Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Metal speciation and characterization of copper complexing ligands in seawater using electrospray ionization… Ross, Andrew R. S. 1998

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

Item Metadata

Download

Media
831-ubc_1998-34620X.pdf [ 7.78MB ]
Metadata
JSON: 831-1.0061569.json
JSON-LD: 831-1.0061569-ld.json
RDF/XML (Pretty): 831-1.0061569-rdf.xml
RDF/JSON: 831-1.0061569-rdf.json
Turtle: 831-1.0061569-turtle.txt
N-Triples: 831-1.0061569-rdf-ntriples.txt
Original Record: 831-1.0061569-source.json
Full Text
831-1.0061569-fulltext.txt
Citation
831-1.0061569.ris

Full Text

METAL SPECIATION AND CHARACTERIZATION OF COPPER COMPLEXING LIGANDS IN SEAWATER USING ELECTROSPRAY IONIZATION MASS SPECTROMETRY by ANDREW R. S. ROSS B.Sc. (Hons.), The University of Surrey, UK, 1993. A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT 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 NOVEMBER 1998 ©Andrew R.S. Ross, 1-998 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 scholarly 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 C. H F H ' S T g V The University of British Columbia Vancouver, Canada Date IS OCTOBER 19^6* DE-6 (2/88) ABSTRACT II Electrospray ionization mass spectrometry (ESI-MS) provides information regarding the structure and distribution of dissolved polar and ionic metal species which cannot be obtained using electroanalytical or spectroscopic techniques. Such information is necessary if the role of metal complexes in biological and geochemical processes is to be fully understood. The utility of ESI-MS for metal speciation is evaluated using a range of well characterized organic ligands and biogeochemically important metal ions. The technique reveals the stoichiometry and metal oxidation state of organic and inorganic complexes and how speciation changes with solution pH. Results for Cu 2 + and 8-hydroxyquinoline (HL) show for the first time that ESI-MS can determine the distribution of dissolved metal species at equilibrium. A detection limit of 13 nM for the complex Cu"L+ demonstrates the potential of this technique for studying metal species in natural waters. Ionization of uncharged complexes is achieved by protonation, deprotonation or electrochemical oxidation of the ligand. N,N-Diethyldithio-carbamate oxidation is confirmed by studies involving alkali and alkaline earth metals, which suggest that ESI-MS detection also depends on the polarity and thermodynamic stability of the complex. Metal complexation by large polyfunctional organic compounds is investigated using tannins as model natural ligands. Simultaneous detection of free and complexed tannins allows the effects of metal concentration, pH and ionic strength to be investigated. Determination of ligand structure and copper binding site by tandem mass spectrometry further illustrates the unique capabilities of this technique for metal speciation. Ill ESI-MS is combined with immobilized metal-ion affinity chromatography (IMAC) to extract and characterize copper complexing ligands from British Columbia coastal waters. Initial UV analysis of IMAC extracts suggests that river and surface waters are significant sources of these ligands. Subsequent ESI-MS analysis requires matrix removal which is achieved on-line by flow injection analysis using a custom made XAD-16 column. A copper ligand of molecular mass 259 is identified in the surface seawater extract with highest UV absorbance. The data are consistent with a dipeptide containing two primary amino groups, a result supported by studies in which IMAC is combined with other methods for characterizing copper ligands in seawater. iv TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables viii List of Figures ix Glossary xiii Preface xvii Acknowledgments xviii CHAPTER 1. INTRODUCTION... 1 1.1 Background 1 1.2 The Importance of Metal .Speciation 4 1.3 Techniques for Metal Speciation 5 1.3.1 Spectroscopic Methods 5 1.3.2 Electroanalytical Methods ...7 1.4. Electrospray Ionization Mass Spectrometry 9 1.4.1 Mass Spectrometry 9 1.4.2 Ionization Methods 10 1.4.3 Electrospray Ionization 11 1.4.4 Advantages and Limitations of ESI-MS 12 1.5 Summary and Thesis Overview 13 CHAPTER 2. DETERMINATION OF DISSOLVED METAL SPECIES BY ESI-MS 15 2.1 Abstract 15 V 2.2 Background 16 2.3 Experimental 19 2.3.1 Materials 19 2.3.2 Mass Spectrometry .........20 2.3.3 Data Acquisition and Treatment 21 2.4 Results and Discussion 22 2.4.1 Optimization of Electrospray Conditions. ...22 2.4.2 8-Hydroxyquinoline Complexes 25 2.4.3 ESI-MS vs. MINEQL Speciation 27 2.4.4 Dimethylglyoxime Complexes 31 2.4.5 A/,A/-Diethyldithiocarbamate Complexes 33 2.4.6 Diphenylthiocarbazone Complexes..... 37 2.4.7 Glutathione Complexes 41 2.4.8 Salicylic and Phthalic Acid Complexes 43 2.4.9 Other Complexes 46 2.4.10 Solvent Effects 47 2.5 Summary 48 CHAPTER 3. ALKALI AND ALKALINE EARTH METAL SPECIATION BY ESI-MS 50 3.1 Abstract .50 3.2 Background 50 3.3 Experimental 53 3.3.1 Materials '...53 3.3.2 Mass Spectrometry 54 3.4 Results and Discussion 55 3.4.1 Ligand Oxidation......... 55 3.4.2: pH '! I..::.......... 59 3.4.3 Metal Ion Properties 63 3.5 Summary 66 CHAPTER 4. DETECTION AND ELECTROCHEMICAL OXIDATION OF DISSOLVED ORGANIC LIGANDS BY ESI-MS ...67 4.1 Abstract ....67 4.2 Background 68 4.3 Experimental 69 4.4 Results and Discussion 69 4.4.1 Acid-Base Chemistry 69 4.4.2 Redox Chemistry 73 4.4.3 Noncovalent Interactions 78 4.4.4 Gas Phase Fragmentation 80 4.5 Summary 81 CHAPTER 5. CHARACTERIZATION OF DISSOLVED TANNINS AND THEIR METAL ION COMPLEXES BY ESI-MS 82 5.1 Abstract 82 5.2 Background 82 5.3 Experimental 85 5.4 Results and Discussion 88 5.4.1 Electrospray Conditions 88 5.4.2 Tannin Ligands 89 vii 5.4.3 Metal Ion Complexes 91 5.4.4 Metal Concentration, pH and Ionic Strength 93 5.4.5 Structural Characterization 97 5.4.6 Extraction and Separation Procedures 100 5.5 Summary 103 CHAPTER 6. EXTRACTION AND CHARACTERIZATION OF COPPER LIGANDS IN SEAWATER BY IMAC AND FIA-ESI-MS 104 6.1 Background 104 6.2 Immobilized Metal Ion Affinity Chromatography 105 6.2.1 The IMAC System 105 6.2.2 IMAC Extraction Procedure 107 6.2.3 Field Studies : 109 6.3 Flow Injection Analysis of Seawater Extracts 112 6.3.1 Design of theXAD-16 Column 113 6.3.2 The FIA System 115 6.3.3 FIA Procedure 117 6.4 Flow Injection Analysis of IMAC Extracts 122 6.4.1 Artificial Seawater Extracts 122 6.4.2 Field Sample Extracts 123 6.4.3 Ligand Characterization 126 CHAPTER 7. CONCLUSIONS AND FUTURE WORK 130 BIBLIOGRAPHY 133 APPENDIX A. COMPLETE DATA SET FOR CHAPTERS 2 AND 3 146 APPENDIX B. SUPPLEMENTARY DATA FOR CHAPTERS 4 AND 5 178 LIST OF TABLES viii Table 2.1 ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions (Cu2+, Ni2 +, Al 3 +, Fe3+) in (a) 50:50 water/acetonitrile and (b) 50:50 water/methanol 26 Table 2.2 Percent relative abundance of copper(ll)/8-hydroxyquinoline complexes detected by ESI-MS in (a) 50:50 water/acetonitrile and (b) 50:50 water/methanol compared with that predicted by MINEQL in water and 50:50 water/dioxane 29 Table 2.3 ESI-MS speciation for dimethylglyoxime (H2L) and selected metal ions (Cu2+, Ni2 +, Co2 +) in 50:50 water/acetonitrile 32 Table 2.4 ESI-MS speciation for /V,/V-diethyldithiocarbamate (HL) and selected metal ions (Cu2+, Ni2 +, Co 2 +, Cd2+) in 50:50 water/acetonitrile 33 Table 2.5 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal . ions (Mn2+, Co 2 +, Ni2 +, Cu 2 + , Zn 2 +, Cd 2 + , Pb2+) in acidic 50:50 water/acetonitrile. 38 Table 2.6 ESI-MS speciation for salicylic acid (H2L) and selected metal ions (Al3+, Cr3 +, Fe3+) in 50:50 water/acetonitrile 44 Table 3.1 ESI-MS speciation for A/,A/-diethyldithiocarbamate (HL) and selected metal ions (Na+, K + , Mg 2 +, Ca 2 + , Mn2+) in 50:50 water/acetonitrile 57 Table 3.2 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions (Na+, K + , Mg2 +, Ca 2 + , Mn2+) in 50:50 water/acetonitrile... 60 Table 3.3 ESI-MS speciation for glutathione (H3L) and selected metal ions (Na+, K + , Mg 2 +, Ca 2 + , Mn2+) in 50:50 water/acetonitrile 61 Table 4.1 Dissociation constants (Kg) for organic ligands in water at 25 °C 70 Table 5.1 Complex ions generated by electrospray ionization of 50:50 water/acetonitrile solutions containing metal nitrates and tannins 92 Table 6.1 UV absorbance (255 nm) of IMAC extracts from British Columbia coastal, waters 111 Table 6.2 Physical properties of Amberlite XAD resins .......114 LIST OF FIGURES ix Figure 1.1 Complexation of a divalent metal ion (M2+) by some of the . ligands found in natural waters 2 Figure 1.2 Kinetics of the physico-chemical processes involving metal species in aquatic systems......: 4 Figure 1.3 Techniques used to study metal complexation in natural waters .6 Figure 1.4 Positive ion mass spectrum of a protonated ligand (HL) and the corresponding 1:1 copper(ll) complex ..9 Figure 1.5 The formation of charged droplets by electrospray.. ......11 Figure 2.1 ESI mass spectrum of copper(ll) and 8-hydroxyquinoline in 50:50 water/methanol at pH 4.4 (a) and 8.9 (b) ..23 Figure 2.2 Intensities of the 8-hydroxyquinoline complex ions [63Cu"L]+ (O) and [63Cu"L2H]+ (•) in 50:50 water/acetonitrile at pH 9.5 as a function of electrospray cone voltage 24 Figure 2.3 Formation of a neutral divalent metal-ion complex by dimethylglyoxime 31 Figure 2.4 ESI mass spectrum of A/,A/-diethyldithiocarbamate complexes in 50:50 water/acetonitrile at pH 9 .....35 Figure 2.5 Proposed formation of unprotonated diphenylthiocarbazone (L) from the thiol tautomer. (H2L) 38 Figure 2.6 ESI mass spectrum of copper and diphenylthiocarbazone in 50:50 water/acetonitrile at pH 4.2, showing the reduction of copper(ll) to copper(l) and the formation of complexes by the unprotonated neutral ligand 40 Figure 2.7 ESI mass spectra of glutathione species in 50:50 water/acetonitrile atpH9 42 Figure 2.8 ESI mass spectrum of aluminum(lll) and salicylic acid in 50:50 water/acetonitrile at pH 3.8, showing the formation of binuclear and mixed ligand complexes under acidic conditions 45 Figure 3.1 ESI mass spectra for alkali metals and A/,A/-diethyldithiocarbamate in 50:50 water/acetonitrile at pH 5.4 56 X Figure 3.2 ESI mass spectra showing species formed by potassium (a) and calcium (b) with acetate, nitrate and acetonitrile (ACN) in 50:50 water/acetonitrile at pH 4.5 ..58 Figure 3.3 Complexation of sodium ions by the phthalate ligand (H2L)4 in 50:50 water/acetonitrile at pH 4.1(B) and pH 6.7 (•) ...62 Figure 3.4 ESI mass spectra for alkaline earth metals and diphenylthiocarbazone in 50:50 water/acetonitrile at pH 4.5.. 64 Figure 4.1 Organic ligands dimethylglyoxime (a), salicylic acid (b), phthalic acid (c), N,N-diethyldithiocarbamate (d), glutathione (e) and oxine (f) 70 Figure 4.2 Theoretical ES ion intensity I (—) and observed intensity (•) for protonated dimethylglyoxime [H2LH]+ (m/z 117) 72 Figure 4.3 Oxidized ligands /v,/V-diethyldithiocarbamate (a) and glutathione (b)....74 Figure 4.4 ESI mass spectra of glutathione with (a) sodium and (b) potassium in 50:50 water/acetonitrile at pH 4.2, showing the Fe 2 + complex (m/z 667-671) formed as the result of iron contamination 75 Figure 4.5 ESI mass spectrum of phthalic acid in 50:50 water/acetonitrile at pH 8.6, showing the [(H2L)jHLX" ions formed as the result of intermolecular hydrogen bonding 77 Figure 4.6 Percent relative abundance of glutathione species [H3LH]+ (•), [(H3L)2H]+ (o), and [(H3L)3H]+ (•) in 50:50 water/acetonitrile as a function of solution pH 79 Figure 4.7 Distribution of salicylic acid complex ions [Cu'(HL)2]~ (•), [Cu'(HL)(HL-C02)r (A) and [Cu'(HL-C02)2] (•) as a function of electrospray cone voltage 81 Figure 5.1 Schematic of the LC-ESI-MS system 86 Figure 5.2 Configuration of the Quattro II tandem mass spectrometer 87 Figure 5.3 ESI mass spectrum of 100 uM tannic acid with (a) 0 and (b) 400 uM copper(ll) nitrate in 50:50 water/acetonitrile at pH 6 89 Figure 5.4 Hydrolysis of tannin L2 to tannin U and gallic acid 90 Figure 5.5 Formation of methyl-4-ketoglucose by methylation and NAD+ oxidation of D-glucopyranose 90 xi Figure 5.6 ESI mass spectra showing the distribution of L2* ligand and complex ions for 100 uM tannic acid with (a) 0, (b) 100, (c) 200 and (d) 400 uM copper in 50:50 water/acetonitrile at pH 6 94 Figure 5.7 CuL/L' ion intensity ratio versus total copper concentration for tannin ligands (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 6 and (b) pH 8 (ionic strength ~ 1 mM) .95 Figure 5.8 CuL/U ion intensity ratio versus total copper concentration for tannin ligands l_r (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 7 and (b) pH 9 (ionic strength ~ 10 mM) 96 Figure 5.9 ESI-MS/MS daughter ion spectrum for the tannin ligand [L2* - HP (m/z 495) at a collision gas pressure of (a) 0.5 and (b) LOubar 98 Figure 5.10 ESI-MS/MS daughter ion spectrum for the tannin complex [L2* - 2H + 6 3Cuj~ (m/z 495) at a collision gas pressure of (a) 1.0 and (b) 1.5 ubar 99 Figure 5.11 SEC-ESI-MS of 100 uM tannic acid in 50:50 water/acetonitrile containing 0.25 mM TRIS.HCI (pH9), showing the elution of tannins (Ln*) and gallic acid (G) in order of decreasing molecular weight 101 Figure 6.1 Schematic of the shipboard IMAC system .....106 Figure 6.2 Field sample locations in British Columbia coastal waters 108 Figure 6.3 Nitrogen overpressure system for seawater filtration 109 Figure 6.4 UV chromatograms for IMAC extracts collected from the Fraser River estuary (station F2) at 50 m (a) and 5 m (b) 110 Figure 6.5 Structure of XAD macroreticular resin 113 Figure 6.6 Diagram of the packed column used for flow injection analysis 115 Figure 6.7 Schematic of the FIA-ESI-MS system 116 Figure 6.8 FIA-ESI-MS calibration for tannins (•), L2* (•) and L3* (A) in 4 mL of artificial acidified seawater (pH 2) 118 Figure 6.9 FIA-ESI-MS calibration for salicylic acid in 4 mL of acidified artificial seawater (pH 2) ...118 Figure 6.10 System to determine ligand recovery by FIA... 119 xii Figure 6.11 SIR chromatogram (m/z 137) for replicate FIA-ESI-MS analyses of 0.1 uM salicylic acid in 4 mL of acidified artificial seawater 120 Figure 6.12 ESI-MS calibration for salicylic acid (m/z 137) in 10 mL of the FIA eluting solvent 121 Figure 6.13 UV chromatograms for duplicate IMAC extractions of 0.1 u\M tannic acid in 1 liter of artificial seawater 122 Figure 6.14 UV chromatogram (a) for an IMAC extract from surface seawater (station J2) and the m/z 258 SIR chromatogram (b) reconstituted from the FIA-ESI-MS response for the 4 mL IMAC fractions shown in (a) 125 Figure 6.15 Dipeptides of molecular mass 259 127 Figure 6.16 Negative ion ESI mass spectrum of the extracted ligand (a) and the simulated spectrum for the deprotonated dipeptide Lys-lle (b) 128 GLOSSARY Ac acetate ACN acetonitrile ACS American Chemical Society ACSV adsorptive cathodic stripping voltammetry activity free ion concentration ASV anodic stripping voltammetry AUFS absorbance units full scale B overall stability constant binuclear containing two metal ions chelates complexes in which the metal-ligand bonds form closed rings CID collision induced dissociation complexes species containing ionic or covalent metal-ligand bonds covalent involving shared electrons CSS Canadian Survey Ship CTD conductivity/temperature/depth Cu T total copper concentration cyanobacteria bacteria capable of photosynthesis Cys cysteine (amino acid) D dipole moment (Debye) dithizone diphenylthiocarbazone DOM dissolved organic matter ESI-MS electrospray ionization mass spectrometry XIV FIA fulvic acids GC GF-AAS Gin, Glu, Gly HPLC humic acids HXL hydrophilic hydrophobic ICP-MS i.d. He IMAC ionic isotope ratio isotopes ISE K Ka KcOND U LC flow injection analysis high molecular weight DOM which remains soluble at low pH gas chromatography graphite furnace atomic absorption spectrometry glutamine, glutamic acid, glycine (amino acids) high performance liquid chromatography high molecular weight DOM which precipitates at low pH organic ligand with x acidic protons attracted to water repelled by water inductively-coupled plasma mass spectrometry inside diameter isoleucine (amino acid) immobilized metal-ion affinity chromatography carrying a positive or negative charge relative abundance of species containing different isotopes atoms of the same element with different masses, e.g. 6 3Cu, 6 5Cu ion-selective electrode stepwise stability constant acid dissociation constant conditional stability constant tannin ligand with n galloyl groups liquid chromatography Leu leucine (amino acid) Lys lysine (amino acid) MeOH methanol mL milliliters mm millimeters mM millimolar MS/MS tandem mass spectrometry m/z mass-to-charge ratio NAD* nicotinamide adenine dinucleotide (oxidized form) nM nanomolar normal phase LC separation procedure involving non-polar solvents o.d. outside diameter oxine 8-hydroxyquinoline Pa Pascals (1 Pa = 10"5 bar) psi pounds per square inch r2 correlation coefficient radicals species containing unpaired electrons reversed phase LC separation procedure involving polar solvents SD standard deviation SEC size exclusion chromatography siderophores ligands produced by aquatic organisms to facilitate iron uptake SIR selected ion recording SPE solid phase extraction stoichiometry ratio of chemical constituents (e.g. metakligand) Iibar microbars uL microliters urn micrometers uM micromolar V volts ACKNOWLEDGMENTS XVIII This thesis marks the end of a journey which began back in September 1989 with the momentous decision to leave work and return to full time study. I could not have completed that journey without the help and support of some very special people. To Martin, Cait, Lawrence, Ed, Ewa, Jenny and all the class of '93; thanks for the great times we shared at the University of Surrey. Special thanks to my good friends Dave Churchman and Carl Williams for service above and beyond the call of duty, and to Neil Ward for whose initial guidance and inspiration I shall always be grateful. It was Neil and fellow international co-op coordinator Rick Reeve who arranged for me to spend my third undergraduate year at the Pacific Forestry Centre in Victoria. B.C., an opportunity which ultimately led to my return to Canada as a graduate student. Many thanks to Caroline Preston, Kevin McCullough, Tamara Fraser and fellow co-op student Gunter Guth for making my stay at PFC such a memorable one. My decision to attend graduate school at UBC was due largely to the encouragement of Jeff Thompson, who continues to share his knowledge and experience as a member of my supervisory committee. Thanks also to fellow committee members Bill Cullen and Al Lewis, whose 'searching' questions kept me on track during the early stages of my Ph.D. To Kristin Orians and Michael Ikonomou; your guidance and motivation have been an essential part of my training as a research scientist. Thank you both for all your help and advice during the past five years. In carrying out my graduate research at the Institute of Ocean Sciences I have been fortunate enough to work with some outstanding scientists. Special thanks to Dayue Shang, Walter Cretney and Neil', Dangerfield, whose consummate expertise and breaktime conversations have been invaluable! To my mother and father, Gillian and David Ross, who have always encouraged and supported me in everything I do; words cannot begin to express my love and appreciation for all that you have done for me. Thanks also to my brother Rupe 'for being there' and sharing in my earliest scientific endeavours, and to sisters Alison and Jane for putting up with us! Special thanks to my parents-in-law, Maria and Elias Ntzoutis, and to Vicky and Peter Sovenko for all their love and support. This work is dedicated to my daughter Maria, born 15 January 1998 (the International Year of the Ocean), and to my wife Niki, without whose love, patience and understanding none of this would have been possible. Andrew R.S. Ross 14 October 1998 CHAPTER 1. INTRODUCTION 1 1.1 Background Metal cations introduced to natural waters as a result of biogeochemical processes or anthropogenic input tend to combine readily with electron donors known as ligands [1, 2]. Water molecules associated with hydrated metal ions can be displaced by preferred ligands or by adsorption onto particles or colloids. Since these processes are governed by the same principles of thermodynamics and coordination chemistry the resulting species may be collectively referred to as complexes [2]. Metal complexing ligands in natural waters range from simple inorganic anions to organic molecules (peptides, polysaccharides, humic and fulvic acids) and larger'inorganic ligands (oxyhydroxides, clays) [1-3] (Figure 1.1). The distribution of a metal amongst various organic and inorganic forms is referred to as speciation, a term which is also used to describe the identification and determination of metal species [2, 4]. Complexation involves specific metal-ligand interactions which depend on the availability of free, uncomplexed metal ions and ligands [1-3]. Free ion concentrations (i.e. activities) are reduced by non-specific (long range electrostatic) interactions which vary according to temperature, pressure and ionic strength. Metal hydrolysis and the ability of organic ligands to form complexes via acidic or basic groups also depend on pH. These factors are taken intoaccount by using conditional stability constants (KCOND) and side-reaction coefficients(a) to describe the formation of a complex ML from free, uncomplexed metal ions M' and ligands L' in natural waters [1]. KCOND = (ML) / (M"). (U) = KML . a M . a L (1.1) where () denotes concentration and KML is the thermodynamic stability constant. Ligand Complex a) Simple Inorganic Ligands H20 c r M(H20) MCI+ 2+ b) Chelating Organic Ligand O OH N H 2 M N H 2 c) Fulvic Acid (R = organic macromolecule) R R O M d) Iron Oxyhydroxide (S = inorganic particle) —Fe-OH Y O —Fe-OH Fe-0 \ O M Fe-0 / Figure 1.1 Complexation of a divalent metal ion (M ) by some of the ligands found in natural waters [2]. Certain uncharged ligand atoms (e.g. N, 0, S) can form bonds with the metal ion by donating lone pair electrons (->), as shown in (b). 3 Complex stability is determined by the electronic structure of the metal cation and how this is affected by the negative field imposed by the surrounding ligands [5]. In general, the extent of covalent bonding between donor atom and metal ion depends on the degree of mutual polarization between their respective electron shells [1] (Section 3.2, p. 51). For transition metal ions, the energies of the outer d-orbitals relative to each other (and the overlapping s-orbital) are also affected by the ligand field. The resulting distribution of electrons between these orbitals determines the coordination number (m) and preferred oxidation state (n) of the metal ion [5]. Complexation of the hydrated metal ion [M(H20)m]n+ is favored if the overall energy of the occupied orbitals is reduced by the competing ligand field [1,5]. The geometry and stability of the complex are also determined by the number and arrangement of ligand donor atoms. Ligands which bind metal ions via a single atom are termed unidentate, while those containing more than one donor atom are referred to as multidentate [1]. Complexation by the latter results in the formation of cyclic complexes, the most stable of which contain 5- or 6-membered rings (Figure 1.1b). Known as chelates, these complexes are generally more stable than those of the corresponding unidentate ligands, since their formation involves a larger increase in entropy [5], They also remain stable at low concentrations, making them an important class of ligands for trace metal ions (i.e. those present at concentrations less than about 50 nM) [1, 3]. Certain metals (e.g. lead, mercury, tin) can form covalent bonds with ligand carbon atoms as a consequence of biological or abiotic processes [1, 5]. The resulting organometallic complexes are stable in water but partition readily into hydrophobic cell membranes (Section 1.2), and are often highly toxic to aquatic organisms. 4 1.2 The Importance of Metal Speciation Total metal concentrations can be determined by graphite furnace atomic absorption spectrometry (GF-AAS) [6-10] and inductively-coupled plasma mass spectrometry (ICP-MS) [11-13]. Such methods have been used to obtain depth profiles showing the distribution of dissolved and particulate metals in natural waters [14-19]. These data have contributed greatly to our understanding of the biogeochemical cycling of trace metals, particularly since the introduction of clean techniques [18, 19]. Oceanic depth profiles of dissolved Ni2 +, Cd 2 + and Zn 2 +, for example, resemble those of the major nutrients nitrate, phosphate and silicate, suggesting that these metals are involved in biological processes [3]. Others such as Pb 2 + and Al 3 + show a reduction in concentration with depth, reflecting the tendency of these cations to adsorb onto the surfaces of sinking particles. The observed biogeochemical behavior of metal ions depends not on their total concentrations but rather on their distribution between inert and labile species (Figure 1.2) [2]. Process Timescale (s) Species Type oxidation/reduction hydration/dehydration ion pair formation/dissociation < 10 i-3 LABILE chelate formation/dissociation IO-3 - 10' ,6 particle adsorption organic assimilation (bio)alkylation > 103 INERT Figure 1.2 Kinetics of the physico-chemical processes involving metal species in aquatic systems [2, 20]. 5 Inert species can be readily isolated from natural waters without significantly altering their composition [2]. Some of these (e.g. organometallic compounds) can be identified and measured individually using techniques based on chromatographic separation [20-24]. Others (e.g. metals adsorbed onto particles or incorporated into organisms) are characterized by separation into operationally defined fractions based on procedures such as sequential extraction [25-27] and (ultra)filtration [28-31]. Labile species exist in rapid equilibrium with one another, and are largely responsible for regulating the activities of metal ions [2, 32]. This has important consequences for metal bioavailability since the free ion can interact directly with transport sites at the surface of cellular membranes [32-34]. Hence, biological uptake of the metal ion depends on the competition between cell transport sites and complexing ligands in the surrounding medium [35-37], as well as competition between metal ions for the available sites [38-39]. Hydrophobic metal species which are readily adsorbed by the lipid membrane can also be assimilated by aquatic organisms [40, 41]. 1.3 Techniques for Metal Speciation 1.3.1 Spectroscopic Methods A variety of analytical techniques have been used to study metal complexation in natural waters (Figure 1.3). Most, however, are unsuitable for studying labile species. For example, Fourier transform infra-red (FTIR) and nuclear magnetic resonance (NMR) spectroscopy require isolation and preconcentration procedures which are likely to perturb the original speciation [2]. Techniques for elemental analysis (AAS, ICP-MS). are usually insensitive to chemical forms, while the separation methods necessary for speciation may again give rise to alteration of labile complexes (with the possible 6 Technique Parameter Advantages and Disadvantages Measured Fourier transform infra-red (FTIR) spectroscopy Nuclear magnetic resonance (NMR) spectroscopy Electron spin resonance (ESR) spectroscopy Fluorescence spectroscopy Potentiometry Voltammetry bond stretching frequency proton relaxation time spin density and absorption frequency for free and bound radical cation emission intensity and wavelength for ligand (or lanthanide cation) activity (free ion concentration) potential & current for oxidation (or reduction) of the cation (or added ligand complex) indicates the involvement of certain (carboxyl) groups in metal complexation; no further structural information; low sensitivity. can distinguish between electrostatic and covalent metal-ligand interactions involving specific functional groups; limited to radical cations (e.g. Mn2 +, Cu 2 + , Fe3+); cannot identify ligands; low sensitivity. provides information regarding metal oxidation state, binding site location and geometry; has also been used to measure stability constants; limited to radical cations; cannot identify ligands. measures stability constant and total ligand concentration (assuming 1:1 stoichiometry); can also distinguish between different types of metal-ligand interaction; limited to radical cations; cannot identify ligands or binding sites. direct in situ measurement of uncomplexed metal (and hence stability constants); low sensitivity; cannot identify ligands. in situ measurement of (operationally defined) labile metal; can provide kinetic and well as thermodynamic data; highly sensitive; cannot identify ligands. Figure 1.3 Techniques used to study metal complexation in natural waters [2, 39, 40, 41, 45]. 7 exception of size-exclusion chromatography; Section 5.4.6, p. 100) [2]. Even techniques which are sensitive enough for making direct speciation measurements, such as electron spin resonance (ESR) [2, 42, 43] and fluorescence spectroscopy [44-48], require significant development before they can be routinely applied to systems as complicated as natural waters [2]. Such techniques are also unable to identify organic ligands or to determine the proportion of the sample which contributes to the observed signal. 1.3.2 Electroanalytical Methods Electroanalytical techniques (potentiometry and voltammetry) are well suited to the determination of metals in natural waters [1, 49] (Figure 1.3). Potentiometry is currently the only technique which permits unequivocal identification of metal species, namely, free ions [1]. Potentiometric measurements are made using ion-selective electrodes (ISE), which perform well in fresh waters and in culture experiments involving millimolar ion activities [50-53]. Their use in seawater is generally limited to the analysis of major and minor cations, due to the high ionic strength of the medium [1, 2, 54]. However, recent advances in ISE technology have greatly improved the performance of these devices [55], and may eventually allow the bioavailability of trace metals to be estimated from direct in situ activity measurements. The current measured by voltammetric techniques is proportional to the concentration of free metal ions available to the electrode within the timescale of the method [1, 2, 49]. This gives an operationally defined measurement of the amount of metal present as labile species (i.e. free ions, inorganic and weakly bound organic complexes). Direct voltammetric techniques such as differential pulse polarography 8 (DPP) are not usually sensitive enough for trace metal analysis, requiring an electrolytic pre-concentration step of the kind used in anodic- and adsorptive cathodic stripping voltammetry (ASV, ACSV) [9, 49, 56-60]. The latter involves the addition of a synthetic ligand to form reducible complexes with the metal ion. Labile metal is thus redefined as the fraction for which the added ligand can successfully compete, which depends on ligand concentration and the stability of the resulting complex [58-60]. ACSV also permits the analysis of metals ions which cannot be fully reduced within the stability boundaries of water and the mercury electrodes used in voltammetry [49]. As well as providing an operational definition of labile metal speciation, voltammetric methods can be used to determine the stabilities of stronger organic complexes (i.e. those with KCOND > 106) [61]. This is done by titrating the sample with the metal ion of interest and measuring the amount of labile metal at each point in the titration [1, 49, 61, 62]. Linearization of the resulting data gives the total concentration of ligand(s) and the conditional stability constant(s) of the corresponding complex(es) [49, 62, 63]. Voltammetric titration experiments show that metal ions such as Cu 2 + and Fe 3 + form complexes with discrete classes of chelating ligands in seawater [64-66]. Some of these are thought to be released by photosynthetic marine organisms in response to growth-limiting iron concentrations (siderophores) [66-68] or potentially toxic levels of Cu 2 + (copper complexing ligands) [69-71]. Unfortunately the ligands themselves cannot be identified by voltammetric analysis, and so their origin and chemical composition have remained largely speculative [71]. New methodology for characterizing these ligands and their metal-ion complexes in solution is therefore required. 9 1.4 Electrospray Ionization Mass Spectrometry 1.4.1 Mass Spectrometry Of all the detection methods currently available to the analytical chemist it is probably mass spectrometry (MS) which provides the best combination of sensitivity, specificity and adaptability. The use of MS detection with an inductively-coupled plasma (ICP) source, for example, has revolutionized elemental analysis by combining high sensitivity and matrix flexibility with simultaneous multi-element and multi-isotope capability [11-13, 72, 73]. However, one of the greatest advantages of MS is the ability to identify and characterize chemical compounds. Mass spectrometry involves the production of gas phase ions from analyte species in the sample [74]. These ions are separated in vacuo according to their mass-to-charge (m/z) ratios using various combinations of electric and magnetic fields (e.g. quadrupole and magnetic sector analyzers). The current produced as each ion reaches the detector is recorded in the form of a mass spectrum, which displays the m/z ratio and relative abundance of individual ions (Figure 1.4). 100% O c c 3 Si < > «J O 1(HL)H]+ [63CuL] + 61 m/z 10 Figure 1.4 Positive ion mass spectrum of a protonated ligand (HL) and the corresponding 1:1 copper(ll) complex. 1.4.2 Ionization Methods Ionization of analytes introduced in the gas phase is usually achieved by electron impact (El), a 'hard' (high energy) ionization method which results in extensive fragmentation of the analyte. This is useful for identification purposes, since the fragment ions can often be related to a specific analyte. Gas chromatography (GC) reduces the need for structural information, since retention times can be used as a means of identification. However, controlled fragmentation can be useful for confirming the assignment of chromatographic peaks with similar retention times. While GC can be used to resolve smaller non-polar compounds or their volatile derivatives, high performance liquid chromatography (HPLC) is usually required to separate the large polar and involatile or thermally labile species frequently encountered in biological and environmental studies. Whereas ionization of volatile analytes following GC separation is relatively straightforward, HPLC has proved much more difficult to interface with MS. This is because conventional ionization methods such as El rely on the gas phase dispersion of relatively stable molecules, making them incompatible with dissolved labile species. Nevertheless, a number of techniques for the introduction and ionization of liquid samples have been developed [74]. Sudden ionization methods such as fast atom bombardment (FAB) and matrix assisted laser desorption (MALDI) achieve vaporization and ionization almost instantaneously. However, the number of intact molecular ions produced by such methods is relatively small, and tends to decrease markedly with increasing molecular weight. Field desorption techniques use a high electrostatic field to extract ions from a 11 substrate. These 'soft' (low energy) ionization methods are capable of producing substantial quantities of unfragmented analyte ions. Electrospray (ESI) and atmospheric pressure chemical ionization (APcl) involve the desorption of ions into a bath or drying gas. These dynamic ionization methods are capable of handling a continuous flow of sample, and are therefore compatible with on-line separation methods such as HPLC and capillary zone electrophoresis (CZE). 1.4.3 Electrospray Ionization The most successful of the field desorption techniques has undoubtedly been electrospray ionization (ESI) [74, 75]. The use of electrospray to produce gas phase ions from solution for mass analysis was first demonstrated by Malcolm Dole [76]. Fenn and co-workers subsequently developed ESI as a method for interfacing HPLC with mass spectrometry [74, 77]. The electrospray ion source and its utility as a LC/MS interface have been studied in detail by Ikonomou and co-workers [78, 79]. Counter Electrode Electrospray Capillary 9 © © © © © © ©• ® • * © ®e M 0 High Voltage Supply Figure 1.5 The formation of charged droplets by electrospray [78]. 12 Electrospray ionization is achieved by passing the sample solution through a narrow metal capillary to which a potential of several thousand volts is applied [75]. The tip of the capillary is positioned close to a counter electrode held at a potential of a few hundred volts. The resulting electrostatic field causes dissolved ionic species with the same charge as the applied potential to migrate towards the counter electrode, forming a stream of charged droplets containing the analyte ions of interest (Figure 1.5). Solvent evaporation, expedited by the use of a nebulizing and/or drying gas (usually nitrogen), increases the concentration of these ions until the repulsive forces between them overcome the surface tension of the liquid. This causes the electrospray droplets to rapidly subdivide, a process which culminates in the transfer of solute ions to the gas phase. The design of the counter electrode allows these ions to be transmitted to the mass spectrometer via a suitable interface [75] (Section 5.3). 1.4.4 Advantages and Limitations of Electrospray Ionization Electrospray is the 'softest' ionisation method currently available, producing little or no fragmentation of analyte species and allowing solution speciation to be probed directly by mass spectrometry. The ability to generate multiply charge molecular ions (a unique feature of ESI) has revolutionized the analysis of proteins and other biopolymers (Section 5.4.1; p. 88). The technique has also been successfully applied to the study of relatively inert metal species such as organometallic compounds [80, 81] and metalloproteins [82-85]. As well as being highly sensitive and specific, ESI-MS provides information about the structure and distribution of dissolved metal species which cannot be obained using other techniques [84] (Section 1.3). 13 Unfortunately the electrospray ionization mechanism imposes certain restrictions on the analysis of metal complexes in solution. Only polar or ionic species, for example, can be detected by ESI-MS, although in practise the exchange of protons or electrons in solution or during electrospray results in the ionization of most analytes [75, 79, 86, 87]. Electrospray ionization also depends on solvent composition. The high surface tension of aqueous solutions must be reduced by adding a miscible organic cosolvent such as acetonitrile or methanol, while the separation of positive and negative ions during electrospray is suppressed at high ionic strengths [75-78]. ESI-MS analysis of natural waters therefore requires suitable extraction and separation procedures, which must take into account the low sample flow rates (typically 1-40 uL/min) imposed by electrospray [74]. 1.5 Summary and Thesis Overview The bioavailability and geochemical reactivity of dissolved metal ions depend on their distribution between inert and labile species. Voltammetric methods are well suited to the study of metal speciation in natural waters. Such methods show that trace metal ions in seawater are extensively complexed by organic chelating ligands, some of which are thought to be released by marine organisms in order to regulate metal bioavailability. New methodology is needed to identify these ligands and their metal-ion complexes if their role in trace metalbiogebchemistry is to be fully, understood. Electrospray is a soft ionization technique which permits the analysis of dissolved polar and ionic species by mass spectrometry. The use of electrospray ionization mass spectrometry for dissolved metal speciation and characterization of copper complexing ligands in seawater is the subject of this dissertation. 14 In Chapter 2 the utility of ESI-MS for metal speciation is evaluated using model organic ligands and biogeochemically important metal ions. The use of alkali and alkaline earth metals to study ion formation mechanisms is described in Chapter 3, while Chapter 4 explores the relationship between solution chemistry and electrospray ionization for organic ligands with different acid-base and electrochemical properties. In Chapter 5 metal complexation by natural organic ligands (tannins) is studied using tandem mass spectrometry and the principles established in previous chapters. Finally, in Chapter 6, the technique is combined with immobilized metal-ion affinity chromatography and flow-injection analysis for the isolation and characterization of copper complexing ligands in seawater. The work presented in Chapters 2 through 5 has been incorporated into four manuscripts, which appear in the form in which they were originally submitted for publication. In accordance with Section 4.5 of the instructions for thesis preparation (University of British Columbia Library) steps have been taken to reduce the degree of repetition in such areas as methodology and background information. 15 CHAPTER 2. DETERMINATION OF DISSOLVED METAL SPECIES BY ELECTROSPRAY IONIZATION MASS SPECTROMETRY 2.1 Abstract The distribution of metal species in solution was determined using flow injection electrospray ionization mass spectrometry. Complexes formed by selected metal ions with added organic ligands in 50:50 water/acetonitrile and 50:50 water/methanol under acidic, neutral and basic conditions were detected using electrospray ionization conditions optimized to best represent solution-phase interactions. Metal species containing acetate, nitrate and solvent molecules predominated in acidic solution, but became less abundant at higher pH. Interactions between metal ions and added organic ligands became more selective with increasing pH, showing the expected preference of hard and soft ligands for metal ions of the corresponding type. Species distributions also tended towards larger complexes as pH increased. Overall ion yield was greater for aqueous acetonitrile than for aqueous methanol solutions; however, reduction of copper(ll) in aqueous acetonitrile resulted in the detection of copper(l) complexes for certain ligands. Experimental results for copper(ll) and 8-hydroxy-quinoline in 50:50 water/methanol showed good agreement with aqueous speciation predicted using the thermodynamic equilibrium model MINEQL. Detection of neutral complexes was achieved by protonation, deprotonation or electrochemical oxidation during electrospray. 16 2.2 Background By combining mass analysis with the ability to generate gas phase ions from labile polar and ionic species in solution, electrospray ionization mass spectrometry (ESI-MS) has made possible many new and important advances in analytical chemistry [75]. Compatibility with liquid chromatographic separation methods and the ability to analyze proteins and other biopolymers have helped to ensure ESI-MS a role at the forefront of biochemical and environmental research [88]. One of the most productive areas of investigation involving ESI-MS has been the detection and analysis of metal-ion complexes. Applications include the study of gas-phase metal-ion chemistries [89-92] and the formation of metal-ion adducts to assist in the detection and structural elucidation of organic compounds [93, 94]. Dissolved-phase interactions between metal ions and organic ligands form the basis of many essential biochemical processes [1, 5]. ESI-MS is well suited to the study of such interactions, and has been successfully used to determine the metal-binding stoichiometry of proteins such as calmodulin, matrilysin and metallothionein [82-84]. ESI-MS also has potential as a technique for determining the speciation of dissolved metals [95], and their distribution amongst various organic and inorganic forms. The bioavailability and geochemical reactivity of dissolved trace metals in natural aqueous systems are largely controlled by speciation [56]. In coastal and oceanic surface waters, for example, complexes formed by unidentified organic ligands account for over 99% of total dissolved copper, significantly reducing the concentration of 'free', bioavailable Cu 2 + ions [62, 64, 65]. Synthetic organic ligands such as 8-hydroxyquinoline and A/,A/-diethyldithiocarbamate have been used to investigate the uptake of trace metals by aquatic organisms [40, 41]. Modification of dissolved metal 17 speciation using synthetic ligands also has important industrial and biomedical applications, such as inhibiting the formation of insoluble residues [96] or targeting specific organs for chemotherapy [97]. ESI-MS can provide information regarding the structure, stoichiometry and metal oxidation state of dissolved metal-ion complexes which cannot be obtained using electrochemical or spectroscopic techniques [84]. It has been proposed that the intensity distribution of ions detected by ESI-MS closely matches the distribution of pre-formed ions in solution [98]. Detection of analytes which are not intrinsically charged, however, depends on protonation, deprotonation, association with other dissolved ionic species (e.g. Na+) or electrochemical ionization at the electrospray tip [79, 87]. Previous studies also suggest that the distribution of ionic species detected by ESI-MS can be significantly affected by pH changes and gas-phase dissociation reactions during electrospray ionization [96, 99, 100]. The aims of the present study were i) to optimize electrospray ionization conditions for measuring the equilibrium distribution of dissolved metal species, ii) to evaluate ESI-MS as a technique for determining metal-ion speciation, and iii) to investigate the effects of pH and solvent composition on ion yield. The metal ions selected for this study were Mn2 +, Co 2 +, Ni2 +, Cu 2 + , Zn 2 +, Cd 2 + , Pb2+, Fe 3 +, Cr 3 + and Al 3 +, which are of known biological and geochemical importance [1, 5, 101]. Synthetic organic ligands which form extractable metal chelates were chosen to represent different types of metal-ion binding site: viz. 8-hydroxyquinoline (N,0"), A/,A/-diethyldithiocarbamate (S,S~), diphenylthiocarbazone (N,S), and dimethylglyoxime (N,N). The properties of these reagents and their metal-ion complexes are well documented, and can be used for comparison with the results of ESI-MS analysis [102]. The interaction of dissolved metal ions with biogenic organic 18 ligands was also investigated using salicylic acid, phthalic acid and the tripeptide glutathione. Results for copper(ll) and 8-hydroxyquinoline were further evaluated by comparing experimental results with those obtained using a computer-based speciation model (MINEQL). The model uses total concentrations of each component and the stability constants of species formed at equilibrium to set up a series of simultaneous mole balance and mass law equations, which are solved to give the equilibrium concentration of each species [101]. Published stability constants for copper(ll)/ 8-hydroxyquinoline complexes in aqueous solvents [102, 103] were used to model the pH dependence of copper speciation. Copper(ll) is susceptible to reduction in acetonitrile [96], which is routinely added to aqueous samples to optimize electrospray ionization efficiency [78]. We used 50:50 water/acetonitrile (aqueous ACN) and 50:50 water/methanol (aqueous MeOH) to examine the effect of these solvents on speciation and ion yield. The ultimate goal of this research is to combine ESI-MS detection with suitable extraction and separation procedures for investigating dissolved metal speciation in natural waters. Hence, flow injection analysis was used in preference to continuous infusion methods, since the resulting transient signal more closely resembles the elution profiles obtained using on-line liquid chromatographic techniques. Flow injection also reduces analysis time and the risk of cross-contamination (Section 6.3). In the following sections, neutral organic ligands are represented by HXL , where x is the number of exchangeable (acidic) protons. M is used to represent the metal rather than the molecular ion, while formal oxidation states of complexed metals are indicated using 19 Roman numerals, e.g. [M'(H2L)]+, [M"HL.] + Square brackets are used to denote the gas phase ions detected by ESI-MS. 2.3 Experimental 2.3.1 Materials All reagents were certified ACS grade or better. Nitrates of manganese(ll), cobalt(ll), nickel(ll), copper(ll), zinc(ll), cadmium(ll), lead(ll), iron(lll), chromium(lll) and aluminum(lll) were obtained from Aldrich Chemical Co., together with salicylic acid and phthalic acid. Glutathione (reduced form) was purchased from Sigma Chemical Co., and 8-hydroxyquinoline, diphenylthiocarbazone, dimethylglyoxime and /V,/V-diethyldithio-carbamate (diethylammonium salt) from BDH Inc. Organic ligands were used as received, except 8-hydroxyquinoline which was further purified by quartz cold-finger sublimation. Laboratory equipment was cleaned sequentially with 4 M hydrochloric acid (Baker Analyzed), 0.6 M nitric acid (Anachemia Environmental Grade) and doubly-deionized Milli-Q™ water (Millipore) to remove trace metals. Solvents were prepared by dissolving 1 mmol acetic acid (BDH Aristar), ammonium acetate (SigmaUltra) or ammonium hydroxide (Fisher ACS) in 500 mLdeionized water and an equal volume of HPLC grade acetonitrile or methanol (BDH). A concentration of 1 mM was sufficient to achieve a useful study pH range without suppressing ionization, or causing spectral interference by [NH4(NH4N03)i]+ and [N03(NH4N03)i]~ clusters under basic conditions. Equal concentrations of reagent were used to give acidic, neutral and basic solutions of approximately the same ionic strength. Metal and ligand stock solutions (16 mM and 4 mM respectively) were prepared in acid-washed polyethylene bottles. Metal salts and glutathione were dissolved in deionized water, while the remaining ligands were dissolved in HPLC grade methanol. Samples were 20 prepared in 5 mL polypropylene Chemtubes (BioRad, Richmond CA) by combining the volumes of stock solution and solvent required to give metal and ligand concentrations of 100 uM Preliminary studies showed these concentrations to be sufficient to ensure detection of all significant ionic metal species [100]. An 8-hydroxyquinoline concentration of 10 u.M was necessary to avoid memory effects encountered at 100 uM, which may affect speciation. The pH of each sample was measured to an accuracy of ±0.1 units using a combination calomel reference electrode and pH meter (Accumet Model 915, Fisher Scientific). Measurements were made following ESI-MS to avoid possible contamination of the sample prior to analysis. 2.3.2 Mass Spectrometry Samples were analyzed by flow injection ESI-MS using the sample solvent as the carrier. A HPLC pump (Model 126, Beckman Instruments, Fullerton CA) delivered solvent at 10 uL/min to a Rheodyne 7725i multi-port rotary valve (20 u.L loop) connected to the ESI probe by a 100 urn i.d. fused-silica capillary (phenyl-methyl deactivated GC guard column, Restek Corporation, Bellafonte PA). Samples were introduced using a 50 ul glass syringe with a stainless steel needle (Hamilton Co., Reno NA), which was rinsed thoroughly with solvent between injections. Mass analysis was performed using a Quattro triple quadrupole mass spectrometer (Figure 5.2, p. 87) fitted with the pneumatically-assisted electrospray probe supplied (Micromass, Manchester UK). The source temperature was 60 °C and nitrogen (> 95 % purity) was used as the drying and nebulizing gas (5.0 and 0.25 L/min, respectively). Electrospray ionization conditions were as follows: electrospray capillary 21 2.9 kV, high voltage (HV) lens 600 V, cone 30 V, (second) skimmer 30 V During optimization the potential difference between the cone and skimmer (i.e. skimmer offset) remained constant at 0 V regardless of the cone voltage (Figure 2.2, p. 24). 2.3.3 Data Acquisition and Treatment Mass spectra were scanned in the m/z range 20 to 1000 at the rate of 10 s/scan, with an interscan delay of 10 ms. The elution of sample components was monitored by means of single ion chromatograms reconstituted from the recorded mass spectra. These showed that, under neutral and basic conditions, salicylic acid and phthalic acid complexes of divalent metal ions were partially retained in the sample introduction system, and tended to co-elute during consecutive analyses. Results obtained for such complexes were not, therefore, representative of dissolved phase speciation, and were omitted. Scans obtained between 2 and 4 minutes after injection (corresponding to the typical non-retained elution profile) were combined and processed using the Masslynx software (version 2.1). The relative abundance of ionic species was calculated from the corresponding peak intensities in the ESI mass spectrum, a procedure which allowed for changes in overall ionization efficiency between samples. Due to quadrupole transmission bias, relative peak intensities are not necessarily proportional to relative dissolved analyte concentrations. However, transmission bias is constant for a given analyte, so qualitative changes in speciation are not affected [98]. For species containing metals with two or more isotopes, the intensity of the highest peak was recorded and adjusted to reflect the isotopic distribution. Metal speciation was divided into three categories: A) complexes formed with the added ligand (e.g. 8-hydroxyquinoline); B) species formed with other solution components (i.e. solvent, 22 nitrate, acetate); C) distribution of total metal speciation between A and B. Results were expressed as percent relative abundance within each category (Table 2.1). 2.4 Results and Discussion For optimum sensitivity and successful detection of target species, instrument polarity should match that of the dissolved analyte ions [82]. Preliminary investigations showed that ligands and complexes containing basic amino groups (8-hydroxy-quinoline; dimethylglyoxime; A/,/V-diethyldithiocarbamate; diphenylthiocarbazone; glutathione) undergo protonation to form positive ions [100]. In certain instances, neutral A/,A/-diethyldithiocarbamate complexes were ionized via electrochemical oxidation rather than protonation (Section 2.4.5). Ligands and complexes which contain only acidic groups (salicylic acid and phthalic acid) were detected in negative ion mode. Initial studies also showed that, with the exception of copper, interactions between the metal and added ligand were largely unaffected by which organic co-solvent was used. Investigations involving pH changes were therefore carried out using 50:50 aqueous ACN, which afforded greater sensitivity than 50:50 aqueous MeOH. Metal speciation was also determined in neutral aqueous MeOH, but only copper speciation was investigated in aqueous MeOH under different pH conditions. A complete set of results for these experiments is given in Appendix A (pp. 146-177). 2.4.1 Optimization of Electrospray Conditions The electrospray ion source was optimized for solutions containing copper(ll) and 8-hydroxyquinoline (HL), in which 1:1 and 1:2 metal.ligand complexes were detected (Figure 2.1). 23 [CunL]+ ESI mass spectrum of copper(ll) and 8-hydroxyquinoline in 50:50 water/methanol at pH 4.4 (a) and 8.9 (b). Relative abundance of 100% corresponds to 1.0 x 107 and 2.4 x 106 counts, respectively. Peaks at m/z 207 and 267 both contribute to the total recorded ion intensity for the 1:1 complex (Tables 2.1 and 2.2). 24 1.6E+6 j c 1.2E+6 -o o 8.0E+5 -w c <D 4.0E+5 - • c O.OE+0 1 10 15 20 25 30 35 40 ES cone voltage (V) cone high voltage lens } capillary second skimmer Configuration of ES Ion Source Figure 2.2 Intensities of the 8-hydroxyquinoline complex ions [63Cu"L]+(o) and [ 6 3 C U " L 2 H ] + ( « ) detected in 50:50 water/acetonitrile at pH 9.5 as a function of electrospray cone voltage. The [ 6 3Cu"L 2H]+ intensity maximum suggests that dissociation of this complex becomes significant above 30 V. Thermodynamic equilibrium calculations made using MINEQL predicted that the 1:2 complex Cu"l_2 should predominate under basic conditions. However, the relative intensities of complex ions detected in aqueous ACN at pH 9.5 were found to be highly dependent on the cone voltage of the electrospray (ES) ion source (Figure 2.2). 25 While optimum sensitivity for the protonated complex ion [Cu"L2H]+(m/z 352,354) was achieved at 30 V, the intensity of the 1:1 complex ion [Cu"l_]+ (m/z 207,209) increased sharply at higher cone voltages, due to dissociation of [CuML2H]+ in the ESI-MS interface. This was confirmed by MS/MS experiments involving copper(ll) and iron(lll) complexes, which showed that [63Cu"L]+ (m/z 207) and p6Fe'"L2]+ (m/z 344) can be formed by collision-induced dissociation of [63Cu"L2H]+ (m/z 352) and [56Fe'"L3H]+ (m/z 489) in the hexapole collision cell (Ar pressure 0.10 Pa, collision energy 10 V). Detection of the iron(ll) complex [56Fe"L]+ (m/z 200) at higher collision energies indicated that intramolecular electron transfer processes of the kind previously documented for other transition metal complexes are also possible [91, 96]. To minimize such reactions a cone voltage of 30 V was used for all samples. Results subsequently showed that lower cone voltages may be required to preserve dissolved phase interactions, although overall ion yield is significantly reduced (Figure 2.2). 2.4.2 8-Hydroxyquinoline Complexes Deprotonated 8-hydroxyquinoline, or oxine (L~), acts as a bidentate anionic ligand (N,0~), forming chelates with a wide range of metal ions. Complexes of the form ML n (where n = metal oxidation state) have been identified for all the metal ions studied [102]. As with many bidentate organic ligands, the stability constants of divalent metal oxinates follow the Irving-Williams order, Mn 2 + < Fe 2 + < Co 2 + < Ni 2 + < Cu 2 + > Zn 2 + [1]. Only the most stable complexes, i.e. those of nickel(ll) and copper(ll), were detected by positive ion ESI-MS, along with aluminum(lll) and iron(ill) oxinates (Table 2.1). 26 (a) Metal Ion (Mn+) Cu 2 + Ni 2 + A l 3 + Fe3 + pH 4.4 6.8 9.5 4.5 7.3 9.0 4.1 6.5 8.7 3.5 6.5 8.7 (A) E [(M"LAq)Hj]+ 96 39 12 100 58 — [M"l_2H]+ 4 48 74 — 42 100 [M"2L3f _ 13 14 — — ' — [M'"L2]+ 96 59 45 96 86 64 [M'"L3H]+ 4 41 55 4 14 36 (B) £[M'(ACN)if 100 100 — E [M"(ACN)J2+ — — — 80 — — S [M"Y(HAc)j(ACN)i]+ 20 — — (C) SA) 1 30 100 1 100 100 100 100 100 100 100 100 EB) 99 70 — 99 — — (b) Metal Ion (Mn+) Cu 2 + Ni 2 + A l 3 + Fe3 + phi 4.4 6.9 8.9 7.2 6.3 6A_ (A) E [(M"LACj)Hj]+ 95 45 7 100 [M"L2Hf 5 49 81 — [M"2L3f — 6 12 — [M"'L2r 83 92 [MmL3H]+ 17 8 (B) E [M"(MeOH)i]2+ — — — — E [MllY(HAc)j(MeOH)i]+ — — — — (C) EA) 100 100 100 100 100 100 SB) ' . —, — — — n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N0 3 . A) oxinate complexes. B) other species. C) all metal species. Table 2.1 ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions in (a) 50:50 water/acetonitrile and (b) 50:50 water/methanol (% relative abundance). 27 Because the phenolic group of oxine is weakly acidic, the extent of metal complexation is highly dependent on solution pH [102]. This is consistent with the shift in oxine complex distribution towards [MLnH]+ observed for copper(ll), nickel(ll), aluminium(lll) and iron(lll) oxinates with increasing pH (Table 2.1a, Section A). The tendency of copper(ll) to form polynuclear complexes [5] was demonstrated by the detection of [Cu"2L3]+ (m/z 558-562) in neutral and basic solutions. Although cobalt(ll) complexes were not observed, [Co"'L2] + (m/z 347) was detected in basic solution. Cobalt(lll) complexes of dimethylglyoxime and A/,A/-diethyldithiocarbamate were also observed (Tables 2.3 and 2.4). This is probably due to stabilization of cobalt(lll) by NH 3 ligands formed from ammonium ions in neutral and basic solutions [60]. Divalent metal species containing acetate, nitrate and solvent molecules were detected along with oxine complexes in acidic aqueous ACN (Table 2.1a, Section B). Ions corresponding to [(Cu"l_Ac)H]+ (m/z 267,269; Figure 2.1a) and [(Ni"LAc)H]+ (m/z 262,264), in which both oxine and acetate may act as bidentate ligands, were also detected in acidic solutions, and contribute to the recorded I [(M"LACJ)HJ] + values (Tables 2.1 and 2.2). 2.4.3 ESI-MS vs. MINEQL Speciation The pH dependence of copper(ll) speciation was modeled using MINEQL+ (version 2.1) [104]. Thermodynamic equilibrium constants for the formation of ligand species [105] and copper complexes [103] in water were obtained from the literature, and are defined as follows. 28 Ligand: Pi = [HL]/[H+].[L1 logPi 9.81 (2.1) p2 = [HLH+]/[H+f[L1 logp2 14.83 (2.2) Complexes: p, = [Cu"L+]/[Cu2+].[L-] logpi 9.55 (2.3) p2 = [Cu"L2]/[Cu2+].[L-]2 logp2 18.27 (2.4) Stability constants for acetate and hydroxyl species were provided by the MINEQL database. Although constants for nitrate species were not provided, inorganic speciation of copper is unlikely to be significant in the presence of organic chelating ligands such as oxine and acetate and can, therefore, be assumed to constitute part of the predicted 'free' Cu 2 + concentration [1]. Stability constants for ternary and binuclear complexes (e.g. Cu"2L3+) were not found in the usual references [102, 103, 106]; however, such complexes constitute a relatively small fraction of the species detected. Total concentrations of 1 mM acetate, 100 u.M copper(ll) and 10 u.M oxine were used in MINEQL simulations. (Although basic solutions did not actually contain acetate, the model predicted that complexation by acetate is not significant above pH 7). According to the model, 'free' (i.e. solvated/inorganic) copper dominates speciation under acidic conditions. Above pH 6 the relative abundance of oxinate complexes becomes significant, and above pH 8 copper is bound quantitatively by oxinate. These results are in good agreement with the experimental data, which show a similar shift in distribution from [Cu'(ACN)j]+ to copper(ll)/oxinate complexes with increasing pH (Table 2.1a, Section C). Only copper(ll)/oxinate complexes were detected in 50:50 aqueous MeOH, which formed [MeOH2(MeOH)i]+ rather than [Cu"(MeOH)i]2+ clusters under acidic conditions. 29 The determination of copper(ll)/oxine speciation by ESI-MS was evaluated by comparing the predicted distributions of (Cu"L)+ and CuML2 in water (Eqtns. 2.3 and 2.4) and 50:50 water/dioxane (logp, = 13.49; logp2 = 26.22) [102] with values for [Cu"L]+ and [Cu"l_2H]+ obtained experimentally (Table 2.2). PH ESI-MS MINEQL Complex Dissolved Ion mean ± 1SDa Complex 50% dioxane water 4.4 S [(Cu"LACj)Hj]+ 96.5 ±0.6 (Cu"L)+ 98.1 99.7 [Cu"L2H]+ 3.5 ± 0.6 Cu"L2 1.9 0.3 6.8 [CU"L]+ 44.6.± 0.4 (CU"L)+ 59.0 63.3 [CU"L2H]+ 55.4 ±0.4 CU"L 2 41.0 36.7 9.5 [Cu'Lf 15.1 ± 1.1 (Cu"L)+ 0.2 0.3 [CU"L2H]+ 84.9 ±1.1 Cu"L2 99.8 99.7 4.4 E [(Cu'^c^Hj]* 96.2 ± 0.0 (Cu"L)+ 98.1 99.7 [Cu"l_2H]+ 3.8 ±0.0 Cu"L2 1.9 0.3 6.9 [CU"L]+ 47.5 ± 0.3 (CU"L)+ 52.0 56.4 [CU"L2H]+ 52.5 ± 0.3 CU"L 2 48.0 43.6 8.9 [CU"L]+ 8.4 ±0.3 (CU"L)+ 0.9 1.0 [Cu"l_2H]+ 91.6 ±0.3 CU"L 2 99.1 99.0 a Duplicate analyses, j = 0 or 1. Table 2.2 Percent relative abundance of copper(ll)/8-hydroxyquinoline complexes detected by ESI-MS in (a) 50:50 water/acetonitrile and (b) 50:50 water/methanol compared with that predicted by MINEQL in water and 50:50 water/dioxane. 30 While results for aqueous ACN and aqueous MeOH show qualitative agreement, those for aqueous MeOH fit the model more closely under neutral and basic conditions. Differences between theoretical and experimental values for basic solutions may be due to (i) gas-phase dissociation of [Cu"L2H]+, (ii) a drop in pH during electrospray [99], or (iii) low availability of protons, since ESI-MS detection of Cu"L2 depends on protonation of the neutral complex, probably at one of the nitrogen atoms (Figure 2.1b). Results obtained during optimization showed that [Cu"l_]+ was not detected in basic solution when a cone voltage of 20 V was used (Figure 2.2). To see if this holds true for other metal ions, the basic iron(lll) oxine solution was re-analyzed at 20 V. Under these conditions [Fe"l_2]+ was not observed, indicating that Fe m L3 is actually the dominant species in solution at pH 8.7. A cone voltage of less than 30 V is therefore recommended to ensure that dissolved-phase interactions are accurately represented. MINEQL calculations also showed that for acidic aqueous solutions containing 10 pM oxine and 1 to 50 u.M copper, the concentrations of total copper (x^ and Cu"l_+ (y,) are linearly related (y, = 9.91 x 10"3.Xi + 4.02 x 10'9; r2 = 0.9994 at pH 4.2). Likewise, [63Cu l lL]+ intensity (y2), as measured from the reconstituted m/z 207 ion chromatogram, increases linearly with 1 to 50 u.M copper (x2) for acidic 50:50 water/ acetonitrile solutions containing 10 uM oxine (y2 = 2.38 x 1011.x2 + 1.53 x 106; r2 = 0.9990). The average blank response plus 3 standard deviations [107] corresponds to a copper concentration of 0.9 u.M, at which the signal:noise ratio for m/z 207 is approximately 5:1. This gives a limit of detection for Cu"L+ of 13 nM, which lies within the concentration range for copper complexing ligands in seawater [62, 64]. 31 Apparent low-level iron contamination resulted in the detection of [Fe l l lL2]+ at copper concentrations below 10 U.M, illustrating the importance of competition between different metal ions when studying metal speciation (Section 1.2). 2.4.4 Dimethylglyoxime Complexes Dimethylglyoxime (H2L) behaves as a weak monobasic acid, forming (M"HL)+ and M"(HL)2 complexes with several divalent transition metal cations and MIM(HL)3 complexes with trivalent cobalt [102]. In M"(HL)2 complexes the metal ion is coordinated by four nitrogen atoms in a square planar arrangement, which is further stabilized by intramolecular hydrogen bonding (Figure 2.3). Figure 2.3 Formation of a neutral divalent metal-ion complex by dimethylglyoxime. The most stable dimethylglyoxime complexes in 50:50 water/dioxane are those of copper(ll), nickel(ll) and cobalt(lll) [102]; these were the only complexes detected at and above neutral pH by ESI-MS (Table 2.3). Oxidation of cobalt(ll) to cobalt(lll) was apparent in neutral and basic solution {cf. 8-hydroxyquinoline, section 2.4.2), while polynuclear cobalt(lll) and nickel(ll) complexes were detected under basic conditions. 32 Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.9 6.8 8.9 4.5 6.8 8.9 4.6 6.8 8.5 [M'(HL)H]+ 100 72 — [M"HL(HY)j]+ — — — 54 — — 56 — — [M"(HL)2H(HY)j]+ — 28 — 46 100 96 42 — — [M"2L(HL)(HY)j+ _ _ _ _ _ _ 2 - -[M"2(HL)4H]+ _ _ _ - - 4 - _ -[Mm(HL)2]+ — 93 55 [M"'2L2(HL)r - 1 -[MMI2L(HL)3l* — 6 27 [MmM"2L2(HL)2]* — — 17 n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N03~ Table 2.3 ESI-MS speciation for dimethylglyoxime (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). Cobalt(lll) and nickel(ll) complexes were also observed in neutral aqueous MeOH. Although [Cu'(HL)H]+ (m/z 179,181) dominated copper speciation in aqueous ACN, only copper(ll) complexes were detected in aqueous MeOH. The selectivity shown by dimethylglyoxime under neutral and basic conditions contrasts sharply with its behavior at low pH. Under acidic conditions all divalent metal ions except copper(ll) gave rise to [M"HL(HY)j]+ and [M"(HL)2H(HY)j]+ species (Y = acetate or nitrate). Although complexes of iron(lll) and other trivalent metals were not observed, iron(ll) complexes such as [Fe"(HL)2H]+ (m/z 285-289) and [Fe"2(HL)4H]+ (m/z 569-577) were detected in acidic solutions containing no added iron. The formation of Fe 2 + by electrochemical reduction of Fe 3 + during electrospray is discussed in the following section. 33 2.4.5 W,W-Diethyldithiocarbamate Complexes The diethylammonium salt of /V,/V-diethyldithiocarbamic acid (HL) dissolves freely in water to give an alkaline solution [102]. Consequently the pH of solutions containing this compound tended to be somewhat higher than those of other ligands. ty/V-diethyldithiocarbamic acid (pKa = 3.95) forms bidentate anionic ligands (S,S~) with a preference for soft metal ions [102]. Metal-ion coordination via both sulfur atoms to give four-membered chelate rings has been confirmed by X-ray crystallography [106]. The order of stability for M"L2 chelates in dimethylsulfoxide is given as Zn 2 + < Pb 2 + ~ Cd 2 + < Ni 2 + < Cu 2 + [108]. Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + Cd 2 + pH 4.9 7.4 9.2 5.3 7.9 8.9 5.3 7.9 9.1 5.4 7.4 9.2 [M' 4L 3r 4 - -[M'sUr 4 - -[M"L]+ — 10 — 28 — — 19 — — 7 — — [MML 2f 92 88 92 44 40 21 64 — — 16 — — [M"l_2H]+ _ _ _ 8 22 48 — — — _ _ _ [M"2L3]+ — 2 7 — — 5 _ _ _ _ _ _ [M"2UH]+ _ _ _ - - 2 _ _ _ _ _ _ [M"3L5r - - 1 - - _ _ _ _ _ _ _ [M"(L2)L]+ — — ,— 20 19 23 17 — — 77 100 100 [MmL2]+ _ 85 81 [M"'L 3f • - 12 3 [M'"L3H]+ _ _ 8 [M'"2L5r , - - 3 [M'^L^L,]* _ 3 5 n = initial oxidation state. (L2) = neutral dimer (see text). Table 2.4 ESI-MS speciation for /V,/V-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 34 The preference of A/,/V-diethyldithiocarbamate for these metals was illustrated by the fact that cadmium, nickel and copper complexes were detected in aqueous ACN (Table 2.4) whereas trivalent chromium, iron and aluminum complexes were not observed. The tendency of A/,A/-diethyldithiocarbamate to form complexes containing metals with unusually high formal oxidation states [5] was demonstrated by ESI-MS. In acidic aqueous ACN, for example, divalent metals gave ions corresponding to [ M n l L 2 ] + . The expected oxidizing nature of higher valence transition metals such as Cu 3 + , together with the known ease of oxidation of L~, suggest that it is the ligand rather than the metal ion which is oxidized [5] (Section 3.4.1). Such ionic complexes are therefore best represented as radical cations, i.e. [M"L2]'+- Copper(ll), nickel(ll) and cobalt(lll) A/,A/-diethyldithiocarbamates are known to be electroactive, and ESI-MS has been used to detect ionic complexes formed by electrochemical oxidation of Cu "L 2 , NiNL 2 and CoL L LL 3 in acetonitrile/0.1M tetrabutylammonium hexafluorophosphate [109]. The same species were detected in 50:50 water/acetonitrile without prior electrolysis (Figure 2.3), presumably as a result of the electrolysis process inherent in electrospray [110-112]. Comparison of experimental oxidation potentials for these complexes (0.17, 0.37 and 0.55 V, respectively) [109] with the standard half-cell potential for reduction of Fe 3 + to Fe 2 + (0.77 V) [105] suggests that Fe 3 + reduction can bring about spontaneous oxidation of the neutral complexes. For example: Cu"L 2 + Fe 3 + ^ Fe 2 + + (Cu"L2)*+ E = +0.60 V (2.5) 35 100-n (a) [Co raL2]+ 355 N [(L2)H]+ 297 371 387 o .in \ c—N [ComL3H]+ 504 g 1 0 0 c ro c < CD > '•«-> _ro CD _ 0 100 (b) 297 [NinL2Hf 355 [Ni nL 2r 354 . V l L ,n—C, (c) III T + [CuuL2) 359 V <s>-«; [Nin(L2)L]+ 502 >-<><>-< J~l-275 3b0 325 350 375 4 6 6 425 450 475 560 525 S S I ^ Figure 2.4 ESI mass spectra of /V,/V-diethyldithiocarbamate complexes in 50:50 water/acetonitrile at pH 9. Oxidation of cobalt(ll) in basic solution is illustrated by the protonated cobalt(lll) complex at m/z 504 (a). Assignment of the peak at m/z 354 to the electrochemical oxidation product [58Ni"L2]'+ (b) is corroborated by detection of the copper(ll) analogue at m/z 359 (c). Strong copper-ligand interactions are indicated by the absence of [(L2)H]+ (m/z 297) from the copper(ll) spectrum (c). Relative abundance of 100% corresponds to 1.0 x 107 counts. 36 Oxidation by Fe 3 + was explored by adding small quantities of iron(lll) nitrate to nickel(ll) complexes in basic aqueous ACN. Addition of 100 nM Fe 3 + increased the [MNiY2]'+/[58Nil'L2H]+ (m/z 354/355) intensity ratio from 0.5 to 1.2, suggesting that Fe 3 + reduction facilitates electrochemical oxidation. Higher concentrations of added Fe 3 + produced no further increase in intensity, probably due to the formation of stable iron(lll) oxyhydroxides. An increase in 'free', unhydrolysed Fe 3 + at lower pH is therefore expected to promote oxidation of Ni"L2 to [NiML2]'+, which is consistent with the observed increase in [58Ni"L2]'+/[58Ni"L2H]+ intensity ratio with decreasing pH (Table 2.4). The appearance of iron(ll) complexes in acidic solutions containing no added iron provides further evidence for Fe 3 + reduction (Section 4.4.2, p. 73). It is therefore apparent that neutral A/,A/-diethyldithiocarbamate complexes are ionized during electrospray as a result of dissolved Fe 3 + reduction. Contact between acidic solutions and stainless steel components (e.g. syringe needle, electrospray capillary) under oxic conditions seems the most likely source of Fe 3 + contamination [86]. Increased selectivity of the ligand at higher pH was generally accompanied by a shift in speciation towards larger complexes. For example, nickel, cadmium and lead gave [M"(L2)L]+ complexes containing the neutral dimer (L2) (see below). Polynuclear complexes [ C O I " 2 L . 5 ] + (m/z 858) and [Ni"2L3]+ (m/z 560-564) were also detected, while oxidation of cobalt(ll) was confirmed by [Co , l lL3H]+ (m/z 504; Figure 2.3a). Protonation of neutral complexes such as Ni"L2 probably occurs at one of the tertiary amino groups (Figure 2.3b). Copper showed a strong tendency to form polynuclear complexes with A/,A/-diethyldithiocarbamate. Copper(l) complexes [Cu'4L3]+ (m/z 696-704) and [Cu'sL^f 37 (m/z 907-917) were only observed in acidic aqueous ACN, probably as a result of strong eopper-ACN interactions (Section 2.4.10). The binuclear copper(ll) complex [Cu"2L3]+ (m/z 570-574) was detected in both solvents, but only in aqueous MeOH under acidic conditions. Stabilization of copper(ll) in aqueous MeOH was confirmed by . [Cu"L2H]+ (m/z 360,362). A distinctive feature of A/,/V-diethyldithiocarbamate ESI mass spectra was the appearance of low intensity peaks 16 and 32 m/z units higher than those corresponding to ligand and complex ions (Figure 2.3a). Such peaks may be due to the incorporation of oxygen atoms as a result of ligand hydrolysis [113]. The formation of disulfide bonds to give the neutral dimer bis(diethylthio-carbamyl) disulfide [105] (L2) was confirmed by [(L2)H]+ (m/z 297) (Figure 2.3) and [M"(L2)L]+ complexes. Peaks corresponding to [L]+ (m/z 148) and diethylammonium (m/z 74), the ligand counterion, were also observed. Ramping down the cone voltage from 30 V reduced the intensity of [L]+ while that of [(L2)H]+ reached a maximum at 15 V, suggesting that [(L2)H]+ undergoes facile collision-induced dissociation via heterolytic cleavage of the disulfide bond: [(L2)Hf - [Lf •+ [HL] (2.6) 2.4.6 Diphenylthiocarbazone Complexes Diphenylthiocarbazone, or dithizone, is expected to behave as a dibasic acid (H2L) due to the presence of thiol and imino protons; however, only a single dissociation constant has been determined (pKa = 4.47), which is consistent with an intramolecular hydrogen-bonded thiol structure [102]. Under acidic conditions dithizone should act as a bidentate anionic ligand (N,S~) with a preference for soft metal cations. 38 Metal Ion (Mn+) Mn 2 + Co 2 + Ni 2 + Zn 2 + Cd 2 + Pb 2 + PH 4.5 4.5 4.5 4.5 4.5 4.4 E [M"LY]+ 67 74 70 51 43 7 [M"L(HL)]+ 4 1 4 5 2 15 [M"L2]2+ 18 13 15 13 — 14 I [M"l_2Y]+ 6 7 4 7 8 12 [M"L2(HL)]+ 1 1 1 7 3 6 2 [M"L3Y]+ 4 4 5 18 42 42 [MMU]2+ — — — — 3 5 n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . (HL) = deprotonated H2L (see text). Table 2.5 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in acidic 50:50 water/acetonitrile (% relative abundance). N > H _ H 2 L (256 Da) L (254 Da) Figure 2.5 Proposed formation of unprotonated diphenylthiocarbazone (L) from the thiol tautomer (H2L). 39 The order of stability for MM(HL)2 complexes in very dilute aqueous solution is given as Mn 2 + < Zn 2 + ~ Co 2 + < Pb 2 + ~ Ni 2 + < Cd 2 + < Cu 2 + [102]. Dithizone complexes formed by these metals were detected by ESI-MS under acidic conditions (Table 2.5); however, the structures did not correspond to those proposed in the literature. For example, [Cu'L]+, [CU'L2]+, [M"L2]2+ and [MHL4]2+ complexes contain an alternative form of the ligand (L), in which the two exchangeable protons are apparently replaced by an additional nitrogen-nitrogen bond (Figure 2.5). This is consistent with free ligand ions [LH]+ (m/z 255), [L2H]+ (m/z 509) and [L3H]+ (m/z 763), and the mixed ligand complexes [M"LAC]+ and [MMLN03]+ detected in acidic solution. Binding of the metal ion in such complexes probably occurs via coordinate covalent bonds formed by donation of nitrogen and/or sulfur lone-pair electrons (Figure 2.6). Detection of [M"L(HL)]+ and [MnL2(HL)]+ implies that H2L and L forms can coexist. Trivalent metal-ion complexes of dithizone were hot observed. The binuclear copper(l) complex [Cu'2LN03]+ (m/z 442-446) was detected in acidic and neutral aqueous ACN (Figure 2.6). Stabilization of copper(ll) in acidic aqueous MeOH was indicated by [Cu"l_]2+ (m/z 158.5,159) and [Cu"l_iN03]+ species, although no copper complexes were observed at higher pH. With the exception of [Cu'Li]+ complexes in aqueous ACN, none of the metals investigated formed detectable complexes with dithizone at or above neutral pH in either solvent. 40 100n CD O c CO "D c _ < CD > ro CD _ [cu^r 317 [Cu^CACN)]* 358 [Cu^LNOs]* 4421 [ C i r L z f 571 325 350 375 400 425 450 475 500 525 550 575 600 Figure 2.6 ESI mass spectrum of copper and diphenylthiocarbazone in 50:50 water/acetonitrile at pH 4.2, showing reduction of copper(ll) to copper(l) by acetonitrile and the formation of complexes by the unprotonated neutral ligand (L). Relative abundance of 100% corresponds to 1.2 x 106 counts. 41 Although dithizone complexes tend to be brightly colored, most solutions were similar in color to the free ligand (i.e. green at low pH, orange at neutral pH and above), suggesting that no significant complexation had taken place. The colors of zinc(ll), cadmium(ll) and lead(ll) solutions, however, were characteristic of the respective primary dithizone complexes [102]. The apparent inability of ESI-MS to detect such complexes may be due to the formation of colloids and/or precipitates, which were frequently observed when ligand and metal solutions were combined. 2.4.7 Glutathione Complexes The reduced form of glutathione (y-Glu-Cys-Gly) is involved in several important biochemical processes, many of which depend on the reactivity of the cysteinyl thiol group [114]. The tripeptide (H3L), which is essential in maintaining the ferrous state of hemoglobin, is formed in red blood cells by NADPH reduction of oxidized glutathione (H4L2), the disulfide-bonded dimer [115]. Both forms were detected, as [(H3L)H]+ (m/z 308) and [(H4L2)H]+(m/z 613), under neutral and basic conditions by ESI-MS of the reduced ligand solution. The analogous oxidation of cysteine to cystine has also been reported [116]. Other reduced ligand species [(H3L)2H]+ (m/z 615) and [(H3L)3H]+ (m/z 922) indicate significant ligand-ligand interaction without disulfide bond formation (Section 4.4.3). Oxidation of H3L to H4L2 may be facilitated by Fe 3 + reduction, since the corresponding standard electrode potential (0.23 V) [115] is favorable and iron(ll)/ glutathione complexes were detected in acidic solutions containing no added iron (cf. A/,A/-diethyldithiocarbamate, section 2.4.5). 42 100 (a) [Cu n(H3L 2)] + 674 V-~Y~NH 2 HO <^ r P l T > f<v,../fo/VMfcFV> rMr, .h-r* fifft A, i d) o c (0 "O c < 0) a: 100 (b) N C [Nin(H3L2)]+ 669 i f > N H 2 H O -" \ l h 100-, (c) [(H4L2)H]+ 613 615 ' [(H3L)2H]+ i66 ' 610 620 ' 630 646 650 660 670 680 690 700 Figure 2.7 ESI mass spectra of glutathione species in 50:50 water/acetonitrile at pH 9. Copper(ll) and nickel(ll) form 1:1 complexes with the oxidized ligand (H4L2) in basic solutions (a) and (b). Both oxidized and reduced glutathione (m/z 613 and 615, respectively) were detected in a solution containing no added metal ions (c). Relative abundance of 100% corresponds to 8.0 x 105 counts. 43 Divalent transition metals formed doubly charged complexes [M"(H3L)]2+, [M"3(HL)2.]2+, [M"3HL(H2L)2]2+ and [MM4(HL)3]2+ in acidic solution, indicating that these metals displace protons more readily from certain binding sites than from others. This observation is consistent with the presence of carboxyl, amino and thiol groups in (H3L). In contrast, 'cadmium(ll), lead(ll), aluminum(lll) and chromium(III) formed singly charged complexes containing one metal ion. Hydrated chromium(lll) complexes [Cr"'HL(H20)]+ (m/z 373-377) and [Cr"'HL(H20)2]+ (m/z 391-395) reflect the tendency of Cr 3 + to interact with oxygen-donating ligands [1]. Iron(lll) complexes of glutathione were not observed. The only glutathione complexes detected in basic solutions were those of copper(ll) and nickel(ll), which are known to form stable square-planar complexes with tripeptides [117]. In both cases the metal was bound exclusively by oxidized glutathione, giving [Ni"(H3L2)]+ (m/z 669,671) and [Cu"(H3L2)]+ (m/z 674,676). The high stability of these complexes was indicated by the absence of free ligand ions in the ESI mass spectra (Figure 2.7a,b) and by stabilization of copper(ll) in the presence of acetonitrile (cf. 8-hydroxyquinoline, section 2.4.2). Additional copper(ll) complexes [Cun(H3L2-Glu)]+ (m/z 545,547) and [Cu,,(H4L2)]2+ (m/z 337.5,338) were observed in aqueous MeOH. 2.4.8 Salicylic and Phthalic Acid Complexes These compounds are often used to model the. complexation of metal ions by large natural organic ligands such as humic and fulvic acids [1]. Both are dibasic acids (H2L), forming bidentate ligands (0~,0") with an expected preference for hard metal 44 ions [1]. Deprotonated free ligand ions [HL]~ were detected in salicylic acid and phthalic acid solutions (m/z 137 and 165, respectively) by ESI-MS. Phthalic acid also gave rise to [(H2L)HL]~ (m/z 331), [(H2L)2HLJ- (m/z 497) and [(H2L)3HL]~ (m/z 663) ions. These are similar to free ligand ions observed for glutathione in positive-ion mode, and are probably formed, as a result of hydrophobic or hydrogen bonding interactions in aqueous solvent (Section 4.4.3, p. 78). Under acidic conditions, transition metal complexes containing up to three salicylate and five phthalate ligands were detected. The role of these metal ions in forming large, multi-ligand complexes was demonstrated by the fact that; unlike phthalic acid, salicylic acid did not form oligomeric free ligands (Section 4.4.3). The tendency of both ligands to interact with hard metal ions was illustrated by aluminum(lll), chromium(lll) and iron(lll) complexes (Table 2.6). Metal Ion (Mn+) Al 3 * Cr3* Fe3 + pH 3.8 6.1 3.9 6.2 3.6 6.1 59 — 4 — 27 — 10 — n = initial oxidation state, i, j = 0, 1 or 2. [M'"L(N03)2]~ 55 - 26 -E[M I I IL(N0 3) 2(H 20)J" — — 40 — [MLLL(HL)(N03)3]- - - _ 79 Z [ M \ 2 ( H N 0 3 ) J ] ~ — — 26 21 £ ^ " 'MHNO^HzCOJ" 35 — 8 — [M"'2L2(N03)3]" 4 - - -[MMM"L 2(N0 3) 2]" — — — _ [M"' 2 L 3 N0 3 ]" 6 — — — Table 2.6 ESI-MS speciation for salicylic acid (H2L) and selected metal ions; in 50:50 water/acetonitrile (% relative abundance). 45 100n o c CO "D c _ < CD > ro cu f_ [ArL(N03)2] 287 * [AlmL2], 299 ^ [Alm 2L 3N0 3)] 524 [Alm2L2(N03)3] 512 2 7 5 3 0 0 3 2 5 3 5 0 3 7 5 4 0 0 4 2 5 4 5 0 4 7 5 5 0 0 5 2 5 5 5 0 m/z Figure 2.8 ESI mass spectrum of aluminum(lll) and salicylic acid in 50:50 water/acetonitrile at pH 3.8, showing the formation of binuclear and mixed ligand complexes under acidic conditions. Relative abundance of 100% corresponds to 1.5 x 106 counts. 46 In acidic aqueous ACN, salicylate complexes [M"lL(N03)2]" and [M l l lL2(HN03)j]" were observed for all three trivalent metal ions, along with hydrated chromium(lll) and binuclear aluminum(lll) complexes (Figure 2.8). The absence of aluminum(lll) and iron(lll) complexes at higher pH is probably due to the formation of insoluble oxyhydroxides, while chromium(lll) complexes were detected at pH 6.2 in both solvents. Similar results were obtained for phthalic acid, except that addition of iron(IH) yielded iron(ll) complexes. Most of the mixed ligand salicylate and phthalate complexes detected under acidic conditions contained nitrate rather than acetate, showing that nitrate remains strongly associated with metal ions in acidic solution (Section 2.4.9). Copper(ll) was not stabilized by salicylic acid or phthalic acid aqueous ACN, forming [Cu'(HL)2j" complexes with both ligands under acidic conditions. In aqueous MeOH, however, [Cu"L(HL)(HN03)j]~ complexes indicated that the higher oxidation state had been preserved. Partial fragmentation of salicylate and phthalate ligands was indicated by [(HL-C02)]~ ions, and by copper and chromium complexes containing (L-C02)2~. The intensities of these ions were highly dependent on electrospray cone voltage, suggesting that decarboxylation occurs as a result of collision-induced dissociation in the ESI-MS interface (Section 4.4.4, p. 80). 2.4.9 Other Complexes Dissociation of acetic acid (pKa = 4.76) at experimental pH produces significant quantities of acetate (Ac~), a bidentate oxygen-donating ligand [5]. This competes with the added organic ligands for dissolved metal ions, forming complexes such as [MMAc]+, [Mm(Ac)2]+ and [(MnLAc)H]+ (L = oxine) in acidic solutions. Neutral solutions, which contain the same total concentration of acetate, do not yield such species, which 47 strongly suggests that their formation is a result of solution- rather than gas-phase interactions. A decrease in the abundance of acetate complexes at higher pH reflects increased competition by ligands with larger dissociation constants and/or higher affinities for particular metal ions. The use of nitrate salts also gave rise to inorganic complexes such as [M"N0 3] + and [Mm(N03)2]+ in acidic solution, suggesting that incomplete dissociation of acetate and other organic ligands allows nitrate to compete for the metal ions. Negative-ion ESI-MS of acidic solutions detected [M"(N03)3r and [Cr'"(N03)30H]~ species in the absence of acetate analogues, indicating that nitrate forms complexes with higher ligand:metal ratios than the larger acetate ligand. 2.4.10 Solvent Effects For the purposes of metal speciation it is important that organic co-solvents do not interfere significantly with the equilibria established in aqueous solution. A 50:50 mixture of water and organic solvent was used in an attempt to preserve aqueous speciation while achieving optimum electrospray conditions. During these investigations, aqueous ACN consistently gave higher overall ion yields than aqueous MeOH (typically 2- to 4-fold in positive-ion mode), especially under neutral and basic conditions. However, ACN also reduced copper(ll), as indicated by the detection of [Cu'(ACN)i]+ clusters in acidic aqueous ACN, rather than the [M"(ACN)i]2+ ions observed for other divalent cations. ESI-MS analysis also showed that strong interactions between copper(ll) and organic ligands such as 8-hydroxyquinoline and glutathione apparently preserve the higher oxidation state in aqueous ACN, while other 48 ligands such as salicylic and phthalic acid formed copper(l) complexes in the same solvent. In contrast, almost all the copper complexes detected in aqueous MeOH contained copper(ll). The effect of these organic co-solvents on ion yield can be rationalized in terms of their molecular properties. The dipole moment of ACN (3.92 D) is much greater than that of MeOH (1.71 D) or H 20 (1.85 D) [105]. This enhances the basicity of the nitrogen lone-pajr, allowing ACN to displace H 20 molecules coordinated to metal ions in aqueous solution more effectively than MeOH (Section 1.1). Consequently [M(ACN)j]n+ ions were detected in preference to [M(H20)j]n+ in aqueous ACN, while only in preliminary experiments using unbuffered solvents (pH ~2) were [M(MeOH)j]n+ clusters observed. The tendency of ACN to interact strongly with metal cations (particularly Cu2+) may affect speciation in cases where metal-ligand interactions are relatively weak (e.g. hard ligands and soft metal ions). Solubility and electrostatic metal-ligand interactions are both influenced by the dielectric constant (s) of the solvent [118]. The addition of ACN or MeOH (e = 38.8 and 32.6, respectively) to water (e = 78.5) increases the solubility of neutral complexes, which may help to promote electrospray ionization. 2.5 Summary ESI-MS is a powerful method for characterizing dissolved metal-ion complexes in aqueous organic solvents, including those commonly used in reversed-phase liquid chromatography. Accurate determination of dissolved metal-ion speciation, however, requires mild electrospray conditions to preserve solution-phase interactions, and depends on the quantitative ionization of neutral species by protonation, deprotonation 49 or the electrolysis process inherent in electrospray. The latter, which appears to involve Fe 3 + reduction, can also affect the speciation of electrochemically labile complexes through ligand oxidation and the formation of Fe2 +. Optimum electrospray conditions and the preservation of copper(ll) in aqueous acetonitrile depend on the thermodynamic stabilities of the species involved. Detection limits are compatible with the levels of organic ligands found in natural waters, although competition between different metal ions for ligands at low concentrations merits further investigation. CHAPTER 3. ALKALI AND ALKALINE EARTH METAL SPECIATION BY ELECTROSPRAY IONIZATION MASS SPECTROMETRY 50 3.1 Abstract Alkali and alkaline earth metal ions were used to investigate how ligand oxidation, solution pH and metal ion properties affect electrospray ionization of dissolved metal species. Metal ions and selected organic ligands were combined in 50:50 water/acetonitrile and analyzed using electrospray conditions optimized to preserve solution phase interactions. Alkali metal ions formed detectable complexes with bidentate oxygen-donating ligands and ligands containing both amino and thiol groups under acidic and neutral conditions. Greater complexation of alkali metals at higher pH was indicated by a corresponding decrease in the relative abundance of the free metal ion M+. In accordance with thermodynamic equilibrium theory the number of alkali metal ions bound by glutathione and phthalic acid ligands also increased with increasing pH. Alkaline earth metal species were detected only in acidic solutions, the absence of 8-hydroxyquinoline complexes being attributed to their relatively low thermodynamic stability The use of metals with fixed oxidation states demonstrated that ionization of neutral A/,A/-diethyldithiocarba-mate complexes takes place via oxidation of the ligand rather than the metal ion. 3.2 Background The ability of electrospray ionization mass spectrometry (ESI-MS) to analyze large dissolved organic compounds intact has revolutionized the study of proteins and other biomacromolecules [75]. When coupled with separation methods such as high performance liquid chromatography and capillary zone electrophoresis, the technique can be used to 51 detect and quantify such compounds in a variety of biological and environmental matrices [88]. Interactions between dissolved species can also be investigated; for example, the structures of noncovalent biomolecular complexes [119] and the metal binding stoichiometry of certain proteins [82-84] have been successfully determined using ESI-MS. The ability of many organic compounds to form adducts with metal ions in solution can assist in their identification and structural elucidation [93, 94, 120, 121]. These and other applications involving the detection of metal ion complexes suggest that ESI-MS has considerable potential as a technique for metal speciation [91,100]. Realization of this potential requires an understanding of the processes by which dissolved metal species are ionized and detected by ESI-MS. The following describes how the unique properties of alkali and alkaline earth metal ions have been used to study the effects of ligand oxidation, pH and metal ion characteristics on electrospray ionization. In common with other class A cations (e.g. Al3+), alkali and alkaline earth metal ions have hard, spherically symmetrical electron shells which are not easily polarized [1]. Such cations tend to form complexes with ligands having oxygen as the donor atom via predominantly electrostatic interactions. In contrast, class B cations (e.g. Cd 2 + and Pb2+) have soft, easily polarized electron shells, and tend to form covalent bonds with donor atoms of low electronegativity, such as sulfur [1]. Transition metal ions are generally of intermediate character, although Cr3 + and Fe 3 + show a greater affinity for oxygen than for sulfur, while Mn 2 + has a relatively low propensity for covalent bonding [1]. Differences in metal ion reactivity have important implications for biological and environmental aqueous systems [122]. For example, Cd 2 + and Pb2+ are known to disrupt calcium metabolism in 52 vertebrates [123], while relatively high concentrations of Ca 2 + affect both the rate and extent of trace metal complexation by marine dissolved organic matter [124]. New evidence also suggests that Ca 2 + and Mg 2 + play a significant role in the formation of marine colloids, which constitute up to 40% of the dissolved organic matter found in seawater [125]. The predominance of Na+, K+, Mg2 + and Ca 2 + in natural waters [1, 3] makes it important to be able to determine speciation for these major cations as well as the trace metal ions with which they compete. In an earlier study of the complexes formed by divalent and trivalent metal ions with selected organic ligands (Chapter 2), it was shown that dissolved metal speciation determined by ESI-MS compares favorably with that predicted by thermodynamic equilibrium theory, provided that the appropriate electrospray ionization conditions are used. Some results, however, proved inconclusive due to the choice of metal ions, which was made primarily on the basis of their importance as trace elements in natural aqueous systems [1, 3]. The formation of radical cations by transition metal A/,A/-diethyldithio-carbamate complexes, for example, could be due to oxidation of either the ligand or the metal ion. The apparent reluctance of metals other than copper to form detectable species at or above neutral pH also limited the investigation of changes in organic and inorganic speciation under different solution conditions. In the present study, the same organic ligands and experimental conditions are used to detect the species formed by alkali and alkaline earth metal ions. Such ions are highly resistant to electrochemical oxidation (i.e. electron removal), as indicated by the corresponding ionization energies (e.g. 496 and 4562 kJ/mol for Na and Na+, respectively) [126]. Preliminary studies show that Na+ and K+ also have a strong tendency to form detectable complexes under neutral as well as acidic 53 conditions, allowing pH related changes in speciation to be observed. The objectives of the present study were i) to identify the species formed by Na+, K+, Mg 2 + and Ca 2 + in solution, ii) to study the effect of ligand oxidation on speciation by using metal ions with fixed oxidation states, and iii) to investigate pH related changes in the distribution of organic and inorganic species formed by Na+ and K+. Results from earlier investigations (Chapter 2) are used where appropriate (e.g. Tables 3.1-3.3) to clarify ion formation mechanisms. A complete data set is given in Appendix A (pp. 147-177). 3.3 Experimental 3.3.1 Materials All reagents were certified ACS grade or better. Nitrates of sodium, potassium, magnesium and calcium were obtained from Aldrich Chemical Co., together with salicylic acid and phthalic acid, Glutathione (reduced form) was purchased from Sigma Chemical Co., and 87hydroxyquinoline (oxine), diphenylthiocarbazone, dimethylglyoxime and A/^-diethyldithiocarbamate (die.thylammonium salt) from BDH Inc. Acidic, neutral and basic solvents of similar ionic strength were prepared by dissolving 1 mmol acetic acid, ammonium acetate or ammonium hydroxide in 500 mL of doubly deionized Milli-Q water (Millipore) and an equal volume of HPLC grade (BDH) acetonitrile (ACN). Stock solutions were combined in 4 mL of each solvent to give metal and ligand concentrations of 100 u.M, with the exception of oxine, for which a concentration of 10 u.M was necessary to avoid memory effects. The pH of each sample was measured to an accuracy of ±0.1 unit using a combination calomel reference electrode and pH meter (Accumet Model 915, Fisher Scientific) following ESI-MS analysis. 54 3.3.2 Mass Spectrometry Samples were analyzed by flow injection ESI-MS in the manner described previously (Section 2.3.2). Electrospray conditions were as follows: capillary 2.9 kV, high voltage (HV) lens 600 V, cone 30 V, (second) skimmer 30 V. Positive ion mode was used for all ligands except salicylic acid and phthalic acid, which form anionic species at experimental pH. Mass spectra were scanned in the m/z range 20-1000 at a rate of 10 s/scan, with an interscan delay of 10 ms. The elution of sample components was monitored by means of reconstituted single ion chromatograms. These showed that under basic conditions Na+ (m/z 23) and K+ (m/z 39) were partially retained in the sample introduction system while organic ligands eluted without retention. Although interaction between M+ and deprotonated hydroxyl sites was suspected, replacement of the fused-silica capillary with polyetheretherketone (PEEK) or stainless steel tubing had little effect. Results for basic solutions were therefore omitted. Likewise, partial retention of Mg 2 + and Ca 2 + complexes formed by salicylic acid and phthalic acid in neutral solutions resulted in cross-contamination, invalidating results for these samples. For the remaining samples, scans corresponding to the typical non-retained elution profile were combined and processed using the Masslynx software (version 2.1). The relative abundance of dissolved ionic species was calculated from the corresponding peak intensities in the mass spectrum, as previously described (Section 2.3.3, p. 21). Metal speciation was divided into three categories: A) complexes formed with the added ligand (e.g. glutathione); B) other metal species (i.e. [M]+, [M"(ACN)i]2+, nitrate and acetate complexes); C) distribution of total metal speciation between A and B. Results were expressed as percent relative abundance within each category (Tables 3.1-3.3). The 55 representation of each ligand in neutral form indicates the number of exchangeable (acidic) protons; viz. oxine and A/,/V-diethyldithiocarbamate (HL); diphenylthiocarbazone, dimethylglyoxime, salicylic acid and phthalic acid (H2L); and glutathione (H3L). Oxidized A/,A/-diethyldithiocarbamate, which contains a disulfide bond in place of thiol protons, is represented as (L2), while diphenylthiocarbazone forms complexes with alkali and alkaline earth metals as the unprotonated neutral ligand (L) (Section 2.4.6). 3.4 Results and Discussion 3.4.1 Ligand Oxidation ESI mass spectra for A/,A/-diethyldithiocarbamate (HL) and alkali metal ions are shown in Figure 3.1. Peaks corresponding to [NaL]'+ {m/z 171) and [KL]'+ (m/z 187) suggest that ionization of neutral rvl'L complexes takes place via electrochemical oxidation: M'L ^ [M'L]'+ + e- (3.1) Likewise, magnesium and calcium ions yield [M"L2]'+ (Table 3.1, Section A). In both cases, detection of radical cationic complexes containing metals of fixed oxidation state (Section 3.2) indicates that it is the ligand which is oxidized rather than the metal ion. Oxidation presumably takes place as a consequence of the electrolysis process inherent in electrospray [110-112]. Simultaneous detection of Fe 2 + complexes for certain ligands (e.g. glutathione; Section 2.4.5) suggests that contaminant Fe 3 + acts as the required electron acceptor. 56 Figure 3.1 ESI mass spectra for alkali metals and /V,/V-diethyldithiocarbamate in 50:50 water/acetonitrile at pH 5.4. Detection of [M'l_]'+ complexes for sodium (a) and potassium (b) is consistent with electrochemical oxidation of the ligand. Relative abundance of 100% corresponds to 2.5 x 106 counts. 57 Metal Ion (Mn+) Na+ K+ Mg2 + Ca 2 + Mn 2 + PH 5.4 8.0 5.4 8.0 5.3 8.1 5.3 8.1 5.3 8.0 A) [M'Lf 34 61 35 50 [M'(L2)]+ 66 39 65 50' [M"L]+ — — — — 44 — [MML 2f 68 — 82 — 44 — [M"(L2)L]+ 32 — 18 — 12 — B) [M]+ 65 92 64 87 2 [M'2Y]+ 35 8 36 13 L [M"(ACN)i]2+ 30 — 47 — 56 — 2 [M"Y(HAc)j(ACN)i]+ 70 — 53 — . 44 — C)EA) 37 22 27 10 6 — 6 — 14 — £ B) 63 78 73 90 94 — 94 — 86 — n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N0 3 . (L2) = neutral dimer (see text). A) Carbamate complexes. B) Other species. C) All metal species. Table 3.1 E S I - M S speciation for A/,A/-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). A/,A/-Diethyldithiocarbamate can itself be oxidized to give the disulfide (L2): 2(HL) ^ (L2) + 2H + + 2e~ (2) The oxidized ligand is subsequently detected as [(L2)H]+ (m/z 297). Protonation likely occurs at the tertiary amino group (Figure 3.1a). Alkali metal ions form analogous species [Na(L2)]+ (m/z 319) and [K(L)2]+ (m/z 335), suggesting that sodium and potassium are also coordinated via nitrogen (cf. diphenylthiocarbazone; Table 3.2, Section A ) . The polarity of such complexes is somewhat greater than if the metal were bound by sulfur, due to the position of the metal cation. This effect is even more pronounced for [M'L ] ' + complexes, in which dissociation of the thiol group results in a zwitterion (Figure 3.1b). 58 100, N MX10 (a) 136.8 [Cu^ACN)]* 103.8 ti (b) [Ca(ACN)2r 60.9 [CaAc(HAc)]H 158.8 [CaAc]+ 98.8 u [CaAc(ACN)]+ 139.8 80 ' 100 1 2 0 1 4 0 160 180 i m/z Figure 3.2 ESI mass spectra showing species formed by potassium (a) and calcium (b) with acetate, nitrate and acetonitrile (ACN) in 50:50 water/acetonitrile at pH 4.5. Peaks at m/z 104,106 indicate contamination by copper, which is stabilized in the +1 oxidation state by acetonitrile. Relative abundance of 100% corresponds to 5.5 x 107 counts. 59 Any increase in polarity is likely to enhance solubility in aqueous solvent, and hence electrospray ionization efficiency. Alkali metal ions are larger and considerably more polarizable than the isoelectronic alkaline earth metal ions [5], which explains why dissolved alkali metal species are so amenable to electrospray ionization. This also applies to uncomplexed alkali metal ions, which are detected as [M]+ in both acidic and neutral solutions (Figure 3.2a). In contrast, solvated magnesium and calcium ions are only observed under acidic conditions, as [M"(ACN)i]2+ (Figure 3.2b). The detection of uncomplexed metal in neutral solutions appears to be characteristic of monovalent cations (cf. Cu+, Figure 3.2a), and allows pH related changes in the distribution of free and complexed metal to be studied using ESI-MS. 3.4.2 pH Solution pH can affect metal speciation in two ways; (i) directly, as a result of the competition between H + and metal ions, and (ii) indirectly, through its effect on ligand oxidation. Oxidation of A/,A/-diethyldithiocarbamate, for example, involves deprotonation of the reduced ligand (Eqtn. 3.2), and is therefore likely to be more extensive at higher pH (Section 4.4.2, p. 73). This process effectively reduces the number of available ligands, and since alkali metal complexes [M'L]'+ and [M'(L2)]+ both have 1:1 metal:ligand stoichiometry the amount of sodium and potassium bound by N,N-diethyldithiocarbamate actually decreases with increasing pH (Table 3.1, Section C). The pH of acidic A/,A/-diethyldithiocarbamate solutions (c. 5.4) is somewhat higher than for other ligands, so that the amount of acetate present as acetic acid (HAc) is relatively small (less than 20%). A further increase in pH therefore has little effect on the concentration of free acetate (Ac-), and hence the abundance of [M2'Ac]+ 60 complexes (Figure 3.2a). Since nitric acid dissociates almost completely in aqueous solution [127], complexation by nitrate also remains largely unaffected by pH. Consequently, most of the alkali metal ions released by /V,/V-diethyldithiocarbamate at pH 8.0 remain uncomplexed, as indicated by an increase in the relative abundance of [M]+ (Table 3.1, Section B). For ligands other than A/,A/-diethyldithiocarbamate, the pH of acidic solutions is below the pKa of acetic.acid (4.76) [105], so that raising pH significantly increases complexation by acetate, reducing the relative abundance of [M]+ (Tables 3.2 and 3.3, Section B). Metal Ion (Mn+) Na* K+ Mg2 + Ca 2 + Mn 2 + pH 4.5 7.3 4.5 7.3 4.6 7.3 4.5 7.3 4.5 7.3 A) .[M'Lf 92 92 90 92 [M'L2]+ 8 8 10 8 £[MMLY]+ 94 — 96 — 67 — [M"l_(HL)]+ — — — — 4 — [MHL2]2+ — — — — 18 — 2[M"L 2Yf 6 — 4 — 6 — [M"L2(HL)]+ — — — — 1 — 2[M"L3Y]+ — — — — 4 — B) [M]+ 96 79 96 85 £[M'2Y]+ 4 21 4 15 S [M"(ACr\l)i]2* 62 — 62 — 55 — 2 [M"Y(HAc)j(ACN)i]+ 38 — 38 — 45 — C) £A) 12 12 6 7 1 — 1 — 1 — ZB) 88 88 94 93 99 — 99 — 99 — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . (HL) = deprotonated H 2 L A) Dithizonate complexes. B) Other species. C) All metal species. Table 3.2 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 61 Unlike A/,A/-diethyldithiocarbamate, which binds alkali metal ions in both oxidized and reduced form, the extent of metal complexation by other organic ligands (e.g. diphenylthiocarbazone, glutathione) should, according to thermodynamic equilibrium theory, increase with increasing pH. However, complexation by acetate appears to limit the proportion of total alkali metal bound by these other ligands at higher pH (Tables 3.2 and 3.3, Section C). Metal Ion (Mn+) Na* K+ Mg2 + Ca 2 + Mn 2 + PH 4.2 6.6 4.2 6.6 4.2 6.7 4.3 6.7 4.3 6.7 A) [M'x(H4-xL)]+ 89 100 89 100 [M'X(H4.XL)(H3L)]+ 10 — 9 — [MIX(H4.XL)(H3L)2]+ 1 — 2 — • [M"(H2L)]+ 25 — 23 — 30 — [M"(H3L)]2+ 68 — 71 — 61 — .[M"x(H4.2xL)f 2 — 2 — 2 — [M" 3(HL) 2] 2 + - 3 — 2 — 4 — [M"3(HL)(H2L)2]2+ • 1 — 1 — 2 — [M"4(HL)3]2t • > ' ' 1 — 1 — 1 — B) [M]+ 100 77 100 97 S [M'2Y]+ — 23 — 3 E [M"(ACN)i]2+ 73 — 75 — 67 — I [M"Y(HAc)j(ACN)i]+ 27 — 25 — 33 — C)ZA) 11 15 3 7 4 — 5 — 4 — ZB) 89 85 97 93 96 — 95 — 96 — n = initial oxidation state. i. j = 0, 1 or 2. Y = A c - or N0 3~. x = 1,2 or 3. A) Glutathionate complexes. B) Other species. C) All metal species. Table 3.3 ESI-MS speciation for glutathione (H3L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 62 Competition between H + and M + can also be expected to influence the number of alkali metal ions bound by polyfunctional organic ligands under different pH conditions. This is demonstrated by phthalic acid (H2L), for which noncovalent interactions (Section 4.4.3, p. 78) apparently give rise to large ligands capable of binding several alkali metal ions. The relative abundance of sodium complexes formed by (H2L)4, for example, shows the expected change in distribution with pH (Figure 3.3). 50 1 2 3 4 5 6 7 Number of Na+ions bound (x) Figure 3.3 Complexation of sodium ions by the phthalate ligand (H2L)4 in 50:50 water/acetonitrile at pH 4.1 (•) and pH 6.7 (•). Percent relative abundance was calculated from the intensities of [M^H/.xU)]-complex ions detected by ESI-MS. Similar results are obtained in positive ion mode for glutathione (H3L), which can bind up to three alkali metal ions in place of acidic protons (see Appendix A, p. 164). 63 3.4.3 Metal Ion Properties Results for Na+ and K+ suggest that alkali metal ions are more likely to interact with oxygen and nitrogen donor atoms than with sulfur [1] (Section 3.2). This trend, which is characteristic of class A metal ions, is also observed for Mg 2 + and Ca 2 + . When combined with diphenylthiocarbazone, for example, these metal ions do not form the [MnL2]2+ and [M"l_i(HL)]+ species observed for divalent transition and class B cations (Section 2.4.6]. Instead, the alkaline earth metals form mixed ligand complexes containing nitrate or actetate (Table 3.2, Section A), which is consistent with coordination by nitrogen and oxygen atoms only (Figure 3.4b). Likewise, magnesium and calcium do not form [MML]+ complexes with /V,/V-diethyldithiocarbamate (Table 3.1, Section A), suggesting that alkaline earth metals are bound via a different mechanism, i.e. by interaction with nitrogen rather than chelation by sulfur. For glutathione, which contains carboxyl, thiol and amino groups [115], the formation of doubly as well as singly charged ionic complexes with magnesium and calcium (Table 3.3, Section A) suggests that alkaline earth metals displace protons more easily from certain binding sites than others. Similar results are obtained for divalent transition metal ions (Section 2.4.7]. In contrast, class B cations Cd 2 + and Pb2+, which have a high affinity for sulfur [122], are more efficient in displacing protons from thiol groups and do not, therefore, give rise to doubly charge complexes. 64 100-, MX10 O c CO " D C 3 < ro (b) 255 240 [MgLAc]+ 3 3 7[MgLN0 3f / 340 MX10 [CaLN03]+ 356 [CaLAc]+ 353 v. 2 6 0 2 8 0 Ibo 3 2 0 3 4 0 3 6 0 m/z Figure 3.4 ESI mass spectra for alkaline earth metals and diphenylthio-carbazone in 50:50 water/acetonitrile at pH 4.5. Formation of the neutral, unprotonated ligand (L) is indicated by the free ligand ion [LH]+ at m/z 255 and the mixed ligand complexes formed by magnesium (a) and calcium (b). Relative abundance of 100% corresponds to 1.5 x 107 counts. 65 Detection of complexes formed by nitrogen-donating ligands oxine (HL) and dimethylglyoxime (H 2L) depend on the relative magnitudes of their thermodynamic stability constants. For example, no oxine complexes were observed for alkali or alkaline earth metals, although the ligand is known to form M N L 2 complexes with magnesium and calcium [102]. Similar;results are observed for Mn 2 + which, like class A metal ions, has a relatively low propensity for covalent bonding [1]. This is illustrated by conditional stability constants for oxine complexes Mg"l_2, Ca"l_2 and Mn"l_2 in aqueous dioxane (log 02 = 11.8, 13.2 and 15.5, respectively) [102]. These values are significantly lower than those for the corresponding Cu 2 + and Ni 2 + complexes (log (32 = 26.22 and 21.38, respectively), which are readily detected by ESI-MS [102] (Table 2.1). Alkali and alkaline earth metal complexes of dimethylglyoxime are not described in the literature [102]. Nevertheless, complex ions [M"HL(HY)J]+, [M"(HL)2H(HY)J]+ and [M"2L(H]L)(HY)J]+ (where Y = acetate or nitrate) were observed for magnesium and calcium in acidic solution. This is consistent with the affinity of Mg 2 + and certain other divalent metal ions for square-planar nitrogen-donating ligands such as tetrapyrroles (e.g. chlorophyll) [5]. The planar geometry of these systems presumably allows for extensive interaction between the complexed metal ion and peripheral organic and inorganic ligands (Y). The distributions of Co 2 + and Ni 2 + dimethylglyoxime complexes detected by ESI-MS show a significant shift from [M"HL(HY)J]+ to [M"(HL)2H(HY)j]+ when compared with magnesium, calcium and Mn 2 + speciation. This is consistent with the high stability of Co"(HL) 2 and Ni"(HL)2 relative to Mn"(HL)2 (log p2 = 21.0, 21.7 and 11.8, respectively, in aqueous dioxane) [102]. 66 3.5 Summary Investigations involving alkali and alkaline earth metal ions show how ligand oxidation, solution pH and the properties of the metal ion affect metal speciation by electrospray ionization mass spectrometry. Results also suggest that detection of dissolved metal species depends on their polarity and thermodynamic stability. Detection of Fe 2 + complexes for certain ligands demonstrates the utility the method for simultaneous speciation of class A and transition metal ions. 67 CHAPTER 4. DETECTION AND ELECTROCHEMICAL OXIDATION OF DISSOLVED ORGANIC LIGANDS BY ELECTROSPRAY IONIZATION MASS SPECTROMETRY 4.1 Abstract The relationship between solution chemistry and electrospray ion yield was investigated for a group of organic ligands previously used in metal speciation studies. Ligands containing basic groups were detected as protonated cations and those containing only acidic groups as deprotonated anions in 50% acetonitrile under acidic, neutral and basic conditions. The abundance of dimethylglyoxime, salicylate and phthalate ions at experimental pH was predicted using the thermodynamic equilibrium model MINEQL and the relevant dissociation constants. Differences between theoretical and experimental ion yield were attributed to the displacement of acid-base equilibria during electrospray. Oxidation of glutathione and /V,/V-diethyldithiocarbamate via the electrolysis process inherent in electrospray was indicated by detection of the corresponding disulfide compounds in solutions containing the reduced ligand. The formation of iron(II) complexes by glutathione suggests that Fe 3 + is involved in the electrolytic process. Noncovalent (hydrogen bonding) interactions between glutathione and phthalic acid ligands were also observed, although mild electrospray conditions were necessary to maintain the structural integrity of dissolved ligand species. Results suggest that electrospray ionization of uncharged metal-ion complexes in solution depends primarily on the acid-base and electrochemical properties of the ligand. 68 4.2 Background Electrospray ionization mass spectrometry (ESI-MS) is now widely regarded as the method of choice for analyzing labile organic compounds in solution, due to the combination of sensitivity, specificity and 'soft' ionization characteristics [75, 78, 88] (Section 1.4). ESI promotes the efficient conversion of dissolved polar and ionic species to gas phase ions while helping to preserve solution phase interactions [74, 75]. Structural information can also be obtained by controlled fragmentation of the intact molecular ions produced by electrospray [93, 97, 119, 128, 129]. These attributes make ESI-MS an ideal method for studying dissolved phase interactions between metal ions and organic ligands, which form the basis of many important biological and geochemical processes [1, 5]. Electrospray ionization of analytes which are not intrinsically charged depends on proton or electron transfer [79, 87]. This has important consequences for the use of ESI-MS as a technique for metal speciation. Detection of the neutral 1:2 coppenligand complexes formed by 8-hydroxyquinoline (oxine), salicylic acid and A/,A/-diethyldithio-carbamate, for example, relies on protonation, deprotonation and electrochemical oxidation, respectively (Sections 2.4.2, 2.4.5 and 2.4.8). Ionization of neutral complexes via these processes is dependent on solution conditions; sensitivity towards the copper(ll)/oxine complex CuL2, for example, is reduced at high pH since fewer protons are available to form CuL 2H + (Section 2.4.2, p. 30) The purpose of this study was to investigate how acid-base and electrochemical properties influence the detection of organic ligands, and hence their metal-ion complexes, by ESI-MS. Ligands in acidic, neutral and basic 50% acetonitrile were analyzed by ESI-MS, and the observed ion yield compared with the distribution of ionic 69 species predicted using the appropriate dissociation constants and reduction potentials. The tendency of certain ligands to form hydrogen bonds in solution or undergo fragmentation in the gas phase was also investigated. Complexes of sodium, potassium and copper were used to clarify the role of the ligand in forming some of the ionic species detected by ESI-MS. 4.3 Experimental Experimental conditions and analytical procedures were the same as those used previously to determine dissolved metal speciation by ESI-MS (Section 2.3, p. 19). Square brackets are again used to denote the gas phase ions detected by electrospray, while curved brackets symbolize the concentrations of dissolved species. As before, neutral ligands are represented as HXL, where x indicates the number of acidic protons, while protonation of basic groups is expressed as HXLH+. 4.4 Results and Discussion 4.4.1 Acid-Base Chemistry Ionization of dissolved ligands by protonation or deprotonation depends on solution pH and the acid-base properties of the ligand. Dimethylglyoxime (H2L), which contains basic amino groups (Figure 4.1a), undergoes protonation to form H2LH+. The abundance of [H2LH]+ (m/z 117) decreases with increasing pH in accordance with the following equilibrium: H 2LH+ ^ H2L + H + (4.1) 70 OH OH 0 (e,307) ( f-145) Figure 4.1 Organic ligands dimethylglyoxime (a), salicylic acid (b), phthalic acid (c), A/,A/-diethyldithiocarbamate (d), glutathione (e) and oxine (f) (molecular mass shown in brackets). Ligand (fully protonated form) 1 PKa 2 3 4 Dimethylglyoxime (H2LH+) 8.7 10.6 11.9 — Salicylic Acid (H2L) 3.0 13.4 — — Phthalic Acid (H2L)a 2.9 5.5 — — A/,A/-Diethyldithiocarbamate (HL) 4.0 — — — Glutathione (H 3LH+) b (Giu)2.2 (Gly)2.4 (Cys)8.3 (Giu) 9.7 Oxine (HLH+) 5.0 9.8 — — a o-phthalic acid. b appropriate values for the free constituent amino acids. Table 4.1 Dissociation constants (Kg) for organic ligands in water at 25°C [102, 105, 115]. 71 The equilibrium distribution of H 2LH+ in solution at experimental pH was modeled using the MINEQL program and the appropriate dissociation constants (Table 4.1). Use of this data to predict the abundance of [H2LH]+ ions detected by ESI-MS was based on the following assumptions: (i) [H2LH]+ intensity is proportional to (H2LH+), and (ii) dimethylglyoxime is fully protonated in the acidic solution (pH 4.5). The former has been demonstrated for the analyte concentrations used in this study (i.e. 100 nM or less) [78], while the latter can be confirmed by MINEQL calculations. Predicted [H2LH]+ intensity (I) was subsequently calculated using the relationship: I = (H 2 LH + ) x l p H 4 5 / (H 2 LH + ) p H 4 . 5 (4.2) where I p H 4.5 is the.observed ion intensity at pH 4.5 and (H2LH+)p H 4.5 and (H2LH+) are calculated concentrations at pH 4.5 and other pH values, respectively. Experimental results showed qualitative agreement with the predicted speciation (Figure 4.2). However, the observed [H2LH]+ intensity dropped more quickly with increasing pH than expected. ESI-MS detection of anions in acidic solution and cations in basic solution is frequently observed [130]. The formation of such ions has been attributed to the transfer of protons and\or a change in the relative basicities of certain compounds in the gas phase [79, 99]. If the pH corresponding to experimental [H2LH+] intensities is increased by a fixed amount (i.e. 2 units) the data show good agreement with the predicted values (Figure 4.2). This suggests that the acid-base equilibrium (eqn. 4.1) is displaced during electrospray, possibly as the result of a decrease in the basicity of dimethylglyoxime on passing from solution into the gas phase. 72 4 5 6 7 8 9 10 11 pH Figure 4.2 Theoretical ES ion intensity I (—) and observed intensity (•) for protonated dimethylglyoxime [H2LH]+ (m/z 117). Displacement of the acid-base equilibrium (eqn. 4.1) during electrospray is suggested by the close agreement between I and experimental values when the corresponding pH is increased by 2 (o). Salicylic acid and phthalic acid (H2L) (Figure 4.1b,c) form anionic species in solution which are detected by negative ion ESI-MS. H2L ^ HL- + H + (4.3) HL- ^ . L2- , + H + (4.4) The equilibrium distributions of HL~ and L2" were calculated using the same procedure as for dimethylglyoxime. Qualitative agreement between ES ion yield and theoretical speciation was again observed. For example, [L]2- was detected in neutral and basic solution for phthalic acid (m/z 82) but not for salicylic acid (m/z 68), since the phenolic group of salicylic acid (pKa = 13.4, Table 4.1) does not dissociate under experimental conditions (i.e. pH 4-9). However, differences between predicted and 73 observed abundance for [HL]~ again suggest that proton exchange can occur during electrospray. /V,/V-Diethyldithiocarbamate (HL), glutathione (H3L) and oxine (HL) (Figure 4.1d-f) undergo protonation in a manner analogous to dimethylglyoxime. Unfortunately oxidation of the thiol ligands A/,/V-diethyldithiocarbamate and glutathione (Section 4.4.2) precludes straightforward interpretation of their acid-base chemistry, while the tendency of oxine to adhere to system components leads to memory effects which invalidated results for this ligand. Such effects do, however, appear to be mitigated as a result of metal complexation, since the addition of copper(ll) decreases the abundance of the free ligand ion [HLH]+ (m/z 146). Consequently, ESI-MS speciation of the copper complexes formed by oxine shows good agreement with the equilibrium distribution predicted by MINEQL (Section 2.4.3, p. 29). 4.4.2 Redox Chemistry Organic compounds containing thiol groups (RSH) can be oxidized to form disulfides [131]: 2RSH ^ RS—SR + 2H + + 2e" (4.5) Oxidation of /V,/V-diethyldithiocarbamate (HL) and glutathione (H3L) during electrospray is indicated by detection of the corresponding disulfide compounds [L2H]+ (m/z 297) and [H4L2H]+ (m/z 613) (Figure 4.3) in solutions containing the reduced ligand. 74 (a, 296) Figure 4.3 Oxidized ligands A/,/V-diethyldithiocarbamate (a) and glutathione (b) (molecular mass shown in brackets). It has been demonstrated that the electrolysis process inherent in electrospray results in the sequential oxidation (or reduction, in negative ion mode) of solution components based on their relative concentrations and reduction potentials [110]. However, ESI-MS detection of Fe 2 + complexes in solutions containing no added iron (Figure 4.4) suggests that ligand oxidation also involves the reduction of contaminant Fe 3+ 75 100, CD O c CO •o c zs < CD > 100, ro <u /a> [(H3L)2H]+ (b) 615 [Na(H3L)2]+ 637 [Fen(H2L)(H3L)]+ 669 [Na2(H2L)(H3L)]+ 659 F J,,,WYA-^AT>J [Na3(HL)(H3L)]+ 681 W L i i i i i i i i i i i |-669 [K(H3L)2]+ 653 60b 610 6 2 0 630 640 650 660 670 680 6* 690 Figure 4.4 ESI mass spectra of glutathione with (a) sodium and (b) potassium in 50:50 water/acetonitrile at pH 4.2, showing the Fe 2 + complex (m/z 667-671) formed as the result of iron contamination. 76 A comparison of standard reduction potentials for glutathione (-0.23 V) [115] and Fe3 +/Fe2 + (0.77 V) [105] shows that the following reaction (which is analagous to the biological regeneration of ferrous hemoglobin) [115] (Section 2.4.7, p. 41) proceeds spontaneously. 2(H3L) + 2Fe 3 + ^ 2Fe 2 + + H4L2 + 2H + E = + 0.54 V (4.6) Similar calculations indicate that transition metal /V,/v-diethyldithiocarbamate complexes can also be oxidized by Fe 3 + to form the radical cations detected by ESI-MS (Section 2.4.5, p. 34). Electrochemical oxidation of iron in the stainless steel electrospray capillary seems the most likely source of Fe 3 + contamination [86]. Charge transfer reagents (e.g. quinones) have been used to facilitate ionization and ESI-MS detection of compounds such as polyaromatic hydrocarbons, which are not amenable to protonation or deprotonation [132]. The present study suggests that oxidation by contaminant (or added) Fe 3 + may be effective in ionizing compounds with suitable reduction potentials, although the analytical utility of this approach requires further investigation [133]. Oxidation of A/,A/-diethyldithiocarbamate and glutathione involves dissociation of thiol groups (eqn. 4.5), and is therefore pH dependent. Detection of [L2H]+ at pH >5 is consistent with a pKa of 4.0 for /V,/V-diethyldithiocarbamate (Table 4.1). Oxidized glutathione [H4L2H]+, however, was not observed at pH <7, since the corresponding (cysteinyl) thiol group has a pKa of 8.3. Because ESI-MS detection of L2 and H4L2 depends on protonation, the abundance of these ions is expected to decrease with increasing pH. 77 Figure 4.5 ESI mass spectrum of phthalic acid in 50:50 water/acetonitrile at pH 8.6, showing the [(H2L)iHL]~ ions formed as the result of intermolecular hydrogen bonding. Relative abundance of 100% corresponds to 2 x 107 counts. 78 However, [L2H]+ shows the opposite trend, suggesting that oxidation is the dominant process governing the formation of this ion. An increase in A/,A/-diethyldithio-carbamate oxidation with increasing pH is confirmed by studies involving alkali metal ions (Section 3.4.1), which also show that ligand oxidation is responsible for the ionization of ML„ carbamate complexes. 4.4.3 Noncovalent Interactions The formation of [(H2L)jHL]~ ions by phthalic acid (H2L) indicates a substantial degree of noncovalent interaction between these ligands (Section 2.4.8). Noncovalent association of organic compounds in aqueous solution can occur as a result of hydrophobic interactions or intermolecular hydrogen bonding. Salicylic acid does not give rise to [(H2L)jHL]~ ions, despite the structural similarities between salicylic and phthalic acids (Figure 1b,c). Intramolecular hydrogen bonding is sterically favored in salicylic acid molecules (Figure 1b), whereas the additional carbon atom inhibits this process in phthalic acid. Consequently phthalic acid is much more likely to form intermolecular hydrogen bonds [134], as confirmed by ESI-MS analysis (Figure 4.5). Peptides such as glutathione (H3L) also participate in hydrogen bonding [115], as indicated by [(H3L)2H]+ (m/z 615) and [(H3L)3H]+ (m/z 922). Hydrogen bonding by amino acid residues depends on pH; glutamic acid, for example, only acts as a H-donor when protonated [115]. Relative ion intensities suggest that hydrogen bonding between glutathione ligands increases with increasing pH (Figure 4.6). However, rapid electrochemical oxidation to H4L2 at higher pH may also decrease the relative abundance of H3L, assuming that (H3L)j species remain unaltered during electrospray. 79 PH Figure 4.6 Relative abundance of glutathione species [H3LH]+ (•), [(H3L)2H]+ (o) and [(H3L)3H]+ (•) in 50:50 water/acetonitrile as a function of solution pH. The biggest change in relative abundance occurs on going from acidic to neutral pH (Figure 4.6). This coincides with the onset of glutathione oxidation (Section 4.4.2), supporting the above hypothesis. Detection of [(H3L)jH]+ ions in basic solution suggests that dissociated carboxyl groups are involved (as H-acceptors)- in the formation of intermolecular hydrogen bonds. Once occupied byrnetar ions, however, these groups can no longer participate in hydrogen bonding. This can be demonstrated by adding 100 u.M alkali metal to buffered glutathione solutions. At pH 4.2 the concentrations of H + and M + are similar, and only some of the carboxyl groups are occupied by M+. Hydrogen bonding between glutathione ligands is therefore maintained, as indicated by the [Mx(H(4.x)L)(H3L)i]+ 80 complexes detected by ESI-MS (Figure 4.4). At pH 6.7 the ratio of metal ions to protons is much greater (about 500:1), and most of the carboxyl groups are occupied by M+. Consequently only [Mx(H(4-x)L)]+ complexes are detected under neutral conditions (Table 3-3). , f ... ; . ; > 4.4.4 Gas Phase Fragmentation Fragmentation of oxine complexes in the gas phase was apparent during metal speciation studies, but was reduced by lowering the cone voltage of the electrospray ion source (Section 2.4.1). Similar experiments performed on ligand-only solutions showed that salicylate and phthalate ligand ions [HL]~ also undergo fragmentation to give [HL-C02r (m/z 93 and 121, respectively). The distributions of [HL]- and [HL-C02]-at different cone voltages suggest that decarboxylation becomes significant above 20 V. A similar experiment was performed after adding copper to salicylic acid (Figure 4.7). In this instance, fragmentation resulted in sequential decarboxylation of the complex ion [Cu'(HL)2]- (m/z 337,339). Stepwise removal of C 0 2 suggests that the bonds between the metal ion and the carboxyl groups each ligand are of different strengths. This is also true of dissolved 1:1 and 1:2 coppensalicylate complexes, which have stability constants of 10 1 1 5 and 10 1 9 3, respectively, in aqueous solution [1]. Thus, controlled fragmentation during electrospray illustrates the stepwise formation and dissociation of metal-ion complexes [1, 2, 5]. 81 1.6E+6 • : ES cone voltage (V) Figure 4.7 Distribution of salicylic acid complex ions [Cu'(HL)2r (•). [Cu^HLXHL-COz)]" (A) and [Cu'(HL-C02)2r (•) as a function of electrospray cone voltage. 4.5 Summary ESI-MS detection of organic ligands, and hence their uncharged metal-ion complexes, depends on solution pH and the acid-base properties of the ligand. Ionization by protonation or deprotonation is, however, affected by the displacement of acid-base equilibria during electrospray, possibly as a result of electron transfer in the gas phase. Electrochemical oxidation of glutathione and A/,/V-diethyldithiocarbamate by Fe 3 + and/or the electrolysis process inherent in electrospray depends on pH and the dissociation constants of the respective thiol groups. Noncovalent interactions between glutathione and phthalic acid ligands are attributed to intermolecular hydrogen bonding. Results show that the distribution of complex ions detected by ESI-MS depends on electrospray conditions as well as the intrinsic properties of the ligand and metal ion. 82 CHAPTER 5. CHARACTERIZATION OF DISSOLVED TANNINS AND THEIR METAL ION COMPLEXES BY ELECTROSPRAY IONIZATION MASS SPECTROMETRY 5.1 Abstract Electrospray mass spectrometry was used to identify the organic ligands obtained by dissolving tannic acid in aqueous solution. Ligand ion intensities were proportional to tannic acid concentration, with an estimated detection limit of 5 nM for individual phenolic compounds in 50:50 water/acetonitrile. Complexes formed by tannin ligands with copper and other biogeochemically important metals were also identified, allowing metal binding stoichiometry and oxidation state to be directly observed. Simultaneous detection of free and complexed tannins enabled the effects of metal concentration, pH and ionic strength on complexation to be investigated. The structure and principal copper binding site for one of the tannins were also determined, providing unique information with regard to metal complexation by these polyfunctional ligands. The technique is compatible with procedures for the extraction and on-line liquid chromatographic separation of dissolved compounds. Results demonstrate the potential of this approach for characterizing natural organic ligands which play an important role in trace metal biogeochemistry. 5.2 Background Complexation by organic ligands can have a profound effect on the bioavailability and geochemistry of dissolved metal ions [4, 19, 32, 40, 56]. In oceanic water columns, trace metals such as copper, iron and zinc are associated almost exclusively with dissolved organic matter (DOM) [58, 61, 62, 65, 66]. At least two 83 distinct classes of organic ligand responsible for binding copper have been recognized [62, 65]. The stronger and more specific of these is found in surface waters and is rich in both primary amines and carbohydrates, suggesting recent in situ production by photosynthetic marine organisms (phytoplankton, cyanobacteria) [71]. The other predominates in deeper water and shows levels of fluorescence characteristic of aromatic compounds, a feature associated with refractory organic matter of high oceanic residence time [3, 71]. Concentrations of DOM in estuaries and coastal waters can be substantially higher than those found in the open ocean [3]. A significant proportion of this material is of terrestrial origin and includes phenolic compounds such as lignins and tannins [3, 135]. The deleterious effect of these compounds on the growth of marine algae has been attributed to the complexation of biolimiting trace metals [135]. Previous studies have also shown that organic complexation can reduce copper toxicity and increase the availability of iron to certain marine organisms [36, 51, 68, 136]. However, many of the ligands and complexes involved in such processes have yet to be identified [71, 137]. New methodology for detecting and analyzing these species is therefore needed to provide a better understanding of their role in trace metal biogeochemistry. Electrospray (ESI) is a 'soft' ionization technique which permits the analysis of dissolved polar and ionic species by mass spectrometry (MS) [75, 78, 79] (Section 1.4). The technique can be applied to ionic species and those capable of acquiring charge via protonation, deprotonation, association with other ionic solutes or electron transfer [79, 87, 96]. The nature of the electrospray ionization process (Section 1.4.3) allows large, fragile compounds to be analyzed intact, a feature which has revolutionized the study of proteins and other biopolymers. Dissolved phase interactions between such 84 compounds and metal ions form the basis of many essential biochemical processes, including respiration, detoxification and photosynthetic electron transport [4, 5 ]. ESI-MS is well suited to the study of such interactions, having been used to investigate metal complexation by proteins and the association of iron with apoferrodoxin [82-85, 88]. These and other applications involving the detection of metal ion complexes suggest that the technique has considerable potential for dissolved metal speciation [95]. This potential has been explored using a range of synthetic organic ligands and metal ions of known biogeochemical importance (Chapter 2). The results of ESI-MS analysis compare favorably with aqueous speciation predicted by thermodynamic equilibrium calculations (Section 2.4.3, p. 27), provided that suitable electrospray conditions are used. This methodology has been adapted to study metal complexation by tannic acid, a component of natural DOM [3, 135]. Preliminary investigations show that tannic acid dissolves in aqueous solution to yield tannins and other potential organic ligands [138]. The aims of this study were (i) to identify and measure the distribution of dissolved organic compounds derived from tannic acid, (ii) to investigate metal complexation by these compounds under different solution conditions (pH, ionic strength, total metal concentration), and (iii) to determine the structures of ligand and complex species detected by ESI-MS. The dependence of ion yield on solvent composition and electrospray conditions (cone voltage, source temperature, drying gas flow rate) was also investigated. Application of the technique to natural water samples requires suitable extraction and separation procedures. Solid phase extraction (SPE) and size exclusion chromatography (SEC) were evaluated for the recovery and separation of tannins prior to ESI-MS analysis. 85 5.3 Experimental Tannic acid (Merck Index 11,9023) and metal nitrates (ACS grade or better) were obtained from Aldrich Chemical Co. and used without further purification. Solvents were prepared by dissolving 1 mmol ammonium acetate (SigmaUltra) or ammonium hydroxide (Fisher ACS) in 500 mL doubly deionized Milli-Q water (Millipore) and an equal volume of HPLC grade acetonitrile (ACN) or methanol (MeOH). Dilute aqueous solutions of tannic acid (0.2-u.m filtered) and metal nitrate were combined in each solvent to give metal and/or ligand concentrations of 400 and 100 u.M, respectively, unless otherwise indicated. Solution pH was measured using a combination calomel reference electrode and pH meter (Accumet Model 915, Fisher Scientific). Flow injection analysis was carried in the manner previously described (Section 2.3.1). Briefly, 20 u.L of sample were introduced via a rotary valve and delivered to the ESI probe at 10 u.L/min by a carrier stream consisting of the sample solvent. Solid phase extraction was performed by drawing 500 mL of sample through C18 SPE cartridges (J.T. Baker) under vacuum. Extracted compounds were eluted with 5 mL volumes of 10 mM ammonium hydroxide, to which 5 mL of acetonitrile were added prior to flow injection analysis. Size exclusion chromatography was carried out by installing the column (Ultraspherogel SEC 2000, Beckman) between the valve and ESI probe and increasing the solvent flow to 0.5 mL/min. A post-column splitter reduced the eluent flow to 10 pL/min for on-line ESI-MS (Figure 5.1). 86 Figure 5.1 Schematic of the LC-ESI-MS system. A HPLC pump B sample loop C LC column 1 solvent 2 waste 3 sample 4 ESI probe Mass analysis was performed using a Quattro triple quadrupole mass spectrometer (Micromass) fitted with the pneumatically assisted electrospray probe supplied (Figure 5.2). Electrospray conditions (negative ion mode) were as follows; ES capillary 2.9 kV, high voltage (HV) lens 600 V, cone 60 V, (second) skimmer 60 V, temperature 60 °C, drying and nebulizing gas (N2), 7.0 and 0.25 L/min, respectively. During flow injection analysis, mass spectra were scanned by the first quadrupole (MS1) in the m/z range 60-1200 at 10 s/scan with an interscan delay of 10 ms. Scans corresponding to the sample elution profile were combined and processed using the spectrometer software. The distribution of dissolved species was determined from the relative intensities of the corresponding ions in the resulting mass spectrum. 87 A B D E F G F C H G Figure 5.2 Configuration of the Quattro tandem mass spectrometer. A electrospray capillary B high voltage (HV) lens C cone D skimmer E ion transfer lens (hexapole) F mass analyzer (quadrupole) G detector (photomultiplier) H collision cell (hexapole) 1 liquid sample 2 nitrogen supply 3 vent 4 rotary pump 5 diffusion pump For MS/MS experiments MS1 was set to transmit the selected parent ion to the empty collision cell, and the instrument tuned to give the maximum parent ion signal at the second quadrupole (MS2) detector. Collision induced dissociation (CID) of the parent ion was subsequently carried out using a collision cell potential of 80 V and a collision gas (Ar) pressure of 0.05-0.15 Pa. Daughter (fragment) ion spectra were obtained by scanning MS2 in the m/z range 100-600. During SEC experiments selected ion recording (SIR) was used to monitor the elution of individual sample components. 88 5.4 Results and Discussion 5.4.1 Electrospray conditions Electrospray ionization of compounds containing several basic or acidic groups can give rise to multiply charged molecular ions depending on the extent of protonation or deprotonation, respectively [75]. Using an appropriate algorithm (e.g. 'MaxEnf, Micromass) the correct charge state can be assigned to each ion, allowing the molecular weights df large biopolymers to be determined with unprecedented accuracy [139]. Distribution of the total molecular ion intensity between several peaks does, however, tend to complicate the resulting mass spectrum. Electrospray conditions which give satisfactory results for monomeric organic ligands such as oxine (Section 2.4.1) produced intense doubly charged tannin ions. Multiple charging was, however, greatly reduced by increasing the cone voltage to 60 V. Unlike oxine and other smaller ligands, the distribution of singly charged tannin species was apparently unaffected by the increase in cone voltage. This may be due to stabilization through intramolecular hydrogen bonding (Section 4.4.3, p. 78), promoted by the greater size and flexibility of. the tannin ligands. Gas phase ions formed via electrospray retain solvent to a degree depending on source conditions (temperature, drying gas flow rate) and the solvation energy of the analyte [75]. Tannins of molecular mass >1000 were detected primarily as solvent adducts. Raising the source temperature from 60 to 120 °C significantly increased the intensity of desolvated molecular ions; however, thermal degradation of larger tannins was suspected from an increase in the relative abundance of gallic acid (m/z 169). Using a 7.0 L/min drying gas flow rate instead of raising the source temperature gave satisfactory results for larger tannins but had little effect on overall ion yield. 89 D-I'-HT 343 ioo-. 8 c (0 c -Q < 0 <D 100i ro _ - Hr 49S Ji| i ii 11111^  I .r|» (a) [ L , * - H -647 t„i\X [ U - H T 787 [ U ' - H T 799 [ U - H T 939 • v |MM|ll..|l..l||...|l^l|lMI|l.ll|llll|llll|MII|lll1|1ll.|IIM|llll| I U - H T 1091 (b) [L1*-2H + Cu,T LT.ul.i tL 2*-2H + Cu 'r | 557 J50 400 4 5 0 5 0 0 5 M 6 0 0 8 6 0 700 750 800 [U* - 2H + Cu1]-i i T | L ^ M 9 H + C U ' R I U " 2 0 H 1 + C U , R I U " 2 H + C U , R j)4JU»rW) A, 1.1 n | frifr Jl|ii •  i w nfrft 11 M «| i * l i f t . "1' $ * 1153. f*<\ 900 950 1000 10S0 1100 1150 tm/z Figure 5.3 ESI mass spectrum of 100 uM tannic acid with (a) 0 and (b) 400 uM copper(ll) nitrate in 50:50 water/acetonitrile at pH 6. 5.4.2 Tannin Ligands When analyzed under optimum electrospray conditions (Section 5.3), tannic acid solutions yielded two series of molecular ions (Ln and Ln*) separated by 12 mass units (Figure 5.3a). Compounds in each series were separated by 152 mass units, which corresponds to the addition or removal one galloyl moeity (Figure 5.4). This confirms that Ln and Ln* are gallotannins [140, 141]. Ions corresponding to pyrogallol (m/z 125), gallic acid (m/z 169) and ellagic acid (m/z 301) [141] were also detected. 90 (483) (331) (170) Figure 5.4 Hydrolysis of tannin L 2 to tannin Li and gallic acid [141] (molecular mass shown in brackets). The relative abundance of [L„ - H] ~ and [Ln* - H]~~ ions was essentially independent of electrospray conditions, indicating that these compounds were either present in the starting reagent or formed by hydrolysis of tannic acid in solution [141]. The structure of gallotannins (Ln) is based on a D-glucopyranose ring, which suggests that Ln* tannins differ only in the structure of this component. A difference of 12 mass units precludes simple methylation or dehydration of D-glucopyranose, but is consistent with methylation and oxidation of two separate hydroxyl groups (Figure 5.5). OH OH OH (180) (194) (192) Figure 5.5 Formation of methyl-4-ketoglucose by methylation and NAD+ oxidation of D-glucopyranose (molecular mass shown in brackets). 91 Methylation of D-glucopyranose usually occurs at the C1-OH group to give the corresponding glycoside [131]. Such compounds do not normally undergo oxidation because C1-OH is converted to C1-OMe. However, in vivo oxidation of similar compounds (e.g. UDP-glucose) by NAD+ occurs at the C4-OH rather than C1-OH group [141]. This strongly suggests that Ln* tannins, which are based on the modified sugar (methyl-4-ketoglucose), are biologically derived and therefore present in the starting reagent. A peak corresponding to methyl-4-ketoglucose (m/z 191) was actually observed, although the unmodified sugar (m/z 179) was not detected. The formation of larger tannins depends on substitution at glucopyranose Cn-OH groups as well as galloyl phenolic groups [140] (Figure 5.4). Methyl-4-ketoglucose has two less Cn-OH groups than D-glucopyranose; consequently, the abundance of Ln* decreases relative to Ln as n increases (Figure 5.3a). Analysis of solutions containing different concentrations of tannic acid in 50% ACN (pH 6) showed that the intensities of individual tannin ions are proportional to tannic acid concentration over the range 0.5-10 p.M. This suggests that relative ion intensities reflect the distribution of dissolved tannins, although their concentrations cannot be determined directly due to mass dependent transmission by the quadrupole [98, 142]. Calibration curves established using salicylic acid (m/z 137) showed that detection limits of around 5 nM can be achieved for individual phenolic compounds. 5.4.3 Metal Ion Complexes The formation of 1:1 copper:tannin complexes in 50% ACN was confirmed by isotopic species with the characteristic 69:31 6 3Cu: 6 5Cu ion intensity ratio (Figure 5.3b). The intensity distributions of ligand and complex ions peaked at m/z 343 and m/z 557, 92 respectively, suggesting that L2* binds copper most effectively. Relative ion intensities also suggest that copper is bound preferentially by Ln* tannins. The m/z ratios of complex ions detected in 50% ACN indicate that they contain copper(l); for example, [L2* - 2H + Cu 1] - (m/z 557, 559). Previous studies suggest that reduction of copper(ll) by acetonitrile depends on the strength of the competing ligand (Section 2.4.10). Salicylic acid and phthalic acid, which are bidentate oxygen-donating ligands, also yield copper(l) complex ions (Section 2.4.8), suggesting that tannins bind copper in a similar fashion. Only copper(ll) complexes were observed in 50% MeOH; for example, [l_2* - 3H + Cu"] - (m/z 556, 558). However the yield of complex ions was extremely low compared with 50% ACN (Section 2.4.10), requiring an 8-fold excess of the metal salt to achieve a detectable signal. The apparent formation of nitrate adducts such as [L2* - 2H + N03]~ (m/z 556) under these conditions made identification of copper(ll) complexes even more difficult. Aqueous acetonitrile was therefore used as the solvent for all subsequent experiments. The formation of tannin complexes>by other metal ions was also investigated (Table 5.1). Metal Ion (Mn+) Complex Ion Na+ K+ Cu 2 + Mg 2 + Ca 2 + Mn 2 + Co 2 + Ni 2 + Zn 2 + Cd 2 + Pb: fl-n ~4H + M"'j ~ [L n-3H + Fe"j~ [L„ - 3H + 2M'] i h -.2+ Table 5.1 Complex ions generated by electrospray ionization of 50% acetonitrile solutions containing metal nitrates and tannins (Ln). 93 Detection of 2:1 metal:tannin complexes containing sodium and potassium is consistent with the multiple binding of alkali metal ions by other polyfunctional organic ligands (e.g. glutathione) (Section 3.4.2). The apparent complexation of added iron(lll) as iron(ll) may be due to (i) redox reactions involving dissolved organic compounds [143] or (ii) the reductive electrolytic process associated with negative ion electrospray [110]. 5.4.4 Metal Concentration, pH and Ionic Strength Complexation of copper (Cu) by a tannin ligand (L) is described by the following equation (charges omitted for clarity): Cu + L ^ CuL (5.1) The equilibrium constant for this process is equivalent to the conditional stability constant of the complex (Section 1.1, p. 1): KCOND = (CUL)/(CU,).(U) (5.2) where (Cu') and (L') are the concentrations of uncomplexed metal and ligand. Since KCOND is a constant it follows that (CuL)/(L') is proportional to (Cu'), provided that ( L ' ) » 0 . .. CuL and L' are detectable in 50% A C N as [LN* - 2H + Cu']~ and [Ln* - H]~ respectively. Assuming that electrospray ion intensity is proportional to analyte concentration (Section 5.4.2), (CuL)/(L') can be calculated from the observed ion intensity ratio [Ln*- 2H + Cu'p[L n* - H]~. Unfortunately (Cu') could not be detected simultaneously in negative ion mode, so the relationship between (CuL)/(L') and total copper concentration, Cu T, was investigated. 94 100-1 n p i H T i n ^ r [L2* - 2H + Cu1] 557 TVi |W>iT i , , ? r | ( i i i | f i n r <f/V^"r>"|i|'r'iri 7T?MiTT*i|ii — [L 2 - 2H + CuT" 545 ,(irn^!i 460 490 500 510 520 530 540 550 560 570 580 Figure 5.6 ESI mass spectra showing the distribution of L2* ligand and complex ions for 100 uM tannic acid with (a) 0, (b) 100, (c) 200 and (d) 400 uM added copper in 50:50 water/acetonitrile at pH 6. 95 (a) 0.8 T 0.6 -_ l 0.4 A (b) Cu T (mM) Figure 5.7 CuL/L' ion intensity ratio versus total copper concentration for tannin ligands Li* (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 6 and (b) pH 8 (ionic strength ~1 mM). 96 Figure 5.8 CuL/L' ion intensity ratio versus total copper concentration for tannin ligands'Li* (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 6 and (b) pH 8 (ionic strength ~10 mM). 97 Solutions containing 100 n.M tannic acid and 0-400 |iM copper in 50% ACN were analyzed by ESI-MS (Figure 5.6), and the CuL/L' ion intensity ratio for each tannin ligand plotted against Cu T (Figure 5.7). In solutions buffered with 1 mM ammonium acetate (pH 6.0 ± 0.4) CuL/L' increased linearly with Cu T, as predicted for (CuL)/(L') vs. (Cu') (eqtn. 5.2). For solutions containing 1 mM ammonium hydroxide (pH 7.6 ± 0.9), CuL/L' increased more rapidly at lower Cu T, suggesting that reduced competition by H + at higher pH increases copper complexation by tannin ligands. The trend towards constant CuLVL' values at higher Cu T indicates that tannin ligands are not fully titrated by excess copper, since (L') remains significant. This is probably due to hydrolysis of the excess copper [1] which progressively reduces pH and thus inhibits further complexation. Using ammonium acetate and ammonium hydroxide concentrations of 10 mM gave better pH control (6.9 ± 0.3 and 9.4 ±0.1, respectively). However, the resulting increase in ionic strength reduced both the extent of copper complexation (Figure 5.8) and overall ion yield (Section 1.4.4). ML/L' ion intensity ratios for other metal ions also increased with total metal concentration. The increase was most pronounced for nickel which, like copper, tends to form stable complexes with many organic ligands [1]. 5.4.5 Structural Characterization Structural information can be obtained by MS/MS experiments involving collision induced dissociation (CID) of the molecular ions generated by electrospray. CID was performed on the most abundant complex ion [L2* - 2H + 63Cu']~ (m/z 557) and that of the corresponding ligand [L2* - H]~ (m/z 495). Optimization of the molecular ion signal (Section 5.3) involved a decrease in MS1 resolution, which resulted in a certain amount 98 Figure 5.9 ESI-MS/MS daughter ion spectrum for the tannin ligand [L2* - H] (m/z 495) at a collision gas pressure of (a) 0.05 and (b) 0.10 Pa. of peak broadening. Nevertheless, CID of [L2* - H]~ was able to confirm the gallotannin structure by identifying gallic acid (m/z 169) and methyl-44<etoglucose (m/z 191) as structural components (Figure 5.9). Fragmentation of [L2* - 2H + 6 3Cu'] - showed that copper remains attached to gallic acid (m/z 231) following stepwise removal of the second galloyl moeity (m/z 405) and methyl-44<etoglucose (Figure 5.10). Hence copper is bound predominantly by the galloyl moeity attached to the pyranose ring. Similarities in the behavior of salicylate and tannin ligands (Section 5.4.3) suggest that copper is indeed bound to L2* via phenolic groups. 99 _ ( b ) & J1 0 [i • i i • i _' 1 • l i J • » • 1 • «-i m/z Figure 5.10 ESI-MS/MS daughter ion spectrum for the tannin complex [L2* - 2H + 6 3Cu'] - (m/z 557) at a collision gas pressure of (a) 0.10 and (b) 0.15 Pa. Electroanalytical and spectroscopic techniques can only infer the identities of metal binding sites by measuring and comparing stability constants or spectral changes associated with metal complexation [2] (Figure 1.3). Positive identification of metal binding sites further illustrates the unique capabilities of ESI-MS for studying metal complexation by the large polyfunctional organic ligands often found in natural waters (Figure 1.1c). 100 5.4.6 Extraction and Separation Procedures Natural organic ligands should ideally remain in solution throughout extraction and separation to minimize alteration of their original structure [144]. The compatibility of ESI-MS with high performance liquid chromatography (HPLC) allows dissolved compounds to be resolved and analyzed on-line by mass spectrometry [84, 93, 96, 97, 120]. Tannins are hydrophilic compounds and are therefore not well suited to reversed phase HPLC. This was confirmed by preliminary experiments using a C8 column (Lichrospher RP8, Merck) and 0.2% acetic acid in 50% ACN (pH 3.5) as the mobile phase [139]. Normal phase separations [140] require non-polar solvents which are generally incompatible with electrospray ionization, although post-column derivatization has proved successful for compounds which form positively charged Na+ adducts [145]. Gel permeation or size exclusion chromatography (SEC) separates compounds on the basis of hydrodynamic radius rather than chemical composition, and has been successfully applied to the analysis of tannins [139]. The procedure does, however, require derivatization of acidic groups to achieve significant resolution, a process which alters the composition and metal binding properties of the original compounds. The analysis of underivatized tannins by high performance SEC with on-line ESI-MS detection was attempted using 50% ACN containing 1 mM ammonium acetate (pH 6) as the mobile phase. The compounds detected were the same as those identified in the same solvent by flow injection analysis (Section 5.4.2), but were apparently eluted from the SEC column without retention. Slightly better results were achieved using TRIS.HCI buffer (pH 9), as recommended by Swift and Posner [146]. 101 100-1 [L/-Hr 01! TV I ' 10Q1 [U*-2H]2 0'| | n i •!• i i i | • l i • | i • i i | lOOi r i I 399.00 4.46e5 Figure 5.11 SEC-ESI-MS of 100 \iM tannic acid in 50:50 water/acetonitrile containing 0.25 mM TRIS.HCI (pH 9), showing the elution of tannins (Ln*) and gallic acid (G) in order of decreasing molecular weight (mass-to-charge ratio and relative abundance are shown to the right). 102 Elution of compounds in order of decreasing molecular weight was confirmed by coelution of singly and doubly charged ions formed by the same compound, e.g. [L2* - H]~ (m/z 495) and [L2* - 2H]2~ (m/z 247) (Figure 5.1.1). Peaks were about 1 min wide and separated by 0.1 min, which meant that individual tannins could not be resolved. However, this procedure may prove more successful for compounds with larger mass differences and/or for SEC columns with greater resolution at low molecular weights. Organic ligands in natural waters must be extracted before ESI-MS analysis can be carried out (Section 1.4.4, p. 12). Efficient adsorption of organic compounds onto hydrophobic resins such as XAD-2 requires acidification [147], which may denature hydrolyzable compounds such as gallotannins [148]. However, dissolved organic compounds can be isolated from unmodified aqueous samples by C18 solid phase extraction (SPE) [149]. This procedure was applied to deionized water samples containing tannic acid at a concentration similar to that of DOM in a typical estuary (i.e. 6 pM, or 10 mg/L) [3]. Organic solvents (ACN, MeOH) apparently failed to remove hydrophilic compounds from the SPE cartridges. However, small amounts of the retained material were eluted with 10 mM ammonium hydroxide. Flow injection analysis of the extracts showed that Ln* tannins are eluted preferentially from C18 SPE cartridges by aqueous ammonium hydroxide. Care was taken to prevent the retained compounds from drying out during extraction, a process which can lead to irreversible changes in the structures of natural organic ligands [144]. 103 5.5 Summary Dissolved tannins were identified and their metal binding properties investigated using electrospray ionization mass spectrometry. Metal binding stoichiometry, oxidation state and the degree of complexation under different solution conditions were directly observed, while MS/MS experiments confirmed the structure and identified the principal copper binding site for one of the tannin ligands. These results demonstrate the unique capabilities of ESI-MS for studying metal complexation by natural organic ligands. Conventional methods for extracting and separating tannins without matrix modification or derivatization met with limited success. However, mass analysis allowed these oligomeric compounds to be resolved without prior chromatographic separation. Methods for the selective extraction and flow injection ESI-MS analysis of copper complexing ligands in seawater are currently under investigation. 104 CHAPTER 6. EXTRACTION AND CHARACTERIZATION OF COPPER LIGANDS IN SEAWATER BY IMMOBILIZED METAL-ION AFFINITY CHROMATOGRAPHY AND FLOW INJECTION ELECTROSPRAY IONIZATION MASS SPECTROMETRY 6.1 Introduction An ideal method for extracting natural organic ligands should be selective for the compounds of interest and involve minimal alteration of the analyte or sample matrix. Established procedures for isolating DOM from natural waters (e.g. XAD-2 or C18 adsorption) tend to discriminate against hydrophilic compounds [150], a group likely to include metal complexing ligands. Moreover, separation of extracted compounds is usually limited to broad, operationally defined fractions based on solubility (i.e. fulvic and humic acids) or molecular weight (e.g. dialysis, ultrafiltration) [2, 150]. Results obtained for tannic acid (Section 5.4.6) highlight some of the problems associated with extracting and separating dissolved organic ligands in their natural state. Selective extraction of metal complexing ligands under ambient conditions can, however, be achieved using immobilized metal-ion affinity chromatography (IMAC). Originally developed for isolating proteins from biological samples [151], IMAC was first applied to the study of copper complexing ligands in seawater by Gordon [152]. The technique has been used in conjunction with electrophoresis, voltammetry, fluorescence spectroscopy and chemical analysis to characterize natural organic ligands [54, 137, 152, 153]. The following describes the development and application of a shipboard IMAC system designed to extract copper complexing ligands from seawater for subsequent ESI-MS analysis. 105 6.2 Immobilized Metal-Ion Affinity Chromatography IMAC involves passing the filtered sample solution through a chelating column to which the metal ion of interest (e.g. Cu2+) is attached [152]. Compounds with an affinity for the metal (e.g. copper complexing ligands) are selectively retained while the rest of the sample passes through the column. The extracted compounds can be recovered by (i) stripping the metal from the column or (ii) displacing the compound(s) from the immobilized metal ions. The former requires elution with a competing ligand such as glycine [152], which is likely to complicate analysis by its presence in the IMAC extracts [137]. The latter can be achieved using an acidic eluent. Seawater acidified with hydrochloric acid is ideal as it avoids the introduction of foreign species. This approach has been used successfully to isolate copper complexing ligands from seawater [54] and was therefore adopted for the present study. 6.2.1 The IMAC System The shipboard IMAC system (Figure 6.1) consisted of two 5 mL Hi-Trap Chelating Sepharose columns (part no. 17-0409-01, Pharmacia), a two-channel peristaltic pump (model 2120, LKB) and a UV detector (model 2138, LKB) fitted with an 8-pL flow cell and a 255 nm filter [54]. These components were secured by mounting them on a single sheet of marine ply fixed to a benchtop in the ship's laboratory. The pump, which required silicone rubber tubing (3.0-mm-o.d., 1.2-mm-i.d.), was positioned directly upstream of the two columns. This minimized pump noise at the detector and allowed the columns to be temporarily disconnected while the lines were primed. 106 5 5 Figure 6.1 Schematic of the shipboard IMAC system. A peristaltic pump (1 mL/min) B IMAC columns (5 mL) C UV detector (255 nm) D chart recorder (0.05 AUFS) 1 deionized water (Milli-Q) 2 copper sulfate solution (10 mM) 3 acidified seawater (10 mM HCI) 4 EDTA solution (50 mM) 5 filtered sample solution 6 waste 7 fraction collector (5 mL Chemtube) the IMAC procedure required a number of different solutions for conditioning, eluting and regenerating the column. These were prepared in 1 liter acid-washed polyethylene bottles (Nalgene) and delivered via a 4-way valve and associated tubing of the type used in a HPLC pump (model 126, Beckman). Teflon tubing (1.8-mm-o.d., 1.2-mm-i.d.) was use throughout the rest of the system. A2-way valve (model LV-3, Pharmacia) connected to each pump channel was used to switch between the 4-way valve and the sample. This arrangement allowed the two columns to be eluted independently (a requirement when using the single UV detector). 107 6.2.2 IMAC Extraction Procedure The pump speed was first adjusted to give a flow rate of approximately 1 mL/min. The exact flow rate for each channel was then measured using a stopwatch and 10 mL graduated cylinder. The measured flow rate was subsequently used to time the delivery of reagent and sample solutions. After rinsing with 30 mL of deionized water the IMAC columns were loaded with 30 pmoles of Cu 2 + ions by passing 3 mL of 10 mM copper (II) sulfate solution through each column. The columns were then washed with 10 mL of deionized water and conditioned with 20 mL of seawater acidified with 10 mM hydrochloric acid (pH 2). The pump was stopped briefly while both 2-way valves were switched to permit simultaneous extraction of the two samples. Extraction was allowed to proceed until a small amount of sample remained, at which point the pump was turned off. The column already connected to the UV detector was then eluted by switching the corresponding 2-way valve and restarting the pump. The elution of extracted compounds was monitored by means of a chart recorder (model 2210, LKB) connected to the UV detector. The recorder input was initially set to give 0.05 absorbance units full scale (AUFS), but was adjusted to accomodate samples with higher UV absorbances (Section 6.4.1). Fractions corresponding to the UV absorption maxima were collected in 5 mL Chemtubes (Bio-Rad) for subsequent analysis, while extraction of the second sample continued. When elution of the first column was complete, the pump was again turned off while the second column was connected to the detector and the corresponding 2-way valve switched. The second column was then eluted in the same way. 108 Finally, both columns were regenerated by removing Cu 2 + ions and any residual organic compounds with 15 mL of 50 mM EDTA solution before repeating the procedure. -124' -123' km M'Hi i i ? a^1 0 5 10 Figure 6.2 Field sample locations in British Columbia coastal waters. 109 6.2.3 Field Studies The shipboard IMAC system was used to extract copper complexing ligands from British Columbia coastal waters during a cruise aboard the CSS Vector (28 July to 1 August, 1997). Samples were collected at various depths from stations in Saanich Inlet, Jervis Inlet and the Fraser River estuary (Figure 6.2). Some of the water collected at 100 m in Saanich Inlet was acidified with hydrochloric acid and used as the IMAC eluent. A Figure6.3 Nitrogen overpressure system for seawater filtration. A polyethylene Swagelok fitting B Go-Flo bottle (10 liters) C Teflon filter holder (142 mm dia.) D membrane filter (0.45 nm) E filter support 1 nitrogen supply (<104 Pa) 2 filtrate collector (Nalgene bottle) 110 Seawater was collected using 10 liter Go-Flo bottles (General Oceanics) mounted on a Kevlar line and tripped with a teflon messenger. Samples were filtered through 0.45 pm mixed cellulose membrane filters (type HA; Millipore) in a teflon holder (model YY-142-00, Millipore), using nitrogen overpressure to avoid cell lysis [4] (Figure 6.3). Between 800 and 1000 mL of each filtered sample were extracted on board using the IMAC system while the remainder was immediately frozen for storage in 1 or 2 liter acid-washed polyethylene bottles (Nalgene). < CD O c CO _ o CO < > 0.05-1 0.04 H 0.03 0.02 i 0 . 0 1 ] 0.00 (b) 10 15 0 5 10 Elution Volume (mL) 15 Figure6.4 UV chromatograms for IMAC extracts collected from the Fraser estuary (station F2) at 50 m (a) and 5 m (b). Small peaks at 0-2 mL were usually observed after connecting the second column to the detector (Section 6.2.1). UV elution profiles for many of the IMAC extracts showed two peaks of roughly proportional intensity (Figure 6.4). The first of these occurred after about one column volume (5 mL) had been eluted, and presumably corresponds to material bound weakly by the immobilized copper ions [137]. The second peak, which typically eluted at about two column volumes [152, 153], was usually much larger than the first. IMAC fractions corresponding to these peaks were frozen on board for subsequent ESI-MS analysis. Location Station Depth (m) Max. Absorbance (AU x 103) Saanich Inlet S1 (48° 35.1' N, 123° 30.4' W) 4 43.0 (oxic above 100 m)a ( 50 30.0 100 24.2 Fraser Estuary F1 (49°07 . r N, 123° 11.1' W) 5 49.3 (halocline at 10-15 m)a F2 (49° 05.5' N, 123° 20.2' W) 5 47.4 50 27.0 F3 (49° 04.1' N, 123° 23.6' W) 10 40.8 100 18.2 F4 (49° 02.5' N, 123° 27.1'W) 10 31.4 Jervis Inlet J1 (50° 03.5' N, 123° 49.1' W) 4 33.3 (thermocline at 4m)a 100 24.0 J2 (49° 48.2' N, 124° 02.6' W) 4 52.6 100 34.2 650 20.7 Blank _ — 2.2 (deionized water) real-time measurements (CTD data acquisition system inoperative). Table 6.1 UV absorbance (255 nm) of IMAC extracts from British Columbia coastal waters (corrected for sample volume). 112 The absorbance of the second peak for each sample was measured relative to the initial baseline and corrected for the volume of sample extracted (Table 6.1). The data showed that maximum UV absorbance was always greater for surface water than for deeper water at a given location. For Fraser estuary samples, maximum absorbance also decreased with increasing distance from the mouth of the river. Assuming that UV absorbance is proportional to the amount of extracted material these results suggest that river and surface waters are both significant sources of strong copper complexing ligands. This is consistent with the production of such ligands by photosynthetic marine organisms [62, 64, 65] and the known metal binding properties of terrestrial DOM [1-3]. Seasonal variation in UV absorbance for IMAC extracts has also been reported [153]. 6.3 Flow Injection Analysis of Seawater Extracts The high salt content of the IMAC extracts precludes direct analysis by ESI-MS (section 1.4.4). A procedure for transferring organic ligands from the seawater matrix to a solvent suitable for electrospray was therefore required. Flow injection analysis (FIA) is a rapid and highly reproducible method of sample introduction for continuous flow analytical techniques such as ICP-MS [154] and ESI-MS . By entraining the sample in a moving stream FIA ensures that the analyte remains in solution and that contamination is minimized [155]. On-line matrix removal and analyte preconcentration can also be achieved by FIA using a suitable arrangement of valves and packed reactors or columns [154]. The extraction of iron complexing ligands (siderophores) from artificial seawater cultures has been successfully accomplished by acidification and adsorption onto Amberlite XAD-16, a hydrophobic macroreticular resin [68, 156] (Figure 6.5). 113 \ HydrophlUe Hoad Hydrophobic Tall Figure 6.5 Structure of a hydrophobic macroreticular resin [157]. This approach appears well suited to the removal of organic ligands from acidic IMAC fractions. The following describes the development and application of a FIA system incorporating a custom made XAD-16 column for on-line matrix removal and ESI-MS analysis of seawater. 6.3.1 Design of the XAD-16 Column Conventional LC columns require a post-column splitter to achieve a flow rate compatible with electrospay ionization [158] (Section 1.4.4). This procedure results in a very small fraction of the total eluent (typically <5 %) reaching the detector, a situation which may prevent minor sample components from being detected. This is of particular concern in the present case, where the identities and relative concentrations of the compounds extracted by IMAC are initially unknown. Microbore columns operate at much lower flow rates than conventional columns [93, 96, 120], allowing up to 100% transfer efficiency. However, microbore columns containing XAD-16 (which is intended for preparative rather than high performance LC) are not commercially available. It was therefore necessary to use a custom made XAD column for the FIA system. 114 Amberlite Resin Mesh Size Surface Area (Mean m2/g) Pore Diameter (Mean A) XAD-2 20/60 300 90 XAD-4 20/60 725 40 XAD-7 20/60 450 90 XAD-16 20/60 800 100 Table 6.2 Physical properties of Amberlite XAD resins [158]. The size of the resin places a lower limit on the diameter of the column. XAD resins (Supelco) are supplied with particles of 250-840 pm diameter (i.e. 20/60 mesh; Table 6.2). A trial column made from tubing large enough to accomodate these particles (approx. 2.0-mm-i.d.) required a flow rate of 1.0 mL/min to obtain chromatographic peaks sharp enough for accurate integration. A 70 mm x 0.1-mm-i.d. fused-silica capillary (Restek) was used to divert 90 % of the eluent from this column, giving an acceptable ESI flow rate of 0.1 mL/min (Section 6.3.2). However, the maximum possible efficiency of this system remained only 10%. To reduce column size, the resin was ground (ceramic mortar and pestle, Coors) and sieved (standard testing sieves, W.S. Tyler) to obtain particles of 150-250 pm diameter (60/100 mesh). This process should not alter significantly the properties of the macroreticular resin, since the beads are themselves composed of much smaller particles (Figure 6.5). The ground resin was suspended in methanol and loaded into a 100 mm x1.2-mm-i.d. teflon tube to which the appropriate fittings had already been attached (Figure 6.6). The resin was retained by glass fibre plugs held between the ' column and valve fittings. This column was small enough to give acceptable peak 115 Si B B Figure 6.6 Diagram of the packed column used for flow injection analysis. A threaded connector B valve fitting C column fitting D teflon tube (100 mm x 1.2-mm-i.d.) E XAD-16 resin (60/100 mesh) F glass fiber plug widths at a flow rate of 0.1 mL/min, obviating the need for a post-column splitter and providing 100% transfer efficiency. 6.3.2 The FIA System The column was loaded and eluted via a 2-position, 6-port rotary valve (model 5020; Alltech Associates). Conditioning, loading and rinsing of the column was achieved using a peristaltic pump (model 2120, LKB) and a pair of 2-way valves connected in series (Figure 6.7). The eluting solvent, supplied by a HPLC pump (model 590, Waters) at 0.1 mL/min, consisted of 1 mM ammonium acetate in 70% acetonitrile (pH 7.6). This combination ensured efficient elution and electrospray ionization of the analyte(s) as well as rapid mixing with the deionized water use to rinse the column. Peak broadening was minimized by backflushing the column with eluting solvent. Eluted compounds were transferred directly to the ESI source via a 0.1-mm-i.d. fused-silica capillary (Restek) connected to the rotary valve by a specially made fitting. 116 LOAD WASTE INJECT WASTE Figure6.7 Schematic of the FIA-ESI-MS system. A peristaltic pump (0.7 mL/min) B HPLC pump (0.1 mL/min) C XAD column 1 acidified artificial seawater 2 sample solution 3 deionized water 4 FIA eluting solvent Electrospray parameters were the same as those used in previous LC experiments (Section 5.3), with the exception of source temperature. The relatively high flow rate from the XAD column (Section 1.4.4) required a minimum temperature of 120 °C to prevent solvent clusters from obscuring the mass spectrum. However, rapid dissipation of heat by evaporation of the excess solvent apparently avoided thermal decomposition at this temperature (Section 5.4.1). Eluted compounds were detected by 117 selected ion recording (SIR) of the corresponding deprotonated molecular ions, and quantified using SIR peak areas determined by the spectrometer software (MassLynx version 2.1). 6.3.3 FIA Procedure A stop watch and graduated cylinder were again used to determine the peristaltic pump flow rate, which was restricted to about 0.7 mL/min by column back-pressure. The measured flow rate was then used to time the delivery of conditioning, sample and rinsing solutions. With the 6-port rotary valve in the LOAD position (Figure 6.7), the column was first conditioned with 4 mL of seawater acidified with 10 mM hydrochloric acid. A known volume of sample (typically 4 mL) was then passed through the column by switching the appropriate 2-way valve. Salts were removed by rinsing the column with 4 mL of deionized water after switching the second 2-way valve. Finally, compounds retained by the XAD-16 resin were eluted by turning the 6-port rotary valve to the INJECT position. Conditioning of the XAD column with acidified seawater before each run was necessary to flush out the eluting solvent and equilibrate the resin with the sample matrix. Acidified seawater taken from field samples gave strong background signals (due presumably to the bulk DOM) which were likely to obscure the compounds selectively extracted by IMAC. Artificial seawater, consisting of 0.5 M sodium chloride and 2.3 mM sodium bicarbonate in deionized water (pH 8) was therefore used to condition the column, following acidification with 10 mM hydrochloric acid. The performance of the FIA system was evaluated using standard solutions containing tannic acid in acidified artificial seawater (Figure 6.8). 0 0.5 1 1.5 2 2.5 Tannic Acid Concentration (pM) Figure 6.8 FIA-ESI-MS calibration for tannins L i * ( • ) , L 2 * (•) and L 3 * (A) in 4ml_ of acidified artificial seawater (pH 2). Figure 6.9 FIA-ESI-MS calibration for salicylic acid in 4ml_ of acidified artificial seawater (pH 2). 119 A linear response was obtained for tannins L i * , L 2 * and L 3 * (r2 = 0.9950, 0.9858 and 0.9894, respectively). To estimate detection limits for individual compounds, standards containing salicylic acid were also analyzed by FIA. The resulting calibration (r2 = 0.9964) gave a detection limit of 0.4 nM, based on 3 SD of the blank [107]. This corresponds to 1.6 picomoles of salicylic acid in 4 mL, the typical volume of an IMAC fraction (Section 6.2.1). 3 WASTE ESI-MS WASTE Figure 6.10 System to determine ligand recovery by FIA A peristaltic pump (0.7 mL/min) B HPLC pump (0.1 mL/min) C XAD-16 column D sample loop (10 \iL) 1 acidified artificial seawater 2 acidified seawater standard 3 deionized water 4 FIA eluent 5 FIA eluent standard 120 To estimate the % recovery of organic ligands by the FIA procedure, a second rotary valve (model 2691, Rheodyne) fitted with a 10-pL sample loop was connected between the FIA system and the ESI source (Figure 6.10). A solution of 0.1 pM salicylic acid in acidified artificial seawater was prepared, along with a set of standards containing 0-50 pM salicylic acid in the FIA eluting solvent. FIA was first performed on 4 mL volumes of the artificial seawater solution (Figure 6.11). 1 0 0 o c CO "O c D < CD > _cg CD 01 -i—i—i—i—i—|—i—i—i—i—r 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 i 1 1 • • i 8 0 . 0 0 -i—i—i i in 1 0 0 . 0 0 Time Figure 6.11 SIR chromatogram (m/z 137) for replicate FIA-ESI-MS analyses of 0.1 pM salicylic acid in 4 mL of acidified artificial seawater (each analysis takes about 30 minutes). Then, using a glass syringe, each of the standard solutions was introduced via the second rotary valve. All SIR peak areas were measured using the same integration parameters. Since peak area is proportional to the quantity of analyte introduced, recovery of all the salicylic acid in the 4 mL sample (i.e. 0.4 nanomoles) should give a 121 response equal to the 40 nM standard (Figure 6.12). The average response was in fact 0.0E+0 -I 1 1 - i 1 = 1 0 10 20 30 40 50 Concentration (M.M) Figure 6.12 ESI-MS calibration for salicylic acid (m/z 137) in 10 y.L of the FIA eluting solvent (r2 = 0.9975). The recovery of hydrophilic compounds from acidified artificial seawater was further investigated using a 10 |iM solution of glutathione, a tripeptide with short, relatively polar side chains. This compound forms positive ions under most solution conditions (Section 2.4.8); however, none of the expected ions were detected by either positive or negative ion ESI-MS. This confirms that adsorption by XAD-16 resin depends on the presence of hydrophobic structural features, such as aliphatic side chains or the aromatic rings present in tannin and salicylate ligands. Natural organic ligands which lack such features may not, therefore, be efficiently recovered by FIA. 122 6.4 Flow Injection Analysis of IMAC Extracts 6.4.1 Artificial Seawater Extracts The FIA procedure was first applied to 4 mL IMAC fractions obtained by extracting 1 liter volumes of artificial seawater (pH 8) containing 0.1 or 0.2 njmoles tannic acid. The corresponding UV chromatograms suggested that tannins are effectively and reproducibly extracted by immobilized Cu 2 + ions (Figure 6.13). 0.05 0.04 H Z) < CD O 0.03 c CO o t/) 0.02 -Q < > 3 0.01 0.00 J 2x > < 2x * < 10 15 20 25 0 Elution Volume (mL) 10 15 20 25 Figure 6.13 UV chromatograms for duplicate IMAC extractions of 0.1 nM tannic acid from 1 liter artificial seawater (50 (imoles immobilized Cu2+). Maximum absorbance depended both on tannic acid concentration and the quantity of immobilized copper [71], with 50 iimoles Cu 2 + giving roughly twice the absorbance observed for 30 jimoles. However, FIA did not detect any tannins in these extracts, even though concentrations should be orders of magnitude higher than the 123 standards previously analyzed. Extracts spiked with known amounts of tannic acid gave results similarto the equivalent standards (Figure 6.8), indicating that (i) tannins were not extracted by IMAC or (ii) tannins extracted by IMAC were not detected by FIA. Earlier titration experiments showed that addition of copper(ll) to tannic acid in aqueous acetonitrile gives yellow solutions in which soluble coppentannin complexes are readily detected (Section 5.4.3, p. 91). The more concentrated IMAC extracts were similar in color, suggesting that the extracted tannins form complexes with Cu 2 + ions in the IMAC column. Further acidification with hydrochloric acid removed the yellow color from these extracts but did not increase tannin recovery. This suggests that complexation of tannin ligands by surface-bound Cu 2 + ions leads to irreversible changes in their structure and /or composition (a process thought to contribute to the formation of humic substances in natural waters) [159]. Previous studies also showed how complexation of added copper reduces the concentration of free tannin ligands (Section 5.4.4, p. 93). However, dissolved copper concentration appeared to have little effect on the recovery of tannins or salicylic acid by FIA. This further suggests that refractory complexes formed with immobilized Cu 2 + ions differ from complexes formed with dissolved copper, and are probably responsible for inhibiting the recovery of organic ligands. 6.4.2 Field Sample Extracts Notwithstanding these results, the FIA procedure was applied to IMAC extracts from some of the field samples in anticipation that other copper complexing ligands would be more amenable to flow injection analysis. Since the molecular weights of these compounds were initially unknown, the spectrometer was operated in scan rather 124 than SIR mode. A series of compounds separated by 44 mass units and with m/z ratios ranging from 351 to 747 were observed in most IMAC extracts. These compounds were similar to those observed in unextracted seawater (Section 6.3.3), but were not initially detected in (blank) deionized water IMAC extracts. However, if the deionized water was filtered and stored in the same manner as the field samples (Figure 6.3) then the ions were again observed, showing them to be artifacts of the sample handling procedure. A constant mass difference of 44 suggests that these ions are derived from polyfunctional or polymeric compounds for which the repeating units form carbon dioxide (C02), fluoroethyne (C2HF), propane (C3H8), acetaldehyde or ethylene oxide (C2H40) upon eliminatitih [<160]. Possible candidates include n-alkanes or plasticizers from the bottles used to store the frozen samples [161, 162] or polyethoxylates, which are ubiquitous contaminants both in the laboratory and in the natural environment [145]. To increase the probability of detecting natural copper complexing ligands, subsequent experiments focused on IMAC extracts of surface water from Jervis Inlet (station J2), which showed particularly high UV absorbance (Table 6.1). Replicate IMAC extractions were performed on the subsamples previously frozen at sea (Section 6.2.2) using 50 [imoles immobilized copper to optimize extraction efficiency [71] (Figure 6.14a). FIA-ESI-MS of fractions collected from these extracts identified a compound of m/z 258 which was not present in the blank and which followed an elution profile similar to the UV absorbance. This compound was subsequently classified as a possible copper complexing ligand. Fractions from the remaining IMAC extracts were analyzed using the mass spectrometer in SIR mode to maximize sensitivity (Figure 6.14b). 125 O.OE+0 H 1 1 1 - i — — 1 1 0 5 10 15 20 25 30 Elution Volume (mL) Figure 6.14 UV chromatogram (a) for an IMAC extract from surface seawater ' (station J2) and the m/z 258 SIR chromatogram (b) reconstituted from the FIA-ESI-MS response for the 4 mL IMAC fractions shown in (a). Data points (•) correspond to the mean elution volume for each fraction. 126 The apparent offset between UV and SIR chromatograms suggests that the sample contains compounds which are stronger chromophores than the detected ligand but which have a lower affinity for copper and are not amenable to FIA-ESI-MS (Section 6.4.1). The SIR response for the first fraction, which consisted of about 75% extracted sample and 25% eluent, was slightly higher than the second. This could be due to (i) poor extraction efficiency or (ii) uneven elution of the IMAC column. The first is in keeping with previous assessments of the IMAC extraction procedure [137], while the second may be caused by 'channeling' or by rapid mixing of the eluting solvent with the residual sample; the latter also explains the removal of weakly bound ligands in less than one column volume of eluent (Figure 6.14a). Extensive mixing of sample and eluent was confirmed by pH measurements, which showed that the acidity of initial IMAC fractions (up to 20 mL elution volume) was significantly higher than the eluting solvent (pH 2). However, FIA of acidified and unacidified seawater samples containing the same quantity of salicylic acid gave the same response, suggesting that the recovery of ligands from IMAC fractions should not be significantly affected by any increase in pH. 6.4.3 Ligand Characterization A mass^o-charge ratio of 258 for the molecular ion [L-H]~ gives a relative molecular mass of 259. This indicates that the proposed ligand contains an odd number of nitrogen atoms, and may therefore be a peptide or other amino acid derivative. A combination of the amino acids glutamine (Gin) or lysine (Lys) and leucine (Leu) or isoleucine (lie) gives the required molecular mass (Figure 6.15). 127 o H2N .0 H2N .0 N H N H OH OH NH 2 O NH 2 Gin-Leu Lys-Ile Figure 6.15 Dipeptides of molecular mass 259 (also Gin- lie and Lys-Leu). Long, hydrophobic side chains suggest that the resulting dipeptides are more amenable to FIA than glutathione, although recoveries are likely to be modest compared with model phenolic ligands (Section 6.3.3). The presence of primary amino groups is also consistent with the description of ligands previously extracted from seawater using IMAC [71]. Simulated isotopic distributions for the proposed ligands show good agreement with the observed mass spectrum (Figure 6.16). Such information is, however, of limited utility since the resolution obtained using the quadrupole mass analyzer is not sufficient to differentiate between these compounds and others with the same unit mass. Dipeptides containing Gin and Lys (Figure 6.15), for example, differ in mass by 258.182 - 258.145 = 0.037, requiring a resolution of 258 / 0.037 - 7000 to distinguish between them. Such resolution would require a magnetic sector analyzer fitted with an electrospray ion source, which was not available during this research. Positive identification of the extracted compound(s) may be 128 (a) 100-[L-H]-258 %-I r ' i r r (b) 12,13,14C:12 1,2H:24 14,15N:3 16,17,180:3 Separation:1000 Min Frac Abun:1.00 Num Charges:l Resolution:200 Nominal Mass:258 Monoisotopic Mass:258.182 Average Mass:258.339 100% 258,182 90 j 80i 70J 60 50J A O J 30J 20J 10 245 ' " 250 ' ' 255 2&6 ' ' 265 ' ' ^O ' ' 2)5 ' ' 280 ' ' 285 ' ' 290 " ' 295 ' ' 36c Figure 6.16 Negative ion ESI mass spectrum of the extracted ligand (a) and the simulated spectrum for the deprotonated dipeptide Lys-lle (b). 129 possible using tandem mass spectrometry (Section 5.4.5), provided that the appropriate standards are available for comparison. Unfortunately the quantity of material isolated using the above procedures was not sufficient for structural analysis by MS/MS. Collection of additional samples and/or further development of extraction and separation procedures are therefore required. 6.5 Summary Organic ligands must be isolated from natural waters before they can be analyzed by ESI-MS. Copper complexing ligands may be selectively extracted from seawater under ambient conditions using immobilized Cu 2 + ion affinity chromatography. The UV absorbance of IMAC extracts from field samples suggests that rivers and surface waters are major sources of copper complexing ligands in British Columbia coastal waters. Elution of IMAC extracts with acidified seawater avoids contamination by interfering compounds but necessitates the removal of salts and residual metal ions. Flow injection analysis using a custom made XAD-16 column facilitates on-line matrix removal and ESI-MS analysis of model hydrophilic compounds (e.g salicylic and tannic acids). Unfortunately complexation of these compounds by immobilized Cu 2 + ions apparently renders them undetectable by FIA. However, analysis of the field sample with highest UV absorbance identified a compound of molecular mass 259, corresponding to a dipeptide with structural characteristics similar to copper ligands previously extracted from seawater by IMAC. Positive identification of the isolated compound(s) by tandem mass spectrometry requires further development of extraction procedures and acquisition or synthesis of the appropriate standards. CHAPTER 7. CONCLUSIONS AND FUTURE WORK 130 Dissolved metal speciation can be determined using electrospray ionization mass spectrometry. The identities and distributions of metal species in solution are established from the m/z ratios and relative abundance of the corresponding ions generated by electrospray. Organic and inorganic speciation for class A, class B and transition metal ions compares favorably with that predicted by thermodynamic equilibrium theory, provided that mild electrospray conditions are used. These findings pave the way for future studies of the interactions between different metal ions and ligands in solution (e.g. complexation of copper by siderophores [52]; extracellular ligands vs. transport proteins [32]). Detection of metal species by ESI-MS depends, however, on their polarity and thermodynamic stability. Ionization of uncharged species by protonation, deprotonation or oxidation also depends on solution pH, as well as the acid-base and electrochemical properties of the ligand. The structural analysis of intact molecular ions by electrospray ionization tandem mass spectrometry provides unique information with regard to metal complexation by large, polyfunctional organic ligands such as tannins. Nanomolar detection limits and compatibility with liquid chromatographic separation methods further demonstrate the utility of ESI-MS for characterizing ligands isolated from natural waters. However, conventional LC techniques remain limited in their ability to isolate and resolve these ligands without matrix modification or derivatization. Immobilized metal-ion affinity chromatography can be used to extract copper complexing ligands from filtered seawater under ambient conditions. Ligands eluted from IMAC columns using acidified seawater are recovered by adsorption onto XAD-16 131 resin, although the efficiency of this procedure depends on the hydrophobicity of the ligand. Subsequent analysis by ESI-MS gives the molecular weight and isotopic distribution of the extracted ligand(s), from which elemental composition and structure may be inferred. Unfortunately structural confirmation by tandem mass spectrometry requires molecular ions of greater abundance than those obtained using these procedures, although repeated extraction of larger seawater samples may help to alleviate this problem [137]. Difficulties encountered with the recovery of model ligands confirm the need for further development of extraction and separation procedures. Although specific for the metal complexing ligands of interest, IMAC may not be effective in isolating weaker ligands, which can be displaced during extraction [71], or very strong ligands, which may already be fully complexed. Furthermore, complexes formed with the immobilized metal ions may be responsible, at least in part, for inhibiting recovery of certain organic ligands from IMAC fractions (Section 6.4.1). Problems have also been encountered with the use of IMAC for isolating iron complexing ligands from seawater, due to the insolubility of Fe 3 + ions [163]. For these reasons the use of IMAC in subsequent investigations should be reviewed. While more efficient than the older XAD-2 resin [157], XAD-16 remains highly selective for compounds with hydrophobic properties. Recovery of hydrophilic ligands may be improved by using XAD-7 resin in the FIA system [157]. Alternatively, porous graphitic carbon (PGC) has been shown to be highly effective in recovering and separating water-soluble compounds (including metal-ion complexes) prior to ESI-MS [97]. The use of PGC columns (e.g. Hypercarb-S, Shandon Scientific) in future studies is therefore highly recommended. 132 The results presented in this dissertation demonstrate the enormous potential of electrospray mass spectrometry for studying metal complexation in natural waters. It is the only technique other than potentiometry which permits unequivocal identification and determination of dissolved metal species. As such, ESI-MS is destined to become a valuable tool for marine chemists in their quest to understand the processes by which metal ions are utilized and distributed within the ocean. BIBLIOGRAPHY 133 1. W. Stumm and J.J. Morgan, Aquatic Chemistry. Chemical Equilibria and Rates in Natural Waters, 3rd ed., Wiley, New York (1996). 2. J. Buffle, Complexation Reactions in Aquatic Systems. An Analytical Approach, Ellis Horwood, Chichester (1990). 3. S.M. Libes, An Introduction to Marine Biogeochemistry, Wiley, New York (1992). 4. T.M. Florence, "The Speciation of Trace Elements in Waters", Talanta, 29, 345-364(1982). 5. F.A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 4th ed., Wiley, New York (1980). 6. L.-G. Danielsson, B. Magnusson, S. Westerlund and K. Zhang, "Trace Metal Determinations in Estuarine Waters by Electrothermal Atomic Absorption Spectrometry After Extraction of Dithiocarbamate Complexes into Freon", Anal. Chim. Acta, 144, 183-188 (1982). 7. W.K. Fletcher, G.S. Holmes and AG . Lewis, "Geochemistry and Biological Availability of Iron and Trace Elements in the Upper Fraser River Estuary", Mar. Chem., 12, 195-217 (1983). 8. F. Laumond, G. Copin-Montegut, P. Courau and E. Nicholas, "Cadmium, Copper and Lead in the Western Mediterranean Sea", Mar. Chem., 15, 251-261 (1984). 9. K.W. Bruland and K.H. Coale, "Analysis of Seawater for Dissolved Cadmium, Copper and Lead. An Intercomparison of Voltammetric and Atomic Absorption Methods", Mar. Chem., 17, 285-300 (1985). 10. K.J. Orians and K.W. Bruland, "The Biogeochemistry of Aluminium in the Pacific Ocean", Earth Plant. Sci. Lett, 78, 397-410 (1986). 11. K.J. Orians and E.A. Boyle, "Determination of Picomolar Concentrations of Titanium, Gallium and Indium in Sea Water by Inductively Coupled Plasma Mass Spectrometry following an 8-Hydroxyquinoline Chelating Resin Preconcentration", Anal. Chim. Acta, 282, 63-74 (1993). 12. B.A. McKelvey and K.J. Orians, "Dissolved Zirconium in the North Pacific Ocean", Geochim. Cosmochim. Acta, 57, 3801-3805 (1993). 134 13. B.A. McKelvey and K.J. Orians, "The Determination of Dissolved Zirconium and Hafnium from Seawater using Isotope Dilution Inductively Coupled Plasma Mass Spectrometry", Mar. Chem. 60, 245-255 (1998). 14. K.J. Orians and K.W. Bruland, "Dissolved Aluminum, in the Central North Pacific", Nature, 316, 427-429 (1985). 15. K.J. Orians and K.W. Bruland, "Dissolved Gallium in the Open Ocean", Nature, 332,717-719(1988). 16. K.J. Orians and K.W. Bruland, "The Marine Geochemistry of Dissolved Gallium. A Comparison with Dissolved Aluminium", Geochim. Cosomchim. Acta, 52, 2955-2962 (1988). 17. K.J. Orians, E.A. Boyle and K.W. Bruland, "Dissolved Titanium in the Open Ocean", Nature, 348, 322-325 (1990). 18. K.W. Bruland, K.J. Orians and J.R Cowen, "Reactive Trace Metals in the Central North Pacific", Geochim. Cosmochim. Acta, 58, 3171-3182 (1994). 19. J.R. Donat, In "Trace Elements in the Oceans", CRC Press, Boca Raton, (1995), 247-281. 20. W.R. Cullen and K.J. Reimer, "Arsenic Speciation in the Environment", Chem. Reviews, 89, 713-764 (1989). 21. K.J. Reimer and J.A.J. Thompson, "Arsenic Speciation in Marine Interstitial Water. The Occurence of Organoarsenicals", Biogeochem., 6, 211-237 (1988). 22. R.P. Mason, W.F. Fitzgerald, J. Hurley, A.K. Hanson Jr., P.L. Donaghay and J.M. Sieburth, "Mercury Biogeochemical Cycling in a Stratified Estuary", Limnol. Oceanogr., 38, 1227-1241 (1993). 23. H. Iwata, S. Tanaba, T. Mizuno and R. Tatsukawa, "High Accumulation of Toxic Butyltins in Marine Mammals from Japanese Coastal Waters" Environ. Sci. Technol:, 29, 2959-2962 (1995). 24. Y.K. Chau, R.J. Maguire, M. Brown, F. Yang, S.P. Batchelor and J.A.J. Thompson, "Occurrence of Butyltin Compounds in Mussels in Canada", Appl. Organometal. Chem., 11, 903-912 (1997). 25. A. Tessier, P.G.C Campbell and M. Bisson, "Sequential Extraction Procedure for the Speciation of Particulate Trace Metals", Anal. Chem., 51, 844-851 (1979). 26. J.-M. Martin, P.M.V. Nirel and A.J. Thomas, "Sequential Extraction Techniques: Promises and Problems", Mar. Chem., 22, 313-341 (1987). 135 27. Y. Babukutty and J. Chacko, "Chemical Partitioning and Bioavailability of Lead and Nickel in a Estuarine System", Environ. Toxicol. Chem., 14, 427-434 (1995). 28. S.B. Moran and R.B. Moore, "The Distribution of Colloidal Aluminium and Organic Carbon in Coastal and Open Ocean Waters off Nova Scotia", Geochim. Cosmochim. Acta, 53, 2519-2527 (1989). 29. S.A. Skrabal, W.J. Ullman and G.W. Luther III, "Estuarine Distributions of Dissolved Titanium", Mar. Chem., 37, 83-103 (1992). 30. I.L. Kuchler, N. Miekeley and B.R. Forsberg, "Molecular Mass Distributions of Dissolved Organic Carbon and Associated Metals in Waters from Rio Negro and Rio Solimoes", Sci. Tot. Environ., 156, 207-216 (1994). 31. N.M Price, B.A. Ahner and F.M.M. Morel, "The Equatorial Pacific Ocean. Grazer-controlled Phytoplankton Populations in an Iron-limited Ecosystem", Limnol. Oceanogr., 39, 520-534 (1994). 32. W.J. Sunda, "Trace Metal Interactions with Marine Phytoplankton", Biol. Oceanogr., 6, 411-442 (1991). 33. W. Sunda and R.R.L. Guillard, "The Relationship between Cupric Ion Activity and the Toxicity of Copper to Phytoplankton", J. Mar. Res., 34, 511-529 (1976). 34. R.J.M. Hudson and F.M.M. Morel, "Iron Transport in Marine Phytoplankton. Kinetics of Cellular and Medium Coordination Reactions", Limnol. Oceanogr, 35, 1002-1020(1990). 35. R.J.M. Hudson and F.M.M. Morel, "Trace Metal Transport by Marine Microorganisms. Implications of Metal Coordination Kinetics", Deep-Sea Res. I, 40, 129-150(1993). 36. M. Azenha, M.T. Vasconcelos and J.P.S. Cabral," Organic Ligands Reduce Copper Toxicity in Pseudomonas Syringae", Environ. Toxicol. Chem., 14, 369-373(1995). 37. J.M. Santana-Casiano, M. Gonzalez-Davila, J. Perez-Pena and F.J. Millero, "Pb2 + Interactions with the Marine Phytoplankton Dunaliella tertiolecta", Mar. Chem., 48, 115-129(1995). 38. WG. Sunda and S.A. Huntsman, "Effect of Competitive Interactions between Manganese and Copper on Cellular Manganese and Growth in Estuarine and Oceanic Species of the Diatom Thalassiosira", Limnol. Oceanogr., 28, 924-934 (1983). 39. J.M. Santana-Casiano, M. Gonzalez-Davila, L.M. Laglera, J. Perez-Pena, L. Brand and F.J. Millero, "The Influence of Zinc, Aluminum and Cadmium on the Uptake Kinetics of Iron by Algae", Mar. Chem., 59, 95-111 (1997). 136 40. P.G.C. Campbell, in Metal Speciation and Bioavailability in Aquatic Systems, A. Tessier and D.R. Turner Eds., Wiley, New York (1995), 45-192. 41. J.T. Phinney and K.W. Bruland, "Uptake of Lipophilic Organic Cu, Cd and Pb Complexes in the Coastal Diaton Thalassiosira weissflogif, Environ. Sci. Technol., 28, 1781-1790 (1994). 42. P. MacCarthy and J.A. Rice, In "Humic Substances in Soil, Sediment and Water", G.R. Aiken, D.M. McKnight, R.L. Wershawand P. MacCarthy Eds., Wiley, New York, (1985), 527-560. 43. PR. Wittbrodt and CD. Palmer, "Reduction of Cr(VI) in the Presence of Excess Soil FulvicAcid", Environ. Sci. Technol., 29, 255-263. 44. D.K. Ryan and J.H. Weber, "Fluorescence Quenching Titration for Determination of Complexing Capacities and Stability Constants of FulvicAcid", Anal. Chem., 54, 986-990 (1982). 45. J.C. Dobbs, W. Susetyo, F.E. Knight, M.A. Castles, L.A. Carreira and L.V. Azarraga," Characterisation of Metal Binding Sites in Fulvic Acids by Lanthanide Ion Probe Spectroscopy", Anal. Chem., 61, 483-488 (1989). 46. S.E. Cabaniss, "Synchronous Fluorescence Spectra of Metal-Fulvic Acid Complexes", Environ. Sci. Technol., 26, 1133-1139 (1992). 47. R.L. Cook and C.H. Langford, "Metal Ion Quenching of Fulvic Acid Fluorescence Intensities and Lifetimes. Nonlinearities and a Possible Three-Component Model", Anal. Chem., 67, 174-180 (1995). 48. J.C.G.E. da Silva and A.A.S.C. Machado, "Characterization of the Binding Sites for Al(lll) and Be(ll) in a Sample of Marine Fulvic Acids", Mar. Chem., 54, 293-302(1996). 49. C.M.G. van den Berg, in Chemical Oceanography, Vol. 9, Academic Press, (1988), 197-245. 50. K.C. Swallow and J.C. Westall," Potentiometric Determination of Copper Complexation by Phytoplankton Exudates", Limnol. Oceanogr., 23, 538-542 (1978). 51. W.G. Sunda and J.A.M. Lewis," Effect of Complexation by Natural Organic Ligands on the Toxicity of Copper to a Unicellular Alga, Monochrysis lutherf', Limnol. Oceanogr. 23, 870-876 (1978). 52. D.M. McKnight and F.M.M. Morel," Copper Complexation by Siderophores from Filamentous Blue-Green Algae", Limnol. Oceanogr., 25, 62-71 (1980). 137 53. S.E. Cabaniss and M.S. Schuman," Copper Binding by Dissolved Organic Matter. I. Suwannee River Fulvic Acid Equilibria", Geochim. Cosmochim. Acta, 52, 185-193 (1988). 54. T. Midorikawa, E. Tanoue and Y Sugimura," Determination of Complexing Ability of Natural Ligands in Seawater for Various Metal Ions using Ion Selective Electrodes", Anal. Chem., 62, 1737-1746 (1990). 55. S. Borman, "Fix for Leaking Electrodes", Chem. Eng. News, 75, 13-14 (1997). 56. T.M. Florence, G.M. Morrison and J.L. Stauber, "Determination of Trace Element Speciation and the Role of Speciation in Aquatic Toxicity", Sci. Tot. Environ., 125, 1-13(1992). 57. K.W. Bruland and K.H. Coale, "Analysis of Seawater for Dissolved Cadmium, Copper and Lead: An Intercomparison of Voltammetric and Atomic Absorption Methods", Mar. Chem., 17, 285-300 (1985). 58. E.L. Rue and K.W. Bruland, "Complexation of Iron(lll) by Natural Organic Ligands in the Central North Pacific As Determined by a New Competitive Ligand Equilibration/Adsorptive Cathodic Stripping Voltammetry Method", Mar. Chem., 50, 117-138 (1995). 59. J. Wu and G.W. Luther III, "Complexation of Fe(lll) by Natural Organic Ligands in the Northwest Atlantic Ocean by a Competitive Ligand Equilibration Method and a Kinetic Approach", Mar. Chem, 50, 159-177 (1995). 60. M. Vega and C.M.G. van den Berg, "Determination of Cobalt in Seawater by Catalytic Adsorptive Cathodic Stripping Voltammetry", Anal. Chem, 69, 874-881 (1997). 61. K.W. Bruland, "Oceanic Zinc Speciation. Complexation of Zinc by Natural Organic Ligands in the Central North Pacific", Limnol. Oceanogr, 34, 267-283 (1989). 62. K.H. Coale and K.W. Bruland,"Copper Complexation in the Northeast Pacific", Limnol. Oceanogr., 33, 1084-1101 (1988). 63. L.J.A. Gerringa, P.M.J. Herman and T.C.W Poortvliet," Comparison of the Linear Van den Berg/Ruzic Transformation and a Non-Linear Fit of the Langmuir Isotherm Applied to Cu Speciation Data in the Estuarine Environment", Mar. Chem., 48, 131-142(1995). 64. J.R. Donat, K.A. Lao and K.W. Bruland, "Speciation of Dissolved Copper and Nickel in South San Francisco Bay. A Multi-method Approach", Anal. Chim. Acta, 284, 547-571 (1994). 138 65. J.W. Moffett, R.G. Zika and L.E. Brand/'Distribution and Potential Sources and Sinks of Copper Chelators in the Sargasso Sea", Deep-Sea Res., 33, 27-36 (1990). 66. C.M.G. van den Berg, "Evidence for Organic Complexation of Iron in Seawater", Mar. Chem., 50, 139-157 (1995). 67. S.W. Wilhelm and C.G. Trick, "Iron-limited Growth of Cyanobacteria. Multiple Siderophore Production is a Common Response", Limnol. Oceanogr., 39, 1979-1984(1994). 68. C.G. Trick and S.W. Wilhelm, "Physiological changes in the Coastal Marine Cyanobacterium Synechococcus sp. PCC 7002 exposed to Low Ferric Ion Levels", Mar. Chem., 50, 207-217 (1995). 69. Z. Kozarak, M. Plavsic, B. Cosovic and C. Vilicic, "Interaction of Cadmium and Copper with Surface Active Organic Matter and Complexing Ligands Released by Marine Phytoplankton", Mar. Chem., 26, 313-330 (1989). 70. J.W. Moffett and L.E. Brand,"Production of Strong, Extracellular Cu Chelators by Marine Cyanobacteria in Response to Cu Stress", Limnol. Oceanogr., 41, 388-395(1996). 71. T. Midorikawa and E. Tanoue, "Extraction and Characterisation of Organic Ligands from Oceanic Water Columns by Immobilised Metal Ion Affinity Chromatography", Mar. Chem., 52, 157-171 (1996). 72. Y. Liu, "Capillary Electrophoresis Coupled On-Line with Inductively Coupled Plasma Mass Spectrometry for Elemental Speciation", Anal. Chem, 67, 2020-2025(1995). 73. M.J. Powell, "Determination of Chromium Species in Enviromental Samples Using High-Pressure Liquid Chromatography Direct Injection Nebulization and Inductively Coupled Plasma Mass Spectrometry", Anal. Chem, 67, 2474-2478 (1995). 74. J.B. Fenn, M. Mann, CK. Meng, S.F. Wong and CM. Whitehouse, "Electrospray lonization-Principles and Practice", Mass. Spectom. Revs., 9, 37-70 (1990). 75. P. Kebarle and L. Tang, "From Ions in Solution to Ions in the Gas Phase. The Mechanism of Electrospray Mass Spectrometry", Anal. Chem., 65, 972A-986A (1993). 76. M. Dole, L.L. Mack, R.L. Hines, R.C. Mobley. L.D. Ferguson and M.B. Alice, "Molecular Beams of Macroions", J. Chem. Phys., 49, 2240-2249 (1968). 139 77. M. Yamashita and J.B. Fenn, "Electrospray Ion Source. Another Variation on the Free-Jet Theme", J. Chem. Phys., 88, 4451-4459 (1984). 78. M.G. Ikonomou, AT. Blades and P. Kebarle, "Electrospray-lon Spray. A Comparison of Mechanisms and Performance", Anal. Chem., 63, 1989-1998 (1991). 79. M.G. Ikonomou, A.T. Blades and P. Kebarle, "Investigations of the Electrospray Interface for Liquid Chromatography/Mass Spectrometry", Anal. Chem., 629, 957-967 (1991). 80. K.W. M. Siu, G.J. Gardner and S.S. Berman, "Atmospheric Pressure Chemical Ionization and In-Spray Mass Spectrometry of Some Organotin Species", Rapid Commun. Mass Spectrom., 2,;201-204 (1988). 81. K.W. M. Siu, G.J. Gardner and S.S. Berman, "Ion Spray Mass Spectrometry/ Mass Spectrometry. Quantitation of Tributyltin in a Sediment Reference Material for Trace Metals", Anal. Chem., 61, 2320-2322 (1989). 82. P. Hu, Q.-Z. Ye and J.A. Loo, "Calcium Stoichiometry Determination for Calcium Binding Proteins by Electrospray Ionization Mass Spectrometry", Anal. Chem., 66,4190-4194(1994). 83. R. Feng, A.L. Castelhano, R. Billedeau and Z. Yuan, "Study of Noncovalent Enzyme-Inhibitor Complexes and Metal Binding Stoichiometry of Matrilysin by Electrospray lonisation Mass Spectrometry", J. Am. Soc. Mass Spectrom., 6, 1105-1111 (1995). 84. X. Yu, M. Wojciechowski and C. Fenselau, "Assessment of Metals in Reconstituted Metallothioneins by Electrospray Mass Spectrometry", Anal. Chem., 65, 1355-1359 (1993). 85. Q. P. Lei, X. Cui, D:M. Kurtz Jr., I.J. Amster, I.V. Chernushevich and KG. Standing," Electrospray Mass Spectrometry Studies of Non-Heme Iron-Containing Proteins", Anal. Chem., 70, 1838-1846 (1998). 86. A.T. Blades, M.G. Ikonomou and P. Kebarle, "Mechanism of Electrospray Mass Spectrometry. Electrospray as an Electrolysis Cell", Anal. Chem., 63, 2109-2114 (1991). 87. X. Xu, S.P. Nolan and R.B. Cole, "Electrochemical Oxidation and Nucleophilic Addition Reactions of Metallocenes in Electrospray Mass Spectrometry", Anal. Chem., 66, 119-125 (1994). 88. A.L. Burlingame, R.K. Boyd and S.J. Gaskell, "Mass Spectrometry (Review)", Anal. Chem., 68, 599R-651R (1996). 140 89. P. Jayaweera, AT. Blades, M.G. Ikonomou and P. Kebarle, "Production and Study in the Gas Phase of Multiply Charged Solvated or Coordinated Metal Ions", J. Am. Chem. Soc, 112, 2452-2454 (1990). 90. AT. Blades, P. Jayaweera, M.G. Ikonomou and P. Kebarle," First Studies of the Gas Phase Ion Chemistry of M 3 + Metal Ion Ligands", Int. J. Mass Spectrom. Ion Processes, 101, 325-366 (1990). 91. AT. Blades, P. Jayaweera, M.G. Ikonomou and P. Kebarle, "Ion-Molecule Clusters Involving Doubly Charged Metal Ions", Int. J. Mass Spectrom. Ion Processes, 102, 251-267 (1990). 92. AT. Blades, P. Jayaweera, M.G. Ikonomou and P. Kebarle, "Studies of Alkaline Earth and Transition Metal M + + Gas Phase Ion Chemistry", J. Chem. Phys., 92, 5900-5906 (1990). 93. M. Kohler and J.A. Leary, "LC/MS/MS of Carbohydrates with Postcolmn Addition of Metal Chlorides using a Triaxial Electrospray Probe", Anal.Chem., 67, 3501-3508(1995). 94. CL. Gatlin,. R.D. Rao, F. Turecek and T. Vaisar, "Carboxylate andAmine Terminus Directed Fragmentations in Gaseous Dipeptide Complexes with Copper (II) amd Diimine Ligands Formed by Electrospray", Anal. Chem., 68, 263-270(1996). ' ' 95. G.R. Agnes and G. Horlick, "Electrospray Mass Spectrometry as a Technique for Elemental Analysis. Preliminary Results", Appl. Spectrosc, 46, 401-406 (1992). 96. CL . Gatlin, F. Turecek and T. Vaisar, "Determination of Soluble Cu(l) and Cu(ll) Species in Jet Fuel by Electrospray lonisatipn Mass Spectrometry", Anal. Chem., 66,3950-3958(1994). 97. H.C. Ehrsson, I. B. Wallin and A.S. Andersson, "Cisplatin, Transplatin, and their Hydrated Complexes. Separation and Identification Using Porous Graphitic Carbon and Electrospray lonisation Mass Spectrometry", Anal. Chem., 67, 3608-3611 (1995). 98. G. Wang and R.B. Cole, "Effects of Solvent and Counterion on Ion Pairing and Observed Charge States of Diquaternary Ammonium Salts in Electrospray Ionization Mass Spectrometry", J. Am. Soc. Mass Spectrom., 7, 1050-1058 (1996). 99. CL . Gatlin and F. Turecek, "Acidity Determination in Droplets Formed by Electrospraying Methanol-Water Solutions", Anal. Chem, 66, 712-718 (1994). 100. A.R.S. Ross, M.G. Ikonomou and K. J. Orians, in Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, (1995), 703. 141 101. F.M.M. Morel, Principles of Aquatic Chemistry, Wiley, New York (1983). 102. K.L. Cheng, K. Ueno and T. Imamura, CRC Handbool of Organic Analytical Reagents, CRC Press, Boca Raton (1982). 103. D.D. Perrin, Stability Constants of Metal-Ion Complexes. Part B. Organic Ligands, Pergamon Press, Oxford (1979), 650. 104. W.D. Schecher and D.C. McAvoy, MINEQL+, User's Manual, Environmental Research Software, Edgewater MD (1991). 105. D.R. Lide, CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton (1993). 106. F.P. Dwyer and D.P. Mellor, Chelating Agents and Metal Chelates, Academic Press, New York (1964). 107. J.C. Miller and J.N. Miller, Statistics for Analytical Chemistry, Ellis Horwood, Chichester (1988), 115. 108. J. Sachinidis and M.W. Grant, "Metal Exchange Reactions Between Divalent Metal Ions and their Dithiocarbamate Complexes in Dimethyl Sulphoxide. A Kinetic and Mechanistic Study", Aust. J. Chem., 34, 2195-2215 (1981). 109. A.M. Bond, R. Colton, A. D'Agostino, A.J. Downard and J.C. Traeger, "A Role for Electrospray Mass Spectrometry in Electrochemical Studies", Anal. Chem., 67, 1691-1695(1995).. 110. G.J. van Berkel and F. Zhou, "Characterisation of an Electrospray Ion Source as a Controlled-Current Electrolytic;Cell", Anal. Chem., 67, 2916-2923 (1995). 111. F. Zhou and G.J. van Berkel," Electrochemistry Combined Oh-Line with Electrospray Mass Spectrometry", Anal. Chem., 67, 3643-3649 (1995). 112. G.J. van Berkel and F. Zhou, "Electrospray as a Controlled Current Electrolytic Cell. Electrochemical Ionization of Neutral Analytes for Detection by Electrospray Mass Spectrometry", Anal. Chem., 67, 3958-3964 (1995). 113. E.S. Gould, Inorganic Reactions and Structure, Holt, New York (1955). 114. I.M. Arias and W.B. Jakoby, Glutathione. Mechanism and Function, Raven Press, New York (1976). 115. L. Stryer, Biochemistry, Freeman, New York (1981). 116. C.L. Gatlin, F. Turecek and T. Vaisar, "Copper(ll) Amino Acid Complexes in the Gas Phase", J. Am. Chem. Soc, 117, 3637-3638 (1995). 142 117. P. Hu, C. Sorensen and M.L. Gross, "Influences of Peptide Side Chains on the Metal Ion Binding Site in Metal lon-Cationized Peptides: Participation of Aromatic Rings in Metal Chelation", J. Am. Soc. Mass Spectrom., 6, 1079-1085 (1995). 118. G.H. Morrison and H. Frieser, in Classical Analysis, Volume IA, CL . Wilson and D.W. Wilson Eds., Elsevier, Amsterdam (1959), 147. 119. E. Bayer, T. Bauer, K. Schmeer, K. Bleicher, M. Maier and H.-J. Gaus, "Analysis of Double-Stranded Oligonucleotides by Electrospray Mass Spectrometry", Anal. Chem., 66, 3858-3863 (1994). 120. Y. Hua, W. Lu, M.S. Henry, R.H. Pierce and R.B. Cole, "On-Line High Performance Liquid Chromatography-Electrospray lonisation Mass Spectrometry for the Determination of Brevetoxins in 'Red Tide' Algae", Anal. Chem., 67, 1815-1823(1995). 121. G. Neubauer and R.J. Anderegg, "Identifying Charged States of Peptides in Liquid Chromatography/Electrospray Ionization Mass Spectrometry", Anal. Chem, 66, 1056-1061 (1994). 122. E. Nieboer and D.H.S. Richardson, "The Replacement of the Nondescriptive Term Heavy Metals by a Biologically and Chemically Significant Classification of Metal Ions", Env. Poll. Ser. B, 1, 3-26 (1980). 123. E. Laws, Aquatic Pollution, Wiley, New York (1981). 124. J.G. Hering and F.M.M. Morel, "Slow Coordination Reactions in Seawater", Geochim. Cosmochim. Acta, 5, 611-618 (1989). 125. W.-C. Chin, M.V Orellana and P. Verdugo, "Spontaneous Assembly of Marine Dissolved Organic Matter into Polymer Gels", Nature, 391, 568-572 (1998). 126. P.W. Atkins, Physical Chemistry, 3rd ed., Freeman, New York (1986), 823. 127. P. Sykes, A Guidebook to Mechanism in Organic Chemistry, 6th ed., Longman, Harlow (1993), 55. 128. Y.-T. Li, Y.-L. Hseih, J. Henion, M.W. Senko and F.W. McLafferty, "Mass Spectrometric Studies on Noncovalent Dimers of Leucine Zipper Peptides" J. Am. Chem. Soc, 115, 8409-8413 (1993). 129. N. Viseaux, E. de Hoffman and B. Domon, "Structural Analysis of Permethylated Oligosaccharides by Electrospray Tandem Mass Spectrometry", Anal. Chem., 69,3193-3198(1997). 130. J.C.Y. Le Blanc, K.W.M. Siu and R. Guevremont, in Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, (1995), 903. 131. R.J. Fessenden and J.S. Fessenden, Organic Chemistry, 3rd ed., Brooks/Cole, Monterey (1986). 132. G.J. Van Berkel, S.A. McLucky and G.L. Glish, "Preforming Ions in Solution via Charge-Transfer Complexation for Analysis by Electrospray Ionization Mass Spectrometry", Anal. Chem., 63, 2064-2069 (1991). 133. A.R.S. Ross, M.G. Ikonomou and K.J. Orians, in Proceedings of the 46th ASMS Conference on Mass Spectrometry and Allied Topics, (1998), 149. 134. L.F. Fieser and M. Fieser, Advanced Organic Chemistry, 1st ed., Reinhold, New York (1961), 1047. 135. J.A. Herrera-Silvera and J.Ramirez-Ramirez, "Effects of Natural Phenolic Material (Tannin) on Phytoplankton Growth", Limnol. Oceanogr., 41, 1018-1023 (1996). 136. S.W. Wilhelm. "Ecology of Iron-Limited Cyanobacteria. A Review of Physiological Responses and Implications for Aquatic Systems", Aquat. Microb. Eco/., 9, 295-303 (1995). 137. J.R. Donat, R.A. Kango and A.S. Gordon, "Evaluation of Immobilized Metal Affinity Chromatography (IMAC) for Isolation and Recovery of Strong Copper Complexing Ligands from Marine Waters", Mar. Chem., 57, 1-10 (1997). 138. A.R.S. Ross, M.G. Ikonomou and K.J. Orians, In 'Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics" (1996), 161. 139. B.T. Chait and S.B.H. Kent, "Weighing Naked Proteins. Practical, High Accuracy Mass Measurement of Peptides and Proteins", Science, 257, 1885-1894 (1992). 140. C. Viriot, A. Scalbert, C.L.M.H. du Penhoat, C. Rolando and M. Moutounet, "Methylation, Acetylation and Gel Permeation of Hydrolysable Tannins" J. Chromatogr. A, 662, 77-85, (1994). . 141. K.B.G. Torsell, Natural Product Chemistry. A Mechanistic and Biosynthetic Approach to Secondary Metabolism, Wiley, New York (1983). 142. A.N. Verentchikov, W. Ens and K,G, Standing, "Reflecting Time-of-Flight Mass Spectrometer with an Electrospray Ion Source and Orthogonal Extraction", Anal. Chem., SB, 126-133(1994). 143. G.R. Harvey and D.A. Boran, in Humic Substances in Soil, Sediment, and Water, G.R. Aiken, D.M. McKnight, R.L. Wershaw and P. MacCarthy Eds., Wiley, New York (1985), 233-248. 144 144. J.A. Leenheer, in Humic Substances in Soil, Sediment, and Water, G.R. Aiken, D.M. McKnight, R.L. Wershaw and P. MacCarthy Eds., Wiley, New York (1985), 409-429. 145. D. Shang, M.G. Ikonomou and R. Macdonald, In "Proceedings of the 46th ASMS Conference on Mass Spectrometry and Allied Topics" (1998), 365. 146. R.S. Swift and A.M. Posner, "Gel Chromatography of Humic Acid", J. Soil. Sci., 22, 237-249(1971). 147. D.E. Slauenwhite and P.J. Wangersky, "Extraction of Marine Organic Matter on XAD-2. Effect of Sample Acidification and Development of an In Situ Pre-• Acidification Technique", Mar. Chem., 54, 107-117 (1996). 148. C. Steelink, in Humic Substances in Soil, Sediment, and Water, G.R. Aiken, D.M. McKnight, R.L. Wershaw and P. MacCarthy Eds., Wiley, New York (1985), 457-476. 149. J.-M. Martin, W.W. Huang and YY Yoon, "Level and Fate of Trace Metals in the Lagoon of Venice (Italy)", Mar. Chem., 46, 371-386 (1994). 150. R. Benner, J.D. Pakulski, M. McCarthy, J.I. Hedges and P.G. Hatcher, "Bulk Chemical Characteristics of Dissolved Organic Matter in the Ocean", Science, 255,1561-1564(1992). 151. T W Hutchens, R.W. Nelson, CM. Li and T.-T. Yip, "Synthetic Metal-Binding Protein Surface Domains for Metal Ion Dependent Interaction Chromatography," J. Chromatogr., 604, 125-141 (1992). 152. A.S. Gordon, "Isolation of Compounds with Affinity for Copper from Seawater using Immobilized Copper Ion Affinity Chromatography", Mar. Chem., 38, 1-12 (1992). 153. A.S. Gordon, B.J. Dyer, R.A. Kango and J.R. Donat, "Copper Ligands Isolated from Estuarine Water by Immobilized Metal Affinity Chromatography. Temporal Variability and Partial Characterization", Mar. Chem., 53, 163-172 (1996). 154. J. Ruzicka and E.H. Hansen, in Flow Injection Analysis, 2nd ed., Wiley, New York (1988). 155. H. Nicolidakis, "A Method for the Extraction and Analysis of Dissolved Trace Metals from Seawater using On-Line FI-ICP-MS", Ph.D. dissertation, University of British Columbia (1995). 145 156. B.L. Lewis, P.D. Holt, S.W. Taylor, S.W. Wilhelm, G.G. Trick, A. Butler and G.W. Luther III, "Voltammetric Estimation of Iron(lll) Thermodynamic Stability Constants for Catecholate Siderophores Isolated from Marine Bacteria and Cyanobacteria", Mar. Chem., 50, 179-188 (1995). 157. C. Crescenzi, "Determination of Acidic Pesticides in Water by a Benchtop Electrospray Liquid Chromatography Mass Spectrometer", Anal. Chem., 67, 1968-1975(1995). 158. Supelco Chromatography Catalogue, 30th Anniversary ed., Sigma-Aldrich (1997), 256-257. 159. M.H.B. Hayes, in Humic Substances in Soil, Sediment, and Water, G.R. Aiken, D.M. McKnight, R.L. Wershawand P. MacCarthy Eds., Wiley, New York (1985), 329-362. 160. Eight Peak Index of Mass Spectra, Vol. 1, Part 1, 3rd ed., Royal Society of Chemistry (1983), 3. 161. R. Gabler, R. Hegde and D. Hughes, "Degradation of High Purity Water on Storage", J. Liquid Chromatogr., 6, 2565-2570 (1983). 162. A.G. Douglas and P.J. Grantham, "Docosane in Rock Extracts. A Possible Contaminative Source", Chem. Geo/., 12, 249-255 (1973). 163. D.A. Hutchins, personal communication. APPENDIX A. COMPLETE DATA SET FOR CHAPTERS 2 AND 3 147 .2+ Metal Ion (Mn+) Na* K* Ca'* Mgz pH 4.5 7.4 9.2 4.5 7.4 9.4 4.5 7.4 9.1 4.5 7.4 9.0 (B) S[M'(ACN)i]+ Z [M'jY^Ac^ACN)^ Z [M"(ACN)|]2+ Z [M"Y(HAc)j(ACN)J+ 68 79 — 89 84 — 32 21 — 11 16 — 66 34 — 62 — 38 (C) SA) ZB) 100 100 — 100 100 — 100 — — 100 Metal Ion (Mn*) Mn2* Zn2* Cd2* Pb2* pH 4.5 7.3 9.0 4.5 7.1 8.9 4.5 7.2 9.1 4.4 6.7 9.2 (B) Z[M"(ACN)i] 2+ 61 — — Z [M"Y(HAc)j(ACN)i]+ 39 — — 48 — 52 — 60 — 40 — — 100 — (C) ZA) ZB) 100 — 100 _ 100 — — 100 Metal Ion (Mn*) Cu2* Ni2* Al 3 * Fe3* pH 4.4 6.8 9.5 4.5 7.3 9.0 4.1 6.5 8.7 3.5 6.5 8.7 (A) E [(M"l_ACj)Hj]+ 96 39 12 100 58 — [M"L2H]* 4 48 74 — 42 100 [M"2L3]* — 13 14 — — — [MmL2]* 96 59 45 96 86 64 [M'"L3H]* 4 41 55 4 14 36 (B) Z[M'(ACN)i]+ 100 100 — Z [M"(ACN)i]2* — — — 80 — — Z [M"Y(HAc)j(ACN)i]* 20 — — (C) ZA) 1 30 100 1 100 100 100 100 100 100 100 100 Z B) 99 70 _ 99 — — — — — — — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . A) oxinate complexes. B) other species. C) all metal species. Table A.1 ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 148 Metal Ion (Mn+) Co2 + Cr3* pH 4.5 7.3 9.1 4.0 6.5 8.8 (A) I [(M"LAq)Hj* . - - -[M"L2H]+ — — — [Mn2l*f -[M'"L2r — — 100 — [M L3H]+ (B) Z [M"(ACN)J2+ 68 — — I [M"Y(HAc)j(ACN)i]+ 32 — — (C) E A) — — 100 S B) 100 — — n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N0 3 . A) oxinate complexes. B) other species. C) all metal species. Table A.1 (contd.) ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 149 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg 2 + PH 7.2 7.3 7.3 7.3 E [M'(MeOH)i]+ 53 65 £ [M'zY^Ac^MeOHJi]* 47 35 SA) — — — — EB) 100 100 ' ' ' Metal Ion (M"+) Mn 2 + • Zn 2 + Cd 2 + Pb2 + PH 7.2 7.1 7.2 6.7 ZA) —_ — — EB) Metal Ion (Mn+) Cu 2 + Ni 2 + A l 3 + Fe3 + pH 4.4 6.9 8.9 72 6^ 6.4 (A) E [(M"LAq)HJ+ 95 45 11 100 [M"L2H]+ 5 49 77 — [M"2L3r 6 12 — [MmL2r 83 92 [M'"L3H]+ 17 8 (C) EA) 100 100 100 100 100 100 SB) — — — — — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . A) oxinate complexes. B) other species. C) all metal species. Table A.2 ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions in 50:50 water/methanol (% relative abundance). 150 Metal Ion (Mn+) . Co 2 t Cr 3 + phi 7J3 6.5 (A) S [(M"LAq)HJ* -[MnL2H]+ — [M"2L3]+ -[MmL2r — 100 [M'"LaHr — — (C) SA) S B) n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N0 3 . A) oxinate complexes. B) other species. C) all metal species. 100 Table A.2 (contd.) ESI-MS speciation for 8-hydroxyquinoline (HL) and selected metal ions in 50:50 water/methanol (% relative abundance). 151 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg 2 + pH 4.6 7.3 9.4 4.5 7.3 9.4 4.5 7.2 9.1 4.6 7.3 9.1 (A) S [M"HL(HY)j]+ 90 — — 88 — — . Z [M"(HL)2H(HY)j* 9 — — 10 — — [M"2L(HL)(HY)3]+ 1 — — 2 — — (B) Z[M'(ACN)i]+ 92 60 — 96 68 — Z [M'zYflHAc^ACNOif 8 40 — 4 32 -Z [M"(ACN)i]2+ 66 — — 62 — — Z [M'YfHAc^ACN))]* 34 — — 38 — — (C) ZA) — — — — — — 20 — — 18 — — ZB) 100 100 — 100 100 — 80 — — 82 — — Metal Ion (Mn+) Mn 2 + Zn 2 + Cd 2 + Pb2 + pH 4.6 7.2 8.8 4.6 7.0 8.9 4.6 7.0 8.9 4.5 6.7 9.0 (A) Z [M"HL(HY)j]+ 89 Z [M"(HL)2H(HY)j]+ 8 — — 17 — — 11 [M"2L(HL)(HY)3]+ [M"2(HL)4H]+ 81 — — 89 — — 100  3 — — 2 (B) Z [M"(ACN)i]2+ 61 — — 48 — — 73 — — _ _ — Z [M"Y(HAc)j(ACN)i]+ 39 — — 52 — — 27 — — 100 — — (C) ZA) 22 — — 43 — — 52 — — 40 — — ZB) 78 — — 57 — — 48 — — 60 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . A) glyoximate complexes. B) other species. C) all metal species. Table A.3 ESI-MS speciation for dimethylglyoxime (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 152 Metal Ion (Mn+) pH Al |3+ Cr .3+ Fe 3+ 4.1 6 . 4 8 . 6 4 .1 6 . 4 8 . 6 3 . 6 6 . 3 8 . 6 (A) Z [M"(HL)2H(HY)J+ (B) S [M"(AGN)i]2+ £ [M"Y(HAc)j(ACN)i]+ (C) £ A) " SB) 1 0 0 1 0 0 — — — 2 8 — _ _ 7 2 — Metal Ion (Mn+) PH Cu ,2+ Ni 2* Co 2+ 4 . 9 6 . 8 8 . 9 4 . 5 6 . 8 8 . 9 4 . 6 6 . 8 8 . 5 (A) [M^HLJH]* £ [M"HL(HY)J* £ [M"(HL)2H(HY)J+ [MM2L(HL)(HY)3]+ [M"2(HL)4H]+ [M'"(HL)2r [M'"2L2(HL)]+ [MM2L(HL)3r [M"'M"2L2(HL)2]+ (B) £[M'(ACN)ir Z [M'CHACJCACN)^ I2+ S [M"(ACN)i] £ [M'VCHAc^ACN),]* (C) £A) £ B) 1 0 0 7 2 — — 2 8 — 7 3 1 0 0 2 6 — 1 — 1 9 9 7 9 3 5 4 — — 4 6 1 0 0 9 6 — — 4 81 — — 1 9 — — 1 3 1 0 0 1 0 0 8 7 — — 5 6 — — 4 2 — — 2 — — — 9 3 — 1 — 6 5 5 2 7 1 7 7 3 — — 2 7 — — 2 0 1 0 0 1 0 0 8 0 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . A) glyoximate complexes. B) other species. C) all metal species. Table A.3 (contd.) ESI-MS speciation for dimethylglyoxime (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 153 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg 2 t pH 7.2 7.2 7.2 7.2 E [M'(MeOH)i]+ 63 82 £ [Ml2Y(HAc)i(MeOH)J+ 37 18 £ A) — — — — EB) 100 100 Metal Ion (Mn+) Mn2 + Zn 2 + Cd 2 + Pb 2 + PH 7.0 7.2 6.7 EA) — — EB) Metal Ion (Mn+) A l 3 + Cr* Fe3 + PH 6.4 6.5 6.4 E A) — — EB) Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + ' pH 4.9 6.7 8.9 6.7 6.7 £ [M"HL(HY)j]+ 24 __ — E [M"(HL)2H(HY)j]+ 70 100 100 100 — [M"2(HL)3]+ 6 — — — — [M'"(HL)2]+ 94 [Mm2L(HL)3]+ 6 (C) EA) EB) 100 100 100 100 100 n = initial oxidation state, i, j = 0, 1 or 2. Y = Ac or N0 3 . A) glyoximate complexes. B) other species. C) all metal species. Table A.4 ESI-MS speciation for dimethylglyoxime (H2L) and selected metal ions in 50:50 water/methanol (% relative abundance). 154 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg2 + phi 5.4 8.0 9.6 5.4 8.0 9.7 5.3 8.1 9.1 5.3 8.1 9.3 (A) [M'Lf 34 61 — 35 50 — [M'(L2)]+ 65 39 — 65 50 [MML 2f 68 — — 82 [M"(L2)L]+ 32 — — 18 (B) 2 [M'(ACN)i]+ 65 92 — 64 87 — Z [M^YfHAc^ACNOif 35 8 — 36 13 — Z [M"(ACN),]2+ 30 . — — 47 Z [M"Y(HAc)j(ACN)J+ 70 — — 53 (C) SA) 37 22 — 27 10 — 6 — — 6 ZB) 63 78 — 33 90 — 94 — — 94 Metal Ion (Mn+) Mn 2 + Zn 2 + Cd2* Pb 2 + pH 5.3 8.0 9.2 : 5.3 7.4 9.2 5.4 7.4 9.2 5.2 7.4 9.3 (A) [M"l-]+ , , 44 — — 55 — — 7 — — 22 — — [MML 2f 44 — — 29 — — 16 — — 22 — — [M"(L2)L]+ 12 — — 16 — — 77 100 100 56 — 100 (B) Z [M"(ACN)i]2+ 56 — — _ _ _ _ _ _ _ _ _ Z [M"Y(HAc)j(ACN)if 44 — — 100 — — 100 — — 100 — — (C) ZA) 14 — — 35 — — 78 100 100 69 — 100 ZB) 86 — — 65 — — 22 — — 31 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . L2 = neutral dimer. A) carbamate complexes. B) other species. C) all metal species. Table A.5 ESI-MS speciation for A/,A/-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 155 Metal Ion (Mn+) A l 3 + Cr3* Fe3 + pH 4.5 6.9 9.0 4.6 7.0 9.0 4.0 6.8 . 8.9 (C) SA) SB) Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.9 7.4 9.2 5.3 7.9 8.9 5.3 7.9 9.1 (A) [M4L3]"* 4 — — [M'sUr 4 - -[M"L]+ — 10 — 28 — — 19 — — [M"L2r 92 88 92 44 40 21 64 — — [M"L2H]+ _ _ _ 8 22 48 — — — [M" 2L 3f _ 2 7 — — 5 — — — [M"2L,Hf _ _ _ — — 2 _ _ _ [M"3L5]+ - - 1 - - - _ _ _ [M"(L2)L]+ — — — 20 19 23 17 — — [MML2r — 85 81 [M I M L 3 f — 12 3 [M IML3H]+ — — 8 [MM I 2L5]+ - - 3 [M"'(L2)L2]+ — 3 5 (B) S[M'(ACN)i]+ 99 — — S [M'(HAc)(ACN)j]+ ' — — — S [M"(ACN)i] 2 + _ _ _ 66 — — 60 — — S [MVCHAC^ACN)^ 1 — — 34 — — 40 — — (C) SA) 10 100 100 21 100 100 28 100 100 SB) 90 — — 79 — — 72 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . L2 = neutral dimer. A) carbamate complexes. B) other species. C) all metal species. Table A.5 (contd.) ESI-MS speciation for /v,/V-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 156 Metal Ion (Mnt) Na+ K+ Ca 2 + Mg 2 + PH 8.1 8.2 8.1 8.2 [M'Lf 51 50 [M'(L2)]+ 49 50" S [M'(MeOH)i]+ 70 85 S [M'2Y(HAc)j(MeOH)i]+ 30 15 LA) 7 3 — — EB) 93 97 Metal Ion (Mn+) Mn 2 + Zn 2 + Cd 2 + Pb 2 + pH 8.0 7.5 7.9 7.4 SA) _ — SB) n = initial oxidation state. i, j = 0, 1 or 2. Y = A c - or N0 3 ~ L2 = neutral dimer. A) carbamate complexes. B) other species. C) all metal species. Table A.6 ESI-MS speciation for /V,/V-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/methanol (% relative abundance). 157 Metal Ion (Mn+) A l 3 + Cr 3 + Fe3 + pH 4.5 6.9 9.0 4.6 7.0 9.0 4.0 6.8 8.9 (C) SA) S B) Metal Ion (Mn+) Cu 2 + Ni 2 + Co2 pH 4.8 7.3 9.2 7.9 8.1 (A) [MML]+ 39 — — — — [MML 2f 56 96 41 33 — [M"L2H]+ — — 51 51 — [M"2L3]+ 6 4 8 ^ — [M"2UH]+ — [M"3L5]+ 1 - — — — [M"(L2)l_r - - - 15 -[M"'L2]+ 86 [M"'L3r 4 [MmL3H]+ 8 [M'"2L5]+ -[Mm(L2)L2]+ 2 (C) SA) 100 100 100 100 100 SB) — — — — — n = initial oxidation state. L2 = neutral dimer. A) carbamate complexes. B) other species. C) all metal species. Table A.6 (contd.) ESI-MS speciation for /V,/V-diethyldithiocarbamate (HL) and selected metal ions in 50:50 water/methanol (% relative abundance). 158 Metal Ion (Mn) PH Na+ K+ Ca 2 + Mg' 4.5 7.3 8.3 4.5 7.3 9.2 4.6 7.3 9.3 4.5 7.3 9.1 ,2+ (A) [M'L]+ [M'L2r [M"LY]+ [M"l_2Y]+ 92 92 8 8 90 10 92 — 8 — 94 6 96 4 (B) Z[M'CACN),f Z [M'2Y(HAc)j(ACN)j]+ Z [M"(ACN)J2+ Z [M"Y(HAc)j(ACN)i]+ 94 79 4 21 96 85 4 15 62 38 — 62 — 38 (C) ZA) ZB) 12 12 88 88 6 7 94 93 1 99 — 1 — 99 Metal Ion (Mn+) PH 2+ Mn 4.5 7.3 8.9 ,2+ Zn' 4.5 7.0 8.8 Ccr 4 Pb' 2+ 4.5 7.1 9.0 4.4 6.7 9.1 (A) [MnLY]+ [M"L(HL)]+ [M"L2]2+ [MML2Y]+ [M"L2(HL)]+ [M"L3Y]+ [M"U]2+ 67 — — 51 — — 43 — — 7 4 — — 5 — — 2 — — 15 18 — — 13 — — — — — 14 6 — — 7 — — 8 — — 12 1 — — 7 — — 3 — — 6 4 _ _ 18 — — 42 — — 42 (B) Z [M"(ACN)i]2+ 55 — Z [M"Y(HAc)j(ACN)i]+ 45 — 38 — — 62 — — 36 64 100 — (C) ZA) Z B) 99 — — 96 — — 4 96 84 16 n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . (HL) = deprotonated H2L. A) dithizonate complexes. B) other species. C) all metal species. Table A.7 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 159 Metal Ion (Mn+) A l 3 + Cr3* Fe3 + JDH 4.1 6.5 8.7 4.1 6.5 8.6 3.6 6.4 8.8 (C) ZA) _ _ _ _ _ _ _ _ _ ZB) _ _ _ _ _ _ _ _ _ Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.2 6.8 9.0 4.5 7.2 8.8 4.5 7.3 8.9 (A) [M'Lf 77 16 — [M'L2]+ 20 70 100 [M'2LY]+ — 4 — [M'2L2Y]+ 2 10 -[M'3L2Y2]+ 1 — — [M"l_Y]+ — — — 70 ' — — 74 — — [M"L(HL)]+ _ _ _ 4 — — 1 — — [M"L2]2+ _ _ _ 1 5 _ _ 1 3 — — [M"L2Y]+ — — — 4 — — 7 — — [MML2(HL)]+ _ _ _ 1 — — 1 — — [M"L3Y]+ _ _ _ 5 — — 4 — — (B) Z[M'(ACN)ir 100 100 — Z [M'(HAc)(ACN)J+ — — — Z [M"(ACN)i]2 • _ _ _ 73 — — 58 — — Z [MnY(HAc)j(ACN)i]+ _ _ _ 27 — — 42 — — (C) ZA) 4 57 100 1 — — 2 — — ZB) 96 43 — 99 — — .98 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . (HL) = deprotonated H2L. A) dithizonate complexes. B) other species. C) all metal species. Table A.7 (contd.) ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 160 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg2+ PH 7.3 . 7.2 7.3 7.3 [M'L]+ 91 89 [M'L2]+ 9 11 2 [M'(MeOH)i]+ 90 92 2 [M'2Y(HAc)j(MeOH)j]+ 10 8 2 A) 6 6 — — 2 B) 94 94 Metal Ion (Mn+) Mn2+ Zn2+ Cd 2 + Pb2+ PH 7.3 7.1 7.2 6.6 2 A) — 2B) n = initial oxidation state. i, j - 0 , 1 or 2 . Y = Ac" or N03. A) dithizonate complexes. B) other species. C) all metal species Table A.8 ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in 50:50 water/methanol (% relative abundance). 161 Metal Ion (Mn+) A l 3 + Cr3* Fe3 + pH 6.3 6^ 6 6.3 (C) SA) - - -SB) — — ' — Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.2 6.8 9.0 7.3 7.2 [M'L]+ 7 — — [M'L2]+ 33 — — [M"LY]+ 8 — — — — [M"L2]2+ 48 — — — — [M"L2Y]+ 3 — — — — [M"L3Y]+ 1 — — — — (C) SA) 100 SB) — n = initial oxidation state. Y = Ac or N0 3 . A) dithizonate complexes. B) other species. C) all metal species. Table A.8 (contd.) ESI-MS speciation for diphenylthiocarbazone (L) and selected metal ions in 50:50 water/methanol (% relative abundance). 162 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg2 + pH 4.2 6.6 8.9 4.2 6.6 8.8 4.2 6.7 9.0 4.3 6.7 8.9 (A) £ [M'x(H4-xL)]+ 89 100 — 89 100 — I [M'x(H4-xL)(H3L)]+ 10 - - 9 - -£ [M'x(H4-xL)(H3L)2]+ 1 - - 2 - -[M"(H2L)]+ 25 — — 23 — — tM"(H3L)]2+ 68 — — 71 — — £ [M"x(H4-2xl_)]+ 2 — - 2 — — [M"3(HL)2]2+ 3 — — 2 — — [M"3(HL)(H2L)2]2+ 1 — — 1 — — [MM4(HL)3]2t 1 - - 1 - -(B) S[M'(ACN)i]+ 100 77 — 100 97 — £ [M'2Y(HAc)j(ACN)i]+ — 23 — — 3 — S [MM(ACN)i]2+ 73 — — 75 — — £ [MnY(HAc)j(ACN)j]+ 27 — — 25 — — (C) SA) 11 15 — 3 7 — 4 — — 5 — — IB) 89 85 — 97 93 — 96 — — 95 — — Metal Ion (Mn+) Mn 2 + Zn 2 + Cd2* Pb2 + pH 4.3 6.7 8.5 4.3 6.4 8.2 4.3 6.4 8.5 4.2 6.5 8.5 (A) [M"(H2L)f 30 — — 70 — — 96 — — 82 — — [M"(H3L)]2+ 61 — — 18 — — _ _ _ _ _ _ £ [M"x(H4-2XL)]+ 2 - - 4 - - 4 - - 1 8 - -[M"3(HL)2]2+ 4 - - 4 - - _ _ _ _ _ _ [M"3(HL)(H2L)2]2+ 2 - - 2 - - - - - - - -[MM4(HL)3]2+ 1 - - 2 - - - - - - - -(B) £ [M"(CAN)i]2+ 67 — — 50 — — 61 — — — — — £ [M"Y(HAc)j(ACN)i]+ 33 — — 50 — — 39 — — 100 — — (C) £ A) 4 — — 2 — — 2 — — 27 — — £ B) 96 — — 98 — — 98 — — 73 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . x = 1,2 or 3. A) glutathionate complexes. B) other species. C) all metal species. Table A.9 ESI-MS speciation for glutathione (H3L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 163 Metal Ion (Mn+) A l 3 + Cr3* Fe3* pH 4.0 6.2 8.4 4.0 6.0 8.8 3.8 6.1 8.4 (A) [MIMHL]+ 95 — — 22 — — 100 — — E [MMIHL(H20)il* — — — 74 — — — — — [M'"HL(H3L)]+ 5 — — 4 — — — — — (C) SA) 100 '— — 100 — — 100 — — SB) _ _ _ _ _ _ _ _ _ Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.0 6.4 8.5 4.3 6.4 8.5 4.3 6.5 8.7 (A) S [M'x(H4.xL)]+ 100 - -[M"(H2L)]+ — — — 29 — — 50 — — [M"(H3L)]2+ — — — 63 — — 37 — — S [M"x(H4.2xL)r _ _ _ 3 - - 5 - -[M"3(HL)2]2+ _ _ _ 2 - - 4 - -[M"3(HL)(H2L)2]2+ _ _ _ 2 - — 2 — — [M"4(HL)3]2+ _ _ _ 1 - - 2 - — [M"(H3L2)r — 100 100 — — 100 — — — (B) S[M'(ACN)i]+ 100 100 — S [M'(HAc)(ACN)i]+ — — — S [M"(ACN)i]2+ _ _ _ 84 — — 66 — — . S [M"Y(HAc)j(ACN)i]* _ _ _ 6 — — 34 — — (C) SA) 1 11 100 2 — 100 2 — — SB) 99 89 — 98 — — 98 — — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . x = 1,2 or 3. (H3L2) = oxidized ligand. A) glutathionate complexes. B) other species. C) all metal species. Table A.9 (contd.) ESI-MS speciation for glutathione (H3L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 164 Metal Ion (M+) Na+ K+ pH 4 . 2 6 . 6 8 . 9 4 . 2 6 . 6 8 . 8 [M'(H3L)]+ 9 0 6 2 — 8 4 6 5 — [M'2(H2L)]+ 1 0 2 3 — 1 6 2 6 — [M'3(HL)]+ — 1 5 — — 9 — Table A. 10 Multiple binding of alkali metal ions by the glutathione ligand (H3L) in 50:50 water/acetonitrile. 165 Metal Ion (Mn+) Na+ K+ Ca 2 + Mg 2 + pH 6.7 6.7 6.7 6.7 2 [M'x(H4-xL)]+ 100 100 2 [M'x(H4.xL)(H3L)]+ — ' 2 [Mlx(H4-xL)(H3L)2]+ — — 2 [M'(MeOH)i]+ 75 91 2 [M'2Y(HAc)j(MeOH)j]+ 25 9 2 A) 16 8 — — 2B) 88 92 Metal Ion (Mn+) Mn 2 + Zn 2 + Cd 2 + Pb2 + PH 6.7 6.5 6.6 6.5 2 A) — __ 2B) Metal Ion (Mn+) Al 3 + Cr* Fe3 + pH 6.1 6.2 6.2 (C) 2 A) _ 2B) Metal Ion (Mn+) Cu 2 + i. Ni 2 + Co 2 + pH 4.0 6.4 8.7 6.6 6.7 (A) [M"(H3L2)]+ 34 100 68 — [M"(H3L2-Glu)]+ 11 — 21 [M"(H4L2)]2+ 40 '— 11 (C) 2 A) 1 11 100 2 B) 99 "89 — n = initial oxidation state, i, j = 0,1 or 2. Y = Ac or N0 3 . x = 1,2 or 3. (H 3L 2) = oxidized ligand. A) glutathionate.complexes. B) other species. C) all metal species. Table A. 11 ESI-MS speciation for glutathione (H3L) and selected metal ions in 50:50 water/methanol (% relative abundance). 166 Metal Ion (Mn+) pH Na+ K+ Ca 2 + Mg2+ 4.2 6.6 8.9 4.1 6.7 9.0 4.1 6.6 8.9 4.1 6.7 8.8 (A) [M'(HL)Y]_ [M'(HL)2] [M"LN03(HN03)]" [MM2L(N03)3]~ E[M"L(HL)(HN03)j]' [M"(HL)3] 74 64 26 36 74 68 26 32 83 — 6 — 9 — 2 — — 60 — 9 — 26 — 5 iv i -(B) E[M'Y2] [M'2(N03)3] [M"(N03)3]" 92 100 8 — 93 100 7 — 100 — 100 — — (C) E A) EB) 55 23 45 77 44 21 56 79 71 29 71 29 Metal Ion (Mn) PH Mn 2+ Zn 2+ Cd 2* Pb 2+ 4.1 6.7 8.5 4.1 6.6 8.4 4.1 6.7 8.6 4.1 6.4 8.7 (A) [M"LN03(HN03)]~ 72 — — 69 — — 75 — — 87 [M"2L(N03)3]~ 6 — — 4 — — 18 — — — E[M"L(HL)(HN03)j]~ 7 — — 11 — — 3 — — 13 E [M'^MNO^HNO^]- 1 2 - - 14 - - 3 - - -E [M"xL2(HL)(N03)2x.4] 3 - - 2 - - 1 - - -(B) [M"(N03)3]' 100 — — 100 — 100 — 100 — (C) EA) EB) 75 — 25 — — 77 — 23 — 81 — 19 — 77 — 24 n = initial oxidation state, j = 0,1 or 2. Y = Ac or N0 3 . x = 2 or 3. A) salicylate complexes. B) other species. C) all metal species. Table A. 12 ESI-MS speciation for salicylic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 167 Metal Ion (Mn+) Al 3 + Cr3* Fe3* pH 3.8 6.1 8.3 3.9 6.2 8.3 3.6 6.1 8.3 (A) [M^NOa^ ] - 55 — — 26 — — 59 — — Z[M l l lL(N0 3) 2(H 20)i]~ _ _ _ 40 — — 4 — — [M I"(HL)(N03)3] - _ _ _ _ 79 _ _ _ _ E[M l"L 2 (HN0 3 ) j ]" 35 — — 26 21 — 27 — — S[M l l lL 2(HN0 3) j(H 20)i]~ — — — 8 — — — — — [M"'2L2(N03)3] _ 4 - - - - - - - -[M'"M"L2(N03)2] _ _ _ _ _ _ 10 — — [M"'2L3N03]' (B) [M'"(N03)30H] — — — 100 — — — — — (C) 2 A) 100 — — 35 100 — 100 — — EB) _ _ _ 65 — — — — — n = initial oxidation state, i, j = 0, 1 or 2. A) salicylate complexes. B) other species. C) all metal species. Table A.12 (contd.) ESI-MS speciation for salicylic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 168 Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.1 6.5 8.5 4.1 6.7 8.6 4.1 6.7 8.6 (A) ; ,[M'L]~ • ' 20 — — [M'(L-C02)N03]~ ' — 61 — [M^HLXNC^]" , . — • ' 3 — [M'(HL)2]~ 30 — 100 [M'(HL)(HL-.C02)]~ s 50 — — [M'2L(HL-C02)]~ — 32 — [M'2(HL)2N03]" — 3 — [M'2(HL)3]" - 1 -[M"LN03(HN03)] —. — — 84 — — 74 — — [MM2L(N03)3]~ — — — 7 — — 10 — — E [M"L(HL)(HN03)j]~ — — — 4 — — 5 — — E [MV2(N03)2x.3(HN03)jr - - - 4 - - 9 - -E [M"xL2(HL)(N03)2x.4] - - - 1 - - 1 — — (B) [M'(N03)2]~ 57 — — E [M'(N03)2(HY)] ~ 43 — — [M"(N03)3]~ — — — 100 — — 100 — — (C) EA) 4 8 1 0 0 1 0 0 65 — — 91 — — E B) 52 — — 35 — — 9 — — n = initial oxidation state, j = 0, 1 or 2. Y = Ac or N0 3 . x = 2 or 3. A) salicylate complexes. B) other species. C) all metal species. Table A. 12 (contd.) ESI-MS speciation for salicylic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 169 Metal Ion (Mn*) Na* K* Ca2* Mg2* PH 6.6 6.7 6.6 6.6 [M'L]" 75 75 [M'(HL)2] 25 25 Z[MY 2 ]~ 100 100 2 A) 24 23 — — 2 B) 76 77 Metal Ion (Mn*) Mn2* . Zn2* Cd2* Pb2* PH 6.6 6.5 6.7 6.5 2 A) __ — — 2 B) • ; . — ~ ~ " ' n = initial oxidation state. Y = Ac"or N 0 3 " A) salicylate complexes. B) other species. C) all metal species. Table A. 13 ESI-MS speciation for salicylic acid (H2L) and selected metal ions in 50:50 water/methanol (% relative abundance). 170 Metal Ion (Mn+) A l 3 + Cr 3 + Fe3 + pH 6.2 6.2 ' 6.1 (A) [MIM(HL)(N03)3]" - 9 -[MmL2] — - . 9 1 — (C) SA) — 1Q0 — SB) — — — Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + pH 4.1 6.5. 8.5 6.7 6.7 (A) [Ml2(HL)(N03)-2]"_ 91 - -S [M"L(HL)(HN03)j] 2 — — — — S[M"2L2(N03)(HN03)j]~ 5 — — — — [M"2L2(HL)] 2 - - - — [M"(N03)3]" . 100 — — — — (C) SA) 60 — — — — SB) 40 — — — — n = initial oxidation state, j = 0, 1 or 2. x = 2 or 3. A) salicylate complexes. B) other species. C) all metal species. Table A. 13 (contd.) ESI-MS speciation for salicylic acid (H2L) and selected metal ions in 50:50 water/methanol (% relative abundance). 171 Metal Ion (Mn+) Na+ K+ pH 4.1 6.7 8.7 4.1 6.7 8.5 [M'i_r 11 — 17 [M'(HL)Y] 16 29 — 15 19 — [M'(HL)2] 59 24 — 67 33 — E [M^UHLXHNO:,),]" 8 12 — 5 9 — E [Ml3L2(HN03)j]" 3 11 — 1 9 — E [M'xGHsJ-a)]" 10 8 — 9 8 — E [M'x(H7-xU)] 4 5 — 3 5 — Ca 2 + Mg2 + 4.1 6.7 8.7 4.1 6.7 8.3 (A) _ 74 — 60 — — E [M"LN03(HN03)i] q _ _ 1 4 _ — E [M"2L2(N03)(HN03)j] 7 _ _ 17 _ -[M"(HL)3]- ' _ _ 3 _ _ E [ M ' ^ H L X H N O^r _ _ 1 _ _ [M"2L(HL)3] \ __ __ . _ _ E [M"3L3(HL)(HN03)j]~ 1 _ _ _ _ [M"2(HL)5] 1 (B) E[M'Y 2]- _ 92 100 - 93 100 -[M'2(N03)3] 8 - - 7 - - 1 Q 0 _ _ [M"(N03)3]- 1 0 0 1 0 0 (C) T. X\ 47 67 - 44 62 - 33 - - 21 - -(C> I* 53 33 - 56 38 - 67 79 n = initial oxidation state, j = 0,1 or 2. Y = Ac or N0 3 . x = 1 to 7. A) phthalate complexes. B) other species. C) all metal species. Table A. 14 ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 172 Metal Ion (Mnt) Mn 2 + Zn 2 + . Cd 2 + Pb2* pH 4.1 6.7 8.3 4.1 6.7 8.4 4.1 6.7 8.3 4.1 6.6 8.3 (A) Z [MYN0 3(H^I0 3) j^ £ [M\(HL)(HN03)j]~ E [M"2L2(N03)(HN03)j] [M"(HL)3] E [M'^HLKHNO^] [M"2L(HL)3] E [M"3L3(HL)(HN03)j] [M"2(HL)5] 67 — — 10 — — 9 — — 7 — — 4 — — 1 — — 1 — — 1 — — 55 — — 32 — — 4 — — 4 — — 2 — — 1 — — 1 — — 1 — — 61 — — 31 — — 8 — — 22 — — 65 — — 12 — — (B) [M"(N03)3]" 100 — — 100 — — 100 — — — —. — (C) S A) EB) 38 — — 62 — — 54 — — 46 — — 71 — — 29 — — 100 — — n = initial oxidation state, j = 0,1 or 2. A) phthalate complexes. B) other species. C) all metal species. Table A. 14 (contd.) ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 173 Metal Ion (Mn+) A l 3 + Cr3* Fe3 + PH 3.9 6.2 8.1 3.9 6.2 8.1 3.6 6.2 8.1 E [MYN0 3(HN0 3)J]~ 64 — — E [M"L(HL)(HN03)J] - 21 — — [M"(HL)3]" 12 — — [M"2L(HL)3] 2 — — [M"2(HL)5] 1 — — [MmL(N03)2] 37 E[MmL(N03)2(H20) i]" — — — 48 — — — — — E [MmL(HL)(N03)(H20)i] — — — 34 24 — — — — E[Mm 2L2(N03)3(H20)i]~ — — — 3 11 — — — — [M'"L2] 53 — — 3 — — — — — [M'"L(L-C02)] — — — — 35 — — — — [Mm(L-C02)2]" — — — — 30 — — — — [M'"L(HL)2] 1 E [M'^ CH^zCHzO);] — — — 12 — — — — — [MIM2L3N03] 4 E [M'^U^g^zO);]" 3 S [Mm2L2(HLh(H20)] 2 [M'"(N03)3OH]~ — — — 100 — — — — — EA) 100 — — 77 100 — 100 — — EB) — — 23 — — — — — n = initial oxidation state, i, j = 0, 1 or 2. A) phthalate complexes. B) other species. C) all metal species. Table A. 14 (contd.) ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 174 Metal Ion (Mn+) Cu 2 + Ni 2 + Co 2 + jpH 4.1 6.6 8.5 4.1 6.7 8.5 4.1 6.7 8.4 (A) [M'L]~ 66 77 — [M'(HL)2]" 19 23 100 IfM'zLfHLKHNOa),]- 12 — — E [M'aLsCHNOaJJ- 2 — — SlM^Hs-xLa)]" 1 - -E [M"LN03(HN03)j] — — — 61 — — 59 — — E [M"L(HL)(HN03)j]~ — — — 6 — — 15 — — E (Mll2L2(N03)(HN03)j]~ — — — 15 — — 12 — — [M"(HL)3]" . — — — 9 — — 4 — — E[M"2L2(HL)(HN03)J" — — — 6 — — 6 — — [M"2L(HL)3]~ — — — 1 — — 2 — — E [M"3L3(HL)(HN03)j]"" — — — 1 — — 1 — — [M"2(HL)5] - - - 1 — - 1 — — (B) [M'(N03)2]" 100 — — — — — — — — [M"(N03)3]~ — — — 100 — — 100 — — (C) EA) 77 100 100 37 — — 85 — — EB) 23 — — 63 — — 15 — — n = initial oxidation state, j = 0,1 or 2. x = 1 to 5. A) phthalate complexes. B) other species. C) all metal species. Table A.14 (contd.) ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/acetonitrile (% relative abundance). 175 Metal Ion (M+) PH Na+ K+ 4.1 6.7 8.7 4.1 6.7 8.5 92 66 — 7 18 — 1 17 — [M'(HL)2]~ 85 52 — [M'2L(HL)]~ 11 26 — [M'3(L)2]" 4 22 -Table A. 15 Multiple binding of alkali metal ions by the phthalate ligand (H2L)2 in 50:50 water/acetonitrile. Metal Ion (M+) Na+ K+ pH 4.1 6.7 8.7 4.1 6.7 8.5 [M'(HL)2(H2L)] 8 4 — 24 6 [M2(HL)3] 80 40 — 70 51 [M'3L(HL)2]~ 6 24 — 6 27 [M'4L2(HL)]~ 6 21 — — 9 [M' 5L 3]" — 11 — — 7 Table A. 16 Multiple binding of alkali metal ions by the phthalate ligand (H2L)3 in 50:50 water/acetonitrile. Metal Ion (M+) Na+ K+ pH 4.1 6.7 8.7 4.1 6.7 8.5 [M'(HL)2(H2L)2r [M'2(HL)3(H2L)] 41 9 — 56 11 — [M'3(HL)4] 42 28 — 30 37 — [M'4L(HL)3] 17 29 — 14 28 — [M'5L2(HL)2] — 18 — — 12 — [M'6L3(HL)] 11 — — 6 — [M'7L4] 5 — 6 — Table A: 17 Multiple binding of alkali metal ions by the phthalate ligand (H2L)4 in 50:50 water/acetonitrile. 176 Metal Ion (Mn+) Na+ Ca 2 + Mg 2 + pH 6.7 6.7 6.7 6.7 [Ml]"* 22 27 [M'(HL)Ac]~ 18 12 [M'(HL)2]" 24 38 [M'2L(HL)] 12 6 [M'3L2] 13 9 2 [M^Hs-xLs)] 11 8 2[M'(Ac)2]- 100 100 2 A) 60 61 — — SB) 40 39 Metal Ion (Mn+) Mn 2 + Zn 2 + Cd 2 + Pb2 + PH 6.7 6.6 6.6 6.5 2 A) 2 B) — - . n = initial oxidation state. x = 1 to 5. A) phthalate complexes. B) other species. C) all metal species. Table A. 18 ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/methanol (% relative abundance). 177 Metal Ion (Mn+) A l 3 + Cr3* Fe3 + pH , 6.6 6.4 6.2 [MML(Ac)2]" . 16 [Ml"(L-C02)(Ac)2] — 23 — [M'"L2] 18 — [M'"L(L-C02)] — 15 — [M"'(L-C02)2]" — 22 — [MmL(HL)2] — 6 — I A ) — 100 Metal Ion (NT) PH (A) S [MYN03(HAC)J] [M"(L-C02)N03] [MY(HL)] - _ [M"2L2(N03)] [M"(HL)3]~ (B) [M"(N0 3) 3]" (C) 2 A) SB) Cu 2 + 4.2 6.5 8.5 41 — — 44 — — 13 — — 100 — — 70 — — 30 — — n = initial oxidation state, j = 0, 1 or 2. A) phthalate complexes. B) other species. C) all metal species. Table A. 18 (contd.) ESI-MS speciation for phthalic acid (H2L) and selected metal ions in 50:50 water/methanol (% relative abundance). APPENDIX B. SUPPLEMENTARY DATA FOR CHAPTERS 4 AND 5 (a) 1.6E+8 2 3 4 5 6 7 8 9 P H Figure B.1 Theoretical ES ion intensity I (—) and observed intensity for (a) salicylate [HL]" ion (•) and (b) phthalate [HL]" (•) and [L]2~ (•) ions as a function of solution pH (Section 4.4.1). 180 (a) (b) 1.5E+7 c 1.2E+7 H 8 9.0E+6 •g 6.0E+6 H J 3.0E+6 -| ~~ 0.0E+0 (C) 0 10 20 30 40 Total copper concentration (u.M) 50 Figure B.2 Intensity of (a) [HLH]+ (•), (b) [Cu"L]+ (A) and (c) [Cu'AC'N]* (•) ions generated from a solution of 10 u.M 8-hydroxyquinoline (HL) and 100 u.M copper(ll) nitrate in 50:50 water/acetonitrile (pH 4.2) (Section 4.4.1). Lines show (a) best fit and (b, c) linear regression for data points from 1-50 pM total copper (r2 = 0.9990 and 0.9961, respectively). 181 8.0E+6 -Mi 6.0E+6 •4—* c O o ^ 4.0E+6 w c CD c 2.0E+6 O.OE+0 4 5 6 7 8 9 pH Figure B.3 ES ion yield for oxidized /V,/V-diethyldithiocarbamate [(L2)H]+ (A) and oxidized glutathione [(H4L2)H]+ (•) in 50:50 water/acetonitrile as a function of solution pH (Section 4.4.2). Lines show best fit. 1.0E+8 ES cone voltage (V) Figure B.4 Distribution of salicylate ions [(HL)f (!) and [(HL-C02)f (•) as a function of electrospray cone voltage (Section 4.4.4). Lines show best fit. Figure B.5 Collision induced dissociation of the salicylate complex [Cu'(HL)2] at ES cone voltages greater than 20 V (Section 4.4.4). 2.0E+5 w 1.5E+5 £Z ZJ. o o ^ 1.0E+5 CO c CD c 5.0E+4 0.0E+0 30 40 50 ES cone voltage (V) 60 Figure B.6 Relative intensity of singly- and doubly-charge tannin ions [L2*-Hf (•) and [L2*-2H]2" (•). generated from a solution of 100 pM tannic acid in 50:50 water/acetonitrile (pH 6) as a function of electrospray cone voltage (Section 5.4.1). Source temperature - 60 °C, drying gas flow rate = 8.3 L/min. 183 (a) (b) (c) Figure B.7 10 n 8 J L N J ill i f 60 80 100 120 ES Source Temperature (°C) Relative abundance of (a) gallic acid, (b) Li* and (c) L2* ions generated from a solution of 100 u.M tannic acid in 50:50 water/acetonitrile (pH 6) as a function of ES source temperature and drying gas flowrate (black 5.0, grey 6.7, white 8.3 L/min). ES cone voltage = 60 V (Section 5.4.1). 184 (d) (e) (f) 60 80 100 ES Source Temperature (°C) 120 Figure B.7 (contd.) Relative abundance of (d) L4*, (e) L5* and (f) L6 tannin ions generated from a solution of 100 u.M tannic acid in 50:50 water/acetonitrile (pH 6) as a function of ES source temperature and drying gas flowrate (black 5.0, grey 6.7, white 8.3 L/min). ES cone voltage = 60 V (Section 5.4.1). 185 OH OH O (a, 126) (b, 170) (c, 302) Figure B.8 Phenolic compounds other than tannins detected by ESI-MS in a solution of 100 u,M tannic acid in 50:50 water/acetonitrile (pH 6-8) (Section 5.4.2); (a) pyrogallol, (b) gallic acid and (c) ellagic acid (molecular mass shown in brackets). 1.0E+6 - 2.0E+5 0.0E+0 Concentration (uM) Figure B.9 ESI-MS calibration for salicylic acid (m/z 137) in 50:50 water/ acetonitrile (pH 6) using the same conditions as for tannic acid (Section 5.4.2). Line shows linear regression (r2 = 0.9995). 186 (a) 0.08 j 0.06 --_j 0.04 -o (b) 0 -I , 1 . 1 0 0.1 0.2 0.3 0.4 Cu T (mM) Figure B.10 CuL/L' ion intensity ratio versus total copper concentration (added as Cu"CI2) for tannin ligands (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 6 and (b) pH 8 (ionic strength ~1 mM; Section 5.4.4). Lines show 2nd order polynomial regression. 187 (a) 0.16 0.12 _ — 0.08 -J O 0.04 0 (b) 0.16 0.12 _ —J 0.08 O 0.04 0 0 0.1 0.2 0.3 0.4 Cu T (mM) Figure B.11 CuL/L1 ion intensity ratio versus total copper concentration (added as Cu'CI) for tannin ligands U* (•), L2* (•) and L3* (A) in 50:50 water/acetonitrile at (a) pH 6 and (b) pH 8 (ionic strength ~1 mM; Section 5.4.4). Lines show (a)>2nd and (b) 3rd order polynomial regression. 

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.831.1-0061569/manifest

Comment

Related Items