , and [Dy(H3TRNS)2]3", O . Error Bars Represent Linewidths at Half Height. Figure 3.7. Enthalpies and Entropies for the Ln(III) - H 3 T R N S 3 " 104 Equilibria: Nd(III), • , Gd(ni), • , Ho(III), E3 , Yb(ffl), m . Figure 3.8. Tightening the Hydrophobic Belt: the 106 Ln(III) - H 3 T R N S 3 " Equilibria Viewed in Terms of Hydrophobic Interactions. Figure 3.9. log K n vs 1 / r (Ionic Radii for C N = 6): 107 K : ( • ),K 2( O ). Figure 3.10. Comparative Binding Affinities of TAMS 6", TAPS6", 110 and H 3 T R N S 3 - for Ln(III): log fa ([Ln(TAMS)]3" • ; log fa ([Ln(TAPS)]3" ^ ; log P n ([Ln(H3TRNS)2]3" H ; n = 1 for La(IH), n = 2 for Nd(III), Gd(ni), Ho(III), Yb(III). Figure 3.11. Comparative pM Values vs Z for H 3 T R N S 3 " ([Ln(III)]tot 111 = 1 mM), ([Ln(III)]tot = 1 uM), - 0 - , TAPS6" ([Ln(III)]tot = 1 mM), TAMS 6" ([Ln(III)]tot = 1 mM), - » - . Figure 3.12. Speciation of Yb(III) in the Presence of H 6 T A M S (Top), 112 H 6 TAPS (Middle), and H 6 TRNS (Bottom); [Yb(III)]tot = 1 mM, [ligand] to t = 2 mM. Figure 4.1. ORTEP Representation of the Molecular Structure of 132 the [Co(bped)]+ Cation in [Co(bped)]PF 6 *CH 3 CN»H 2 0 (33% probability thermal ellipsoids). vii Figure 4.2. Titration Curves (pH vs. a; a = mol OH" / mol bped) for 134 H 2 bped»2HCl in the Presence and Absence of Equimolar (2 mM) Al(m), Ga(III), and In(III). Figure 4.3. Variable pH UV and NMR Titrations of H 2bped. Top: 135 Molar Extinction Coefficient at 260 nm, (Right Axis) and Chemical Shift Difference of H(l), # , (Left Axis) vs pH. Bottom: Chemical Shift Difference of H(6), H(7), - 0 - , and H(8), - B - , vs pH. Figure 4.4. % NMR Spectrum (500 MHz) of [In(bped)]Cl in D 2 0 at 136 pD = 3.2. Figure 4.5. HMBC (Top) and HMQC (Bottom) Spectra of 137 [In(bped)]Cl in D 2 0 at pD = 3.2. *H and 1 3 C Labelling as in Chart 4.1. Vertical Axis (125.8 MHz), Horizontal Axis (500.1 MHz). Figure 4.6. *H NMR spectra (300 MHz) of Equimolar (-70 mM) 140 Mixtures of Ga(III) - H 2 bped«2HCl: Bottom, 2 Equivalents of Base (pD = 1.5); Middle, 4 Equivalents of Base (pD = 3.1); Top, 5 Equivalents of Base (pD = 6.6). Figure 4.7. Speciation diagrams (%M(III) vs. pH) Calculated for 144 Equimolar (2 mM) Al(III) - H 2 bped»2HCl (top), Ga(III) -H 2 bped«2HCl (Middle), and In(III) - H 2 bped«2HCl (Bottom). Figure 4.8. Plot of %M(III) Bound to bped as a Function of pH 151 Calculated for a 1 M(III): 1 edta : 1 bped = 2 mM Mixture, for Each of Al(III), Ga(III), and In(III). Figure 5.1. Comparison of neutral donors for Ln(III) chelation 159 after Thompson et al. The amide value is taken from Paul-Roth and Raymond. viii Figure 5.2. Titration Curves (pH vs. a; a = mol OH" / mol bped) 166 for H 2 bped»2HCl ( ), and Equimolar (2 mM) bped -Ln(III)Solutions: La(III) ( ), Nd(III) ( ), Gd(III) ( ), Ho(III)( ),Yb(m)( ). Figure 5.3. Plot of Dysprosium Induced NMR chemical Shift 169 (Dy.I.S.) of H 2 1 7 0 vs. [Dy(UI)] for [Dy(bped)]+ at 20 °C, pH 4.5. Error bars represent linewidths at half height. Figure 5.4. Plot of A ' / C D vs. <S Z>/CD (top) and 171 A'/<SZ> vs. C D / <S Z > (bottom). Figure 5.5. Top: plot of p K a of bound water vs. inverse ionic radius 174 (CN = 9) for [Ln(bped)]+, and [Ln(hedta)], - » - . Bottom: plot of log K J ^ L v s - inverse ionic radius (CN = 9) for [Ln(bped)]+, [Ln(edda)]+, - B - , and [Ln(edta)]", ix List of Tables Table 1.1. Rate Constants for Water Exchange at 25 °C and Ionic 3 Radii (CN 6) for Some Trivalent Metal Ions. Table 1.2. Ionicity in Bonding of Lewis Acids and Lewis Bases. 6 Table 1.3. Formation Constants (log K{) for Some Unidentate 7 Ligands with Trivalent Metal Ions. Table 1.4. Equilibrium Constants (log K = M L * H 2 / H 2 L » M ) for 8 Bidentate Anionic Oxygen Ligand Complexes with Trivalent Metal Ions. Table 1.5. Formation Constants (log Kj) and Formation Constant 10 Differences for Some Tridentate Ligands with Trivalent Metal Ions. Table 2.1. Deprotonation Constants. 40 Table 2.2. Metal - Ligand Formation Constants. 53 Table 2.3. Comparative pM Values Calculated at pH = 7.4 for 66 10 | iM Ligand : 1 | iM Metal. Table 3.1. Thermodynamic Parameters for the L n 3 + - H 3 TRNS 3 " 88 Complexation Equilibria. Table 3.2. Log Formation Constants for Ln(III) with T A M S 6 - and 90 TAPS 6 - at 25 °C, \i = 0.16 M NaCl. Table 3.3. Thermodynamic Values for the First Three 93 Deprotonations of H 6 TRNS. Table 4.1. Selected Bond Lengths (A) and Angles (deg) for the 125 [Co(bped)]+ Cation in [Co(bped)]PF 6 «CH 3 CN«H 2 0. Table 4.2. *H (300 MHz) Spectral Data for H 2bped and its Co(III) 129 and In(III) Complexes. x Table 4.3. 1 3 C (75.5 MHz) Spectral Data for H 2bped and its Co(IH), 130 In(III), and Ga(III) Complexes. Table 4.4. Deprotonation Constants of H 2 bped«2HCl (Charges 133 Omitted for Simplicity) at 25 °C. Table 4.5. Log Formation Constants for Al(III), Ga(III), and In(III) 134 with bped 2' and edta4'. Table 4.6. Binary Metal-Ligand Formation Constants and 142 Calculated pM Values (pH 7.4) for Al(III), Ga(III). and In(III) Complexes of Ethylenediamine Based Hexadentate Ligands (see Chart 4.1 for Ligand Abbreviations). Table 5.1. Log Formation constants for Ln(III) with bped at 166 25 °C, ]i = 0.16 M NaCl. Table 5.2. 2 H (300 MHz) Spectral Data for H2bped and its La(III), 167 Sm(III)/ ¥(111), and Lu(III) Complexes in D 2 0 . Table 5.3. 1 3 C (75.5 MHz) Spectral Data for H 2bped and its La(III), 168 Sm(III), Y(III), and Lu(III) Complexes in D 2 0 . Table 5.4. 1 7 0 Lanthanide Induced NMR Chemical Shifts of D 2 0 172 by [Ln(bped)]+ at pD = 4.6 at 20 °C. Table 1.1. Crystallographic data for [Co(bped)]PF 6 »CH 3 CN«H 2 0. 188 Table 1.2. Bond Lengths (A) for [Co(bped)]PF 6 «CH 3 CN«H 2 0. 188 Table 1.3. Bond Angles (°) for [Co(bped)]PF 6 «CH 3 CN»H 2 0. 189 xi List of Abbreviations Abbreviation Meaning a moles of hydroxide ion per mole of ligand A angstrom, 1 x 10"10 metre APT attached proton test Anal analysis OAc acetate (3n overall formation constant of n stepwise equilibria (3" beta particle B M Bohr magneton H2bbpen N,N'-bis(2-hydroxybenzyl)-N,N'-bis(2-pyridylmethyl)-ethylenediamine H2bped N,N'-bis(2-methylpyridyl)ethylenediamine-N,N'-diacetic acid br broad (spectral) Calcd Calculated °C degrees Celsius cm"1 wavenumber(s) (reciprocal centimeter) C N coordination number COSY correlated spectroscopy (NMR) cyclen 1,4,7,10-tetraazacyclododecane 8 chemical shift in parts per million (ppm) downfield from tetramethylsilane (NMR); vibrational in-plane bending mode (IR) A8 change in chemical shift in parts per million (ppm) D deuterium xii d doublet (NMR) H4dcta frflns-l^-diaminocyclohexane-N/N/N'/N'-tetraacetic acid dd doublet of doublets (NMR) DMF N,N'-dimethylformamide DMSO dimethylsulfoxide DMSO-rfg deuterated dimethylsulfoxide • H4dota l,4/7/10-tetraazacyclododecane-N/N'/N",N"'-tetraacetatic acid H 3 D03A l/4,7/10-tetraazacyclododecane-N/N'/N"-triacetatic acid H 2dpa 2,6-dipicolinic acid DSS sodium 2,2-dimethyl-2-silapentane-5-sulfonate H5dtpa N^^'^'^N'-diethylenetriaminepentaacetic acid H3dtpa-bm N,N"-dimethyl-diethylenetriamine-N/N',N"-triacetic acid H3dtpa-bma diethylenetriamine-N,N"-bis(methylamide)-N,N'/N"-triacetic acid H3dtpa-bp diethylenetriamine-N/N"-bis(2-methylpyridyl)-N,N'/N"-triacetic acid T| absolute hardness e extinction coefficient (UV) in M^cm" 1 H 4 E C N,N'-ethylene-di-L-cysteine H 4 EDDASS N/N'-bis(2-mercaptoethyl)ethylenediamine-N,N'-diacetic acid H4edta N/N/N^N'-ethylenediaminetetraacetic acid H4EF£PG N,N'-ethylenebis[2-(o-hydroxyphenyl)glycine] H 2 E N D A - H P N/N'-bis(3-hydroxy-6-methyl-2-pyridylmethyl)-ethylenediamine-N,N'diacetic acid EPR electron paramagnetic resonance xiii eq equivalent fwt formula weight g gram(s) h hour(s) H 4 HBED N/N'-bis(2-hydroxybenzyl)ethylenediamine-N/N'-diacetic acid H 2hedda N,N'-bis(2-hydroxyethyl)ethylenediamine-N,N'-diacetic acid H2hedta N/-(2-hydroxyethyl)ethylenediamine-N/N'/N'-triacetic acid H2hida (2-hydroxyethyl)iminodiacetic acid HMBC 1H-detected multiple bond heteronuclear multiple quantum coherence H 3 HP-D03A l,4,7,10-tetraazacyclododecane-N-2-hydroxypropyl-N'/N",N"'-triacetatic acid H 4 HPED N/N'-bis(2-hydroxyphenyl)ethylenediamine-N,N'-diacetic acid -1 HMQC H-detected heteronuclear multiple quantum coherence HSAB hard and soft acids and bases Hz hertz H 2 ida iminodiacetic acid IR infrared J coupling constant (NMR); Joule(s) K Kelvin K n stepwise formation constant of the n-th equilibrium X wavelength \ n a x wavelength at peak maximum (UV) xiv L litre(s); ligand LSIMS liquid secondary ion mass spectrometry | i micro- (10"6); ionic strength |i eff effective magnetic moment in BM m milli- (10~3); multiplet (NMR); moderate (IR) M molarity; metal m I z mass-to-charge ratio (in mass spectrometry) MeOH methanol min minute(s) mol mole mp melting point MRI magnetic resonance imaging v stretching vibration (IR) ft moles of bound ligand per mole of metal ion nm nanometer Na3saltachs trisodium cis, cis - l,3,5-tris(((2-hydroxy-5-sulfonato-benzylidene)amino)cyclohexane Na3saltames trisodium l,l,l-tris(((2-hydroxy-5-sulfonato-benzylidene)amino)methyl)ethane Na3saltaps trisodium l,2,3-tris(((2-hydroxy-5-sulfonato-benzylidene)amino) propane Na3saltrens trisodium tris(((2-hydroxy-5-sulfonato-benzylidene) amino)ethyl)amine N M R nuclear magnetic resonance H 3nota l/4/7-triazacyclononane-l,4,7-triacetic acid H 3nta nitrilotriacetic acid H 3nta-ma nitrilotriacetic acid monoamide xv H2oda oxydiacetic acid ORTEP Oak Ridge Thermal Ellipsoid Program pD negative log of the concentration of D3O"1" pH negative log of the concentration of H^O4" H~2pida (2-methylpyridyl)iminodiacetic acid p K a negative log of the acid dissociation constant K a p M negative log of the concentration of a metal aquo ion at a particular pH ppm parts per million (NMR) pro-ligand the organic molecule from which the ligand bound to the metal is derived rt room temperature s singlet (NMR); strong (IR) H4Sbad l,10-bis(2-hydroxy-5-sulfonylbenzyl)-l,4/7/10-tetraazadecane H4Sbbpen N,N'-bis-(2-hydroxy-5-sulfonylbenzyl)-N,N'-bis(2-methylpyridyl)ethylenediamine H 6 SHBED N,N'-bis(2-hydroxy-5-sulfonylbenzyl)ethylenediamine-N,N'diacetic acid t triplet (NMR) T temperature (K) tach cis, cis 1,3,5-trisaminocyclohexane H3TACN-HP l/4,7-tris(3-hydroxy-6-methyl-2-pyridylmethyl)-l,4,7-triazacyclo-nonane H3TACN -TX l,4,7-tris(3/5-dimethyl-2-hydroxybenzyl)-l/4/7-triazacyclo-nonane xvi H 6 T A C S cis, cis - l,3,5-tris(((2-hydroxy-5-sulfobenzyl)amino)-cyclohexane tame 1,1,1 -tr is (aminomethyl) ethane H 6 T A M S l/l/l-tris(((2-hydroxy-5-sulfobenzyl)amino)methyl)ethane tap 1,2,3-triaminopropane H 6 T A P S l,2,3-tris(((2-hydroxy-5-sulfobenzyl)amino)propane Tf transferrin TMS tetramethylsilane tren tris(aminoethyl)amine H 6 T R N S tris(((2-hydroxy-5-sulfobenzyl)amino)ethyl)amine U V ultraviolet vis visible vs very strong (IR) V T variable temperature w weak (IR) W i / 2 peak width at half height xvii Acknowledgments I must first acknowledge Professor Chris Orvig for his enthusiasm, encouragement, and patience. He allowed me the scientific freedom to go off on tangents which usually ended up in some black oleaginous material, but occasionally came to joyful fruition. I have had opportunity for some fun collaborative research, and I thank my colleagues Hamid Hoveyda, Shuang Liu, Tomas Hedlund, Staffan Sjoberg, Ika Setyawati, Ernest Wong, Parisa Mehrkodavandi, and Mark Lowe for our joint scientific endeavours. I especially thank Professor Staffan Sjoberg for allowing me to study in Umea for a time and for impressing upon me the marvels of n curves. Cheers to all the past and present members of Equipe Orvig Team for making my time in Vancouver an enjoyable one. My thanks also to the Fringe chemists, notably Mark, Guy, Marco, Caro, and honoured visiting Fringe chemist, Rich Hooper. The creative genius of Viz is to be acknowledged for Top Tips, Finbar Saunders, and Billy the Fish. The assistance of the departmental support staff is gratefully acknowledged, especially Ms. Liane Darge, Ms. Marietta Austria and Mr. Peter Borda for his patience. NSERC is thanked for the cash. I am grateful to my sisters and my parents for their continuous support, uninterrupted for over twenty-six years. Thanks always to Vera. xviii I dedicate this thesis to E. R. Hayes, a scientist and a humanist. "Indeed there is no end to this madness, and the yahoos never sleep." - Hunter S. Thompson xix Chapter 1: General Introduction The aqueous chemistry of the group 13 elements Al(III), Ga(III)/ and In(III), and that of the lanthanides^, Ln(III), has burgeoned within the last twenty years in part because of the biological significance of these elements. However none of these metals are essential for any known organism. The interest lies in understanding the toxic role of aluminum, and in exploiting the physico-chemical properties of gallium, indium, and the lanthanides for use in medicine and biology. Aluminum constitutes 7.5 % of the Earth's crust by weight,1 yet aluminum is a non-essential element. Williams has suggested that the reasons for this are 1) the lack of redox chemistry of Al(III); 2) the slow ligand exchange reaction rates of Al(III) which hinders its use as a Lewis acid catalyst; 3) the low concentration of soluble Al(III) at neutral pH (10'n M at p H 7.0); and 4) the strong Lewis acidity which results in.high Al(III) - ligand binding constants which in turn would block binding of essential ions such as Mg(II) to relevant substrates. The free A l 3 + concentration in equilibrium with a precipitate will increase as [H + ] 3 . The extensive use of acidifying fossil fuels has decreased soil and water pH levels in many areas, and as a consequence dramatically increased aluminum mobility.3 Abnormally high aluminum levels are linked to dementia dialysis, iron-adequate microcytic anemia, osteomalacia, and possibly Alzheimer's disease.4 As a result of this, there is a need to understand the coordination chemistry of Al(III) with biomolecules to ascertain how and why aluminum manifests its toxicity, and to develop selective aluminum chelators for the treatment of Al(III) overload. t Throughout this thesis, the term lanthanide, abbreviated Ln, shall refer to all the elements from La - Lu (Z = 57 - 71) as well as yttrium. Lanthanum and yttrium are included in the generic term because of their physical and chemical properties which are similar to the rest of the series. 1 The biological chemistry of gallium and indium is dominated by the use of 6 7 G a , 6 8 G a , and 1 1 1 In in diagnostic nuclear medicine.5"7 The design of new radiopharmaceuticals containing an isotope of gallium or indium must satisfy, at first, two criteria: the metal complex must be stable with respect to hydrolysis to the metal hydroxide, and it must be stable with respect to demetallation by the serum protein transferrin. This stability can be thermodynamic, or kinetic if ligand exchange is slow in comparison to the time course of the radiopharmaceutical's biodistribution. The lanthanides, along with lanthanum and yttrium, possess an even more diverse biological/medicinal chemistry than the group 13 metal ions mentioned above. P~ emitting isotopes such as 9 0 Y , 1 5 3 S m , and 1 6 9 Y b are being used in therapeutic nuclear medicine.6 Gadolinium(III) and dysprosium(III) complexes are used as contrast agents in magnetic resonance imaging. ' Dysprosium and thulium complexes of multidentate phosphonate ligands have been used as shift reagents in 2 3 N a NMR allowing for the determination of intra- and extracellular sodium content.1 0'1 1 Linking a luminescent lanthanide such as Eu(III) to a chelator that contains an organic sensitizer yields a strongly luminescent complex; if this complex is then linked to a biomolecule, it can act as a fluoroimmunoassay.12 Because of the similarity in ionic radii 1 3 and donor preferences between calcium and the lanthanides, the latter can isomorphously replace calcium in biological macromolecules.14 Here, the diverse physical properties of the lanthanides can be exploited to probe the role of calcium in biochemistry and biology.1 5 For instance, Eu(III) and Tb(III) have useful luminescent properties 1 6 , 1 7 which enable the determination of the hydration A radiopharmaceutical is defined as a chemical substance containing a radionuclide that is suitable for use in nuclear medicine. 2 number about the lanthanide, the Ln - Ln distance in a macromolecule binding two or more ions, as well as more established techniques of optical spectroscopy such as determining equilibrium constants and rate constants. Lanthanide induced shifts of NMR resonances can be used to probe the calcium binding sites of proteins.18 For most of these medicinal/biological applications, thermodynamic stability of the metal chelate is a unifying theme. The removal of toxic metal ions requires a ligand that can bind the ion selectively and strongly so that the resultant complex can be excreted. Similarly, a metal chelate used in diagnostic or therapeutic nuclear medicine should be stable with respect to metal ion hydrolysis or transmetallation if the localization of the radiopharmaceutical is governed by the metal chelate structure. Thermodynamic stability is a necessary requisite in the development of MRI contrast agents since these are used in high concentrations (gram quantities), and release of the metal ion would lead to toxic side effects. Likewise, the use of lanthanide complexes in fluoroimmunoassays requires that the lanthanide remain bound to the fluorophore. Table 1.1. Rate Constants for Water Exchange at 25 ° C 2 1 and Ionic Radii 1 3 (CN 6) for Some Trivalent Metal Ions. Al(III) Fe(III) Ga(III) In(III) Gd(III) Yb(III) k (s"1) 1.29 1.6 x 102 4.0 x 102 4.0 x 104 8.3 x 108 4.7 x 107 r (CN 6) 0.535 A 0.645 A 0.620 A 0.800 A 0.938 A 0.868 A Complexes which are kinetically inert would also be desirable. A comparison of the water exchange rates (Table 1.1) indicates that the trivalent 3 group 13 and lanthanide ions are very labile (with the exception of Al(III) which has been termed "sluggishly labile")19 A multidentate ligand of high denticity has favourable attributes for the applications described above. The chelate effect should impart an added stability to a multidentate chelate complex relative to a complex composed of similar ligands of lower denticity. The rate of ligand dissociation should be slower for a multidentate ligand since more metal - ligand bonds must be broken compared to the dissociation of a ligand of lower denticity.20 Furthermore, if the metal : ligand stoichiometry is 1 : 1, then the complex will be stable with respect to dilution (cf. an M L 6 species where complex formation has a 1/[L]6 dependence as opposed to a 1/[L] dependence for ML). This thesis attempts an understanding of multidentate (usually hexadentate) ligand complexes of Al(III), Ga(III), In(III), and Ln(III) in aqueous solution. How does the choice of donor atoms in a given ligand affect the stability and selectivity (the difference in stability between ions) in complexes of these ions? What effect does keeping the donor set constant, but varying the ligand framework have on metal complex stability and selectivity? How well do solid state structures correlate with solution structures? How important are ternary complexes, eg. H M L or ML(OH)? These are some of the questions that are addressed in this thesis. There are several chemical similarities among the metal ions treated in this thesis. All are tripositive ions that are considered "hard" in the Pearson2 2 sense or class a according to Ahrland, Chatt, and Davies 2 3 There are group similarities: the group 13 metal ions behave similarly as do the lanthanide ions, and these group similarities form the basis of the Periodic Table and the format of inorganic chemistry texts. 2 4 , 2 5 Within the context of "hard" metal ions, these ions can be further subdivided. Hancock and 4 Marscicano ' have developed a semi-empirical relationship to predict the metal ion stability constants of unidentate ligands, eq. 1. Here, E and C refer to the electrostatic and covalent contributions to the formation constant, respectively, and D represents a steric parameter. The subscripts A and B refer to the Lewis Acid (metal ion) and the Lewis Base (ligand) respectively. The values of D are zero for large cations and small anions, and only become important when correcting for the weak stability of small cation - large anion interactions. Martell and Hancock proposed that the ratio, E / C = I, be used as a measure of hardness. In Table 1.2, there are values of I, E, C, and D listed for a series of Lewis Acids and Lewis Bases of interest in this thesis. Parr and Pearson have defined absolute hardness, n, as where A is the ionization potential and E is the electron affinity of the molecule or ion. Absolute softness is the inverse of equation 2. Values of n are also recorded in Table 1.2. Pearson's n values are based on gas phase properties, and to a first approximation correlate with the aqueous solution behaviour of these ions. A better representation of the solution behaviour of these metal ions is given by the I A parameter (IA = E A / C A ) . Using the I A criterion Al > La > Lu > Fe > Ga > In in terms of increasing ionic character in bonding. l o g K i = E A E B + C A C B + D A D B (1) T | = (A-E)/2 (2) 5 Hoveyda has stated that "In effect, oxygen and nitrogen are the only donor atoms that nature parsimoniously bequeaths to the group 13 coordination chemist keen on the biological relevance of his/her work."32 Table 1.2. Ionicity in Bonding of Lewis Acids and Lewis Bases.a Lewis Acid IA E A C A D A Al(III) 45.77 10.50 6.90 0.657 2.0 Fe(III) 12.08 7.22 6.07 0.841 1.5 Ga(in) 17 7.07 5.72 0.809 1.5 In(lll) 13 6.30 4.49 0.714 0.5 La(HI) 15.39 10.30 3.90 0.379 0 Lu(III) 12.12 10.07 4.57 0.454 0 Lewis Base IB E B c B D B F" — 0 0 1.0 0 0 C H 3 C O O - — 0 0 4.76 0 OH" — 0 0 14.00 0 N H 3 8.2 -0.088 -1.08 12.34 0 C5H5N 5.0 -0.102 -0.74 7.0 0 a r| values taken from ref. 30; I, E, and C values taken from ref. 31 This statement is generally true, although Martell and Welch 3 3" 3 7 have recently done some nice work on thiol-containing ligands with Ga(III) and In(III) which has shown the anionic thiol to be an effective donor, and substituted phosphines may also prove useful as donor types for these ions. 6 This thesis deals only with oxygen and nitrogen donors. In Table 1.3 there are some equilibrium constants listed for the reaction of M(III) + L - M L K T (3) some trivalent metal ions with unidentate ligands in aqueous solution. The only ligand to form a stronger complex with Al(III) (the hardest acid) than the other trivalent ions is fluoride (the hardest base). Hefter3 8 has Table 1.3. Formation Constants (log K{) for Some Unidentate Ligands with Trivalent Metal Ions.a Lewis Acid o p r F" C H 3 C O O " N H 3 b C 5 H 5 N C Al(III) 8.45 6.42 1.51 0.8 -0.5 Fe(III) 11.27 5.17 3.38 3.8 1.4 Ga(III) 10.6 4.47 3.8a 4.1 1.4 In(III) 9.5 3.73 3.5 4.0 1.7 La(HI) 4.67 3.00 1.8 0.2 -0.2 Lu(III) 5.81 3.95 1.84 0.7 -0.2 a Data taken from ref. 39, 25 °C \i = 0.1 M. b Estimated, taken from ref. 31. c Calculated using equation 1 and the data in Table 1.2. shown that log K4 for F" correlates with z 2 / r for a range of cations (z = cation charge, r = ionic radius) which indicates that metal - fluoride bonds can be reasonably modelled in terms of electrostatics. Although the bonding between hydroxide and these tripositive metal ions should be largely ionic in 7 nature, log for hydroxide binding has Al(III) ranked below Fe(III), Ga(III), and In (III). The value of E B in Table 1.2 is zero, suggesting that the differences in the thermodynamics of hydroxide coordination between these hard metal ions can be modelled in terms of a covalent interaction. Crumbliss and Garrison 4 0 reached the same conclusion. They showed that the p K a of bound water for a variety of aquo ions correlated with z%m (z = cation charge, % m = metal ion electronegativity41) and this was interpreted as increasing in M - O covalent bond character due to an increase in metal ion electronegativity which resulted in a greater polarization of the MO - H bonds. The I A values listed in Table 1.2 correlate somewhat with the formation constants for acetate, ammonia, and pyridine. Fe(lll), Ga(III), and In(III) all have a similar affinity for anionic carboxylate oxygen or neutral nitrogen pyridyl and amine donors, while Al(III) behaves more like the lanthanides and has much less affinity for these donors. Based on the values in Table 1.3, chelating anionic oxygen ligands should form stable complexes with the trivalent metal ions listed above. This is indeed the case and now the stability order is Ga > Al > In (Table 1.4). Tiron and chromotropic acid differ in that upon complexation, tiron forms a complex with a 5-membered chelate ring, while chromotropic acid forms a complex with a 6-membered ring. The larger ring size results in a decrease in stability, but this decrease is least for the smallest metal ion, Al(III), and this has consequences for altering the selectivity of a ligand for a given metal ion. "03S tiron chromotropic acid (ca) 8 T a b l e 1.4. Equilibrium Constants (log K = M L * H 2 / H 2 L « M ) for Bidentate Anionic Oxygen Ligand Complexes with Trivalent Metal Ions. tiron ca A l o g K Al(III) -3.2 -3.81 0.61 Ga(HI) -0.93 -2.73 1.80 In(III) -3.74 -4.89 1.15 In order to form a hexadentate ligand, it is usually necessary to include neutral donors. This enables the maximum number of 5- and 6-membered chelate rings to be formed upon complexation (cf. edta), and also lowers the overall charge on the metal complex, which is important since highly negative charged complexes tend to have rapid renal excretion. The values for pyridine and ammonia are estimates since no aqueous complexes exist for the M(III) ions listed with ammonia or pyridine as ligands. However incorporation of these neutral nitrogen donors into a chelating ligand can force the nitrogen atoms to coordinate. Shown below are oxydiacetate (oda), iminodiacetate (ida), and 2,6-dipicolinate (dpa). The stability constants of these ligands with some M(III) ions are listed in Table 1.5 and are represented by the equilibrium in equation 4: M 3 + + L 2 " •» ML 9 Kj (4) M L + + 2 H + Ki (5) but the actual equilibrium process is better represented by equation 5. It is easy to compare dpa and oda since both of these ligands have comparable basicity (pK a l + p K a 2 = 6.76, 6.75, respectively). However ida is much more basic ((pK a l + p K a 2 = 11.97). To account for this difference in basicity, Table 1.5 also contains the difference in stability between two ligands, e.g. ida - oda = log K^* (ida) -log Kj* (oda) (6) The pyridyl donor is a better ligating group than an ether in all cases. It is also a better donor than a secondary amine for the lanthanides at any pH, and for Al(III) and Fe(III) at a pH where the secondary amine of ida is protonated. > Table 1.5. Formation Constants (log K{) and Formation Constant Differences for Some Tridentate Ligands with Trivalent Metal Ions.3 Lewis Acid odab ida b dpab ida-odac dpa-odac dpa-idac Al(III) 3.16 8.10 4.87 -0.27 1.71 1.99 Fe(III) 5.04 11.13 10.91 0.88 6.09 5.00 Ga(in) — 12.76 — — — — In(III) 5.0 10.14 — -0.07 — — La(DI) 5.79 5.88 7.94 -5.12 2.15 7.28 Lu(lll) 6.27 7.42 8.83 -4.06 2.56 6.63 M ^ + + H 2 L a Data taken from ref. 39, 25 °C |i = 0.1 M . b Equilibrium constant for eq. 4. c Equilibrium constant difference for the equilibrium in eq. 5. 10 The preceding discussion shows some similarities between Al(III) and Ln(III) in terms of their tendencies to ionic bonding. The major differences between Al(III) and Ln(III) lie in the large size difference between the ions, the much slower kinetics of Al(III) ligand substitution reactions, and the greater tendency of Al(III) to hydrolyze. Crumbliss and Garrison 4 0 have reviewed the similarities in the aqueous chemistry of Al(III) and Fe(III). A closer relationship exists between high spin Fe(III) and Ga(III): they have the same charge, similar ionic radii (0.645 A and 0.620 A, respectively), are unperturbed by ligand field effects, have similar ligand substitution kinetics (Table 1.1), and similar affinity for a variety of ligands (Tables 1.2, 1.3, 1.5). As a result, Raymond and coworkers 4 2 , 4 3 have often used Ga(III) as a diamagnetic NMR probe to study the chemistry of biologically relevant Fe(III) systems, while Green and Welch have drawn upon the much more developed field of Fe(III) coordination chemistry in the design of new gallium(III) radiopharmaceuticals. The similarities among the metal ions in a group can be found in texts 2 4 , 2 5 , 4 4 Much of this thesis deals with metal - ligand formation constants, and these can often be misleading since the convention is to define the equilibrium in terms of the deprotonated ligand. However the actual metal -ligand complex will have a pH dependent speciation, since, thermodynamically, there is a second competing reaction of ligand protonation. Differing ligands have differing basicity and it is often difficult to directly compare formation constants. For example the naturally occurring iron binding ligands (siderophores) enterobactin and ferrioxamine E both bind strongly to Fe(III). For enterobactin, a tris(catecholate) ligand, the log stability constant is 49; for the tris(hydroxamate) ferrioxamine E, log K = 32.5. 11 . The 17 orders of magnitude difference in stability would only be relevant if life existed at pH 14, since catechols are much more basic than hydroxamates. A better comparison is to use pM (-log [Mfr e e]) values under a given set of conditions. At pH 7.4, at a ligand to metal ratio of 10 fiM : 1 | iM, pM = 32.5 for enterobactin and pM = 27.7 for ferrioxamine E; the difference in stability has dropped to 5 orders of magnitude and will be even smaller with decreasing pH. Another example is the effect of substituting an acetate arm on edta 3 9 with a 2-oxybenzyl group to give HBET, 4 5 and finally H B E D . 4 6 The stability constants for Fe(lll), Ga(lll), and In(lll) all increase as the acetate groups are 12 substituted, but the ligand basicity also increases. An effective stability constant, Keff, can be defined as the product of the thermodynamic stability constant, K, and the mole fraction, an, of the deprotonated pro-ligand. Thus K eff increases with pH to a maximum value K. In Figure 1.1, Keff is plotted versus p H for edta, HBET, and HBED with Fe(III), Ga(III), and In(III). Substitution of the 2-oxybenzyl group for an acetate group augments the stability of the Fe(III) and Ga(III) complexes but diminishes that for In(III). This thesis examines the aqueous coordination chemistry of two classes of pro-ligand with the trivalent lanthanide and group 13 metal ions. In Chapter 2, the tripod-like aminophenolate pro-ligands shown below are prepared and their formation constants and solution structures with Al(III), Ga(III), and In(III) are studied. In Chapter 3 the coordination chemistry of the lanthanides with these pro-ligands (with the exception of TACS6") is examined. Chapters 4 and 5 constitute solution and solid state studies on the pro-ligand bped in an effort to measure the affinity of the pyridyl moiety for these metal ions. Chapter 6 summarizes the work and looks to where this research might lead. 13 Pro-ligands H2bped 14 15 References 1) Williams, R. J. P.; Frausto da Silva, J. J. R. The Natural Selection of the Chemical Elements; Clarendon Press: Oxford, 1996. 2) Williams, R. J. P. Coord. Chern. Rev. 1996 ,149 , 1. 3) Sposito, G. E. The Environmental Chemistry of Aluminum; CRC: Boca Raton, FL, 1989. 4) Corain, B.; Bombi, G. G.; Tapparo, A.; Perazzolo, M.; Zatta, P. Coord. Chern. Rev. 1996, 149, 11. 5) Weiner, R. E.; Thakur, M. L. Radiochimica Acta 1995, 70, 273. 6) Jurisson, S.; Berning, D.; Jia, W.; Ma, D. Chern. Rev. 1993, 93, 1137. 7) Green, M . A.; Welch, M. J. Nucl. Med. Biol. 1989,16, 435. 8) Lauffer, R. B. Chern. Rev. 1987, 87, 901. 9) Watson, A. D. / . Alloys Compds. 1994, 207/208, 14. 10) Springer, C. S. Ann. Rev. Biophys. Biophys. Chern. 1987 ,16, 375. 11) Sherry, A. D.; Geraldes, C. F. G. C ; Cacheris, W. P. Inorg. Chim. Acta 1987,139,137. 12) Li, M . ; Selvin, P. R. / . Am. Chern. Soc. 1995, 117, 8132. 13) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. 14) Martin, R. B.; Richardson, F. S. Q. Rev. Biophys., 1979,12, 181. 15) Lanthanide Probes in Life, Chemical, and Earth Sciences; Biinzli, J.-C. G.; Choppin, G. R., Eds.; Elsevier: Amsterdam, 1989, . 16) Biinzli, J.-C. G. Inorg. Chim. Acta 1987, 239, 219. 17) Horrocks, W. D. J.; Albin, M. Prog. Inorg. Chern., 1984, 31, 1. 18) Geraldes, C. F. G. C. Methods Enzymol. 1993, 227, 43. 19) Orvig, C.; Berthon, G. Handbook of Metal-Ligand Interactions in Biological Fluids: Bioinorganic Chemistry ;Ber\ihon, G., Ed.; Marcel Dekkar: New York, 1995; Vol. 2, 1266. 16 20) Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes; 2nd ed.; VCH: Weinheim, 1991. 21) Lincoln, S. F.; Merbach, A. E. Adv. Inorg. Chern. 1995, 42, 1. 22) Pearson, R. G. /. Am. Chern. Soc. 1963, 85, 3533. 23) Ahrland, S.; Chatt, J.; Davies, N. R. Q. Rev., Chern. Soc. 1958,12, 265. 24) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; John Wiley & Sons: New York, 1988. 25) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements; Pergamon: Oxford, 1984. 26) Hancock, R. D.; Marsicano, F. Inorg. Chern. 1978,17, 560. 27) Hancock, R. D.; Marsicano, F. Inorg. Chern. 1980,19, 2709. 28) Hancock, R. D.; Martell, A. E. Chern. Rev. 1989, 89, 1875. 29) Parr, R. G.; Pearson, R. G. /. Am. Chern. Soc. 1983, 205, 7512. 30) Pearson, R. G. Inorg. Chern. 1988, 27, 734. 31) Martell, A. E.; Hancock, R. D. Metal Complexes in Aqueous Solution; Plenum: New York, 1996. 32) Hoveyda, H. R., Ph.D., University of British Columbia, 1993. 33) Sun, Y.; Anderson, C. J.; Pajeau, T. S.; Reichert, D. E.; Hancock, R. D.; Motekaitis, R. J.; Martell, A. E.; Welch, M. J. /. Med. Chern. 1996, 39, 458. 34) Li, Y.; Martell, A. E.; Hancock, R. D.; Riebenspies, J. H.; Anderson, C. J.; Welch, M . J. Inorg. Chern. 1996, 35, 404. 35) Ma, R.; Welch, M. J.; Reibenspies, J.; Martell, A. E. Inorg. Chim. Acta 1995,236, 75. 36) Sun, Y.; Motekaitis, R. J.; Martell, A. E.; Welch, M . J. Inorg. Chim. Acta 1995,228, 77. 37) Anderson, C. J.; John, C. S.; Li, Y. J.; Hancock, R. D.; McCarthy, T. J.; Martell, A. E.; Welch, M. J. Nucl. Med. Biol. 1995, 22, 165. 17 38) Hefter, G. Coord. Chem. Rev. 1974,12, 221. 39) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum: New-York:, 1974-1989; Vol. 1-6. 40) Crumbliss, A. L.; Garrison, J. M . Comments Inorg. Chem. 1988, 8, 1. 41) Allred, A. L. /. Inorg. Nucl. Chem. 1961,17, 215. 42) Kersting, B.; Telford, J. R.; Meyer, M.; Raymond, K. N. /. Am. Chem. Soc. 1996,118,5712. 43) Borgias, B. A.; Barclay, S. J.; Raymond, K. N. /. Coord. Chem. 1986,15, 109. 44) Comprehensive Coordination Chemistry; Wilkinson, G., Gillard, R.D., McCleverty, J.A., Eds.; Pergamon: Oxford, 1987; Vol. 3,. 45) Ma, R.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1995, 233, 137. 46) Ma, R.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1994, 224, 151. 18 Chapter 2: Tripodal Aminophenolate Ligand Complexes of Al(III), Ga(III), and In(III) in Water 2.1 Introduction In the quest for new chelators of gallium and indium with the intent of developing radiopharmaceuticals based on 6 7 G a , 6 8 G a , or m I n , it is germane to employ a multidentate ligand. A hexadentate ligand offers two advantages: a higher thermodynamic stability than its bi- or tridentate analogues, and a metal complex stability indifferent to a dilution effect, important considering the low concentrations (nM) involved in nuclear medicine. The gallium or indium complex should be stable with respect both to hydrolysis and to demetallation by the serum protein transferrin. As the radionuclides involved are short lived, rapid complexation kinetics when labelling are desirable, while demetallation kinetics should be slow. Another useful characteristic is the ease of modification of the ligand backbone in order to change the lipophilicity of the complex (and possibly its biodistribution), or to tag the complex to a biologically important molecule (antibody, protein, etc.) via a bifunctional linkage. Harris and coworkers have determined binding constants for a variety of metal ions with the human serum protein transferrin.1-3 For tripositive metal ions such as gallium and indium, transferrin is the dominant endogenous chelating agent in the blood. A comparison of binding constants between the metal chelate proposed for radiopharmaceutical trial and the transferrin metal complex would offer a rational approach to radiopharmaceutical screening, and this approach has been employed successfully by Martell, Welch, and coworkers.4,5. 19 The same principles apply to Al(III) sequestration. Aluminum, a nonessential element, can cause neurotoxicity, dialysis encephalopathy, and has been shown to accumulate in some of the hallmark lesions of Alzheimer's Disease.6 The only approved chelator for aluminum(III) overload is the tris(hydroxamate) siderophore desferrioxamine, an donor. It has been recognized that the most effective ligands for Al(III) chelation are O-donor ligands, and that the stability of Al(III) complexes generally decrease Q via the replacement of O-donor groups with N-donor groups . Consequently, the aqueous coordination chemistry of Al(III) with 0,N-donor ligands is not nearly so well established as is that of Ga(III) and In(III), both of which have higher affinities for the neutral nitrogen donor.9 The aminophenolates reported in this study, while unlikely to present themselves as efficient Al(III) chelators for clinical application, will serve to further the understanding of aqueous Al(III) coordination chemistry by addressing a number of factors. Scheme 2 . 1 There are a variety of topologies available for the design of a multidentate ligand: macrpcycles, macrobicycles, macrocycles with pendant 20 donors, linear, branched, and tripod ligands10 (Scheme 2.1). Because of its cyclic nature, a macrocycle-derived ligand generally offers more selectivity for a metal ion of appropriate ionic radius. Macrocycle based ligands also offer 1 1 high thermodynamic stability (macrocycle effect); while the metallation kinetics are often slower than for a non-macrocyclic analogue, the demetallation kinetics are even slower. In contrast, flexible linear and branched ligands often bind rapidly, but consequences of their flexibility is indiscriminate metal binding, and more rapid demetallation. With the tripodal ligand, a combination of the favourable properties of both acyclic and cyclic ligands may be realized. This approach has been taken with Schiff bases derived from various tripodal amines. 1 3 - 1 5 There is a wide array of substituted salicylaldehydes available with the potential to alter the biodistribution of the complexes. Furthermore, the Schiff base condensation reaction is rapid, clean, and gives high yields. Although metallated Schiff base complexes generally have low aqueous solubility, Evans and Jukubovic have studied complexes derived from the condensation of sodium 5-sulfosalicylaldehyde with various amines and determined pM values as a measure of thermodynamic stability. 1 6 , 1 7 One drawback to Schiff bases is that in water they undergo partial hydrolysis at the imine linkage, and this may be the source of in vivo degradation of their metal complexes. Reduction of the imine eliminates the hydrolysis problem and results in a more flexible ligand. In the case of the Schiff base saltren (the Fe complex of which is depicted schematically in Scheme 2.2), the reduced form of the ligand resulted in three different coordination geometries with Al(III), Ga(III), and In(III)18 (shown schematically in Scheme 2.2); each geometry differed from the analogous saltren complexes.19 This diversity was a consequence of employing a ligand with seven potential 21 donors. Ligands with potential N3O3 donor sets (e.g. nonsulfonated analogs of H 6 T A M S and H 6 TAPS in Scheme 2.3) give the expected octahedral group 13 metal complexes.20'21 Previous work in this laboratory has involved the synthesis and characterization of amine phenols derived from the amines 1 D T I tren (tris(aminoethyl)amine), tame (tris(aminomethyl)ethane), and tap 9 0 (trisaminopropane), and their aluminum, gallium, and indium 1 ft 9 1 9 0 complexes. ' ' Roundhill and coworkers have adopted the same approach using the amine cis, cis - 1,3,5-trisaminocyclohexane, and ubiquitously reported the Al(III), Ga(III), In(III), and Fe(III) complexes 2 2 - 2 4 This chapter explores the effect of backbone variations on the selectivity of multidentate aminophenolate ligands among the trivalent metal ions Al(III), Ga(III), and In(III) in water. The study involves the preparation of the 5-sulfonic acid derivatives of the amine phenols (to ensure aqueous solubility) shown in Scheme 2.3, the in situ characterization of the Al(III), Ga(III), and In(III) complexes, the determination of ligand Scheme 2.2 22 deprotonation constants, and the determination of metal complex formation constants. For X = SCV, H 6 TAMS For X = SO3-, H 6 TRNS For X = S0 3 - , H 6 TACS For X = SO3, H 6 TAPS Scheme 2.3 2.2 Experimental Materials. Sodium borohydride, tris(2-aminoethyl)amine (tren), salicylaldehyde, aniline, tris(hydroxymethyl)ethane, sodium azide, lithium aluminum hydride, diethylene glycol, glycerin, and cis, cis 1,3,5-cyclohexane triol, sodium deuteroxide (NaOD, 40%), deuterium chloride (DO, 12M) and the aluminum, gallium, and indium atomic absorption standards were obtained from Aldrich. Hydrated aluminum, gallium, and indium nitrates were obtained from Alfa. Deuterium oxide (D 20) was purchased from Isotec Anhydrous sodium carbonate was obtained from BDH. All were used 23 without further purification. Sodium 5-sulfosalicylaldehyde, tris(aminomethyl)ethane (tame),21 trisaminopropane (tap),26 and cis, cis 1,3,5-trisaminocyclohexane (tach) were prepared according to literature preparations. Instrumentation. XW NMR (200 and 300 MHz) and 1 3 C NMR (50.3 and 75.5 MHz) spectra were referenced to TMS and recorded on Bruker AC-200E and Varian XL 300 spectrometers, respectively. 2 7 A1, 7 1 G a , 1 1 5 In NMR spectra were recorded on the Varian XL 300 spectrometer at 78.2, 91.5, and 65.7 MHz, respectively, and were referenced to the appropriate 0.1 M metal nitrate in 6 M nitric acid. Mass spectra were obtained with a Kratos Concept II H32Q (Cs+, LSIMS) instrument. UV-vis spectra were recorded on a Shimadzu UV-2100 spectrophotometer. Infrared spectra were obtained as KBr disks in the range 4000 - 400 cm"1 using a Galaxy Series FTIR 5000 spectrophotometer, and were referenced to polystyrene. Analyses of C, H , and N were performed by Mr. Peter Borda in this department. Melting points were measured on a Mel-Temp apparatus and are uncorrected. Trisodium tris(((2-hydroxy-5-sulfonato-benzylidene) amino)ethyl)amine (Na3saltrens). To a suspension of sodium sulfosalicylaldehyde (6.72 g, 30 mmol) in methanol (150 mL) was added tren (1.46 g, 10 mmol). This suspension immediately turned from a pale yellow to a brilliant canary yellow. The mixture was allowed to stir for an hour, after which time it was cooled to 4 °C in the fridge. The fine bright yellow powder was collected on a Biichner funnel and washed with ethanol (10 mL), followed by ether (10 mL). Yield: 5.61 g (83%). X H NMR (300 MHz, DMSO -d6): 2.88 (s, 6H, ethylenic CH 2 ) , 3.66 (s, 6H, ethylenic CH 2 ) , 6.78 (d, 3H, ring H(3), 3 J = 9.3 Hz), 7.55 (dd, 3H, ring H(4), 3 J = 9.3 Hz, 4 J = 2.3 Hz), 7.69 (d, 3H, ring H(6), 4 J = 2.3 Hz), 8.53 (s, 3H, imine CH). 1 3 C NMR (50.3 MHz, DMSO -.24 d6): 54.87,56.32,116.41,116.99,129.15,130.22,138.17,162.42,165.55. Mass spectrum (LSIMS): m/z = 743 ([M - Na + 2H] +, [ C 2 7 H 2 9 N 4 N a 2 0 1 2 S 3 ] + ) , 765 ([M + H] + , [C27H28N4Na301 2S3]+), 787 ([M + Naf , [C 2 7H 2 7 N 4 Na 4 0 1 2 S3] + ) . IR (cm"1, KBr disk): 3600-3400 (b s, v 0 _ H ) , 1642 (s, v c = N ) , 1518 (s, v c= c), 1186 (vs, v s = 0 ) , 1107,1035 (s, v c_ c). Tris(((2-hydroxy-5-sulfobenzyl)amino)ethyl)amine monohydrate (H 6 TRNS»H 2 0). Sodium borohydride (1.6 g, 4 mmol) was added to a suspension of saltrens (7.65 g, 1.0 mmol) in methanol over a period of 30 min. After the addition was complete, the solvent was removed to give a white solid. Methanol (60 mL) was added and removed under reduced pressure. This was repeated twice more. The resultant solid was dissolved in a minimal amount of water (15 mL) and the pH adjusted to about 2 with concentrated HC1; a white solid precipitated and was collected on a Buchner funnel. It was purified by redissolving it in a mildly basic solution (pH~8), filtering, and acidifying the filtrate to give a white precipitate as the zwitterion. This was filtered and dried overnight in vacuo at 95 °C to yield 5.0 g (71%); mp > 250 ° C Anal. Calc. (found) for C 2 7 H 3 6 N 4 0 1 2 S 3 « H 2 0 : C, 44.87 (44.86); H , 5.30 (5.46); N, 7.75 (7.59). Potentiometric analysis was consistent with this formulation. X H NMR (200 MHz, D 2 0 , pD=7.1): 2.80 (t, 6H, ethylenic C H 2 , 3 J = 6 Hz), 3.08 (t, 6H, ethylenic C H 2 , 3 J = 6 Hz), 4.18 (s, 6H, benzylic CH 2 ) , 6.70 (d, 3H, ring H(3), 3 J = 9 Hz), 7.61 (dd, 3H, ring H(4), 3 J = 9 Hz, 4 J = 1 Hz), 7.63 (d, 3H, ring H(6), 4 J = 1 Hz). 1 3 C NMR (75.48 MHz, D 2 0 , pD = 8.0): 45.99,51.81,52.06,121.54,122.47,130.67,130.81,130.89,170.42. Mass spectrum (LSIMS): m/z = 705 ([M+H]+, [ C 2 7 H 3 6 N 4 0 1 2 S 3 ] + ) . IR (cm"1, KBr disk): 3500-2600 (b s, v N . H / 0-H)> 1730 (m, 5N_H), 1604 (s, 5N_H), 1502,1442,1375 (s, v c = c ) , 1200 (vs, v s = 0 ) . UV ( A ^ nm, (e M ' W 1 ) : pH=2: 276 (4800), 232 25 (26000), 203 (72000); pH=12: 290 (11000), 256 (47000). It is slightly soluble in acid and soluble in neutral to basic pH. Trisodium l,l,l-tris(((2-hydroxy-5-sulfonato-benzylidene)amino) methyDethane (Na3saltames). To a suspension of sodium 5-sulfosalicylaldehyde (6.72 g, 30 mmol) in methanol (150 mL) at room temperature was added tame (1.10 g, 10 mmol) in 10 mL of methanol. The pale yellow suspension immediately brightened to a lemon yellow hue. The suspension was stirred for an hour, cooled to 4 °C in the fridge while a lemon coloured powder deposited. The powder was collected on a Biichner funnel, washed with ethanol (10 mL), followed by ether (10 mL), and dried overnight in vacuo. Yield: 6.00 g (82%). 1 H NMR (300 MHz, DMSO - dg): 1.04 (s, 3H, methyl CH 3 ) , 3.62 (s, 6H, methylenic CH 2 ) , 6.86 (d, 3H, ring H(3), 3 J = 9.3 Hz), 7.58 (dd, 3H, ring H(4), 3 J = 9.3 Hz, 4 J = 2.3 Hz), 7.73 (d, 3H, ring H(6), 4 J = 2.3 Hz), 8.66 (s, 3H, imine CH). 1 3 C NMR (75.48 MHz, DMSO - d6): 20.03, 63.78, 115.76,117.23,129.03,130.03,139.08,160.96,167.25. Mass spectrum (LSIMS): m/z = 670 ([M - 3Na + 4H] +, [ C 2 6 H 2 8 N 3 0 1 2 S 3 ] + ) , 692 ([M - 2Na + 3H] +, [C 2 6 H 2 7 N 3 Na0 1 2 S 3 ] + ) , 714 ([M - Na + 2H] +, [ C 2 6 H 2 6 N 3 N a 2 0 1 2 S 3 ] + ) , 736 ([M + H] + , [ C 2 6 H 2 5 N 3 N a 3 0 1 2 S 3 ] + ) , 758 ([M + Na] +, [ C 2 6 H 2 4 N 3 N a 4 0 1 2 S 3 ] + ) . IR (cm"1, KBr disk): 3450 (b s, v 0 - H ) , 1639 (s, v c = N ) , 1484 (m, v c = c ) , 1191 (vs, v s = 0 ) , 1111, 1039 (s, v c . c ) . l,l,l-tris(((2-hydroxy-5-sulfobenzyl)amino)methyl)ethane dihemihydrate ( H 6 T A M S«2 .5H 2 0) . Methanol (60 mL) was cooled to 0 °C in an ice bath and sodium borohydride (0.11 g, 2.9 mmol) was added with stirring. To this was added Na3saltames (0.54 g, 0.74 mmol) in small portions (~ 50 mg each). After each portion was added, the reaction was allowed to proceed until the solid Schiff base had all dissolved and the bright yellow colour dissipated. After all the Na3saltames had been added (~ 1 hr), the 26 solution was allowed to stir for an hour. The solution was then acidified with 6 M HC1, which gave a white precipitate. The methanol was then boiled down to 20 mL, another 100 mL methanol was added, and the methanol again boiled down to 20 mL. The white suspension was allowed to cool to room temperature, and the white solid was collected on a medium frit. The white solid was then dissolved in water, and this solution was passed down a Rexyn 101 cation exchange column ( H + form) and eluted with deionized distilled water until the eluant was no longer acidic. Solvent was removed from the eluant under reduced pressure at 65 °C to give a white solid. The solid was dried at 95 °C in vacuo overnight to yield 0.35 g (71%); mp > 250 °C. Anal. Calc. (found) for C26H33N30 1 2S3«2.5H20: C, 43.33 (43.23); H , 5.31 (5.21); N, 7.83 (7.68). Potentiometric analysis was consistent with this formulation. : H NMR (200 MHz, D 2 0 , pD=12): 2.80 (t, 6H, ethylenic C H 2 , 3 J = 6 Hz), 3.08 (t, 6H, ethylenic C H 2 , 3 J = 6 Hz), 4.18 (s, 6H, benzylic CH 2 ) , 6.70 (d, 3H, ring H(3), 3 J = 9 Hz), 7.61 (dd, 3H, ring H(4), 3 J = 9 Hz, 4 J = 1 Hz), 7.63 (d, 3H, ring H(6), 4 J = 1 Hz). 1 3 C NMR (75.48 MHz, D 2 0 , pD = 12): 22.55,40.02,52.78,57.04,121.13, 128.75,129.01,129.63,130.38. 171.73. Mass spectrum (LSIMS): m/z = 676 ([M+H]+, [C 2 6 H 3 4 N 3 0 1 2S3] + ). IR (cm"1, KBr disk): 3500-2600 (b s, vN_H, O-H)> 1604 (s, 8N_H), 1501,1433 (s, v c = c ) , 1210 (vs, v s = G ) , 1100, 1033 (s, v c . c ) . UV (Vax nm, (e M ' W 1 ) : pH=2: 276 (4420), 232 (31000); pH=12: 288 (10400), 256 (52000). It is soluble in water and hot DMSO. Trisodium l,2,3-tris(((2-hydroxy-5-sulfonato-benzylidene)amino) propane (Na3saltaps). To a suspension of sodium 5-sulfosalicylaldehyde (6.72 g, 30 mmol) in methanol (150 mL) at room temperature was added tap (0.90 g, 10 mmol) in 10 mL of methanol. The pale yellow suspension immediately became bright yellow. The suspension was stirred for an hour, then cooled to 4 °C in the fridge while a bright yellow coloured powder deposited. The 27 powder was collected on a Biichner funnel, washed with ethanol (10 mL), followed by ether (10 mL), and dried overnight in vacuo. Yield: 5.80 g (82%). X H NMR (200 MHz, DMSO - d6): 3.38 (m, 1H, methine CH), 3.98 (s, 4H, methylenic CH 2 ) , 6.82 (d, 2H, ring H(3), 3 J = 9.67 Hz), 6.83 (d, 1H, ring H(3'), 3 J = 9.33 Hz), 7.55 (dd, 2H, ring H(4), 3 J = 9.67 Hz, 4 J = 2.0 Hz), 7.56 (dd, 1H, ring H(4'), 3 J = 9.33 Hz, 4 J = 2.0 Hz), 7.71 (d, 3H, ring H(6) and ring H(6'), 4 J = 2.0 Hz), 8.68 (s, 2H, imine CH), 8.70 (s, 1H, imine C H center arm). 1 3 C NMR (75.5 MHz, DMSO - d6): 60.94, 68.57, 115.77,117.03,117.09, 128.90, 128.94,129.97, 130.04,138.86,138.94, 160.65,160.86,166.63, 167.38. Mass spectrum (LSIMS): m/z = 664 ([M - 2Na + 3H] +, [ C ^ H ^ ^ N a O ^ D , 686 ([M - Na + 2H] +, [C 24H 2 2N3Na 20 1 2S3] +), 708 ([M + H] + , [C 2 6 H 2 1 N 3 Na 3 0 1 2 S3] + ) , 730 ([M + Na] +, [C26 H 2oN 3 Na 4 Oi 2 S 3 ] + ) . IR (cm -1, KBr disk): 3456 (b s, v 0 _ H ), 1638 (s, v c = N ) , 1514,1492 (m, v c = c ) , 1190 (vs, v s = 0 ) , 1108, 1036 (s, v c . c ) . l,2,3-tris(((2-hydroxy-5-sulfobenzyl)amino)propane dihemihydrate ( H 6 T A P S » 2 . 5 H 2 0 ) . Methanol (100 mL) was cooled to 0 °C in an ice bath and sodium borohydride (0.45 g, 12.0 mmol) was added with stirring. To this was added Na3saltaps (2.1 g, 3.0 mmol) in small portions (~ 10 mg). After each portion was added, the reaction was allowed to proceed until the solid Schiff base had all dissolved and the bright yellow colour dissipated. After all the Na3saltaps had been added (~ 1 hr), the solution was allowed to stir for an hour. The solution was then acidified with 6 M HC1; this gave a white precipitate. The methanol was then boiled down to 20 mL, another 100 mL methanol was added, and the methanol again boiled down to 20 mL. The white suspension was allowed to cool to room temperature, and the white solid was collected on a medium frit. The white solid was then dissolved in water, and this solution was passed down a Rexyn 101 cation exchange column ( H + form) and eluted with deionized distilled water until the eluant 28 was no longer acidic. The solvent was removed from the eluant under reduced pressure at 65 °C to give a white solid. The solid was dried at 95 °C in vacuo overnight to yield 1.00 g (49%); mp > 250 °C. Anal. Calc. (found) for C 2 4 H 3 9 N 3 0 1 2 S 3 » 2 . 5 H 2 0 : C, 41.61 (41.72); H, 4.95 (4.70); N, 6.07 (6.05). Potentiometric analysis was consistent with this formulation. 1 H NMR (200 MHz, D 2 0 , pD=12): 2.60 (d, 4H, methylenic C H 2 , 3 J = 5.7 Hz), 2.91 (q, IH, methine C H , 3 J = 5.7 Hz), 3.63 (s, 6H, benzylic CH 2 ) , 6.58 (d, 3H, ring H(3), 3 J = 9.3 Hz), 7.42 (dd, 3H, ring H(4), 3 J = 9.3 Hz, 4 J = 2.7 Hz), 7.47 (d, 3H, ring H(6), 4 J = 2.7 Hz). 1 3 C NMR (75.48 MHz, D 2 0 , pD = 12): 49.70,51.99,52.95,59.05,121.17, 121.30,128.79,128.87,129.07,129.20,129.61,129.94,130.46,130.51,171.57,171.70. Mass spectrum (LSIMS): m/z = 648 ([M+H]+, [ C 2 6 H 3 4 N 3 0 1 2 S 3 ] + ) . IR (cm 4 , KBr disk): 3500-2600 (b s, v N _ H / Q -H)/ 1 6 1 1 (S/ oN_H), 1493,1439 (s, v c = c ) , 1210, 1170 (vs, v s = 0 ) , HOO, 1033 (s, v c . c ) . UV ( ^ a x nm, (e M^cm"1): pH=2: 277 (4400), 232 (28900); pH=12: 288 (11800), 256 (47000). It is soluble in water and hot DMSO. Trisodium cis, cis - l,3/5-tris(((2-hydroxy-5-sulfonato-benzylidene) amino)cyclohexane (Na3saltachs). To a suspension of sodium 5-sulfosalicylaldehyde (2.13 g, 9.5 mmol) in methanol (50 mL) at room temperature was added tach (0.41 g, 10 mmol) in 10 mL of methanol. The pale yellow suspension immediately brightened to a lemon yellow hue and dissolved with stirring. The solution was stirred for an hour, 50 mL absolute ethanol was added, and the suspension cooled to 4 °C in the fridge while a lemon coloured powder deposited. The powder was collected on a Biichner funnel, washed with ethanol (10 mL), followed by ether (10 mL), and dried overnight in vacuo. Yield: 1.90 g (80%). *H NMR (200 MHz, DMSO - de, see Scheme 2.5): 1.81 (q, 3H, H(Aax), 2 J A B = 12 Hz, 3 J A C = 12 Hz), 2.06 (d, 3H, H(Be(?), 2 J A B = 12 Hz, 3 J B C = 2.5 Hz), 3.72 (t, 3H, H(ce(?), 3 J A C = 12 Hz, 3 J B C = 2.5 29 Hz), 6.83 (d, 3H, ring H(3), 3 J = 9.0 Hz), 7.57 (d, 3H, ring H(4), 3 J = 9.0 Hz), 7.74 (s, 3H, ring H(6)), 8.72 (s, 3H, imine CH). 1 3 C NMR (75.48 MHz, DMSO - d6): 62.46,115.68, 117.18,128.96,129.86, 138.89, 160.88, 164.60. Mass spectrum (LSIMS): m/z = 682 ([M - 3Na + 4H] +, [ C 2 7 H 2 8 N 3 0 1 2 S 3 ] + ) , 704 ([M - 2Na + 3H] +, [C27H27N 3 Na0 1 2 S3] + ) , 726 ([M - Na + 2H] +, [C 27H2 6N3Na20 1 2S3]+), 748 ([M + H] + , [ C 2 7 H 2 5 N 3 N a 3 0 1 2 S 3 ] + ) , 770 (M + Na] +, [ C 2 7 H 2 4 N 3 N a 4 0 1 2 S 3 ] + ) . IR (cm"1, KBr disk): 3454 (b s, v 0 - H ) , 1639 (s, v c = N ) , 1524,1382 (m, v c = c ) , 1186 (vs, v s = 0 ) , 1108, 1035 (s, v c_ c). cis, cis - l,3,5-tris(((2-hydroxy-5-sulfobenzyl)amino)cyclohexane tetrahydrate ( H 6 T A C S » 4 H 2 0 ) . Methanol (100 mL) was cooled to 0 °C in an ice bath and sodium borohydride (0.28 g, 7.4 mmol) was added with stirring. To this was added Na3saltachs (1.30 g, 1.7 mmol) in small portions (~ 10 mg each). After each portion was added, the reaction was allowed to proceed until the solid Schiff base had all dissolved and the bright yellow colour dissipated. After all the Na3saltachs had been added (~ 1 hr), the solution was heated to reflux and stirred for 30 min. The solution was then acidified with 6 M HC1, resulting in a white precipitate. The methanol was then boiled down to 20 mL, another 100 mL methanol was added, and the methanol again boiled down to 20 mL. The white suspension was allowed to cool to room temperature, and the white solid was collected on a medium frit. The white solid was then dissolved in water, and this solution was passed down a Rexyn 101 cation exchange column ( H + form) and eluted with deionized distilled water until the eluant was no longer acidic. The solvent was removed from the eluant under reduced pressure at 65 °C to give a white solid. The solid was dried at 95 °C in vacuo overnight to yield 0.73 g (55%); mp > 250 °C. Anal. Calc. (found) for C 2 4 H 3 9 N 3 0 1 2 S 3 « 4 H 2 0 : C, 42.68 (42.89); H , 5.44 (5.40); N, 5.53 (5.49). Potentiometric analysis was consistent with this 30 formulation. 2 H NMR (200 MHz, D z O , pD = 12, see Scheme 2.5): 0.97 (q, 3H, H(A f l X), 2 J A B = 11 Hz, 3 J A C = 11.7 Hz), 2.26 (d, 3H, H(Be(?), 2 J A B = 11 Hz, 3 J B C = 2.0 Hz), 2.67 (t, 3H, H(c^), 3 J A C = 11.7 Hz, 3 J B C = 2.0 Hz), 3.67 (s, 6H, benzylic CH 2 ) , 6.56 (d, 3H, ring H(3), 3 J = 8.6 Hz), 7.39 (dd, 3H, ring H(4), 3 J = 8.6 Hz, 4 J = 2.6 Hz), 7.46 (d, 3H, ring H(6), 4 J = 2.6 Hz). 1 3 C NMR (75.5 MHz, D 2 0 , pD = 12): 40.28, 43.71, 49.03, 55.37,121.14,128.72,128.99,129.54,130.72.171.53. Mass spectrum (LSIMS): m/z = 688 ([M+H]+, [C 2 6 H 3 4 N 3 0 1 2 S3] + ) . IR (cm"1, KBr disk): 3500-2600 (b s, v N _ H , OH)> 1607 (s, 8N_H), 1503,1433 (s, v c = c ) , 1210,1171 (vs, v s = 0 ) , HOI, 1033 (s, v c_ c). UV (Xmax nm, (e M _ 1 cm _ 1 ): pH=2: 276 (4960), 232 (33600); pH=12: 287 (12000), 256 (48800). It is soluble in water and hot DMSO. Metal Complexes The metal complexes of these sulfonated ligands were prepared in situ; no attempts were made to isolate them; however, many structural reports exist for the analogous nonsulfonated ligand metal complexes. 1 8 , 2 0" 2 4 The method used was the same in all instances, and the complexes were characterized by 1 H and in some cases 1 3 C , 2 7 A1, 7 1 G a , or 1 1 5 In NMR. As complex formation is pD dependent, the usual protocol for complex synthesis was to take the appropriate stoichiometry of metal nitrate and ligand precursor (30 - 50 mM) and adjust the pD to the desired level using NaOD or DC1 solutions. In order to compare the spectral behaviour (in D 2 0) with the thermodynamic constants determined (in H 2 0) , the relationship pD = p H + 0.40 was used 2 8 Na 3[Al(TAMS)]. X H NMR (300 MHz, D 2 0 , pD = 10.0): 0.68 (s, 3H, methyl CH 3 ) , 2.58 (d, 3H, methylenic C H , 2 J A B = 14.3 Hz), 2.63(d, 3H, methylenic C H , 2 J A B = 14.3 Hz), 3.27 (d, 3H, benzylic C H , 2 J A B = 13.5 Hz), 4.19 31 (d, 3H, benzylic C H , 2 J A B = 13.5 Hz) 6.86 (d, 3H, ring H(3), 3 J = 8.4 Hz), 7.56 (d, 3H, ring H(6), 4 J = 2.4 Hz), 7.67 (dd, 3H, ring H(4), 3 J = 8.4 Hz, 4 J = 2.4 Hz). 2 7 A1 NMR (D 20, pD = 10): 8 =+5 ppm, W 1 / 2 = 870 Hz. N a 3 [ G a (TAMS) ] . *H NMR (300 MHz, D 2 0 , pD = 9.3): 0.69 (s, 3H, methyl CH 3 ) , 2.69 (d, 3H, methylenic C H , 2 J A B = 13.3 Hz), 2.77 (d, 3H, methylenic C H , 2 J A B = 13.3 Hz), 3.37 (d, 3H, benzylic C H , 2 J A B = 15.0 Hz), 4.23 (d, 3H, benzylic C H , 2 J A B = 15.0 Hz) 6.82 (d, 3H, ring H(3), 3 J = 9.7 Hz), 7.56 (s, 3H, ring H(6)), 7.68 (d, 3H, ring H(4), 3 J = 9.7 Hz). 7 1 G a NMR (D 20, pD = 9.3): 8 = +34 ppm, W 1 / 2 = 3400 Hz. N a 3[In ( T A M S ) ] . J H NMR (300 MHz, D 2 0 , pD = 9.7): 0.70 (s, 3H, methyl CH 3 ) , 2.76 (d, 3H, methylenic C H , 2 J A B = 15.0 Hz), 2.90 (d, 3H, methylenic C H , 2 J A B = 15.0 Hz), 3.57 (d, 3H, benzylic C H , 2 J A B = 14.3 Hz), 4.08 (d, 3H, benzylic C H , 2 J A B = 15.0 Hz) 6.62 (d, 3H, ring H(3), 3 J = 8.7 Hz), 7.49 (d, 3H, ring H(6), 4 J = 2.3 Hz), 7.63 (dd, 3H, ring H(4), 3 J = 8.7 Hz, 4 J = 2.3 Hz). 1 1 5 In NMR (D 20, pD = 9.7): 8 = +113 ppm, W 1 / 2 = 26000 Hz. N a 3 [ A l ( T A P S ) ] . % NMR (200 MHz, D 2 0 , pD = 9.0): 2.52 (d, 1H, methylenic C H , 2 J A B = 13 Hz), 2.86 (d, 1H, methylenic CH', 2 J A B = 13 Hz), 3.05 (s, 1H, methine CH), 3.36 (d, 1H, methylenic C H , 2 J A B = 13 Hz), 3.48 (d, 1H, benzylic C H 2 J A B = 13 Hz), 3.53 (d, 1H, benzylic C H 2 J A B = 13 Hz), 3.73 (d, 1H, methylenic CH', 2 J A B = 13 Hz), 3.77 (d, 1H, benzylic C H , 2 J A B = 13 Hz), 3.93 (d, 1H, benzylic CH", 2 J A B = 13 Hz), 4.24 (d, 1H, benzylic CH", 2 J A B = 13 Hz), 4.35 (d, 1H, benzylic CH', 2 J A B = 13 Hz), 5.91 (d, 1H, ring H(3), 3 J = 8.8 Hz), 6.82 (d, 1H, ring H(3'), 3 J = 8.5 Hz), 6.98 (d, 1H, ring H(3"), 3 J = 8.5 Hz), 7.38 (d, 1H, ring H(6'), 4 J = 2.0 Hz), 7.39 (dd, 1H, ring H(4), 3 J = 8.5 Hz, 4 J = 2.0 Hz), 7.49 (d, 1H, ring H(6), 4 J = 2.0 Hz), 7.57 (d, 1H, ring H(6"), 4 J = 2.0 Hz), 7.65 (dd, 1H, ring H(4'), 3 J = 8.5 Hz, 4 J = 2.0 Hz), 7.67 (dd, 1H, ring H(4"), 3 J = 8.5 Hz, 4 J = 2.0 Hz). 2 7 A1 NMR (D 20, pD = 9.0): 8 = +16 ppm, W 1 / 2 = 660 Hz. 32 Na3[Ga(TAPS)]. J H NMR (200 MHz, D 2 0 , pD = 9.4): 2.53 (d, 1H, methylenic C H , 2 J A B = 13 Hz), 2.94 (d, 1H, methylenic CH', 2 J A B = 13 Hz), 3.07 (s, 1H, methine CH), 3.46 (d, 1H, methylenic C H , 2 J A B = 13 Hz), 3.50 (d, 1H, benzylic C H 2 J A B = 13 Hz), 3.57 (d, 1H, benzylic C H 2 J A B = 13 Hz), 3.76 (d, 1H, methylenic CH', 2 J A B = 13 Hz), 3.76 (d, 1H, benzylic C H , 2 J A B = 13 Hz), 3.97 (d, 1H, benzylic CH", 2 J A B = 13 Hz), 4.33 (d, 1H, benzylic CH", 2 J A B = 13 Hz), 4.43 (d, 1H, benzylic CH', 2 J A B = 13 Hz), 5.81 (d, 1H, ring H(3), 3 J = 8.5 Hz), 6.88 (d, 1H, ring H(3'), 3 J = 8.5 Hz), 7.02 (d, 1H, ring H(3"), 3 J = 8.5 Hz), 7.38 (d, 1H, ring H(6), 4 J = 2.0 Hz), 7.40 (dd, 1H, ring H(4), 3 J = 8.5 Hz, 4 J = 2.0 Hz), 7.48 (d, 1H, ring H(6'), 4 J = 2.0 Hz), 7.58 (d, 1H, ring H(6"), 4 J = 2.0 Hz), 7.66 (dd, 1H, ring H(4'), 3 J = 8.5 Hz, 4 J = 2.0 Hz), 7.68 (dd, 1H, ring H(4"), 3 J = 8.5 Hz, 4 J = 2.0 Hz). 7 1 G a NMR (D 20, pD = 9.4): 8 = +57 ppm, W 1 / 2 = 1230 Hz. Na3[In(TAPS)]. a H NMR (200 MHz, D 2 0 , pD = 9.6): 2.62 (d, 1H, methylenic CH), 3.03 (d, 1H, methylenic CH'), 3.11 (s, 1H, methine CH), 3.50 (d, 1H, benzylic CH), 3.61 (d, 1H, methylenic CH), 3.61 (d, 1H, benzylic CH'), 3.74 (d, 1H, methylenic CH'), 3.83 (d, 1H, benzylic CH), 3.94 (d, 1H, benzylic CH"), 4.41 (d, 1H, benzylic CH"), 4.41 (d, 1H, benzylic CH'), 6.09 (d, 1H, ring H(3), 3 J = 8.5 Hz), 6.99 (d, 1H, ring H(3'), 3 J = 8.6 Hz), 7.05 (d, 1H, ring H(3"), 3 J = 8.6 Hz), 7.37 (d, 1H, ring H(6), 4 J = 2.4 Hz), 7.42 (dd, 1H, ring H(4), 3 J = 8.5 Hz, 4 J = 2.4 Hz), 7.62 (d, 1H, ring H(6'), 4 J = 2.3 Hz), 7.62 (d, 1H, ring H(6"), 4 J = 2.3 Hz), 7.68 (dd, 1H, ring H(4'), 3 J = 8.6 Hz, 4 J = 2.3 Hz), 7.68 (dd, 1H, ring H(4"), 3 J = 8.6 Hz, 4 J = 2.3 Hz). 1 1 5 In NMR (D 20, pD = 9.6): 8 = +175 ppm, W 1 / 2 = 22000 Hz. Na3[Ga(TRNS)]. 1 3 C NMR (75.5 MHz, D 2 0 , pD = 10.0): 42.14,42.81, 43.06,44.02,46.01,48.54,48.81,50.96,51.14,51.69,52.00,52.25, 55.43,55.61,57.74, 119.30,119.55,120.27,120.83,120.97,123.46,127.25,127.49,127.60,127.82,128.16, 128.54,128.99,130.57,131.82,165.11,165.77. 33 Na3[In(TRNS)]. X H NMR (300 MHz, D 2 0 , pD = 9.0): 2.55 (m, 6H, ethylene CH 2 ) , 3.62 (d, 3H, benzylic C H , 2 J A B = 11.4 Hz), 5.08 (d, 3H, benzylic C H , 2 J A B = 11.4 Hz), 6.87 (d, 3H, ring H(3), 3 J = 8.1 Hz), 7.57 (d, 3H, ring H(6), 4 J = 2.4 Hz), 7.64 (dd, 3H, ring H(4), 3 J = 8.1 Hz, 4 J = 2.4 Hz). Na3[Ga(TACS)]. Four equivalents of Ga(N0 3 ) 3 were added to a H 6 T A C S solution at pD = 9.3 and the solution heated to 70 °C for 30 min. A white solid was filtered off and the H and Ga NMR spectra were obtained for the filtrate. The filtrate was found to contain about 70% complex and 30% unreacted ligand. 1 H NMR of the complex, (300 MHz, D 2 0 , pD = 9.3, see Scheme 2.5): 1.92 (d, 3H, H(Aax), 2 J A B = 15 Hz), 2.34 (d, 3H, H(Becj), 2 J A B = 15 Hz), 3.28 (s, 3H, H(Ce(?)), 3.76 (d, 3H, benzylic C H , 2 J A B = 12.8 Hz), 4.07 (d, 3H, benzylic C H , 2 J A B = 12.8 Hz), 6.41 (d, 3H, ring H(3), 3 J = 8 Hz), 7.56 (d, 3H, ring H(6), 4 J = 2.4 Hz), 7.60 (dd, 3H, ring H(4), 3 J = 8 Hz, 4 J = 2.4 Hz). 7 1 G a NMR (D 20, pD = 9.3): 5 = +18 ppm, W 1 / 2 = 1000 Hz. Potentiometric Equilibrium Measurements. Potentiometric measurements of the pro-ligands in the absence, and presence, of Al(III), Ga(III), and In(III) were performed with a Fisher Acumet 950 pH equipped with an Orion Ross glass and calomel reference electrodes. The electrode was calibrated before each titration, and often afterwards, by titrating a known amount of aqueous HC1 with a known concentration of NaOH. A plot of mV (measured) vs p H (calculated) gave a working slope and intercept so that the pH could be read as -log [H+] directly. A Metrohm automatic burette (Dosimat 665) was used for the NaOH additions and the burette and pH meter were interfaced to a PC such that each titration was automated. The temperature of the solutions in covered, water jacketed beakers was kept constant at 25.0 ± 0.1 °C by a Julabo circulating bath. The ionic strength was fixed at 0.16 M NaCl. 34 Argon, which had been passed through 10% NaOH, was bubbled through the solutions to exclude C O 2 . The pro-ligands were checked for purity by NMR and elemental analysis before titration. Titrations were also employed to ensure that the molecular weight obtained by elemental analysis for the required stoichiometry was the same as that determined by titration. Fresh pro-ligand solutions were used as the aminophenolates slowly oxidize in aerated solutions, especially in the presence of base. The metal ion solutions were prepared by dilution of the appropriate atomic absorption (AA) standards. The exact amount of acid present in the A A standards was determined by titration of an equimolar solution of metal standard and N a 2 H 2 E D T A . The amount of acid present was determined by Gran's method,2 9 and this was equal to the amount of acid in the A A standard plus the two equivalents of acid liberated from the complexed EDTA. NaOH solutions (0.1M) were prepared from dilution of 50% NaOH with freshly boiled distilled, deionized water and standardized potentiometrically against potassium hydrogen phthalate. The ratios of ligand to metal used were 1:1 < L:M < 3:1. Concentrations were in the range 0.5 - 2.5 mM. A minimum of five titrations were performed for each metal - ligand combination, each titration consisting of about one hundred data points. The ligand solutions were titrated over the range 2 < pH < 11.5, while the metal - ligand solutions were titrated over the p H range 2 -11. Complexation was usually rapid (1-3 min per point to give a stable pH reading) in the Ga(III) and In(III) studies; however, caution was taken to ensure that no trace hydrolysis or precipitation was occurring by monitoring for up to 30 minutes for pH drift. 35 For the Al(III) stability constant studies, and for those involving Ga(III) with H^TACS, equilibration was too slow for the automated titration procedure and a batch method was employed instead. Here, a series of 24 (0.16 M NaCl) C02-free solutions was prepared with a 1.1 : 1 L : M ratio and varying amounts of NaOH were added. These solutions were equilibrated at 25 °C until the pH reading of each stabilized. Two batches of titrations were performed for each of Al-TAMS and Al-TAPS, and these were fully equilibrated within three days. For Ga-TACS, equilibrium was not reached after five weeks, and TACS could not prevent some degree of Ga(OH) 3 formation even after standing for three months. Any solutions containing precipitate were excluded from the refinement of data. Computations - Potentiometry. The protonation constants for the ligand and the lanthanide-ligand stability constants were determined by using the program BEST. 3 0 This program sets up simultaneous mass-balance equations for all the components present at each addition of base and calculates the pH at each data point according to the current set of stability constants and total concentrations of each component. The stability constants are iteratively varied to minimize the sum of the square of the difference between observed and calculated pH. An indication of the fit is given by a, where a is a = E(pH c a i c -pH 0 bs) 2 / (N- l ) where N is the total number of data points. In all the titrations considered, o < 0.01. The hydrolysis constants used were taken from those in Baes and Mesmer for |i = 0.10 M and extrapolated to | i = 0.16 M using the equation listed for each metal;31 in the case of In(III), formation constants with chloride 36 were also included in the model. The deprotonation constants listed for the pro-ligands refer to the equilibrium: H n L ( 6 . n ) . > H ^ L ^ - + H + (1) pK a ( 7-n) = [Hn-iL^7 _ n^] »[H + ] / [HnTJ6"n)"] The equilibrium constants reported in this thesis refer to reactions involving the fully deprotonated pro-ligand, L n ' , unless otherwise noted (charges and [] omitted for clarity). M + L « ML (2) K M L = ML/M«L M + H + L « H M L (3) K H M L = HML/M«H»L M + L « M L O H + H (4) K M L O H = MLOH«H/M»L Therefore, the dissociation constant for a protonated metal complex, (eq. 5), H M L - ML + H (5) P K H M L = M L . H / H M L = log K H M L - log K M L ML(OH 2) » M L O H + H (6) p K a = ML(OH).H/ML(OH 2 ) = log K M L - log K M L O H 37 and the p K a of a coordinated water molecule (eq. 6) can be related to the equilibria 2-4. The data were modeled initially with only a ML species. To this simple model were added protonated species, H X M L , and/or hydroxo species, ML(OH), sequentially. All of the metal complex equilibria could be explained satisfactorily with ML and sometimes HML; the inclusion of other species only worsened the fit. Unless otherwise noted, the stability constants listed in this thesis represent l a in the final decimal place. 2.3 Results and Discussion Ligand Syntheses Evans and Jukobovic ' had previously reported on the Al , Ga, In, and Fe complexes of saltrens and saltames. However, the authors prepared their complexes in situ, by a template method and did not isolate free ligand. The Schiff bases reported here were all easily prepared by mixing the appropriate amine with sodium 5-sulfosalicylaldehyde in methanol. Although these compounds hydrolyzed readily in water, they were stable in deuterated DMSO for weeks. They were clearly marked by the distinctive imine resonance at 8.7 ppm (1H) and 166 ppm ( 1 3C) in the NMR spectrum, as well as the C=N stretch at 1640 cm"1. The Schiff bases were isolated as trisodium salts as evinced by their mass spectra which give the parent ion in a pattern typical of three sodium atoms present, i.e. M - 3Na + + 4H + , M - 2Na + + 3 H + , M + Na + . Apart from H 6 TRNS, which was sufficiently insoluble in acidic aqueous solution to allow its recrystallization, no suitable solvent was found for the recrystallization of the sulfonated amine phenols. However cation exchange chromatography proved sufficient for analytical purity. It was 38 important to remove the boron as the volatile trimethoxyboron before chromatography, otherwise a boric acid impurity resulted. The sulfonated amine phenols were all isolated as inner salts. Positive LSIMS showed the parent ion plus one, and the elemental analyses were in agreement with the proposed formulations; there was no evidence for a sodium salt. The NMR spectra of these compounds were straight forward owing to the two- or three-fold symmetry present in the molecules. Loss of the imine resonance at 8.7 ppm ( H) and 166 ppm ( C) and the appearance of a benzylic resonance at ca. 3.6 ppm ( H) and 55 ppm ( C) was diagnostic of conversion of the Schiff base to the amine phenol. The infrared spectra of the series of Schiff bases were about 90% superimposable indicating the dominance of the 2-hydroxy-5-sulfonato-benzylidene moiety on the spectra. Likewise, the series of amine phenols were virtually superimposable. The most dramatic difference between an amine phenol and its corresponding Schiff base was the exceedingly broad absorption between 3500 and 2600 cm"1 in the amine phenols, indicative of extensive hydrogen bonding in the solids; with the Schiff bases, there existed only a 50 cm wide band centred at 3450 c m . The loss of the imine linkage was also detected by IR where the C=N stretch at 1640 cm"1 is replaced by an N - H bend at 1605 cm"1 after treatment of the Schiff base with borohydride. Deprotonation Constants and Conformational Changes The deprotonation constants of H 6 TRNS, H 6 TAMS, H 6 TAPS, and H 6 T A C S are listed in Table 2.1. The pK as of the four pro-ligands treated here all showed similar behaviour. The three phenolate moieties had pK a s between 6 and 9 in all four instances. This is demonstrated in Figure 2.2 where absorbance at 256 nm is plotted against pH. Since the phenolic group is the only chromophore in the molecules in question, any change in the UV 39 Table 2.1. Deprotonation Constants.3 p K a H 6 T R N S H 6 T A M S . H 6 T A P Sb H 6 T A C S 6 11.2 (1) N-H 11.19 (4) N-H 11.24 (9) N D - H 11.24(5) N - H 5 10.6 (1) N-H 9.81 (4) N-H 9.77(6) N Q - H 9.84(2) N-H 4 9.59 (3) N-H 8.91 (2) O-H 8.73 (4) O r H 8.95 (1) O-H 3 8.07 (3) O-H 7.95 (3) O-H 7.78 (3) 0 0 - H 7.85 (2) O-H 2 7.29 (3) O-H 6.56 (2) O-H 6.54 (2) 0 0 - H 7.08 (1) O-H 1 6.17 (3) O-H 2.92(2) N-H 1.7(1) N r H 6.01 (3) N - H a The numbers in parentheses refer to l a . N-H and O-H refer to ammonium and phenol deprotonation, respectively. The subscripts o and i refer to the outer and inner arms, respectively. 200 220 240 260 280 300 320 ^(nm) Figure 2.1. Variable pH UV Spectra of H 6 TAPS (~2 mM). Uncorrected for Dilution. 40 spectrum should be attributable to phenol deprotonation. When phenols are deprotonated, there is a bathochromic shift in the absorption bands (232 nm -> 256 nm, 276 nm -» 288 nm) and a concomittant increase in e m a x . 3 2 Figure 2.1, a variable pH UV titration of HgTAPS, is typical of the four pro-ligands discussed in this chapter. The band at 256 nm was chosen since this represented the largest degree of change in the spectrum with pH; however graphs with the same pH profile could be generated from other wavelengths between 200 - 300 nm. 40-S 30-| x </"! CN CO 20. 10-TRNS o TAPS + TACS A TAMS • o of o °+ o ra ° es + o O 0 + o P + O 0+ O 0+ O 0+ O Q + o Q 0 + o O 0 I 8 10 pH + Figure 2.2. Plot of Molar Extinction Coefficient (e, M'^cm"1) at 256 nm vs. p H for H 6 TRNS, H 6 TAPS, H 6 TACS, and H 6 T A M S . Variable pD XB. NMR was also used to probe the deprotonation of the ligands. In the case of H 6 TAMS, H 6 TAPS, and H 6 TACS, the first 41 deprotonation occurred at an ammonium centre, whereas for H 5 T R N S the first deprotonation was a phenol hydrogen. Comparison of the deprotonation constants for the parent amines tame, tap, tach, and tren gives a useful empirical relationship. If the amine has a pKa < 8 then this pKa becomes lower in the amine phenol, whereas a pKa > 8 in the parent amine becomes elevated in the corresponding amine phenol. Because HfcTAPS lacks the three-fold symmetry of the other molecules, more information can be gleaned about its deprotonation. The central arm of H^TAPS is distinguishable from the outer arms by NMR, and this allows the assignment of the pKas in Table 2.1. Figure 2.3 shows the influence of p H on the chemical shifts of various 1 H resonances in the molecule. It is clear from the benzyl (b and b') and the amine backbone shifts (a and a') that the inner arm ammonium is the first site to deprotonate. The shift of the H resonances ortho to the hydroxyls (c and c') indicate that the outer arm phenol moieties deprotonate first, followed by the inner phenol moiety. This deprotonation scheme is expected since the inner amine offers a site to H -bond to the inner phenol, and this should raise the pKa of the inner phenol group, relative to the outer phenol groups. Somewhat surprising behaviour was noted for the benzyl, methylene backbone, and methine backbone resonances, especially when the molecule was in the protonation state H5TAPS". Figure 2.4 shows the 1 H NMR for these resonances at various pH. When the molecule is hexaprotonated (pD = 0.9), the two benzylic peaks almost overlap and are seen as sharp singlets. The methine resonance is the expected quintet, while the methylene peak appears as a doublet. Raising the pD to 4.3, reveals the outer benzylic peak split into an AB doublet ( J A B = -12.9 Hz), while the inner benzylic resonce remains a sharp singlet. The methine and methylene signals now form a complex 42 Figure 2.3 Variable pH J H NMR Titration of H 6 TAPS. 43 multiplet of an AA'BB'C spin system ( 2 J A B = 2 J A . B . = -12 Hz, 3 J A C = 3JA>C = 8 Hz, 3 J B c = 3 J B ' c = 4 Hz). Upon deprotonation of a phenol group (pD = 6.7), the spectrum once again simplifies and the methylene hydrogen atoms are seen as a doublet once more. When all three phenol moieties are deprotonated (pD 9.7), there is again evidence for rigidity. In this instance, the methylene resonance shows the same splitting pattern as in the pD = 4.3 spectrum, but now the benzylic resonances are both singlets. At the point of complete deprotonation (pD = 12.2), the spectrum is again simplified. Here the benzyl resonances overlap and the methylene resonance is once again a doublet. To further examine this rigidity, the sample at pD = 4.3 was heated in 15 °C increments to 85 °C. The only change in the spectra was the migration of the HOD peak which overlapped the benzylic AB quartet. Figure 2.5 shows the spectrum of the pD = 4.3 sample at 25 °C (middle), its simulation using the coupling constants listed above (bottom), and at 85 °C (top). At 85 °C the HOD resonance has shifted to lower frequency to a point such that the outer arm benzyl resonance can once again be seen. There is very little change between the 85 °C and 25 °C spectra. The separation of the outer arm A and B benzyl resonances decreases slightly at the higher H5TAPS" temperature, and the inner arm benzylic hydrogen atom resonance shifts slightly to higher frequency. However the massive HOD resonance occurring in the 85 °C spectrum makes phasing difficult, and quantitation of small shift changes is obviated. There are some minor resonances located about 3.3 ppm in the 25 °C spectrum that are not present in the simulation. These could be either an impurity or a minor conformer. A minor conformer would be 44 pD 6.7 pD 12.2 pD 0.9 pD9.7 U J I i i i i i I i i i i i | i i i i i | i i i i i | i i | i i i i | i i i i | i i i i | i i i i | 4.6 4.1 3.6 3.1 2.6 4.5 4.0 3.5 3.0 2,5 ppm ppm Figure 2.4. Aliphatic and Benzylic Portion of 1 H NMR Spectrum (200 MHz) of H 6 TAPS at Various pD. 4 5 expected to have a temperature dependence on its intensity if there is a conformational equilibrium. There is such a change, which suggests the presence of a minor conformer, and the minor conformer appears to be disfavoured at higher temperatures. These observations lead to the proposal of a conformation of H5TAPS" as shown above. Deprotonation of the inner arm ammonium leads to the formation of a hydrogen bond network of three H-bonds among the three nitrogen atoms giving rise to one six- and two five-membered rings, whereby the outer arm ammonium nitrogen atoms share their hydrogen atoms with the inner arm amine nitrogen atom. Such a solution structure would explain the 1 H NMR spectrum which still shows two-fold symmetry. The inner arm benzylic resonance appears as a singlet since both hydrogen atoms point toward an outer arm benzylic hydrogen atom. The outer arm benzylic carbon atoms have one attached hydrogen atom facing toward the inner arm benzyl group, while the second hydrogen atoms faces towards the other outer arm benzyl group; this would give the observed AB quartet. This is also consistent with the AA'BB'C pattern seen for the backbone, and the coincidence of the chemical shifts of 5 A = SA' and 8g = 83'. Loss of another proton, this time from an outer arm phenol, to give H 4 TAPS leads to singlets for both benzylic resonances, and a doublet for the methylene group. Presumably this is because there is more than one hydrogen bonding network available to the molecule, and these H-bonding arrangements are close in energy. For instance, in addition to the arrangement shown above (with a phenol deprotonated now), there should be a strong charged hydrogen bond available between the phenolate oxygen atom and the outer arm ammonium. This is also available to the second outer arm ammonium, and rapid conversion between the three would 46 account for the spectrum shown in Figure 2.4. As the molecule is successively deprotonated, more arrangements exist for the formation of intramolecular hydrogen bonds and this eliminates the presence of one rigid solution structure. The presence of this strong intramolecular hydrogen bonding which remains intact at 85 °C in H 2 0 argues strongly for the preorganized nature of H x T A P S n " in intermediate states of deprotonation. This should contribute a favourable entropic effect to the complexation of metal ions. A similar effect should exist with HgTRNS. The apical nitrogen atom is known to be very acidic (pK a < 1.5). Because of this, there can exist the possibility of the four nitrogen atoms sharing the three hydrogen atoms through a series of alternating H-bonds involving the secondary and tertiary nitrogen atoms (Scheme 2.4). A similar scheme can be envisaged for HsTAMS", since the first p K a of H 6 T A M S (2.91) is also quite acidic. However, because both H x TRNS n ~ and H x T A M S n " both possess three-fold symmetry, the benzylic, ethylenic, or methylenic H atoms will always appear equivalent, and the presence of this effect cannot be confirmed by 1 H NMR. + H 2 N N N H 2 + + H 2 N N N H 2 + i / W W V ^ W V <SWVt y ^ . A A A / 1 NH HN-.|_jt -N N H 2 + + H 2 N / ^N-Cj^NH NHc+ ' N H 2 + Scheme 2.4 48 The high acidity of an amino group in the three molecules discussed above can be thought of in two ways. Firstly, the protonation of all the amino groups would result in a large local charge increase, and the ease of proton loss is a result of reducing coulombic repulsion. However the deprotonated molecule may also exhibit added stability because of the possibility of the formation of hydrogen bonds between an ammonium group and a neighbouring amine. If this occurs (and the 1 H NMR of H5TAPS" supports this), then these molecules can exhibit some degree of preorganization in forming a tripodal cavity. The case of HfcTACS is somewhat different from the three compounds discussed above. Both the parent amine tach, and the amine phenol derivative described here exist in the equatorial conformation, whereby the amino groups point away from each other. This is clearly delineated by the coupling constants obtained for the three distinct cyclohexane ring hydrogen atoms. In Scheme 2.5 the two possible conformers along with the corresponding Newman projections are presented. If the equatorial conformer is present, Scheme 2.5 49 the methine hydrogen (He) should exhibit a large (8 -12 Hz) vicinal coupling to the axial hydrogen atom, H A , and an intermediate (2-4 Hz) vicinal coupling to the equatorial hydrogen atom, H B . 3 3 An axial conformation would give coupling constants in the range J A c = 2 - 4 Hz and J A B = 1-3 H z . 3 3 The observed constants are 3 J A c = H-7 Hz and 3 J A B = 2 Hz confirming the equatorial conformer. This is expected since the parent amine tach and the Schiff base saltachs both exist in the equatorial conformation. The difference in energy between the axial and equatorial conformers for tach has been calculated to be 11 kcal/mol in the gas phase,34 and this has been o r confirmed in a recent MM3 study. The equatorial conformer of the amine phenol ^ T A C S , because of the steric bulk of the benzyl groups, should be even more stable relative to the axial conformer. The equatorial conformer is present over the p H range studied in the 1 H NMR titration (pD 1-13). A pD = 12 solution was heated to 90 °C with no change. The equatorial conformation has ramifications in the magnitude of the pK a s of H 6 T A C S . The first deprotonation occurs at one of the ammonium moieties, but this ammonium is considerably less acidic (pK a = 6) than the analogous H^TAMS (pK a = 2.9) ammonium. This is because the physical separation of the ammonium groups is greater than in H 6 T A M S , and also because the deprotonated species H5TACS doesn't have the opportunity to form an intramolecular H-bond between an amine and ammonium group. The higher basicity of this ligand compared to TAMS 6" or TAPS6" should manifest itself in the ability to bind metal ions in acidic media. Since metal complexes of TACS derivatives all have the ligand in an axial conformation, complexation must involve a ring flip. No evidence of conformational interconversion was observed at 90 °C. This kinetic barrier to conformational change, coupled with the fact that the axial conformer is 50 higher in energy than the equatorial conformer, may result in slow metal complexation kinetics. Metal Complexes. M(III) - TAMS. The most straightforward complexation reactions were those involving PLjTAMS. The gallium(III) and indium(III) complexes formed rapidly (1-2 minutes per data point) and were completely formed by p H 4 and pH 5, respectively. The aluminum(III) complex formed slowly, necessitating a batch titration method. With [Al(TAMS)]3", p H readings were stable after one day; however, the stability of the aluminum complex was insufficient to prevent some precipitation of Al(OH)3. Ga(III) and In(III) only form [M(TAMS)]3" complexes, whereas Al(III) forms this species as well as a [M(HTAMS)]2" species. The most stable species here is [Ga(TAMS)]3" and this is clearly seen in the speciation diagram (Figure 2.6) which compares the complexation behaviour of TAMS 6" with the three metals. Notably Al(III) complexation begins at about the same p H as for In(III), because of the formation of the protonated Al complex; however, [Al(TAMS)]3" is much less stable than is [In(TAMS)]3", and this leads to some precipitation of Al(OH)3 (the region where Al(OH) 3 coexists is denoted by dashed lines). Another point to note from the speciation diagrams is the formation of chloro complexes of indium. Using an ionic medium of sodium chloride gives rise to chloro complexes at low pH, and the formation of these species suppresses the formation of In(OH)3. The result is a larger experimental window (about one p H unit) for the determination of indium stability constants. Use of nitrate or perchlorate as a background electrolyte would have introduced complications because of the competition from In(OH)3 formation during the titrations. 51 Figure 2 .6. Speciation Diagram of 2 mM M(III) : 2 mM H 6 T A M S . Dashed Lines Indicate the Region Where Precipitation Occurs. Charges Omitted for Simplicity. 52 The H NMR spectra of the TAMS complexes are similar; the Al , Ga, and In complexes all show the expected 3-fold symmetry in solution. Furthermore, the complexes are rigid in solution, as evinced by the inequivalence of the two benzylic hydrogen atoms and the inequivalence of the two methylene hydrogen atoms. This is analogous to the spectra observed in the nonsulfonated complexes which were obtained in DMSO-rfg. In the [M(TAMS)] complexes, the chemical shift separation of the methylene AB doublets increased down the the group (Al vs. Ga vs. In), whereas the chemical shift separation of the benzylic AB doublets decreased down the same group. This is likely a size effect (observed and discussed previously in the solid state structures of the nonsulfonated analogs21) which results in a change in the torsion angles about the benzylic and methylene carbon atoms. Table 2.2 Metal - Ligand Formation Constants.3 logK Al(III) Ga(III) In(III) M(HTRNS) 2" 36.90 (7) 34.9 (1) M(TRNS) 3" 28.55 (5) 29.3 (1) M(HTAMS) 2 " 29.3 (1) M(TAMS) 3 " 22.5 (1) 31.83 (5) 28.49 (3) M(HTAPS) 2" 29.0 (1) 35.15 (4) 31.93 (4) M(TAPS) 3" 22.8 (1) 31.54 (4) 27.56 (4) a K M H L = MHL/M»H»L; K M L = M L / M * L . 53 Figure 2.7. Speciation Diagram of 2 mM M(III): 2 mM H 6 TAPS. Dashed Lines Indicate the Region Where Precipitation Occurs. Charges Omitted for Simplicity. 54 M(III) - TAPS. The stability constants of the [M(TAPS)]3" complexes were of the same order of magnitude as their analogous TAMS complexes, but were slightly less (Table 2.2). In addition, all three metals formed stable protonated complexes. This is clearly delineated by the speciation diagram in Figure 2.7. The relative stability of the protonated complex decreases as the group is descended. Again, a batch titration was necessary to determine the A l binding constants, and as with TAMS, TAPS6" was unable to prevent some degree of aluminum hydroxide formation. This is in accord with the mol O H / m o l L Figure 2.8. Experimental Titration Curves for 2 mM M(III) : 2 mM H 6 L . Dashed Lines Represent H 6 TAPS Titrations, Solid Lines Represent H 6 T A M S . observations of Liu et al.20'21 who noted that it was difficult to prepare analytically pure samples of the Al(III) complexes of the uhsulfonated TAMS and TAPS analogs without some occlusion of Al(OH) 3. Figure 2.8 shows 55 experimental 1:1 M:L titration curves for Ga(III) and In(III) along with curves for the free ligands, H 6 T A M S and H 6 TAPS. The coincidence of each pair of curves shows the similar binding ability of the two ligands. Although the stability constant of [Ga(TAPS)]3" is slightly less than that of [Ga(TAMS)]3", the formation of the protonated species [Ga(HTAPS)]2" enables Ga sequestration at a lower pH. The : H NMR of the [M(TAPS)]3" complexes were similar to each other and to those reported for nonsulfonated analogs.20 All three arms are inequivalent in the complexes, giving rise to 20 distinct resonances. COSY experiments allow partial assignment of the spectra. One striking feature of the 1 H NMR spectra of the [M(TAPS)]3" complexes is the resonance of the ring 3' hydrogen (H(3) ortho to the oxy group). Two of the three arms have resonances in the chemical shift range of the other coordinated phenolates in this study (6.6 - 7.1 ppm), but the third arm is significantly deshielded, resonating between 5.9 and 6.1 ppm in the Al, Ga, and In complexes. Examination of the 5'-methoxy substituted Ga complex, as well as models of the complexes in this work, suggest that this is one of the outer arms. 1 H NMR was used to probe the solution structure of [Ga(HTAPS)]2", but since this species is always in equilibrium with free H5TAPS" and /or [Ga(TAPS)] , the observed spectrum was a series of overlapping peaks. A Ga -TAPS UV pH titration was carried out in another effort to elucidate the coordination sphere of [Ga(HTAPS)]2". In the H x TAPS n " system, the protonated phenols have an absorption maximum at 232 nm; upon deprotonation to the phenolate anion this maximum shifts to 256 nm. Titration of a 1.1 Ga : 1 TAPS solution saw the appearance of a new band at 249 nm, assignable to the coordination of a phenolato group to gallium. 56 Figure 2.9 shows a partial speciation diagram (only the Ga-TAPS species are shown for simplification), along with the absorbance at 249 nm as a function of pH. The [Ga(HTAPS)] species can have an N3O2 or an N2O3 donor set from the ligand, or both cases can exist if there is more than one isomer. The rise in absorbance mirrors exactly the curve for the formation of [Ga(HTAPS)]2" + [Ga(TAPS)]3", suggesting that [Ga(HTAPS)]2" has the ligand coordinated in an N2O3 manner, ie. with an amine still protonated. The change being monitored here is the extent of phenolate coordination. This Figure 2.9. Partial Speciation Diagram ([Ga(OH) x]n + Species Omitted for Clarity) for 1 Ga(III) : 1 H 6 TAPS (30 |xM) and the Molar Extinction Coefficient (e, M - W 1 ) at 249 nm vs. pH. [Ga(HTAPS)]2"- - ^ [Ga(TAPS)]3" / ([Ga(HTAPS)]2" + [Ga(TAPS)f) z (249 nm) • 57 change is complete when all of the ligand has been coordinated to Ga in one form or another ([Ga(HTAPS)]2" + [Ga(TAPS)]3"). If the protonated species were an N3O2 isomer with a protonated phenol, one would expect to see a continued increase at 249 nm over another pH unit, until formation of [Ga(TAPS)]3" is complete. M ( I I I ) - T R N S . With H 6 TRNS, the unsulfonated analogs showed diverse coordination chemistry in the solid state (Scheme 2.2), with Al being coordinated by an N3O3 donor set, Ga by an N4O2 donor set, and In by an 1 ft N 4 O 3 donor set. Titrations of the Al - TRNS system yielded little useful information. There is no complexation up to pH 6 or above pH 10. UV titrations suggested that there may be some interaction between the phenolate groups and A l between pH 6 and 10. H NMR titrations yielded spectra that were identical to the free ligand. If there is complexation, then exchange between [Al(HxTRNS)]n" and H x TRNS n " is rapid. Any changes in the chemical shifts with respect to those in the free ligand were small (< 0.04 ppm). Formation of Al(OH)3 obviated any potentiometric experiments. Ga complexation by TRNS was slower than that by TAMS or TAPS (5 -lOmin/pt). Two species were formed: a [Ga(HTRNS)]2" complex which was 100% present at pH 5.5, and a [Ga(TRNS)]3" complex. 1 H NMR clearly demonstrated the lack of C 3 symmetry in solution - a series of overlapping signals were seen in the aliphatic, benzylic, and aromatic regions. The only region of the 1 H NMR spectrum that could be assigned was between 6-7 ppm. In this region, three doublets corresponding to the 3' hydrogen atoms were observed. It was shown above (Figure 2.3, bottom) and in previous work that these resonances were sensitive to phenol deprotonation. Examination of a series of spectra (1 Ga : 1 TRNS) showed only one of these doublets having a p H dependent chemical shift. In a UV titration, a plot of 58 change in absorbance at 250 nm, corresponding to phenol deprotonation, versus pH gave a sigmoidal curve. These results are summarized in Figure 2.10. This suggests that Ga is coordinated by an N4O2 donor set in the same manner as was observed in the crystal structure of the unsulfonated analog.1 8 The spectroscopic data clearly show that the deprotonation of [Ga(HTRNS)]2" is taking place at a phenol. Since the chemical shifts of two of the 3' hydrogen atoms do not change, it is unlikely that this deprotonation involves a change in coordination geometry. Scheme 2.6 is a pictorial representation of the deprotonation process. The pK value of 8.35 obtained from both the UV and 0.20-0.15->S> 0.10-< 0.05 0.00-A , A A A A • O A on • o o • o • o • I 8 pH -44 o • 42 * o 1 -40 10 Figure 2.10. Plot of Change in 2 H Chemical Shift (Left Axis) vs. pH for H ortho to Hydroxyl Group (H(a) A , H(b) o , H(c) • as in Scheme 2.6) and Plot of Molar Extinction Coefficient (e, M'^cm"1) at 250 nm (Right Axis) vs. pH (Filled Symbols) for [Ga(TRNS)]3". 59 NMR data is in accord with that obtained by potentiometry, and is close to the value of the p K a of 4-sulfophenol (8.60).9 A 1 3 C NMR spectrum of the [Ga(TRNS)]3" complex showed 26 of the expected 27 resonances (the 2' C region showed only 2 resonances, presumably because of overlap). This suggests that there is only one major (> 90%) isomer present. The In - TRNS system was more straightforward. Equilibrium was established rapidly (1-2 min/pt) and TRNS complexed In over the same p H R = S03-Scheme 2.6 range as it did Ga. This is in contrast to the TAMS and TAPS systems where the Ga complexes were four orders of magnitude more stable than their In analogs. There is also a protonated species present, [In(HTRNS)]2", but it exists over a narrow p H range and never as the sole In containing species. The differences in the speciation of Ga(III) and In(III) with TRNS are highlighted in Figure 2.11. The 2 H NMR spectrum of [In(TRNS)]3" is characteristic of a rigid complex (the benzylic H atoms are inequivalent) with three-fold symmetry (only one set of resonances is observed for the three arms). The symmetry indicates that the donor set must be N4O3, or N3O3 with the apical nitrogen atom nonbonding. The former seven coordinate 60 complex seems more likely. This was the case in the solid state structure of the unsulfonated analog.18 Coordination of the apical nitrogen should be favored because it results in the formation of three 5-membered chelate rings. Also, [In(TRNS)]3- should be more basic if there is an uncoordinated amine. Figure 2.11. Speciation Diagram of 2 mM M(III) : 2 mM H 6 TRNS. Charges Omitted for Simplicity. M(III) - TACS. No stability constants could be determined with H^TACS because of its exceedingly slow complexation kinetics with Al, Ga, and In. A batch titration with gallium did not equilibrate even after one month. During this period, a slight excess of TACS (1.1 TACS : 1 Ga) could 61 not prevent the precipitation (of some degree) of gallium hydroxide. There was evidence of complex formation, however. Using an excess of of gallium (4 eq) at a pD of 9.3 showed two species in the 1 H NMR. There was a 70 : 30 ratio of (presumably) [Ga(TACS)]3": H x TACS n ". The Ga complex showed spectral similarities to similar nonsulfonated complexes reported by Roundhill and coworkers.23 The gallium complex has the expected three-fold symmetry in solution, and also displays rigidity as indicated by the inequivalence of the benzylic hydrogen atoms. The nitrogen donors of the ligand now occupy the axial positions of the cyclohexane ring, which is clear from the coupling constants. There was no evidence for an In - TACS complex utilizing the same preparative conditions. The slow rate of complexation is undoubtably because of the large energy barrier involved in the ring flip conformational change (vide supra). Furthermore, because TACS is in the wrong conformation for binding, this will also be reflected in a much reduced thermodynamic binding constant. Metal NMR. When possible, the metal NMR was recorded for the aforementioned complexes. 2 7 A1 and 7 1 G a are both quadrupolar nuclei (I = 5/2 and 1 = 3/2, respectively) with sufficiently low quadrupole moments such that solution NMR spectra can be recorded. Higher symmetry about the nuclei in question gives narrower resonances. 7 1 G a NMR spectra (Figure 2.12) were obtained for all the gallium complexes reported here with the exception of [Ga(TRNS)]3". [Ga(TRNS)]3" has an asymmetric N 4 0 2 coordination geometry, whereas [Ga(TAMS)]3", [Ga(TAPS)]3", and [Ga(TACS)] all have facial N 3 0 3 coordination geometries. Using the same conditions employed for the N 3 0 3 complexes ([Ga] = 30 mM, 10000 pulses, 90° flip angle, no delay time), no signal was observed for [Ga(TRNS)]3", presumably because of the lowered N4O2 symmetry about the gallium centre. 62 [Ga(TAPS)]3" resonated at +57 ppm relative to G a ( H 2 0 ) 6 3 + with W 1 / 2 = 1230 Hz. The [Ga(TACS)]3" resonance appeared at a lower frequency, +18 ppm (Wj/2 = 1000 Hz); both of these spectra were acquired with 1000 pulses. [Ga(TAMS)]3" displayed markedly different behaviour. It resonated at +34 ppm (Wi/2 = 3400 Hz), and because of the broader linewidth required 10000 pulses. An increase in the linewidth is indicative of an exchange process, a change in the correlation time T C , or a lower symmetry about the metal ion. There was no indication of intermolecular exchange in these complexes from their 1 H NMR spectra. Therefore the broader linewidth observed for [Ga(TAMS)]3" must be a result of a less symmetric environment about the gallium nucleus, or an increase in xc. Examination of the crystal structures of the corresponding unsulfonated Ga complexes 2 0 , 2 1 , 2 3 shows that the o [Ga(TAPS)] analog suffers the greatest deviation from octahedral geometry. 9 7 It has been observed that the Al linewidths in a series of tris(N-alkyl-3-oxy-4-pyridinonato)aluminum(III) complexes increased with increasing alkyl o n chain length, whereas the chemical shift remained unchanged. Likewise, Parker and coworkers noted in a 7 1 G a NMR study of 1,4,7-triazacyclononane-l,4,7-triyltrimethylenetris(alkylphosphinato)gallium(III) complexes that increasing the size of the alkyl group on the phosphorus atom increased the 71 Ga linewidth. Within each of these cases, the environment (electric field gradient) about the quadrupolar nuclei should be very similar, and the linewidth change can be ascribed to a change in the rate of molecular tumbling. With [Ga(TAMS)]3", the electric field gradient at the gallium nucleus should not be markedly different than in [Ga(TAPS)]3" or Q [Ga(TACS)] ; the greatest difference is in the overall shape of the complexes and these shape differences will affect xc. 63 i 1 1 1 1 r~ 0 -100 -200 -300 -400 -500 ppm 300 200 100 Figure 2.12. 7 1 G a NMR Spectra (91.5 MHz) of [Ga(TAPS)]3" (Top, 1000 Transients), [Ga(TACS)]3" (Middle, 1000 Transients), and [Ga(TAMS)]3" (Bottom, 10000 Transients); [Ga] = 30 mM, pD = 9.3,20 °C. 64 The chemical shifts for the gallium complexes lie in the range expected 7 1 for octahedral gallium(III) complexes; Cole et al. have reported Ga chemical shifts for Ga in an N3O3 environment. For a tris(amino)tris(carboxylato) environment, a shift of +171 ppm was observed and this shifted to lower frequency, between +130 and +140 ppm when the carboxylate groups were replaced by substituted phosphinato groups. With three phenolates as oxygen donors, the chemical shift range moves to an even lower frequency. The 2 7 A l NMR spectra of [Al(TAMS)]3" and [Al(TAPS)]3" showed the same trends as did their gallium congeners. Both appeared in the shift range expected for octahedral aluminum complexes. [Al(TAMS)]3" resonated at a lower frequency than [Al(TAPS)]3" and, as in the gallium complex, it had a larger linewidth. 1 1 5 In NMR were also recorded for [In(TAMS)]3" and [In(TAPS)]3"; the linewidths were outrageous (26 kHz and 22 kHz, respectively). Despite this, similar trends were observed - [In(TAMS)]3" resonated to lower frequency of [In(TAPS)r and again had a larger linewidth. Comparisons with Other Multidentate Ligands. The in situ work of Evans and Jakubovic16 allows some comparisons to be made on the effect of changing from an imine to an amine nitrogen donor. In their studies on the group 13 complexes of saltames6", they found that [Ga(saltames)]3" and [Al(saltames)] were rigid complexes in D 2 0 solution, whereas [In(saltames)]3" and [Tl(saltames)]3" showed fluxional behaviour. The larger metal ions are less easily accommodated by the N3O3 tripodal cavity. Because of the imine linkages, the saltames6" ligand is much more rigid than TAMS 6 " and presents a tripodal cavity that is more defined within a narrow size range. This can be seen in a comparison of the pM values (Table 2.3). [Ga(saltames)]3" is 7.6 orders of magnitude more stable than [In(saltames)]3". 65 Upon reduction of the imines to amines in TAMS ", the difference in stability between [Ga(TAMS)]3" and [In(TAMS)]3" drops from 7.6 to 4.1 orders of magnitude. The more flexible TAMS 6" ligand is better able to accommodate the larger indium(III) ion. [Ga(TAMS)]3" is only 0.5 orders of magnitude more stable than [Ga(saltames)]3" which suggests that the saltames6" cavity is well suited to the size of the Ga(III) ion. [Al(saltames)]3" is more stable than [Al(TAMS)]3" which again shows the size preference of the saltames6" cavity for the smaller ion, and that Al(III) prefers coordination by imine donors over amine donors. Evans and Jakubovic16 also gave a pM range for [In(saltrens)] , but their upper limit lies below that of the pM for T a b l e 2.3. Comparative pM Values Calculated at pH = 7.4 for 10 uM Ligand : 1 uM Metal ligand Al(III) Ga(III) In(III) TRNS 6 " 22.4 22.1 saltrens6" 17.9 < pM < 21.8a Sbad4" 22.9b 20.1b T A M S 6 " 15.2 25.1 21.0 saltames6" 16.9a 24.6a 17.0a TAPS 6" 15.8 24.4 20.4 transferrin 14.5C 20.9d 18.76 TACN-HP 29.6f 18.4f TACN-TX 25.9S 15.78 a ref. 16;b ref. 41; c ref. 2; d ref. 1;e ref. 3; f ref. 43; S ref. 42. 66 [In(TRNS)] ". Again this is likely because of the greater flexibility of the amine versus the imine, and the fact that TRNS6" acts as an N4O3 donor toward In(III), whereas saltrens acts as an N3O3 donor. A comparison of the N3O3 ligands TAMS 6" and TAPS 6" shows similar complexation abilities with the three metal ions studied; log K for [M(TAMS)]3" and [M(TAPS)]3" are both of the same order of magnitude for a given group 13 metal. The order of stability for both of these ligands with the three group 13 metals is Ga(III) > In(III) > Al(III). This is expected based on literature reports of similar complexes. Martell and coworkers have shown that the oxybenzyl donor displays a large preference for coordinating to R X = N, R = H, TACN-HP X = CH, R = C H 3 , TACN-TX gallium(III) over indium(III). Hancock and coworkers have given estimates for the formation constant of ammonia with the group 13 metals40 and these show log K for M ( N H 3 ) 3 + formation as In(III) ~ Ga(III) » Al(III). On the basis 67 of these reports, one can rationalize the preference of an aminophenolate binding to Ga(III) over In(III) in terms of the presence of the oxybenzyl donor. In the case of Sbad4", which is an N4O2 donor 4 1 (see below), a coordination preference for Ga(III) over In(III) was also observed, however the pM values (Table 2.3) were decreased by 2 orders of magnitude for Ga(III) and one order of magnitude for In(III). The lowering of the stability constants is probably because of the replacement of an anionic oxybenzyl donor by a neutral amino group, and this also leads to a ligand which is less selective for Ga(III) over In(III) relative to T A M S 6 - or TAPS 6 - . The same trend (Ga(III) > In(III)) is also observed with T A C N - T X 4 2 and T A C N - H P 4 3 (see above). For these N3O3 ligands, the amine backbone is the macrocycle 1,4,7-triazacyclononane and the selectivity for Ga(III) over In(III) is close to 10 orders of magnitude (Table 2.3). Comparing the pM values of saltames6" versus TAMS 6" shows the dramatic effect of ligand rigidity on the stability of the metal complex where [In(TAMS)]3" is 104 times more stable than [In(saltames)]3". Presumably, the macrocyclic ring in TACN-HP and TACN-TX is providing a similar effect, ie. the macrocycle has a clear preference for the smaller Ga(III) ion. This is supported by the fact that, although polyaminopolycarboxylates generally favour In(III) complexation over Ga(III) complexation, the aminocarboxylate nota shows a large (105) selectivity for Ga(III)44 The affinity of the oxybenzyl donor for Al(III) has not been probed in detail; however, the affinity of anionic oxygen donors for Al(III) is axiomatic. The low stabilities of the Al(III) complexes studied here, relative to their Ga(III) and In(III) analogs is likely because of the low affinity of Al(III) for the neutral amine donor. A consequence of this weak amine affinity is that the protonated complexes, [Al(HTAPS)]2" and [Al(HTAMS)]2", exist over a broader p H range than do their Ga(III) or In(III) analogs. By analogy with 68 [Ga(HTAPS)] ", one of the amine groups remains protonated, and Al(III) is coordinated by an N2O3 donor set with the last octahedral coordination site occupied by a water molecule. The preference of Al(III) for anionic oxygen donors, coupled with the fact that TAMS 6 " and TAPS6" failed to prevent some Al(III) hydrolysis makes it hardly surprising that no stability constants in the Al-TRNS system were obtained. The crystal structure of the nonsulfonated analog (Scheme 2.2) had one of the secondary amines protonated and uncoordinated. A result of this N3O3 donor set is that the stable six-membered chelate ring formed with 2-oxybenzylamino coordination is lost for one of the three arms. This lack of chelate ring stabilization coupled with the larger size of the TRNS backbone make this ligand useless for Al(III) coordination. The factors which disfavor Al(III) complexation by TRNS6" serve to make TRNS6" the best ligand in this series for the coordination of In(III). Indium is large enough to accommodate all seven donor atoms, resulting in the formation of three 6- and three 5-membered chelate rings. The optimization of all seven donors, coupled with the larger cavity size of TRNS6" and the higher affinity of In(III) for amine donors results in a ligand that coordinates In(III) as well as it does Ga(III). The coordination of Ga(III) by TRNS6" gives an N4O2 coordination geometry, and this results in a pM value lower than that in the N3O3 donors studied here, but of the same order of magnitude as in Sbad4", another N2O4 aminophenolate donor. 4 1 The Ga(III) ion is too small to be seven coordinate. By binding in an N4O2 fashion, the complex optimizes the number of 5- and 6-membered chelate rings formed (three 5- and two 6-membered rings). Although an N 3 0 3 donor set would be better suited to Ga(III), TRNS6" can only accomplish this by breaking up part of its chelate ring network. 69 This may represent a useful result in the selectivity of In(III) over Ga(III). Recently Martell, Welch, and coworkers have focused on developing chelators with a high affinity for In(III), and have had some success by incorporating thiol donors.4 5"4 9 Ligand design incorporating seven or eight donors would enhance In(III) selectivity by yielding an In(III) complex with a coordination number greater than six, and lead to steric congestion by the noncoordinating donors in the analogous Ga(III) complex. This has also been shown recently in stability constant studies involving dtpa derivatives.50 All of the ligands reported here, with the exception of TACS 6", form more stable complexes with Ga(III) and In(III) at pH 7.4 than does transferrin. From a thermodynamic standpoint, the aminophenolates based on tame and tap represent the best candidates for a ligand framework with which to design Ga radiopharmaceuticals. The tren backbone is better suited for In(III), but the use of oxybenzyl donors should be avoided for m I n radiopharmaceuticals and these should be replaced with thiolates. FI^TACS proved to be the poorest candidate of this series for group 13 metal sequestration. The only metal examined was Ga(III); based upon the results obtained with the Ga-TAMS and Ga-TAPS titrations, it is expected that Ga(III) might form a stable group 13 metal complex with TACS 6". The complexation kinetics were very slow, however, and the ligand was incapable of preventing some precipitation of Ga(OH) 3 upon standing for three months. Using the solubility product of amorphous Ga(OH) 3 3 1 and the deprotonation constants for H 6 TACS, an upper limit for [Ga(TACS)]3" formation can be determined to be log K = 30. From a thermodynamic viewpoint only, this means that complexation might not start until above pH 4, and that Ga(OH)4" would start to form at pH 10. This is a result of the higher basicity of TACS 6" relative to TAMS 6". Furthermore, complexation must take place with the 70 energetically disfavored axial conformer. If the free ligand existed in the axial conformer, there would be a gain in energy of complex formation in the amount of energy difference between the two conformers. This has been shown for M(II) complexes of tame and tach. 3 4 , 5 1 Both of these tris(amine) ligands bind in a facial manner and both have stability constants of the same order of magnitude for Ni(II), Cu(II), and Zn(II). 3 4 , 5 1 Paoletti and coworkers showed that the similarity in binding constants is because of a coincidence between an unfavorable enthalpy change and a favorable entropy change when tach coordinates to a metal. The amine tach, when in the axial conformation, is preorganized for binding, and this results in a more positive entropy relative to tame. In order to achieve this conformation, tach must undergo a ring flip to a conformer higher in energy; it is this unfavorable enthalpy process which nullifies the gain in entropy. Parker and coworkers have recently prepared the conformationally biased c/s-cz's-2,4,6-trimethyl-1,3,5-triaminocyclohexane ligand in which the three amino groups are in the axial conformation. They have also prepared the tris(2-hydroxybenzyl) derivative, as well as the aminocarboxylate and aminophosphinate derivatives. An aminophenolate ligand based on this amine should exhibit high formation constants with trivalent metal ions because of the favorable entropy associated with its preorganization, while the unfavorable enthalpic effect in tach derivatives is eliminated because no conformational change is necessary for binding. The lack of of a conformational change on binding should also serve to increase the kinetics of complexation. 2.4 Conclusions The tripodal framework offers intermediate chelation properties between open chain ligands and those based on macrocycles. 1 H NMR gives 71 some evidence for preorganization in the TAPS system, and this is likely present in the TRNS and TAMS systems as well. The order of stabilities for a given ligand are Ga > In > Al. The low affinity of Al(III) for amine donors coupled with slow complexation kinetics, makes these ligands unsuitable for Al(III) sequestration. The presence of the oxybenzyl donor imparts a selectivity for Ga(III) over In(III); TAMS 6" and TAPS6" are best suited for Ga(III) complexation. The larger ligand cavity and the seven donor atoms of TRNS 6" make it the best ligand for In(III). The solid state structures determined for the Ga(III) and In(III) complexes of the nonsulfonated analogs of TRNS, 1 8 T A M S , 2 1 and T A P S 2 0 are retained in aqueous solution. This parallel is absent for Al(III); Al(III) has a greater propensity for forming protonated complexes with TAMS and TAPS, and no stable complex was observed with TRNS, reinforcing the importance of water and/or hydroxide as a ligand in aqueous Al(III) chemistry. TACS exists in the wrong conformation for metal ion complexation, and, as a result, exhibits exceedingly slow complexation kinetics and weaker binding. 2.5 References 1) Harris, W. R.; Pecoraro, V. L. Biochemistry 1983, 22, 292. 2) Harris, W. R.; Sheldon, J. Inorg. Chem. 1990, 29, 119. 3) Harris, W. R.; Chen, Y. C.; Wein, K. Inorg. Chem. 1994, 33, 4991. 4) Mathias, C. J.; Sun, Y.; Welch, M. J.; Green, M. A.; Thomas, J. A.; Wade, K. R.; Martell, A. E. Nucl. Med. Biol. 1988,15, 69. 5) Motekaitis, R. J.; Martell, A. E.; Welch, M. J. Inorg. Chem. 1990, 29, 1463. 6) Doll, R. Age and Ageing 1993, 22, 138. 7) Martell, A. E.; Motekaitis, R. J.; Smith, R. M. Polyhedron 1990, 9, 171. 8) Clevette, D. J.; Orvig, C. Polyhedron 1990, 9, 151. 72 9) Martell, A. E.; Smith, R. M . Critical Stability Constants; Plenum: New York:, 1974-1989; Vol. 1-6. 10) Martell, A. E.; Hancock, R. D. Metal Complexes in Aqueous Solution; Plenum: New York, 1996. 11) Cabbiness, D. K.; Margerum, D. W. /. Am. Chern. Soc. 1969, 92, 6540. 12) Lindoy, L. F. The Chemistry of Macrocyclic Ligand Complexes; Cambridge University Press: Cambridge, 1989. 13) Green, M . A.; Welch, M. J.; Huffman, J. C. /. Am. Chern. Soc. 1984, 206, 3689. 14) Green, M. A.; Welch, M. J.; Mathias, C. J.; Fox, K. A. A.; Knabb, R. M.; Huffman, J. C. /. Nucl. Med. 1985, 26, 170. 15) Green, M. A. /. Labelled Comp. Radiopharm. 1986, 23, 1227. 16) Evans, D. F.; Jakubovic, D. A. Polyhedron 1988, 7, 1881. 17) Evans, D. F.; Jakubovic, D. A. /. Chern. Soc. Dalton Trans. 1988, 2927. 18) Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1992, 32, 5400. 19) Cook, D. F.; Cummins, D.; McKenzie, E. D. /. Chern. Soc. Dalton Trans. 1976,1369. 20) Liu, S.; Wong, E.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1993, 32, 4268. 21) Liu, S.; Wong, E.; Karunaratne, V.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1993,32,1756. 22) Bollinger, J. E.; Mague, J. T.; Roundhill, D. M. Inorg. Chern. 1994, 33, 1241. 23) Bollinger, J. E.; Mague, J. T.; O'Connor, C. J.; Banks, W. A.; Roundhill, D. M . /. Chern. Soc. Dalton Trans. 1995, 1677. 24) Bollinger, J. E.; Mague, J. T.; Banks, W. A.; Kastin, A. J.; Roundhill, D. M . Inorg. Chern. 1995, 34, 2143. 73 25) Botsivali, M.; Evans, D. F.; Missen, P. H.; Upton, M . W. J.Chem. Soc. Dalton Trans. 1985, 1147. 26) Henrick, K.; McPartlin, M.; Munjoma, S.; Owston, P. G.; Peters, R.; Sangokoya, S. A.; Tasker, P. A. / . Chem. Soc. Dalton Trans. 1982, 225. 27) Fleisher, E. B.; Gebala, A. E.; Levey, A.; Tasker, P. A. /. Org. Chem. 1971, 36,3042. 28) Glasoe, P. K.; Long, F. A. /. Phys. Chem. 1960, 64, 188. 29) Gran, G. Acta Chem. Scand. 1950, 4, 559. 30) Motekaitis, R. J.; Martell, A. E. Can. }. Chem. 1982, 60, 2403. 31) Baes, C. F. Jr.; Mesmer, R. E. Hydrolysis of Cations; Wiley-Interscience: New York, 1976. 32) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification of Organic Compounds; 4 ed.; Wiley: New York, 1981. 33) Abraham, R. J.; Gatti, G. J. /. Chem. Soc. B 1969, 961. 34) Childers, R. F.; Wentworth, R. A. D.; Zompa, L. J. Inorg. Chem. 1971,10, 302. 35) Ivery, M.; Weiler, L. S. Unpublished results . 36) Caravan, P.; Hedlund, T.; Liu, S.; Sjoberg, S.; Orvig, C. /. Am. Chem. Soc. 1995, 227,11230. 37) Akitt, J. W. Prog. NMR. Spectros. 1989, 22, 1. 38) Nelson, W. O. Ph.D. thesis; University of British Columbia, 1988. 39) Cole, E.; Copley, R. C. B.; Howard, J. A. K.; Parker, D.; Ferguson, G.; Gallagher, J. F.; Kaitner, B.; Harrison, A.; Royle, L. /. Chem. Soc. Dalton Trans. 1994,1619. 40) Mulla, F.; Marsicano, F.; Nakani, B. S.; Hancock, R. D. Inorg. Chem. 1985, 24,3076. 74 41) Wong, E.; Caravan, P.; Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1 9 9 6 , 35,715. 42) Clarke, E. T.; Martell, A. E. Inorg. Chim. Acta 1 9 9 1 , 186, 103. 43) Motekaitis, R. J.; Sun, Y.; Martell, A. E. Inorg. Chim. Acta 1992,198-200, 421. 44) Clarke, E. T.; Martell, A. E. Inorg. Chim. Acta 1991,181, 273. 45) Anderson, C. J.; John, C. S.; Li, Y. J.; Hancock, R. D.; McCarthy, T. J.; Martell, A. E.; Welch, M . J. Nucl. Med. Biol. 1 9 9 5 , 2 2 , 165. 46) Li, Y.; Martell, A. E.; Hancock, R. D.; Riebenspies, J. H.; Anderson, C. J.; Welch, M . J. Inorg. Chern. 1 9 9 6 , 35, 404. 47) Ma, R.; Welch, M. J.; Reibenspies, J.; Martell, A. E. Inorg. Chim. Acta 1995,236,75. 48) Sun, Y.; Motekaitis, R. J.; Martell, A. E.; Welch, M . J. Inorg. Chim. Acta 1995,228, 77. 49) Sun, Y.; Anderson, C. J.; Pajeau, T. S.; Reichert, D. E.; Hancock, R. D.; Motekaitis, R. J.; Martell, A. E.; Welch, M. J. /. Med. Chern. 1 9 9 6 , 39, 458. 50) Hancock, R. D.; Cukrowski, I.; Cukrowska, E.; Hosken, G. D.; Iccharam, V.; Brechbiel, M . W.; Gansow, O. A. /. Chern. Soc. Dalton Trans. 1 9 9 4 , 2679. 51) Sabatini, A.; Vacca, A. /. Chern. Soc. Dalton Trans. 1 9 8 0 , 519. 52) Fabbrizzi, L.; Micheloni, M.; Paoletti, P. /. Chern. Soc. Dalton Trans. 1 9 8 0 , 1055. 53) de Angelis, S.; Batsanov, A.; Norman, T. J.; Parker, D.; Senanyake, K.; Vepsalainen, J. /. Chern. Soc, Chern. Commun. 1 9 9 5 , 2361. 75 Chapter 3: Aqueous Solution Studies of Tripodal Aminophenolate Ligands with the Lanthanides 3.1 Introduction There has been a proliferation in the chemistry of multidentate ligands with lanthanide(III) ions (Ln(III)) in the past decade. Much of this has been spurred by the application of Gd-dtpa1 (dtpa = diethylenetriaminepentaacetate) and Gd-dota2 (dota = l,4/7/10-tetraazacyclododecane-N,N,,N"/N",-tetraacetate) and their derivatives as magnetic resonance imaging (MRI) contrast agents.3 The similarity in size of Ln(III) and Ca(II) has lead to various lanthanides being substituted for Ca(U) in biological systems with the physical properties of the lanthanide being used as a probe,4"6 e.g. Eu(III) luminescence,7'8 Gd(III) as an EPR probe,9 and various paramagnetic lanthanides as NMR shift probes.10"12 This flurry of activity has been a boon to those interested in studying lanthanide coordination chemistry. The work done on multidentate ligand complexation of lanthanides for MRI has greatly increased the knowledge base in the kinetic and thermodynamic stability of multidentate metal complexes, and the exploitation of the physical properties of the lanthanides as probes has expanded the arsenal of techniques with which to study metal complexes. The aqueous chemistry of Ln(III) ions is dominated by their oxophilicity,1 3 , 1 4 with most ligands studied containing carboxylato or phosphonato oxygen donors. Amines alone are poor donors for lanthanides in water; however, when incorporated into aminocarboxylates or aminophosphonates, they form quite stable Ln(UI) complexes, much more stable than the analogous oxo-acids. This can be ascribed to the initial coordination of the anionic oxygen which disrupts the hydration sphere and serves to anchor the ligand while the amine moiety coordinates secondarily.13 Excluding catecholato derivatives,15'16 there has been little attention paid to the phenolato group as an 76 anionic oxygen donor for lanthanides. Ligands such as bis(2-hydroxybenzyl)ethylenediamine diacetic acid (HBED) and 5-sulfosalicylic 1 S acid have been studied, but these contain carboxylate groups which greatly facilitate phenol complexation. Coordination of lone phenols is hindered by the weak acidity of the hydroxyl group (pKa of phenol ~ 10),19 and since the aqueous lanthanide ions are themselves weak Lewis acids which precipitate as hydroxides above pH 8, they generally cannot compete with the hydronium ion for complexation of phenols in water. Hence a ligand consisting of only amines and phenols would not be expected to be a particularly good ligand for aqueous lanthanide chelation because it would simply be too basic. a H 3TRN: Ri = R 2 = H H 3trac H3TRNCI: R1 = H; R 2 = CI H3TRNBr: R^ = H; R 2 = Br H3TRNOMe: R, = OMe; R 2 = H H 3TRNS3 _ : R^ = H; R 2 = S03" Scheme 3.1 N 4 0 3 Pro-ligands Orvig and coworkers have, however, recently prepared a number of lanthanide complexes of tren based amine phenols (Scheme 3.1) and have discovered three different coordination geometries (Scheme 3.2) with coordination numbers of 9,8 and 6, which have been described as capped,2' 77 encapsulated dimer, and bicapped, respectively. The tripodal Schiff tribase ligand trac3" forms a 1:1 seven coordinate encapsulated complex with Yb . 2 3 encapsu la ted dinner b icapped Scheme 3.2 Although all preparations were in nonaqueous solvents, the capped and bicapped complexes were stable (at least kinetically) in weakly acidic, neutral and weakly basic aqueous solutions (the encapsulated Schiff base complex hydrolyzed, and the encapsulated dimer was insoluble). As the capped 78 complexes were prepared by stoichiometric metathesis of lanthanide nitrate with the neutral form of the ligand (no added base), it appeared that the nitrogen atoms were acting as an internal base source. The encapsulated dimer was prepared by adding an excess of hydroxide. Based on these results it was of interest to see which complexes, if any, were present in aqueous solution, and if the coordination geometries were pH dependent. Furthermore, could phenolato coordination provide an anchoring effect which would enable the amino nitrogen atoms to coordinate in water? It has been recognized that the stability trends for lanthanides with various ligands fall into two categories:24 1) one in which the stability increases with increasing charge to ionic radius ratio, as with edta and 2) another in which the stability increases across the series, reaches a plateau, and then declines, as with dtpa. The enthalpies and entropies of formation show no apparent trends. A plot of AH or AS of complex formation versus atomic number may be S-shaped (AS, hedta), U-shaped (AS, dtpa), or linear (AS, dcta). 1 3' 2 5' 2 6 Often, a change in coordination number (either the degree of inner sphere hydration or a change in the ligand binding mode) has been suggested for the inflections in the curves; however, recent NMR studies have cast doubts on this assumption 2 7 , 2 8 Since the parent amine phenol was insoluble in the pH range 7-11, the sulfonic acid derivative tris(((2-hydroxy-5-sulfobenzyl)amino)ethyl) amine (H6TRNS) was employed, Scheme 3.1. In this chapter the stability constants of this ligand with a variety of lanthanides are determined potentiometrically, the enthalpic and entropic contributions to these stability constants are obtained via calorimetry, and the variable pH solution NMR spectra of these complexes ( 1H, O) are reported. What emerges is an unexpected picture of a ligand which has an unprecedented selectivity for heavier lanthanides over lighter rare earths and which has an increasing tendency to favour bicapped complexation with the 79 heavier lanthanides, so much so that K 2 > Kj. These results and the overall trends will be discussed with respect to ionic radius and solvation. In a further effort to gain some insight into the aqueous chemistry of HfcTRNS with the lanthanides, the aqueous lanthanide coordination chemistry of two other smaller tripodal pro-ligands, HgTAMS and H^TAPS (see Chapter 2, Scheme 2.3), was also investigated. There has been no structural chemistry reported for either the Ln - TAMS or Ln - TAPS systems. However, as was seen in Chapter 2, variations in the number of potential donor atoms, the number of chelate rings formed upon coordination, and the size of the chelate rings formed (5- or 6-membered rings) can have a profound effect upon metal ion selectivity and coordination geometry. HgTAMS and ^ T A P S react with Ln(III) ions in the presence of base to form complexes wherein all 6 donor atoms of the ligand (i.e. N3O3 coordination) coordinate to the lanthanide ion. This change in coordination mode relative to HgTRNS also results in a lower selectivity for heavy lanthanide chelation. 3.2 Experimental Materials. H 6TRNS, H 6 TAMS, and H 6 TAPS were prepared as in Chapter 2. Sodium deuteroxide (NaOD, 40%), deuterium chloride (DC1,12M) and the lanthanide atomic absorption standards were obtained from Aldrich. Hydrated lanthanide nitrates were obtained from Alfa. Deuterium oxide (D20) was purchased from Isotec. All were used without further purification. Instrumentation. *H NMR spectra (200 and 300 MHz) were referenced to DSS and recorded on Bruker AC-200E and Varian XL 300 spectrometers, respectively. 1 3 C NMR (75.5 MHz, referenced to DSS, standard pulse sequence), natural abundance 1 7 0 NMR (40.7 MHz, referenced to H 2 0 , zero delay time), and 1 3 9 L a NMR (42.4 MHz, zero delay time, referenced to 0.1 M La(C104) in 1 M 80 HCIO4) spectra were recorded on the latter instrument. Infrared spectra were obtained as KBr disks in the range 4000 - 400 cm"1 on a Bomem MB-100 spectrophotometer, and were referenced to polystyrene. Analyses of C, H , and N were performed by Mr. Peter Borda in this department. Preparation of Metal Complexes. Metal complexes were prepared in situ by mixing stoichiometric amounts of ligand, metal and hydroxide solutions. 1 H NMR (200 MHz, D 2 0) for [Lu(TRNS)2]3": 2.40,2.70 (AA'BB' coupling pattern, 24H, ethylenic CH 2 ); 3.88,4.44 (AB doublet of doublets, 12H, benzylic C H 2 , 2 J = 14 Hz); 5.98 (d, 6H, H(3), 3 J •= 8 Hz); 7.62 (d of d, 6H, H(4), 3 J = 8 Hz, 4 J = 1 Hz); 7.70 (d, 6H, H(6), 4 J = 1 Hz). [La(TRNS)]: X H NMR (200 MHz, D 20): 2.8 (broad, 12H, ethylenic CH 2 ) , 4.2 (broad, 6H, benzylic CH 2 ) , 6.81 (d, 3H, H(3)), 7.70 (m, 6H, H(4) and H(6)). Long needles were obtained by allowing the solution to stand, however none proved suitable for x-ray diffraction. Anal. Calc. (found) for C 2 7 H 3 6 L a N 4 0 1 2 S 3 « 1 1 . 5 H 2 0 : C, 30.95 (30.92); H , 5.39 (5.34); N, 5.35 (5.36). IR (cm"1, KBr disk): 3500-2600 (b s, v N . H / O-H)/ 1 5 9 5 > U 8 0 > 1 2 9 0 (s> *t=c)/ 1 1 6 0 (vs, vs=o)-The : H NMR and 1 3 C NMR spectra of [Lu(TAPS)]3" in D 2 0 at pD 9 showed a series of broad overlapping resonances characteristic of fluxional behaviour. The % NMR spectra of [Lu(TAMS)]3", [La(TAMS)]3", and [La(TAPS)] were similar to those of the free ligand suggesting fast exchange. A 1 3 9 L a NMR study of [La(TAPS)]3" as a function of pH showed one resonance at 0 ppm, and the linewidth of this resonance increased with pH: pH 7.16 (2250 Hz), pH 7.31 (3400 Hz), p H 7.46 (3600 Hz) NMR Measurements. The variable pH 1 H NMR spectra of the complexes were run in D 2 0 with the pD values being measured by a Fisher Accumet 950 pH meter employing an Accumet Ag/AgCl combination microelectrode. The 81 pD values were converted to pH by adding 0.40 to the observed reading. The variable pH O NMR spectra were run in H 2 0 and were referenced to 0.16 M NaCl. The Dy(III) experiments were recorded at 21°C while the spectra for the Gd(IJJ) titration were run at 25.0 + 0.3 °C. With the 1 7 0 NMR, a spectral window of 1000 Hz, with a 90° pulse width of 18 Lis, and an acquisition time of 0.256 s was usually employed; this gave 512 data points. 2000 transients were collected per spectrum. The O linewidths for H 2 0 were about 60 Hz. Concentrations employed ranged from 1 to 40 mM. The dysprosium induced shifts (Dy.I.S) were obtained from the observed shift by making a correction for the bulk magnetic susceptibility of the solution.30 Stock solutions were prepared from metal nitrates in D 2 0 (H 20) and the metal-ligand solutions were prepared by pipetting required amounts of stock solution and adjusting pH with acid or base. Dy(H 3TRNS) 2 was prepared by adding a fivefold excess of Na 3 H 3 TRNS to Dy(N0 3) 3 . In the case of equilibrium measurements, the ionic strength was controlled by addition of NaCl. Potentiometric Equilibrium Measurements. The procedure was the same as employed in Chapter 2. The measurements were made at 25.0 ± 0.1°, |i, = 0.16 M NaCl. The pK as of the pro-ligands were checked whenever a different synthetic batch of pro-ligand was used, and fresh pro-ligand solutions were always employed. The lanthanide solutions were prepared by dilution of the appropriate atomic absorption standards. Since the lanthanides do not hydrolyze below pH 6, the excess acid in the solutions could be obtained by titrating with standard NaOH and analyzing for the strong acid by the method of Gran. 3 1 The ratio of pro-ligand to metal used was 1:2 < L:M < 4:1. Concentrations were in the range 0.5 - 2.5 mM. A minimum of five titrations were performed for each metal. The metal-H6TRNS solutions were titrated to just beyond three 82 equivalents NaOH/H 6 TRNS (about pH 7) as the complexes were found to undergo slow hydrolysis beyond this point (vide infra). In the case of La where the 1:1 complex precipitates, data points corresponding to it < 0.8 (n = moles of bound L / moles of M) were used. The metal - H 6 TAPS and metal - H 6 T A M S solutions were titrated to just beyond six equivalents NaOH /HsTAPS (HgTAMS), again because of slow hydrolysis beyond this point. Although complexation was rapid (1-3 min per point to give a stable pH reading), care was taken to ensure that no trace hydrolysis or precipitation was occurring by monitoring up to 30 minutes for pH drift. Calorimetry. The automated titration calorimeter used was essentially the o n o o same as described previously ' with some of the instrumentation modernized so that it is controlled by a PC. The gold reaction vessel had a volume of 104 mL and the volume was kept constant during the titrations by removing lmL of solution and then adding lmL of titrant via automatic burettes (Metrohm Dosimat and Hamilton, respectively). The initial solutions typically consisted of 2 mM H 6TRNS, 1 - 2 mM L n 3 + , and 3 mM H + in an ionic medium of 0.16M NaCl. The solutions were titrated with a standardized carbonate free NaOH solution, with each titration consisting of 18 data points. The calorimeter was calibrated twice for each titration, once before any base was added and again after the final addition of base. The calibration consisted of supplying a known quantity of electrical heat and recording the thermocouple reading in the solution. Eleven calibration points were measured, which spanned the heat range produced during the actual titration, and a linear calibration was obtained. The ligand titrations in the presence, and absence, of Ln(III) ions were repeated two or three times, and in some cases the H 6TRNS : Ln(III) ratios were varied. 83 Computations - Potentiometry. The protonation constants for the lanthanide-ligand stability constants were determined by using the program BEST 3 4 as outlined in Chapter 2. H 6TRNS can be considered a hexaprotic acid and its 6 pKas were determined directly by potentiometry (Chapter 2); however, in the case of the lanthanide complexes, the ligand coordinated in a tridentate fashion and was treated as a triprotic acid in the computations. By relating the equilibrium constants to H3TRNS 3 " , as in equations 1-3, the values of the equilibrium constants become more meaningful. L n 3 + + H3TRNS 3 - - [Ln(H3TRNS)] (1) Kj = p! = {[Ln(H 3TRNS)]}/{Ln 3 +}«{H 3TRNS 3-} [Ln(H3TRNS)] + H 3TRNS 3" « [Ln(H3TRNS)2]3" (2) K 2 = {[Ln(H 3TRNS) 2] 3-}/{[Ln(H 3TRNS)]}«{H 3TRNS 3-} L n 3 + + 2 H 3TRNS 3" « [Ln(H3TRNS)2]3" (3) P 2 = { [Ln(H 3 TRNS) 2 ] 3 l / {Ln 3 + }«{H 3 TRNS 3 - } 2 HgTRNS could be treated as a triprotic acid because the lanthanide titrations never reached a pH that was high enough for ammonium deprotonation. The validity of this assumption was tested by including L(OH) and ML 2(OH) species in the calculation. These did not improve the fit of the rt o r * data. The Ln(OH) species were also included in the calculation, but these do not form in any appreciable amount. The best fit of the data consisted simply of the two metal-ligand species [Ln(H3TRNS)] and [Ln(H3TRNS)2]3". Protonated metal complexes were tried but these did not improve the fit. 84 H 6 TAMS and H 6 TAPS, on the other hand, both reacted with Ln(UI) to coordinate as hexadentate ligands, liberating six equivalents of acid per ligand. The equilibrium conventions used in Chapter 2 apply to these two systems, i.e. L n 3 + + TAMS 6" - [Ln(TAMS)]3" (4) K L n ( T A M S ) = {[Ln(TAMS)]3"}/{Ln3+}.{TAMS6"} L n 3 + + H + + TAMS 6" - [Ln(HTAMS)]2" (5) K L n ( H T A M S ) = {[Ln(HTAMS)]3"}/{Ln3+}.{H+}*{TAMS6"} L n 3 + + TAPS6" * [Ln(TAPS)]3" (6) K L n ( T A P S ) = {[Ln(TAPS)]3"}/{Ln3+}.{TAPS6"} L n 3 + + H + + TAPS6" - [Ln(HTAPS)]2" (7) K L n ( H T A P S ) = {[Ln(HTAPS)]3"}/{Ln3+}.{H+}.{TAPS6"} Typically 100 data points were collected with about 80-90% of the points being in the buffer region of metal-ligand complexation, with the remainder in the strong acid region being used as a check of excess acid concentration. Computations - Calorimetry. A version of the program SOLGASWATER was used to calculate the molar amounts of each species present at each data point based on the equilibrium constants, the initial concentrations of Ln(III), H 6 TRNS, and acid, the base concentration, and the volumes added and removed. The following equilibria comprised the observed heat change: 85 H + + OH" - H 2 0 (8) H 3 T R N S 3 - + H + - H 4 T R N S 2 - (9) H 4 TRNS 2 _ + H + « H5TRNS" (10) H 5TRNS" + H + - H 6 TRNS (11) H 3 T R N S 3 " + L n 3 + •> [Ln(H3TRNS)] (12) H 3 T R N S 3 - + LnL •» [Ln(H3TRNS)2]3" (13) As the titrations began in an acidic medium, the heat measured for the first few data points was solely because of equation (8) and the heat of dilution of the NaOH titrant. Correcting for the heat of dilution and using only these initial points, the heat of neutralization at |i = 0.16M NaCl was obtained. This value of -55.92 kj'mol , which was in agreement with the data of Grenthe and coworkers,38 was then fixed for the other calculations. In the case of the H 6 TRNS only titrations, eqs. 8-11 defined the system. The observed heat, Q 0 b s , can be expressed as the sum of the contributions from each equilibrium, Q o b s = An 8 AH 8 + An 9 AH 9 + A n 1 0 A H 1 0 + A n n A H n (14) where An is the change in the number of moles of a given species after an addition of base. Since An can be calculated, and A H 8 is known, the other three molar heats can be solved by linear algebra. Similarly for the metal-ligand titrations, AH for the ligand protonations was fixed, and A H calculated for the 86 metal complexes. The unknowns were solved using matrices with the aid of the Matlab suite of programs.39 The largest source of error would arise from an error in concentration or in the stability constants used to calculate the relative amounts of species. To estimate the error, the heats were calculated using concentrations that were purposefully incorrect by 5% and stability constants incorrect by 0.2 log units; however, since the calculation involves changes in species concentrations, in the buffer region this will not differ greatly. The error was estimated to be + 0.3 kjmol" . 3.3 Results Formation Constants. Figure 3.1 (top) shows the experimental titration curves for the lanthanides studied at a ratio of 2 HgTRNS : 1 Ln. The flatness of the curves (except for La) is indicative of a predominant 2 :1 complex. This is also shown by an n plot, n = moles of bound ligand per mole of metal ion, (Figure 3.1, bottom); a typical stepwise equilibrium in which K4 » K 2 would show a plateau at fi = 1, whereas in this system, n rises directly to two. In the plot of n vs log [H3TRNS ] (Figure 3.1, bottom), all the data for a given lanthanide ion coincide to a single curve independent of the ligand or metal concentrations or ratios proving that only mononuclear complexes are formed.4 0 If the curves curled back upon themselves, or if at different ratios the curves did not overlap, polynuclear complex formation or protonated metal complex formation, respectively, would be indicated. Furthermore, at ratios less than 2 : 1, there was rapid hydrolysis after a = 3 (a = mol OH"/mol ligand), whereas with a ratio of 2 :1 or greater, hydrolysis was quite slow, which suggests a stable 2 :1 complex. Analysis of the potentiometric data gave the formation constants listed in Table 3.1. Figure 3.1 (bottom) shows the agreement of the calculated constants with the experimental data in the n plot. It was found that the second formation 87 constant was greater than the first (K 2 > K )^ for the heavier lanthanides, with the crossover point coming at neodymium. Because there is only a small amount of 1:1 complex formed, the error associated with Kj and hence K 2 is larger than that for p 2. For lanthanum, only the 1:1 constant could be determined because the 1:1 complex precipitated. With Nd(III), if the ratio of H 6 TRNS : Nd(III) was less than 1.3, a hydrolysis/polymerization took place at a ~ 1.5, most likely to a slightly soluble 1:1 complex. Table 3.1. Thermodynamic Parameters for the L n 3 + - H 3 TRNS 3 " Complexation Equilibria.a , b L n 3 + log K x logK 2 log (32 AH1 A H 2 A H p 2 ASX AS 2 ASp2 La 5.65 — — — — — — — — Nd 6.41 6.34 12.75 -20.34 +6.63 -13.71 54 143 197 Gd 6.67 7.69 14.36 -12.11 -15.15 -27.26 88 97 185 Ho 7.67 8.75 16.42 -21.33 -14.06 -35.39 74 121 195 Yb 8.53 9.73 18.26 -23.30 -21.81 -45.11 85 113 198 Estimated errors: log K l x log K 2 ± 0.1; log ± 0.03; A H n ± 0.3; AS n ± 3. bUnits: AH (kjmor1); AS (JK^mor 1 ) 88 Figure 3.1. Top: Experimental Lanthanide Titration Curves at 2 mM H 6TRNS : 1 mM Ln(III). Bottom: Experimental Plots of n vs. log [H3TRNS ] (Symbols) and Curves Generated from the Calculated Stability Constants (Solid Lines). 89 The titrations of H 6 TAMS and H 6 TAPS in the presence of Ln(IU) ions were much different. Figure 3.2 shows experimental titration curves for the lanthanides with H 6 TAMS (top) and H 6 TAPS (bottom) at a ratio of 2 mM Ln(III) : 2 mM pro-ligand. The curves now show plateaus which extend up to a = 6 indicating that the ligands are coordinating in a hexadentate fashion, coordinating through all six ionizable donor atoms. This is further emphasized in the n plots (Figure 3.3). The n curves rise to n = 1 and then plateau even in the experiments with excess ligand. Analysis of the potentiometric data gave the stability constants listed in Table 3.2. It was necessary to include protonated species to the model to improve the fit of the Table 3.2. Log Formation Constants for Ln(HI) with TAMS 6" and TAPS6" at 25 °C, it = 0.16 M NaCl. Ln(III) TAMS 6" TAPS6" ML/M»L HML/ML»H ML/M»L H M L / M L * H La 9.17(1) — 11.33 (3) 7.14 (2) Nd 11.19 (6) — 13.59 (3) 6.54 (3) Gd 11.86 (9) 6.55 (9) 14.50 (1) 6.38 (4) Ho 12.71 (10) 6.69 (4) 14.71 (4) 6.44 (9) Yb 13.78 (1) 6.33 (3) 15.15 (3) 6.39 (4) data, although these only form to a small extent (maximum ~ 25% of total Ln(III)). Both TAMS 6" and TAPS6" are selective for the heavier lanthanides, but much less so than H 3TRNS 3"' 90 a Figure 3.2. Experimental Lanthanide Titration Curves at 2 mM H 6 T A M S : 2 mM Ln(ni), Top, and 2 mM H 6 TAPS : 2 mM Ln(III), Bottom. 91 1.2-1 Figure 3.3. Experimental (Symbols) Plots of n vs. log [TAMS 6 -] (Top) and n vs. log [TAPS6"] (Bottom). The Solid Lines are Generated from the Calculated Stability Constants, K L n ( T A M S ) and K L n ( T A P S ) . 92 Calorimetry. The calorimetric results for HgTRNS showed values consistent with those measured for other phenols/ and are listed in Table 3.3. Table 3.3. Thermodynamic Values for the First Three Deprotonations of H 6TRNS. Equilibrium pK a AH (kjmol'1) AS ( J K ^ m o r 1 ) [H3TRNS 3-] [H+] / [H4TRNS 2-] 8.07 (3) [H4TRNS 2-] [H+] / [H5TRNS"] 7.29 (3) [H5TRNS-][H+]/[H6TRNS] 6.17(3) 24.8 (3) 27.6 (3) 30.7 (3) 35 (3) 36 (3) 62 (3) The thermodynamic measurements on the Ln - H3TRNS system are summarized in Table 3.2. The heats of formation were all exothermic, except that for the Nd(TRNS)2 complex, which was slightly endothermic. The heats of formation of the 1:1 complexes were more exothermic than those of the 2:1 complexes, except for gadolinium. The entropies for the second equilibrium (AS2) were greater than those of the first (AS1) but this difference decreased across the series. Overall (i.e. in terms of p2)/ as the lanthanide series was traversed from Nd(IU) to Yb(IJI), the entropy term was fairly static at about 195 1 1 J K " mol , while the enthalpic term became more and more exothermic. Multinuclear NMR ( 1 H, 1 3 C , 1 3 9 La) The % NMR spectrum of HgTRNS was typical of that of the other amine phenols based on tren . 4 1 , 2 2 The threefold symmetry of the pro-ligand yields only six of a possible eighteen 1 H resonances. Upon complexation to Lu(III), this threefold symmetry is retained and the molecule became rigid. The rigidity is apparent in that the benzylic resonance splits from a singlet in the free ligand to an AB quartet in the complex (Figure 3.4,1:1 and 2 :1 spectra). Likewise, the two triplets representing the 93 ethylenic backbone hydrogen atoms split upon complexation to give an AA'BB' system. The three ring hydrogen atoms (assigned by the magnitude of their coupling constants, J o r ^ 0 > ]meta > ]para) a ^ s o s r iift upon coordination; however, the hydrogen ortho to the phenolate oxy group undergoes the greatest 3:1 - i 1 1 I 900 650 600 750 Hz Figure 3.4. *H NMR Spectrum (200 MHz) of the Benzylic Region for Various H3TRNS3": L u 3 + Ratios. 94 coordination induced shift. In the H NMR of the Lu complex as a function of ligand added (at a = 3), the spectrum was unchanged up to H3TRNS : Lu = 2 after which free ligand peaks appeared (Fig. 3.4), corroborating the 2 :1 stoichiometry determined by potentiometry. The La complex also had C3 symmetry but was not as rigid, the hydrogen atoms in the backbone displaying fluxionality. The 1:1 complex precipitated, obviating quantitative study of this fluxionality. As predicted from the potentiometric data, free ligand peaks appear in the La-TRNS 1 H NMR spectra at H 3 T R N S 3 " : L a 3 + ratios greater than one. The X H NMR spectra of the Yb(UI), Ho(III), Gd(III), and Nd(IU) complexes all showed broad unassignable resonances; however, the spectra in all cases showed evidence of 2:1 complex formation. At ratios greater than 2:1, free ligand resonances appeared. The spectra also showed rigid species which did not change upon heating to 70 °C. The 1 H NMR spectra of the soluble [Ln(H3TRNS)2]3" complexes run at a > 3 showed the resonances from the complex being slowly replaced by those of the free ligand. There was no indication of different complexes being formed (involving amine coordination, for instance), although one instance of a possible different N4O3 isomer was observed. The spectrum of a pH 12,2 H 3 T R N S 3 " : 1 Lu showed about 95% free ligand, but there was also a C 3 symmetric species present. In the capped and bicapped complexes, the hydrogen atom ortho to the phenolate was shifted to lower frequency. The minor species has this resonance shifted to higher frequency, which is analogous to the In(III) complex which has the metal encapsulated.41 This minor isomer is a kinetic product which eventually gives way to lutetium hydroxide, and was not further apprehended. Solution NMR studies on the Ln(III) - H 6 TAMS and Ln (III) - H 6 TAPS systems were less revealing. The 1 H NMR and 1 3 C NMR spectra of [Lu(TAPS)]3" in D 2 0 at pD 9 showed a series of broad overlapping resonances characteristic of 95 fluxional behaviour. The X H NMR spectra of [Lu(TAMS)]3-, [La(TAMS)]3-, and [La(TAPS)]3- were similar to those of the free ligand suggesting fast exchange. A 1 3 9 L a NMR study of 30 mM La(III): 30 mM H 6 TAPS as a function of pH showed only one resonance at 0 ppm, the chemical shift of La( a q ) 3 + . The linewidth of this resonance increased with pH suggesting that [La(TAPS)] is in exchange with La(aq) • 1 7 0 N M R The natural abundance 1 7 0 NMR spectrum of water in the presence of a lanthanide ion and ligand gives a qualitative picture of complexation. A plot of lanthanide induced shift (LIS) versus pH showed complexation occurring as the ligand becomes deprotonated. As the number of metal sites for water exchange decreased with complexation, so did the LIS. 1 7 Figure 3.5 (top) shows a plot of the Gd(III) induced O shift of water versus p H in a titration of 4 mM HsTRNS : 2 mM Gd(III), roughly the same conditions employed in the potentiometric titrations. The bottom half of Figure 3.5 is the speciation of Gd(III) under these conditions, calculated based on the stability constants determined in this work. Peters and coworkers have exploited the dysprosium induced shift of water (Dy.I.S.) to estimate quantitatively the number of bound water molecules AO A'X 9ft associated with various lanthanide complexes. ' ' The Dy.I.S. of water was measured at varying dysprosium concentrations. The plot of Dy.I.S. versus [Dy(III)] was linear with a slope of -358 ppm/M (Figure 3.6). It had been previously established that the contact contribution of a paramagnetic Ln(III)-induced shift of a Ln(III)-bound O nucleus is almost independent of the nature of the probed O-containing ligand in question and of other co-ligands coordinated to the lanthanide.4 3 , 4 4 Since the 1 7 0 shift is predominantly contact in nature, the slope of a plot of Dy.I.S versus [Dy(III)] should be proportional to the number of bound water molecules associated with the complex. If the 96 hydration would be: number of Dy(III) is taken as eight, then a slope of -358/8 = -45 indicative of one bound water and each multiple of 45 corresponds to Figure 3 . 5 . Plot of Gd.I.S vs. pH for 4mM TRNS : 2mM Gd(III) (Top), and the Relevant Speciation Diagram Calculated from the Equilibrium Constants Shown in Table 3.1 (Bottom). 97 one water. Figure 3.6 shows the Dy.I.S. versus [Dy(III)] for Dy(a q) , [Dy(TAMS)]3-, [Dy(TAPS)]3", and [Dy(H 3TRNS)2] 3 _. The slope of -358 ppm/M for Dy( a q ) 3 + is in excellent agreement with that obtained by Alpoim et al.42 (-357 ppm/M) and by Reuben and Fiat 4 6 (-360 ppm/M). The error bars show the linewidths at half height (60 Hz), however, the precision was + 5 Hz. All four plots were linear with correlation coefficients of greater than 0.999. The slope of the plot for [Dy(H 3TRNS) 2] 3 _ was -50 ppm/M, or 1.1 H 2 0 , very close to the value reported for [Dy(dtpa)]2", -52 ppm/M under the same conditions.42 The slopes for [Dy(TAMS)]3- and [Dy(TAPS)]3" were -128 ppm/M (2.8 H zO) and 123 ppm/M (2.7 H 2 0) , respectively. 0 10 20 30 [Dy] (mM) Figure 3.6. Plot of Dy.I.S. vs. [Dy(III)] (mM) for D y ( a q ) 3 + , A , [Dy(TAMS)]3", • , [Dy(TAPS)]3", • , and [Dy(H3TRNS)2]3', O . Error Bars Represent Linewidths at Half Height. 98 The low concentration of the 1:1 complex, [Dy(H3TRNS)], made a confident estimation of the number of bound waters associated with it impossible. The 1:1 species will always exist with large amounts of either the 2 : 1 species, or the aquo ion, and will be pH dependent. Since the majority of Q AT) aqueous lanthanide complexes are 9 coordinate in solution, ' and since a 9 0 crystal structure of an analogous 1:1 complex has been reported, it is likely that the coordination number of the 1:1 complex reported here is 9 with 6 associated water molecules. Studies with H 3 T R N S 3 " : Dy(ffl) ratios as high as 25 showed the same Dy.I.S.: [Dy(III)] ratio, indicating that the maximum number of ligands bound is two. Similarly, ratios of TAMS 6": Dy(III) and TAPS 6": Dy(III) as high as 8 showed a limiting stoichiometry of 1 TAMS 6 ": 1 Dy(III) and 1 TAPS 6": 1 Dy(IH). 3.4 Discussion In previous studies on tripodal N4O3 ligands with lanthanide(III) ions, the 9 1 ease of formation of the monocapped complexes was noticed. Simply mixing the free base form of the pro-ligand with a lanthanide salt yielded the 1:1 capped complex. With the tren based amine phenol, addition of base gave a 2 : 2 encapsulated dimer. This dimerization can be prevented by employing the same ligand with methoxy substituents in the 3 positions; however, reaction of this ligand with lanthanide salts in the presence of excess NaOH resulted in the 9 9 formation of bicapped 6-coordinate complexes. In light of the current solution chemistry results, this reaction can be explained by the relative inertness of the bicapped complex, coupled with its high formation constant. The 1:1 capped complex disproportionated in the presence of base to give Ln(OH) 3, which was filtered out, and the bicapped 2 :1 complex which was isolated (Scheme 3.3). 99 2 capped bicapped Scheme 3.3 The work described here supports a solution structure for the 2 :1 bicapped complex, [Ln(H 3TRNS)2]3-, that is similar to that determined by X-ray OO crystallography in the 3'-methoxy substituted Gd(III) complex. From the proton stoichiometry, only three donor atoms were deprotonated, and these donors were shown by variable pH 1 H NMR and UV spectroscopies to be all phenolato oxygen donors. The H NMR of the [Lu(H3TRNS)2] complex shows a rigid complex possessing C 3 symmetry, which rules out a non-apical N / 0 3 donor set per ligand. The 1 7 0 NMR studies on the [Dy(H3TRNS)2]3" complex showed it to contain one inner sphere coordinated water molecule, hence a coordination number of 7 was obtained in solution. The [Gd(TRNOMe)2j complex22 was obtained from a nonaqueous solvent and contained methoxy substituents ortho to the phenolate oxygen atoms which probably contributed to 1 7 its low coordination number (6). The O NMR study of the bicapped dysprosium complex, [Dy(H3TRNS)2] , indicated one bound water (the slope of the Dy.I.S. plot being the same as that of the dtpa complex which had been AT ASK established to have one inner sphere water molecule based on Tb*' and Eu*° fluorescence), but the H NMR of the Lu complex is consistent with a bicapped structure owing to the threefold symmetry. To account for the H and O 1 0 0 H N - H J H - — N H / H - N H results, a seven coordinate solution structure similar to the bicapped crystal structure is proposed in which there is one water molecule in the equatorial plane. Since the rate of water exchange in Ln complexes is very fast, 4 9 , 5 0 , 5 1 the water molecule will be averaged about the equatorial plane on the NMR time scale, thus preserving the threefold symmetry seen in the Lu(III) complex. 3 . Less can be said about the 1:1 complex, [Ln(H3TRNS)], since it was formed in such small amounts. Variable pH and variable M : L ratio 1 H NMR studies on the Lu(III) system, showed only one set of resonances for the complex, suggesting that the 1:1 complex had the same structure as the 2 :1 complex, and that removal of one capping ligand did not effect the chemical shift of the remaining bound ligand, or that the bound ligand in the 1:1 complex is in fast exchange with the free ligand. The coordination number could not be determined for the ( H - , - N H H N - H J H — N H -N 1:1 complex. The small percentage relative to either the aquo ion or the 2 :1 complex coupled with its pH dependence made an accurate determination of the 1 7 hydration number of the capped complex by O NMR impossible. A variable pH luminescence study on the europium or terbium system may prove sensitive enough to ascertain the coordination number of the capped complex. However, the proton stoichiometry places the ligand denticity at 3 and the likely coordination number is 9 (6 bound waters) in accord with the crystal structure of the related [Gd(H3trac)(N03)3].20 The high formation constant of [Ln(H3TRNS)2]3" relative to that for [Ln(H3TRNS)] was surprising, and no examples could be found in the lanthanide literature. It is a well known fact of coordination chemistry that the formation 101 constants for stepwise equilibria generally decrease with increasing substitution due to statistical, electrostatic, and steric effects, but exceptions to this general rule have been known for a long time. Most notable perhaps is Bjerrum's work on silver ammine complexes where K 2 > 5 2 , 5 3 Other instances include K 4 > K 3 in the Hg(II)-Cl equilibria19 and K 3 > K 2 in the Fe(II)-phenanthroline equilibria.19 The Fe(II)-phen anomaly is attributed to the gain in free energy from the change in electronic configuration on going from Fe(phen)2 (high spin) to Fe(phen)3 (low spin), which should be an enthalpic effect. The other anomalies can also be attributed to enthalpy changes from reduction of hydration number associated with the formation of highly stable linear (Ag(I), Hg(II)) complexes. For instance, HgCl 2 is a linear species, whereas [HgCl 3(H 20)] _ and HgCL/' are both tetrahedral. Thus, on going from HgCl 2 to [HgCl3(H20)]" a molecule of solvent is required, while the fourth addition of chloride to give HgCi4 displaces this bound water. In the Ag(I)-ammine system, the first equivalent of ammonia displaces one water from a tetrahedral A g + ion, while the second displaces three waters to give the linear [Ag(NH 3) 2] +. Presumably, the change in hybridization from sp3 to sp results in the anomalous order observed.54 With a lowering of coordination number (more bound H 2 0 liberated), a positive entropy change should accompany complex formation; however, in these instances the entropy change is zero or negative, and the increase in stability can be traced to the enthalpy term. This low, or negative, entropy change which occurs despite the net increase in the number of molecules in the reaction: [Ag(NH 3)(H 20) 3] + + N H 3 . [Ag(NH 3) 2] + + 3H 2 0 (15) emphasizes the fact that thermodynamic quantities are measured globally and cannot always be rationalized locally. 102 In the H3TRNS 3 " : Ln(III) system, the anomaly is likely to be predominantly entropic. It is well known that lanthanide(III) bonding is primarily electrostatic in nature. In the H3TRNS 3 " : Ln(III) system, a tripositive lanthanide reacts with a trinegative ligand to give a neutral 1:1 complex, which then reacts with a second trinegative ligand to give a trinegative complex. There should be no favourable enthalpy associated with K 2 (relative to K4) based on electrostatic arguments, and even less so if the coordination number decreases. The coordination number of lanthanide(III) aquo ions has been the subject of numerous investigations, with eight or nine being the generally accepted number, 4 9 , 4 5 with a change in coordination number from 9 for the light lanthanides to 8 for the heavy lanthanides.55 Based on a crystal structure of an analogous compound and the preponderance of nine coordinate lanthanide complexes in solution and the solid state, the 1:1 complex should have a coordination number 9, [Ln(H3TRNS)(H20)3]. The 2 :1 dysprosium complex, [Dy(H3TRNS)2(H20)]3" has a coordination number of seven, based on 1 7 0 NMR shifts of water. A similar bicapped complex had been obtained which was six 99 coordinate; however, it contained methoxy groups ortho to the phenolate donors which may have forced such a low coordination number on the lanthanides. It appears a change in coordination number is occurring as the equilibrium shifts from monocapped to bicapped. For the equilibria considered here, the first equivalent of H3TRNS 3 " displaces 3 waters, while the second equivalent should displace 5 waters. This second equilibrium should increase the translational entropy of the system more than the first, and this will be seen in the large K 2 . This is most likely a steric effect which arises from placing six phenolato groups about the metal and crowding out any further coordination sites. 103 The argument is supported by the calorimetric measurements, which show AS 2 > ASj for each Ln(III) under consideration (Figure 3.7). The difference (AS2 - ASi) decreases across the series, consistent with the fact that the first equilibrium involves two triply charged species coming together to give a neutral 1:1 complex which should result in a positive entropic contribution. The second equilibrium involves no such net change in charge; therefore, one would expect ASi to increase across the series as the effective nuclear charge increases. The overall entropic effect, ASp2, is fairly static across the series studied (Fig. 3.7), which is to be expected if the degree of solvation of the 2 :1 complex is constant across the lanthanide series. Figure 3.7. Enthalpies and Entropies for the Ln(III) - H 3 TRNS 3 _ Equilibria: Nd(III), • , Gd(III), • , Ho(III), ITl , Yb(III), m . 104 An alternative argument for the anomalous K 2 > Ki effect is based upon the hydrophobic effect.56,57 Consider the solvation of a gaseous hydrocarbon in water at 25 °C. This process involves a small negative enthalpy of solvation, but a larger negative entropy of solvation; it is thermodynamically disfavoured because of entropy.57 The aggregation of apolar solutes is then driven by entropy such that the water molecules avoid entropically unfavourable interactions with the apolar solute molecules. The ion H 3 TRNS can be thought of as an amphiphilic ion with charged polar regions and three apolar aryl rings. Describing the two equilibria, Ki and K 2 , pictorially as in Figure 3.8 leads to a hydrophobic interpretation of the two complexation reactions. The areas shaded in grey represent the hydrophobic aryl portions of the molecules. In the first step (Ki) one ion with a hydrophobic region combines with a lanthanide ion to give a neutral molecule with a hydrophobic region. The second step (K2) is the combination of an ion and a neutral, each with a hydrophobic region, combining to give an ion with only one hydrophobic region. This minimization of solvent (H 20) accessible hydrophobic regions, or "a tightening of the hydrophobic belt", should be reflected in a more positive entropy for K 2 relative to Kj, as is observed. Both steps are also enthalpically favoured by the formation of Ln - O (phenolate) bonds. backbone with three phenylphosphinate groups attached. At low p H (1.5), (shown below) with Al(III), Ga(III), and In(III) in water. H 3ppma is a zwitterion that is structurally very similar to comes from recent work by Lowe et air on the equilibria of the ligand H 3ppma Further evidence for this effect H 3ppma H 3TRNS 3". It has an N-methylated tren 105 Viewed in Terms of the Hydrophobic Effect. H 3ppma reacts with Al(III), 03(111), and In(III) to form capped and bicapped complexes by coordinating to the metal through the phosphinato oxygen atoms. Again, the second stepwise equilibrium constant is greater than the first. Since the aquo ions of Al(III), Ga(III), and In(III) are known to be six coordinate and the bicapped complexes contain octahedral ions, the argument presented above for an inner sphere desolvation and lowering of coordination number does not apply. However this anomalous behaviour can be rationalized by the hydrophobic effect as shown in Figure 3.8. The second surprising phenomenon was the high selectivity that H 3 TRNS displayed for the heavier lanthanides, clearly evident in Figure 3.1 106 which shows the Yb complex forming at a pH one unit lower than the Nd complex. A change in the overall stability constant P2 of 105'5 was observed on going from Nd(III) to Yb(JJI). This compares with 10 2 9 for EDTA, 10 1 9 for NTA, or 10 1 0 for D T P A . 1 9 The calorimetric data for AHp 2 showed an increasing exothermicity on traversing the lanthanide series (Fig. 3.7), while ASp2 was fairly static. This result is in accord with an electrostatic bonding model, in which the formation constants should increase with effective nuclear charge (or 1/r) 5 9 as they do here (Fig. 3.9). This selectivity may prove useful in the separation of lanthanides by chelation ion chromatography.60 When H3TRNS 3 " binds to a lanthanide(III) in a tridentate fashion, three sixteen membered chelate rings are formed. There should be no chelate effect on forming such large rings; however, the three protonated amine functionalities in 10-, 00 o 5 -La Nd Gd Ho Yb 1/r ( A " 1 ) Figure 3.9. log K n v s 1/r (Ionic Radii for C N = 6, ref. 59): K x ( # ) , K 2 ( O ). 107 this ligand make it different from a tris(phenol) linked by nine carbon chains. The elevated pKas of the amines coupled with the depressed pKas of the phenols (relative to their molecular substituents, cf. tren and 4-sulfophenol) imply that an intrastrand hydrogen-ion bond exists between the protonated N atom of the amine, its acidic H atom, and the deprotonated phenolato oxygen atom (as has been seen in several solid state structures ' ' ' ) . Interstrand H-bonding may also be postulated if only because of coordination to the metal. In Chapter 2 H -bonding within H 5 T A P S " was shown to be strong. Given the relatively high stability of these capped 16-membered ring complexes, there must be an effect which predisposes the ligand to a binding posture. In a related ligand crystal structure interstrand H-bonding, as well as intrastrand H-bonding was noted.6 1 This interstrand H-bonding, if present, is certainly not expected to be strong and there will likely be several local energy minima (the ligand 1 H NMR spectra show C 3 symmetry). The flexibility imparted by a loose H-bond network coupled with the large chelate ring size results in a tridentate ligand which should have little or no strain energy created in accommodating different Ln(IU) ions; thus the increase in stability is purely electrostatic and increases with the inverse ionic radius of the lanthanide considered (Fig. 3.9). Although the results obtained for overall formation (i.e. (32) of the 2 :1 complex can explained in terms of a 7-coordinated species, the ASi and AHj (and hence the AS 2 and AH 2) data show no apparent trends (Fig. 3.7). If there is a change in coordination number in the 1:1 complex because of the steric requirements of the ligand (e.g. 5 bound H 2 0 for the late lanthanides, 6 for the early lanthanides), then one would expect to see two trends in ASi (and A H ^ as the Ln(III) series is traversed. Figure 3.9 shows a linear relationship between log K 2 and 1/r (rz = 0.999), whereas there appears to be a discontinuity in the log K4 vs 1/r data at Gd. The lack of smooth trends in the enthalpic and entropic data is 108 consistent with the aqueous calorimetric literature on the lanthanides. ' ' Whether this can be interpreted as a change in inner sphere coordination appears dubious, since a recent calorimetric and multinuclear NMR study demonstrated that a series of Ln complexes with constant geometry and constant coordination number can also give rise to this phenomenon of a discontinuity in the enthalpic n o and entropic terms. In order to further explore the effect of large chelate ring size on Ln(III) selectivity, solution studies with LI^TAMS and E^TAPS were instigated. If these two pro-ligands reacted in the same manner as H^TRNS, then lanthanide complexes containing 14-membered and 13-membered chelate rings would be formed. Instead of coordinating solely through the phenolato donor atoms, TAMS 6 ' and TAPS 6 - coordinated through the three amino nitrogen and three phenolato oxygen donor atoms. A major difference between HgTRNS and HsTAMS or H 6 TAPS is the microscopic order of deprotonation. In Chapter 2 it was shown that the first three deprotonation events of H 6 TRNS occur at phenolic sites, whereas H 6 TAMS and H 6 TAPS are firstly deprotonated at an ammonium site, followed by three phenol sites, and then the remaining two ammonium groups. The first deprotonation of HgTAMS and H 6 TAPS occurs at a p H much lower than that at which Ln(III) complexation occurs. Hence coordination to this amino group should be facile. Coordination to one amino group would necessarily bring the remaining ammonium groups closer to the metal ion to allow for proton displacement and lanthanide coordination to give the observed N3O3 ligand donor set. The 1 7 0 NMR study of both [Dy(TAMS)]3" and [Dy(TAPS)]3" indicated the presence of three inner sphere water molecules, which implies a 9-coordinate Dy(III) in each of the complexes. 109 20-, Figure 3.10. Comparative Binding Affinities of TAMS 6", TAPS6", and H 3 T R N S 3 " for Ln(ni): log fa ([Ln(TAMS)]3" • ; log fa ([Ln(TAPS)]3" ^ ; log fi n ([Ln(H3TRNS)2]3" • ; n = 1 for La(III), n = 2 for Nd(in), Gd(III), Ho(III), Yb(III). This change in coordination mode from H 3 T R N S 3 " to TAMS 6" and TAPS6" has a profound effect on the metal ion selectivity. This is readily appreciated in Figure 3.10. There is a large increase in stability upon going from La(IU) to Nd(III) for all three ligands. However, on going from Nd(III) to Yb(III), H 3 T R N S 3 " exhibits a selectivity of about 2 log units per lanthanide studied. TAMS 6" exhibits lesser selectivity, about one log unit per lanthanide studied, whereas TAPS 6" has a much lower selectivity between Gd(III) and Yb(III). Figure 3.10 is a little misleading when representing the relative affinities of the ligands studied for a given lanthanide, since the competition with hydrogen ion for the ligand is not considered. A better way of analyzing the data is to calculate pM values where pM = -log [Mf r e e]. This gives an impression of the relative sequestering ability of the ligands under a standard set of conditions. In Figure 3.11, pM values are calculated at p H 7.4 for a ligand to 110 metal ratio of 10 :1. The total concentration of Ln(III) is set at 1 mM; however since the stability constants for H 3TRNS have an inverse square dependence on [H3TRNS 3 "], the pM values for [Ln] t o t = lp:M have also been calculated to highlight this dilution effect. At millimolar concentrations and above, H3TRNS is the best ligand for complexing Nd(III) —» Yb(III), and its sequestering ability increases with atomic number, Z. The much flatter curve for TAPS6" indicates that it is less able to discriminate between the lanthanides. The major structural difference between TAPS6" and TAMS 6" is that TAPS6" coordinates to a lanthanide forming four 6- and two 5-membered chelate rings, whereas TAMS 6" forms only 6-membered chelate rings upon coordination. An established tenet of coordination chemistry is that 5-membered chelate rings are more stable than 6-membered chelate rings and this difference in stablity increases with increasing metal ion size.62 This effect is manifested here where Figure 3.11. Comparative pM Values vs Z for H3TRNS 3 " ([Ln(ffl)]tot = 1 mM), ([Ln(III)]tot = 1 uM), - 0 - , TAPS6" ([Ln(III)]tot = 1 mM), TAMS 6" ([Ln(III)]tot= 1 mM), I l l pH Figure 3.12. Speciation of Yb(UI) in the Presence of H 6 TAMS (Top), H 6 TAPS (Middle), and H 6TRNS (Bottom); [Yb(III)]tot = 1 mM, [ligand]tot = 2 mM. 112 [Ln(TAPS)] complexes are 1 - 2 orders of magnitude more stable than the analogous [Ln(TAMS)]3" complexes. As a final visual aid into the comparative behaviour of the three ligands treated in this chapter, the speciation of Yb(III) with each ligand is shown in Figure 3.12. This further emphasizes the order of stability among the ligands: H 3 T R N S 3 " > TAPS6" > TAMS 6 ", as well as emphasizes the minor significance of protonated complexes. 3.5 Conclusions. The tripodal aminophenolate ligand H3TRNS 3 " acts as a tridentate ligand toward lanthanide ions in aqueous solution, binding through the phenolato oxygen atoms to form large 16-membered chelate rings. The formation constant for the 2 :1 complex (K2) is larger than that for the 1:1 complex ( K 4 ) , and this is believed to be an entropic effect based on the desolvation of the 1:1 complex because of steric crowding of the aryl rings about the metal centre, an effect which also results in a minimization of hydrophobic contacts with water. H 3 T R N S 3 " is highly selective for the heavier lanthanides and this has been determined to be an enthalpic effect which agrees well with the electrostatic nature of Ln(III) bonding. The ligands TAPS6" and TAMS 6" coordinate to Ln(IH) in a hexadentate fashion through three amino nitrogen and three phenolato oxygen donors. The ability of TAPS6" and TAMS 6" to coordinate through the amino groups may be a consequence of the low pK a of one of the ammonium moities in these compounds. The selectivity for the late lanthanides by TAPS6" and TAMS 6" is considerably muted compared to H3TRNS 3 ". 1 7 0 NMR established that the coordination number for [Dy(H3TRNS)2(H20)]3" is seven 113 (07), for [Dy(TAMS)(H20)3]3- is nine (N 30 6), and for [Dy(TAPS)(H20)3]3- is nine (N 30 6). 3.6 References 1) Watson, A. D. /. Alloys Compds. 1994,207/208,14. 2) Kumar, K.; Jin, T.; Wang, X.; Desreux, J. F.; Tweedle, M . F. Inorg. Chem. 1994, 33,3823. 3) Lauffer, R. B. Chem. Rev. 1987,87,901. 4) Martin, R. B.; Richardson, F. S. Q.Rev. Biophys. 1979,12,181. 5) Meares, C. F.; Wensel, T. G. Acc. Chem. Res. 1984,17,202. 6) Lanthanide Probes in Life, Chemical, and Earth Sciences; Biinzli, J.-C. G.; Choppin, G. R., Eds.; Elsevier: Amsterdam, 1989. 7) Biinzli, J.-C. G. Inorg. Chim. Acta 1987, 239,219. 8) Horrocks, W. D. J.; Albin, M. Prog. Inorg. Chem. 1984,32, 1. 9) Stephens, E. M. in ref. 6. 10) Gupta, R. K.; Gupta, P. /. Mag. Reson. 1982,47,344. 11) Pike, M. M.; Springer, C. S. /. Mag. Reson. 1982,46,348. 12) Sherry, A. D.; Geraldes, C. F. G. C.; Cacheris, W. P. Inorg. Chim. Acta 1987, 139,137. 13) Choppin, G. R. Pure Appl. Chem. 1971,27,23. 14) Moeller, T. The Lanthanides; Moeller, T., Ed.; Pergamon: Oxford, 1973; Vol. 4, p i . 15) Zhu, D.-H.; Kappel, M. J.; Raymond, K. N. Inorg. Chim. Acta 1988, 247,115. 16) Freeman, G. E.; Raymond, K. N. Inorg. Chem. 1985,24,1410. 17) Ma, R.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1994,224,151. 18) Lajunen, L. H. J. Finn. Chem. Lett. 1976,36. 114 19) Martell, A. E.; Smith, R. M . Critical Stability Constants; Plenum: New York:, 1974-1989; Vol. 1-6.. 20) Smith, A.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1988,27,3929. 21) Liu, S.; Gelmini, L.; Rettig, S. J.; Thompson, R. C ; Orvig, C. /. Am. Chern. Soc. 1992,114,6081. 22) Liu, S.; Yang, L.-W.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1993,32, 2773. 23) Berg, D. J.; Rettig, S. J.; Orvig, C. /. Am. Chern. Soc. 1991,113,2528. 24) Tse, P.-K.; Powell, J. E. Inorg. Chern. 1985,24,2727. 25) Anderegg, G. Coordination Chemistry; Martell, A.E., Ed.; Van Nostrand, Reinhold: New York, 1971; Vol. 1,427. 26) Ashcroft, S. J.; Mortimer, C. T. Thermochemistry of Transition Metal Complexes; Academic Press: London, 1970. 27) Peters, J. A. Inorg. Chern. 1988,27,4686. 28) Huskens, J.; Peters, J. A.; van Bekkum, H.; Choppin, G. R. Inorg. Chern. 1995, 34,1756. 29) Glasoe, P. K.; Long, F. A. /. Phys. Chern. 1960, 64, 188. 30) Bertini, I.; Luchinat, C. NMR of Paramagnetic Molecules in Biological Systems; Benjamin/Cummings: Menlo Park, 1986; Vol. 3. 31) Gran, G. Acta Chern. Scand. 1950,4,559. 32) Ginstrup, O. Chern. Scr. 1973,3,97. 33) Danielsson, D.; Ginstrup, O.; Ingri, N. Chern. Scr. 1973,3, 81. 34) Motekaitis, R. J.; Martell, A. E. Can. }. Chern. 1982, 60,2403. 35) Baes, C. F. Jr.; Mesmer, R. E. Hydrolysis of Cations; Wiley-Interscience: New York, 1976. 36) Eriksson, G. Anal. Chim. Acta 1979, 212,375. 37) Rossini, F. D. National Bureau of Standards Circular No. 500; U.S. Government Printing Office: Washington, 1952. 115 38) Grenthe, I.; Ots, H.; Ginstrup, O. Acta Chern. Scand. 1979, A24,1067. 39) MATLAB; 4.0; The MathWorks Inc., 1993. 40) Rossotti, H . The Study of Ionic Equilibria; Longman: London, 1978. 41) Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1992,32,5400. 42) Alpoim, M. C.; Urbano, A. M.; Geraldes, C. F. G. C.; Peters, J. A. /. Chern. Soc. Dalton Trans. 1992,463. 43) Huskens, J.; Kennedy, A. D.; van Bekkum, H.; Peters, J. /. Am. Chern. Soc. 1995,227,375. 44) Peters, J. A.; Kieboom, A. P. G. Reel. Trav. Chim. Pays-Bas 1983, 202,381. 45) Helm, L.; Foglia, F.; Kowall, T.; Merbach, A. E. /. Phys.: Condens. Matter 1994, 6, A137. 46) Reuben, J.; Fiat, D. /. Chern. Phys. 1969,52,4909. 47) Chang, C. A.; Brittain, H. G.; Telser, J.; Tweedle, M . F. Inorg. Chern. 1990,29, 4468. 48) Bryden, C. C ; Reilley, C. N. Anal. Chern. 1982,54,610. 49) Cossy, C ; Helm, L.; Merbach, A. E. Inorg. Chern. 1988,27,1973. 50) Micskei, K.; Helm, L.; Brucher, E.; Merbach, A. E. Inorg. Chern. 1993,32, 3844. 51) Powell, D. H.; Gonzalez, G.; Tissieres, V.; Micskei, K.; Brucher, E.; Helm, L.; Merbach, A. E. /. Alloys Compd. 1994,207/208,20. 52) Bjerrum, J. Metal Ammine Formation in Aqueous Solution; P. Haase and Son: Copenhagen, 1957. 53) Maeda, M.; Arnek, R.; Biedermann, G. /. Inorg. Nucl. Chern. 1979,41,343. 54) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; 4th ed.; Wiley: New York, 1980. 55) Cossy, C ; Helm, L.; Powell, D. H.; Merbach, A. E. New }. Chern. 1995, 29,27. 116 56) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological Membranes; John Wiley & Sons: New York, 1973. 57) Blokzijl, W.; Engberts, J. B. F. N. Angew. Chem. Int. Ed. Engl. 1993,32,1545. 58) Lowe, M . P.; Rettig, S. J.; Orvig, C. /. Am. Chem. Soc. 1996, In Press. 59) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. 60) Kumar, M. Analyst 1994,119, 2013. 61) Liu, S.; Wong, E.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1993, 32,4268. 62) Hancock, R. D.; Martell, A. E. Chem. Rev. 1989, 89,1875. 117 Chapter 4: On the Effect of Pyridyl Donors in the Chelation of Al(III), Ga(III), and In(III). 4.1 Introduction The ethylenediamine unit represents an ideal fragment with which to construct a multidentate ligand. The diamino moiety allows facile functionalization to incorporate other donor atoms. Martell and Hancock have pointed out that the close proximity of the donor atoms maximizes the number of 5- and 6-membered chelate rings formed upon metal ion complexation, and this, in turn, maximizes the entropic contribution to a high thermodynamic stability.1 As a result, a number of ligands have been prepared offering a variety of donor atom sets (some of these2"13 are displayed in Chart 4.1). Accordingly, a rich coordination chemistry has developed, and there exists a body of structural and thermodynamic - n O "Q o^o-bped edta X = CH,R = H: HBED X = CH,R = s0 3 " : SHBED X = N,R = CH 3: ENDA-HP R=H: bbpen R = s0 3 " : Sbbpen 118 Chart 4.1 information for a variety of metal complexes of this type. This chapter focuses on the chemistry of the group 13 metal ions, Al(III), Ga(III), and In(III) with these types of multidentate ligands. The aqueous chemistry of aluminum(III) is dominated by the oxophilicity of the ion, a consequence of which is coordination by anionic oxygen donors and/or water. There is little known about the affinity of Al(III) for hexadentate ligands containing donors other than anionic oxygen, outside of edta and its derivatives. The reason for this is two-fold: firstly, substitution kinetics with Al(III) are sluggish and long equilibration times (hours) are required to obtain reliable complex formation constants; secondly, the affinity of Al(III) toward the neutral nitrogen donor is known to be low. 1 4 By anchoring a ligand via strong oxygen-aluminum bonds and exploiting the chelate effect, neutral nitrogen donors can be forced to coordinate 1 5 - 1 7 and aluminum complex formation constants may be determined in order to gauge the effect of neutral nitrogen donor coordination. The aqueous chemistry of Ga(III) and In(III) is much richer than that of Al(III) owing to a higher affinity of these two ions for softer donor types, such as neutral nitrogen or anionic sulfur. Faster ligand exchange kinetics allows for easy determination of complex formation constants by automated potentiometric procedures. Because of the rapid proliferation of 6 7 G a , 6 8 G a and m I n in nuclear medicine,19 there has been widespread interest in the coordination chemistry of Ga(III) and In(III) with multidentate ligands. The Ga(III) or In(III) chelates must be kinetically and/or thermodynamically stable with respect to hydrolysis and demetallation by the serum protein transferrin. 2 0 , 2 1 One means of assessing these criteria is to determine the metal complex stability constant. This provides a thermodynamic basis for the in vivo stability of the intact chelate and also provides insight into factors concerning better ligand design. 119 A remarkable reversal of stability occurs when two of the carboxylate groups on edta are substituted by two 2-oxybenzyl groups to give HBED (Chart 4.1). The affinity of In(III) for edta is four orders of magnitude higher than that for Ga(III) (log K M L = 24.9 (In), 21.0 (Ga))22, while [Ga(HBED)]" is almost 10 orders of magnitude more stable than [In(HBED)]" (log K M L = 27.8 (In), 38.5 (Ga)).6 Further replacement of the remaining two carboxylate groups by methylpyridyl moieties to give Sbbpen (the sulfonato groups are added for aqueous solubility) almost completely mutes any selectivity for Ga(III) or In(III) (log K M L = 34.9 (In), 35.3 (Ga)).4 Three factors involved in this comparison are: the donor atom type, the size of the metal chelate ring formed upon coordination, and the overall coordination number of the metal ion. The first two are obvious; the effect of the third is less so. Hancock has argued that ligands containing five membered chelate rings coordinate with less strain to larger metal ions.2 3 For hexadentate ligands such as edta which contain five 5-membered chelate rings upon coordination, larger metal ions such as In(III) may adopt a coordination number of seven (by coordinating a water molecule); the longer M-O and M-N bond lengths in a seven coordinate complex impart less strain in the ligand. 2 4 Smaller ions such as Al(III) or Ga(III) which cannot adopt a higher coordination number of seven may alleviate some ligand strain by having one of the donor atoms replaced by solvent. In this case a direct comparison of ligands coordinating to Al(III), Ga(III), and In(III) becomes more complex. In an effort to understand the factors governing metal ion selectivity among group 13 metal complexes, the ligand N,N'-bis(2-methylpyridyl)ethylenediamine-N,N'-diacetic acid (H 2bped) 7 has been prepared in improved yield. Its deprotonation constants have been determined and assigned, and the formation constants of this ligand with 120 Al(III), Ga(III), and In(III) have been measured. In addition, the Co(III) complex has been prepared and its structure determined by X-ray crystallography. The Co(III) complex is used as an aid in the determination of the solution structures of the group 13 metal chelates. 4.2 Experimental Materials. Ethylenediamine-N,N'-diacetic acid, 2-(chloromethyl)pyridine hydrochloride, gallium(III) nitrate nonahydrate, indium(III) nitrate pentahydrate, cobalt(II) chloride, sodium perchlorate, sodium hexafluorophosphate, sodium deuteroxide (NaOD, 40%), deuterium chloride (DO, 12M) and the aluminum, gallium, and indium atomic absorption standards were obtained from Aldrich. Cetyltrimethylammonium bromide and sodium acetate were obtained from BDH, and aluminum trichloride hexahydrate from Anachemia. Deuterium oxide (D 20) was purchased from Isotec. All were used without further purification. Instrumentation. 1 H NMR (200 and 300 MHz) and 1 3 C NMR (50.3 and 75.5 MHz) spectra were referenced to DSS and recorded on Bruker AC-200E and Varian XL300 spectrometers, respectively. 2 7 A1 NMR spectra were recorded at 78.2 MHz on the Varian XL300 spectrometer and are referenced to 0.1 M A1C13 in 1 M H O . % nOe difference spectra were recorded on a Bruker WH-400 spectrometer. The 1H-detected heteronuclear multiple quantum coherence (HMQC) and 1H-detected multiple bond heteronuclear multiple quantum coherence (HMBC) experiments were run on a Bruker AMX-500 spectrometer. Mass spectra were obtained with a Kratos Concept IIH32Q (Cs+, LSIMS) instrument. UV-vis spectra were recorded on a Shimadzu UV-2100 spectrophotometer. Infrared spectra were obtained as KBr disks in the range 4000 - 400 cm"1 on a Galaxy Series FTIR 5000 spectrophotometer, and were 121 referenced to polystyrene. Analyses of C, H , and N were performed by Mr. Peter Borda in this department. The relationship pD = p H + 0.4025 was used to relate the acid concentration of the NMR samples (pD) to the potentiometric results (pH). N,N'-Bis(2-methylpyridyl)ethylenediamine-N,N'-diacetic acid dihydrochloride (H2bped»2HCl). To a yellow suspension of 2-(chloromethyl)pyridine hydrochloride (16.40 g, 0.1 mol), ethylenediamine-N,N'-diacetic acid (8.81 g, 0.05 mol), and hexadecyltrimethylammonium bromide (1.80 g, 0.005 mol) in water (400 mL) was added sodium hydroxide pellets until a pH of 12 was obtained. Addition of base gave a red orange solution. The solution was stirred at room temperature under nitrogen for 36 hr. During this time, the solution was maintained at pH 11 by occasional (over 4 hr) addition of sodium hydroxide pellets. Upon completion of the reaction, the reaction mixture was passed down a Rexyn 101 cation exchange column ( H + form) and eluted with deionized distilled water until the eluent was no longer acidic. The solvent was removed from the eluate under reduced pressure to give an orange solid. Hot ethanol (200 mL) was added and any insoluble material filtered off. The orange ethanol solution was acidified with 6M HC1, and the solvent removed under reduced pressure. The resultant solid was taken up in a minimum amount of hot ethanol (300 mL) and acetone was added to precipitate a white powder. The powder was collected by vacuum filtration, washed with cold ethanol (2 x 10 mL) and dried in vacuo at 60 °C for 24 hr. If excess acid is used, the compound may be obtained as the trihydrochloride salt. Yield: 16.45 g (76%). Anal. Calc. (found) for C 1 8 H 2 2 N 4 0 4 . 2 H C 1 : C, 50.12 (49.94); H , 5.61 (5.66); N, 12.99 (12.67). Mass spectrum (+LSIMS): m/z = 359 ([M+l]+, [ C 1 8 H 2 3 N 4 0 4 ] + ) . IR (cm - 1, KBr disk): 3500-2500 (b s, v 0 _ H , v N _ H ) , 1728 (s, v c = 0 ) , 1632 (m, 8 N .H), 1612,1548,1521,1463 122 (ms, v c = c , v c = N ) , 1409 (m), 1388 (m), 1227 (bs), 1161 (s), 765 (m). UV (kmax nm, (e M ' W 1 ) : pH=2: 259 (10900); pH=10: 261 (7400). [Co(bped)]PF6. To a solution of H 2 bped«2HCl (210 mg, 0.49 mmol)) in 20 mL of methanol was added 65 mg of C0CI2 (0.50 mmol) to yield a mauve suspension. Sodium acetate (170 mg, 2.07 mmol) was then added whereupon the suspension gave way to a red brown solution. Air was bubbled through the solution overnight and the color changed to a cherry red. A solution of tetra(n-propyl)ammonium hexafluorophosphate (180 mg, 0.54 mmol) in 10 mL of methanol was added to the red solution, and a red precipitate formed after allowing the solution to stand overnight at room temperature. The precipitate was recrystallized from acetonitrile and dried in vacuo to yield 103 mg (38%). Crystals suitable for X-ray crystallographic analysis were obtained by slow evaporation of a solution of [Co(bped)]PFg in acetonitrile. Anal. Calc. (found) for [CoC 1 8 H 2 0 N 4 O 4 ]PF 6 «0 .5CH 3 CN: C, 39.29 (39.16); H , 3.73 (3.78); N , 10.85 (10.79). Mass spectrum (+LSIMS): m/z = 415 ([M]+, [ C o C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm 4 , KBr disk): 1675 (bs, v c = 0)/1614,1467 (m, v c = c , v c = N ) , 1443 (m), 1351 (m), 1303 (s), 1065 (m), 928 (m), 842 (vs, v P. F), 763 (m), 558 (s, v P . F). [Al(bped)]PF6. To a solution of H 2 bped«2HCl (142 mg, 0.33 mmol)) in 50 mL of methanol was added 82 mg of A 1 0 3 « 6 H 2 0 (0.33 mmol). Sodium acetate (110 mg, 1.32 mmol) was then added to the colorless solution. The solution was refluxed for 30 minutes, allowed to cool to room temperature; sodium hexafluorophosphate (55 mg, 0.33 mmol) was added and the resulting solution was filtered into a 50 mL beaker. After standing overnight at room temperature, a white powder formed. The powder was collected on a frit, washed with methanol (1x5 mL), then acetone (2x5 mL), and dried overnight in vacuo at 65 °C to yield 92 mg (53%). Mass spectrum (+LSIMS): m/z = 383 ([M]+, [ A l C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm"1, KBr disk): 1690 (s, v c = 0 ) , 1650, 123 1618 (s, v c = C / V O N ) , l 4 7 ^ (m), 1442 (m), 1391 (m), 1308 (m), 1161 (m), 937 (s), 844 (vs, V p . F ) , 780 (m), 558 (s, v P. F). [Ga(bped)]PF6. To a solution of H 2 bped»2HCl (189 mg, 0.44 mmol)) in 50 mL of methanol was added 185 mg of G a ( N 0 3 ) 3 « 9 H 2 0 (0.44 mmol). Sodium acetate (145 mg, 1.32 mmol) was then added to the colorless solution. The solution was refluxed for 30 minutes and allowed to cool to room temperature; sodium hexafluorophosphate (74 mg, 0.44 mmol) was added and the resulting solution filtered into a 50 mL beaker. After standing overnight at room temperature, a white powder formed. The powder was collected on a frit, washed with methanol (2x5 mL), then acetone (2x5 mL), and dried overnight in vacuo at 65 °C to yield 115 mg (46%). Anal. Calc. (found) for [GaC 1 8 H 2 0 N 4 O 4 ]PF 6 : C, 37.86 (38.26); H , 3.53 (3.49); N, 9.81 (9.50). Mass spectrum (+LSIMS): m/z = 425 ([M]+, [ G a C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm - 1, KBr disk): 1681 (s, v c = 0 ) , 1643,1616 (s, v c = c , v c = N ) , 1468 (m), 1441 (m), 1345 (m), 1306 (m), 1035 (m), 921 (s), 842 (vs, v P. F), 777 (m), 558 (s, v P. F). [In(bped)]PF6. To a solution of H 2 bped«2HCl (100 mg, 0.23 mmol)) in 20 mL of methanol was added 71 mg of I n C l 3 » 5 H 2 0 (0.23 mmol). Sodium acetate (75 mg, 0.92 mmol) was then added to the colorless solution. The solution was stirred for 20 minutes at room temperature. Sodium hexafluorophosphate (39 mg, 0.23 mmol) was added and the resulting solution filtered into a 50 mL beaker. After standing overnight at room temperature, a white powder formed. The powder was collected on a frit, washed with methanol (1x5 mL), then acetone (2x5 mL), and dried overnight in vacuo at 65 °C to yield 85 mg (59%). Anal. Calc. (found) for [ InC 1 8 H 2 0 N 4 O 4 ]PF 6 .H 2 O: C, 34.09 (34.47); H, 3.50 (3.40); N, 8.83 (8.44). Mass spectrum (+LSIMS): m/z = 471 ([M]+, [ InC 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm 4 , KBr disk): 124 1631 (s, v c = 0 ) , 1609,1561 (s, v c = c , v c = N ) , 1492 (m), 1430 (m), 1382 (m), 1318 (m), 1027 (m), 844 (vs, v P. F), 770 (m), 558 (s, v P. F). X-Ray Crystallographic Analysis of [Co(bped)][PF 6 ]«CH 3 CN»H 2 0. The crystal structure analysis was carried out by Dr. Steven J. Rettig of the Departmental X-ray Facility. Details of the data collection and structure refinement are given elsewhere.26 Selected bond lengths and bond angles appear in Table 4.1. Complete tables of crystallographic data, bond lengths, and bond angles are included in Appendix I. Final atomic coordinates, equivalent isotropic thermal parameters, anisotropic thermal parameters, torsion angles, intermolecular contacts, and least-squares planes are given elsewhere. Table 4.1. Selected Bond Lengths (A) and Angles (deg) for the [Co(bped)]+ Cation in [Co(bped)]PF 6 «CH 3 CN«H 2 0. Co - O l 1.878 (2) Co -03 1.888 (2) Co - N l 1.937 (2) Co -N2 1.941 (2) Co -N3 1.960 (2) Co -N4 1.958 (2) O l -Co- -03 178.53 (8) O l - C o - N l 87.36 (9) O l -Co -N2 92.51 (8) O l - C o - N3 92.34 (8) O l -Co -N4 86.40 (8) 03 - C o - N l 94.10 (9) 03 -Co -N2 87.39 (9) 03 - C o - N3 87.99 (8) 03 -Co -N4 92.14 (8) N l - C o - •N2 88.87 (9) N l -Co -N3 82.14 (9) N l - C o - •N4 169.12 (9) N2 -Co -N3 169.58 (9) N2 - C o - •N4 82.51 (9) N3 -Co -N4 107.01 (9) 125 Potentiometric Equilibrium Measurements. The equilibrium constants were determined by potentiometric methods as described in Chapter 2. The ionic strength was fixed at 0.16 M NaCl, and the solutions were maintained at 25 ± 0.5°C. Argon, which had been passed through 10% NaOH, was bubbled through the solutions to exclude C O 2 . H 2bped was checked for purity by NMR and elemental analysis before titration. Titrations were also employed to verify the molecular weight obtained by elemental analysis. Metal ion solutions were prepared by dilution of the appropriate atomic absorption (AA) standards. The exact amount of acid present in the A A standards was determined by titration of an equimolar solution of metal standard and Na 2H 2edta. The amount of acid present was determined by Gran's method,2 7 and this is equal to the amount of acid in the A A standard plus the two equivalents of acid liberated from the complexed edta. Alternatively, hydrated metal salts (chlorides or nitrates) were dissolved in deionized distilled water and the exact metal ion concentration was determined by passing an aliquot down a Rexyn 101 cation exchange column ( H + form). The liberated acid was determined by titration with sodium hydroxide (one equivalent of M 3 + displaces three equivalents of H + ) . The ratio of ligand to metal used was 1:1 < L:M < 1.5:1. Concentrations were in the range 1.0 - 2.5 mM. Six titrations (three titrations from two batches of ligand) defined the H 2bped system, while five titrations were performed to determine the Ga(III) - H 2bped equilibria, each titration consisting of about one hundred data points. The ligand solutions were titrated over the range 2 < pH < 10, while the metal - ligand solutions were titrated over the range 2 -10.5 Complexation was usually rapid (1-3 min per point to give a stable pH reading) in the Ga(III) and In(III) studies; however, 126 caution was taken to ensure that no trace hydrolysis or precipitation was occurring by monitoring up to 30 minutes for pH drift. The [In(bped)]+ complex was completely formed by p H 2. Potentiometric titrations as described above gave no evidence for a [In(Hbped)]2+ species, but did allow determination of a [In(bped)(OH)] species. The UV spectrum of the binary species [In(bped)]+ remained unchanged until below pH 1 suggesting a lower limit of log K M L = 20.5. Based on this evidence, the tris(aminophenolato) ligand TAPS 6 - was chosen for a ligand-ligand competition study. TAPS 6" forms complexes with In(III) of sufficient stability (log K H M L = 31.93, log KjyiL = 27.56) that ligand exchange should take place above pH 6. A series of 49 solutions was prepared, 2 mM each in H 2bped, H 6 TAPS, and In(III), and varying amounts of NaOH were added to each solution. The solutions were equilibrated for two days and the pH of each solution was recorded. For the Al(III) stability constant studies equilibration was too slow for the automated titration procedure and a batch method was employed instead. Here, a series of 24 (0.16 M NaCl) C02-free solutions was prepared with a 1.1 : 1 L : M ratio and varying amounts of NaOH was added. These solutions were equilibrated for two days at 25 °C until the pH reading of each stabilized. The data were refined using the program BEST. 2 8 The hydrolysis constants used were taken from Baes and Mesmer,2 9 and in the case of In(III), formation constants with chloride were also included in the model. Although Ga(III) was completely complexed at pH 2, the complex was unstable to hydrolysis to [Ga(OH)4]" at pH > 9.5. The [Ga(OH)4]" formation constant was used to obtain the absolute [Ga(bped)]+ stability constant.30 The relative stability constant for equation (1) 127 [In(bped)]+ « [In(bped)(OH)] + H + (1) K M L O H = [In(bped)(OH)][H+]/[In(bped)]+ could be determined by direct titration. The stability constant for the equilibrium, equation (2), In 3 + + bped2" - [In(bped)]+ (2) K M L = ([In(bped)]+)/[In3+][bped2"] was determined by competition with the ligand TAPS6" (charges omitted for simplicity) as shown in equation (3): [In(bped)] + H X TAPS * [In(TAPS)] + H ybped + (x-y)H (3) The calculation involved keeping fixed the stability constants of [In(HTAPS)]2", [In(TAPS)]3", the deprotonation constants of H 6 TAPS and H 2bped, and the relative stability constant ( K M L O H ) / while iteratively varying the constant ( K M L ) from equation (2) until the observed pH agreed with the calculated. 4.3 Results H2bped. N,N'-Bis(2-methylpyridyl)ethylenediamine-N,N'-diacetic acid had been prepared previously by Lacoste et al? as the monosodium salt. The previous method involved attaching a nitrile function to each of the secondary aliphatic amines of N,N'-bis(2-methylpyridyl)ethylenediamine, and then hydrolyzing the nitrile groups to carboxylic acids. The product was obtained in a relatively low yield (22%) and was characterized only by 128 Table 4.2. X H (300 MHz) Spectral Dataa'b for H 2 bped c and its Co(III)d and In(III)e'f Complexes. H2bped [Co(bped)]+ [In(bped)]+ [In(bped)(OH)] H(l) 8.59 (4.4) 8.95 (5.4) 8.73 (5.0) 8.74 (5.2) H(2) 7.76 (4.4, 7.7) 7.84 (5.4, 7.5) 7.77 (5.0, 7.7) 7.71 (5.2, 7.4) H(3) 8.28 (7.7, 7.8) 8.30 (7.5,8.0) 8.22 (7.7, 7.9) 8.17 (7.4, 7.9) H(4) 7.85 (7.8) 7.80 (8.0) 7.70 (7.9) 7.65 (7.9) H(6a) (Pro R) 4.43 4.73 (-15.0) 4.03 (-16.9) 4.04 (-16.0) H(6b) (Pro S) 4.37 (-15.0) 4.45 (-16.9) 4.30 (-16.0) H(7a) (Pro R) 3.34 3.94 (-9.4) 3.05 (-10.2) 2.97 (-10.7) H(7b) (Pro S) 3.56 (-9.4) 3.18 (-10.2) 3.04 (-10.7) H(8a) (Pro R) 3.67 4.00 (-17.9) 3.32 (-16.8) 3.23 (-16.8) H(8b) (Pro S) 3.55 (-17.9) 3.45 (-16.8) 3.34 (-16.8) a For labelling, see Chart 4.1; Numbers in parentheses refer to coupling in Hz (for H(l) = 3 J 1 2 ; H(2) = 3 J 1 2 , 3 J 2 3 ; H(3) = 3 J 2 3 , ^ H(4) = 3 J 3 4 ; H(6) = 2 J a b ; H(7) = 2 J a b ; H(8) = 2 J a b ; c H2bped at pD = 2.3;d [Co(bped)]PF6 in DMSO-rf 6; e [In(bped)]Cl at pD=3.2;f [In(bped)(OH)] at pD = 9.7. elemental analysis and alkimetric titration. The preparation described in this chapter is a one-pot method which gives a much higher yield (76%). H 2bped has been fully characterized by a variety of spectral methods. The 1 H NMR spectrum shows seven resonances with the correct integral ratios (Table 4.2). There are nine resonances observed in the 1 3 C NMR spectrum (Table 4.3). The NMR chemical shifts were assigned on the basis of HMQC, HMBC, and 129 Table 4.3. 1 3 C (75.5 MHz) Spectral Data3 for H 2 bped b and its Co(m),c In(m) / d ' e ' f and Ga(III)g Complexes. H 2 L CoL + InL + InL(OH) [InL]PF6 [GaL]PF6 C(l) 147.52 152.65 151.00 151.29 148.30 (148.40) 148.13 C(2) 129.65 126.68 128.83 128.43 125.45 (125.90) 125.65 C(3) 146.53 141.15 145.52 144.84 141.81 140.22 C(4) 128.91 124.75 128.23 128.02 125.36 (125.27) 124.84 C(5) 152.94 162.77 155.23 155.19 153.45 (153.77) 154.03 Q6) 58.90 64.96 61.05 60.86 57.76 (57.90) 63.26 C(7) 50.03 63.04 51.78 51.74 49.76 (49.66) 55.72 C(8) 57.40 63.27 58.48 58.60 57.26 (57.01) 59.65 C(9) 175.18 178.52 178.57 178.95 170.47 (170.24) 170.41 a For labelling, see Chart 4.1;b L 2 " = bped 2 -, H 2 L at pD = 2.3;c [CoL]PF6 in DMSO-d 6 ; d [InL]Cl at pD=3.2;e [InL(OH)] at pD = 9.7;f [InL]PF6 in DMSO-d 6, numbers in parentheses refer to the minor species; § [GaL]PF6 in DMF-^7 at 120 °C. COSY experiments. Positive LSIMS clearly showed the parent monocation at m/z = 359. The infrared spectrum showed a broad absorption at 3500 - 2500 cm" indicative of extensive hydrogen bonding in the solid. Bands consistent with carboxylate and pyridyl functions were also observed. The UV spectrum was typical of a substituted pyridine and exhibited a hyperchromic effect with decreasing p H . 3 1 Metal Complexes. The group 13 complexes of bped2" were all prepared in a similar manner. These cations, [M(bped)]+, were isolated as either the 130 hexafluorophosphate or perchlorate salts from methanol or ethanol. The Co(III) complex was prepared by the aerial oxidation of the Co(II) complex, followed by precipitation with hexafluorophosphate or perchlorate. All of the metal complexes gave acceptable elemental analyses, with the exception of the Al(III) complex which was also found to be quite hygroscopic. Positive LSIMS mass spectral studies showed strong parent ion peaks for [M(bped)]+ with the expected isotopic pattern. The infrared spectra were similar for all four [M(bped)]+X complexes with strong bands typical of noncoordinated hexafluorophosphate (842, 558 cm"1)3 2 or perchlorate (1092 cm"1).33 The carboxylato asymmetric stretch undergoes a considerable bathochromic shift upon coordination from 1728 cm"1 (H2bped*2HCl) to 1690 cm"1 ([Al(bped)]+) to 1631 cm"1 ([In(bped)]+). The J H and 1 3 C NMR spectra of the Co(III) and In(III) complexes display two-fold symmetry showing only 10 of a possible 20 1 H resonances (Table 4.2) and 9 out of a possible 18 1 3 C resonances (Table 4.3). In addition, the methylene hydrogen atoms on carbons 6 and 8, and the ethylene hydrogen atoms on C(7) are inequivalent indicating rigid solution structures. The 1 H and 1 3 C NMR spectra of the Ga(III) and Al(III) complexes were complicated by fluxional processes and possibly by the presence of additional isomers. Crystals of [Co(bped)]PF6 suitable for X-ray crystallographic analysis were grown by slow evaporation of an acetonitrile solution at room temperature. The complex crystallized with one water molecule and an acetonitrile solvate. An ORTEP diagram of the [Co(bped)]+ cation is shown in Figure 4.1, and the bond lengths and angles in the Co(III) coordination sphere are listed in Table 4.1. The Co(III) ion is hexacoordinated via two neutral pyridyl nitrogen atoms, two neutral tertiary aliphatic amino nitrogen atoms, and two anionic carboxylato oxygen atoms in a distorted octahedral 131 C(9) C(10) Figure 4.1. ORTEP Representation of the Molecular Structure of the [Co(bped)]+ Cation in [Co(bped)]PF 6 »CH 3 CN»H 2 0 (33% probability thermal ellipsoids). 132 coordination geometry. The carboxylato moieties coordinate in a trans fashion with the O l - Co - 03 angle being 178.53 (8)°. The largest deviation from octahedral geometry occurs among the N - Co - N angles. The angle defined by the two pyridyl nitrogen atoms and the cobalt atom, N3 - Co - N4, is opened to 107°, while the angles defining the aliphatic amine - pyridyl chelate rings, NI - Co - N3 and N2 - Co - N4, are compressed to about 82°. The Co - N and Co - O distances are comparable to those found in other chelating aminopyridyl- and aminocarboxylato cobalt(III) complexes.34"40 Potentiometric Titrations. The titration of H 2 bped»2HCl (Figure 4.2) is indicative of two fairly strongly acidic sites and two weakly acidic sites. The pK as determined (at the 3a level) are listed in Table 4.4 along with those reported by Lacoste et al? Variable pH 1 H NMR and UV spectroscopy of the ligand allows assignment of the deprotonation scheme (Figure 4.3). The first deprotonation occurs at a pyridyl nitrogen. The assignment of the second deprotonation is more ambiguous. The largest degree of chemical shift change about pH 3.11 occurs at the methylene hydrogens of the acetate group Table 4.4. Deprotonation Constants of H 2 bped«2HCl (Charges Omitted for Simplicity) at 25 °C. H = 0.16 M NaCl a \i = 0.10 M K N Q 3 b H 3 L . H / H 4 L (pKa 1) 1.5 (2) 2.34 H 2 L . H / H 3 L (pKa 2) 3.11 (2) 3.02 H L . H / H 2 L ( p K a 3 ) 5.53(3) 5.63 L«H/HL (pK a 4) 8.67(8) 8.84 a This work; b ref 7. 133 I 0 1 2 3 4 5 6 7 8 9 10 a Figure 4.2. Titration Curves (pH vs. a; a = mol OH" / mol bped) for H2bped»2HCl in the Presence and Absence of Equimolar (2 mM) Al(III), Ga(IH), and In(III). Table 4.5. Log Formation Constants for Al(III), Ga(III)/ and In(III) with bped2" and edta4". L = bped(2-)a L = edta(4-)b Equilibrium Al(III) Ga(in) In(III) Al(III) Ga(IH) In(III) H M L / M L * H 1.90 (2) 2.5 1.8 1.5 ML/M»L 10.85 (10) 19.89 (10) 22.6 (1) 16.5 21.0 24.9 ML(OH).H/M»L 6.37 (8) 15.62 (3) 15.44 (3) 10.67 15.42 16.41 ML/ML(OH)«H 4.48 (8) 4.27 (3) 7.16 (3) 5.83 5.58 8.49 a This work; b ref 22. (H(8)). The third deprotonation takes place at a second pyridyl nitrogen, while the final deprotonation occurs at an aliphatic ammonium nitrogen. 134 Figure 4.3. Variable pH UV and NMR Titrations of H 2bped. Top: Molar Extinction Coefficient at 260 nm, B , (Right Axis) and Chemical Shift Difference of H(l), (Left Axis) vs pH. Bottom: Chemical Shift Difference of H(6), H(7), - 0 ~ , and H(8), - H - , vs pH. Titration curves for H 2bped in the presence of Al(III), Ga(III), and In(III) are shown in Figure 4.2. The 1 : 1 curves with Al(III) and Ga(III) both display an inflection at a (mol OH" / mol bped ") = 5 indicative of the formation of a [M(bped)(OH)] species. The 1 : 1 titration curve with In(III) shows an inflection at a = 4, a second buffer region between pH 7 and 8, and another inflection at a - 5, suggesting a [In(bped)]+ species being converted to 135 [In(bped)(OH)] above pH 7. The formation constants of the group 13 metal bped complexes are shown in Table 4.5, along with those of edta4" for comparison. Solution NMR studies. The *H and 1 3 C NMR spectra of [Co(bped)]+ (Tables 4.2 and 4.3, respectively) were assigned on the basis of HMQC, HMBC, and 1 H nOe difference experiments. The HMQC and HMBC experiments established the connectivity in the molecule. The individual hydrogen atoms of each AB pattern for the two methylene resonances, H(6) and H(8), and the those of the AA'BB' (8A = 5 a-;8R = 8R0 spin system, H(7), were assigned on the basis of 1 H nOe difference spectra. There were nOe enhancements between H(4) and H(6b), H(6a) and H(7a), and H(6b) and H(8b). Using the coordination geometry from the crystal structure and these nOe correlations, the individual a and b hydrogen atoms were assigned. PPM 8 7 1 5 4 3 Figure 4.4. X H NMR Spectrum (500 MHz) of [In(bped)]Cl in D 2 0 at pD = 3.2. 136 1 3 24 -10013, - 100 13, Figure 4.5. HMBC (Top) and HMQC (Bottom) Spectra of [In(bped)]Cl in D 2 0 at pD = 3.2. J H and 1 3 C Labelling as in Chart 4.1. Vertical Axis (125.8 MHz), Horizontal Axis (500.1 MHz). 137 The complex [In(bped)]+ could be prepared in situ by mixing appropriate amounts of ligand, metal and hydroxide, and it displayed H and 1 3 C NMR spectra in D2O identical with those of the isolated and redissolved complex. The 1 H NMR spectrum of the latter, as a CI" salt, is shown in Figure 4.4. The assignments were again made on the basis of HMQC, HMBC, and nOe difference spectra. The HMQC and HMBC spectra are shown in Figure 4.5 There were nOe enhancements between H(4) and H(6b), H(6a) and H(8b), H(6b) and H(7b), and H(7a) and H(8a). Preparation of the [In(bped)(OD)] complex by reaction of a 1 : 1 mixture of In(III): H2bped»2HCl with five equivalents of sodium deuteroxide in D 2 0 gave *H and 1 3 C NMR spectra very similar to those of [In(bped)]+ (Tables 4.2 and 4.3 respectively). The chemical shifts change slightly, but the two-fold symmetry is preserved, and the methylene hydrogen atoms on carbons 6, 7, and 8 remain inequivalent. If [In(bped)]PFg is dissolved in DMSO-dg, two species are observed in the 1 H NMR spectrum in a 4.7 : 1 ratio at 20 °C. The assignment of the spectrum is marred by overlapping resonances between 3-3.5 ppm. Both species have two-fold symmetry as evinced by a single set of pyridyl resonances per species and by one set of H(6) resonances which are split in an AB pattern. C NMR is consistent with this symmetry as 9 major and 9 minor resonances are observed (Table 4.3); the tentative assignment of the resonances is by comparison between the spectra of the In(III) complex in D 2 0 and the Co(III) complex in DMSO-dg. The resonances for H(l) and H(6a) are sufficiently separated for each species that their ratio can be calculated as a function of temperature. For the equilibrium: major ^- minor 138 -1 the thermodynamic parameters are AH = +7.9 ± 0.9 kJ»mol"\ AS = +14 ± 3 J/K^mol" 1, which indicates that the minor isomer is favored at elevated temperatures for entropic reasons. The *H NMR spectrum of solid [Ga(bped)]PF6 dissolved in any of D 2 0 , DMSO-dg, or DMF-^7 is described by a series of broad overlapping resonances. Heating a DMF-dj solution to 120 °C caused the spectra to simplify, consistent with fluxionality. No individual isomers (or resonances) could be assigned at any temperature because of overlap and line broadening. There was no evidence for dissociation at high temperature. The C NMR spectrum at -14 °C displayed a multitude of resonances: 12 peaks were observed between 50 and 65 ppm, the chemical shift region for C(6), C(7), and C(8); at least three carboxylate, C(9), resonances were observed clustered about 171 ppm. Heating to 120 °C gave rise to a much simpler spectrum of only nine 1 3 C resonances (Table 4.3), however these resonances are broadened to about 35 Hz. A series of equimolar solutions of H 4 bped«2HCl and Ga(N0 3)3 were prepared and molar equivalents of sodium hydroxide were added ranging from 0 to 6. The solutions were evaporated to dryness, taken up in D 2 0 , evaporated again, and dissolved in D 2 0 to give ca. 60 mM metal - ligand solutions. The 1 H NMR spectra were recorded, and some of these spectra are shown in Figure 4.6. The spectra of solutions containing up to three equivalents of base (pD = 1.8) indicated three distinct pyridyl environments. The 0-2 equivalent base spectra (pD 0.9 -1.5) contained two sharp singlets at 3.36 and 3.37 ppm. At 4 equivalents base (pD = 3.1), all of the resonances had broadened excessively. At 5 equivalents base (pD = 6.6), the resonances remained broadened, but had shifted, and there was now a new sharp resonance at 3.39 ppm. Adding a 6th equivalent of base (pD = 10.2) resulted in 139 pD = 6.6 pJD = 3.1 pD=1.5 J 1 1 1 1 1 ) .n 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ' i ' i • • • • i • 1 •11 9 8 7 6 5 4 3 PPM Figure 4.6. *H NMR spectra (300 MHz) of Equimolar (-70 mM) Mixtures of Ga(ITJ) - H 2 bped«2HCl: Bottom, 2 Equivalents of Base (pD = 1.5); Middle, 4 Equivalents of Base (pD = 3.1); Top, 5 Equivalents of Base (pD = 6.6). 140 the same spectrum as at pD 6.6, however some free ligand resonances were now observed. A similar series of experiments were performed with aluminum. At pD = 2.8 (3 equivalents of base), hydrogen resonances corresponding to free 9 7 ligand and metal complex were observed; the Al NMR spectrum showed a broad peak (W1/2 = 2600 Hz) at 40 ppm, and a sharp peak at 0 ppm. At four -1 equivalents of base (pD = 3.5), the H NMR spectrum showed a series of broad resonances, while the peak at 40 ppm in the A l NMR spectrum remained. At slightly less than five equivalents of base (pD = 4.6), the hydrogen atom resonances became sharper and two distinct pyridyl environments could be distinguished, however the peaks between 3 - 4.5 ppm could not be assigned; a C NMR spectrum at this pD contained only 18 resonances. Formation of the hydroxo complex also resulted in a high frequency shift of the 2 7 A1 NMR resonance to 46 ppm (Wj/2 = 2800 Hz). Further addition of hydroxide (pD = -1 8.9) increased the resolution of the H spectrum; metal complex peaks corresponding to those at pD = 4.6 and also free ligand hydrogen resonances were observed. The broad peak in the 2 7 A l spectrum moved higher in frequency to 54 ppm and there was also a sharp peak at 80 ppm corresponding to [Al(OH) 4]-. 4 1 4.4 Discussion The synthesis of H 2bped has been improved from the previously reported method. The agreement among the deprotonation constants is reasonable except for those of the first deprotonation (Table 4.4). Lacoste et al. assigned the deprotonation steps by comparing the basicities of the various donors. The variable p H X H NMR and UV spectroscopic titrations of H 2bped (Figure 4.3) indicate clearly that the first and third deprotonations 141 occur at the pyridinium moieties, while the final p K a corresponds to a tertiary ammonium deprotonation. The second deprotonation is assigned to a carboxylic acid. The high acidity of one of the tertiary ammonium nitrogen atoms in this molecule is to be expected because of coulombic repulsion from the second tertiary ammonium group, and the presence of the pyridyl and carboxylate groups over which the positive charge can be delocalized. The stability constants determined for bped with the group 13 metal ions (Table 4.5) follow the trend expected for this ligand, based upon comparison of ligands with similar donor atoms. 2,2'-Bipyridine has a slightly higher affinity for In(III) than Ga(III), while aminocarboxylates such Table 4.6. Binary Metal-Ligand Formation Constants and Calculated pM Values (pH 7.4) for Al(III), Ga(III). and In(III) Complexes of Ethylenediamine Based Hexadentate Ligands (see Chart 4.1 for Ligand Abbreviations). ligand log K A 1 L (pAl) tog K G A L (pGa) log K J ^ L (pin) bped 10.85 (13.4) 19.89 (22.7) 22.6 (22.7) edtaa 16.5 (16.3) 21.0 (21.0) 24.9 (23.1) ENDA-HP b 29.2 (21.8) 28.0 (20.6) Sbbpenc 35.3 (27.4) 34.9 (26.9) H B E D d 38.5 (29.6) 27.8 (18.9) SHBED e 37.5 (29.4) 29.4 (21.2) HPED f 25.8 (18.8) 32.0 (25.0) 26.3 (19.3) EDDASSS 35.6 (30.0) 37.0 (31.3) E C h 31.5 (25.6) 33.0 (27.1) a ref. 2; b ref. 9; c ref. 4; d ref. 6; e ref. 12;f ref. 13; 8 ref. 10; h ref. 8; 142 as nitrilotriacetate bind in the order In(III) > Ga(III) > Al(III). Martell and coworkers have shown6 that replacing two of the carboxylato donors in edta with two 2-oxybenzyl moieties reverses the selectivity of In(III) over Ga(III) by 14 orders of magnitude (Table 4.6), whereas Wong et al.A have shown that by replacing the remaining two carboxylate groups by methylpyridyl units gives a ligand with approximately equal affinity for gallium and indium ions (Table 4.6). By replacing the 2-oxybenzyl groups of HBED with 2-methylpyridyl moieties to give bped, it was expected that bped would be selective for In(III). This is indeed the case, but the selectivity is not as great as expected. Figure 4.7 shows the speciation of Al(III), Ga(III), and In(III) with bped2" at a ligand to metal ratio of 2 mM : 2 mM, the same conditions used for the titrations. Both Ga(III) and In(III) are completely complexed at pH 2, whereas Al(III) begins to be complexed at about pH 3. The striking feature of the speciation diagrams is the high stability of the monohydroxo species of Ga(III) and Al(III). It is known that larger metal ions bind to smaller chelate rings with less ligand strain than do smaller metal ions . 2 3 , 4 2 ' 4 3 Furthermore, Hancock has pointed out that a hexadentate ligand such as edta, which forms only five-membered chelate rings, often coordinates to larger metal ions with the addition of a unidentate ligand such as water 4 4 The increase in coordination number will result in longer metal - ligand bond lengths and will reduce the strain introduced into the ligand upon coordination. If the metal ion is too small (e.g. Ga(III) or Al(III)) to accommodate seven donors, then the strain induced by coordinating all six donors in a ligand which only forms five membered rings may be alleviated by displacement of one of the ligand donor atoms by solvent (i.e. the ligand lowers its denticity). This is implied from the titration data with both edta and bped whereby Al(III) and 143 2 3 4 5 6 T T 7 8 9 10 11 pH Figure 4.7. Speciation diagrams (%M(III) vs. pH) Calculated for Equimolar (2 mM) Al(III) - H 2 bped«2HCl (top), Ga(III) - H 2 bped»2HCl (Middle), and In(III) - H 2 bped«2HCl (Bottom). 144 Ga(III) both form stable monohydroxo complexes at much lower p H than does In(III). A ligand such as bped can coordinate in an octahedral fashion to a metal ion with three possible geometries: trans O, O, trans N p y , N p y / or cis O, O and cis N p y , N p y (Scheme 4.1). Despite trying a variety of anions and Scheme 4.1 solvents, crystalline complexes of In(III), Ga(III), or Al(III) were never isolated. The [Co(bped)]PF6 complex was prepared as an inert structural probe of the group 13 metal complexes. Even if all three isomers of the cobalt(III) complex formed, it was hoped that the inert nature of Co (III) would allow separation of the isomers. Only one isomer formed and this was shown to have the trans O, O configuration, both in the solid state and in solution. The 1 H and 1 3 C NMR spectra of the [Co(bped)]+ cation indicated two-fold symmetry in solution. This can only be satisfied by the trans O, O or the trans N p y , N p y isomer. The nOe enhancements observed in the 1 H spectrum were consistent with those expected from analysis of the crystal structure. Furthermore, models show that a trans N p y , N p y isomer should show a strong nOe correlation between H(l) and H(8); there are no nOe enhancements with H(l). The assignment of the 1 H and 1 3 C spectra in the Co(III) system was then used to aid in the estimation of the possible solution structures of [In(bped)]+ 145 and In(bped)(OH)]. Both [In(bped)]+ and [In(bped)(OH)] displayed two-fold symmetry by NMR (Figure 4.4). The nOe enhancements were the same as those seen in the [Co(bped)]+ cation, although the chemical shift order was different (i.e. H(6a) was more deshielded than H(6b) for [Co(bped)]+, the reverse order was seen with [In(bped)]+). Again no nOe enhancement was observed with the H(l) resonance ruling out a trans N p y , N p y isomer. The H NMR data thus implied a solution structure for [In(bped)]+ similar to that of the [Co(bped)]+ cation in the solid state. Based on the potentiometric data, a 1 In(III): 1 Fybped solution was prepared and the pH was raised to above 9 where the [In(bped)(OH)] complex dominates. The two-fold symmetry in the H NMR spectrum was retained, as was the rigidity of the molecule. The hydrogen atom resonances for H(6), H(7), and H(8) all remained as AB quartets suggesting that all six bped donor atoms remain coordinated. If hydroxide was to displace one of the donor atoms, more resonances would be expected because of the lowered symmetry, and a singlet should appear for H(6) or H(8) if a pyridyl or acetato arm becomes displaced. The only effect observed on forming [In(bped)(OH)] is a slight change in the 1 H and 1 3 C NMR chemical shifts. This implies that the [In(bped)(OH)] complex contains a seven-coordinate In(III) ion. The angle defined by the two pyridyl nitrogen atoms and the cobalt atom, N3 - Co - N4, spanned 107° in the crystal structure of [Co(bped)]PF6. Shannon 4 5 lists the ionic radii for Co(III) (low spin, C N = 6) and In(III) (CN=6) as 0.545 A and 0.800 A respectively. Substituting the larger In(III) for Co(III) should result in longer In - N and In - O bond lengths and an even more obtuse N3 - In - N4 angle than in the cobalt structure. Insertion of an aquo or hydroxo ligand between the two pyridyl donors of the indium complex would result in a pentagonal bipyramidal-like geometry which would still maintain 146 a C2 axis and hence be consistent with the NMR results. In addition, a crystal structure of Na3[In(edta)(SC>3)] has been reported in which a seven-coordinate In(III) is bound to the six donors of edta and a monodentate sulfito ligand. 4 6 The *H and 1 3 C NMR spectra of a solution of [In(bped)]PF6 in DMSO-rf6 each indicate the presence of two species. Again, two-fold symmetry is observed in the 1 3 C NMR spectrum; the 1 H NMR spectrum reveals only one environment for the pyridyl resonances of each species and H(6) is again split into an AB quartet. The resonances for H(7) and H(8) are obscured by overlap and broadening. The ratio of minor species to major species increases with temperature (AH = +7.9 ± 0.9 kj-mol"1, AS = +14 + 3 J/K-moT 1). The lack of a well defined AB system for H(8) suggests that this process may involve the displacement of one or both of the carboxylato donors by DMSO, and this displacement increases with temperature (Scheme 4.2). major minor Scheme 4.2 The gallium complex, [Ga(bped)]PF6, is fluxional in DMF-rf7, with the required number of resonances for all three isomers being observed at -14 °C. Alternately, the spectrum could represent bped coordinating with a lower denticity and the presence of solvent donors. 147 The variable pH H NMR spectra of the Ga - bped system are much more revealing (Figure 4.6). There is evidence of complexation at very low pH. In the region where the protonated complex obtains, there are two sharp singlets at 3.36 and 3.37 ppm, which, together integrate to 2 hydrogens, and are assigned to the methylene hydrogens adjacent to a noncoordinated carboxylic acid group. This is analogous to the crystal structure of [Ga(Hedta)(H20)], where one of the carboxylate groups is protonated and noncoordinating, and a water molecule occupies the sixth site of the coordination octahedron. Between 0-2 equivalents of base, three environments are observed for the pyridyl hydrogen atoms, implying the existence of two or three isomers. At four equivalents of base (pD = 3.1), the spectrum shows fluxional behavior, but the singlet corresponding to the free carboxylic acid has vanished. At pD = 6.6, where the hydroxo complex should dominate, the resonances have shifted but remain broad, with the exception of a sharp singlet at 3.39 ppm, again attributed to a noncoordinated carboxylate. At higher pH, the complex resonances remain the same but diminish in intensity as free ligand resonances appear and increase in intensity. The variable pH 1 H and 2 7 A1 NMR spectra are consistent with the potentiometric results. If less than a stoichiometric amount of base (3 equivalents pD = 2.8) to produce [Al(bped)]+ is added, both metal complex and free ligand resonances are observed in the 1 H NMR spectrum. At pD = 3.5, where [Al(bped)]+ should be the dominant species, the 1 H NMR spectrum is similar to that of [Ga(bped)]+, a series of broad resonances. At pD = 4.6 there should be a mixture of [Al(bped)]+ and [Al(bped)(OH)]. At this point, the hydrogen atom resonances have become sharper and two distinct pyridyl environments can be distinguished, however the peaks between 3 - 4.5 ppm could not be assigned; the 18 resonances in the 1 3 C NMR spectrum suggest 148 the presence of only one isomer, since [Al(bped)(OH)] cannot have C 2 symmetry if a hydroxide is coordinated to the aluminum center and aluminum is six-coordinate. A sharp singlet corresponding to the -1 displacement of an acetato group was not observed in the H NMR spectrum. It was expected that one of the pyridyl donors would be displaced by hydroxide, and this would result in a sharp singlet between 4 - 4.5 ppm corresponding to H(6), but this was not observed. A pyridyl arm may have been displaced, but the methylene hydrogen atoms remain diastereotopic. In the gallium system, rotation enabled H(8a) and H(8b) to have identical chemical shifts, and appear as a singlet; the methylpyridyl moiety is much larger than an acetato group and may suffer hindered rotation. Displacement of a pyridyl donor breaks one five-membered chelate ring. Alternately, a tertiary nitrogen may have been displaced, but this would result in the loss of three 5-membered chelate rings; however loss of the chelate effect to coordinate stronger donors has precedence in aluminum chemistry.15 Displacement of a carboxylate is less likely given the affinity of Al(III) for oxygen donors and the lack of a singlet at about 3.4 ppm. Further addition of base (pD = 8.9) sharpens the H NMR resonances of the complex and resonances corresponding to free ligand are observed. 9 7 A l chemical shifts have been used to assign the coordination number of the aluminum ion in a given complex. 4 1 , 4 8 Octahedral complexes resonate between -30 and 50 ppm, tetrahedral complexes between 60 - 120 ppm, five-coordinate complexes are intermediate between octahedral and tetrahedral, while aluminum alkyls resonate to even higher frequency, from 100 - 300 ppm. The [Al(bped)]+ cation resonates at 40 ppm which is at the extreme of chemical shifts reported for octahedral Al(III) complexes. The 2 7 A1 spectra complement the potentiometric results, in that both [A1(H 20) 6] 3 + and 149 [Al(OH)4]" are observed at the extremes of the pD range measured, in 9 7 accordance with the stability constants reported. The A l chemical shifts of aluminum complexes of edta, £ra«s-l,2-diaminocyclohexanetetraacetate (edta), and 1,3-propylenediaminetetracetate (pdta) all occur between 40 -41ppm. 4 9 The alumichromes, aluminum complexes of tris(hydroxamato) cyclic peptide ligands, also resonate at 41 ppm. 5 0 In the Al(III) study with edta, pdta, and edta, it was noted that formation of a hydroxo complex resulted in a low frequency shift of about 3 ppm. In the Al(III) - bped system, addition of base resulted in a high frequency shift to 46 ppm (pD = 4.6) for a mixture of [Al(bped)]+ and [Al(bped)(OH)]. Full conversion to [Al(bped)(OH)] (pD = 8.9) 9 7 results in a Al chemical shift of 54 ppm. This may suggest that [Al(bped)(OH)] formation also brings about a lowering in the coordination number of aluminum from six to five. The aluminum - glycolate (ga) system r--| has been studied, and the Al chemical shifts of mononuclear and oligomeric octahedral aluminum glycolates were determined. In addition, at p H > 9, two complexes, [Al(H_iga)2]~ and [Al(H.1ga)(OH)2]", were tentatively assigned as having tetrahedral aluminum, on the basis of chemical shifts of 60.5 and 55.5 ppm, respectively. Chemical shifts in the octahedral and tetrahedral range were also seen in the aluminum - tartrate system5 2 and in the aluminum - catechol system.53 The high frequency (tetrahedral) resonances were only observed above pH 9. While the chemical shift of [Al(bped)(OH)] lies in the range of five-coordinate aluminum, it should be noted that the vast majority of octahedral aluminum complexes studied by 9 7 Al NMR have the Al(III) ion coordinated by only oxygen donors or by polyaminopolycarboxylates; no shifts of octahedral aluminum complexes containing pyridyl donors have been reported until now. Further 2 7 A1 NMR studies on octahedral Al(III) with different donors would certainly be useful. 150 The stability constants listed in Table 4.5 are lower for the bped complexes than for the edta complexes with the same metal ion, with the exception of that for [Ga(bped)(OH)] which is slightly higher than that for [Ga(edta)(OH)]. The differences between metal - bped and metal - edta stability constants are relatively small for gallium or indium. Since the pK a s are different, a comparison of edta and bped was made by calculating the speciation of a 1 M(III): 1 edta : 1 bped = 2 mM system from pH 2 -10. In Figure 4.8, the percent metal bound to bped is plotted against pH. It is seen that replacing two carboxylates of edta with 2-methylpyridyl units of bped results in a slightly better ligand for gallium, has little effect on indium (the metal is distributed almost equally), but has a detrimental effect on aluminum. Figure 4.8. Plot of %M(III) Bound to bped as a Function of pH Calculated for a 1 M(III): 1 edta : 1 bped = 2 mM Mixture, for Each of Al(III), Ga(III), and In(III). Comparing the complexing properties of bped among the three metal ions studied, the order is In(III) > Ga(III) > Al(III). However, the high stability 151 of the [Ga(bped)(OH)] complex means that gallium competes with indium at higher pH, and this is the reason for the similarity in pM values at p H 7 (Table 4.6). A comparison of bped with the other ligands listed in Table 4.6 indicates that it is not as effective at binding gallium or indium at p H 7.4 as are ligands containing thiolato or phenolato donors. This is a consequence of the lower basicity of the ligand, which makes it more effective at lower pH. An interesting case is the ligand ENDA-HP 9 which is identical to bped, but with the addition of a methyl group in the 3-position and a hydroxyl group in the 6-position of the pyridine ring. ENDA-HP can potentially coordinate through either the pyridyl nitrogen or the anionic oxygen of a deprotonated hydroxyl on the aromatic ring. It was assumed that the ligand coordinates via an N2O4 donor set since two protonated complexes were formed with protonation constants in the same range as pyridine derivatives. Unfortunately no structural or metal complex NMR results have been presented to delineate between the various isomers possible. The pM values for Ga(III) and In(III) with ENDA-HP (Table 4.6) are lower than those listed for bped, while the pM value for Ga(III) - ENDA-HP is much lower than that for Ga(III) - HBED. This suggests some sort of steric interference that makes the 3-methyl-6-hydroxypyridyl moiety less effective than either the 2-methylpyridyl or 2-hydroxybenzyl moieties at coordinating gallium when linked to an ethylenediaminediacetic acid framework. Of the ligands (Chart 4.1) listed in Table 4.6, all have a charge of (4-) or greater, with the exception of bped2". Although the binary metal - bped complexes are cationic, they still possess high thermodynamic stability. Lacoste et al. found the same behavior in their study of divalent metal - bped complexes, and noted that [Cu(bped)] was over 1000 times more stable than 152 [Cu(edta)] . The ability of a 2-methylpyridyl group to substitute for a carboxylate without loss of thermodynamic stability in Ga(III) and In(III) complexes has important ramifications in the design of multidentate chelators for these metals. Ligands with an established high affinity for Ga(III) and In(III) such as EDDASS, 1 0 ' 1 1 E C , 8 ' 4 4 and H B E D 6 ' 5 4 can be altered by substitution of one or both of the carboxylate groups to yield neutral and cationic complexes respectively. This substitution should not have too deleterious an effect on the new gallium complexes and may actually improve the stability of the indium complexes. The work done on Sbbpen4 supports this. 4 . 5 Conclusions The ligand bped2" coordinates to Co(III) in a distorted octahedral fashion with trans carboxylato groups. With Al(III), Ga(III), and In(III) ions, complexes of the type [M(bped)]+ and [M(bped)(OH)] are observed in aqueous solution. The indium complexes are believed to contain seven-coordinate In(III) with a bound water or hydroxo ligand in the seventh position. The gallium complexes are likely 6-coordinate. A protonated species [Ga(Hbped)]2+ was detected and it exists in two or three isomeric forms with a noncoordinated carboxylic acid. The [Ga(bped)(OH)] complex has a noncoordinated carboxylate. The donor types about the [Al(bped)(OH)] complex could not be determined, but only one isomer is present. The carboxylato donors remain coordinated to Al(III), while 2 7 A1 NMR suggests that the complex may be five-coordinate. The order of stability of the binary complexes, [M(bped)]+, is In(IJJ) > Ga(III) > Al(III). The gallium(III) ion circumvents its inability to be 7-coordinate by forming the most stable hydroxo complex of the three and 153 displacing a carboxylate donor. Pyridyl donors can be substituted for carboxylate donors in multidentate ligands without significant loss of thermodynamic stability in the Ga(III) or In(III) complexes. 4.6 References 1) Martell, A. E.; Hancock, R. D. Metal Complexes in Aqueous Solution; Plenum: New York, 1996. 2) Schwarzenbach, G.; Gut, R.; Anderegg, G. Helv. Chim. Acta 1954, 37, 937. 3) Wong, E.; Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1995, 34, 3057. 4) Wong, E.; Caravan, P.; Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chern. 1996, 35, 715. 5) Vaughn, O. J.; Gibson, J. F. Polyhedron 1990, 9, 1593. 6) Ma, R.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1994, 224, 151. 7) Lacoste, R. G.; Christoffers, G. V.; Martell, A. E. /. Am. Chern. Soc. 1965, 87,2385. 8) Li, Y.; Martell, A. E.; Hancock, R. D.; Riebenspies, J. H.; Anderson, C. J.; Welch, M . J. Inorg. Chern. 1996, 35, 404. 9) Motekaitis, R. J.; Sun, Y.; Martell, A. E. Inorg. Chim. Acta 1992,198-200, 421. 10) Sun, Y.; Motekaitis, R. J.; Martell, A. E.; Welch, M . J. Inorg. Chim. Acta 1995, 228, 77. 11) Sun, Y.; Anderson, C. J.; Pajeau, T. S.; Reichert, D. E.; Hancock, R. D.; Motekaitis, R. J.; Martell, A. E.; Welch, M . J. /. Med. Chern. 1996, 39, 458. 12) Motekaitis, R. J.; Sun, Y.; Martell, A. E. Inorg. Chim. Acta 1989, 259, 29. 13) Ma, R.; Martell, A. E. Inorg. Chim. Acta 1993, 209, 71. 14) Mulla, F.; Marsicano, F.; Nakani, B. S.; Hancock, R. D. Inorg. Chern. 1985, 24,3076. 154 15) Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1 9 9 2 , 31, 5400. 16) Liu, S.; Wong, E.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1 9 9 3 , 32, 4268. 17) Liu, S.; Wong, E.; Karunaratne, V.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1993,32,1756. 18) Caravan, P.; Orvig, C. Inorg. Chem. 1 9 9 6 . In Press. 19) Weiner, R. E.; Thakur, M. L. Radiochimica Acta 1 9 9 5 , 70, 273. 20) Harris, W. R.; Pecoraro, V. L. Biochemistry 1 9 8 3 , 22, 292. 21) Harris, W. R.; Chen, Y. C ; Wein, K. Inorg. Chem. 1 9 9 4 , 33, 4991. 22) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum: New York:, 1974-1989; Vol. 1-6. 23) Hancock, R. D. Prog. Inorg. Chem. 1 9 8 9 , 37, 231. 24) Hegetschweiler, K.; Hancock, R. D.; Ghisletta, M.; Kradolfer, T.; Gramlich, V.; Schmalle, H . W. Inorg. Chem. 1 9 9 3 , 32, 5273. 25) Glasoe, P. K.; Long, F. A. /. Phys. Chem. 1 9 6 0 , 64, 188. 26) Caravan, P.; Rettig, S. J.; Orvig, C. Submitted to Inorg. Chem. 1 9 9 6 . 27) Gran, G. Acta Chem. Scand. 1 9 5 0 , 4, 559. 28) Motekaitis, R. J.; Martell, A. E. Can. f. Chem. 1 9 8 2 , 60, 2403. 29) Baes, C. F. Jr.; Mesmer, R. E. Hydrolysis of Cations; Wiley-Interscience: New York, 1976. 30) Motekaitis, R. J.; Martell, A. E. Inorg. Chem. 1 9 8 0 , 19, 1646. 31) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification of Organic Compounds; 4 ed.; Wiley: New York, 1981. 32) Morrison, R. M.; Thompson, R. C. Can. }. Chem. 1 9 8 2 , 60, 1048. 33) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds; John Wiley & Sons: New York, 1978. 34) Weakliem, H. A.; Hoard, J. L. /. Am. Chem. Soc. 1 9 5 9 , 81, 549. 155 35) Riley, P. E.; Pecoraro, V. L.; Carano, C. J.; Raymond, K. N. Inorg. Chern. 1983,22,3096. 36) Miyanaga, A.; Sakaguchi, U.; Morimoto, Y.; Kushi, Y.; Yoneda, H. Inorg. Chern. 1982, 21, 1387. 37) Figgis, B. N. ; Kucharski, E. S.; White, A. H. Aust. J. Chern. 1983, 36, 1563. 38) Bombieri, G.; Polo, A.; Benetollo, F.; Tobe, M . L.; Humanes, M.; Chatterjee, C. Acta Cryst. C1987,43, 1866. 39) Birse, E. F.; Williams, P. A.; Stephens, F. S.; Vagg, R. S. Inorg. Chim. Acta 1988,148, 63. 40) Stoeckli-Evans, H.; Brehm, L.; Pousaz, P.; Bermauer, K.; Burgi, H.-B. Helv. Chim. Acta 1985, 68, 185. 41) Akitt, J. W. Prog. NMR. Spectros. 1989,21, 1. 42) Hancock, R. D.; Martell, A. E. Chern. Rev. 1989, 89, 1875. 43) Hancock, R. D. Acc. Chern. Res. 1990, 23, 253. 44) Anderson, C. J.; John, C. S.; Li, Y. J.; Hancock, R. D.; McCarthy, T. J.; Martell, A. E.; Welch, M. J. Nucl. Med. Biol. 1995, 22, 165. 45) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. 46) Agre, V. M.; Kozlove, N. P.; Trunov, V. K.; Ershova, S. D. Zh. Strukt. Khim. 1979, 22, 138. 47) Kennard, C. H. L. Inorg. Chim. Acta 1967, 2, 347. 48) Delpeuch, J. J. NMR of Newly Accessible Nuclei; Laszlo, P. Ed.; Academic Press: 1983, Vol. 2. 49) Iyer, R. K.; Karweer, S. B.; Jain, V. K. Magn. Res. Chern. 1989, 27, 328. 50) Llinas, M.; De Marco, A. /. Am. Chern. Soc. 1980,102, 2226. 51) Venema, F. R.; Peters, J. A.; van Bekkum, H. / . Chern. Soc. Dalton Trans. 1990,2137. 156 52) Venema, F. R.; Peters, J. A.; van Bekkum, H. Inorg. Chim. Acta 1992, 292, 261. 53) Mhatre, S. M.; Karweer, S. B.; Pradhan, P.; Iyer, R. K.; Moorthy, P. J. /. Chern. Soc. Dalton Trans. 1994, 3711. 54) Motekaitis, R. J.; Martell, A. E.; Welch, M . J. Inorg. Chern. 1990, 29, 1463. > 157 Chapter 5: Cationic Lanthanide Complexes of N,N'-Bis(2-methylpyridyl) ethylenediamine-N,N'-diacetic acid (H2bped). 5.1 Introduction The aqueous chemistry of the lanthanide(III) ions (Ln(III)) is dominated by their oxophilicity. In aqueous solution, saturated or unsaturated neutral nitrogen donors have a low affinity for lanthanides; however when incorporated into a chelate ring containing an oxygen donor, the neutral nitrogen donor readily coordinates to lanthanides. There are numerous reports of stability constants for Ln(III) complexes of picolinate and aminocarboxylate derivatives. The neutral nitrogen donor is ubiquitous in macrocyclic lanthanide complexes through the use of polyaza-macrocycles (e.g. cyclen), Schiff base condensation reactions to give imines, and 2,6-pyridinedicarboxaldehyde as a macrocyclic building block.3"6 O I X O" bped COO" X = CH2COCr, dtpa X = CH 3 , dtpa-bm X = C(0)NHCH3, dtpa-bma X = 2-methylpyridyl, dtpa-bp COO" "O. .0 -ooc-N; X I" X = CH 2COO\ dota X = H, D03A X = CH3CH(OH)CH3, HP-D03A o" o-X = COO", nta X = OH, hida X =Y = COO", edta X = OH, Y = COO", hedta X = C(0)NH2, nta-ma X = Y = OH,hedda Chart 5.1 158 C\ p. H N — L n 3 + L / rr Thompson et al. compared the difference in stability constant between N-substituted iminodiacetic acids and iminodiacetic acid (ida). Their results are summarized in Figure 5.1. The added stability imparted by the fourth donor was approximately the same for each lanthanide. The value for amide substitution was taken from Paul-Roth and Raymond,8 who compared the stabilities of Gd(III) dtpa-bm and dtpa-bma complexes (see Chart 5.1 for abbreviations). The neutral donors which give the greatest increase in stability are alcoholic oxygen, pyridyl nitrogen, and amide oxygen. Figure 5.1. Comparison of neutral T h e use of Gd(III) complexes in donors for Ln(III) chelation after magnetic resonance imaging (MRI) Thompson et al.7 The amide value is h a s spawned many studies on taken from Paul-Roth and Raymond.8 lanthanide complexes of polyaminocarboxylates. The two most studied ligand frameworks are based on dota9 and dtpa 1 0 (Chart 5.1). However, these ligands form anionic and dianionic complexes, respectively, with gadolinium. In order to reduce osmolality and concomitant osmotic shock upon intravenous administration, one of the carboxylato donors in 159 / x X A log K / " " " ^ 0.0 1.6 / ^OH 2.9 rO 3.3 / N = / N H 2 3.38 / 0 dota has been replaced with an alcohol or an amide, to yield an overall neutral Gd(III) complex. The same approach has been taken with dtpa, where two carboxylato donors are replaced with two amide (dtpa-bma), or 2-j methylpyridyl (dtpa-bp) moieties. Although the lanthanide coordination chemistry with alcohol and amide derivatives of dota and dtpa have been well studied, less is known about the 2-methylpridyl derivative. Cacheris et al}1 showed that while the Gd(III) complexes of dtpa-bma and dtpa-bp have similar formation constants, dtpa-bp had a much lower selectivity for Gd(III) over Zn(II), which in turn led to a higher toxicity in rats because of Gd(III) displacement by endogenous Zn(II). Presumably, this negative result has precluded the use of the 2-methylpyridyl group in other ligands for Gd(III) MRI applications. Because of the established affinity of the 2-methylpyridyl moiety for Ln(III), the bis(2-methylpyridyl) derivative of edta (bped) was synthesized. Stability constants for some Ln - bped complexes have been determined to support the previous evidence regarding the replacement of an acetato group by a 2-methylpyridyl group. Multinuclear ( 1H, 1 3 C , 1 7 0) NMR studies have been carried out in order to examine further the effect of changing donor atom types with regard to hydration number, number of isomers present, and rigidity of the complexes. 5.2 Experimental Materials. H 2 bped»2HCl was prepared as described in Chapter 4. Sodium perchlorate, sodium hexafluorophosphate, sodium deuteroxide (NaOD, 40%), deuterium chloride (DC1, 12M) and the lanthanide atomic absorption standards were obtained from Aldrich. Hydrated lanthanide 160 nitrates and chlorides were purchased from Alfa. Deuterium oxide (D2O) was purchased from Isotec. All were used without further purification. Instrumentation. 1 H NMR (200 and 300 MHz) spectra were referenced to external DSS and recorded on Bruker AC-200E and Varian XL300 spectrometers, respectively. O NMR (40.7 MHz, referenced to external H 2 0) and C NMR (75.5 MHz, referenced to external DSS) spectra were recorded on the Varian XL300 spectrometer. The O NMR spectral parameters have been listed in Chapter 3. Mass spectra were obtained with a Kratos Concept II H32Q (Cs+, LSIMS) instrument. Infrared spectra were obtained as KBr disks in the range 4000 - 400 cm"1 on a Galaxy Series FTIR 5000 spectrophotometer, and were referenced to polystyrene. Analyses of C, H , and N were performed by Mr. Peter Borda in this department. The relationship pD = p H + 0.4012 was used to relate the acid concentration of the NMR samples (pD) to the potentiometric results (pH). Synthesis of Metal Complexes. Caution! Pet-chlorate salts of metal complexes are potentially explosive and should be handled with care and only in small amounts. The synthetic procedure for the [Ln(bped)]X (X = CIO4, PF6) complexes was the same in all cases. Only one representative example is given for each counter ion. [Lu(bped)]PF6. To a solution of H 2 bped«2HCl (150 mg, 0.35 mmol)) in 25 mL methanol was added 164 mg of Lu(N03)3»6H 2 0 (0.35 mmol), followed by 115 mg sodium acetate (1.40 mmol, 4 eq.) to yield a colourless solution. The solution was filtered into a 50 mL Erlenmeyer flask and sodium hexafluorophosphate (60 mg, 0.35 mmol) was added with stirring. The resulting colourless solution was filtered into a 30 mL beaker and allowed to stand. A white precipitate formed after the solution stood overnight at room temperature. The precipitate was collected on a Buchner 161 funnel and dried in vacuo overnight at 60 °C to yield 118 mg (50%). Anal. Calc. (found) for [LuC 1 8H 2oN 404]PF 6: C, 31.97 (31.90); H , 2.98 (3.21); N, 8.28 (8.21). Mass spectrum (+LSIMS): m/z = 531 ([M]+, [ L u C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm" 1 , KBr disk): 1599 (b vs, v c = 0 ) , 1447 (s), 1421 (s), 1308 (m), 1262 (m), 1090 (m), 1019 (m), 934 (m), 843 (vs, v P. F), 766 (s), 558 (s, v P. F). [Sm(bped)]C104. To a solution of H 2 bped«2HCl (83 mg, 0.19 mmol)) in 25 mL methanol was added 72 mg of SmCl 3*6H 20 (0.19 mmol), followed by 63 mg sodium acetate (0.77 mmol, 4 eq.) to yield a colourless solution. To this solution, a sodium perchlorate solution in methanol (30 mg, 0.25 mmol, in 1 mL) was added drop wise while a white precipitate gradually formed. After the addition was complete, the white suspension was stirred for 30 minutes and cooled to 4 °C for 1 hr. The off-white solid was collected on a Hirsch funnel, washed with acetone (5 mL), and dried overnight in vacuo. Yield 60 mg (49%). Anal. Calc. (found) for [ S m C 1 8 H 2 0 N 4 O 4 ] C l O 4 » C H 3 O H : C, 35.75 (36.13); H , 3.79 (3.83); N, 8.78 (8.87). Mass spectrum (+LSIMS): m/z = 508 ([M]+, [SmC 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm'1, KBr disk): 1604 (b vs, v c = 0 ) , 1444 (s), 1410 (s), 1326 (m), 1310 (m), 1121,1108 (s, v c l . 0 ) , 1088 (s), 1013 (m), 926 (m), 769 (s), 624 (s). [Dy(bped)]PF6. Yield 145 mg (66%). Anal. Calc. (found) for [ D y C 1 8 H 2 0 N 4 O 4 ] P F 6 « 0 . 5 H 2 O : C, 32.13 (32.26); H, 3.15 (3.33); N, 8.33 (8.07). Mass spectrum (+LSIMS): m/z = 520 ([M]+, [ D y C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm"1, KBr disk): 1591 (b vs, v c = 0 ) , 1446 (s), 1422 (s), 1309 (m), 1263 (m), 1086 (m), 1019 (m), 914 (m), 842 (vs, v P. F), 764 (s), 558 (s, v P. F). [Gd(bped)]C104. Yield 35 mg (28%). Anal. Calc. (found) for [ G d C 1 8 H 2 0 N 4 O 4 ] C l O 4 « C H 3 O H : C, 35.37 (35.57); H, 3.75 (3.79); N, 8.68 (8.86). Mass spectrum (+LSIMS): m/z = 514 ([M]+, [ G d C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm"1, KBr 162 disk): 1605 (b vs, v c = 0 ) , 1445 (s), 1408 (s), 1326 (m), 1310 (m), 1121,1108 (s, v c l . o), 1086 (s), 1014 (m), 926 (m), 768 (s), 624 (s). [La(bped)]C104. Yield 85 mg (60%). Anal. Calc. (found) for [LaC 1 8 H 2 oN404]C104«CH 3 OH: C, 36.41 (36.28); H , 3.50 (3.86); N, 8.94 (8.90). Mass spectrum (+LSIMS): m/z = 495 ([M]+, [ L a C 1 8 H 2 0 N 4 O 4 ] + ) . IR (cm"1, KBr disk): 1602 (b vs, vc==o)/1443 (s), 1410 (s), 1329 (m), 1120,1107 (s, v c l _ 0 ) , 1087 (s), 1011 (m), 929 (m), 767 (s), 624 (s). Multinuclear NMR Studies. The hydration number of [Dy(bped)]+ was determined by the method of Alpoim et a/. 1 3 An equimolar (35 mM) solution of D y 3 + and H 2 bped»2HCl was prepared and a stoichiometric amount (4 eq.) of standardized NaOH was added so that the complex was fully formed. Five solutions of differing dysprosium concentrations were prepared by serial dilution of the stock solution. The 17O NMR spectra were recorded at 20 °C. Solutions of [Ln(bped)]+ in D 2 0 were prepared by adding an equimolar amount of the appropriate lanthanide salt to a D 2 0 solution of Na2bped. The solutions were approximately 90 mM at pD ~ 4.5. The 1H, 1 3 C , and 1 7 0 NMR spectra were recorded at 20 °C. The 1 H and 1 3 C NMR spectra of [La(bped)]+ and [Lu(bped)]+ were identical to those obtained by dissolving solid samples of [Ln(bped)]PF6 (Ln = La, Lu). Potentiometric Equilibrium Measurements. The equilibrium constants were determined by potentiometric methods described in Chapters 2 and 3. The ionic strength was fixed at 0.16 M NaCl, and the solutions were maintained at 25 °C. Argon, which had been passed through 10% NaOH, was bubbled through the solutions to exclude C 0 2 . The ligand was checked for purity by NMR and elemental analysis before titration. Titrations were also employed to verify the molecular weight obtained by elemental analysis. The metal ion solutions were prepared by 163 dilution of the appropriate atomic absorption (AA) standards. The exact amount of acid present in the A A standards was determined by Gran's method. 1 4 The ratio of ligand to metal used was 1 : 2 < L : M < 1 . 2 : 1 . Concentrations were in the range 1.0 - 2.5 mM. At least four titrations were performed on the Ln(III) - H 2bped equilibria, each titration consisting of about one hundred data points. The metal - ligand solutions were titrated over the p H range 2-11 Complexation was usually rapid (1-3 min per point to give a stable pH reading); however, caution was taken to ensure that no trace hydrolysis or precipitation was occurring by monitoring up to 30 minutes for pH drift. The data were refined using the program BEST. 1 5 The stepwise deprotonation constants for H 2 bped»2HCl at 0.16 M NaCl and 25 °C are 1.5, 3.11, 5.53, and 8.67 (Chapter 4). The hydrolysis constants used were taken from Baes and Mesmer.1 6 The data could be satisfactorily explained by the presence of only two species, [Ln(bped)]+, and [Ln(bped)(OH)]. 5 . 3 Results Metal Complexes. The lanthanide complexes of bped2" were all prepared in a similar manner. These cations, [Ln(bped)]+, were isolated as either the hexafluorophosphate or perchlorate salts from methanol. Better yields were obtained for the heavier lanthanides with hexafluorophosphate, while the lighter lanthanide complexes were better isolated as perchlorate salts. Lanthanide nitrate or chloride salts could be used to prepare the Gd - Lu complexes, but only chloride salts were used for the lighter lanthanides, otherwise the complex was isolated as a nonstoichiometric salt of the added counter ion and nitrate. The use of ethanol as a solvent, or heating the 164 reaction mixture both led to the inclusion of inorganic salts. The metal complexes prepared in methanol gave acceptable elemental analyses, with the perchlorates being isolated as methanol solvates. Positive LSIMS mass spectral studies showed strong parent ion peaks for [Ln(bped)]+ with the expected isotopic patterns. The infrared spectra were similar all the [Ln(bped)]+X complexes with strong bands typical of noncoordinated hexafluorophosphate (842, 558 cm"1)1 7 or perchlorate (1120, 1108 cm"1).18 There was a small splitting in the perchlorate Cl-O stretch which may indicate some weakly bonding interaction, although this splitting was much less than that of coordinated perchlorate.18 The carboxylato asymmetric stretch underwent a considerable bathochromic shift upon coordination, from 1728 cm"1 (H 2 bped«2HCl) to about 1600 cm"1 ([Ln(bped)]+). The complexes did not readily dissolve in water or DMSO; however, upon standing or heating, they did dissolve and had considerable solubility ( > 0.2 M), suggesting that the solid state structures may be polymeric. Potentiometric Titrations. Titration curves for H2bped»2HCl in the presence of La(III), Nd(III), Gd(III), Ho(III) and Yb(III) are shown in Figure 5.2. The 1 : 1 curves display an inflection at a (mol OH" / mol bped2") = 4, as expected for the formation of a [Ln(bped)]+ complex. The curves diverge beyond a = 4, with a second buffer region occurring at decreasing pH as the lanthanide series is traversed indicating the formation of a [Ln(bped)(OH)] complex which becomes increasingly stable with increasing atomic number. The formation constants of the Ln - bped complexes are listed in Table 5.1. 165 Figure 5.2. Titration Curves (pH vs. a; a = mol OH" / mol bped) for H 2 bped»2HCl ( ), and Equimolar (2 mM) bped - Ln(III) Solutions: La(III) ( ), Nd(IH) ( ), Gd(III) ( ), Ho(III) ( ), Yb(m) ( ). Table 5.1. Log Formation constants for Ln(III) with bped2" (abbreviated L) at 25 °C, li = 0.16 M NaCl. Ln(III) La(IH) Nd(III) Gd(III) Ho(III) Yb(Jil) ML/M«L 10.81 (4) 11.99 (2) 12.37 (3) 12.31 (2) 13.42 (8) ML(OH)»H/M«L 0.06 (8) 1.45 (8) 2.10 (3) 3.00 (2) 4.43 (4) ML/ML(OH)»H 10.75 (4) 10.45 (8) 10.25 (3) 9.31 (2) 8.98 (4) Solution NMR studies. The % and 1 3 C NMR spectra of [Ln(bped)]+ complexes (Tables 5.2 and 5.3, respectively) were assigned on the basis of 166 COSY spectra and by comparison of the chemical shifts with those determined for [In(bped)]PF6 and [Co(bped)]PF6 (Chapter 4). X H and 1 3 C NMR spectra of the [Ln(bped)]+ complexes indicated the presence of only one isomer, and this isomer had two-fold symmetry in solution. Only 9 of a possible 18 1 3 C NMR resonances and 10 out of a possible 20 1 H NMR resonances were observed for each complex. The splitting of H(6), H(7), and H(8) into three AB patterns suggests that all six donor atoms are coordinated and that the complex is rigid Table 5.2. % (300 MHz) Spectral Dataa'b for H 2bped, c and its La(III),d Sm(III),d Y(III),d and Lu(III)d Complexes in D 2 0 . H2bped [La(bped)]+ [Sm(bped)]+ [Y(bped)]+ [Lu(bped)]+ H(l) 8.59 (4.4) 8.66 (5.1) 8.39 (3.4) 8.71 (4.6) 8.78 (5.1) H(2) 7.76 (4.4, 7.7) 7.48 (5.1, 7.8) 7.23 (3.4, 7.5) 7.48 (4.6, 7.8) 7.57 (5.1, 7.1) H(3) 8.28 (7.7, 7.8) 7.95 (7.8, 7.5) 7.62 (7.5, 7.8) 7.95 (7.8, 7.5) 8.05 (7.1,8.2) H(4) 7.85 (7.8) 7.49 (7.5) 6.90(7.8) 7.47 (7.5) 7.55 (8.2) H(6a) 4.43 4.33 (-15.0) 4.15 (-17.2) 4.20 (-15.3) 4.24 (-15.5) H(6t) 3.83 (-15.0) 3.34 (-17.2) 4.10 (-15.3) 4.16 (-15.5) H(7a) 3.34 2.84 (-10.4) 1.46 (-9.3) 3.04 (-10.1) 3.18 (-10.2) H(7b) 2.39 (-10.4) 1.28 (-9.3) 2.82 (-10.1) 3.06 (-10.2) H(8a) 3.67 3.30 (-16.8) 3.13 (-14.6) 3.51 (-17.3) 3.57 (-17.3) H(8b) 3.11 (-16.8) 3.03 (-14.6) 3.32 (-17.3) 3.31 (-17.3) a For labelling, see Chart 5.1; Numbers in parentheses refer to coupling in Hz (for H(l) = 3 J 1 2 ; H(2) = 3 J 1 2 , 3 J 2 3 ; H(3) = 3 J 2 3 , H(4) = 3 J 3 4 ; H(6) = 2 J a b ; H(7) = 2 J a b ; H(8) = 2 J a b ; c pD = 2.3;d pD=4.6. 167 on the NMR time scale. As the ionic radius decreases from La(III) to Y(III) to Lu(III), some marked changes occur in the 1 H NMR spectra. The resonance for H(l) shifted to lower frequency with decreasing ionic radius. The chemical shift separation between H(6a) and H(6j-,) decreased from 0.50 ppm (La) to 0.10 ppm (Y) to 0.08 ppm (Lu); the same trend was observed for the chemical shift separation between H(7a) and H(7b): from 0.45 ppm (La) to 0.22 ppm (Y) to 0.12 ppm (Lu). However H(8a) and H(8b) remained separated by 0.19 ppm for all three. Of the paramagnetic lanthanide complexes, the only spectra assignable were those of [Sm(bped)]+. The other lanthanide complexes exhibited broad resonances in the H NMR spectrum which were shifted both upfield and downfield. Table 5.3. 1 3 C (75.5 MHz) Spectral Dataa for H 2bped, b and its La(III),c Sm(III),c Y(III),C and Lu(III)c Complexes in D 2 0 . H2bped [La(bped)]+ [Sm(bped)]+ [Y(bped)]+ [Lufbped)] C(l) 147.52 151.45 151.74 152.00 152.30 C(2) 129.65 127.12 125.71 126.72 127.10 Q3) 146.53 142.37 142.15 143.02 143.87 C(4) 128.91 126.58 125.05 126.72 126.95 C(5) 152.94 159.50 158.01 159.38 160.43 C(6) 58.90 64.89 67.00 66.20 64.72 C(7) 50.03 54.40 57.65 59.48 60.98 C(8) 57.40 63.00 66.87 64.52 63.95 C(9) 175.18 182.58 188.92 182.46 182.81 a For labelling, see Chart 5.1;b pD = 2.3;c pD=4.6. 168 Figure 5.3 shows a plot of the dysprosium induced shift (Dy.I.S.) of H 2 O versus concentration of [Dy(bped)]+. The slope of the line is proportional to the hydration number of the complex. ' For Dy (aq)/ the slope is -358 p p m / M . The hydration number Dy (aq) has been shown to be 8, ' which yields a slope of 45 ppm/M per bound water molecule. The slope of the line in Figure 5.3 is -131 ± 3 ppm/M (r = 0.998), which corresponds to a hydration number of 2.93 + 0.07. Figure 5.3. Plot of Dysprosium Induced NMR chemical Shift (Dy.I.S.) of H 2 1 7 0 vs. [Dy(UI)] for [Dy(bped)]+ at 20 °C, pH 4.5. Error bars represent linewidths at half height. The shift, A, induced at a nucleus of a ligand binding to a Ln(III) cation can be expressed as the sum of the diamagnetic shift (A^), the contact shift (Ac), the pseudocontact shift (Ap), and the shift due to the bulk magnetic susceptibility (Ay), equation (1). 169 A = A(j.+ A c + Ap + A, 'X (1) The diamagnetic shift is usually relatively small, and this has been estimated using the shift of the La(III) complex. Since the magnetic moments of the Ln(III) ions are relatively constant, the bulk magnetic susceptibility shift can be estimated from equation (2) which applies to a superconducting solenoid, where C is the concentration (mM) of Ln(III), |ieff is the effective magnetic moment for Ln(III), and T is the temperature (K). Calculated |ieff values were taken from Figgis. The contact and pseudocontact shifts can be expressed by equation (3) where A c and A p are each expressed as the product of two terms. The first term (<SZ> or C D ) is characteristic of the lanthanide, but independent of the ligand, while the second term (F or G) is characteristic of the ligand in question, but independent of the Ln(III) cation. Values for the lanthanide dependent contact term, <SZ>, and pseudocontact term, C D , have been o r o n calculated. Equation (3) can be separated into two linear forms, equations (4) and (5). Although, both (4) and (5) are mathematically identical, Reilley et o n al. have advocated the A z = 47iC(u.eff/2.84)2/3T (2) A' = A - (Ad + A x) = A c + A p = <SZ>F + C U G (3) A ' / C u = F(<SZ>/CU) + G (4) 170 A'/<SZ> = G(CU/<SZ>) + F (5) use of equation (4) when F » G (and equation (5) when G » F) since the dependence on theoretical C D (or <SZ>) will be minimized by a small intercept. In Figure 5.4 (top), the lanthanide induced chemical shifts (L.I.S.) 1 7 for the O nucleus in D 2 0 (after correcting for (-366 ppm for La) and A )^ are listed in Table 5.4 and have been plotted according to equation (4) and -1 1 1 1 1 1 r -0.5 0.0 0.5 1.0 ' 1.5 2.0 2.5 <S Z>/CU Yb I I I 1 1 --2 0 n 2 4 6 C /<%> Figure 5.4. Plot of A ' / C D vs. <S Z>/C D (top) and A'/<SZ> vs. C D / <S Z > (bottom). 171 according to equation (5) in Figure 5.4 (bottom). Following the procedure of Peters et aZ.31 the shifts have been extrapolated to [Ln(bped)]+ = [D20]; the values for C D and <SZ> were taken from reference 21. Plotting the L.I.S. data according to equation (4) gives F = -245 + 5, G = -9.2 + 4, r = 0.999; equation (5) gives F = -268 ± 13, G = -13.4 ± 3, r 2 = 0.83. Table 5.4. 1 7 0 Lanthanide Induced NMR Chemical Shifts of D 2 0 by [Ln(bped)]+at pD = 4.6 at 20 °C. Ln(III) [Ln(III)] (mM) L.I.S. (ppm)a L.I.S. (ppm) La 92.0 -0.67 0 Ce 92.4 -0.14 +288 Pr 90.8 +1.03 +934 Nd 92.1 +1.42 +1135 Eu 85.4 -5.10 -2618 Gd 92.7 -14.27 -7331 Tb 89.9 -16.21 -8650 Dy 92.1 -13.79 -7119 Ho 91.9 -10.94 -5587 Er 90.5 -9.08 -4653 T m 91.5 -6.20 -3021 Yb 91.0 -2.76 -1150 a L.I.S. measured at [Ln(III)] given in column 2. b L.I.S extrapolated to [Ln(III)] = [D 20], and corrected for the diamagnetic contribution by +366 ppm (the shift for La(III)). 172 5.4 Discussion The metal complexes were formed readily by mixing H2bped«2HCl with the appropriate lanthanide salt in methanol in the presence of sodium acetate. As was the case with the Al(III), Ga(III)/ In(III), and Co(III) complexes (Chapter 4), the lanthanide complexes could be isolated by the addition of the appropriate anion, either hexafluorophosphate or perchlorate. No crystalline samples could be prepared despite trying a variety of solvents (methanol, ethanol, isopropanol, water), solvent mixtures (addition of acetone, diethyl ether, acetonitrile), and counter ions (chloride, bromide, iodide, tetrafluoroborate, perchlorate, sulfate, and nitrate). The infrared spectra of the lanthanide complexes were superimposable across the series suggesting similar solid state structures. The asymmetric C=0 stretch shifted upon coordination from 1728 cm"1 (H 2bped»2HCl) to 1690 cm"1 ([Al(bped)]+) to 1631 cm"1 ([In(bped)]+) to 1600 cm"1 ([Ln(bped)]+). The +LSIMS spectra gave strong peaks for [Ln(bped)]+ in the correct isotopic pattern, as expected for monocationic complexes. The potentiometric measurements substantiate the stoichiometry of these complexes. The titration curves show a buffering effect up to four equivalents (a = 4) of hydroxide. The stability constants (Table 5.1, Figure 5.5 (bottom)) increase from La(III) to Nd(III) and then plateau at Gd(III) and Ho(III), before increasing again to Yb(III). This is a common effect in Ln(III) solution chemistry. Although the formation constant for a lanthanide complex with a given ligand usually increases across the series, this increase is rarely monotonic; plateaus and decreases often appear somewhere in the lanthanide series in a plot of free energy of formation versus atomic number. As may be expected from a hydrated monocationic complex, further addition of hydroxide leads to the formation of a ternary 173 Figure 5.5. Top: plot of pK a of bound water vs. inverse ionic radius (CN = 933) for [Ln(bped)]+, and [Ln(hedta)], Bottom: plot of log K M L vs. inverse ionic radius (CN = 93 3) for [Ln(bped)]+, -Q~, [Ln(edda)]+, -Br-, and [Ln(edta)]", —A—. [Ln(bped)(OH)] complex. The stability of this complex increases across the lanthanide series from La(III) to Yb(III). In Figure 5.5 (bottom), the stability constants for La(III), Nd(III), Gd(III), Ho(III), and Yb(III) for edda2", bped2", and edta2" are plotted against inverse ionic radius (coordination number 9). The formation constants for 174 [Ln(bped)]+ are intermediate between [Ln(edda)]+ and [Ln(edta)]" which is to be expected on the basis of the results presented in Figure 5.1. The mean change in stability constant upon adding two 2-methylpyridyl groups to edda to give bped2" is Alog K = 4.1 ± 0.3; conversion of the 2-methylpyridyl groups in bped2" to two acetato groups in edta4" gives Alog K = 5.3 ± 0.8. The same effect is seen on going from ida to pida (Alog K = 2.1 ± 0.1) to nta (Alog K = 2.6 ± 0.1) except the magnitude of the effect is halved. Thompson et al. estimated the stability of certain donors (Figure 5.1), by considering a series of N-substituted iminodiacetic acid derivatives. They plotted log K i (Ln) versus the second deprotonation constant of the ligand for a series of nonbonding substituents (hydrogen, methyl, phenyl, and benzyl) and obtained a straight line. The substituted ligands in Figure 5.1 were treated in the same manner, and the deviation from this line was used to infer the added stability from the substituent. This allows for the differences in basicities of the varying ligands. There is little available data on lanthanide complexes of N-substituted edda derivatives with which to make an analogous comparison. To account for differing ligand basicity the following equilibrium (eq. 6) H 3 L + + L n 3 + - L n L + + 3 H + K 6 (6) can be considered, where L = edda2" or bped2". Equation 6 corrects for competition with H + . For this equilibrium, the difference in Ln(III) stability 9 9 between bped and edda is Alog K 6 = 5.3 ± 0.3. This gives an added stability of 2.65 + 0.15 log units per 2-methylpyridyl group. The same approach used in a comparison of edda and hedda 3 4 of La(III) (see Chart 5.1), gives Alog K 6 = 5.0, 175 or 2.5 log units per 2-hydroxyethyl group. Although these numbers are lower than those given in Figure 1, the trend is the same. In Figure 5.5 (top), the hydrolytic tendencies of [Ln(bped)]+ and [Ln(hedta)] are compared. Plotting the p K a of bound water versus inverse ionic radius shows that both complexes become more acidic as the lanthanide ionic radius decreases. Somewhat surprising is the fact that neutral [Ln(hedta)] is more acidic than cationic [Ln(bped)]+. 1 rj The Dy.I.S. of H 2 O (Figure 5.3) is indicative of a complex containing three bound water molecules, i.e. [Dy(bped)(H20)3]+. In order to determine whether or not the coordination number of [Ln(bped)]+ was constant across the lanthanide series, i.e. [Ln(bped)(H2 0) 3 ] + , the L.I.S. of D 2 l z O was determined for the lanthanide series (excluding Pm). The 1 7 0 NMR shifts were greater than 75% contact in nature for all the lanthanides. It was expected that if the hydration number changed across the series, there would be a break in a plot of A ' / C D versus <SZ>/CD. The linearity of this plot (Figure 5.4, top) suggests that the hydration number is constant across the series. It q c 1 has been shown that a coordinated O nucleus has an F value of -70 ± 11 at 346 K. Since the x/0 NMR shifts are predominantly contact, and since F is linearly proportional to 1/T, 2 5 F = -83 per coordinated 1 7 0 at 20 °C. The value of F determined from equation (4) was -245 ± 5. Dividing this by -83 yields 2.95 ± 0.06 coordinated D 2 O molecules, which is in excellent agreement with the value obtained from the Dy.I.S. experiment, 2.93 + 0.07. Separating the data in Figure 5.4 into heavy (Tb - Yb) and light (Ce - Eu) lanthanides does not yield a better result; in fact the nature of the scatter is such that IFI is slightly larger for the heavier lanthanides. Peters and coworkers31 have shown that for the lanthanide - glycolate (ga) sytem, the 1 : 3 complex exists as [Ln(ga)3(D20) 3 ] , and this complex has a constant hydration number across the 176 lanthanide series. They obtained a value for F of 231 (converted to 20 °C) in good agreement with the result for [Ln(bped)(D20)3]+. In the Ln(III) - edta system, it is believed that there is a hydration equilibrium (equation (7)) occurring across the series in the vicinity of Eu(III): [Ln(edta)(H20)xr - [Ln(edta)(H20)x.!]- + H 2 0 (7) Two species are observed in the UV-vis spectrum of [Sm(edta)]", [Eu(edta)]", and [Gd(edta)]".36 This has been interpreted in terms of a hydration equilibrium where one water molecule is displaced (eq. 7). 3 7' 3 8 Ots 3 9 ' 4 0 has given thermodynamic arguments for a hydration equilibrium. Luminescence spectroscopy studies have given hydration numbers of 2.541 and 3.542 for [Eu(edta)]" and 2.443 and 2.644 for [Tb(edta)]". A Dy.I.S. study of [Dy(edta)]" gave a slope of -90 ppm/M which corresponds to two bound water molecules. X-ray crystal structures of M(I)[Ln(edta)]»nH 20 show a water coordination number of three for Ln = L a , 4 5 Pr , 4 6 Sm, 4 6 G d , 4 6 Dy , 4 7 and H o , 4 8 and two for Ln = Yb. This suggests that the hydration number changes from three to two across the series (i.e. x = 3 in eq. 7). Bryden et al. reported lanthanide induced 1 7 0 NMR shifts of D 2 0 for [Ln(edta)]" complexes 4 9 After extrapolating the shifts to [Ln] = [D 20], and correcting for the temperature difference, F = 206 for the lighter lanthanides (Pr - Eu) and F = 163 for the heavier lanthanides (Tb - Yb). This gives a hydration number of 2.5 for the lighter lanthanides and 2.0 for the heavier lanthanides. These values support the argument for a change in coordination number, and also illustrate the difference in hydration number between [Ln(edta)]" and [Ln(bped)]+. 177 The H and C spectra indicate that the [Ln(bped)]+ complexes have two-fold symmetry in aqueous solution. All six donor atoms are coordintaed since H(6), H(7), and H(8) are split into AB quartets in the lanthanide complexes. This also suggests that the complexes are rigid at 20 °C on the NMR time scale. As the ionic radius decreases from L a 3 + (1.216 A, C N = 9) 3 3 to Y 3 + (1.075 C N = 9, intermediate between D y 3 + and H o 3 + ) 3 3 to L u 3 + (1.032, C N = 9), 3 3 there is no change in the number of peaks in the 1 H NMR spectrum, but the chemical shifts vary (Table 5.2). The largest effect is the chemical shift difference between H(6a) and H(6b). These two resonances are well separated for [La(bped)]+ at 0.50 ppm; however as the ions become smaller, this shift separation drops to 0.10 ppm for Y(III), and 0.08 ppm for Lu(III). A similar effect is observed for H(7a) and H(7b). Since the coordination number does not change across the series, the chelate ring angles must adjust to the smaller size of the ion. The coordination number of the Ln(III) aquo ions is known to change from 9 to 8 across the lanthanide series.50 The complexes, [Ln(bped)(H 20)3]+ have a constant hydration number of three. In the formation of [Ln(bped)(H20)3]+, three factors are at play: the dehydration of Ln 3 + ( a q ) which loses six waters for the early lanthanides, and five waters for the late lanthanides; the increasing charge to radius ratio as the lanthanides become heavier; and any changing steric effects within the bped2" ligand as the lanthanide is changed. This combination is responsible for the plateau in stability constant in the middle of the lanthanide series. However the formation of [Ln(bped)(OH)(H20)n] should depend primarily on the acidity of the ion, and this is what is observed. 178 5.5 Conc lus ions The neutral pyridine donor is an effective ligating group for the lanthanide(III) ions, and offers an alternative to the neutral amide or neutral alcoholic oxygen donor in the design of stable lanthanide complexes. The Ln(III) complexes formed with bped are intermediate in stability between [Ln(edda)]+ and [Ln(edta)]". In contrast to [Ln(edta)]" which undergoes a change in hydration number across the lanthanide series from three to two, [Ln(bped)]+ has a constant hydration number of three. The acidity of bound water is strongly dependent on the lanthanide in question; the p K a of bound water varies by two orders of magnitude from La(III) to Yb(III). 5.6 References 1) Choppin, G. R. Pure Appl. Chem. 1971,27, 23. 2) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum: New York:, 1974-1989; Vol. 1-6. 3) Alexander, V. Chem. Rev. 1995, 95, 273. 4) Bombieri, G. Inorg. Chim. Acta 1987, 239, 21. 5) Fenton, D. E.; Vigato, P. A. Chem. Soc. Rev. 1988, 27, 69. 6) Vallarino, L. M . /. Less-Common Met. 1989, 249, 121. 7) Thompson, L. C.; Shafer, B. L.; Edgar, J. A.; Mannila, K. D. Adv. Chem. Ser. 1967, 72, 169. 8) Paul-Roth, C.; Raymond, K. N. Inorg. Chem. 1995, 34, 1408. 9) Kumar, K.; Jin, T.; Wang, X.; Desreux, J. F.; Tweedle, M . F. Inorg. Chem. 1994,33,3823. 10) Watson, A. D. /. Alloys Compds. 1994, 207/208, 14. 11) Cacheris, W. P.; Quay, S. C.; Rocklage, S. M. Mag. Res. Imag. 1990, 8, 467. 12) Glasoe, P. K.; Long, F. A. /. Phys. Chem. 1960, 64, 188. 179 13) Alpoim, M . C ; Urbano, A. M.; Geraldes, C. F. G. C.; Peters, J. A. /. Chern. Soc. Dalton Trans. 1992, 463. 14) Gran, G. Acta Chern. Scand. 1950, 4, 559. 15) Motekaitis, R. J.; Martell, A. E. Can. J. Chern. 1982, 60, 2403. 16) Baes, C. F. Jr.; Mesmer, R. E. Hydrolysis of Cations; Wiley-Interscience: New York, 1976. 17) Morrison, R. M.; Thompson, R. C. Can. }. Chern. 1982, 60, 1048. 18) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds; John Wiley & Sons: New York, 1978. 19) Huskens, J.; Kennedy, A. D.; van Bekkum, H.; Peters, J. /. Am. Chern. Soc. 1995, 227, 375. 20) Caravan, P.; Hedlund, T.; Liu, S.; Sjoberg, S.; Orvig, C. /. Am. Chern. Soc. 1995, 227, 11230. 21) Cossy, C ; Barnes, A. C ; Enderby, J. E.; Merbach, A. E. /. Chern. Phys. 1989, 90,3254. 22) Kowall, T.; Foglia, F.; Helm, L.; Merbach, A. E. /. Am. Chern. Soc. 1995, 227,3790. 23) Peters, J. A.; Huskens, J.; Raber, D. J. Prog. NMR Spectrosc. 1996,28, 283. 24) Figgis, B. N. Introduction to Ligand Fields; Robert E. Krieger Publishing Co.: Malabar, Florida, 1986. 25) Golding, R. M.; Halton, M. P. Aust. }. Chern. 1972, 25, 2577. 26) Pinkerton, A. A.; Rossier, M.; Spiliadis, S. J. Magn. Reson. 1985, 64, 420. 27) Bleaney, B. /. Magn. Reson. 1972, 8, 91. 28) Bleaney, B.; Dobson, C. M.; Levine, B. A.; Martin, R. B.; Williams, R. J. P.; Xavier, A. V. / . Chern. Soc, Chern. Commun. 1972, 791. 29) Golding, R. M.; Pyykko, P. Mol. Phys. 1973,26, 1389. 180 30) Reilley, C. N.; Good, B. W.; Allendoerfer, R. D. Anal. Chem. 1976, 48, 1446. 31) Peters, J. A. P.; Nieuwenhuizen, M . S.; Kieboom, A. P. G.; Raber, D. J. /. Chem. Soc. Dalton Trans. 1988, 717. 32) Choppin, G. R. in Lanthanide Probes in Life, Chemical, and Earth Sciences; Biinzli, J.-C.G., Choppin, G. R., Ed.; Elsevier: Amsterdam, 1989. 33) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. 34) Courtney, R. C.; Gustafson, R. L.; Chabarek, S. Jr.; Martell, A. E. /. Am. Chem. Soc. 1958, 80, 2121. 35) Nieuwenhuizen, M . S.; Peters, J. A.; Sinnema, A.; Kieboom, A. P. G.; van Bekkum, H . /. Am. Chem. Soc. 1985, 107, 12. 36) Geier, G.; Karlen, U.; von Zelewsky, A. Helv. Chim. Acta 1969,52, 1967. 37) Geier, G.; Jorgensen, C. K. Chem. Phys. Lett. 1971, 9, 263. 38) Kostromina, N. A.; Tananaeva, N. N. Russ. }. Inorg. Chem. (Engl. Trans.) 1971, 16, 1256. 39) Ots, H. Acta Chem. Scand. 1973, 27, 2344. 40) Ots, H. Acta Chem. Scand. 1973, 27, 2350. 41) Bryden, C. C.; Reilley, C. N. Anal. Chem. 1982, 54, 610. 42) Horrocks, W. D. Jr.; Sudnick, D. R. /. Am. Chem. Soc. 1979,101, 334. 43) Brewer, J. M.; Carreira, L. A.; Irwin, R. M.; Elliot, J. I. /. Inorg. Biochem. 1981,14, 33. 44) Chang, C. A.; Brittain, H . G.; Telser, J.; Tweedle, M. F. Inorg. Chem. 1990, 29,4468. 45) Hoard, J. L.; Lee, B.; Lind, M. D. J. Am. Chem. Soc. 1965, 87, 1612. 46) Templeton, L. K.; Templeton, D. H.; Zalkin, A.; Ruben, H . W. Acta Cryst. 1982, B38, 2155. 181 47) Nassimbeni, L. R.; Wright, M . R. W.; van Niekerk, J. C ; McCallum, P. A. Acta Cryst. 1979, B35,1341. 48) Templeton, L. K.; Templeton, D. H.; Zalkin, A. Acta Cryst. 1985, C41, 355. 49) Bryden, C. C ; Reilley, C. N.; Desreux, J. F. Anal. Chern. 1981,53, 1418. 50) Cossy, C ; Helm, L.; Powell, D. H.; Merbach, A. E. New J. Chern. 1995, 29, 27. 182 Chapter 6: Conclusions and Further Thoughts 6.1 General Conclusions What I hope to have shown in the preceding chapters is the complementary relationship between solid state structures of metal complexes and the behaviour of complexes in aqueous solution. In Chapter 2, it was shown that the aqueous solution structures of Ga(III) and In(III) complexes of sulfonated aminophenolates were similar to the solid state structures of the analogous unsulfonated Ga(III) and In(III) complexes. The X-ray crystal structures however gave no indication of the relative stabilities of the metal complexes. Stability constant determinations allowed a ranking of the efficacy of the ligands for sequestering a given metal. Furthermore it enabled the detection of other metal complexes of differing H X M V L Z stoichiometries, and the nature of these complexes could be probed by NMR and/or UV spectroscopy. The aluminum complexes were in general the least stable. There was little evidence to support a [Al(HTRNS)]2" complex in aqueous solution, although Liu 1 had reported the crystal structure of the unsulfonated analogue. The synthetic studies2 ,3 on analogues of H 6 T A P S and HfcTAMS with Al(III) failed to recognize the importance of the protonated complexes, [Al(HTAPS)]2' and [Al(HTAMS)]2". The usefulness of the solution studies presented in Chapter 2 lie in further ligand design. The tren backbone may be suitable for further elaboration of tripodal ligands with which to complex In(III), but it is clearly too large for Ga(III) or Al(III). The tame and tap backbones are much better suited to Ga(III) as is the 2-oxybenzylamino chelating moiety. l,3,5-cis,cis-triaminocyclohexane will only be useful if the amine is altered such that the 183 amino groups are all axial. As expected from the donor atom survey in Chapter 1, the neutral nitrogen amine donor has a low affinity for Al(III). The coordination geometries of tripodal aminophenolate complexes of the lanthanides were more varied. Chapter 3 summarized the different coordination geometries seen in the solid state, and three of these were manifested in solution. H 3 TRNS 3 " acted as a tridentate ligand with the lanthanides coordinating solely through the phenolate groups - monocapped and bicapped complexes were observed. TAMS 6" and TAPS6" behaved as hexadentate ligands coordinating with an N3O3 ligand donor set. 1 7 0 NMR proved a useful tool in determining the overall coordination number of the complex by giving an estimate of the hydration number. Three important findings came out of Chapter 3: the high selectivity of H 3 TRNS 3 " for the heavier lanthanides; the reversal in the magnitude of the stepwise formation constants for the capped and bicapped complexes of H 3 TRNS 3 " with Ln(III); and the marked effect on coordination mode seen by subtly changing the pro-ligand structure and the deprotonation order of the ligand donor atoms. Chapter 4 showed the utility of the 2-methylpyridyl moiety as a neutral substitute for the anionic carboxylate donor, allowing a means of controlling the charge of Ga(III) and In(III) complexes while still maintaining high thermodynamic stability. The complex [Co(bped)]PF6 proved a useful structural probe in examining the solution structures of the group 13 metal complexes. The formation of [Al(bped)]+(aq) was also significant since it represents an Al(III) complex with an N 4 0 2 donor set that is stable in water, albeit under a restricted pH range. The pyridyl group may prove a better neutral donor than the amino group in further development of multidentate Al(III) chelators. 184 The 2-methylpyridyl group was shown in Chapter 5 to be an effective neutral donor in the chelation of lanthanides. 1 7 0 NMR again was useful in determining the hydration number of the Ln(III) complexes and suggested that the complexes have the same coordination number across the lanthanide series. 6.2 Suggestions for Future Work The tripodal aminophenolates are versatile ligands for the trivalent metal ions presented in this chapter. The biodistribution of 6 7 G a complexes of derivatized aminophenolates prepared from the amines tap and tame should be studied to explore the potential of these complexes as radiopharmaceuticals. The tripod concept can be expanded to include different types of donors. Since In(III) has a lower affinity for phenolates than Ga(III), a ligand containing thiol donors may be better. The tripod based on tren was the most effective chelator for In(III) and pro-ligand such as LI and L2 may prove selective for In(III). L I R Tripodal ligands based on the amines tame and tap seemed to have the best size match for the Al(III) ion, however the use of three aliphatic amines in H 6 T A M S and H 6 TAPS caused these pro-ligands to be quite basic. Employing aromatic amines, as shown in L3, should lower the basicity of the 185 amino groups. Using a 2-aminophenolate fragment would also increase the number of 5-membered chelate rings formed upon metal ion complexation. L3 would likely bind Ga(III) quite potential of the Ga(III) complex may be a useful means of varying the [b /GaL3] biodistribution. The factors influencing the selectivity of H 3 T R N S 3 " for the heavier lanthanides should be examined in more detail, along with the proposed hydrophobic explanation for the reversal in the stepwise stability constants. This could be probed by using ethers, amides, or alkyl chains instead of amino groups. The hydrophobicity of the molecule could be increased by adding hydrophobic substituents to the aryl rings. Another useful addition would be a fluorine atom ortho to the phenolic oxygen (R = o-F) which would result in lower phenol deprotonation constants and hence augmented coordination to lanthanides at lower pH. The determination of the stability constants of H 3 T R N S 3 " with the entire lanthanide series would indicate how selective this ligand is on an element to element basis. Good selectivity may merit further studies on the separation of lanthanide mixtures. For instance, control of p H and q [ H 3 T R N S ] might enable the lanthanides to be eluted off a cation exchange column from heaviest to lightest as [Ln(H 3TRNS)2] 3". Derivatizing q H 3 T R N S to add long alkyl chains may make it useful for extraction of heavier lanthanides into an organic phase. R R oxidation) and its electronic behaviour may be controlled by the use of other substituents on aryl ring. Variation of the redox strongly as well. A molecule such as L3 is likely to be redox active (susceptible to L3 186 Finally, the use of the pyridyl donor as part of a multidentate ligand should be considered in the design of new ligands for the trivalent metal ions discussed in this thesis. 6.3 References 1) Liu, S.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1992, 32, 5400. 2) Liu, S.; Wong, E.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1993, 32, 4268. 3) Liu, S.; Wong, E.; Karunaratne, V.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1993, 32,1756. 187 Appendix I: X-Ray Crystallographic Data for [Co(bped)][PF 6 ]»CH 3 CN»H 2 0. Table 1.1. Crystallographic data for [Co(bped)]PF 6«CH 3CN«H 2G\ compound [Co(bped)]PF 6 .CH 3 CN»H 2 0 y,deg 106.96 (1) empirical formula C2nH25CoF6N505P v, A 3 1231.4 (3) formula weight 619.35 z 2 crystal system triclinic Pcalo g'cm3 1.670 space group PI T,°C 21 a, A 10.611 (2) X,A 0.71069 b,A 12.720 (2) \i (MoKoc), cm"1 8.50 c A 9.868 (1) transm factors 0.95 -1.00 a, deg 102.70 (1) R(F) a 0.041 M e g 93.60 (1) R w (F) a 0.038 3 R(F) = SI IF0I - IFCI l / Z l F 0 l ; R w ( F ) = (Sw(IF0l - I F c l ) 2 / E w F 0 2 ) 1 / 2 Table 1.2. Bond Lengths (A) for [Co(bped)]PF 6 »CH 3 CN»H 2 0. bond length bond length bond length C o - O l 1.878 (2) PI - F5a 1.528 (7) N5 - C19 1.118 (4) C o - N l 1.937 (2) PI - F6a 1.534 (9) C1-C2 1.505 (4) C o - N 3 1.960 (2) 01-C4 1.298 (3) C3-C4 1.512 (4) Co-03 1.888 (2) 02-C4 1.218 (3) C5-C6 1.510 (4) C o - N 2 1.941 (2) 03-C6 1.298 (3) C7-C8 1.501 (4) Co - N4 1.958 (2) 04-C6 1.213 (3) C8-C9 1.379 (4) P l - F l 1.572 (6) N l - CI 1.498 (3) C9 - C10 1.385 (4) 188 bond length bond length bond length PI -F2 1.560 (6) NI -C3 1.504 (3) C10- C l l 1.361 (4) PI -F3 1.53 (1) NI -C7 1.483 (4) C l l - C12 > "1.382 (4) PI -F4 1.544 (6) N2 -C2 1.498 (3) C13- C14 1.502 (4) PI -F5 1.541 (6) N2 -C5 1.498 (3) C14- C15 1.375 (4) PI -F6 1.581 (9) N2- C13. 1.498 (4) C15- C16 1.386 (4) P l - Fla 1.492 (6) N3 -C8 1.350 (3) C16- C17 1.374 (4) P l - F2a 1.577 (5) N3- C12 1.344 (3) C17- C18 1.364 (4) P l - F3a 1.54 (1) N4- C14 1.357 (3) C19- C20 1.435 (5) P l - F4a 1.538 (6) N4- C18 1.343 (3) Table 1.3. Bond Angles (°) for [Co(bped)]PF 6 »CH 3 CN»H 2 0. atoms angle atoms angle atoms angle O l - C o l - 03 178.53 (8) Fla -PI -F3a 93.9 (8) N3 -C8- •C7 114.5 (2) O l - C o l - NI 87.36 (9) Fla -PI -F4a 177.3 (6) C7 - C 8 - C9 124.1 (3) O l - C o l - N2 92.51 (8) Fla -PI -F5a 94.4 (6) C9- C10- C l l 119.3 (3) O l - C o l - N4 86.40 (8) Fla -PI -F6a 90.2 (6) N3- C12-• C H 121.5 (3) 03 - C o l - N2 87.39 (9) F2a -PI -F3a 86.6 (7) N4- C14- C13 114.5 (2) 03 - C o l - N4 92.14 (8) F2a -PI -F4a 91.4 (5) C13 -C14-C15 1239. (3) NI - C o l - N3 82.14 (9) F2a -PI -F5a 175.8 (4) C15 -C16 -C17 119.1 (3) N2 - C o l - N3 169.58 (9) F2a -PI -F6a 93.0 (6) N4- C18- C17 122.3 (3) N3 - C o l - N4 107.01 (9) F3a -PI -F4a 86.8 (7) Co - N l - C7 106.2 (2) O l - C o l - N3 92.34 (8) F3a -PI -F5a 89.2 (7) CI - N l - C7 114.4 (2) 03 - C o l - NI 94.10 (9) F3a -PI -F6a 175.8 (8) Co -N2- C2 106.2 (2) 03 - C o l - N3 87.99 (8) F4a -PI -F5a 88.2 (5) Co- N2- C13 105.9 (2) 189 atoms angle atoms angle atoms angle NI - Col -N2 88.87 (9) F4a - P l - F6a 89.1 (6) C2- N2- C13 114.1 (2) NI - Col -N4 169.12 (9) F5a - P l - F6a 91.1 (6) Co -N3--C8 112.1 (2) N2- Col -N4 82.51 (9) Col - O l -C4 116.4 (2) C8- N3- C12 119.1 (2) F l - P l - F2 87.2 (5) Col -03 -C6 115.3 (2) C o - N4- C18 129.4 (2) FI - P l - F3 89.1 (7) Col -NI - C I 106.6 (2) N l - C l -C2 106.7 (2) FI - P l - F4 1754.4 (5) Col - N I -C3 107.5 (2) NI -C3 -C4 111.4 (2) FI - P l - F5 88.6 (6) CI--NI--C3 111.0 (2) O l -C4--C3 114.8 (2) FI - P l - F6 90.6 (6) C3--NI--C7 110.6 (2) N2 -C5 -C6 112.3 (2) F2 - P l - F3 94.4 (8) Col -N2 -C5 107.6 (2) 03 -C6--C5 115.5 (2) F2 - P l - F4 88.9 (4) C2 -N2--C5 111.9.(2) NI -C7 -C8 106.2 92) F2 - P l - F5 173.7 (6) C5- N2- C13 110.6 (2) N3 -C8 -C9 121.4 (3) F2 - P l - F6 84.9 (6) C o - N3- C12 128.8 (2) C8- C9- C10 119.0 (3) F3 - P l - F4 93.7 (7) C o - N4- C14 112.1 (2) C10 - C l l -C12 110.6 (3) F3 - P l - F5 90.3 (8) C14 -N4--C18 118.4 (2) N2- C13 -C14 107.0 (2) F3 - P l - F6 179.2 (9) N2 -C2--CI 107.3 (2) N4- C14 -C15 121.1 (7) F4 - P l - F5 95.1 (7) O l -C4- 02 123.9 (3 C14 -C15 -C16 118.9 (3) F4 - P l - F6 86.6 (5) 02 -C4-•C3 121.2 (3) C16 -C17-C18 119.4 (3) F5 - P l - F6 90.4 (6) 03 -C6-•C4 124.8 (3) N5- C19 -C20 179.5 (4) Fla - P l - F2a 86.1 (5) 04 -C6-•C5 119.8 (3) 190 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059510/manifest

Comment

Related Items