Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of analytical methods for arsenic speciation Pergantis, Spiridon 1994

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

Item Metadata

Download

Media
831-ubc_1994-894369.pdf [ 3.61MB ]
Metadata
JSON: 831-1.0061715.json
JSON-LD: 831-1.0061715-ld.json
RDF/XML (Pretty): 831-1.0061715-rdf.xml
RDF/JSON: 831-1.0061715-rdf.json
Turtle: 831-1.0061715-turtle.txt
N-Triples: 831-1.0061715-rdf-ntriples.txt
Original Record: 831-1.0061715-source.json
Full Text
831-1.0061715-fulltext.txt
Citation
831-1.0061715.ris

Full Text

DEVELOPMENT OF ANALYTICAL METHODS FOR ARSENIC SPECIATION by SPIRIDON PERGANTIS B.Sc., University of loannina, Greece, 1988 A THESIS SUBMI’fliiD IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CHEMISTRY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA APRIL 1994 © Spiridon Pergantis, 1994  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—keM  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  ES,, IS4.  ABSTRACT  Metal speciation has become increasingly important in making risk assessments of metal toxicities. No longer is the knowledge of the total elemental content adequate for such a purpose, it is now essential to know the chemical form and concentration of all species of the element, including oxidation state of the metal and the type and number of substituents. Mass spectrometry is a very powerful tool for the identification of arsenic species in environmental samples. In particular Desorption chemical ionization  -  mass spectrometry  (DCI-MS) produces satisfactory spectra with 100 ng arsenic placed on the DCI filament, making the technique eminently suitable for the investigation of environmental samples. The spectra of the arsenicals exhibit molecular ions as well as characteristic fragment ions. To extend the usefulness of low resolution DCI-MS, an analytical method was developed that permitted the use of mass deficient reference standards for calibration purposes in accurate mass measurements of positive ions under ammonia DCI conditions. This was accomplished by employing a mixture of ammonia and methane as reagent gases. In the high resolution accurate mass measurement experiment, this gas mixture allows for simultaneous detection of the mass spectrum of perfluorokerosene (calibration substance) adequate for calibration purposes, and the spectrum of the analyte which contains molecular weight information. For positive ion accurate mass measurements of higher masses (up to m/z=2300), Fomblin 18/8 oil was used successfully as a reference standard under ammonia, methane, and iso-butane desorption chemical ionization conditions. Both low and high resolution DCI-MS were used to identify arsenic compounds present in Mytilus californianus. This was accomplished by first extracting the arsenicals from the mussel flesh with methanol, and then isolating and purifying the compounds by means of conventional chromatography. Subsequently the mass spectrometric techniques described above were used to analyze the purified materials and provide spectra suitable for  11  the structural characterization of two principal arsenic species present in the mussels; arsenobetaine (AsB) and the tetramethylarsonium ion. Matrix assisted laser desorption ionization  time of flight  -  -  mass spectrometry  (MALDI-TOF-MS) is a very sensitive mass spectrometric technique which can detect as little as 0.3 ng of arsenic or 4 pmole of the arsenical under investigation. This feature suggests that the method could be used for the identification of minor arsenic components in environmental samples. Additional advantages of the technique are its capability of providing molecular ions as well as fragment ions for a variety of arsenicals, and allowing for a certain 2 laser power. Quantitative degree of control in compound fragmentation, by adjusting the N  analysis by MALDI-TOF-MS is still considered a highly unreliable procedure since many variables associated with sample preparation and analytical procedure can seriously affect the results. Hydride generation  -  gas chromatography  -  mass spectrometry (HG-GC-MS),  3 incorporation developed in this work, was used to provide conclusive evidence of -CD e-methyl-d into arsenic compounds produced from arsenate by alga cell from L-methionin 3 cultures. These findings strongly support the notion that the oxidation-reduction pathway involving carbonium ions, as originally suggested by Challenger for the alkylation of arsenic by microorganisms, applies to marine unicellular alga and probably other marine organisms. Micro  -  liquid chromatography (LC) columns (0.32 mm inner diameter) fabricated and  packed in house can be conveniently coupled on-line to a variety of mass spectrometric systems, mainly because of the extremely low flow rates they require. It was shown that a 99 % reduction in the volume of solvent waste is achieved by switching from conventional to micro-LC, and a 87.5 % reduction by switching from microbore to micro-LC. Compared to previous work, lower detection limits for arsenic compounds were also obtained when micro-LC columns were employed in conjunction with ultraviolet (UV) and electrothermal atomic absorption (ETAA) detectors.  HI  The micro-LC columns are able to efficiently separate arsenicals that are used as animal feed additives as well as their potential metabolites. The separation efficiency achieved on this type of column was shown to be similar to those achieved on microbore (1 mm inner diameter) and conventional (4.5 mm inner diameter) high performance liquid chromatography (}{PLC) columns. Simplex optimization was used to efficiently delineate the optimum experimental conditions to be used for the electrothermal atomic absorption spectrometric analysis of arsenic in a standard reference material of marine origin. Four experimental variables, were considered: ashing temperature, atomization temperature, modifier concentration, and atomization ramping time. This combination of methods and materials provides a powerful means of rapidly optimizing the experimental conditions used for the analysis of arsenic in a wide variety of samples of environmental origin. Excellent recoveries of arsenic were obtained when using the optimum electrothermal atomic absorption spectrometry conditions to analyze standard solutions of arsenobetaine, arsenocholine and tetramethylarsonium iodide. This procedure also allowed for the accurate determination of arsenic in Californian mussels collected from a variety of locations along the B.C. coast, which was found to range from 9 to 16 Finally, in Chapter 6 it was demonstrated that methylarsonic acid (MAsA) is a likely precursor to AsB in the marine environment. Mytilus calfornianus exposed for 9 days to a J-. [ H ]-methylarsonic acid, was found to contain 3 [ H seawater system containing 3 -labeled compounds 3 j-arsenobetaine and two unknown H [ H methylarsonic acid along with 3 in the tissue parts of this mussel. A linear increase with time in the specific activity present in the flesh of Mytilus calfomianus was also observed indicating uptake of the labelled compound or its metabolites. The highest specific activity was found in the visceral mass and the gills of the mussel.  ‘V  TABLE OF CONTENTS  ii  ABSTRACT TABLE OF CONTENTS  V  LIST OF TABLES LIST OFFIGURES  xi xii  LIST OF ABBREVIATIONS ACKNOWLEDGMENTS  xviii xxii  CHAPTER 1 GENERAL INTRODUCTION  1  1.1 Arsenic Compounds Present in the Environment 1.1.1 Organoarsenic Compounds Present in Marine Organisms 1.1.2 Arsenic Biomethylation 1.2 Determination of Arsenic Species Present in Environmental Samples 1.2.1 Methods used for the Identification of Arsenic Compounds 1.2.2 Trace Speciation of Arsenic 1.2.2.1 Hydride Generation 1.2.2.2 High Performance Liquid Chromatography  2 2 4 6 7 7 7 8 8  1.3 Objective and Overview of Thesis CHAPTER 2 DEVELOPMENT OF MASS SPECTROMETRIC METHODS FOR ARSENIC  11  SPECIATION 2.1 INTRODUCTION 2.1.1 Scope and Rationale of Work  11 11  2.1.2 Mass Spectrometry 2.1.2.1 Desorption Chemical Ionization 2.1.2.2 Matrix Assisted Laser Desorption Ionization  14  2.2 EXPERIMENTAL 2.2.1 Instrumentation 2.2.1.1 Low and High Resolution DCI MS Instrumentation 2.2.1.2 MALDI-TOF-MS Instrumentation  17  -  V  15 16 17 17 17  2.2.1.3 Hydride Generation Gas Chromatography MS -  -  17  Instrumentation  19  2.2.2 Reagents 2.2.2.1 Reagents used for High Resolution DCI MS  19  2.2.2.2 Reagents used for HG-GC- MS  20  -  -  20  2.2.3 Procedures 2.2.3.1 Low Resolution DCI-MS Procedures  20  2.2.3.2 High Resolution DCI-MS Procedures  20  2.2.3.3 MALDI-TOF-MS Procedures  22  2.2.3.4 HG-GC-MS Procedures  22  2.2 RESULTS AND DISCUSSION  23  2.2.1 DCI MS of Arsenic Compounds of Environmental Interest  23  2.2.2 Accurate Mass Measurements in DCI MS  32  -  -  2.2.2.1 PFK as an Internal Reference Standard for Accurate Mass Measurements in Ammonia DCI-MS  33  2.2.2.2 Fomblin (18/8) as an Internal Reference Standard for Accurate Mass Measurements in Ammonia DCI-MS  36  2.2.3 Analysis of Arsenic Compounds by using MALDI-TOF-MS  41  2.2.4 Structural Characterization of Arsines by using HG-GC-MS  50  2.4 SUMMARY  57  CHAPTER 3 OPTIMIZATION OF ATOMIC ABSORPTION SPECTROMETRIC METHODS FOR THE DETERMINATION OF ARSENIC IN BIOLOGICAL SAMPLES OF MARINE ORIGIN 3.1 INTRODUCTION  59 59  3.1.1 Scope and Rationale of Work  59  3.1.2 Atomic Absorption Spectrometry  60  3.1.2.1 Electrothermal Atomic Absorption Spectrometry and the Determination of Arsenic  61  3.1.2.2 Hydride Generation Atomic Absorption Spectrometry for the Determination of Arsenic  62  3.1.3 Digestion Procedures for the Preparation of Biological Samples for Arsenic Determination  vi  62  3.1.3.1 The use of Microwave Energy for the Digestion of Biological Materials  64 64  3.1.4 Simplex Optimization  67  3.2 EXPERIMENTAL 3.2.1 Instrumentation  67  3.2.2 Reagents  68  3.2.3 Sample Preparation  69  3.2.3.1 Microwave Digestions Prior to ETAAS Analysis 3.2.3.2 Wet Digestion Procedure Prior to HGAAS Analysis  69 69  3.2.4 Procedure for the Simplex Optimization of ETAAS Experimental Conditions  70 72  3.3 RESULTS AND DISCUSSION 3.3.1 Simplex Optimization of Conditions for the Determination of Arsenic in Environmental Samples by using ETAAS  72  3.3.2 Determination of Arsenic in Environmental Samples by using 78  HGAAS 3.3.2.1 Wet Digestion HGAA Determination of Arsenic 3.3.2.2 Microwave Digestion HGAA Determination of  79 80  Arsenic 3.3.3 Comparison of Methods used for the Determination of Arsenic in Environmental Samples  82  3.3.4 Arsenic Concentrations in Californian Mussels Collected from 84  the B.C. Coast  87  3.4 SUMMARY CHAPTER 4 DEVELOPMENT OF HPLC AND MS METHODS FOR THE SEPARATION AND DETERMINATION OF ARSENIC ANIMAL FEED ADDiTIVES AND  88  THEIR METABOLITES  88  4.1 INTRODUCTION 4.1.1 Scope and Rationale of Work  88  4.1.2 Arsenicals used as Animal Feed Additives  90  4.1.3 Micro HPLC used for the Separation of Arsenical Animal Feed -  92  Additives 4.1.4 Interfacing Liquid Chromatography to Mass Spectrometry  vu  94  4.2 EXPERIMENTAL  .96  4.2.1 Instrumentation  96  4.2.1.1 HPLC pumps  96  4.2.1.2 Injectors 4.2.1.3 HPLC Columns  97  4.2.1.4 HPLC Detectors  97  4.2.1.5 LC-MS Interfaces  98  97  4.2.2 Chemicals  99 101  4.2.3 Procedures 4.2.3.1 Fabrication and Packing of Micro HPLC Columns 4.3 RESULTS DISCUSSION  101 103  -  4.3.1 Separation and Determination of Arylarsenicals by using HPLC 104  Thermospray MS 4.3.2 Separation of Arylarsenicals on a Microbore C18 HPLC Column  109  4.3.3 Separation of Arylarsenicals on Micro C18 HPLC Columns  109  4.3.3.1 Effects of Packing Procedure and Particle Size on the Quality of Micro-LC Columns  112  4.3.3.2 Effect of Column Length on the Elution Characteristics of the Arylarsenicals  112  4.3.3.3 Effect of Temperature on the Elution Characteristics of the Arylarsenicals  113  4.3.4 Detectors used for Microscale HPLC 4.3.4.1 ETAAS used as a Detector for Microscale-HPLC  113  4.3.4.2 Liquid Secondary Ion Mass Spectrometry  114 116  4.3.4.3 Direct Liquid Introduction Mass Spectrometry  120  -  4.4 SUMMARY  122  CHAPTER 5 ARSENIC SPECIATION IN MYTILUS CALIFORNIANUS MUSSELS  123 123  5.1 INTRODUCTION 5.1.1 Scope and Rationale of Work  125 126  5.2 EXPERIMENTAL  126  5.2.1 Instrumentation  vu’  5.2.2 Chemicals  .  127 127  5.2.3 Procedures  5.2.3.1 Extraction of Organoarsenic Compounds from Mussels 127 5.2.3.2 Purification Procedures for Arsenic Compounds Present in Mussel Extracts  128  OH 4 5.2.3.3 Isolation of Arsenic Compounds in the NH 129  Fraction 5.2.3.4 Isolation of Arsenic Compounds in the HC1 Fraction  130  5.2.3.5 Detennination of Arsenic in Mussel Shells  131 131  5.3 RESULTS AND DISCUSSION 5.3.1 Purification and Isolation of Arsenic Compounds Extracted  131  from M. califomianus 5.3.2 Identification of an Arsenic Compound Present in the Ammonium Hydroxide Fractions  135  5.3.3 Identification of an Arsenic Compound in the Hydrochloric Acid 138 Fractions 5.3.4 Attempts to Identify an Arsenic Compound in the Water Fractions 5.3.5 Arsenic Speciation in Mussel Shells 5.4 SUMMARY  140 140 142  CHAPTER 6 BIOTRANSFORMANTION OF [ H]-METHYLARSONIC ACID IN A STATIC 3 SEAWATER SYSTEM CONTAINING MYTILUS CALIFORNIANUS  143 143  6.1 INTRODUCTION 6.1.1 Scope and Rationale of Work  143  6.1.2 Methylarsonic Acid in the Marine Environment 6.1.3 Arsenic, accumulation and biotransformation experiments  144  6.2 EXPERIMENTAL  145 147  6.2.1 Instrumentation 6.2.1.1 Liquid Scintillation Counting  147 147  6.2.1.2 High Performance Liquid Chromatography  147  6.2.1.3 Atomic Absorption Spectrometry  147  6.2.2 Chemicals  148  6.2.3 Collection and Storage of Mytilus californianus  148  Ix  H-Labeled Compounds 6.2.4 Procedures for the Speciation of 3 Extracted From Mytilus californianus H In Mussel Tissue and Shells 6.2.5 Determination of Total 3  149 150 152  6.3 RESULTS AND DISCUSSION H-Labeled Compounds Extracted From M. 6.3.1 Speciation of the 3  152  californianus H]-AsB in mussel 3 6.3.1.1 Identification of [ H]-MAsA and [ 3  154  extracts 6.3.1.2 Attempts to identify Peaks #1 and #2 6.3.2 Determination of Total 3 H in Mussel Tissue and Shells  160 160 162  6.4 SUMMARY CHAPTER 7 SUMMARY  164  REFERENCES  168  APPENDIX A Thermospray mass spectra of arylarsenicals APPENDIX B LSIMS mass spectra of arylarsenicals  183 186  APPENDIX C DLI mass spectra of test standards using heated capillary interface.. 189  x  LIST OF TABLES PAGE Some of the major arsenic compounds present in environmental and biological system  3  2.1  DCI filament currents at which arsenic compounds were desorbed  25  2.2  Positive ion accurate mass measurements of arsenic and other compounds in DCI mode, using PFK or Fomblin 18/8 as internal reference standards  35  2.3  HG-GC-MS experimental parameters  52  3.1  Most common digestion methods used for the preparation of biological samples for 63 metal analyses  3.2  Operating conditions for the Hydride Generation Atomic Absorption System  68  3.3  Range and precision of experimental variables examined  71  3.4  Simplex experiments and ETAAS response  74  3.5  Furnace heating program  76  3.6  Arsenic concentration found in NEST oyster tissue  76  3.7  Arsenic concentration of solutions containing organoarsenicals  77  3.8  Recovery of arsenic from DOLT-i and DORM-i  81  3.9  Concentration of arsenic in Californian mussels  85  3.10  Arsenic disthbution in Californian mussel tissue parts  86  4.1  Classification of Microscale HPLC  93  4.2  Summary of micro-column length vs resolution (R) obtained for arsenicals  112  5.1  Weight and arsenic contents of extracted soft tissue, methanol extract, and residue  132  6.1  Summary of chromatographic experiments  158  1.1  XI  LIST OF FIGURES PAGE 1.1 Mechanism for the biomethylation of arsenic, proposed by Challenger. The intermediates in { } are unknown as monomeric species. They can be isolated AsJ in the laboratory, but not from any ) 3 [(CH O AsO) and 2 3 as (CH biological system  5  1.2 Structure of S-adenosylmethionine (SAM)  5  2.1 Desorption probe equipped with a tungsten filament wire, 0.08 mm diameter, 0.45 mm loop diameter, 8 loops  18  2.2 Negretti SS-needle valve (Model 145.207.2P2PK.IXE) system used to form specific gas mixtures  18  2.3 Hydride generation gas chromatography mass spectrometry experimental setup. Dotted lines represent gas lines  19  2.4 Loading positions of sample and internal reference substance (Fomblin) on a Kratos DCI probe filament  21  2.5 DCI mass spectrum of dimethylarsinic acid (0.2 big), ammonia reagent  24  2.6 DCI mass spectrum of arsenobetaine (0.2 pg), ammonia reagent  24  2.7 DCI mass spectrum of tetramethylarsonium iodide (0.2 tg), ammonia reagent  26  2.8 DCI mass spectrum of arsenocholine (0.2 .tg), ammonia reagent  26  2.9 DCI mass spectrum of trimethylarsine oxide (0.2 pig), ammonia reagent  27  2.10 DCI mass spectrum of disodium methylarsonate (0.2 .tg), ammonia reagent  27  2.11 DCI mass spectrum of sodium arsenate (0.2 rig), ammonia reagent  29  2.12 DCI mass spectrum of sodium arsenite (0.2 jig), ammonia reagent  29  2.13 DCI mass spectrum of o-arsanilic acid (0.2 jig), ammonia reagent  30  2.14 DCI mass spectrum of cL-toluenearsonic acid (0.2 jig), ammonia reagent  30  2.15 DCI mass spectrum of triphenylarsine (0.2 jig), ammonia reagent  31  -  -  XII  2.16 PFK mass spectrum obtained under mixed gas CI conditions. The ammonia to methane ratio was monitored at m/z 18, and 17, 29 respectively. The relative intensities and peak areas of these ions are presented above 34 2.17 Positive ion DCI mass spectrum of Fomblin using i-C 12 as the reagent gas. H 4 The three major ion series observed are labelled as A, B, and C 37 2.18 Structural assignment of Fomblin fragment ions (A=69+n166; B=135+n166, and C=185+n166), obtained under positive ion DCI conditions, using i 12 as reagent gas H 4 C  38  2.19 Positive ion DCI mass spectrum of Fomblin using NH 3 as the reagent gas. The three major ion series observed are labelled as A’, B’, and C’ 38 2.20 Positive ion DCI mass spectrum of Fomblin using CH 4 as the reagent gas  39  2.21 MALDI-TOF mass spectrum of arsenobetaine, DHB used as matrix (*:matrix ions)  44  2.22 MALDI-TOF mass spectrum of arsenobetaine, no matrix used  44  2.23 Effect of laser power (%) on the intensity of ions observed in MALDI-TOF mass spectrum of arsenobetaine, m/z 179 is the protonated molecular ion, m/z 135 is the ion formed after loss of CO . Each data point represents the 2 average ion intensities 45 2.24 MALDI-TOF mass spectrum of tetramethylarsonium iodide, DHB used as matrix. (*: matrix ions, DHB)  47  2.25 MALDI-TOF mass spectrum of arsenocholine, DHB used as matrix. (*: matrix ions, DHB)  47  2.26 MALDI-TOF mass spectrum of disodium methylarsonate, DHB used as matrix. (*: matrix ions, DHB)  48  2.27 MALDI-TOF mass spectrum of p-arsanilic acid, DHB used as matrix. (*: matrix ions, DHB)  48  2.28 MALDI-TOF-MS calibration graph obtained from the analysis of MB (m/z 179 corresponds to [M+H]). The molecular ion of DHB (mlz 154) was used as an internal standard  50  2.29 Mass spectra of arsine, methylarsine, and dimethylarsine are presented. A solution containing standard arsenate, methylarsonic acid, and dimethylarsinic acid (50 ng of arsenic for each compound) was used to generate the arsines 53  XIII  2.30 A: Mass spectrum obtained from culture medium inoculated with L methionine-methyl-d and arsenate, average of 100 scans. B: Mass spectrum 3 obtained from culture media inoculated only with arsenate, average of 100 scans  55  2.31 Full scan and Selected Ion Chromatograms (SIC) obtained from the analysis of culture media inoculated with both L-methionine-methyl-d3 and arsenate. a. 56 Full scan mlz 74-115 b. SIC of m/z 112 c. SIC of m/z 109 d. SIC of m/z 93 3.1 A simplex for two variables. A, B: simplex moves in a direction opposite the 1 response surface and reaches optimum worst point; C: simplex “climbs’  66  3.2 Schematic diagram of hydride generation assembly  67  3.3 Wet Digestion apparatus. A: Teflon® cylindrical plugs; B: Teflon® diffusion funnel; C: Teflon® stopper with capillary; D: 250 mL round-bottomed flask, containing sample and acid mixture. Adapted from the design described by ° 17 Bajo et al.  70  3.4 The absorbances obtained from 33 optimization experiments plotted as a function of the four experimental parameters studied. Dotted lines indicate the estimated highest response level at this value of experimental parameter, 75 irrespective of the values of the other three parameters 3.5 Typical calibration plot obtained by ETAAS. Dotted line indicates 95% confidence limits; r: correlation coefficient; b: slope; a: intercept; S: standard deviation  82  3.6 Typical calibration graph obtained by HGAA. Dotted line indicates 95% confidence limits; r: correlation coefficient; b: slope; a: intercept; S: standard deviation  83  4.1 Structures of arsenic compounds used as animal feed additives  89  4.2 Continuous flow LSIMS probe and Cs ion beam  98  4.3 DLI nebulizers; A. diaphragm nebulizer with desolvation chamber, and B. heated capillary nebulizer  100  4.4 Fabrication of the micro-HPLC column  102  4.5 Schematic diagram of the apparatus used for packing micro-HPLC columns  103  4.6 Methods used for the separation and identification of animal feed arsenicals  104  xiv  4.7 A: Three dimensional graph showing the variation of resolution between 3NHPAA and 4-NPAA when the methanol content of the mobile phase and the flow rate are varied. B: UV Q=254 nm)-chromatogram of arylarsenicals (22.5 ng As of each compound) 106 4.8 Thermospray mass spectrum of p-arsanilic acid. Temperature settings: Probe at 129 °C, vaporizer at 185 °C, source block at 223 °C, and jet at 120 °C  108  4.9 Calibration plot obtained by using Thermospray-MS for p-arsanilic acid and 4nitrophenyl arsonic acid. The error bars represent the signal range obtained from three injections of the analyte  108  4.10 UV (=254 nm)-chromatogram of arylarsenicals (2.5 ng As of each compound); Microbore C 18 column, 80 pL.miir 1 flow rate, 85% water (0.1% TFA) and 15% methanol  111  4.11 UV (A=254 nm)-chromatogram of arylarsenicals (0.8 ng As of each compound); micro-HPLC column, 10 pL.min 4 flow rate, 85% water (0.1% TFA) and 15% methanol. A: Compounds separated: p-ASA, 3-NHPAA, and 4-NPAA B: 4-hydroxyphenylarsonic acid added to analyte mixture of A  111  4.12 Effect of temperature on the arylarsenical retention times  113  4.13 ETAAS.-chromatograms of arsenicals, separated on a: A. conventional-LC column, 500 pL fractions collected, B. microbore-LC column, 40 j.tL fractions collected, C. micro-LC column, 5 i.iL fractions collected 115 4.14 LSIMS mass spectra of 3-NHPAA; A. Negative ion detection, and B. Positive ion detection (*: NBA matrix ions) 118 4.15 Background subtracted CF-LSIMS mass spectrum of p-ASA; mobile phase: 89% water (0.1% TFA), 10% methanol, and 1% glycerol (Glyc); flow rate; 5 1 pL.min  119  4.16 DLI-MS of 3-NHPAA, mobile phase consisted of 90% water (0.1%TFA), and 10 % methanol, 20 iL.min 4 flow rate, negative ion detection  121  4.17 DLI-MS of 4-NPAA, mobile phase consisted of 90% water (0.1 %TFA), and 10 % methanol, 20 iL.min’ flow rate, negative ion detection  121  5.1 Proposed mechanism for the biosynthesis of arsenosugars by marine algae  124  5.2 Proposed mechanisms for the biosynthesis of arsenobetaine  126  5.3 Structure of an arsenic containing phospholipid present in brown alga  133  xv  5.4 Sephadex LH-20 ETAA chromatogram; mobile phase: water; 7 mL fractions collected; 20 i.iL of each injected into the ETAA spectrometer; each fraction analyzed in triplicate  134  5.5 Dowex 50Wx8(H+) ETAA chromatogram; mobile phase: 120 mL water, 120 mL 5% ammonium hydroxide, 50 mL water and 200 mL 2M HC1; 7 mL fractions were collected; 20 pL of each fraction was injected into the ETAA spectrometer; each fraction analyzed  134  -  -  5.6 HPLC-ETAAS chromatogram of: A. Synthetic arsenic standards: AsB, As(llI), and DMAsA; 20 pL samples containing 500 ng of each arsenical were injected onto the column, B. arsenic containing material collected off the strong cation exchange column 136 5.7 DCI (ammonia reagent gas) mass spectrum of arsenic containing material, collected off a strong cation exchange resin in the ammonium hydroxide fractions  137  5.8 DCI (ammonia reagent gas) mass spectrum of arsenic containing material, collected off a strong cation exchange resin in the hydrochloric acid fractions... 139 5.9 HG-GC-AAS chromatogram of arsenic containing species, originating from Californian mussel shells  141  6.1 Structures of glycerylphosphorylarsenocholine (6. 1A), and phosphatidy larsenocholine (6. 1B). Both of these compounds were present in yellow-eye mullet following oral administration of arsenocholine  146  6.2 Seawater tanks used for mussel storage and exposure experiment: A: 200 L seawater tank with continually flowing seawater, used for mussel storage; B: 15 L experiment tank containing mussels and [ H]-MAsA; *: aeration lines 3 for bothAandB  148  6.3 Oxygen Combustion Flask (3L); A. platinum basket, B. paper sample wrapper, C. pressure relief in the form of a rubber balloon  151  6.4 Sephadex LH-20 liquid scintillation chromatogram; water mobile phase; 9.5 mL fractions collected; 0.5 mL of each fraction mixed with 5 mL scintillate and 152 counted 6.5 Dowex 50Wx8 (Hj liquid scintillation chromatogram; mobile phase 200 mL water, 200 mL 5% ammonium hydroxide, 40 mL water and 150 mL 2M HC1; 9.5 mL fractions collected; 0.5 mL of each fraction mixed with 5 mL scintillate and counted  xvi  153  6.6 HPLC conditions for A, B, C: Waters Protein Pak DEAE column; mobile phase 5 mM sodium acetate, pH adjusted to 4 with acetic acid; flow rate 1 ml.min; fractions collected every 1 mm; A. HPLC-ETAA chromatogram of standards a: arsenobetaine (250 ng As), b:dimethylarsinic acid (500 ng As) c: methylarsonic acid (500 ng As). B. HPLC-liquid scintifiation chromatogram of Peak #3; 1 mL fraction mixed with 5 mL scintillate and counted. C. HPLC-liquid scintillation chromatogram of Peak #4; 1 mL fraction mixed 155 with 5 mL scintillate and counted ,  6.7 HPLC conditions for A and B: Waters Protein Pak DEAE column; mobile . 4 phase 5 mM ammonium acetate, pH 6.8; flow rate 1 mLmin A. HPLC-ETAA chromatogram of standards a: arsenobetaine (250 ng As), b:dimethylarsinic acid (250 ng As); 0.5 mL fractions collected every 0.5 mm. B. HPLC-liquid scintillation chromatogram of Peak #4; 1 mL fraction mixed with 5 mL scintillate and counted  156  18 reversed phase column; 6.8 HPLC conditions for A, B, C: Bondclone 10 C reagent 5 mM ion-pair water-methanol 95:5; mobile phase . 4 tetrabutylammonium nitrate; flow rate 1 mLmin of standards chromatogram a: arsenobetaine, b: A. HPLC-ETAA dimethylarsinic acid, C: methylarsonic acid; fractions collected every 30 sec. B. HPLC-liquid scintillation chromatogram of Peak #3; 1 mL fraction collected and mixed with 5 mL scintillate and counted. C. HPLC-liquid scintillation chromatogram of Peak #4; 1 mL fraction collected and mixed with 5 mL scintillate and counted  157  H activity within mussel parts, sampled on day 9 6.9 Distribution of 3  162  xvii  LIST OF ABBREVIATIONS  AA  atomic absorption  AAS  atomic absorption spectrometry  AES  atomic emission spectroscopy  As(ffl)  arsenite and/or arsenious acid  As(V)  arsenate and/or arsenic acid  AsB  arsenobetaine  AsC  arsenocholine  BC  British Columbia  CF  continuous flow  CI  chemical ionization collisional induced dissociation  CMS  composite modified simplex  Da  dalton  DCI  desorption chemical ionization  DHB  2,5-dihydroxybenzoic acid  DLI  direct liquid introduction  DMAsA  dimethylarsinic acid and/or dimethylarsinate  DMAsE  dimethylarsinylethanol  El  electron ionization  ES  electrospray  ET  electrothermal  ETAAS  electrothermal atomic absorption spectrometry  FAAS  flame atomic absorption spectrometry  FAB  fast atom bombardment  FD  field desorption  xv”  FID  flame ionization detection  g  gram  GC  gas chromatography  GF  graphite furnace  Glyc  glycerol  GPAC  glycerylarsenocholine  GPC  gel-permeation chromatography  h  hour(s)  HCL  hollow cathode lamp  HG  hydride generation  HGAAS  hydride generation atomic absorption spectrometry  HPLC  high performance liquid chromatography  ICP  inductively coupled plasma  ID  inner diameter  kg  kilogram  L  liter  LOD  limit of detection  LSIMS  liquid secondary ion mass spectrometry  mm  minute  MALDI  matrix assisted laser desorption ionization  MAsA  methylarsonic acid and/or methylarsonate  MBI  moving belt interface  MDA  minimum detectable amount  mm  millimeter  mM  millimolar  MS  mass spectrometry  3-NHPAA  3-nitrophenyl-4-hydroxyphenylarsonic acid  NIST  National Institute of Standards and Technology  NMR  nuclear magnetic resonance  4-NPAA  4-nitrophenylarsonic acid  NRCC  National Research Council of Canada  OD  outer diameter  p-ASA  p-arsanilic acid  p-UPAA  p-ureidophenylarsonic acid  PEG  polyethylene glycol  PFK  perfluorokerosene  ppb  parts per billion  ppm  parts per million  R  resolution  RSD  relative standard deviation  s  second  SAM  S-adenosylmethionine  SIC  selected ion chromatogram  SIM  selected ion monitoring  SRM  standard reference material  tR  retention time  TDFC  thermal desorption filament current  TFA  trifluoroacetic acid  TIC  total ion chromatogram  TLC  thin layer chromatography  TMAs +  tetramethylarsonium ion  TMAsO  trimethylarsine oxide  xx  TOF  time of flight  TSP  thermospray  UV  ultraviolet wavelength microliter  xxi  ACKNOWLEDGMENTS I would like to express my deepest and most sincere gratitude to my research supervisors, Dr. W. R. Cullen and Dr. G. K. Eigendorf, for introducing me to this exciting research and sharing their knowledge and enthusiasm for it. Their guidance, support and encouragement during the course of my studies and research at the UBC are sincerely appreciated. I would also like to thank Dr. A. P. Wade for introducing me to experimental design, in particular simplex optimization, and helping me apply it to my research project. My gratitude is extended to the technical staff of the department and the staff of the mechanical workshop, electronic workshop, and mass spectrometry laboratory for their invaluable assistance on many occasions. I especially would like to thank Lina L. Madilao, Chris A. Emond, and Marshall Lapawa for their valuable suggestions and assistance in performing many of the mass spectrometric experiments, as well as for the friendly environment they created while I worked in the mass spectrometry laboratory. My greatest thanks go to my parents, brother and sister for their continued belief in me and for their support, which by no means was weakened by the geographical distance put between us. Finally, I would like to express my special thanks to my wife Alexandra for her love, patience, and encouragement during my studies, as well as to acknowledge the sacrifices she has made by following me in pursuit of my Ph.D. degree.  xxil  CHAPTER 1  GENERAL INTRODUCTION Arsenic occurs ubiquitously in the atmosphere, hydrosphere, pedosphere, lithosphere, and biosphere of the earth.’ The arsenic content of the earth’s crust is quoted to range from 1.5 to 3 mg.Kg 1 on average, however, this may vary depending on geographical 2 Arsenic is found in many types of mineral deposits, particularly those locations.” containing sulfides and sulfosalts. The principal arsenic containing minerals are: realgar ), orpiment (As S 4 (As ), arsenopyrite (FeAsS), and arsenides of copper, iron, cobalt, 3 S 2 and lead. 3 Arsenic is mobilized into the environment mainly by natural processes such as weathering, volcanic, and biological activity. The environment also receives arsenic as a result of anthropogenic activities such as smelting and burning of fossil fuels, industrial use, and mining activities.’ 4’5 The ratio of natural to anthropogenic atmospheric inputs has been estimated to be 60:40.6 Arsenic is well known for the toxic properties of some of its compounds. Arsenic trioxide (white arsenic, tasteless and colourless powder) was the favoured homicidal 7 through the Middle ages and until the nineteenth century. During the first World poison War, Lewisite gas (C1 AsCH=CHC1), a highly vesicant organoarsenical compound, was 2 used with severe effects. However, arsenic has also been used for medicinal purposes: as early as the fourth century BC a paste of realgar was used as a treatment for ulcers. During the early 1900s, arsenic compounds were found to be effective in the treatment of syphilis and other diseases. 7 Today, there is little use of arsenicals in human medicine. ’ 4 ) which is 3 0 2 Industrial use of arsenic is, however, widespread. 5 Arsenic trioxide (As the usual commercial form of arsenic, is used at the level of 50,000-53,000 tons per year worldwide. Arsenicals are being used as wood preservatives, desiccants and herbicides in  1  cotton production. Even though various health and environmental protection organizations and agencies have been trying to gradually restrict the use of arsenic, the production of high-purity metallic arsenic for gallium arsenide semiconductors has been increasing. In addition, the use of arsenic compounds in animal feeds is still continued. Current scientific interest in regard to the arsenic biogeochemical cycle was initiated at the beginning of the century when high levels of the element were found in marine organisms. This interest is well justified, taking into account the fact that marine organisms, via the food chain, are a major source of arsenic for human beings 8 and that the various arsenic species exhibit different toxicities.  1.1 Arsenic Compounds Present in the Environment 1.1.1 Organoarsenic Compounds Present in Marine Organisms During the 1920s it was established that marine organisms naturally contain high levels of arsenic, relative to the levels present in 9 seawater. ” ° In 1925 Cox reported that within 24 h of a person eating fish, arsenic was detected in the urine at levels previously thought to be associated with chronic arsenic poisoning. 11 The first successful isolation and identification of an organoarsenical originating from a marine animal was reported by Edmonds et at.. 12 This arsenical, present in the tail muscle of the western rock lobster, was characterized by using X-ray analysis as being arsenobetaine (AsB), Table 1.1. Since then, AsB has been confirmed to be the major arsenical present in marine animals such as fish, crustaceans, molluscs, and 1320 Meanwhile, toxicological studies have established the non-toxic echinoderms.L nature of MB. 22 The second arsenical reported to be present in marine animals is ’ 21 arsenocholine (AsC), however conflicting reports concerning its presence have been 2327 made.  Other  arsenic  compounds  2  identified  in  marine  animals  are  29 The major s) and trimethylarsine oxide (TMAsO). 28 tetramethylarsonium ion (TMA arsenic compounds of environmental interest are listed in Table 1.1.  Table 1.1 Some of the major arsenic compounds present in environmental and biological systems. No.  Name  Formula  Abbreviation  1  Arsenite  3 As(OH)  As(ffl)  2  Arsenate  sO4 H A 3  As(V)  3  Methylarsonic acid  sO(OH) 2 A 3 CH  MAsA  4  Dimethylarsinic acid  sO(OH) (CH A 2 ) 3  DMAsA  5  Trimethylarsine oxide  sO (CH A ) 3  TMAsO  6  Tetramethylarsonium ion  s+ (CH A 4 ) 3  TMAs+  7  Arsenocholine  H H H H s-t-C (C A ) 3 C 2 O  AsC  8  Arsenobetaine  H H OO s-i-C 2 (C A ) 3 C  AsB  9  Dimethylarsinylethanol  CH H H H s(O) (C A ) 3 C 2 O  DMAsE  Arsenosugars  R  x  Y  10  s(O)(CH A 2 ) 3  -OH  -OH  11  s(O)(CH A 2 ) 3  -OH  H H)C H PO CH H(O -O H 3 C 2 O  12  s(O)(CH A 2 ) 3  -OH  H 3 -SO  13  s(O)(CH A 2 ) 3  2 -NH  H 3 -SO  14  s(O)(CH A 2 ) 3  -OH  H 3 -OSO  15  s+(CH A ) 3  -OH  H 3 -OSO  3  The ubiquity of dimethylarsinyiribosides (arsenosugars) in marine macroalgae has been well established. 3039 These compounds have also been found in a clam kidney (Tridacna maxima), probably derived from symbiotic unicellular green algae within the clam tissue. 35 Most of the arsenic speciation studies concerning arsenosugars have been carried out on Japanese edible seaweeds. The water soluble arsenicals extracted from brown algae, Ecklonia radiata 38 were identified as the ’ 37 37 and Hizikia fusiforme ’ 36 arsenosugars 10-14 (Table 1.1). The first tetraallcyl arsenic compound to be isolated from a Recent work on the kidney of the giant  marine algae is the trimethylarsonioriboside  clam Tridacna maxima has shown the presence of an arsenic containing nucleoside, in addition to the arsenosugars. ° 4 Attempts have been made to identify lipid-type arsenic (ether-soluble arsenic, chloroform-soluble arsenic) in marine algae. These compounds were identified as 41 ,42 phospholipids containing a dimethylarsinyiribosyl moiety.  1.1.2 Arsenic Biomethylation The biological methylation of arsenic, has been observed to occur in a variety of 43 microorganisms, as well as in higher organisms such as plants, mice, monkeys and man. Challenger was the first to investigate the mechanism of biomethylation in detail.’ In his As) the poisonous gas produced by 3 pioneering work he identified as trimethylarsine (Me molds growing on wall paper coloured with arsenic containing pigments. The mechanism he proposed for the methylation of arsenic (Figure 1.1) involves sequential reduction, and CH donor to 45 + arsenate.’ S oxidative transfer of a methyl group from a 3 adenosylmethionine (SAM) (Figure 1.2) is the likely source of the methyl group. A number of compounds such as thiols and dithiols including cysteine, glutathionine, dithiothreitol, and lipoic acid have been found to be capable of carrying out the reduction 47 ’ 46 via a two electron transfer.  4  AsO 3 H 4  2&  As(OH) 3 {CH } 2  f 3 [CH  > 3 AsO H  ] 3 [CH  AsO. ) 3 (CH  ] 3 [CH  AsO(OH) 3 CH 2  AsO(OH) 2 ) 3 (CH  2&  2&  2e  As(OH)} 2 ) 3 {(CH  As ) 3 > (CH  Figure 1.1 Mechanism for the biomethylation of arsenic, proposed by Challenger. The intermediates in { } are unknown as monomeric species. They can be isolated as As] in the laboratory, but not from any biological system. ) 3 [(CH 0 (CH3ASO)n and 2  0$  Figure 1.2 Structure of S-adenosylmethionine (SAM).  5  1.2 Determination of Arsenic Species Present in Environmental Samples For a complete understanding of the biogeochemical cycle and the toxicological significance of arsenic, it is necessary to know all the chemical forms of the element that are involved. Three general approaches have been employed for the speciation of environmental arsenic compounds. The first approach involves the separation of the arsenic species from a large quantity of sample. After the compounds are purified and isolated, their structures are determined by using mass spectrometric techniques, together with in some cases NMR spectroscopy, X-ray crystallography, and elemental analysis. These procedures allow for unequivocal identification, however they also demand large amounts of sample and much time and effort. In addition there is the possibility that isolation and work up procedures can produce artifacts which do not exist in the original source. The second approach for the identification of arsenic species, involves the use of efficient and mild separation techniques coupled to highly selective and sensitive detection methods. Such methods are suitable for both qualitative and quantitative analysis, but only if appropriate arsenic standard compounds are available. ° However, compounds can be 485 mis-identified if they happen to coelute. The third approach involves the use of similar chromatographic techniques coupled on-line to mass spectrometric systems which are capable of offering fairly high sensitivity, as well as providing a moderate degree of chemical information.  6  1.2.1 Methods used for the Identification of Arsenic Compounds Novel arsenic compounds have been discovered and identified in various marine 3 plants and animals only after they had been completely isolated from their natural matrix.’ 20 Various types of chromatography have been used for this purpose: i.e. gel permeation, ion-exchange, thin layer, and high performance liquid chromatography. A variety of stationary and mobile phases have been used in conjunction with these systems. X-ray diffraction analysis can provide complete structural identification of the arsenical, providing that a large amount of sample is available (usually about 1 mg) in C crystalline form. NMR spectroscopy also provides much structural information, but for 13 NMR studies usually several tens of micrograms of sample are required. Mass spectrometry (MS) can provide structural information adequate for identification purposes with submicrogram down to sub-nanogram amounts of material. The amount required is highly dependent on the MS technique employed. For complete identification, it is also necessary to use high resolution MS.  1.2.2 Trace Speciation of Arsenic 1.2.2.1 Hydride Generation A number of arsenic compounds can form hydrides upon treatment with sodium , however, 3 borohydride under acidic conditions. Arsenite and arsenate both form AsH 5 53 arsenite can be selectively determined by carefully adjusting the pH of the reduction. sH while dimethylarsinic acid (DMAsA) forms 2 A 3 CH Methylarsonic acid (MAsA) forms , 5456 s, upon reduction with sodium borohydride. (CH A ) )AsH, and TMAsO forms 3 3 (CH The hydride generation (HG) technique has been coupled with a variety of spectrometric techniques, particularly atomic  absorption spectrometry (HG-AAS), allowing for  ° 576 extremely low limits of detection for the arsines.  Major improvements in  the  speciation capabilities of the method were made when a gas chromatograph was  7  incorporated between the hydride generator and the detector. 62 The use of a mass ’ 61 spectrometric detector allowed for additional structural information regarding the arsines 63 produced. A major limitation associated with the technique is its inability to detect some environmentally important arsenicals such as AsB, TMAs+, AsC, and arsenosugars, because these compounds do not form hydrides upon treatment with sodium borohydride. However, the problem can be eliminated by using an appropriate digestion procedure prior 64 and on-line UV to hydride generation. Recently on-line microwave digestion 65 procedures have been introduced to extend the usefulness of hydride decomposition generation methods.  1.2.2.2 High Performance Liquid Chromatography HPLC is a very efficient separation technique for arsenic compounds. Ionic arsenic species can be easily separated by using anion or cation exchange chromatography, 6668 as well as on reverse phase ion pair HPLC. ° Appropriate counter ions, such as the 687 tetraalkylammonium cation or heptanesulfonate anion are added to the mobile phase. Gelpermeation HPLC has been used for the separation of arsenosugars. 69 ’ 27 The use of HPLC coupled to highly selective and sensitive detectors has lead to significant advances in element speciation research. The most common detectors used are: flame atomic absorption spectrometry (FAAS), 72 electrothermal atomic absorption ’ 71 spectrometry ’ 70 (ETAAS), 7 3 atomic emission spectrometry (AES), and plasma MS, 4 particularly inductively coupled plasma mass spectrometry (ICP-MS). 4850  1.3 OBJECTIVE AND OVERVIEW OF THESIS The main objective of this work was to develop analytical methods for the speciation of arsenic at trace levels. Applications of these methods were made on samples of  8  environmental interest, in order to advance our understanding of the arsenic cycle in the environment. Chapter 2 describes the development and evaluation of a variety of mass spectrometric techniques such as Desorption Chemical Ionization (DCI) Assisted Laser Desorption Ionization (MALDI) -Time of Flight (TOF)  -  -  MS, Matrix  MS. and HG-GC  MS, for use in the structural identification of arsenic compounds of environmental interest. The advantages of DCI-MS for analyzing this type of material were explored. For conclusive identification of unknown arsenic compounds, a high resolution accurate mass measurement DCI method for positive ions, was developed. A relatively new mass spectrometric method, MALDI-TOF-MS was also evaluated for its potential to provide improved sensitivity over other MS methods used for the analysis of arsenicals. Finally, a HG-GC system was interfaced to a mass spectrometer and was used to identify deuterium labeled arsenicals. Chapter 3 outlines the optimization of AAS methods, that are used for the accurate determination of arsenic in samples of marine origin. A Simplex optimization procedure was used in conjunction with a standard reference material (SRM) of marine origin in order to efficiently delineate the optimum experimental conditions for the analysis of arsenic by using ETAAS. Four experimental variables were considered: ashing temperature, atomization temperature, modifier concentration, and atomization ramping time. Microwave dissolution techniques were also evaluated for their ability to digest marine samples containing arsenic.  Chapter 4 describes the fabrication and packing of micro-HPLC columns and their interfacing to mass spectrometers. The primary concern of this work was to develop analytical micro methods capable of separating and determining organoarsenicals used as animal feed additives. A number of analytical systems were evaluated in this study.  9  In Chapter 5, the speciation of arsenic in Californian mussels (Mytilus californianus) is described. DCI-MS, both low and high resolution (described in Chapter 2) was used to characterize arsenicals that were obtained from purified extracts of mussel tissue. In Chapter 6, investigations concerning the biotransformation of an arsenic compound in the marine environment are described. This particular study is concerned with the uptake and biotransformation of 3 H]-MAsA) in a 3 H-labeled methylarsonic acid ({ static seawater system containing Californian mussels, Mytilus californianus.  10  CHAPTER 2  DEVELOPMENT OF MASS SPECTROMETRIC METHODS FOR ARSENIC SPECIATION  2.1 INTRODUCTION  2.1.1 Scope and Rationale of Work A brief account of arsenic compounds present in the marine and terrestrial environment has been given in Chapter One of this Thesis. A great amount of work has gone into the development of analytical methods capable of identifjing these compounds. To a large extent this work has been successful and has advanced our understanding of the arsenic 3 However, these investigations need to be continued in order J 1 cycle in the environment. to complete our understanding of the arsenic cycle, and also to assess toxicity on the basis of the arsenic species present rather than just on the total arsenic content. This is of particular interest, because the toxicity of arsenic greatly depends on the form in which the element is present. Arsenic speciation continuous to be a challenge for the analytical chemist. The work presented in this chapter contributes to the improvement of analytical techniques used for the speciation of trace elements, and it is hoped that this effort will allow for further advancements in arsenic speciation research. A variety of mass spectrometric techniques and procedures were developed and evaluated for their use in the speciation of environmental arsenic compounds.  11  Elemental speciation is defined as the identification and quantitation of all the physico chemical forms of an element present in the 76 sample. 7 ’ 7 These forms include gaseous compounds, solid forms or phases and dissolved forms, depending on the nature of the sample. Also, the chemical form of an element refers to both the oxidation state, and the type and number of substituents. Mass spectrometry (MS) has proven to be a very powerful tool for the identification of arsenic compounds. The most informative mass spectra have been obtained by using field desorption (FD), 237879 fast atom bombardment (FAB), 29367880 atmospheric pressure chemical ionization, ’ 81 8 electrospray 8 ’ 2 and thermospray MS. 74 The main advantage in utilizing these soft ionization techniques is the detection of the compound’s molecular ion M  +.  .  or, more usually, a simple molecular adduct [M+X]  + ,  where X= H, Na, K, etc. The  even-electron protonated molecule [M+Hf is the most common ion containing the unfragmented molecule formed by these soft ionization techniques. However, the disadvantage in many cases, is the absence of characteristic fragment ions, particularly at lower masses. These features are reversed in spectra obtained by using electron impact (El) MS, which yields mass spectra in which molecular ions M+ are rare and fragmentation 48 In order to overcome these deficiencies other MS techniques, such as extensive. Desorption Chemical Ionization (DCI), Matrix Assisted Laser Desorption Ionization of Flight (MALDI-TOF), and Hydride Generation  -  Gas Chromatography  -  -  Time Mass  Spectrometry (HG-GC-MS) have been studied in this Thesis. These techniques provide both molecular and structurally important fragment ions and have improved sensitivities for detection of analytes. Frequently the high detection limits of commonly used methods do not allow the identification of minor arsenic containing species found in environmental samples. ,29,36,7880 23  12  DCI-MS, is a technique that is easily accessible in most laboratories and capable of providing both molecular ions as well as characteristic fragment ions, for the type of compounds investigated here. Furthermore a procedure has been developed which allows for high resolution accurate mass measurement of positive ions in DCI mode. 83 Data obtained by using this method allow for calculation of the elemental composition of the compounds investigated. Mass deficient reference standards [i.e. perfluorokerosene (PFK)J, normally used for calibration purposes in mass spectrometry, do not provide adequate mass spectra under ammonia chemical ionization conditions. In order to overcome this problem a procedure employing a mixture of ammonia and methane as reagent gas has been developed. In high resolution accurate mass measurement experiments, this gas mixture allows for the simultaneous detection of mass spectra of PFK adequate for calibration purposes, and of spectra containing molecular weight information for the analyte. For positive ion accurate mass measurements of higher masses (up to m/z=2300), perfiuoropropylene oxide (Fomblin 18/8) was used successfully as a reference standard under ammonia, methane, and iso-butane DCI conditions. Extremely high sensitivity has been reported for MALDI-TOF-MS, and this stimulated us to evaluate the technique for the analysis of organoarsenicals of environmental interest. A number of matrices, normally used in MALDI, were evaluated for their potential to promote desorption and ionization of organoarsenicals. Other factors which affect the quality of the mass spectra were also investigated, these include: sample and matrix concentration, as well as the effect of laser power on compound fragmentation. Attempts were also made to aquire quantitative data. Finally, a HG-GC-MS system, capable of identifying arsenic hydrides, was developed and used to identify deuterium labelled arsenicals. This analytical technique was employed in  13  conjunction with a feeding experiment in order to establish the methylating agent responsible for arsenic methylation in biological systems.  2.1.2 Mass Spectrometry Mass spectrometry had its beginning in the work of J. 3. Thomson (19 10-20) and subsequently of F. W. Aston (1920s), concerning the behaviour of ions in magnetic 85 This enabled the determination of naturally occurring isotopic abundances of ’ 84 fields. many elements. Since then, the evolution of MS has been extraordinary. MS is now being used in a wide variety of analytical applications and is perhaps the most frequently encountered of all analytical techniques. This can be viewed as a consequence of the technique’s ability to provide both qualitative and quantitative information on inorganic and organic analytes in complex mixtures, information concerning the structures of complex molecules, the structure and composition of solid surfaces, and isotopic ratios of elements in samples. In addition, for many purposes MS has not only the required very wide dynamic range but is also the most sensitive analytical method available today. The operating principle of the method is basically very simple. First, the sample components are converted into ions (positive and/or negative), which are then separated on the basis of their mass-to-charge ratios, and finally a suitable detector is used to convert the beam of ions into an electrical or optical signal. The most critical aspect of MS is the ionization process. A great number of different ionization methods have been developed in order to accommodate the special requirements set forward by the variety of different compound classes. In selecting the most appropriate ionization technique for a particular sample, a number of factors must be taken into consideration: the physical and chemical properties of the sample, the ionization efficiency of the ion source, and the kind of information required from the analysis.  14  2.1.2.1 Desorption Chemical Ionization  DCI-MS was first introduced in 1973.86 The technique had an immediate impact on the analysis of involatile biological substances, because it was more convenient and easy to use than field desorption. The technique has been used to provide structural information on 87 highly polar and nonvolatile compounds, such as oligopeptides, 86 flavonoid glycosides, 88 oligosaccharides, fullerenes, 89 alkaloids, 90 macrocycles with molecular weights up to 4400 Da, 91  and various  synthetic polymers (polystyrenes,  polyethylene glycols,  polysiloxanes, and polynorbornene), 92 as well as on a whole variety of other compounds. Some of the most advantageous features of this technique are the following: very fast analysis time, good sensitivity, both molecular weight as well as structural information are obtained, and the method is easy to use and free of chemical noise. The main disadvantage of the technique has been the relatively short duration of the signal (2-5 sec), due to fast desorption times. In the past this feature resulted in poor reproducibility of the mass 93 The introduction of fast scanning mass analyzers and data systems has allowed spectra. rapid data acquisition over wide mass ranges, thus solving this problem. During the DCI process, the sample is desorbed from a wire as a result of rapid heating. 93a,b This process takes place inside an ionization source in an atmosphere of reagent gas. The gaseous sample molecules are ionized by collision with reagent gas ions produced by an electron beam. The CI source is constructed in order to maintain a reagent gas pressure of  1 torr inside the source and also maintain the analyzer region of the  spectrometer at a pressure below I 0 torr. A variety of different reagent gases have been 9495 In the positive ion mode the most widely used gases have been those used in DCI. which yield Brönsted acids, some of which are: ammonia, methane, propane, and iso-butane. The major reaction mode of these gases is proton transfer to the analyte molecule. In negative ion mode gases capable of forming BrOnsted bases are used successfully.  15  2.1.2.2 Matrix Assisted Laser Desorption Ionization  MALDI-TOF-MS was first introduced in 1988, by two independently working 97 and has since 96 and Tanaka and co-workers, groups, those of Karas and Hillenkamp, 99 98 synthetic polymers, been used with great success for the analysis of biopolymers, s. The virtually unlimited mass 102 des, and carbohydrate 10 oligonucleoti 100 ’ peptides, range, and very high sensitivity are the two main features which have made this technique so 103 The latter feature, the high sensitivity, attractive for various analytical applications. stimulated us to evaluate the technique for the analysis of organoarsenicals of environmental interest. Some of the disadvantages of the technique are the poor resolution achieved (reflecting the velocity of ejection of the ablated material), the unpredictability of the matrix selection for different compound types, and its poor reproducibility which makes quantitation marginal at best. The ionization process that is involved in MALDI is still not completely understood. The matrix used consists of an organic compound of low molecular weight (< 200 Da), capable of absorbing radiation in the 266-532 nm range. The analyte and matrix molecules 2 laser emitting at 337 tim) are intermixed in a crystallized lattice. Laser radiation (typically N is used to excite this lattice. The matrix is thought to absorb the energy and transfer it as protonation to the analyte molecules. In some cases when the analyte contains certain chromophores capable of absorbing the laser radiation a matrix may not be required for obtaining ionization. The ions are formed in an electric field from which they are accelerated into a field-free drift region. This region is the TOF mass analyzer. Separation of the ions is based on the fact that their velocities in this region depend on their mass-to-charge ratio. 103 Thus small ions will reach the detector faster than larger ions of the same charge.  16  2.2 EXPERiMENTAL 2.2.1 Instrumentation 2.2.1.1 Low and High Resolution DCI MS Instrumentation -  A quadrupole mass  spectrometer (Delsi/Nermag RiO-lOB) consisting of a desorption probe, an El/Cl source, a quadrupole mass filter, and a channeltron detector, was used for the low resolution DCI analysis of the arsenic compounds. The desorption probe (Figure 2.1) carries a coiled tungsten filament wire, to which a heating current is supplied in order to thermally desorb the sample after the probe has been inserted into the evacuated ion source chamber. High resolution (R=10,000, 10% valley definition) DCI measurements were carried out on a double focusing mass spectrometer (Kratos Concept II H / Mach3 Data system). A Negretti stainless steel-Needle Valve system was constructed and used to supply gas mixtures of suitable ratios (Figure 2.2). A Kratos single reagent gas inlet system was employed to introduce the gases into the ion source and to control the ion source pressure.  2.2.1.2 MALDI-TOF-MS Instrumentation Two MALDI-TOF mass spectrometers were used for these experiments, one was a Kompact MALDI (Kratos Analytical) and the other was a VG TOF (Fisons). Both were 2 lasers emitting at 337 nm, 3 ns pulses. 100% laser power is equivalent to equipped with N  200 .iJ. The Kompact MALDI was capable of operating in the reflectron mode.  2.2.1.3 Hydride Generation Gas Chromatography MS Instrumentation -  -  A quadrupole mass spectrometer (Delsi!Nermag Rb-bC) interfaced to a gas chromatograph (Varian, Vista 6000) was used in this study. A Porapak-PS (Waters, Milford, MA 01757) packed GC column was used for the arsine separations. Data acquisition and processing were performed by using a PC based data system (Teknivent, Vector 2) interfaced to the mass spectrometer. The hydride generation apparatus consisted  17  of a peristaltic pump, a gas-liquid separator, a moisture trap, and a hydride trap. The experimental setup of the system is shown in Figure 2.3. Desozption Probe  Thnsten Filament  Figure 2.1 Desorption probe equipped with a tungsten filament wire, 0.08 mm diameter, 0.45 mm loop diameter, 8 loops.  Kratos Single Reagent Gas Inlet  H’ 4 i-C  >  4 CH  3 NH  Negretti SS-Needle Valves  Figure 2.2 Negretti SS-needle valve (Model 145.207.2P2PK.1XE) system used to form specific gas mixtures.  18  Peristaltic Pump  Gas/Liquid Separator  Dry-Ice/Acetone Water Trap  Figure 2.3 Hydride generation  Gas Chromatograph Mass Spec. 1. Arsine trap: Liquid Nitrogen 2. Releasing Arsines Water bath -  -  gas chromatography  -  mass spectrometry experimenta]  setup. Dotted lines represent gas lines.  2.2.2 Reagents Arsenobetaine (AsB), 104 arsenocholine (AsC), 105 tetramethylarsonium iodide 74 and trimethylarsine oxide (TMAsO) çrr&s+I-) 106 were synthesized by using literature methods, all other chemicals were commercially available. Deionized water was used for all applications. Glass and plasticware were cleaned by soaking them overnight in 2% Extran (BDH Inc.) solution, followed by a water rinse, a soak in dilute (2 M) hydrochloric acid and fmally a water rinse.  2.2.2.1 Reagents used for High Resolution DCI MS -  -  Three reagent gases were used: i-butane, methane, and anhydrous ammonia (Linde, Union Carbide Canada Ltd.). Solutions containing various concentrations of Fomblin 18/8 (Edwards, High Vacuum Canada Ltd. Oakville Ont.) were made up in PF-5060 (3M, St.  19  Paul, MN 55144-1000, USA). This solvent was preferred because it has a zero ozone depleting potential.  2.2.2.2 Reagents used for HG-GC- MS Arsenic standards were freshly prepared by performing serial dilutions of stock solutions (1000 ppm of elemental arsenic) of the following compounds: sodium arsenate, 2 (J.T. Baker 7H H 2 Na . 4 0 AsO (MC&B, Norwood Ohio 45212); sodium arsenite, NaAsO sO a (Alfa 6H A CH N 3 . 2 Chem. Co., Phillipsburg); disodium monomethylarsonate, 0 AsO(OH) (Sigma Chem. Co., 2 ) 3 Division, Danvers, MA 01923); dimethylarsinic acid, (CH 4 were freshly prepared St. Louis, MO 63178). Solutions of 1 M HC1 and 2% (w/v) NaBH on a daily basis.  2.2.3 Procedures 2.2.3.1 Low Resolution DCI-MS Procedures The solid sample was dissolved in an appropriate solvent resulting in a 100 ppm arsenic solution. A syringe was used to apply 2 iL of the solution (200 ng of As) to the DCI filament. The solvent (usually MeOH) was allowed to evaporate prior to insertion of the probe into the ion source. After introduction of the ammonia reagent gas a heating current was applied to the desorption filament at a rate of 10 mA.s’. The sample was thermally desorbed, and ionized by reagent gas ions that were formed via collision with an electron beam.  2.2.3.2 High Resolution DCI-MS Procedures In order to use PFK as a reference standard for accurate mass measurements under mixed ammonia-methane DCI conditions, it was first necessary to optimize the ratio of ammonia to methane. This was accomplished by adjusting the Negretti SS-Needle valve  20  i’ 5 system (Figure 2.2) to obtain the desired intensity of the reagent gas ions [CH ] (m/z=29). Once this was done, PFK was 5 H 2 ] (mlz=18), and [C 4 (m/z=17), [NH introduced into the ion source via a heated inlet system. Subsequently the sample was introduced into the ion source by means of a DCI probe equipped with a filament. The filament current was ramped from 30  -  400 mA, in order to thermally desorb the sample.  The filament current and heating rate were determined by the volatility and/or temperature sensitivity of the analyte. This high resolution DCI-MS procedure allows for accurate mass measurements of positive ions in peak matching, narrow scan or full scan mode. A slightly modified procedure was followed for accurate mass determinations of positive ions by using Fomblin as the internal reference standard. Only one reagent gas was used and, because it is very difficult to introduce Fomblin into the ion source via a heated inlet system, the sample and Fomblin were introduced simultaneously by using the DCI probe. The sample solution was placed on the filament coil, while Fomblin, dissolved in PF-5060, was placed on the straight section of the filament (Figure 2.4).  Sample  Ceramic Insulator  Figure 2.4 Loading positions of sample and internal reference substance (Fomblin) on a Kratos DCI probe filament.  21  2.2.3.3 MALDI-TOF-MS Procedures Solutions (1.5 jiL) containing matrix [2,5-dihydroxybenzoic acid (DHB)] and analyte (typically 5 jig  10 ng), were applied on the sample slide (target) and allowed to air dry  prior to their introduction into the mass spectrometer. After an initial “survey” scan across the sample spot, the laser beam was focused on the solid phase matrix-analyte mixture in an area were the signal to chemical noise ratio was optimized (often referred to as the “sweat spot”). The mass spectra reported in this thesis are average spectra obtained from 50 successive laser shots. Internal calibrations were used in this work. Ions at mlz 23 ([Na]j and m/z 177 ({DHB+Na]j provided suitable reference ions for calibration.  2.2.3.4 HG-GC-MS Procedures For the HG-GC-MS analysis, a semi-continuous method was developed based on modifications of the procedures described by Kaise et alJ 07 and Reimer. 61 A peristaltic pump was used to mix the sample (3 mL) with 1 M HC1 and 2% (wlv) NaBH 4 solutions (Figure 2.3). Arsines were generated and swept through a gas-liquid separator by means of a  stream of helium gas and the evolved hydrogen gas, through a moisture trap (Teflon® Utube, 30 cm length x 0.8 cm ID) which was cooled by a dry ice-acetone slurry, and subsequently condensed in a hydride trap (Teflon® U-tube, 30 cm length x 0.4 cm ID) which was immersed in liquid nitrogen (-196°C). After the hydride generation reaction was completed the peristaltic pump was stopped and the hydride trap was warmed by using a water bath (70°C) to volatilize the arsines which were then carried to a Porapak-PS GC column (80-100 mesh, 50 cm length x 0.4 cm ID, silanized with Silyl-8 [Chromatographic Specialties] column conditioner as described by Reimer ). By using a gas chromatograph 61 with a pre-set temperature program, the volatile arsines were separated and detected by using a mass spectrometer.  22  2.2 RESULTS AND DISCUSSION  2.2.1 DCI MS of Arsenic Compounds of Environmental Interest -  In this work, the analysis of a number of arsenic compounds was studied by using DCI-MS. The mass spectra obtained are presented and their features are discussed. The spectrum of DMAs (Figure 2.5) shows the protonated molecule at mlz 139, and also a very characteristic fragmentation pattern. These fragment ions provide additional structural information and also allow for compound identification based on four characteristic signals. This pattern is absent from spectra obtained by using thermospray 74 and atmospheric pressure chemical ionization, 81 since both of these techniques provide spectra which exhibit the even-electron protonated molecule [M+HfF of DMAs and no fragments at lower masses. Thermal desorption of DMAs occurred when the DCI filament current was at 170 mA (Thermal Desorption Filament Current, TDFC= 170 mA). The TDFC values of the compounds studied are listed in Table 2.1. The spectrum of arsenobetaine (Figure 2.6) indicates a protonated molecule [M+H]+ as well as a fragmentation pattern that again allows positive identification of this compound. 74 Other methods such as FAB, 79 and thermospray ’ 78 80 electrospray, 81 field desorption provide very informative spectra containing protonated molecular ions, and dimeric species, but fragmentation in some cases is observed to a much lesser extent. Published electron 48 spectra do not show any molecular ion information. ionization Tetramethylarsonium iodide exhibits a base peak at m/z 135, corresponding to the tetramethylarsonium ion (Figure 2.7), also present are fragment ions that are similar to those obtained for AsB.  23  r  -________  Dimethylarsinic acid 58  <  25..<  OH  243 259  91 186 ,  .  76  Hsc_rto  4 (2M..OH.01  96  118  136  158  178198  218  238  258  278298  Figure 2.5 DCI mass spectrum of dimethylarsinic acid (0.2 pg), ammonia reagent.  +  135  H 3 00 C—Ai-CH 3 H C 2  sH [cH A 3  92  75,  3  Arsenobetaine 56/  179 Z5? 185 [As] I’  40  60  88  168  126  140  i6O  188  280  228  248  •  268  Figure 2.6 DCI mass spectrum of arsenobetaine (0.2 pg), ammonia reagent.  24  MIZ  MfZ  Table 2.1 DCI filament currents at which arsenic compounds were desorbed Thermal desorption filament current (TDFC) Arsenic Compound (mA) Arsenobetaine Trimethylarsine oxide  210  Arsenocholine Tetramethylarsonium iodide  200  Dimethylarsinic acid Disodium methylarsonate  170  100 200 250 270 320  Sodium arsenite Disodium arsenate Triphenylarsine  0  ] plus other [M-H O Arsenocholine exhibits an ion at m/z 147 corresponding to 2 signals that provide additional structural information (Figure 2.8). A low intensity molecular ion peak is observed at m/z 165. Other methods result in molecular ion peaks of a much higher intensity. In DCI-MS the presence of hydroxyl groups seem to mitigate against the observation of intense molecular ions, possibly due to reduced volatility. The base peak of the spectrum of trimethylarsine oxide corresponds to [M+HJ+ at m/z 137 (Figure 2.9). With the exception of triphenylarsine, the TDFC of 100 mA used here is considerably lower than the values recorded for the other organoarsenicals. MMAs, (Na Because of the relatively low volatility of the disodium methylarsonate 2 MW= 184), a high TDFC (250 mA) was required in order to obtain its spectrum (Figure 2.10). This resulted in an elevated temperature which is believed to be responsible for the formation of the arsenic dimers, trimers, and tetramers, present at masses m/z 150, 225, and 300 respectively. Additional structural information can be obtained from the peaks at m/z 137, 121, 106 and 92, as indicated in the spectrum.  25  i8øx  s] E(cH A 4 ) 3 135  s+H] 121 A ) 3 [(CH  H3c—cu3  Tetramethylarsonium Iodide 58x  F  92 185 [Asj, ,  48  68  •1  88  I  h.j. 1  188  •  ?___.t  128  L 188  168  148  288  228  Figure 2.7 DCI mass spectrum of tetramethylarsonium iodide (0.2  1884  s+H]+ I(CH A ) 3  44  248  268  wz  gig), ammonia reagent.  I  x  2  —>  3 CH H3c_cH2cH2oH 3 ca Arsenocholine  ]: [M-H o 2 147  58x  133: sH 2 A 3 [cH ] 0  Z5z  i65  58 Figure  78  98  118  138  158  178  198  218  236  258  278  2.8 DCI mass spectrum of arsenocholine (0.2 ag), ammonia reagent.  26  wz  18B  137 [M+H] C —=c 3 fl  75x Trimethylarsine ox + +  Ocr  25:4  k 1 il 188  486888  128  -  148  188  168  288  228  246  268  WL  Figure 2.9 DCI mass spectrum of trimethylarsine oxide (0.2 rig), ammonia reagent.  75z  188z 75y.  0  [As ] 4 388  II  14.0— A,—ONa  58x  013  25x  Disodium methylarsonate I  296  J.  3i8 338358 378 398 418 438 458 478 496  516  wz  Figure 2.10 DCI mass spectrum of disodium methylarsonate (0.2 Mg), ammonia reagent.  27  Sodium arsenate and sodium arsenite both exhibit the same DCI mass spectra, Figures 2.11 and 2.12 respectively, from which it is clear that DCI-MS is not able to distinguish between these two inorganic arsenic species. The high current of the DCI filament (TDFC arsenate=320 mA, TDFC-arsenite=270 mA) causes their pyrolysis and is probably responsible for formation of polymers. The presence of arsenate and arsenite might be indicated by the appearance of these polymeric species. Similar arsenic dimer, trimer, and tetramer formation has been reported to occur in AsH 3 gas cells in the presence of heated tungsten filaments. 108 o  -  Arsanilic (Figure 2.13) and p  -  arsanilic acid cannot be distinguished by this  technique, since they produce identical spectra. The base peak for both appears at mlz 181 Of, at mlz 94 we observe H 2 [M-2H 6 N 2 [H + 5 Hf, C and at mlz 91 [AsOf. Arsenic dimers, trimers, and tetramers are observed at m/z 150, 225 and 300 respectively. -  Toluenearsonic acid can also be identified by DCI-MS. A molecular ion at m/z 216  (Figure 2.14) is observed, and in addition there are signals at mlz 198 due to [M-H Of and 2 at m/z 91 due to 2 C 5 H 6 [C f H The peaks appearing at m!z 257, 274, 306, 349, 365, and ’. 396 can be accounted for in the same manner as for the arsanilic acids. Triphenylarsine is desorbed from the DCI filament without applying a current. The protonated molecule [M+Hj+ forms the base peak in the spectrum (Figure 2.15), again the fragmentation pattern provides sufficient information to characterize this compound. A satisfactory full scan spectrum of these organoarsenicals is easily obtained with 100 ng arsenic on the DCI filament. The ability of DCI-MS to detect organoarsenicals at such levels makes this technique eminently suitable for the investigation of environmental samples. To further reduce the detection levels one could utilize selected ion monitoring (SIM), which generally improves sensitivity by one to two orders of magnitude. We have used the full scan DCI mode to identify AsB along with TMAs in mussel (Mytilus  28  californianus) extracts which had been purified by conventional liquid chromatographic techniques (Chapter 5, of this Thesis).  ‘88 75 LM1  158  i,or 91 i’  188  18  r  68  h  i  88  1 3 IA3  225 128  r  i-’ii  ,  108  148  168  188  288  228  248  .  ..•  1.1•  268  ] 4 (As  75% 58z Disodium Arsenate  25 298  318  338  358  378 398 410  438  458 478  498  518  MJZ  Figure 2.11 DCI mass spectrum of sodium arsenate (0.2 pg), ammonia reagent. 188z 75% [AsT  58z 25%  I  188d  ] 2 [As  75  158  1 3 [As  91. 18  68  88  188  --  128  168  148  188  288  228  248  268  ] 4 lAS  75 58%  Sodium Arsenite  .  25 298  318338358378398418438458  478  490  518  wz  Figure 2.12 DCI mass spectrum of sodium arsenite (0.2 pg), ammonia reagent.  29  [M-2H O 2 ]  8z 75d’  25  5a 25  Figure 2.13 DCI mass spectrum of o-arsanilic acid (0.2 pg), ammonia reagent.  2 H 6 [C J sCH  188w. 75,  25,  a-Toluenearsonic acid 290  318 338 358 378 398 410  430 458 478  498 SIR  WZ  Figure 2.14 DCI mass spectrum of -to1uenearsonic acid (0.2 pg), ammonia reagent.  30  Attempts were also made to analyze the above arsenicals in the negative ion DCI mode. In most cases, fragments produced did not provide structural information about the compound examined. AsB however exhibited a high intensity ion at mlz 177 ([M-H1  188  i.  152  As.2H] 2 (Ph  25  1.69  51  -r 188  88 188 68 387 [M+Hff  18  75  •1  .-  128  148  168  188  288  •  ASJ+ 2 TLPh  228  248  268  58> 25 Triphenylarsine .  298  318  336  358 378  398  418 438  458 478 498  516  Figure 2.15 DCI mass spectrum of triphenylarsine (0.2 jig), ammonia reagent.  31  wz  2.2.2 Accurate Mass Measurements in DCI MS -  Accurate mass measurements, of both fragment ions and of ions containing the unfragmented molecule (M+., [M+H]+, etc.), are of particular importance for the identification of unknown arsenicals via the information thus provided on elemental compositions of the ions. In suitable cases, generally at lower molecular masses, the elemental composition can thus be specified unambiguously, while in less favourable cases the composition can be limited to a few possibilities. Thus, in order to increase the utility of the DCI mass measurements we investigated the possibility of making high resolution accurate mass measurements under DCI ammonia conditions. This procedure requires the presence of an internal reference compound together with the analyte. However, the problem associated with this requirement is that under ammonia DCI (or CI) conditions none of the normally used mass deficient reference standards,  such as, PFK,  heptacosafiuorotri-N-butylamine, or tris(perfluoroheptyl)-S-triazine, give adequate mass spectra. Nevertheless, PFK has been used successfully with methane in chemical ionization 109 and DCI modes. PFK has also been used as an internal standard for accurate mass (Cl) measurements under alternating El and ammonia CI conditions.’ ° In the latter procedure 1 the analyte is introduced into the ion source under ammonia CI conditions and several scans are acquired. The acquisition is suspended while the ammonia reagent gas is pumped out, PFK is then introduced into the ion source under El conditions. Substances other than mass deficient ones have also been used as calibration references. Ballantine et. aL  reported the use of mixtures of long-chain fatty acid methyl  esters, long-chain alcohol stearates, and carbowax 600 as standards for peak matching under ammonia CI conditions. Because these compounds are not mass deficient, problems can arise in the measurement of unknown ions that have the same nominal mass as the reference standard, unless high enough resolution is employed. Commercially available polyethylene glycols (PEGs) and their mixtures have also been shown to serve as reference compounds  32  for full scan accurate mass calibration under ammonia CI conditions.’ 12 These compounds are also mass sufficient and precautions must be taken in order to avoid confusion. Furthennore, several mixtures must be prepared to cover a wide mass range. The lack of appropriate calibration substances for use in accurate mass measurements of positive ions produced under ammonia DCI conditions has prompted us to investigate the suitability of a number of substances for such a purpose. In this work we have demonstrated the successful use of PFK as a reference standard under CI conditions when using a gas mixture of methane and ammonia. This procedure allows accurate mass measurements of positive ions in the full scan, narrow scan, and peak matching modes for a variety of compounds. Fomblin 18/8 was also established as a reference standard for DCI-MS in combination with either iso-butane, methane, or ammonia as the chemical reagent gas. Routine accurate mass measurement of compounds, with molecular weights up to 2300 Da, can be made in positive ion DCI mode.  2.2.2.1 PFK as an Internal Reference Standard for Accurate Mass Measurements in Ammonia DCI-MS As mentioned above, under CI conditions PFK yields adequate mass spectra with 109 Since it is desirable to analyze methane as the ionization agent, but not with ammonia. arsenic and other compounds under ammonia CI conditions, we investigated the possibility of analyzing PFK in a gas mixture of ammonia and methane. By using defined mixtures of the two gases it is possible to simultaneously obtain mass spectra of PFK adequate for calibration purposes, and satisfactory analyte spectra. The mass spectrum of PFK in a , along with the required reagent gas ion ratios, is presented in 4 mixture of NH 3 and CH 109 The PFK Figure 2.16. Detailed mass assignments of PFK have been reported previously. spectrum obtained under mixed gas CI conditions is similar to the one obtained under El conditions, with the exception that the relative intensities of the PFK ions above mlz 100 are  33  much higher under CI conditions. As expected, an increase in the ammonia to methane ratio resulted in suppression of the PFK ion abundances. Accurate mass measurements of arsenic compounds, along with analytes from other chemical classes, were made in order to validate this procedure. Some of these results are presented in Table 2.2.  [CH xl  x2  i I,.()  [JJ]  ENH ill  I1.5  fl.II  11.141  14.l1  11.141  14.44$  ,...41  wz  Reagent gas ion ratios PEAK PEAK GAS AREA TWTGHT IONS 0.86 0.68 ] 5 [CH 0.43 0.32 [NB II] 1.00 1.00 [IJ,J  100-  .  50-  I81  ‘43 943  493  293  931  L ii  rr  0— 100  800  J!j ’ 1T  iiJiL[IbiiL  800  Li1 1  . 1 T  400  [LII  -  800  I  I  j 800  I  I  I  700  800  M/Z  Figure 2.16 PFK mass spectrum obtained under mixed gas CI conditions. The ammonia to methane ratio was monitored at mlz 18, and 17, 29 respectively. The relative intensities and peak areas of these ions are presented above.  34  Table 2.2 Positive ion accurate mass measurements of arsenic and other compounds in DCI mode, using PFK or Fomblin 18/8 as internal reference standards.  Compound / Formula Arsenobetaine/ s 11 H 5 C A 2 O  Tetramethylarsonium ion / (CH As 4 ) 3  Reference I Reagent gas 3 PFK/NH 4 &CH  PFK/NH 3 4 & CH  Error ppm 0.39  Theoretical mlz 179.0053  Measured m/z 179.0054  sf’ [(CH A 4 ) 3  135.0155  135.0157  1.48  s+H] [(CH A ) 3  120.9998  120.9995  2.89  s] [(CH A 2 ) 3  104.9686  104.9690  4.29  sH 2 A 3 [CH f  91.9607  91.9604  3.59  [Asf  74.9216  74.9215  1.33  sf [(CH A 4 ) 3  135.0155  135.0150  3.70  As+Hj’ ) 3 [(CH  120.9998  121.0003  3.72  sf [(CH A 2 ) 3  104.9686  104.9690  4.29  sH 2 A 3 [CH f  91.9607  91.9604  3.59  [As]  74.9216  74.9212  5.34  Ions Measured [M+Hj  Monosaccharide / 8 6 2 C 2 H 1 0  Fombim / 3 NH  t ] 4 [M+NH  452.1768  452.177  0.43  Macrocycle I 2 3 2rS C 3 H B 5 O  PFK / CH 4  [M+H]t  489.1231  489.131  0.35  Hemicarcerand / 7 5 C 9 H 3 N 8 0  Fomblin / 12 H 4 i-C  [M+H]  1430.7 197  1430.720  0.29  Disaccharide / 4 23 C 3 H 15 F0  3 PFK /NH 4 &CH  t ] 4 [M+NH  598.2147  598.212  5.00  35  2.2.2.2 Fomblin (18/8) as an Internal Reference Standard for Accurate Mass Measurements in Ammonia DCI-MS  Several groups have reported the use of Fomblin as a calibration substance for negative ion 1 Cl-MS. 13-115 In this work we have investigated the use of Fomblin as a calibration substance for high resolution positive ion DCI-MS. Fomblin was analyzed in the presence of three different reagent gases, , , or CH 3 . The spectra obtained 4 12 NH H 4 i-C were evaluated for their suitability as calibration references. When i-C 12 was used as CI reagent gas in the positive ion DCI mode, three major H 4 ion series were observed in the mass spectrum of Fomblin (Figure 2.17). The mlz values of these ion series can be expressed as A=69+n166, B135+n166, and C185+n166 (Figure 2.18). The perfluoropropylene oxide (C 0) units of Fomblin are believed to be 6 F 3 3 (m!z=69), C 0 (m/z=135), 5 F 2 responsible for the 166 segment, while groups such as CF and C 0 (mIz=185) are believed to be the terminating groups of these oligomeric 7 F 3 Fomblin fragments. The observed fragmentation pattern is in agreement with those reported previously. 13,116 The presence of these ion series throughout a wide mass range, from 1 mlz= 100 to mIz=23 00, makes Fomblin (18/8) very suitable for use as an internal reference standard for accurate mass measurements of positive ions under DCI mode, thus supplementing PFK at the high mass end (700 <m/z  <  2300).  When Fomblin was analyzed under NH 3 CI conditions the resulting mass spectrum  exhibited the series A’, B’, C’ (Figure 2.19), similar to those previously described ion series A, B, C but shifted upwards by 17 Da. Accurate mass measurements of the new spectrum revealed the presence of NH 3 adducts. The ion series A, B, and C were also present as minor components. Using CH 4 as the reagent gas, the Fomblin fragmentation pattern differed from the  4 CI conditions a mass shift of the major patterns discussed above (Figure 2.20). Under CH fragment ions was observed. This can be interpreted as being either a downward shift of the  ) 36  major ion series A by 22 Da or an upward shift of the ion series C by 28 Da. A downward shift of ion series A by 22 Da may be attributed to the addition of an oxygen and the loss of two fluorine atoms, from the respective fragment ions (Scheme 2.1). An upward shift of ion series C by 28 Da may be attributed to the capture of CO from the methane gas (Scheme 4 reagent gas for the 2.2). The latter possibility was investigated by analyzing the CH presence of CO. It was found that only a minute amount of CO was present in the gas stream. Thus we believe that the CO in the gas stream is not responsible for the mass shift of the major Fomblin fragment ions under methane-Cl conditions, and that the proposed fragmentation pathway outlined in Scheme 2.1 is most favourable. However, it should be 4 conditions are normally present mentioned that the major Fomblin ions observed under CH 12 H 4 i-C 3 or . as minor signals in El mode as well as under CI conditions using NH A too. 9080B ‘“  A  C  10  6050  A  c  1Z31  2015  40-  A ian  3O  A  B B  7 c 9 ‘  20  c 3 ’ 4  I  800  1000  1563  4 U a .Lk 1400  1200  B  A  a629C’  1600  B C  1800  A  2000  2 as the reagent gas. i-C 1 H Figure 2.17 Positive ion DCI mass spectrum of Fomblin using 4 The three major ion series observed are labelled as A, B, and C.  37  ii  m/z  1  235  5 6 7  899 1065 1231  1  301  5 6 7  965 1131 1297  1  351  5 6  1015 1181  7  1347  FF F FC I (I  As..  n  I i PP  CF 3  FFF  B  O_(— C _ 5 F 2  FPF  PC I I  PC Il  or 3  ii  ii  FF  FP FFp  FFp FC I I _O_(_C_C_O_)_ 7 F 3 C  C  PC C2 5 F  (  ——  PP  FF  Figure 2.18 Structural assignment of Fomblin fragment ions (A=69+n166; B=135+n166, and C=185+n166), obtained under positive ion DCI conditions, using i as reagent gas. iooA’  A’  A’  1012  90A’  80-  A’  1414  1248  70  50  A’  1510  60  1746 616  B’ C’ B’  C’  1912  C’  1530  982  C’  40  A  C’  B’  B’  1364  is’s  1862  fl’ 1976  20  800  1000  1400  1200  1600  1800  2000  3 as the reagent gas. Figure 2.19 Positive ion DCI mass spectrum of Fomblin using NH The three major ion series observed are labelled as A’, B’, and C’.  38  100 90 1043  80 70 1209  60 111 761  1375  977 927  30-  1143 1541  1309  849  I  20-  1475  fl ..i 1i  fbi  10-  800  hi  1000  1200  1600  1400  1800  M1Z 2000  4 as the reagent gas. Figure 220 Positive ion DCI mass spectrum of Fomblin using CH  Scheme 2.1 FFF  FFp  FC I  I  -Ec--c—o)-cp 3 I FF  PC  -2F  —O<C O 4 P 3 )(-C—C—O---CP I I n-i PP  +0  Ion Series A ii  mz  1  235  5 6 7  899 1065  n  m,z  _p  1  213  +0  5 6  877 1043 1209  1231  7  39  Scheme 22 FPp  PFp \lf  7 —E-- —?—o-)— r 3 c  PC  +Co  7 —E-- —.—o-)— p 3 —[c  PP  pP  n  +  co]  Ion Series C  n  m/z  1  351  5 6 7  1015 1181 1347  n  nzfz  5 6 7  1043 12.09 1375  +co  Fomblin solutions of various concentrations were made up in PF-5060. This solvent is volatile (b.p. 56 °C) and desorbs much earlier from the probe filament than Fomblin, thus eliminating an interference which could be caused by the solvent. Manufacturer -C perfluoro compounds. The mass 5 C specifications list this solvent as containing 18 4 which are of the same ), 1 F 6 spectrum of the solvent showed ions up to m/z 328 (C composition as those observed forPFK (m1z69, 100, 119, 131, 169, 219, 319). The procedure described above, using Fomblin 18/8 as reference, was evaluated for accurate mass measurements of positive ions in DCI (CI) mode. Representative data obtained from a number of compounds are included in Table 2.2. Data presented in this study indicate that mass deficient calibration reference compounds such as PFK and Fomblin can be used successfully for high resolution accurate mass measurements in positive DCI (CI) mode. These compounds have a number of attractive properties such as adequate volatility, mass deficiency, and compatibility with the most widely used CI gases or gas mixtures. The spectra consist of a series of reference ions covering a wide mass range. Thus they are generally better reference materials for DCI-MS than other compounds reported in the literature.” 1, 112, 117, 118  40  4 can be used This investigation has shown that a gas mixture of NH 3 and CH 3 or CH 4 for 12 NH H 4 i-C successfully with PFK, while Fomblin can be used with either , accurate mass measurements of positive ions in DCI (CI) mode. 83  2.2.3 Analysis of Arsenic Compounds by using MALDI-TOF-MS As mentioned previously one of the objectives of our work was to investigate the use of the highly sensitive MALDI-TOF-MS technique for the analysis of low molecular weight arsenic compounds. The MALDI-TOF mass spectrum of AsB (Figure 2.21) shows a [M+Hf ion at m/z 179, as well as [M+Naf and [M+Kf ions at mlz 201 and 217, respectively. The base peak of the spectrum appears at mlz 135 which corresponds to a loss of 44 amu (C0 ) from the 2 protonated AsB molecule. It is interesting to note that this four ion pattern is repeated for the AsB dimer ion cluster. Proton, sodium, and potassium adducts of the dimer are present 2 is seen at mlz 357, 379, and 395 respectively, while the protonated dimer after loss of CO at mlz 313. Thus the pattern observed here, consisting of 8 ion peaks, serves as a unique signature for AsB. DHB was used as the matrix component. The approximate minimum detectable amount (MDA) of AsB was determined to be 0.3 ng of arsenic or 4 pmole of compound. These values correspond to the total amount of arsenic or compound respectively in the 1.5 .iL aliquot of solution deposited on the sample target. Because each mass spectrum is only generated from analyte desorbed from a very small area exposed to the laser beam, a large area of the solid phase matrix-analyte mixture is not used and thus does not contribute to the signal obtained. Therefore the MDAs reported here could be improved by either reducing the solution droplet volume loaded onto the target, effectively reducing the amount of analyte loaded, or by reducing the area onto which the solution droplet is deposited, thus allowing for the formation of a solid phase matrix-analyte mixture containing a much higher concentration of analyte.  41  Most publications reporting the use of mass spectrometry in organoarsenical structural determinations have not stated the detection limits of the methods used or even the amount . (We are not referring to ctra 674788 ° spe of sample needed to obtain suitable mass 23293 ICP-MS, which is an element specific detector and only provides definite structural information when used in conjunction with synthetic standards and chromatographic techniques). Therefore it is very difficult to compare the detection limits of the various mass spectrometric methods used to analyze these compounds. One notable exception is the work kers who analyzed organoarsenicals by using atmospheric pressure 2 81 co-wor 8 of Siu and ’ chemical ionization and electrospray MS/MS. Initially they described the introduction of 1100 ppm solutions of arsenic either continuously, or via flow injection. In SIM mode (mlz 81 In more recent 135) they reported a minimum detectable amount of 1 ng arsenic for AsB. work where they used ion-exchange chromatography coupled to electrospray MS/MS they report the minimum detectable amount for AsB to be about 20 pg in the selected reaction 0 As described above DCI 2 H-C j. CH 8 [M+ 2 I monitoring mode (m/z 179 [M+Hf/120 3 MS can provide satisfactory full scan spectra of the organoarsenicals from 100 ng arsenic, when using ammonia as the reagent gas (Section 2.2.1). When comparing these various MS techniques, it is clear that MALDI-TOF-MS has high potential, especially since, as reported, it can provide mass spectra in the full spectrum 8. analyte 9 ’ mode with very small amounts of 96 In order to investigate the effect that the matrix (DHB) has on the desorption and ionization of AsB, mass spectra of the compound were also obtained in the absence of the matrix component. Figure 2.22 shows the mass spectrum of AsB obtained under such experimental conditions. A notable feature of this mass spectrum is the absence of any molecular mass information; only the m/z 135 fragment ion is observed. A higher laser power was required in order to obtain this spectrum, indicating that DuB plays a major role  42  in facilitating the desorption and ionization of AsB. DHB and its alkali impurities are required for protonation and alkali adduct ion formation. Laser photoionization TOF-MS has been previously used for the analysis of some 119 The main feature of these spectra is the arsenic compounds, mainly inorganic. appearance of AsO+ from oxygen containing arsenic compounds; however, no detection limits or sample concentrations were reported. In the present investigation, using MB as a model, the effect of laser power on molecular ion intensities as well as on fragment ion formation was studied. After an initial maximum at 30% of full laser power the intensity of the protonated molecular ion of MB decreases with increasing laser power. The intensity of the fragment ion at m/z 135 increases initially, but finally decreases at very high laser power (Figure 2.23). Both, the total ion current and the ion at mlz 135 show an optimum at approximately 45% laser power. From these results it is evident that, in general, the laser power can be selected so that molecular ions, fragment ions or a combination of both can be observed.  43  I16  fM+B-cOj+  16 16 16 16  Pf9HJ+  76 7D 16  116 116  12M+H.O2J t2M+HI  16 16 16  LM+NaV’ I LM+K)  16 16  2Jl 17  1’4  l0  .11  6  C  2I3  Figure 2.21 MALDI-TOF mass spectrum of arsenobetaine, DHB used as matrix (*: matrix ions). 135  l00  I(dflAs]  16  ¶H3  —-H OO 2 C 3 H c  16 16  -  3 CH  ID 76 70 IC  416 J16  I  lID  IDI41WIiI  a0  lID  Figure 2.22 MALDI-TOF mass spectrum of arsenobetaine, no matrix used.  44  ID  qI  6III2 416  Z  2750 2500 2250 >  1750 1500 1250 1000 750 500 250 0 20  25  30  40  35  45  Laser Power  50  55  60  (%)  Figure 2.23 Effect of laser power (%) [100 % laser power 250 jaJj on the intensity of ions observed in MALDI-TOF mass spectrum of arsenobetaine, mlz 179 is the protonated molecular ion, mJz 135 is the ion formed after loss of CO . Each data point represents the 2 average ion intensities from 50 laser shots. (TIC: total ion current, matrix: DHB)  The tetramethylarsonium cation is another arsenic species which has been reported to exist in marine animals such as clams and mussels. 28 The MALDI-TOF mass spectrum of a sample of tetramethylarsonium iodide is presented in Figure 2.24. The base peak at m/z 135 corresponds  to  the  tetramethylarsonium  species.  An  iodide  adduct  of  the  tetramethylarsonium dimer is observed at m/z 397. In comparing this spectrum with the data obtained from AsB it is clear that SIM 82 are not sufficient to differentiate the various arsenicals. data The presence of arsenocholine in the marine environment has been in dispute since it was first reported in marine animals. 2327 Most of the positive evidence is based on retention times obtained by using HPLC. The reported retention times match those of synthetic standards of arsenocholine, however no additional evidence was provided. The MALDI-TOF mass spectrum of this compound is presented in Figure 2.25. A very strong molecular ion (m/z 165) is observed along with some fragmentation (m/z 147, 135). Such  45  spectral information would be invaluable in confirming the presence of AsC in environmental samples. This could be achieved by isolating HPLC fractions containing AsC and confirming its presence by using MALDI-TOF-MS. Furthermore the development of dynamic MALDI TOF should provide direct on-line MS capabilities. 1 19ac Disodium methylarsonate 2 AsO(ONa) was also analyzed by using MALDI 3 [CH ] TOF-MS (Figure 2.26). The base peak at mlz 139 corresponds to protonated methylarsonate. In addition very intense sodium and potassium adducts are also observed at mlz 161 and 177, respectively. This ion pattern is also repeated for the methylarsonate dimer ions. Another class of arsenicals worthy of investigating consists of the arylarsenicals that are used as animal feed additives for swine and poultry. 120-12 1 Although these materials have been used on large scale to promote production of these domestic animals for the consumer market, very little analytical work has been carried out on these compounds. One of these compounds, p-arsanilic acid, was analyzed in this study (Figure 2.27). The main characteristics of the spectrum are the very intense protonated molecule as well as sodium and potassium adducts of lower relative intensity. These compounds generally are very difficult to analyze by MS methods, such as El or DCI, because they normally decompose or pyrolize on heatable MS probes. It should be noted that we have been able to successfully analyze these compounds by using dynamic FAB, with MDAs at about 10 ng (Chapter 4, -  of this Thesis).  46  I  11  S  laD ‘a 12z (cH) As+IJ+ 4 a a S S Is Is  a  177  S .  Figure 2.24 MALDI-TOF mass spectrum of tetramethylarsonium iodide, DHB used as matrix. (*: matrix ions, DHB) IM]  165  H3  Is Is  H3C_cH2cH2OH 3 Ca  a Is Is a ,aD  a Is S  LM-B20]’  a I’ Is S Is .  ‘  w’  135 147  Li..  1  i,i  7 ‘i1!_’ àá  472  4W  a  Figure 2.25 MALDI-TOF mass spectrum of arsenocholine, DHB used as matrix. (*: matrix ions, DHB)  47  _  139 —  0  LM-2Na+HJ  H L 3 -0Na  ONa  S U  LM-Na1 IM-2Na+K] —  •  161  III  (2M.4Na+H1+  j 177  277  •  L2M-3Na]  299 S  —  257  S  L2M-4Na+KJ 315  $  •  I4  N  • S  lCD  I  JiJk  .  SC  SC  SC  Figure 2.26 MALDI-TOF mass spectrum of disodium methylarsonate, DHB used as matrix. (*: matrix ions, DHB) SC  0  S CD  7’  2 NH.  7. S S  I” ,EID gIl 4D  IM+Ns1 1 1M+Kr  CD K  S  137  CD  /26 240/  CD I •  •  .—,-.—I•  . •  I +1T, SC  SC  I  *  376 aix  SC  Figure 2.27 MALDI-TOF mass spectrum of p-arsanilic acid, DHB used as matrix.  (*: matrix ions, DHB)  48  Attempts were also made to analyze inorganic arsenic salts, (arsenite and arsenate) by using MALDI-TOF-MS. The resulting mass spectra did not provide any meaningful structural information. Quantitative analysis by MALDI-TOF-MS is still considered a highly unreliable procedure since many variables associated with sample preparation and analytical procedure can seriously affect the results. So far, only a small number of reports concerning quantitative analysis by using the technique have been made. 123 In our ’ 122 view the main problem is associated with inhomogeneity of the solid phase matrix-analyte mixture. Thus, the resulting mass spectrum is highly dependent on the position of laser impact. In our experience very poor precision is obtained not only between measurements made on different sample loadings, but also within the same sample deposition. This problem can be resolved to a curtain degree by firing the laser beam at a large number of different spots on the matrix-analyte sample, either automatically or manually, and averaging the resulting mass spectra. Another possible way of overcoming the inhomogeneity would be to recrystallize the solid phase matrix-analyte mixture on the target by using solvents, such as ethanol. Preliminary results from our investigations using either of these procedures did not produce analytically acceptable calibration curves for the organoarsenicals. However, some improvement was observed when the matrix (DHB) molecular ion was used as an internal standard, Figure 2.28 shows a calibration graph obtained after following such a procedure. A recent report 123 has claimed that the addition of nitrocellulose provides improvement in sample-to-sample reproducibility of MALDI ion yields and also improves the precision for peptide quantitation. MS and optical microscopy results suggest that the nitrocellulose modifies the crystallization of matrix-analyte solution to allow even coverage over the sample surface.  49  5.5 5.0 4.5 I  4.0 3.5  i.  3.0  .2 4-  2.5  0  2.0 C -J  1.5 1.0 0.5 0.0 0  1  2  3  4  Ln [amount of As(ng)]  Figure 2.28 MALDI-TOF-MS calibration graph obtained from the analysis of MB (m/z 179 corresponds to [M+H1). The molecular ion of DHB (mlz 154) was used as an internal standard.  All of the above arsenic compounds were analyzed in a variety of different matrices, including DHB, Sinapic (3,5-dimethoxy-4-hydroxycinnamic acid), HABA [2-(-4hydroxyphenylazo)benzoic acidj, and gentisic acid. DHB provided the best results, both in sensitivity as well as spectral information. The other matrices provided poor spectra.  2.2.4 Structural Characterization of Arsines by using HG-GC-MS A need for HG-GC-MS analysis of arsines developed out of the work of H. Li. ” 63  In his work Polyphysa peniculus was grown in artificial seawater in the presence of arsenate and L-methionine-methyl-d3. The HG-GC-MS method described here was used to 3 moiety from L-methionine-methyl-d determine if the -CD 3 is incorporated into arsenic compounds produced by the algae.  50  The principle of the method is based on the fact that inorganic and methylated arsenic compounds, i.e. As(OH) 3 and (CH3)As(O)(OH)3 (n=O-3), can be reduced to their corresponding arsines, (CH3)nAsH3n (n=O-3), by allowing them to react with sodium borohydride. This reduction is a pH dependent process and pH control has been used to speciate arsenate and ’ 5 54 arsenite. 0 ” 1 25 6 Under pH 7 arsenicals can be reduced to the arsines. At pH  >  <  1 conditions, all hydride forming  4, arsenite but not arsenate forms arsine.  The evolved arsines can be separated and detected by using gas chromatography  -  atomic  absorption spectrometry 21 ’ 27 (GC-AA) 3 9 and 2 J 2961 GC-MS. 1 29 7 The majority of studies, reporting the use of a mass spectrometer as a detector, used the selected-ionmonitoring (SIM) mode and demonstrated a capability for characterizing arsine, methylarsine, dimethylarsine and trimethylarsine. In a study involving the analysis of arsenic in NaOH digested tissue extracts of shellfish, fish, crustaceans, and seaweeds, Kaise et al.-° 7 collected the arsines, generated by NaBH 4 reduction, in a liquid-nitrogen cooled trap that was coupled to a GC-MS system. The SIM mode gave detection limits of 0.3 ng arsenic. Earlier, Odanaka et al.’ 27 were able to quantify hydride forming arsenicals by using a combination of a GC-MS system, an off-line hydride generator and a heptane cold trap. The arsines were trapped in the cold heptane and the detection limits were between 10-20 ng of arsenic. In almost all cases the most abundant ions of each arsine were used for the SIM analysis, e.g. mlz 78 [AsH AsH 3 [CH ] , 90 J and 76 [AsH] for arsine; m/z 92 2 3 As] for 3 As] and 76 [AsHI’ for methylarsine; mlz 106 [(CH 3 [CH AsHJ and 90 [CH 2 ) 3 As] for ) 2 dimethylarsine; m/z 120 [(CH AsJ, 105 [(CH ) 3 AsJ and 103 [(CH 2 ) 3 trimethylarsine. In the present study the mass spectrometer was scanned from m/z 74 to 115 at 1 scan per 0.1 s. Scans above m/z 115 were not made because there was no need to investigate for trimethylarsine whose absence had been confirmed by using HG-GC-atomic absorption 24 The wide-scan monitoring mode was used instead of the SIM mode, for a spectrometry.’  51  number of reasons. First of all only the wide-scan mode would allow for the observation of any unsuspected fragment ions resulting from the deuterium labelled arsines. This precaution was taken because the fragmentation patterns of deuterium labelled arsines apart from ° 13 A ) 3 (CD s, were unknown. The other reason for acquiring data in the wide-scan mode was to establish the identity of any other volatile compounds resulting from the cell culture samples. Any such interfering ions could lead to erroneous results in the SIM mode. The HG-GC-MS experimental conditions used are all listed in Table 2.3. These conditions were established after optimizing the hydride generation efficiency, the chromatographic resolution, and the mass spectrometric sensitivity for the arsines of interest. The mass spectra of arsine, methylarsine and dimethylarsine obtained by using the wide-scan monitoring mode are presented in Figure 2.29. The present study was mainly concerned with the detection of dimethylarsine. By using the HG-GC-MS methodology described above it was possible to detect down to 25 ng of arsenic. In the SIM mode the detection limit could easily be improved by at least an order of magnitude.  Table 2.3 HG-GC-MS experimental parameters Hydride Generation  Gas Chromatography  Mass Spectrometer  2% NaBH 4  Initial temperature: 70°C  Interface temperature: 110°C  1 M HC1  Ramp rate: 30°C.min 1  Ion source temperature: 140°C  3 mL Sample  Final temperature: 150°C for 2 mm  Scan every 0.1 s for 4 mm  1 Helium flow rate: 30 mL.min  Scan range: m/z 74-115  52  eu  90%  AsH3  [M-2] +  60% 0% 40% 50% 40% 30% 20% 10%  O%.1 70  75  90  95  95  90  100  105  110  115  M!Z 100*  CH3AsH2  90% 76 70%  [M]+  40%  I  50% 40% 30% 20% 10%  [ 70  75  60  93  90  45  1QQ  105  110  115  M/Z  .  100%  90%  (CH3)2AsH  90% 70% 60% 30% 40% .  30% 20%  16  7  1:: 70  t  I  I  I 75  60  IS  90  95  I  100  I 205  110  115  MIZ  Figure 2.29 Mass spectra of arsine, methylarsine, and dimethylarsine are presented. A solution containing standard arsenate, methylarsonic acid, and dimethylarsinic acid (50 ng of arsenic for each compound) was used to generate the arsines. 53  The data obtained lead to the conclusion that, when L-methionine-methyl-d3 is added 3 label is incorporated into the dimethylarsenic species to a to the medium, the CD considerable extent (in some cases up to 30  %). The mass spectra obtained from samples  sHf [CD A f, mlz 94 3 (CD )AsH 3 [(CH sHf, m/z 109 ) [(CD A 2 ) contain ions at mlz 112 3 sJ+ (Figure 2.30A). These ions are absent from the mass spectra obtained [CD A and mlz 93 3 -methyl-d (Figure 2.30B). Further evidence of 3 from the media containing no L-methionine 3 into the dimethylarsenic species is provided by the single ion the incorporation of the CD chromatograms at m/z 93, 109, and 112, shown in Figure 2.31. It can be seen that peak retention times are identical with the retention time of dimethylarsine. 3 incorporation was determined by comparing the relative peak The percentage of CD F, and 112 (CD A5HT ) 3 [(CH sHf, 109 ) [(CH A 2 ) intensities of the molecular ions, m/z 106 3 AsH] The relative standard deviation of the peak intensity of m/z 106 for 4 2 ) 3t [(CD determinations of standard dimethylarsinic acid was 6%. Also, the assumption that the ionization efficiencies and thus the responses of the deuterated arsines are identical to the responses of the undeuterated arsines  may cause additional errors in the calculated  ° addressed this assumption, 3 incorporation percentages. A previous study by Cullen et at’  and concluded that the deuterated trimethylarsine shows slightly lower sensitivity under El mass spectrometric conditions.  54  100%-  I  A  90% 80%  [(CH3)2ASH]+  70%  )(CH3)ASH] + 3 [(CD  I  —  60%  [(CD3)2AsHJ +  50%  1  [CD3As] 30% 20%  [44]+  II  75 I  0%70  103  Hi. 75  1  191  10%  199  ii 80  85  95  90  100  105  112  110  115  MIZ  J.uu.  [CH3AsJ  90%  B  80% 70% 60% 50% 40%  30%  [(CH3)2AsH)  LAs] 20% 10% 0%70  1  75  I  75  11  80  85  90  I  95  100  105  110  115  MIZ  Figure 2.30  A: Mass spectrum obtained from culture medium inoculated with L  methionine-methyl-d3 and arsenate, average of 100 scans. B: Mass spectrum obtained from culture medium inoculated only with arsenate, average of 100 scans.  55  a li—GO?  0.—DOG C0  0.2  0.’  0.5  0.5  ‘.2  1.0  OOO  z z  0 .  T  .5  l.2  Retention Time (s) Figure 2.31 Full scan and Selected Ion Chromatograms (SIC) obtained from the analysis of culture medium inoculated with both L-methionine-methyl-d3 and arsenate. a. Full scan m/z 74-115 b. SIC of m/z 112 c. SIC of m/z 109 d. SIC of mlz 93.  56  2.4 SUMMARY The work which has been described in this chapter is mainly concerned with the development of mass spectrometric techniques for the analysis of arsenic compounds of environmental interest. In particular, DCI MS was shown to produce mass spectra with characteristic ions and fragmentation patterns, thus providing abundant structural information which could be used to analyze such compounds. This allows identification based on more ions than just the protonated molecular species. Accurate mass measurements can also be made by using this technique in conjunction with a high resolution mass spectrometer. Our investigations have shown that a gas mixture of NH 3 and CH 4 can be used successfiully with PFK as reference, while Fomblin can be used as reference with either , 12 NH H 4 i-C 3 or CH 4 as reagent gases in accurate mass measurements of positive ions in DCI (CI) mode. This technique will be of particular importance in identifying unknown arsenic species, as well as a wide variety of compounds which have been previously analyzed by using low resolution DCI-MS. We have also shown that MALDI-TOF-MS has great potential for the analysis of organoarsenicals of environmental interest. Some of the main features of this technique are as follows: 1. The method is capable of providing molecular ion as well as structural information for a variety of arsenic compounds. Molecular or quasi-molecular ions are observed in all the MALDI mass spectra presented in this work. In addition, in many cases structurally characteristic fragment ions and/or dimer formation along with sodium and potassium adducts were observed. These features allow for the unambiguous identification of the arsenicals investigated so far. 2. The flll spectrum detection limit of the method is extremely good compared to that of other mass spectrometric methods reported. For instance, AsB, was detected at levels as low as 0.3 ng of arsenic (4 pmole of compound).  57  3. Only small volumes of sample are required for each analysis, <1.5 tL. 4. The observed ion intensities are highly dependent on the N 2 laser power. Higher power results in increased fragmentation coupled with a decrease in the intensity of the molecular ions. 5. Not all matrices normally used in MALDI-MS are suitable for the analysis: DHB is the most suitable matrix for the MALDI analysis of organoarsenicals. It is probable that MALDI-TOF-MS could be used, possibly with the same degree of success, for the analysis of other organometallics of environmental concern, particularly those of Sn, Se, and Sb. The development of dynamic LC MALDI-TOF-MS should increase the potential of -  the method to solve problems such as the ones referred to in this work.” 9a-c Finally a HG-GC-MS method was developed which provided conclusive evidence of  -  3 incorporation into arsenic compounds produced from arsenate by alga cell cultures. CD This evidence supports the hypothesis that 5-adenosylmethionine, or some closely related suiphonium compound, is involved in the biological methylation of arsenic.  58  CHAPTER 3  OPTIMIZATION OF ATOMIC ABSORPTION SPECTROMETRIC METHODS FOR THE DETERMINATION OF ARSENIC IN BIOLOGICAL SAMPLES OF MARINE ORIGIN  3.1 INTRODUCTION  3.1.1 Scope and Rationale of Work The main objective of the work reported in this Chapter, is to optimize atomic absorption spectrometric (AAS) methods in order to allow for accurate determinations of total arsenic in samples of marine origin. More specifically, we are concerned with the arsenic content of Mytilus calfornianus (Californian mussels) collected from the B.C. coast. One of the most commonly used AAS methods for the analysis of metals in various 131 Even though samples is Electrothermal atomic absorption spectrometry (ETAAS). ETAAS compares very well in terms of performance with most other analytical techniques, a number of interferences have been reported. These are especially pronounced when analyzing environmental samples. 1 32-134 In order to overcome these interferences, various modifications to parts of the atomic absorption spectrometer have been reported. Platform 36 have been used to overcome chemical interferences. 35 and matrix modifiers’ atomizers’ Zeeman or deuterium background correction systems have been used to compensate for 137 The performance characteristics of ETAAS are molecular absorption or light scattering. affected by a wide range of instrumental components and fhrnace heating conditions. Because of the great variations in experimental parameters it is almost impossible to adopt a  59  set of experimental conditions established for the detennination of arsenic on one AAS sys tem for the determination of the same element on a slightly modified AAS system. Even conditions recommended for a broad category of samples (e.g. biological), may not be optimum for a more specific sample type (e.g. oyster tissue), on the same spectrometer. It is therefore evident that an efficient optimization procedure must be adopted in order to accomplish the accurate determination of arsenic in environmental samples. The majority 38 This has a procedures. 1 ’ of reports to date have used one-factor-at-a-time optimization 66 number of drawbacks; large numbers of experiments are required and the best conditions may be missed if important interactions exist between experimental parameters. ° (CMS)l3 M In this study we have employed the Composite Modified Simplex 9 optimization method in conjunction with a Standard Reference Material (SRM) of marine origin, in order to delineate the optimum experimental conditions for the analysis of arsenic by using ETAAS. Appropriate SRMs containing certified amounts of arsenic were selected based on the similarity of their matrix to the matrix of the marine samples of interest.  3.1.2 Atomic Absorption Spectrometry AAS was first introduced by the Australian physicist Alan Walsh, in 1955.141 This analytical method is based on the conversion of the sample components into gaseous atoms which absorb radiation of wavelengths characteristic of their electronic transitions. Measurement of the amount of absorption allows a quantitative determination of the analyte. After some initial hesitation by scientists to utilize flame AAS for the determination of metals, the method soon saw extensive use in the 1970’s. Further development of the method has resulted in a highly reliable method which probably reached its development plateau several years ago. Meanwhile the introduction of other AAS methods have resulted in advances in the area of trace metal determinations. L’vov first described Electrothermal (ET) or Graphite  60  142 This method offers great improvements in terms Furnace (GF) AAS, three decades ago. of sensitivity for the analysis of metals, and has since become the most widely used analytical method for such purposes.  3.1.2.1 Electrothermal Atomic Absorption Spectrometry and the Determination of Arsenic .iL) is deposited onto the inner wall of a In ETAAS a very small volume (1-100 1 graphite tube, which is aligned in the path of light of an element-specific hollow cathode lamp (HCL). In some cases the solution is deposited on to a platform placed inside the tube. 143 eventually resulting in The graphite tube is then electrically heated in a number of steps, the vaporization of the sample and the formation of an atom cloud in the graphite tube. These atoms absorb light emitted from the HCL. The extremely high sensitivity of the technique for the analysis of metals can be attributed to a number of factors. First, the analyte transport efficiency into the observation path is very high, close to 100%: techniques employing pneumatic nebulizers, such as flame and inductively coupled plasma (ICP), have very low analyte transport efficiencies, typically between 1-10%. Secondly, residence times of the analyte atoms in the observation path are about two orders of magnitude higher than those found in other sources. Thus, it is the high analyte densities present in the observation path in ETAAS that are responsible for the high sensitivity and low detection limits of the technique. ETAAS has been used widely for the determination of arsenic in biological, 31 ,44, 145 The United States Environmental samples. 1 ’ geological, marine and fresh water 1 Protection Agency recommends ETAAS methods for the determination of a number of 146 Low detection limits, good precision, simplicity of op trace elements, including arsenic. eration as well as minimum sample pretreatment are all features which have contributed to the widespread use of the method.  61  3.1.2.2 Hydride Generation Atomic Absorption Spectrometry for the Determination of Arsenic Methods that employ hydride generation (HG) methodology coupled to atomic absorption spectrometry have been used extensively for the determination of hydride ° Such 5 49 and Pb.’ forming elements such as As, Bi, Se, and Sb, 148 Te and Sn,’ 147 Ge, analyses are accomplished by first converting the metal or metalloid into its corresponding hydride, the hydride is then transferred to an atomizer where it is decomposed to the gasphase metal (-bid) atoms. The atomization step takes place in the optical path of the spectrometer. The main advantage of this method is that it separates and preconcentrates the analyte from the sample matrix. This procedures eliminates matrix interferences that are quite commonly encountered when analyzing samples by using other AAS methods, such as ETAAS. HGAAS has been used extensively for the determination of arsenic in environmental samples. 53,151 Because of the nature of the technique only the hydride forming arsenic species will be detected. Since most samples of marine origin also contain non-hydride forming arsenicals, these samples must be digested and converted to, e.g. arsenate, prior to their analysis by using HGAAS. The digestion step is critical for the success of the analysis. If all the non-hydride forming arsenicals have not been converted into hydride forming 64 species, low recoveries are obtained. 3.1.3 Digestion Procedures for the Preparation of Biological Samples for Arsenic Determination Biological tissue samples are usually digested prior to their analysis, since most quantitative analytical methods employing instruments allow for only liquid sample introduction. A variety of digestion methods have been used to convert a largely organic analysis. 52-156 Acid digestions, as well as combustion matrix into a solution suitable for 1 methods have been carried out in both open and closed vessels. A summary of the most  62  commonly used digestion techniques for decomposition of biological materials is given in Table 3.1.  Table 3.1 Most common digestion methods used for the preparation of biological samples  for metal anal ‘ses Common  Vessel  name of  type  Reagents  Conditions  Energy  Advantages  Disadvantages  Large amounts of reagents, contamination problems May not always accomplish  source  technique  Wet or Acid Digestion  Microwave Digestion  Open  Acid mixtures,  Closed  oxidizing agent Acid mixtures, oxidizing  agent  Oxygen or Schoniger  Temperature of acid boiling points  Hot plate  Specialized equipment not required  High pressures, high temperatures  Microwave energy  Very fast, typically 1-10 miii, on-line digestions  complete  Closed  Oxygen  Flame Combustion  Flame in 02 environment  decomposition, specialized equipment Very fast, no Often difficult to reagents dissolve ash, needed poor precision  Open  Oxygen  High temperatures (400-700 °C)  Muffle Furnace  Low amounts of  flask  combustion Dry Ashing  and ashing  aids  reagents required. Can handle large amounts of sample  Slow, possible analyte loss at high temperatures, often difficult to dissolve ash, contamination problems  In order to select the appropriate digestion technique for the preparation of a particular biological sample, a number of factors must be taken into consideration. For example, some analytical methods require that the metal under investigation be completely removed from its organic substituents. Other factors which must be taken into consideration are the following: the digestion vessel must not adsorb the metal of interest onto its walls or  63  allow the evaporation of the metal, also the digestion reagents must not interfere with the analysis or contain impurities which may contaminate the sample under investigation.  3.1.3.1 The use of Microwave Energy for the Digestion of Biological Materials Microwave energy as a heating source for acid digestions was first introduced in 1975.157 At that time these digestions were performed at atmospheric pressure.’ 59 ” 58 Because of the problems associated with open vessel digestions, researchers turned to 2 and Teflon® PFA [(Perfluoro alkoxy) ethylene] bottles 6 ” closed polycarbonate 160 digestion vessels.’ 62465 The main advantage of this procedure is that the digestion takes place under high pressure, which allows the acid mixtures to reach temperatures above the boiling points of the acids. Microwave digestions have been used with great success for 163 and biological samples. 164l65 62 geological materials, sample preparation of steels,’ The basic principle behind microwave heating is that microwave energy causes molecular motion of ions and also dipole rotation, but does not cause changes in molecular structure. Ionic conduction causes a flow of current which results in heat production, due to resistance to ion flow. Molecules with permanent or induced dipole moments are caused to rotate in microwave electromagnetic fields, this results in rapid heating.  3.1.4 Simplex Optimization Most analytical procedures must be optimized prior to their use. A very popular method for such a purpose is the old one-factor-at-a-time method. This optimization procedure investigates the effects of several factors (experimental variables) on a response, by holding all but one factors constant. The responses obtained at the various levels of this factor are evaluated. Successive factors are then allowed to vary, while at the same time keeping the previous factors at their optimum levels. However simple this procedure seems, there are a number of disadvantages associated with it. It is a time consuming and tedious  64  procedure, and most importantly it does not guarantee that an optimum will be reached, especially if there exists an interdependence of the optimum values of the chosen factors. In order to avoid these problems other more advanced statistical methods have been  66 in 1962 has been developed. Simplex optimization, first introduced by Spendley et al.’ 167 were the shown to be quite successful in solving these problems. Deming and Morgan first to employ the procedure to optimize methods in analytical chemistry. To understand how simplex optimization works, let us consider a chemical system for which we want to optimize n variables (n=l,2,3,...). These variables (factors) can be considered as n orthogonal axes, with each experiment being represented by one point in this n dimensional space. A simplex is defined as a geometric figure consisting of (n+1) vertices. So for a 2 variable chemical system, the simplex will be a triangle, for a 3 variable system it will be a tetrahedron, etc. So for n variables, n+1 experiments must be conducted in order to construct the first simplex. To better illustrate how the simplex optimization procedure functions, let us consider a two variable system. The initial simplex will be a triangle, which is defined by the initial trial points labelled 1,2, and 3, as shown in Figure 3. 1A. If we assume that the worst response was obtained at point 3, it would be logical to assume that a better response would be obtained at point 4 (Fig. 3. 1A), which is a reflection of point 3 with respect to the line joining 1 and 2. A new simplex is formed, consisting of points 1,2, and 4. By repeating this procedure the simplex “climbs” the response surface, and thus arrives at the optimum experimental conditions (Fig. 3. lB and C). To increase the chances of reaching a global optimum, as opposed to a local one, the search can be repeated by starting off with different initial vertices. The optimization procedure described above is referred to as the fixed-step or basic simplex method. Because of the limitations associated with this procedure, other more advanced simplex procedures, called modified simplex methods, were developed. These  65  methods allow for the acceleration of the simplex in favorable directions and deceleration in unfavourable directions. The expansion and contraction of the simplex results in faster optimization as the simplex “climbs” up the response surface and then contracts onto the 69 168 In the present work a modified version described by Betteridge et al. was used.’ top.  4  A  0  Variable 1  F  Figure 3.1 A simplex for two variables. A, B: simplex moves in a direction opposite the worst point; C: simplex “climbs” response surface and reaches optimum.  66  3.2 EXPERIMENTAL  3.2.1 Instrumentation An Atomic Absorption Spectrometer (Varian Techtron Model AA 1275) equipped with an arsenic hollow cathode lamp (Spectra AA) operating at 8 mA, and a deuterium back ground corrector were used for the arsenic determinations. The 193.7 rim arsenic resonance line was selected and used with a 1 nm bandwidth. For the ETAAS analysis the spectrometer was equipped with a GTA-95 (Varian) graphite furnace. Pyrolytically coated graphite partitioned tubes and argon purge gas were also used. For the HGAAS analysis a continuous flow system, shown in Figure 3.2, was employed. The experimental conditions used are listed in Table 3.2.  light path  peristaltic pump Sample Acid  regulators gas ,‘ liquid separator gaslines  gas/liquid lines  Figure 3.2 Schematic diagram of hydride generation assembly  67  Table 3.2 Operating conditions for the Hydride Generation Atomic Absorption System Material Uptake tubes  Sample:  2.80 mm I.D.  PVC  HC1:  2.28 mm I.D.  Viton  : 2.29 mm ID. 4 NaBH Uptake flow  PVC  Sample: 7.5 mLmin HC1:  2.0 mL.min 1  : 4.OmL.min’ 4 NaBH Carrier gas  : 100 1 2 N mL.min mixing coil : 25 mL.min 2 N 1 gas-liquid separator  HC1 concentration  4M  4 concentration NaBH  2 % (w/v) in 0.05% NaOH solution  Measurement mode  Run Mean  3.2.2 Reagents A 1000 ppm stock solution of arsenic, as arsenic trioxide, was used to prepare arsenic standard solutions. A 1000 ppm palladium solution, the matrix modifier, was also prepared by dissolving the appropriate amount of Pd powder in the minimum amount of aqua regia, followed by dilution with water containing 2% citric acid to the required volume. Standard solutions containing 100 ppm of arsenic as arsenobetaine, arsenocholine and tetramethylarsonium iodide were also prepared. These compounds were synthesized by using literature methods. 74,104,105 Microelemental analysis and nuclear magnetic reso nance spectroscopy were used to confirm their purity. Sample solutions containing microwave digested oyster tissue were used for the optimization of the ETAA experimental conditions. The freeze dried oyster tissue, standard reference material 1566a, was obtained from the U.S. National Institute of Standards and Technology (NIST).  68  Two other SRMs were also used in order to evaluate arsenic recoveries for the HGAAS method. Dogfish muscle reference material (DORM-i), and dogfish liver reference material (DOLT-i) were obtained from the National Research Council of Canada (NRCC).  3.2.3 Sample Preparation 3.2.3.1 Microwave Digestions Prior to ETAAS Analysis A commercial microwave oven (Sharp Carousel II) was used to digest samples contained in a Teflon® decomposition vessel (Parr Instrument Company, 45 mL). Samples of 150-400 mg each were weighed directly into the digestion vessel and 2 mL of concentrated nitric acid was added. The digestion vessel was assembled and placed in the microwave oven. The microwave program consisted of one 90 s step at high power output (500 W). After cooling the digestion vessel the content was diluted to 50 mL with de-ionized water. Blanks and arsenic standards were also prepared by using 2 mL of nitric acid and the same digestion and dilution procedure.  3.2.3.2 Wet Digestion Procedure Prior to HGAAS Analysis Homogenized mussel flesh or SRMs were freeze dried, 0.25-0.5 g of the sample was put into a 250 mL round bottom flask, along with 3 mL 30% hydrogen peroxide, 3 mL 69% nitric acid, and 1 mL sulfuric acid. A specially designed apparatus (Figure 3.3) was used for the digestion. This apparatus was originally designed and used for the digestion of biological samples for Se, As, and Hg analysis. L70 The digestion apparatus was placed in a heating mantle and heated for 3 h at approximately 250 0 C. The digestion was carried out in a fume hood. Finally the digestate was transferred into a 100 mL volumetric flask.  69  H 15mmFigure 3.3 Wet Digestion apparatus. A: Teflon® cylindrical plugs; B: Teflon® diffusion funnel; C: Teflon® stopper with capillary; D: 250 mL round-bottomed  flask,  550 mm  containing  sample and acid mixture. Adapted from the design described by Bajo et ° 7 al.’  3.2.4 Procedure for the Simplex Optimization of ETAAS Experimental Conditions The simplex optimization was carried out as follows. Four variables were studied for their effect on arsenic absorbance; ashing temperature, atomization temperature, modifier concentration, and atomization ramping time. The experimental variable names were entered into the microcomputer together with their ranges and the precision required for each  70  variable (Table 3.3). All the simplex caiculations were carried out by using the OPT1MA3 computer program. 171 This program was run on several PC/XT and PC/AT compatible IBM microcomputers. The initial set of conditions was entered and the program then  generated the four other sets needed to form the initial simplex and printed worksheets for each experiment. After the  completion of the experiments, the actual variable values  used and the peak area absorbances were entered. The variable values used were kept as close as possible to those suggested by the program.  Table 3.3 Range and precision of experimental variables examined Variable Ashing temperature (°C)  Lower limit  Upper Limit  Precision  600  1800  100  Atomization temperature (°C)  1800  2700  100  Modifier concentration (ppm)  25  500  25  Ramp time (see)  0.5  8  0.5  The program then calculated the next single set of conditions to be investigated and printed another worksheet. This process was continued, the program giving one new experiment each time. For each set of conditions three replicates were analyzed by using ETAAS. Between each one a blank injection was made to correct for possible lamp drift and to also assure that stable repeatable analytical results could be obtained. The optimum conditions obtained from this procedure were then used to analyze standard arsenic solutions and quantify the arsenic present in the microwave digested oyster tissue. These conditions were also used to analyze de-ionized water solutions containing ar senobetaine, arsenocholine and tetramethylarsonium iodide.  71  3.3 RESULTS AND DISCUSSION  3.3.1 Simplex Optimization of Conditions for the Determination of Arsenic in Environmental Samples by using ETAAS Various problems have been encountered when analyzing arsenic by using ETAAS, some of which are volatilization losses, interaction with the graphite tube, vapour phase in terferences, and spectral 1 interferences. 72,173 Furthermore the analysis of arsenic especially in samples of marine origin may pose additional problems. It is well documented that a large number of arsenic compounds are present in marine organisms. 1 Therefore if the appropriate experimental conditions are not selected it is possible that each arsenical may behave in a different way during ETAAS analysis. This results in sensitivity variations for different arsenicals (due to incomplete detection) and therefore leads to results highly dependent on the arsenic species present. These species-dependent effects may be missed if only standard arsenic solutions of artificial origin are examined. Consequently it is necessary to optimize the ETAAS conditions using ‘real” samples. SRMs which have a similar matrix to the environmental samples of interest are ideal for use in optimizing and validating the method. NIST oyster tissue was used in this study. Table 3.4 shows the optimization experiments performed and the responses obtained. Twenty three experiments were required before  establishing optimum conditions for the analysis of arsenic in oyster  tissue. A total of 33 experiments were performed before ending the simplex optimization search. The recorded absorbance from each experiment as a fbnction of variable value is displayed in Figure 3.4. The optimization procedure very quickly predicted the optimum ramping time. This was 0.5 sec (the minimum limit used in our optimization) and was reached after 13 experiments. Instrumental limitations did not permit use of a shorter ramping time. These  72  results indicate that shorter atomization ramping times improve the arsenic absorbance obtained in ETAAS. Longer ramping times probably allow for loss of arsenic at temperatures close to the atomization temperature. The optimum ashing temperature established in this study was 1500 °C. This temperature allows for the removal of matrix components which may otherwise act as interferences for the analysis of arsenic. Higher temperatures result in arsenic loss during the ashing stage, while lower temperatures may result in incomplete removal of various matrix interferences. The optimum atomization temperature established was 2200 °C. We have shown that this temperature allows for complete atomization of all arsenic in the SRM analyzed. The optimum modifier concentration was 500 ppm (upper limit value). Modifier con centrations are not very critical when analyzing standard arsenic solutions, but are extremely critical when analyzing environmental samples. This is probably a consequence of the sample’s matrix. In environmental samples higher concentrations of Pd are required because it interacts with various matrix components and may become unavailable to act as a matrix modifier for arsenic. However, use of higher concentration values for Pd results in higher costs per analysis.  73  Table 3.4 Simplex experiments and ETAAS response Cycle* Expt. Abs. [Pd] Ashing Atomization No. (a.u.) (ppm) (°C) (°C) 1 01 0.018 200 700 2000 2 0.019 200 1300 2000 3 0.015 2400 200 1000 4 0.010 200 1000 2100 5 0.016 1000 2100 500 6 1R 0.025 1200 2000 350 E 7 0.021 400 2000 1300 2R 8 1200 1800 0.000 500 9 C 0.026 2200 300 1100 10 3R 0.010 1200 2100 50 11 C 0.018 450 1000 2100 12 F 0.021 400 1100 2100 13 4R 0.042 450 2200 1600 14 E 0.031 1800 2400 500 15 F 0.038 400 2200 1500 16 5R 0.029 500 1600 2300 17 6R 0.030 450 1800 2300 18 7R 0.032 1800 500 2500 19 8R 0.037 500 1800 2300 20 9R 0.028 400 1800 2200 21 C 0.037 500 1600 2300 22 F 450 1700 0.035 2300 23 1OR 0.049 1500 500 2200 24 E 0.040 1400 500 2200 25 F 0.03 1 1600 2300 500 26 hR 0.026 1300 2000 500 27 C 0.030 500 1700 2400 28 12R 0.027 500 1300 2200 29 E 1700 2300 0.038 500 13R 0.045 2200 30 500 1600 31 14R 0.041 1300 2100 500 0.040 32 15R 500 1300 2100 E 1400 2100 33 0.038 500 * I: Initial cycle, R:reflection, E: expansion, C:contraction, F:fit. Sample standard deviation of 10 blank determinations: 0.002 a.u.  74  Ramp (sec) 4 4 4 7 5 1.5 0.5 1.5 1.5 0.5 3.5 3.0 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5  005 0.04  ,-“  IA ,A A  0  •  I—  <  0.05  A  /  0.03  o  I  A  A  0  f  L  I  <  /  A  0.01  A  A A  A  A A A  * 0.01  I 0.00 1600 1800 20D0 2200 2400 2600  A  A  I  I  1000  1500  0.00  Atomization Temp. (°c) 0.05  A  0.03  0  A  :  -  /  0  ‘0.02  .0  AA  -  0..  &QQ2 •  a, A  0.04-  Ashing Temperature  2000  (  C)  0.05  Al  A’  0.04  0.04  -a  0.03  <  2  o  0.03 ,.‘  a) à..  ,/  0.02 A  (I  .0  V  0 .0  0.01  “  0.01  0.00  A  0  1  2  3  4  5  I  I  6  7  -.,‘  0  Romping Time (s)  V  V  V  I  0.00 8  V  ,,  o  .  A  0.02  ‘‘  V  C q)  I  100 200 300 400 500 600 [Pd] (ppm)  Figure 3.4 The absorbances obtained from 33 optimization experiments plotted as a function of the four experimental parameters studied. Dotted lines indicate the estimated highest response level at this value of experimental parameter, irrespective of the values of the other three parameters.  75  The optimum furnace operating conditions along with the solution volumes used in this study are given in Table 3.5. Table 3.5 Furnace heating programa Step No.: Purpose  1: Drying 2: -II3: -II4: Ashing 5: -II6: -II7: Atomization 8:  -II-  9: Cleanup 10: -II-  Temperature (°C)  Time (s)  Argon Flow (mL.min 1) 3 3 3 3  85 95 120 ramp to 1500 1500 1500 ramp to 2200  5.0 40  2200  2.0  0  2200 ramp to 2600  2.0 0.5  3 3  10 5.0 1.0 2.0  3 3 0  Abs. measured Abs. measured  a: Sample and matrix modifier solutions (20 jiL of each ) injected onto ETAAS. Underlined parameters were optimized.  To check if these conditions also resulted in optimum arsenic recovery, the normal calibration method was used to determine arsenic in the NIST standard. The standard additions method was not required. Three samples were microwave digested and analyzed for arsenic, under the optimum ETAAS conditions. The arsenic recovery results obtained from these experiments, in addition to the certified values of arsenic in the SRM, are displayed in Table 3.6. These results indicate that the experimental conditions allow for quantitative determination of arsenic in NEST oyster tissue. Table 3.6 Arsenic Concentration Found in NEST oyster tissue  [As], j.tg.g  Analyzed by ETAAS  Certified Value  14.5, 14.8, 13.9  14.0  76  ±  1.2  Because a number of arsenic species, most of which exhibit different physical and chemical properties, have been reported to exist in the environment and in digests or ex tracts of environmental samples, the ETAAS conditions must be set so that the sensitivities for all these compounds are equal. In order to evaluate whether the established optimum conditions resulted in equal sensitivities for different arsenicals, standard solutions of arsenobetaine (AsB), arsenocholine (AsC) and the tetramethylarsonium ion (TMAsj, were analyzed. These compounds were selected because of their presence in marine organisms. Their determination using ETAAS was investigated by using the optimized experimental conditions previously established. Inductively coupled plasma-mass spectrometric (ICP MS) analysis was also performed on these solutions for comparison purposes (Table 3.7). Data from the table indicate excellent arsenic recovery for the three arsenicals analyzed.  Table 3.7 Arsenic concentration of solutions containing organoarsenicals [As}, [tg.mL Compound  ETAAS  ICP-MS  Arsenobetaine  0.065  0.003  0.069  ±  0.003  Arsenocholine Tetramethylarsonium Iodide  0.054 ± 0.002  0.058  ±  0.003  0.050 ± 0.00 1  0.050 ± 0.003  ±  ±  Sample standard deviation, n=3.  From this work, it becomes apparent that the application of an optimization procedure such as simplex (in this case the CMS method), in conjunction with a SRM, is of particular value when it is necessary to determine concentrations of a metal or a metalloid such as arsenic in environmental samples. Optimum conditions for various sample matrices can be quickly reached, thus eliminating matrix effects and also sensitivity variations resulting from the different arsenic species present in a particular sample.  77  This optimization procedure may also be used to improve the analysis of arsenic in the presence of various reagents or buffers used in HPLC eluents, e.g. ion-pair reagents such as heptanesulfonic acid. Recently a number of reports have been published on the analysis of arsenic by using ETAAS in conjunction with mixed modifiers, e.g. palladium and magnesium.’ 74 The procedure used here would be ideal for optimizing the concentrations of these modifiers in conjunction with the appropriate furnace heating program. We believe that the optimization procedure used here can easily be applied to in telligent automated ETAAS optimization. Since most ETAAS spectrometers can be programmed, the addition of simplex optimization is feasible. This would then allow for the optimal analysis of a great variety of environmental samples without any great knowledge about the sample matrix and its effects on arsenic absorption.  3.3.2 Determination of Arsenic in Environmental Samples by using HGAAS The first hydride generation method to gain acceptance, was that introduced by 75 This method employed a metal (Zn) / acid (HCI and/or 4 Holak.’ S0 reduction 2 H ) mixture to generate hydrides. A number of other metals, such as, aluminum powder, 176 or magnesium metal and titanium(III) chloride, have also been used to generate arsines in 48 This metallacid reduction technique has a number of J 57 conjunction with an acid. disadvantages, which have limited its current use. These include: slow reaction times, up to 20 mm, 57 and difficulty in automating the system because of the metal addition. As a consequence only As, Sb, and Se hydrides have been examined by using this method. HGAAS methods based on the production of hydrides by using sodium borohydride / acid reduction mixtures have been used with great success for the analysis of arsenic and other hydride forming elements since 1972.177 This reduction procedure has allowed for the automation of HGAA systems. In the present work a continuous flow system was used  78  (Figure 3.2). This system has been previously optimized for the analysis of arsenic by Cullen and Dodd. 178 The arsine generation reaction (Eq. 3.1) and the decomposition of the excess sodium borohydride (Eq. 3.2) occur simultaneously.  4 3) + BH As(OH 4 BH  +  +  0 + H 2 3H  H  >  2 B0 + H 3 AsH + H  (Eq. 3.1)  >  2 HBO + 4H  (Eq. 3.2)  Although many arsenic compounds can be reduced to their corresponding arsines, only arsenite, arsenate, methylarsonate, dimethylarsinate, and trimethylarsine oxide are readily converted to volatile hydrides upon reduction with sodium borohydride under acidic conditions. In order to use HGA.A to determine the total arsenic concentration in samples which contain arsenic compounds other than thOse mentioned above, a digestion method must first be used that is capable of converting all arsenic compounds into hydride forming arsenicals. In this work we have evaluated a number of digestion methods that are used to digest environmental samples prior to their being subjected to arsenic determination by using HGAAS.  3.3.2.1 Wet Digestion HGAA Determination of Arsenic Four mussels (Mytilus cahfornianus) were blended, homogenised and subsequently freeze dried. Four sub-samples of the freeze dried homogenate were wet digested and analyzed for arsenic by using HGAA. The result obtained was 12.9 ± 0.6 ppm arsenic. Four more sub-samples of the mussel homogenate mentioned above were also wet digested and subsequently treated with 1% KI 15 mm prior to the HGAA analysis. The concentration of arsenic present was found to be 12.8 ± 0.5 ppm. The iodide in this case was added in order 4 reduction step. Several reports have to prereduce As(V) to As(III), prior to the NaBH  79  shown that As(V) usually exhibits lower sensitivity in the HG process than does As(III), 179-181 probably because of their different rates of reduction. These differences are especially pronounced when peak heights are compared. Peak area does not, seem to be affected by the oxidation state of arsenic in the solution. Our results indicate that the addition of KI is not required for the complete recovery of arsenic in mussel samples. This is most likely a consequence of the measurement mode. Hydrides were continuously generated and swept into the optical path of the spectrometer, and measurements were taken only after the absorbance signal had stabilized. These mussel samples were also analyzed by using neutron activation analysis. This method gave an average result of 11.4 ± 0.4 ppm arsenic. Standard reference materials were also analyzed. After wet digestion, DOLT-i was found to contain 9.4  ±  0.5 ppm arsenic, the certified value for this material is 10.1  DORM-i was also analyzed and was found to contain 15 value of which is 17.7  ±  ±  ±  1.4.  1 ppm arsenic, the certified  2.1 ppm arsenic.  Wet digestion HGAA was employed for the analysis of byssal threads from Californian mussels. The byssal threads were found to contain 3.2 ppm arsenic. From the results reported above it can be concluded that the wet digestion HGAA method used for the analysis for arsenic provides excellent recoveries of arsenic from biological samples of marine origin.  3.3.2.2 Microwave Digestion HGAA Determination of Arsenic SRMs (0.2-0.4 g sample size) were microwave digested and then analyzed for their arsenic content by using HGAA. A variety of reagent mixtures were used to convert the organoarsenicals present in SRMs into forms which could be readily reduced to hydrides and subsequently detected by using HGAA. Table 3.8 lists a number of different reagent combinations tested in this study, along with the arsenic recoveries obtained.  80  Table 3.8 Recovery of arsenic from DOLT-i and DORM-i Arsenic recovery Duration of Volume (mL) of reagent used HNO  H9SOd  H,O,  radiation (sec)  6  DOLT-i  (%)  DORM-i  120  20.3  10.1  1  1  60  20.1  14.4  2  2  120  23.9  15.4  2  2  i20+120  24.2  15.6  1.5  0.5  90  27.3  17.1  1.5  1.0  90  46.8  28.6  1.5  1.5 1.0  90  52.5  36.2  90  47.1  30.7  1.0  i.0  As can be seen from these data, this procedure generally resulted in very low arsenic 0 increased the recoveries, perhaps 4 S 2 recoveries from the SRMs. The addition of H indicating that larger amounts of the sulfuric acid are needed for the complete decomposition of any organoarsenicals. We did not pursue these conditions further because of the possibility of pressure build-up in the bomb which could have hazardous effects: there was also a possibility of melting the Teflon® digestion vessel. The highest arsenic recovery obtained for DOLT-i was 52.5%, while for DORM-i it was 36.2%. Compared with the wet digestion HGAAS and microwave digestion ETAAS procedures, this method results in very low arsenic recoveries. Apparently the experimental conditions used for the microwave digestion were not rigorous enough to decompose the organoarsenicals. This indicates that a substantial amount of arsenic present in the dogfish flesh is in a form which is not easily reducible, which is in agreement with other studies that have reported 89% of the arsenic in 2 a compound that has been found to be very arsenobetaine, 8 ’ DORM-i to be present as 48 82 experienced similar low recoveries of arsenic when difficult to decompose. Raptis et al.’ attempting to decompose fish tissue extracts with nitric acid. Studies on Mytilus edulis (Blue mussel) have also reported low arsenic recoveries obtained by using this method: 1.5  81  ppm arsenic (wet weight) when analyzed by using microwave digestion-HGAAS and 2.8 ppm arsenic analyzed by wet digestion-HGAASJ 83 The excellent arsenic recoveries obtained after analyzing the microwave digestate by using ETAAS reinforce the idea that incomplete decomposition is the reason for low recoveries when using HGAA.  3.3.3 Comparison of Methods used for the Determination of Arsenic in Environmental Samples A typical calibration plot for arsenic, obtained by using ETAAS is shown in Figure 3.5. Linearity is typically observed in the range of 0.2 to 3.0 ng arsenic. The absolute limit of detection (LOD), defmed as the amount of analyte giving a signal equal to the blank plus three times the standard deviation of the blank, was determined to be 0.1 ng arsenic. The relative standard deviation (RSD) of twenty injections of 0.4 ng arsenic was calculated to be 5.6%. In most cases 20 iL sample solutions were injected into the graphite furnace.  0.32 0.28 ,-...  0.24 r=0.999 b=8.2 a=0.015 S y/x =0.006 Sb = 0.3 S & =0.006  0.20 ‘  0.16 0.12 0.08  0.04 0.00 0.00  0.01  0.03  0.02  0.04  0.05  Arsenic Conc. (p.p.m.) Figure 3.5 Typical calibration plot obtained by ETAAS.  Dotted line indicates 95%  confidence limits; r: correlation coefficient; b: slope; a: intercept; 5: standard deviation  82  0.12 0.10 ‘  0.08  0.06 0.04 0.02 0.00 0.00  0.01  0.02  0.03  0.04  0.05  Arsenic Conc. (p.p.m.) Figure 3.6 Typical calibration graph obtained by HGAA. Dotted line indicates 95% confidence limits; r: correlation coefficient; b: slope; a: intercept; S: standard deviation  In HGAA, approximately 5 mL of sample solution is required for each determination. This method has a LOD of 0.5 ng.m1 1 for arsenic, or an absolute LOD of approximately 2.5 ng arsenic. A typical calibration plot for this method of arsenic analysis is presented in Figure 3.6. These plots are typically linear up to 100 ng.mL arsenic. The RSD of twenty  determinations of a 20 ng.mL 1 arsenic solution was 3.1%. In comparing the two methods that are used for the analysis of arsenic in environmental samples, their advantages and disadvantages are apparent. ETAAS exhibits a  much better absolute LOD (0.2 ng) and thus is best suited for the analysis of samples for which only small volumes of solution are available. HGAA has a much better LOD in terms of arsenic concentrations (0.5 ng.mLl), and thus is better suited for samples of larger volumes (>15 mL). One of the main advantages of HGAA is its inherent ability to eliminate most interferences, since the arsine products are separated from the sample matrix.  83  Interferences are a major problem for ETAAS, thus the experimental conditions used for the analysis must be carefully optimized in order to remove such problems. 184 or to a digestion method e.g. HGAAS can easily be coupled on-line to HPLC 64 This is a very difficult procedure to apply to ETAAS and has not microwave digestion. been used to any great extent. There are major limitations on the chromatographic 73 conditions that can be used, because of the requirement for very low flow rates. The main disadvantage of using HGAAS to determine total arsenic in marine samples is that all organoarsenicals must be decomposed to arsenic compounds which are readily reduced to volatile arsines. 185 As has been demonstrated in this work, this decomposition is not always easily accomplished, and vigorous digestion methods are necessary. ETAAS is not as dependent on the digestion procedure used, because the ashing stage of the heating procedure acts as a very efficient digestion step. Several other analytical techniques have also been used for the analysis of arsenic. 6 is now a widely used (ICP-MS) 8 J Inductively Coupled Plasma-Mass Spectrometry 48 technique and is similar to ETAAS in that it only requires sample dissolution prior to analysis. The two techniques differ in relative ease and cost of operation, and the types of interferences  encountered. 1 72,173  Other less sensitive analytical techniques such as  Neutron Activation Analysis, and ICP-atomic emission spectroscopy, have also been 135 employed to determine total arsenic.  3.3.4. Arsenic Concentrations in Californian Mussels Collected from the B.C. Coast Bivalve mussels have been used in many marine environmental programs, as in situ bioindicators of trace metal exposure. However, it is very difficult to assess the origin of arsenic, as being natural or anthropogenic, by merely determining the total arsenic content of the animal.  84  The main purpose in analyzing marine animals for their arsenic content is for regulatory purposes, because these animals constitute a major human food source. Such determinations enable estimations to be made of arsenic uptake by humans. For example, the daily intake of arsenic in Japan is estimated to be 70-170 jig. day, 87,188 while it is 71 60 j.tg.day in countries such as Canada, UK, USA, and France. 189 The arsenic concentrations in the Californian mussel collected from a variety of locations are listed in Table 3.9. The results were obtained by using the two digestion and detection methods discussed in detail in this Chapter. All concentrations are quoted on a dry weight basis. Three replicate analyses were carried out for each sample.  Table 3.9 Concentration of arsenic in Californian mussels Location on British  No. of mussels  Columbia Coast  pooled  As (jtg.g ) 1 Wet Dig.  -  Microwave Dig.  HGAA  ETAAS  5  14.1± 0.5  14.5± 0.5  5  12.9± 0.6  13.1± 0.5  Anthony Island  6  9.6± 0.6  9.2± 0.4  Langara Island  6  10.1 ± 0.3  10.9 ± 0.5  Tofino  7  15.2 ± 0.6  15.6 ± 0.6  China Beach  4  11.4 ± 0.5  12.2 ± 0.6  Quatsino Sound  -  (Koprino Bay) Quatsino Sound (Varney Bay)  To the best of our knowledge these are the first reports of arsenic concentrations in Californian mussels, thus the levels can only be compared with those of other marine animals. In general the values in Table 3.9 are close to the average values reported for other marine species. Blue mussels (Mytilus edulis) for example, are found to contain 1.8 ± 0.4 78 This is equivalent to approximately 14 -16 j.tg.g’ 1 arsenic, on a wet weight basis.’ .tg.g arsenic, on a dry weight basis, It should be noted that the highest arsenic values came from  85  mussels originating from Quatsino Sound close to a copper mine, and from Tofino, an area where anthropogenic arsenic input is considered very low. Arsenic distributions in the various tissue parts of the Californian mussel were also investigated in order to determine whether any particular mussel organ concentrates arsenic preferentially. For this analysis, five mussels were dissected and their tissue parts were bulked together. Table 3.10 shows the data obtained for such determinations.  Table 3.10 Arsenic distribution in Californian mussel tissue parts Tissue Part  ) 4 As (ig.g  Muscle  7.0±0.9  Visceral mass  12.5  Gill  10.4±0.6  ±  05  The highest arsenic concentration was found in the visceral mass of the mussel. Studies on Mercenaria mercenaria (clam) and Mytilus coruscum have shown similar 28 showed that the gill of the clam 0. In contrast Shiomi et al., 50 distributions 9 arsenic ” Meretrix lusoria contains higher arsenic concentrations (21.5 ig arsenic g fresh tissue) 1 fresh tissue). Similar trends were reported by than other tissue parts (>5.4 jig arsenic g ° for the same clam species. 5 Shibata and Morita  86  3.4 SUMMARY Simplex optimization was used to efficiently delineate the optimum experimental conditions to be used for the ETAAS analysis of arsenic in a SRM of marine origin, which had previously been microwave digested. Four experimental variables, were considered: ashing temperature, atomization temperature, modifier concentration, and atomization ramping time. This combination of methods and materials provides a powerful means of rapidly improving the experimental conditions used for the analysis of arsenic in a wide variety of samples of environmental origin. Excellent recoveries of arsenic were obtained when using the optimum electrothermal atomic absorption spectrometry conditions to analyze standard solutions of arsenobetaine, arsenocholine and tetramethylarsonium iodide. Other methods used to determine arsenic in environmental samples of marine origin were evaluated. HGAAS showed very good arsenic recoveries when used to analyze SRMs which had been previously wet digested. Microwave digestions (acidIH ) followed by O 2 HGAA analysis did not offer acceptable arsenic recoveries. Finally Californian mussels collected from the B.C. coast were found to contain arsenic in the range 9 to 16 .tg.g 4 dry weight basis.  87  CHAPTER 4  DEVELOPMENT OF HPLC AND MS METHODS FOR THE SEPARATION AND DETERMINATION OF ARSENIC ANIMAL FEED ADDITIVES AND THEIR METABOLITES  4.1  INTRODUCTION  4.1.1 Scope and Rationale of Work The work reported in this Chapter is primarily concerned with the development of analytical methods capable of separating and determining organoarsenicals used in animal feeds, as well as identifying their metabolites. Currently four arsenicals are approved as feed additives for domestic animal production: 3-nitro-4-hydroxyphenylarsonic acid (3NHPAA), p-arsanilic acid (p-ASA), 4-nitrophenylarsonic acid (4-NPAA), and  p  A). The development of analytical methodology is (p-UPA93 ureidophenylarsonic acid 1911 necessary in order to provide information about these compounds that will enable us to understand their interactions and fate in the environment. In addition these methods would allow us to identify and quantify arsenic feed additives and their metabolites. This is of particular interest since there is very little information in the scientific literature, regarding the arsenic compound interactions and/or their metabolic fates in the environment. In addition analytical procedures developed here could be utilized for pharmacological investigations. So far only a number of methods have been developed for the analysis of this class of compounds. Most of these, however, are for target analysis of a few specific compounds  88  and are not suitable for the detection of possible metabolites. For example gas chromatography (GC) with flame ionization detection (FID), which has been used for the determination of p-UPAA and p-ASA, is not suitable for the determination of 3-N}{PAA and is also interference prone. 194 A spectrophotometric method for the determination of p UPAA in Carbarsone (proprietary formulation containing 33.6% w/v p-UPAA) has also been reported. 195 This method involves a coupling reaction with N-1-naphthylethylenediarnine. The coloured product that forms is extracted with butanol and subsequently measured photometrically; again interferences can cause difficulties in this determination. Thin layer chromatography (TLC) has been used for the separation and identification of 3N}IPAA, p-ASA, 4-NPAA, and p-UPAA, using colouring reagents for the visualization.’ 96 HPLC has also been used to separate some arylarsenicals, but not specifically those used as animal feed additives. 1 78,197,198  HO-  3-NHPA4  p-ASA  4-NPAA  p-UPAA  Figure 4.1 Structures of arsenic compounds used as animal feed additives.  The four compounds used commercially as feed additives are very polar. Thus, in this work NPLC was selected for study as being probably the most efficient method for achieving their separation, while GC was not considered because of the low volatility of the compounds. Suitable derivatization could improve their volatility but may not allow for the  89  detection of possible unknown metabolites which may occur. Micro-HPLC separation was also evaluated and compared to other conventional and microbore-HPLC systems. The micro-LC columns used in this work were constructed and packed in our laboratory as part of continuing studies on the applications of these techniques, which appear to have many 199 Ultraviolet (UV), ETAAS, Thermospray MS, Liquid secondary ion mass advantages. spectrometry (LSIMS) and dynamic or continuous flow (CF) LSIMS were used as detectors in conjunction with HPLC both in on-line and off-line modes.  4.1.2 Arsenicals used as Animal Feed Additives The use of arsenic in modem human medicine started in 1907 when Dr. P. Erlich discovered Salvarsan. ° This arsenic compound was used as an anti-syphilis drug and 20 related arsenicals were developed for other medicinal applications. In 1930 Morehouse et 201 were able to identifj copper arsenite as the active ingredient present in poultry al. drinking water medication used for the control of coccidiosis. This marked the beginning of studies of inorganic arsenic compounds with regard to their ability to control coccidiosis; however these compounds proved to be too toxic for any practical application. These results directed researchers back to the organoarsenicals that had previously been used in human medicine. In 1945 Morehouse showed that 3-NHPAA was capable of controlling cecal coccidiosis in poultry. 202 While investigating the therapeutic performance of this compound it was also discovered that it acted as a growth promoter, providing growth stimulation, improved feed conversion, better feathering, increased egg production and pigmentation. In the following years a number of other compounds were shown to have similar properties, p-ASA, 4-NPAA, p-UPA.A and benzenearsonic acid. With the exception of the last one, all of these compounds are still in use today. Variation of the substituents on the aromatic ring results in differences in the growth-promoting and disease-controlling effects of the compounds. Thus, 3-NHPAA and p-ASA are approved as animal feed  90  additives for both poultry and swine, whereas 4-NPAA and p-UPAA are approved only for ’ -193 19 controlling blackhead disease in turkey. Recent studies however, have disputed the beneficial effects of these compounds as growth promoters. p-Arsanilic acid was found to be ineffective in improving the biological or economic performance of broiler chicks when added to their diet at the maximum ° This was also the case when 3-NHPAA was supplemented 12 permissible level in Canada. 203 in the diet of growing-finishing pigs. A number of reports have pointed out the absence of microscopic changes in tissues from animals whose diets had been supplemented with these compounds, within the allowable levels. 120,203 If these feed additives are used it has been well documented that, shortly after the arsenicals are removed from the animal diet, they are rapidly metabolized and depleted from the animal and within five days the levels of arsenic in the various animal tissues return to background levels. Thus the use of the arsenical containing feed must be discontinued five days prior to the slaughtering of the animal. A number of reports have pointed out that these additives can cause toxic effects when used at levels higher then the recommended levels. The maximum limit for 3-NHPAA is 50 mg per kg of feed, while the recommended level is 37.5 mg per kg. Reports show that 400 mg of 3-N}IPAA per kg of feed when given to weanly pigs for 4 weeks resulted in 65% 204 200 mg of 3-NRPAA per kg of pig feed did of the animals showing neurological signs: 205 Even 105 mg of 3result in weight gain; however, muscle tremors were also observed. 206 NHPAA per kg of pig feed resulted in toxic effects. In order to evaluate the physiological role of these compounds as well as to assess their impact on the environment a number of areas need to be monitored for arsenic content and arsenic species. Of main concern are: the producer who is exposed to the additives in the feed mill, animal excreta used as a fertilizer and as a nutrient source in animal feeds, and  91  finally the effect on the human population both from a meat consumer point of view as well as from considerations involving sources of drinking water (aquifer).  4.1.3 Micro HPLC used for the Separation of Arsenical Animal Feed Additives -  A great deal of research effort has gone into the development of microscale HPLC. This began in 1967, when Horwáth and co-workers used a 1 mm ID and 2 m length column tides. Further development came in 1973, when Ishii and co-workers 207 to separate nucleo successfully separated polynuclear aromatic hydrocarbons on a 0.5 mm ID x 150 mm length n. There are a number of advantages associated with this particular 208 colum chromatographic technique and these are the main reasons for continuing research in HPLC column miniaturization. Some of the advantages are the following: First of all, the low consumption of both stationary and mobile phases, which is a direct consequence of the decrease in column dimensions, allows the use of expensive packing materials, and exotic mobile phases, such as deuterated solvents. The technique allows for a dramatic reduction in solvent waste, which is a serious concern since today it costs more to dispose of some solvents than to purchase them. Another advantage of the method is the increase in mass sensitivity when using a concentration sensitive detector. This is a consequence of the reduced dilution of the sample injected onto the LC column, compared with that in conventional HPLC. Finally, the one feature that made microscale ITPLC especially appealing for the present application, is the ease with which it is coupled to a mass spectrometer. The latter is a direct consequence of the extremely low solvent flow rates employed in micro-HPLC, typically in the range of 5  -  20 i.tL.min for columns, that can be adequately handled by  most MS systems. Microscale HPLC columns can be classified into three categories depending on the physical type of stationary phases and the packing states of the columns. Table 4.1 lists the  92  types of columns together with their specific characteristics. As can be seen the main difference between densely packed and loosely packed columns is their column-particle ratio (d/d). Open-tubular columns are unique in that their stationary phase is a coated or chemically bonded thin film on the inner surface of the capillary. Densely packed columns can be further classified according to their column inner diameters into three categories: conventional HPLC with 4-6 mm ID columns, semi-micro (or microbore) HPLC with 1-2 mm ID columns, and micro-I{PLC with 0.2-0.5 mm ID. This classification is employed in the remainder of this Chapter. Most of the work presented here has been conducted on micro-HPLC columns (0.32 mm ID), densely packed with 3 or 5 p.m diameter particles.  C 921 HPL 19920 Table 4.1 Classification of Microscale ° Common names of different types of microscale HPLC Densely packed  Differences in  Column LD.  Particle  (jim)  dId  kId  diameter (jim) Id1  250-1000  3-30  50-200  50-200  10-100  2-8  10-60  Coated film  columns  columns  Loosely packed columns  Open-tubular (capillary) columns  93  packing state of  4.1.4 Interfacing Liquid Chromatography to Mass Spectrometry As mentioned in detail in Chapter 2, MS is one of the most widely used analytical techniques used for the detection and structural elucidation of organic, inorganic, and organometallic compounds. Thus it is highly desirable to couple this extremely powerful detector to a separation technique as efficient as HPLC. For over twenty years combined gas chromatography  -  MS has been applied with great success to the analysis of volatile  compounds. Consequently much effort has been devoted to applying combined LC-MS to the analysis of polar and non-volatile compounds. The primary difficulty encountered in the early development of LC  -  MS interfaces  was the major difference in the operating environments of HPLC (liquid under pressure) and MS (sample and eluent vapour under high vacuum). As an example, consider an aqueous reversed-phase LC eluent flow at 1 mL.min’ which generates 1-4 L. In order to overcome such experimental difficulties many different types of LC  -  MS interfaces have  been developed, some of which have become commercially available. The first report of a commercial LC-MS interface was a moving wire interface, in which the LC effluent was deposited on a heated wire. After the removal of the solvent the 2. A similar sample residue was transported into the ion source of the mass spectrometer device, called the moving belt interface (MBI), was more successful and is still commercially available. 212 Eluent is sprayed onto a moving belt, via a heated nebulizer. The belt then passes under infrared heaters and subsequently through a series of vacuum locks, removing all the solvent. Once the sample reaches the ion source it is flash evaporated off the belt and into the ionization chamber. This technique is compatible with El, CI, and FAB ionization modes, and can handle non-volatile buffers. Its main disadvantages are sample carryover (memory effects) and mechanical complexity. Another very interesting approach to solving the LC-MS interface problem is the 213 and Baldwin Direct Liquid Introduction (DLI) method, pioneered by Tal’ Rose et a!.  94  86 Typically, a solvent jet is formed by passing the liquid at a flow rate of and McLafferty. 5-20 i.tL.min 1 through a 2-5 gim diameter hole in a replaceable diaphragm located at the probe tip placed in the ion source. The ionization involves chemical ionization from the mobile phase vapour. This method is restricted to low liquid flow rates, and can only employ volatile buffers. These low flow rates are of course directly compatible with micro HPLC methods suggesting a coupling of micro-HPLC and DLI. This is the approach we have pursued in order to interface our micro-HPLC columns to a mass spectrometer. ce vaporizes 14 a interfa 2 , 4 l In contrast to the DLI interface, the thermospray (TSP) 2 the LC effluent just prior to its introduction into the ion source. This, in addition to auxiliary . A 1 pumping, makes sampling possible at high solvent flow rates, up to 2 mL.min resistively heated stainless steel metal capillary is used to produce a jet flow, which consists of fine mist droplets of column eluent. This jet is carried at a high speed into the heated ion source, where it continuous to vaporize. The mobile phase contains volatile buffers Ac, TFA). The buffer ions form charged droplets which undergo evaporation and (NET O 4 lead to production of sample ions in the gas phase. Sample and buffer ions are directed into the mass analyzer, while the excess vapour (both solvent and sample) is pumped away by an auxiliary vacuum pump. The development of this method has reached a plateau and is now in the hands of users throughout the world, as demonstrated by the significant increase in 215 ’ 74 the number of application reports in recent years. Dynamic or CF FABJLSIMS, developed in the mid 1980s, has shown great promise interfacing. 6,217 The technique is well suited for the analysis of large, polar, ’ for LC-MS 2 and/or thermally labile molecules. Flow rates of approximately 5 iL.min are used with 110% matrix (glycerol, nitrobenzylalcohol, etc.) incorporated into the LC eluent. A flit is normally placed at the end of the LC transfer line situated inside the ion source. This allows for sample dispersion and also provides for an impact area upon which the FAB/LSTMS beam can be focused.  95  Electrospray (ES) I Ionspray MS is the most recent technique which functions as both an LC-MS interface and ionization source. 219 ES operates in the continuous flow (CF) ’ 218 mode, as does CF-FAB, thus the sample is introduced and ionized in the liquid phase. This method has undergone tremendous growth in recent years, and is evolving into the dominant LC-MS interface for a variety of applications. The basic principle of operation of this method is the formation of a fme spray of charged droplets in a strong electrostatic field at the exit of a narrow capillary through which the solution of the analyte passes. Subsequent evaporation of the solvent allows for the formation of single and multiply charged analyte ions. The exact mechanism of ion formation by ES is still under intense discussion and investigation. 222 220  4.2  EXPERIMENTAL  4.2.1 Instrumentation 4.2.1.1 HPLC pumps Two different types of pumps were used in these experiments. A Waters 510 reciprocating pump was used for the conventional HPLC columns. A dual syringe pump (Applied Biosystems, Model 140B) was used to provide the appropriate mobile phase gradient and flow rate for the microbore and micro columns. This system consists of two independently-driven syringe pumps with a volume of 10 mL each. A 200 iL mixer, used to form gradients, is also part of the system. This pump is capable of 1 delivering pulse-free solvent flows at rates as low as I iiL.min . Pulse-free mobile phase 4 flows are extremely advantageous for obtaining stable ion beams for the CF-FABILSIMS and ES methods.  96  4.2.1.2 Injectors  Samples were introduced onto the conventional column via a 20 tL stainless steel loop injector (Rheodyne 7125). This dual-mode injector was operated in the completefilling mode, in order to improve accuracy and precision of the sample volume loaded onto the column. A 0.5 jiL internal-chamber micro injector (Rheodyne 7520) was used for loading samples onto microbore and micro columns.  4.2.1.3 HPLC Columns  8 Spheri-5 RP-18 (Brownlee Labs) C The conventional HPLC column used was a 1 with the following dimensions: 100 mm length x 4.6 mm ID, containing 5im diameter packing material. 8 Spheri-5 RP-18 (BrownLee Labs), C Microbore separations were achieved on a 1 250 nmi length x 1 mm ID, containing 5 I.m diameter packing material. The micro columns in a variety of lengths (8, 14.5, 20 cm length) and 0.32 mm ID 8 packing material (Spherisorb ODS2, 3 and 5 C were fabricated in-house and packed with 1 .tm diameter).  4.2.1.4 HPLC Detectors  For the initial evaluation of the LC columns and mobile phases a UV detector (Waters, Lambda-Max Model 481) equipped with a 1 .tL flow cell was used to monitor the separations, and allow for optimization of their conditions. A Kratos Concept II H mass spectrometer was used for the LSIMS and CF-LSIMS studies. For the DLI-MS experiments a Nermag RiO-lOB quadrupole mass analyzer, equipped with a desolvation chamber and a cryopump, was used. A Kratos MS 80 mass spectrometer was used for the TSP-MS.  97  The ETAAS system, used for the analysis of the LC fractions for arsenic, is described in detail in Chapter 3 of this Thesis.  4.2.1.5 LC-MS Interfaces In order to perform micro-LC-MS experiments three interfaces were evaluated, a CF LSIMS probe and two different DLI couplings. The CF-LSIMS probe (Figure 4.2), located inside the high vacuum ion source chamber, delivers the mobile phase via a quartz capillary tube to a mesh screen at the tip of the probe. The quartz capillary is in firm contact with the mesh, thus allowing the liquid to spread over the mesh by vacuum and capillary actions. A porous steel body in contact with the mesh acts as an absorbent for the excess solvent, and thus prevents potential peak broadening caused by the probe tip. The latter is in contact with the heated ion source block allowing for efficient heat transfer to the mobile phase, thus leading to its stable rate of evaporation. This feature is critical for maintaining a stable ion source pressure and ion currents.  Cs + beam  Fine wire mesh Porous metal body Liquid delivery line  Ion extraction to mass analyzer  I__ High voltage  SS probe shaft  insulator  Figure 4.2 Kratos continuous flow LSIMS probe and Cs ion beam.  Two DLI interfaces were used in these experiments (Figure 4.3). The first DLI probe (Fig.4.3A) was equipped with a nickel diaphragm having a 2-5 .tm orifice. A liquid jet was  98  formed by forcing the mobile phase to pass through this orifice. The probe was inserted into a heated desolvation chamber attached to the ion source. This allowed for the subsequent desolvation of the liquid jet droplets. The second DLI interface (Fig. 4.3B) was constructed in-house. This interface is similar in principle to a thermospray nebulizer, the main difference being its micro dimensions. A quartz capillary (10 urn ID) is heated, thus converting the liquid mobile phase into a vapour jet, which is introduced into the mass spectrometers ion source. For both DLI interfaces an electron beam was used to generate ions for subsequent mass analysis.  4.2.2 Chemicals Stock solutions of 1000 ppm of each of the following organoarsenicals were prepared: 3 -nitro-4-hydroxyphenylarsonic acid (ICN Biochemicals, Cleveland, OH 44128), p-arsanilic acid (EASTMAN Organic Chemicals, Rochester 3, NY), 4-nitrophenylarsonic acid (Aldrich Chemical Co., Milwaukee, Wisconsin), and 4-hydroxyphenylarsonic acid (EASTMAN Kodak Company, Rochester, NY). All solvents (HPLC grade) were filtered and degassed prior to their use as mobile phase by using a 35 mm all glass filter holder fitted with an HA 0.45 urn filter (Millipore) for the aqueous solvents and a FH 0.5 urn filter (Millipore) for the organic solvents.  99  A. Mobile phase  /  Probe  source  I ‘1  Diaphragm  Heated desolvation chamber  B. Heated Cu sleeve with thermocouple SStube  Mobile phase  sffiea  pillaiy  acu  Figure 4.3 DLI nebulizers; A. diaphragm nebulizer with desolvation chamber, and B. heated capillary nebulizer.  100  4.2.3 Procedures  4.2.3.1 Fabrication and Packing of Micro HPLC Columns The fabrication procedure for the micro LC column is schematically outlined in Figure 4.4. Fused silica columns [Al (lIP-ULTRA 2) with internal diameters of 0.32 mm were cut to various lengths. These columns were used as the main body of the HPLC column. A 1 cm plug of silanized glass wool was inserted into the column, 1 cm from the end. This was achieved by slipping a 2 cm piece of Teflon tubing over one end, and using a piece of quartz tubing (0.290 mm O.D.) as a pushing rod to insert the glass wool. Fused silica tubing (B] (0.050 mm I.D., 0.143 mm O.D., Polymicro Technologies) 50-60 cm long was used for coupling the micro-HPLC column to the detector. In order to connect this line to the micro HPLC column (A] a piece of fused silica tubing (C] (0.181 mm I.D., 0.290 mm outer diameter, Polymicro Technologies) 2 cm in length was placed over [B], and cemented into place by using glue (Lapage #8). [B]+(C] were then inserted into (A] up to the glass wool plug. Glue was also used to keep these parts together. In addition a piece of glass tubing (D] 4 cm in length (Accupette pipets) was used to support this junction, by placing it over the connections and gluing it into place. Prior to packing the constructed column was allowed to dry overnight. In order to pack the micro-LC column a packing reservoir was needed, as shown in Figure 4.5. Silica based C 18 packing material (Spherisorb 3 or 5 jim ODS2) was used to prepare a slurry by sonicating 3 0-40 mg of the material in 1 mL of carbon tetrachloride. The slurry was then transferred into the packing reservoir by means of a dropping pipette. The mobile phase (acetonitrile / water 6:4) not mixable with the carbon tetrachioride was filtered and degassed prior to use. In order to achieve packing of the column, a pressure of about 5000 psi was exerted by the pump, while delivering the mobile phase. The packing of the column can be followed visually to completion.  101  [A]  [B]  [C] IEEE  I  : Glue (Lepage #8)  //  : Silanized glass wool  [A] [B+C]  +  Glue  -//  -  1  I  [D]  [A+B+C]  I  Figure 4.4 Fabrication of the micro-HPLC column.  102  I  Tubing  1/16’ SS Coupling  -114’ SS Coupling 18 Spheiisorb 40 mg of 3 pm C 00S2 Paddng Materiai hi 1.5 mL of Carbon tetracNoride  Packing Reservoir  .114’ SS Tubing  (High Pressure)  40% Graphatized Vespel® Ferrule 1/16’ X 0.4 mm ID 0.32 mm ID Capillaiy Column  50 pm ID Quartz glass transfer line  Figure 4.5 Schematic diagram of the apparatus used for packing micro-HPLC columns.  4.3  RESULTS DISCUSSION -  A number of combinations HPLC techniques and detectors were evaluated in this work for use in the analysis of the animal feed arsenicals (Figure 4.6). Since we were interested in comparing the performance characteristics of various HPLC columns of different diameters and flow rates, it was necessary to evaluate the column and detector as one system, rather than each component separately. The main reason for doing this, is that a detector may influence the separation achieved on a particular type of column. For example, the separation efficiency of a micro column will be greatly degraded if it is  103  connected to a UV detector equipped with a larger volume 14 .iL UV cell, rather than a 1 tL UV cell, while the separation efficiency of a conventional column will not be affected if connected to a detector with a 14 j.iL cell. In general when coupling all components in a micro-LC system it is important to avoid dead volumes, stagnation volumes, etc.  DLI-MS CF LSIMS -  Micro Column C-18 Syringe Pump Microbore Column  Reciprocating Pump  —  ::::::::  c-i 8  Conventional Column C-18  LSIMS UV  ETAAS  Thermospray MS Off-line coupling  On-line coupling Figure 4.6 Methods used for the separation and identification of animal feed arsenicals.  4.3.1 Separation and Determination of Arylarsenicals by using HPLC-Thermospray MS The separation of the compounds 3-NHPAA., p-ASA, and 4-NPAA on a conventional 8 column (5 p.m particle size, 4.6 mm ID x 100 mm length) was monitored by using UV C 1 detection. The optimization of this chromatographic separation was carried out by varying the methanol content of the mobile phase, and its flow rate. The resolution (R) 2 ti obtained from , t: retention time, Wb: base peak width], [R —  =  1  (Wb,1 +Wb,2) each experiment for 3-NHPAA and 4-NPAA, is presented in Figure 4.7 along with two representative chromatograms.  104  Thermospray MS was also used for the detection of these compounds. The [M+H]+ ion was the base peak in the mass spectra of all samples analyzed. Figure 4.8 shows the TSP mass spectrum of p-ASA, which is typical for all the arylarsenicals analyzed. The TSP mass spectra of the other arylarsenicals are included in Appendix A. TSP-MS provided little information concerning the analyte structure, but the molecular weight of the analyte can easily be determined. However, structural information via fragmentation can be obtained by operating the TSP ion source in the auxiliary filament ionization mode. In this mode the electron beam, which is produced by the filament, is situated in the TSP jet and can give rise to analyte fragmentation. This mode of ionization is also of interest when high contents of organic solvents are being used. In addition, a discharge can be generated in the source of some TSP-MS systems, leading to some fragmentation. Collisional-induced dissociation (CID) can also be used to generate characteristic fragment ions, if a MS-MS instrument is available. For this study only the TSP-MS was available. Operating temperatures are critical for optimum performance of TSP-MS. Best sensitivity is obtained when sufficient heat is provided to completely vaporize the sample, without causing it to pyrolize or change structure. Other temperature settings such as those of the vaporizer tip (probe) as well as the ion source temperature also influence the sensitivity of the detector. This feature of TSP-MS somewhat limits the utility of the method when it is used for the identification of unknown compounds, because its sensitivity is compound dependent and TSP ionization does not always take place. It has been our 223 that TSP-MS is a very useful technique for target experience, and that of others, analyses, but not as valuable for the identification of unknown compounds. This is because a set of four temperatures, all of which greatly influence the efficiency of ion formation, must be optimized for each compound. In searching for unknown compounds it is almost impossible to perform such an optimization.  105  1.60  A.  C  o 1. 0)  a,  v.94  a:  0.020  p-ASA  B.  0.015  C.  0.020  0.016  0.012 <0010  0.008  0.005  °L  0.000 ‘  2 46802 I I 46 I  0.000  0  Tune (Tnft)  2  4  I  6 8 10 12 14 16 TtAE (mm.)  Figure 4.7 A: Three dimensional graph showing the variation of resolutiop between 3NHPAA and 4-NPAA when the methanol content of the mobile phase and the flow rate are varied. B: UV Q=254 nm)-chromatogram of arylarsenicals (22.5 ng As of each compound), 1 mL.min’ flow rate, 95% water (0.1% TFA) and 5% methanol. C: UV 1 flow rate, 75% water (0.1% TFA) chromatogram (254 nm) of arylarsenicals, 1.2 mL.min  and 25% methanol. 106  The arsenicals studied here, both the animal feed additives and a possible metabolite (4-hydroxyphenylarsonic acid) of 3-NHPAA, give similar responses in terms of signal for the same amount of compound, when using the same set of TSP temperatures. However, the two calibration curves shown in Figure 4.9 exhibit slightly different slopes, thus the sensitivities for the two compounds vary slightly. However, the variation is not sufficiently large to raise concerns about widely different response factors for structurally related arylarsenical feed additives under the LC-MS conditions used here. The linearity of the curves (Figure 4.9) supports the possibility of quantitation within a limited dynamic range. Clearly, the use of a suitable internal standard would greatly improve the precision of the method. The LOD in TSP-MS for these compounds was calculated to be between 30 and 33 ng (1 .5-1.6 ppm) of arsenic, depending on the compound analyzed. The results obtained in this work indicate that TSP-MS is a very attractive technique for the analysis of arsenical animal feed additives. The non-volatile nature of the compounds has made their analysis quite difficult by using other conventional MS techniques, such as DCI (Figure 2.13) and El MS. This is due to the fact that the compounds pyrolyze in the ion source prior to their desorption, and thus afford mass spectra with no molecular or quasi molecular ions.  107  I,.  218 [M+H]  HO-As-OH  c  I,  7.  -..  C,)  C  a?  I.  C a?  4I  -..  a? 21  I.  M/Z  Figure 4.8 Thermospray mass spectrum of p-arsanilic acid. Temperature settings: Probe at 129 °C, vaporizer at 185 °C, source block at 223 °C, and jet at 120 °C. 1 .5e+ 007  1 .Oe+007 a, -x 0  a)  5.Oe+006  0  0.Oe+000  0  5  10  15  20  25  Amount of Arsenic (ng) Figure 4.9 Calibration plot obtained by using Thennospray-MS for p-arsanilic acid and 4nitrophenyl arsonic acid. The error bars represent the signal range obtained from three injections of the analyte.  108  8 HPLC Column C 4.3.2 Separation of Arylarsenicals on a Microbore 1 The first step towards the miniaturization of HPLC systems was the development of ns. We found that the arylarsenicals can be separated by using 08 207 colum 2 microbore LC ’ 8 1 mm ID x 250mm length) at a flow rate of 80 .tL.min , 1 a microbore column (Spheri-5 C 1 (Figure 4.10). The methanol content of the mobile phase has a pronounced effect on the resolution, as noted previously (Figure 4.7). Higher methanol percentage decreased the retention time of p-ASA and 4-NPAA, but also decreased the resolution. Good resolution was achieved by using 10-15% methanol. The separation efficiency of this column is quite similar to that obtained on the conventional HPLC column (Figures 4.10 and 4.7B). However, because of the reduced sample dilution on the microbore column, the LOD for p ASA is 0.22 ng of arsenic, when using UV detection, compared to 1.9 ng of arsenic on the conventional column for the same detector. Another advantage we observed is a considerable reduction in solvent consumption, leading to economic and environmental benefits.  8 IIPLC Columns C 4.3.3 Separation of Arylarsenicals on Micro 1 The elution characteristics of the arylarsenicals on the micro-LC columns (Figure 4.11) are similar to those observed on the conventional (Fig. 4.7) and microbore (Fig. 4.10) columns, both in terms of the overall time required for the separation and the efficiency of separation. As expected the LODs for UV detection were improved by one and two orders of magnitude respectively, over those obtained by using microbore and conventional HPLC columns, respectively. The LOD for p-ASA was calculated to be 0.02 ng of arsenic. Seven runs were carried out in order to evaluate the precision with which retention times can be measured while using the micro-LC columns. The relative standard deviation of the retention times was found to be 0.9 % for p-ASA, 1.2 % for 4-HPAA, 1.6 % for 3-  109  NHPAA, and 1.4 % for 4-NPAA. These results indicate that arsenical identification can be based on retention times if standards are available. In terms of solvent consumption, a typical arsenical separation requires 0.14 mL, 1.12 mL, and 14 mL of solvent respectively, on micro, microbore, and conventional LC columns. This means that a 99 % reduction in the volume of solvent waste is achieved by switching from conventional to micro-LC, and a 87.5 % reduction when switching from microbore to micro-LC. These percentages can translate into extremely large savings, in terms of both solvent consumption, and solvent waste handling and disposal. The use of other types of chromatography for the separation of arylarsenicals has also been reported. Maruo et at. 197 used a 0.5 mm ID column (ID between that of the micro and microbore columns used in this study) packed with an anion exchange stationary phase, to separate three arsenicals, p-ASA and two other compounds that are structurally similar to the arsenic animal feed additives; o-arsanilic acid, and o-nitrophenylarsonic acid. The resulting UV-chromatogram showed a separation that required over 40 mm  and broad  peaks that had baseline widths between 5 and 10 mm. Others have attempted to separate the 197 by using nonsame compounds on columns similar to those used by Maruo et at., suppressed i6n chromatography, which resulted in co-elution of o-arsanilic acid and o 198 Both these reports re-inforce the effectiveness of the LC nitrophenylarsonic acid. approach we have utilized in order to separate arylarsenicals, often referred to as ionsuppressed reverse phase LC. The micro-LC columns were also evaluated for their potential to separate an additional arsenical, 4-hydroxyphenylarsonic acid (4-HPAA), which can be regarded as a possible metabolite of 3-NHPAA, on the basis of structural similarity. Figure 4.1 lB shows that the compound is quite well resolved from its potential precursor.  110  p-ASA  4-NPAA 0  o  2  4  6  8 10121418  TtE (nIn3  Figure 4.10  UV (A=254 nm)-chromatogram of arylarsenicals (2.5 ng As of each compound); Microbore C 18 column, 80 iL.min 1 flow rate, 85% water (0.1% TFA) and 15% methanol.  A.  0.020  B.  p-ASA  0.016  0.012  e 0008  0.004  0.000 0  2  4  6  8 10 12 14 16  ThwlE ()  0  2  4  6  8 10 12 14 18  TRvE  &flifl.)  Figure 4.11 UV (=254 nm)-chromatogram of arylarsenicals (0.8 ng As of each 1 flow rate, 85% water (0.1% TFA) and 15% compound); micro-HPLC column, 10 iL.miir methanol. A: Compounds separated: p-ASA, 3-NHPAA, and 4-NPAA B: 4hydroxyphenylarsonic acid added to the analyte mixture of A.  111  4.3.3.1 Effects of Packing Procedure and Particle Size on the Quality of Micro-LC Columns Micro-LC columns fabricated in-house were packed with a 3 and 5 J.tm diameter spherical porous silica based  18 stationary phase. No noticeable difference in resolution  was observed for these two sizes of packings. Most HPLC applications use silica particles with a 5 or 10 jim size diameter, while 3 im particles, even though commercially available, are only slowly being accepted by the chromatographic community. The main reason for this skepticism is the requirement for spherical shaped particles which result in higher quality packed columns and which are difficult to prepare uniformly. The small diameter particles, which are now more readily available, seem to be spherical. The slurry packing procedure we adopted was evaluated in both the downward and upward direction, because it is not yet established which approach gives the better results. In our hands the two procedures produced columns with similar performance characteristics.  4.3.3.2 Effect of Column Length on the Elution Characteristics of the Arylarsenicals Column length has a profound effect on the separation efficiency of the micro-LC column. Table 4.2 summarizes the resolution (R) obtained on columns of various lengths for two pairs of arsenicals. It is evident that the longest column (200 mm) provides the best resolution, while the 145 mm long column also exhibits good resolution.  Table 4.2 Summary of micro-Column length vs resolution (R) obtained for arsenicals R (p-ASA I 4-HPAA) R(3NI{PAA / 4-NPAA) Column Length (mm) 80  unresolved  1.3  145  1.4  2.2  200  1.6  2.7  112  4.3.3.3 Effect of Temperature on the Elution Characteristics of the Arylarsenicals It is quite easy to control the temperature of the micro columns. In our case, this was accomplished by immersing the column in a water bath; a thermocouple was used to monitor the temperature. Separations were carried out over a range of temperatures, from 18 0 C to 57 °C and the effect of temperature is quite pronounced. Retention times are dramatically reduced (Figure 4.12), as are resolutions, when the column temperature is increased. This we believe is a consequence of an increase in the interaction of the methanol in the mobile phase with the analytes. As mentioned earlier (Section 4.3.1), a similar type of behaviour is observed at constant ambient temperature, when the methanol content of the mobile phase is increased. 20 0 p-ASA • 4—HPAA V 3—NFIPAA v4—NPAA  5 15 v 10  :  70 Column Temperature (°C)  Figure 4.12 Effect of temperature on the arylarsenical retention times.  4.3.4 Detectors used for Microscale HPLC Because UV detectors provide limited structural information about compounds eluting from HPLC columns, it is necessary to develop other types of detectors with this application in mind. In this work we have evaluated a number of detectors for their ability to provide structural and/or elemental information concerning the arsenicals used as animal  113  feed additives. For example, MS techniques such as continuous flow FAB and DLI can provide structural information, while element specific detectors such as ETAAS can be very useful for selectively detecting arsenic. All of these detectors have the potential to be used either off-line or on-line.  4.3.4.1 ETAAS used as a Detector for Microscaie-RPLC ETAAS was used in the off-line mode for the detection of arsenic compounds eluting from LC columns. To date there has not been a LC interface developed for ETAAS which has received general acceptance. 224 This means that fractions of the LC eluent must be collected and subsequently analyzed. This method has been extensively used in conjunction with conventional HPLC columns, 225 where the main disadvantage is the low sensitivity of the method due to dilution effects in the LC mobile phase. In addition chromatographic resolution is degraded. Both of these features make it very difficult to detect minor arsenic components in environmental samples. In this study we have compared the chromatograms obtained by using conventional, microbore, and micro HPLC coupled off-line to an ETAA detector. Fractions were collected every 30 sec from all three different types of column. This procedure resulted in fraction volumes of 500 1 .tL for conventional columns, 40 jiL for microbore, and 5 1 iL for micro columns. The resulting ETAAS-chromatograms are shown in Figure 4.13. The LOD for p-ASA on the three chromatographic systems was calculated to be 250 ng of arsenic for the conventional column, 2.5 ng of arsenic for the microbore column and 0.3 ng of arsenic for the micro column. Although the ETAAS detector is not concentration sensitive an improvement in LOD for the micro column, was observed. This is expected since all of the LC eluent is injected into the graphite furnace while only 2 % and 37.5 % of the eluent from the conventional and the microbore column are injected, respectively.  114  Retention Time (mm.)  o  2  4  6  8  10  12  14  16  18  o  2  4  6  8  10  12  14  16  18  0.3  0.2  0.1  oco 0.15  Q)  0.10 0.05  a.)  oco 0.3 0.2 0.1 0.0  Retention Time (mm) Figure 4.13 ETAAS-chromatograms of arsenicals, separated on a: A. conventional-LC column, 500 j.tL fractions collected, B. microbore-LC column, 40 pL fractions collected, C. micro-LC column, 5 pL fractions collected.  115  The micro column has the disadvantage that the eluent volume allows only for a single analysis of each fraction, instead of two or three as is the case with the other two LC systems. This translates into potential problems in the accuracy and precision of the analysis. Of all three LC systems used, the microbore column seems to be the most suited for this application since its has a good LOD, and also allows for replicate determinations of each fraction collected. However, it should be noted that the LODs obtained in this study can be easily improved by performing multiple injections of each fraction into the AA furnace. The drawback of course would be a dramatic increase in the time required to analyze all the collected fractions. 4.3.4.2 Liquid Secondary Ion Mass Spectrometry FAB-MS has been recognized as a powerful tool in the mass spectral analysis of polar, thermally labile and/or involatile compounds. The operating principle of the method has been described in detail by Barber et a!. P226 In brief; the analyte is dissolved in a viscous matrix of low volatility such as glycerol, and subsequently introduced into the ion source of the mass spectrometer, where it is bombarded by a beam of high energy (6-8 keV) neutral (> 80%) species (Ar, Xe). A similar technique termed liquid secondary ion mass spectrometry (LSIMS) uses high energy Cs+ ions, instead of neutrals. The two methods yield similar results, although some differences can result, since the primary ion beam is generally more focused compared to the neutral beam, is of higher energy, and usually has a higher density of particles. Thus, in general LSIMS tends to be the more sensitive method. 227 and in Spectra obtained in FAB and LSIMS are affected by the nature of the matrix 228 some cases strong matrix-analyte interactions exist in the solution. In this work we have utilized both LSIMS and continuous flow CF-LSIMS to analyze the animal feed arsenicals. LSIMS has the potential to be used as an off-line LC detector, while CF-LSIMS can be used on-line.  116  In LSIMS a variety of different matrices namely, 3-nitrobenzylalcohol (NBA), glycerol, and thioglycerol, were investigated for their ability to promote ionization of the arsenicals. The mass spectra of 3-NHPAA, in both negative and positive ion detection modes, are presented in Figure 4.14: the matrix was NBA. The main feature of the positive ion spectrum is the formation of [M+H]+ ions and protonated reaction adducts of the matrix and the analyte. Negative ion mass spectra generally appear 1 Da below the molecular weight of the compound being analyzed, because of the loss of a proton, which is required for negative ion formation, as opposed to a proton gain which results in positive ion formation. The triple ion pattern ([M+Hj+ ion, and protonated reaction adducts of the matrix and the analyte) observed in these spectra provides molecular weight information concerning the analytes. The general features of the mass spectra of all the arylarsenicals are similar to those observed for 3-N[TPAA in Figure 4.14. Appendix B contains the LSIMS mass spectra for these compounds. The use of CF-LSIMS to analyze the arsenicals is seen in the positive ion mass spectrum obtained for p-ASA (Figure 4.15). In addition to the [M+H]+ ion and the matrix adducts, a number of characteristic fragment ions are also observed. The mobile phase, used 0 2 to deliver the analyte at a flow rate of 5 iiL.min’ to the ion source, consists of 89 % H H, and 1% glycerol. These conditions were selected because they CH O (0.1% TFA), 10 % 3 satisfied both the requirements set forth by the micro-LC system used for the separation of the arsenicals and the LSIMS, which required the presence of a matrix-substance, such as glycerol, in order to perform the analysis. Micro-HPLC chromatography of the arsenicals under these conditions produced UV-chromatograms which showed few effects associated with the presence of 1% glycerol. However, some problems could be avoided by introducing the matrix solution after the LC separation has taken place (post-column mode of introduction). 228a,b  117  Al1TSOJ  •  SIT  (suo!  fl pu ‘UOnp UOT  XTJ1W  yg *) uOnp UOT  yydJ{s- jo uiods ssuz  1 swas’i vrv nth  Relative Intensity 0  -  — —*  --  : :  +  +  :1  —--—  I%  1  —  II’  +  :  1’  +  The concentration of glycerol matrix is much lower than that used in the conventional standard probe LSIMS conditions (described earlier in detail), and this considerably reduces the signal from matrix ions (chemical noise). Another advantage associated with the CF method is that with time the resulting spectrum changes through the cycle, pure matrix to sample plus matrix to pure matrix, as the sample elutes. Again this permits background subtraction, and thus provision of clean mass spectra. The p-ASA spectrum shown in Fig. 4.15 has been cleaned up in this way. The minimum detectable amount (M1)A) of p-ASA was determined to be approximately 10 ng of arsenic, under the experimental conditions used in this study.  i::  [M+H1  p-ASA  80 70 60 50• 40  +  [M+GIyc+H] .Hj4  20  0 2 -H  4  10-  [M.t-2G1yc+2H+MeOHJ  9’  I 100  [Cs] 13  I  291  I  .-f-4 150  200  232  I  MeOHL  ao  .2.. 254  417  300  250  i  292  350  400  j 450  Figure 4.15 Background subtracted CF-LSIMS mass spectrum of p-ASk mobile phase: 89% water (0.1% TFA), 10% methanol, and 1% glycerol (Glyc); flow rate; 5 tL.min.  119  500  4.3.4.3 Direct Liquid Introduction Mass Spectrometry -  DLI-MS is another mass spectrometric method with potential for coupling on-line with micro-LC. Two different nebulizers were used in this study. A heated capillary nebulizer was built in house, while a diaphragm nebulizer with an associated desolvation chamber was obtained commercially. Use of the diaphragm nebulizer (Fig. 4.3A) enabled the production of mass spectra in the negative ion detection mode. The main disadvantage of the method was its requirement for relatively large amounts of analyte, compared to the other mass spectrometric methods discussed in this chapter. Mass spectra were obtained with approximately 1000 ng of arsenic, in the negative ion detection mode. As can be seen from the mass spectra of 3NHPAA (Figure 4.16) and 4-NPAA (Figure 4.17), extensive fragmentation has occurred and the molecular ion is not the base peak as was the case with most of the other MS discussed methods before. This is in accordance with our previous observations that these arylarsenicals are thermally labile and decompose/pyrolize upon extensive heating. The heating in this case is provided by the desolvation chamber. To the best of our knowledge, this is the first reported use of this type of nebulizer for the analysis of organometallic compounds. The heated capillary interface (Fig. 4.3B) was also tested. No meaningful spectra of the arsenicals were produced when using this nebulizer. Again, this can be explained in terms of analyte decomposition, caused by the extensive heating which is required in order to convert the CF-liquid into vapour. Organic test compounds, used for the initial set-up of this system, were analyzed successfiully. These included, caffeine, naphthalene, and 2,6dimethyiphenol. The spectra of these compounds mainly consisted of [M+Hf ions (mass spectra of these compounds are presented in Appendix C).  120  [M-0HJ  lOOx  246  o  [Mi  II  .263  • -..  HO-As-OH  -..  C  :1  2 C1NO  a?  OH a? 25>: 213  thi 1111111111  208  233 H1L’ 1  I  !IL.II.I  221)  IiiiI..Li,iJ tiII  240  pi. 281)  260  .IlIiIll’  M/Z  Figure 4.16 DLI-MS of 3-NHPAA, mobile phase consisted of 90% water (0.1 %TFA),  and 10 % methanol, 20 1 2L.min’ flow rate, negative ion detection.  100>: -  213  [M-20HJ 0  II  .  HO-As-OH  75>: a?  C 50>:  [M-O] 231 [MJ 247  •  a?  150  r‘ 170  2 NO  u 4 .iuJk  190  218  231)  250  ,  270  ji_fl  flu,  291)  F-.  37  M/Z  Figure 4.17 DLI-MS of 4-NPAA, mobile phase consisted of 90% water (0.1 %TFA), 4 flow rate, negative ion detection. and 10 % methanol, 20 L.min  121  4.4 SUMMARY  The work described in this chapter deals with the development of analytical techniques which can be used for the analysis of arsenic compounds added into animal feeds. More specifically, chromatographic methods have been developed for the separation of these arsenicals. In particular micro-LC, which is a relatively new technique, shows a number of advantages for these analyses, especially when interfaced to MS. Improved limits of detection, low solvent consumption, lower use of stationary phase, as well as flow rates easily accommodated by CF-LSIMS and DLI-MS, are some of the advantages we have observed when fabricating, and using these columns. Off-line coupling of micro-LC to ETAAS improves the LOD for arsenic and opens the possibility of on-line coupling, an area which requires extensive research before it becomes reality. Additional advances could also be made in this area by coupling micro-columns to MS-MS systems or alternative detectors such as ICP-MS. These aspects of micro-LC remain to be evaluated for their effectiveness in analyzing not only arylarsenicals but also arsenic compounds encountered in the marine and terrestrial environment. Further use of the methods developed here will allow for a more detailed understanding of the fate and interactions of arylarsenicals in biological and environmental systems.  122  CHAPTER 5  ARSENIC SPECIATION IN MYTILUS CALIFORNIANUS MUSSELS  5.1 INTRODUCTION As mentioned in the General Introduction of this Thesis (Chapter 1), a number of organoarsenicals have been reported to exist in marine organisms. Arsenosugars found in marine algae, along with arsenobetaine (AsB) found in marine animals, are some of the most significant arsenicals in this respect. 113 Francesconi and Edmonds have proposed a mechanism for the biosynthesis of the arsenosugars (Figure 5.1), and for the conversion of arsenosugars into AsB (Figure 5.2).42,229 It is believed that arsenate is incorporated by marine organisms, possibly because of its structural similarity to phosphate, 230 and subsequently methylated according to Challenger’s mechanism 44 (Figure 1.1), thus forming MAsA and DMAsA. According to the proposed mechanism (Figure 5.1) for the biosynthesis of arsenosugars, an arseniccontaining nucleoside (dimethylarsinyladenosine) is produced after S-adenosylmethionine (Figure 1.2) donates an adenosyl group to DMAsA. Dimethylarsinyladenosine has been isolated from the kidney of the giant clam, Tridacna maxima, by Edmonds et a!. 30 A number of dimethylarsinyiribosides identified from algal sources may have resulted following enzymatic hydrolysis of this compound. Although the order of alkylation in the biosynthesis of dimethylarsinyiribosides shown here is the most likely, adenosylation may precede one or both, of the methylation steps.  123  0  OH  ÷  As 4 Me  t  Reduction then methylation  Reduction then Methy1ation?? DMAsA  0  II  Me—As—0  0 Reduction the,n methylation  II  As—O 2 Me  OH MAsA  Reduction then Methylation’9  0  II  As 3 Me  Reduction then adenosylation  Reduction then Adenosylation  ,c5 2  MeSA_-\fl Reduction then Methylation??  HOOH  I  Glycosidation  Glycosidation  A 3 Me  —\ OR 0 L 2 Me ,  OR  Reduction then Methylation?? HO  HOOH  Figure 5.1 Proposed mechanism for the biosynthesis of arsenosugars by marine algae.  124  It has also been proposed that trimethylarsonioribosides are formed following this mechanism, although again in this situation the stage at which the third methyl group is transferred to arsenic (in Figure 5.1 it is shown as the final stage) is still not clear. It is however, possible for the methylation of arsenic to proceed in algae without adenosylation, thus producing a tetramethylarsonium salt. This type of arsenical has not been identified in algae, however it is a common constituent in bivalve mollusks. 1,28,231 It has been suggested that AsB is a degradation product of the arsenosugars. The proposed mechanism of conversion of dimethylarsinylribosides to AsB is outlined in Figure 13 This conversion requires the cleavage of the C3-C4 bond of the ribose ring, 5.2A. followed by oxidation at C4, reduction and further methylation. Limited experimental evidence is available to support this proposed pathway. The identification of a trimethylarsonioriboside in a brown alga, Sargassum thunberg-ii, and in Tridacna maxima kidney, allows for the proposal of an alternative pathway for the formation of AsB, as shown in Figure 5.2B. 13 The degradation of the trimethylarsonioriboside to arsenocholine (AsC) under anaerobic conditions has been shown to occur with virtually quantitative yield. 232 It has also been shown that AsC is converted to AsB when fed to yelloweye mullet. 233  5.1.1 Scope and Rationale of Work The main purpose of work described in this chapter was to isolate and identifj arsenicals present in Californian mussels. A number of purification techniques were adopted in order to carry out this task. Once these compounds were purified they were identified by using mass spectrometric techniques, as well as by using high performance liquid chromatography (HPLC) coupled to atomic absorption spectrometry (AAS).  125  methylarsinylacetic acid  0  II  A.  A 2 Me -COOH s-CH reduction then \iy1ation  ion Anaerobic Decomposition  0  As-CH—-CH 2 Me O H  AstCHT COO Me 3 (AsB)  \sreduction then \,methy1ation  HO  ,,,/ation  Dimethylarsinyi.ribosides  2 A 3 Me O t—CH!—-CH H Arsenocholine  B. Anaerobic Decomposition  OH  t—CH 3 Me — 2 O H -CH  oXidation  Arsenocholine  2 A 3 Me — —CH -COO (AsB)  Figure 5.2 Proposed mechanisms for the biosynthesis of arsenobetaine.  5.2 EXPERIMENTAL 5.2.1 Instrumentation A Varian Techtron Model AA 1275 Atomic Absorption Spectrometer equipped with a Varian GTA-95 accessory, was used to determine arsenic by using Electrothermal Atomic  Absorption Spectrometry (ETAAS). For these determinations a chemical modifier consisting of 500 ppm Pd in 2% citric acid was used. Chapter 3 of this Thesis provides details concerning the method employed, as well as the conditions used for the ETAAS analysis of arsenic.  126  A quadrupole mass spectrometer (DelsifNermag RiO-lOB) incorporating a desorption probe, an electron beam ion source, a quadrupole mass filter, and a channeltron detector, was used for the low resolution DCI analysis of the arsenic compounds. High resolution (R=10,000, 10% valley definition) DCI measurements were carried out using a double focusing mass spectrometer (Kratos Concept II H / Mach3 Data system). Details concerning these two instruments as well as their operating conditions are given in Chapter 2 of this Thesis. The HPLC system used consisted of a Waters M45 pump, a Waters U6K injector, the appropriate column, and an automated fraction collector. Separations were achieved by using two different columns, a Protein Pak DEAE 5PW column (7.5 mm I.D. x 75 mm 18 column (3.9 mm ID. x 300 mm length, length, Waters), and a Bondclone 10 C Phenomenex). The hydride generation gas chromatographic system used in this work has been described in detail in Chapter 2 (Section 2.2.1.3, Figure 2.3). The only difference here is that an AA spectrometer is used as the detector instead of a MS detector. The AA spectrometer used was an Instrumentation Laboratories 351 AAS. The atomization of the arsines took place inside a quartz cuvette, mounted in a hydrogen-air flame. The signal was monitored at 193.7 nm and processed by using a Hewlett-Packard 3390A integrator.  5.2.2 Chemicals The origins of the synthetic arsenic compounds used for this investigation, have been specified in Chapters 2 and 3. 5.2.3 Procedures 5.2.3.1 Extraction of Organoarsenic Compounds from Mussels Frozen Californian mussels, collected from Vancouver Island (China Beach, Sooke), were thawed and their shells were removed. This procedure resulted in 300 g of wet mussel  127  tissue which was homogenized in a blender. Extraction of the arsenic compounds present was carried out by using methanol (2.5 mL per g of wet tissue). The flask containing the tissue-methanol mixture was sealed and placed on a mechanical shaker that was left to operate for two days. Following completion of the extraction and filtering, the arsenic content of the extract was determined by using HGAAS and ETAAS.  5.2.3.2 Purification Procedures for Arsenic Compounds Present in Mussel Extracts An outline of the procedures used for the purification of the arsenic compounds present in mussel extracts is shown in Scheme 5.1. The methanol extract, obtained as described above, was evaporated to dryness and the residue was dissolved in water and extracted with portions of diethyl ether until the ether portion was colourless. The water fraction was reduced to a volume of 15 mL and then applied to a gel-permeation column (Sephadex LH-20, 23 mm I.D. x 415 mm length), and eluted with water. Fractions of 7 mL were collected, and the arsenic content in 20 jiL aliquots of each fraction were determined by using ETAAS. Each fraction was analyzed in triplicate, so a total of 60 .tL of solution was required. The arsenic containing fractions were combined and evaporated down to a volume of 10 mL and then applied to a strong cation exchange column [Dowex 50Wx8(Hj, 200-400 mesh, 25 mm I.D. x 300 mm length]. [The packing material in this column was initially extracted with acetone, until the acetone became colourless. The regeneration of the cation exchanger was accomplished by washing it with de-ionized water (500 mL), 2M sodium hydroxide (300 mL), water (500 mL), 2M HCI (300 mL), and finally with water until neutral]. A step-wise gradient elution procedure was used to elute the arsenic containing compounds. This involved the successive use of the following mobile phases: 120 mL water, 120 mL 5% ammonium hydroxide, 50 mL water and 200 mL 2M HCI. Fractions of 7 mL were collected and analyzed for arsenic. The arsenic containing fractions were combined into three groups, each of which corresponded to a peak eluting  128  off the Dowex 5OWx8(Hj column with water, ammonium hydroxide and HC1, respectively.  Scheme 5.1  Outline of the procedures used for the extraction and purification of  arsenic compounds from mussels.  [MolluskJ  Ether/water  MeOH extract  partitiomng 4—  .1  Dissolve in water  ,11r  Aqueous phase  (Gel-permeation [chromatography  Strong Cation exchange Resin  water  ammoniurn  hydrochloric  5.2.3.3 Isolation of Arsenic Compounds in the NH OH Fraction 4 The arsenic containing fractions, that were eluted from the strong cation exchange column by using ammonium hydroxide, were combined and subsequently concentrated on the rotary evaporator. The solid obtained was dissolved in a minimum amount of water and subjected to gel-permeation chromatography [GPC] (Sephadex LH-20, 23 mm I.D. x 300 mm length) with water as the eluent. Fractions of 7 mL were collected and analyzed for arsenic. The arsenic containing fractions were combined, evaporated to dryness and re  129  dissolved in a minimum amount of water. This material was then placed on a Dowex 2x8(OHj column (15 mm I.D. x 400 mm length), and eluted with 120 mL of water, followed by 200 mL of 0.2 M Rd. Fractions of 10 mL were collected and analyzed for arsenic. The evaporation of the combined arsenic containing fractions resulted in a residue which was re-dissolved in a minimum amount of water. Aliquots of this solution were examined by using thin layer chromatography (TLC) and HPLC ETAAS. TLC was carried -  out on silica-gel plates, using 65% ethanol  -  1% acetic acid  -  34% water, as the mobile  phase. The Rf values of the arsenic compounds in the extract material as well as the Rf values of the standards AsB, AsC, TMA5+1 were determined by scraping 5 mm wide silicagel strips off the TLC plate, extracting the silica powder obtained from each strip individually with 5 mL methanol and finally analyzing each extract for its arsenic content by using ETAAS. The remaining solution was also purified by TLC. Final purification was achieved on an HPLC Protein-Pak DEAE column (7.5 mm I.D. x 75 mm length) with 5 mM ammonium acetate solution, pH 6.65. A 1 mL.min 4 flow rate was used and 0.5 mL fractions were collected. The arsenic containing fractions produced a residue that was further analyzed by low and high resolution DCI-MS.  5.2.3.4 Isolation of Arsenic Compounds in the HCI Fraction The arsenic containing fractions eluting off the strong cation exchange column by using hydrochloric acid were combined and then concentrated on a rotary evaporator. The residue was dissolved in 0.2 M ammonium hydroxide (15 mL) and then chromatographed by using GPC (Sephadex LH-20, 25 mm I]) x 350 mm length). Fractions of 10 mL were collected and analyzed for their arsenic content. The arsenic containing fractions were combined and evaporated down to 10 mL. TLC was then used as described earlier, both for identification and purification purposes. Final purification was accomplished on a Protein  130  Pak DEAE HPLC column. The purified solid was analyzed by using both low and high resolution DCI-MS.  5.2.3.5 Determination of Arsenic in Mussel Shells Shells from Californian mussels were divided into two groups. The shells in both groups were washed with deionized water, air dried, crushed and ground into a fine powder in a mortar. However, the shells from one of the groups had their outer surface sand-blasted prior to being ground into a fine powder. This procedure was employed in order to remove any organic matter from the outer surface of the shell. The powder (10.0 g) was placed in a 250 mL beaker and 40 mL concentrated HC1 was added to dissolve it. After dilution with  water the undissolved material was filtered out. The resulting solution was analyzed for arsenic by using hydride generation -gas chromatography (HG-GC) AAS. -  5.3 RESULTS AND DISCUSSION  5.3.1 Purification and Isolation of Arsenic Compounds Extracted from M calfornianus The extraction and separation procedures used, as described above, were based on work previously 2 24 ,234. 31 ’ 28 published. Soft tissue from Californian mussels was extracted with methanol. The weight and arsenic content of the mussel tissue and extracts are given in Table 5.1. From this table it can be seen that 83% of the arsenic present in the mussel’s soft tissue was extracted into the methanol. After evaporation of the methanol the gum like material which remained was dissolved in water and then extracted with diethyl ether, which removed approximately 4% of the arsenic present in the water solution. This percentage is an indication of the amount  131  of lipid-like arsenicals present in the original methanol extract. We did not carry out any further work on the ether extracts. The percentage of ether soluble arsenic determined to be present in the methanol extract of these mussels, is comparable to those found by others in clam tissue, 235 Western rock lobster, 236 and shell fish. 237 However, higher amounts have been reported for crabs (17%),190,238 and brown alga Undaria pinnatIda (25%).34 An earlier worker suggested that arsenolipids from algae are unstable and are readily hydrolyzed to water-soluble arsenic compounds. 239 Recently however, the arsenical present in the chloroform extract of Undaria pinnat/Ida was isolated by using solvent partitioning, GPC, and HPLC on silica gel, and was shown to be an arsenic containing phospholipid whose structure is shown in Figure 5•334 Evidence for this structural assignment was provided by its ‘H NIVIR spectrum, and by ‘H NMR and GC-mass spectral analysis of its hydrolysis products. It was reported that this compound is labile, readily decomposing during extraction, so its identification is considered a major accomplishment.  Table 5.1  Weight and arsenic contents of extracted soft tissue, methanol extract, and  residue. Californian mussels Weight of soft tissue extracted (g)  100  Arsenic content of soft tissue (ig)  236  Residue after methanol extraction (g)  15  Arsenic content of residue (pg)  43  Arsenic content of methanol extract (ig)  196  Total arsenic extracted (%)  83  132  Following the ether extraction the water soluble arsenic compounds were subjected to GPC. The arsenic containing material eluted from the Sephadex LH-20 between 70 and 130 mL (Figure 5.4). A synthetic standard of AsB eluted within the same elution volume. This is an indication that the water soluble arsenicals present in mussels are not strongly associated with other larger molecules in the extracts. The arsenic containing fractions collected off the gel permeation column were combined, reduced down in volume and then chromatographed on a strong cation exchange resin. This type of resin has been commonly employed for preliminary separation of arsenicals of marine origin. 231 It should be noted that this chromatography separates the arsenicals into “H 0”, “NH 2 OH”, and “HCl” fractions. The arsenic content of the collected 4 fractions eluting off the column were determined by using ETAAS, and the resulting chromatogram is presented in Figure 5.5. Of the total arsenic eluting off the strong cation exchange column, 36% was present in the ammonium hydroxide fractions, 60% in the hydrochloric fractions, and 4% in the water fractions. Similar results have been reported for a variety of clam species. 231 Synthetic arsenic standards were also chromatographed on this column. Methylarsonic acid elutes from the column with water, whilst DMAsA and MB elute with ammonium hydroxide. AsC and TMAs elute with 2M Rd.  Mi  )nMe 2 (CH ) Me 2 (CH OH  Figure 5.3 Structure of an arsenic containing phospholipid present in brown alga.  133  0.30 0.25 0.20 ci)  0 C  0.15  a -  0.10  I—  0 (1)  -o  0.05 0.00 0  14  28  42  56  70  84  98  112 126 140 154 168 182 196 210  Retention Volume (mL)  ETAA chromatogram; mobile phase: water; 7 mL fractions collected; 20 jiL of each injected into the ETAA spectrometer; each fraction analyzed in  Figure 5.4 Sephadex LH-20  -  triplicate. 0.12 0.10 :5  a ci C) C  0.0s  a  0.06  0 C’, .0  0.04  -O  0.02 0.00 0  35  70  105  140  175  210  245  280  315  350  385  420  455  Retention Volume (mL) Figure 5.5 Dowex 50Wx8(H) ETAA chromatogram; mobile phase: 120 mL water, 120 mL 5% ammonium hydroxide, 50 mL water and 200 mL 2M HC1; 7 mL fractions were collected; 20 iiL of each fraction was injected into the ETAA spectrometer; each fraction -  analyzed in triplicate.  134  5.3.2 Identification of an Arsenic Compound Present in the Ammonium Hydroxide Fractions The arsenic containing fractions which eluted off the strong cation exchange column were combined, evaporated to dryness, redissolved in water and then subjected to GPC. Only fractions with retention volumes in the range of 118 to 140 mL were found to contain arsenic. AsB, a compound which is recovered from the ammonium hydroxide fraction, eluted from the gel-permeation column with a similar retention volume. For further purification the arsenic  containing material was  subjected to  chromatography on a Dowex 2x8 (OH-) column. This yielded a single set of arsenic containing fractions which eluted from the column with water. TLC was also employed to further isolate the arsenic containing compound. Methanol extraction of the arsenic-containing TLC spot and subsequent evaporation of the solvent, yielded a residue. This residue was dissolved in water and chromatographed on an HPLC Protein-Pak DEAE column with 5 mM ammonium acetate buffer as the eluent. Both of these chromatographies not only further purified the arsenic containing material but also provided information concerning the compound’s structure. This was accomplished by comparing Rf values and retention times resulting from TLC and HPLC, respectively. On TLC the arsenic containing material gave an Rf value of 0.53, while synthetic MB, which was also subjected to the same chromatographic procedure under the same conditions, gave an Rf of 0.53. Other synthetic organoarsenicals have the following Rf values: AsC, 0.30 and TMAsI, 0.39. On the Protein-Pak column the arsenic containing material elutes with the same retention time, 2.5 3.5 mm, as synthetic AsB (Figure 5.6). -  In order to further confirm the structure of the arsenical as being that of MB, mass spectrometry was used, in particular the DCI-MS technique developed in Chapter 2. The DCI mass spectrum (Figure 5.7) obtained by using ammonia as the reagent gas indicates As], among 2 ) 3 AsJ, and 105 [(CH ) 3 As], 120 [(CH 4 ) 3 ions at m!z 179 [M+H], 135 [(CH  135  other ions. This mass spectrum corresponds very well to the DCI mass spectrum obtained for synthetic AsB, shown in Figure 2.6. The identification of the arsenical was confirmed by using high resolution accurate mass measurements of positive ions in DCI mode. The development of this method has been described in detail in Chapter 2, in particular the procedure followed the analysis as described in Section 2.2.3.2. The accurate mass measurements showed that the m/z 179 ion corresponds to the protonated molecular ion of AsB. The absolute error, [(Theoretical value-Measured)/Theoretical] x 1 o6, the measured mlz value was calculated to be 3.9 ppm. From these data it can be concluded that AsB is a major component in the Californian mussel. This result reinforces the notion of the virtual ubiquity of AsB in marine invertebrates and fish. 1,13  0.12 Cl)  0  0.04  8:8 0.06 0.03 0.00 0  2  4  8  6  10  12  14  16  18  Retention Time (mm.) Figure 5.6 BPLC-ETAAS chromatogram of: A. Synthetic arsenic standards: MB, As(llI), and DMAsA 20 .tL samples containing 500 ng of each arsenical were injected onto .the column, B. arsenic containing material collected off the strong cation exchange column in the ammonium hydroxide fractions. Chromatographic conditions: Waters Protein Pak DEAE column; 5 mM ammonium acetate mobile phase; flow rate 1 mL.min; fractions collected every 30 s and 20 iiL of each one injected into the graphite tube for the determination of arsenic.  136  suopij pxo1p1(q wniuowum  LET  in  rnsi uinpx uoiwo uoiis jjo prnjjo  ‘um unnmuoo usi jo wtuids ssw (s iuai niourni) IDU  LS  ‘!I  Relative Intensity N  x  x  x  I.-’  N  I.  N  N  N  N  N  N  5.3.3 Identification of an Arsenic Compound in the Hydrochloric Acid Fractions The arsenic containing material recovered from the strong cation exchange resin when using HC1 as the eluent, was further subjected to GPC. Arsenic containing fractions eluted from this column between the volumes of 118 to 154 mL. This elution pattern is similar to the one observed for the arsenical which we identified as AsB, present in the ammonium hydroxide fractions. This indicates that the arsenic containing compound present in the HC1 fraction is similar in size to AsB. Further purification of this material was accomplished by using preparative TLC and HPLC. As mentioned earlier TLC and HPLC methods are capable of providing some structural information regarding the arsenical under investigation. In this case TLC was used for such purposes. The Rf of the arsenic containing material was determined to be 0.38, which is in agreement with that obtained for synthetic TMAs+I. To further confirm that the arsenic compound present in the HC1 fractions is the TMAs ion, MS was employed. Again in this case DCI-MS proved to be invaluable. The low resolution DCI mass spectrum is shown in Figure 5.8; with characteristic ions at mlz 135, 121, and 105, corresponding to [(CH Asf, [(CH 4 ) 3 As+Hf, and [(CH ) 3 Asf’, 2 ) 3 respectively. The DCI mass spectrum obtained is similar to the one reported for synthetic TMAsI (Figure 2.7). High resolution DCI measurements of the m/z 135 ion showed that it corresponded to [(CH Asf, with an absolute error of 2.2 ppm from the theoretical 4 ) 3 value. From these results we can conclude that the arsenic compound present in the residue obtained from the HCI fractions is a tetramethylarsonium salt. The presence of such salts in marine animals (clams: Meretrix lusoria), was first reported by Shiomi et al. 28 Since then others have reconfirmed their presence by identi1ying the TMA5+ in a variety of marine 231 Even though the role of the compound is still unclear, it has been proposed that animals.  138  3 it is formed via the methylation of Irimethylarsine, with SAM being the likely CH 3 donor. 1 ’ 1  N  N  L)_________________  —  N Lr,___ ‘-4 —=‘-4  Lt  x  x  Lr  Lt  £suuj  N Aj)J  Figure 5.8 DCI (ammonia reagent gas) mass spectrum of arsenic containing material, collected off a strong cation exchange resin in the hydrochloric acid fractions.  139  5.3.4 Attempts to Identify an Arsenic Compound in the Water Fractions Attempts were also made to identi1’ the arsenic containing compound(s) present in the water fractions. Purification by using GPC, TLC, and HPLC yielded residues for which  no structural assignments could be made based on the DCI mass spectra obtained. The residue appeared to consist of a mixture of compounds and so fhrther purification and isolation seemed necessary. TLC showed the presence of only one arsenic containing spot with an Rf of 0.83. Unfortunately fI.irther investigation of this material was not possible because of its scarcity and/or lack of stability. No previous study has reported the successful identification of an arsenic compound that eluted in the water fraction, perhaps for similar reasons. It has been recently reported that arsenosugars are present in bivalves in addition to 4 Ion-pair reverse-phase HPLC-ICP-MS was used for the identification of the TAsh 8 J 50 arsenosugars. The advantage in using such a method is the elimination of a number of sample preparation steps, such as the use of the strong cation exchange resin, which could change the original structure of the arsenicals or even introduce artifacts. We suggest that, although it is likely that the water fractions from M cabjornianus contain arsenoribose derivatives, these compounds may not nccessarily be those originally extracted from the mussel flesh, because such derivatives are sensitive to extremes of pH conditions and may not survive passage down the strong cation exchange column. In support of this suggestion, some recent work shows that arsenosugars are eluted partially decomposed from a strong ° 24 cation exchange resin in the water fractions.  5.3.5 Arsenic Speciation in Mussel Shells A HG-GC-AA system was used to separate and detect the arsenic species present in mussel shells which form arsines upon reduction with sodium borohydride. As described in the Experimental Section, the shells used for this study were divided into two groups, one  140  contained shells which had the outside part sand blasted and the other remainded untreated. However, once analyzed no pronounced difference was detected in the arsenic content and arsenic species present in these two groups of mussel shells. The shells were found to contain: 105 ng.g inorganic arsenic [As(V)+As(ffl)], 2.1 ng.g’ MeAs(V), and 10.3 ng.g Me As(V). These results are reliable to ± 5%. A typical chromatogram of the 2 arsines is presented in Figure 5.9. Similar arsenic species have been detected in Blue mussel shells, as well as a variety of clam shells. 241  8  Retention Time (mm.) Figure 5.9 HG-GC-AAS chromatogram of arsenic containing species, originating from Californian mussel shells.  141  5.4 SUMMARY  The results reported in this Chapter establish the presence of AsB and a tetramethylarsonium salt in Californian mussels. Initially the arsenic containing material extracted from mussels was purified and isolated by using a combination of chromatographic techniques, such as, TLC, GPC, and [{PLC. Comparison of Rf values, retention volumes, and retention times, of the arsenicals present in the mussel extracts with those of synthetic arsenicals, provided preliminary information regarding their structural identity. Finally the purified residues were analyzed by using MS techniques developed in this work (described in detail in Chapter 2). The use of both low and high resolution DCI MS, allowed for the identification of the arsenicals mentioned. Questions still remain about the identity of the arsenic compound(s) eluting unabsorbed off the strong cation exchange column, even though it is highly possible that these compounds are arsenosugars. The presence of methylated arsenic species in the mussel shells was also established. This was accomplished by using a HG-GC-AAS system and identifying the arsines by comparing their retention times with those of synthetic standards.  142  CHAPTER 6  11]-METI{YLARSONIC ACID II%i A STATIC 3 BIOTRANSFORMANTION OF [ SEAWATER SYSTEM CONTAINING MYTILUS CALIFOR[VIANUS  6.1 INTRODUCTION  6.1.1 Scope and Rationale of Work The main objective of this study was to investigate the biotransformation of an arsenic compound in the marine environment. More specifically, this experiment was designed to H]-MAsA) in a static 3 study the biotransformation of 3 H-labeled methylarsonic acid ([ Hj-MAsA 3 seawater system containing Mytilus cahfornianus, and to identify the [ metabolites accumulated within the mussel flesh. The experiment was conducted in order to provide answers to a number of questions concerning the arsenic cycle in the marine environment,  such  as:  which  marine  organisms  are  responsible for the arsenic  biotransformations, what is the identity of the arsenicals produced along this cycle, and how do these arsenicals interconvert? MAsA was selected as the compound to investigate because it is believed to be a precursor to the more complex arsenicals found in marine organisms, e.g. AsB, arsenosugars, 44 although a direct link has not been established. Another reason for using MAsA is that etc., it is one of the least complex arsenicals present in the marine environment and can easily be labeled with tritium. The main purpose for using the radioactive label is to simplify the detection of MAsA metabolites, even if they occur at very low levels, and also to allow for a simple way of differentiating between the metabolites produced from naturally occurring arsenicals and  143  those from [ H1-MAsA. Tritium was selected, as opposed to other radionuclides, because it 3 is inexpensive and can be detected at very low levels, it is relatively long lived, and it can C, which As and 14 easily be handled with minimum precautions. Other radionuclides, e.g. 74 could have been used as labels, do not possess most of these advantages. Mytilus calfornianus mussels collected off the B.C. coast were used because of their known arsenic species content (Chapter 5 of this Thesis), and also because they can be maintained relatively easily compared to the smaller M edulis species.  6.1.2 Methylarsonic Acid in the Marine Environment Methylarsonic acid and its derivatives have been reported to exist in various parts of the 1 The anthropogenic input marine environment; seawater, marine organisms, and sediments. of MAsA is believed to be relatively small, except for localized, often accidental discharges, 42 45 ’ 5 herbicide. and in spite of the fact that sodium methylarsonate is widely used as a 2 The presence of methylarsenic species in the marine environment is a consequence of the methylation of inorganic arsenic. It has been demonstrated that this methylation occurs via biomethylation, i.e., it involves living organisms, rather than just an environmental methylation involving a chance abiotic reaction between an available metal/metalloid and a 246 biologically formed methylating agent outside the cell. The presence of MAsA in seawater is well documented as the result of the work of 24 on the concentration of MAsA and other arsenic species in 48 and others Andreae 2 ’ 247 the northeast Pacific and Californian coastal waters. The profiles of arsenic species concentrations versus depth show that the arsenate concentration in the photosynthetically active surface waters (euphotic zone) is decreased, with a corresponding increase in arsenite, MAsA and DMAsA concentrations. In deeper waters the concentrations of the methylarsenicals are at or below detection limits (0.002 ppb). This observation suggests that DMAsA and MAsA are produced by phytoplankton or closely related heterotrophs. This is  144  further supported by correlations which were found between typical indicators of primary productivity, i.e., chlorophyll concentrations and/or 14 C uptake, and depth profiles of the concentrations of arsenic species. Similar results were reached by Reimer and Thompson who studied the distribution of arsenic species present in the British Columbia coastal waters during a bloom. The methylarsonate profile they reported also indicated that this compound is produced in the surface water. 249 Methylarsonic acid has also been shown to be present, at very low levels, in interstitial pore water of both natural marine sediments and sediments contaminated with mine-tailings. Laboratory studies have shown that microorganisms present in these sediments are capable of methylating arsenic under both aerobic and anaerobic conditions. 61 By using HG-GC-AAS it has been shown that bivalve shells contain methylarsenicals, probably salts of MAsA and DMAsA. 241 Because of the method of detection, the possibility that these are methylarsenic(III) species cannot be completely ruled out. Traces 48 However, the identification of of MAsA have been reported to exist in dogfish tissue. MAsA was based only on a comparison of the retention times obtained by using HPLC-ICP MS, for both standard arsenicals and those present in the tissue extracts.  6.1.3 Arsenic, accumulation and biotransformation experiments  To date a limited number of accumulation and biotransformation experiments have provided results that have contributed to our knowledge of arsenic cycling in the marine environment. Arsenocholine (AsC) was converted into arsenobetaine (AsB), glyceryiphos phorylarsenocholine (Figure 6. 1A), and phosphatidylarsenocholine (Figure 6. 1B) in yellow 250 Accumulation experiments with mussels (Mytilus eye mullet following oral administration. 229 edulis) suggested that AsB is readily taken up from seawater by these animals. The effects of certain environmental (concentration, temperature, salinity) and biological As-arsenate accumulation and elimination processes in the (tissue parts) variables on 74  145  mussel Mytilus gaioprovincialis have been studied. 251 In this study it was found that arsenic uptake increased with increasing arsenic concentration in the water. The accumulation was As concentrations partially supressed at higher external arsenic concentrations. The highest 74 were recorded in the byssal threads and the digestive gland. It was also shown that increased water temperatures and decreased salinity enhanced arsenic uptake, and that accumulation in the byssal threads contributed to the elimination of the element from the mussels. In a similar experiment Kiump examined the factors influencing the process of arsenic accumulation and elimination in a food chain consisting of Fucus spiralis (macro algae) Littoria littoralis (marine snail)  252 The most Nucella lapillus (predatory marine snail).  As-labels were the following: interesting results obtained from this work which involved 74 diet is the major source of arsenic for L. littoralis and N lapillus, and arsenic is not bio magnified along the food chain. H-labeled arsenic In an experiment similar to the one reported below, water-soluble 3 compounds were extracted with phenol from mussels (Mytilus edulis) and seawater after HJ-AsB was found in both the 3 HJ-MAsA and 253 3 H]-dimethylarsinate. [ 3 [ exposure to [ mussels and the seawater. The results of this experiment indicate that AsB is biosynthesized by microscopic organisms, probably primary producers, in the seawater and that it is bioaccumulated quite rapidly by mussels.  0 +  3 Me  II -O-P-O-CH 2 I I 0  I  OROR  A) R=H B) 3 )CH 2 R=CO(CH Figure 6.1 Structures of glycerylphosphorylarsenocholine (6. 1A), and phosphatidy larsenocholine (6.113). Both of these compounds were present in yellow-eye mullet following oral administration of arsenocholine.  146  6.2 EXPERIMENTAL 6.2.1 Instrumentation 6.2.1.1 Liquid Scintillation Counting A liquid scintillation counter (Packard Tri.carb® 1900 TR) was used to measure the activity of 3 H-labeled compounds present in the various mussel flesh extracts and chromatographic fractions. The counter was standardized daily by using a standard (Packard Standard #22) which contained  0.2 .tCi of  [II1. The scintillation counter was connected to  an Epson LX-810 printer. Plastic scintillation vials capable of holding 6 mL of liquid were used throughout this work. Each sample was counted in triplicate, for 4 minutes each time. The results obtained from the counts were averaged.  6.2.1.2 High Performance Liquid Chromatography The HPLC system consisted of a Waters M45 pump, a Waters U6K injector, the appropriate column and an automated fraction collector. Separations were achieved by using two different columns, a Protein Pak DEAE 5PW column (7.5 mm I.D. x 75 mm length, Waters), and a Bondclone 10 C 18 column (3.9 mm I.D. x 300 mm length, Phenomenex).  6.2.1.3 Atomic Absorption Spectrometry A Varian Techtron Model AA 1275 Atomic Absorption Spectrometer equipped with a Varian GTA-95 accessory, was used to determine arsenic by using Electrothermal Atomic Absorption Spectrometry (ETAAS). For these determinations a chemical modifier consisting of 500 ppm Pd in 2% citric acid was used. Chapter 3 of this Thesis provides details concerning the method used, as well as the conditions used for the ETAAS analysis of arsenic.  147  6.2.2 Chemicals 104 AsC,’° AsB, 5 TMAs, 74 and 254 H]-MAsA were synthesized by using literature 3 [ methods,  all  other  chemicals  glyceiylphosphorylarsenocholine  were  (0.7  commercially  mg)  was  kindly  available.  A  supplied  by  sample Dr.  of K.A.  229 Francesconi.  6.2.3 Collection and Storage of Mytilus cal(fornwnus Mussels (Mytilus californianus) collected from Quatsino Sound, B.C., were stored in holding tanks at the University of British Columbia I Federal Dept. of Fisheries and Oceans, West Vancouver Laboratory, for 4 months prior to the experiment. These facilities are capable of providing a continuous flow of seawater which is subsequently aerated in the holding tank (Figure 6.2).  A)  Seawater from ocean  *  B)  Mussels  *  I7  Seawater into ocean Figure 6.2 Seawater tanks used for mussel storage and exposure experiment: A: 200 L seawater tank with continually flowing seawater, used for mussel storage; B: 15 L experiment tank containing mussels and { H]-MAsA; 3  *:  aeration lines for both A and B.  148  6.2.4  Procedures for the Speciation of 3 H-Labeled Compounds Extracted From  Mytilus calfornianus For the experiment 21 mussels were selected randomly and were placed in a static but H]-MAsA (1.5 3 well aerated tank (Figure 6.2B) containing 15 L of seawater and 34 jiCi [ ppm As). The mussel shells varied in length from 7 to 16 cm. After 9 days, 9 mussels were removed and shucked, providing a total of 200 g of wet tissue. The tissue was homogenized, 500 mL of methanol was added, and the mixture was then placed on a shaker for 2 days. This extraction step was repeated with another 500 mL of methanol. The extracts were combined and evaporated to dryness and the residue was dissolved in water and extracted with portions of diethyl ether until the ether portion was colorless. The water fraction was evaporated down to a volume of 25 mL, applied to a gel permeation column (Sephadex LH-20, 25 mm ID. x 300 mm length), and eluted with water. Fractions of 9.5 mL were collected, and the 3 H activity in each was determined by withdrawing a 500 1 iL aliquot which was mixed with 5 mL of scintillation liquid in a counting vial, before being transferred to a scintillation counter where disintegrations per minute (DPM) were measured. The 3 H containing fractions were combined and evaporated down to a volume of 10 mL and then applied to a strong cation exchange column [Dowex 50Wx8 (Hj, 200-400 mesh, 25 mm I.D. X 300 mm long columnj. The following mobile phases were H containing compounds: 200 mL water, 200 mL 5% ammonium used to elute the 3 hydroxide, 40 mL water, and 150 mL 2M HC1. Fractions of 9.5 mL were collected and analyzed for 3 H activity. The 3 H containing fractions were combined into four main fractions, each of which corresponded to a peak eluting off the Dowex 50Wx8 (Hj column. These fractions were then evaporated to dryness and redissolved in a minimum volume of water. Each of these four solutions was chromatographed on an anion exchange Protein Pak DEAE column. Two mobile phases were used, the first consisted of 5 mM sodium acetate adjusted to pH 4 with acetic acid and the second consisted of 5 mM ammonium acetate, pH  149  6.8. A Bondclone 10 C 18 reversed-phase column was also used for ion-pair reversed-phase liquid chromatography, with water-methanol 95:5 as the mobile phase and 5 mM . In 4 tetrabutylammonium nitrate as the ion pair. The flow rate in all cases was 1 mL.min order to establish retention times for identification purposes a number of standard arsenicals [AsB, methylarsonic acid (MAsA), dimethylarsinic acid (DMAsA), glycerylarsenocholine (GPAC)] were also chromatographed. The standards were detected by using ETAAS: 15 .tL aliquots from 0.5 mL fractions. The sample extracts were monitored by counting aliquots of 1 mL that were mixed with 5 mL of scintillation liquid.  6.2.5 Determination of Total H In Mussel Tissue and Shells On each of the third, sixth and ninth days, 4 mussels (M calfornianus) were removed from the tank and dissected into six parts: gills, adductor muscle, foot, mantle, muscle tissue and visceral mass. The same tissue parts from each of the four mussels were combined. The H determinations. All of the tissue byssal threads and the shells were also set aside for total 3 parts were freeze dried and then ground into a fine powder. H The Oxygen Flask combustion method was used to prepare the mussel tissue for 3 liquid scintillation counting. This method has been used in a wide range of analytical H and 14 C activities in biological applications as well as for the determination of 3 56 samples. 2 ’ 255 The apparatus consists of a 3L round-bottom flask, a stopper and a platinum basket (Figure 6.3). The freeze dried sample (100 mg) was wrapped in a paper sample wrapper and placed into the platinum basket. The flask was filled with oxygen and then the tip of the sample wrapper was ignited and the stopper and basket were inserted into the flask. Once H]-H 3 [ 0 combustion had ceased 5 mL of absolute ethanol was added to absorb the 2 produced. A 2 mL aliquot of the alcohol solution was transferred to a counting vial and 4 mL  150  of scintillation liquid was added prior to counting. This procedure was performed in triplicate for each tissue sample. Shells were washed with deionized water, air dried, crushed and ground to a fine powder in a mortar. The powder (2.0 g) was placed in a 250 mL beaker and 10 mL of 2M HC1 was iL added. After dissolution the solution was filtered to remove the insoluble residue. 500 1 portions were withdrawn and mixed with 5 mL of scintillation liquid and then counted.  ignition  A sample  Figure 6.3 Oxygen Combustion Flask (3L); A. platinum basket, B. paper sample wrapper, C. pressure relief in the form of a rubber balloon.  151  6.3 RESULTS AND DISCUSSION  6.3.1 Speciation of the 3 H-Labeled Compounds Extracted From M. californianus Hj-MAsA. The pitJiininaiy 3 Mussels were exposed for 9 days to seawater containing [ methanol treatment of their tissue extracted 74% of the 3 H-labeled compounds. The extraction efficiency was determined by measuring the 3 H activity of the mussel flesh before and after the extraction. The methanol extract was evaporated to dryness and redissolved in water. Diethyl ether extracted approximately 9% of the counts from this solution which is probably a measure of the amount of lipid-like arsenicals in the original methanol extract. The water soluble 3 H-labeled compounds eluted from Sephadex LH-20 between 100 and 140 mL (Figure 6.4). Standard AsB eluted within the same retention volume. This is an indication that the 3 H-label has been incorporated into water soluble compounds that exhibit similar physical properties (size and/or adsorption characteristics) to those of AsB, on this particular column.  1 400 1200 1000 800 600 400  200 0  0  2  4  6  8  1012141618202224262830  Fraction Number  Figure 6.4 Sephadex LH-20 liquid scmtillation chromatogram; water mobile phase; 9.5 mL fractions collected; 0.5 mL of each fraction mixed with 5 mL scintillate and counted.  152  <—  0————><———NH 2 H 0 4 H———>  <———HCI———>  900 800 700 600  0 400 300 200 1 00 0 04812162024283236404448525660  Fraction Number  Figure 6.5 Dowex 50Wx8 (Hj liquid scintillation chromatogram; mobile phase 200 mL water, 200 mL 5% ammonium hydroxide, 40 mL water and 150 mL 2M HC1; 9.5 mL fractions collected; 0.5 mL of each fraction mixed with 5 mL scintillate and counted.  153  A strong cation exchange resin has been commonly employed to accomplish a preliminary separation of the arsenicals in marine animals into water, ammonium hydroxide 24 This procedure was also used in the 31 34 28 fractions. ’ and hydrochloric acid eluting 2 present study. The radioactivity in the fractions eluting from the Dowex 5OWx8 (Hj was measured and the resulting chromatogram is presented in Figure 6.5. Four peaks containing H-labelecl compounds eluted from this column. The first peak contained a compound (this 3 compound will be referred to as PEAK #1) that essentially showed no interaction with the strong cation exchange resin. The compounds comprising the second (Peak #2) and third (Peak #3) peak were weakly retained by the resin, and were eluted by using water. Finally the compound in the fourth peak (Peak #4) was eluted by a 5% ammonium hydroxide solution. Methylarsonic acid elutes from this type of column with water, whilst DMAsA and AsB elute H activity was with ammonium hydroxide. AsC and TMAs elute with 2M HC1. No 3 detected in any of the fractions when 2M HC1 was used as the eluent for the mussel extracts. H-labeled After obtaining this preliminary information about the properties of the 3 compounds, HPLC was used for further identification.  H]-AsB in mussel extracts 3 H]-MAsA and [ 3 6.3.1.1 Identification of [ The HPLC Protein Pak DEAE chromatograms (5 mM sodium acetate, pH 4, mobile phase), of Peaks #3 and #4, and of a mixture of standard AsB, DMAsA and MAsA are presented in Figure 6.6. Standard MAsA and Peak #3 exhibit the same retention time. Peak #4 also has the same retention time as standard AsB and DMAsA, which are not separated under these conditions. However, these two standards are readily separated on the Protein 74 This is Pak column by using 5 mM ammonium acetate (pH 6.8) as the mobile phase. presented in Figure 6.7, along with the elution profile of Peak #4. Peak #4 exhibits the same retention time as standard AsB.  154  o  0.5  <  0.4  -  0  0  5  10  15  20  25  30  0  5  10  15  20  25  30  0.3  a) 0.1 U)  0  a  0.0 600 400 200 150 1 00  a 50 0  Retention Time (mm) Figure 6.6 HPLC conditions for A, B, C: Waters Protein Pak DEAE column; mobile phase 5  mM sodium acetate, pH adjusted to 4 with acetic acid; flow rate 1 mLmin; fractions collected every 1 mm; A. HPLC-ETAA chromatogram of standards a: arsenobetaine (250 ng As), b:dimethylarsinic acid (500 ng As) c: methylarsonic acid (500 ng As). ,  B. HPLC-liquid scintillation chromatogram of Peak #3; 1 mL fraction mixed with 5 mL scintillate and counted. C. HPLC-liquid scintillation chromatogram of Peak #4; 1 mL fraction mixed with 5 mL scintillate and counted.  155  0  2  4  6  8  10  12  14  16  18  20  0  2  4  6  8  10  12  14  16  18  20  0.5 a) -  a) 0-.  0.4 0.3 0.2  Q 0  U0  80 0  40  0  Retention time (mm)  Figure 6.7 HPLC conditions for A and B: Waters Protein Pak DEAE column; mobile phase 5 mM ammonium acetate, pH 6.8; flow rate 1 mLmin. A. HPLC-ETAA chromatogram of standards a: arsenobetaine (250 ng As), b:dimethylarsinic acid (250 ng As); 0.5 mL fractions collected every 0.5 mm. B. HPLC-liquid scintillation chromatogram of Peak #4; 1 mL fraction mixed with 5 mL scintillate and counted.  156  a  U) L  0  3  6  9  12  15  0  3  6  9  12  15  0.3  -  a  U) 0 —  U) ..D  <  0.1 0.0 500 400 300  o  200  1 00 1 20 90  a...  60  0  30 0  Retention Time (mm) Figure 6.8 HPLC conditions for A, B, C: Bondclone 10 C 18 reversed phase column; mobile phase water-methanol 95:5; ion-pair reagent 5 mM tetrabutylammonium nitrate; flow rate 1 . 1 mLmin A. HPLC-ETAA chromatogram of standards a: arsenobetaine, b: dimethylarsinic acid,  C:  methylarsonic acid; fractions collected every 30 sec. B. HPLC-liquid scintillation chromatogram of Peak #3; 1 mL fraction collected and mixed with 5 mL scintifiate and counted. C. HPLC-liquid scintifiation chromatogram of Peak #4; 1 mL fraction collected and mixed with 5 mL scintillate and counted.  157  Ion-pair reversed-phase liquid chromatography was also used for these identifications. Figure 6.8 presents the elution profile of standard AsB, DMAsA, MAsA and that of Peak #3 and #4. Again Peak #4 is identified as AsB and Peak #3 as MAsA. All of these chromatographic results are summarized in Table 1.  Table 6.1 Summary of chromatographic experiments.  Dowex H (Eluent)  HPLC Anion Exchange, 5mM sodium acetate, pH 4 (tR, mm.)  Compound proposed to be MAsA (PEAK #3) Standard MAsA Compound proposed to beAsB (PEAK #4j Standard AsB Standard DMAsA PEAK #2  0 2 H  15.5±0.5  0 2 H  15.5±0.5  HPLC Anion Exchange, 5mM ammonium acetate, pH 6.8 (tR, mm.) nd  nd  RPLC C-18 Ion-Pair, 5mM tetrabutyl ammonium nitrate, pH 6.8 (tR, mm.) 10.75±0.25  11.0±0.5  5% OH 4 NH  3.5±0.5  3.5 ±0.5  5.5±0.5  5% OH 4 MI 5% OH 4 NH 0 2 H  3.5±0.5  3.25±0.25  5.75±0.25  3.5±0.5  14.25±0.25  8.25±0.25  8.0±0.5  3.5 ±0.5  nd : not determined  158  nd  These results are significant in that they indicate that MAsA is a precursor to AsB. Other experiments, 1,257 of similar nature, have been unsuccessful in demonstrating that a simple 25 arsenic compound, i.e. As]-arsenate, 74 is a precursor to AsB. These experiments involved [ either direct feeding of the precursor arsenate to a marine animal or exposing one specific alga to the precursor in the growth medium. In contrast, the present study used natural seawater, which contains different types of algae, bacteria, phytoplankton, and diatoms, any one of which, or any combination, could be involved in the biotransformations necessary for the production of AsB. Only limited information can be gained from this experiment regarding the stage in the food chain at which the arsenical interconversion takes place, even when other results obtained from this work (Chapter 5) and data of other researchers are considered. AsB has been found to occur naturally in Mytilus californianus, while MAsA has not H]-AsB found 3 been detected (Chapter 5 of this Thesis). Thus the ratio of [ H]-MAsA to [ 3 in the mussel flesh following exposure to [ Hj-MAsA does not reflect the natural ratio. It 3 should be pointed out, however, that the high arsenic concentration (1.5 ppm As) in the static seawater system, may have caused overloading of the mussels thus not allowing them H]-MAsA 3 to function in a natural way. Thus we can only speculate that the conversion of [ into [ Hj-AsB does not take place within the mussel itself This is based on the fact that if 3 H]-MAsA was readily converted to [ 3 [ H]-A5B within the mussel tissue then the measured 3 content of [ H]-MAsA would be much lower. If however the conversion is a slow process 3 we would expect to find naturally occurring MAsA within this mussel and this, as indicated above, is not the case. These results indicate that [ H]-AsB is either accumulated from water and/or food, or 3 may be synthesized from arsenic compounds other than MAsA within the mussel itself Cullen and Nelson 253 reached similar conclusions from studies with Mytilus edulis, a 229 Wrench et. al. 254 studied a three step bivalve that readily takes up AsB from seawater.  159  food chain consisting of an autotroph, a grazer and a carnivore. They concluded that the muscle tissue of the carnivorous shrimp could not itself form “organic arsenic”, which is believed to be synthesized by primary producers.  6.3.1.2 Attempts to identify Peaks #1 and #2 Peaks #1 and #2 have also been chromatographed under all the above conditions. The resulting retention times do not correspond with any of the standard arsenicals available to us. Glyceryiphosphorylarsenocholine (Figure 6. 1A), a compound that has been found to accumulate in yellow-eye mullet following oral administration of arsenocholine, was also ° The reported chromatographic behavior of this compound seemed to be similar to 25 tested. those of Peaks #1 and #2. However, the retention times acquired for this compound on the Protein Pak column did not match those of Peaks #1 and #2. Thus questions still remain about the identity of these two compounds that were not 3 ’ 30 identified. Only recently has it been established that arsenosugars are present in bivalves. Although it is likely that PEAKS #1 and #2 contain arsenoribose derivatives, these may not necessarily be those originally extracted from the mussel flesh, since such derivatives are reported to be sensitive to extremes of pH and may not survive passage down the strong cation exchange ° 24 ’ 259 column.  6.3.2 Determination of Total H in Mussel Tissue and Shells H]-MAsA containing seawater the specific 3 After 3, 6, and 9 days of exposure to [ 1 activities detected in the M calfornianus mussel tissue were 27, 42, and 65 dpm.mg respectively. The distribution of 3 H activity within the mussel was also determined. After the mussels had been exposed for a 9 day period the highest specific activity was found in the visceral mass and the gills of the mussel (Figure 6.9). Similar distribution of the activity within the tissue parts was observed on days 3 and 6.  160  No 3 H activity was detected in the shells after using the sample preparation procedure described, indicating that physical processes such as surface sorption play a minor role in the uptake of arsenic compounds by the mussel shells. Methylarsenicals have been shown to be present at very low levels in shells of M cahfornianus (Chapter 5 of this Thesis), M edulis H activity in the shells we examined, and shells of various clams. 241 The absence of 3 indicates that the incorporation of arsenic compounds in shells is a process that requires more than 9 days to occur. This would be consistent with the idea that the mainly inorganic arsenic species to be incorporated into the shell are produced via biomineralization processes rather As-arsenate added to than just surface sorption. Others have reported that radioactive 74 seawater is found in (or on) the shells of gastropods. In general, there are few reports in the literature regarding the presence of arsenicals in mollusk shells. It has been proposed that byssal threads act as sinks for the removal of certain 260262 This has also been shown radionuclides and heavy metals from contaminated mussels. 251 In our work the byssal As-arsenate) elimination from mussels. to occur with arsenic (as 74 H activity thus supporting the threads were found to contain an appreciable amount of 3 notion of arsenic elimination and removal via the byssal threads.  161  20  >..  t >  ±  15  -I  C,  ++  10 C,  Iii  LI)  0  —  —  4  o  U)  Ct)  _J  ZE  I  — I-  Lii  Lii  C)  I-  0  -  Li  Cfl  -‘  I  U)  0  c_) D  5;  I  0  4 U) Cl) >-  I  -‘ Lii  Ct)  4  Mussel Parts H activity within mussel parts, sampled on day 9. Figure 6.9 Distribution of 3  6.4 SUMMARY A great deal of information concerning the arsenic cycle in the marine environment can be obtained by conducting arsenic accumulation and biotransformation experiments, such as those presented in this chapter. The major fmdings of this study are the following: when Mytilus californianus is [ H 1-MAsA, 3 H]-MAsA in a static seawater system, 3 3 exposed to [ [ H 1-AsB, and two H-labeled compounds accumulate in the tissue parts of the mussel, also the unknown 3 highest specific activity is found in the visceral mass and the gills of the mussels. This work establishes that MAsA is a likely precursor to AsB. It also indicates that  162  strong anaerobic conditions are not required for the formation of AsB, as has previously 29 proposed. 2 ’ been 42 Because experiments of this type can add to our knowledge of arsenic cycling in the environment, additional studies should be conducted to address a number of still unanswered questions, for example: (a) Once suitable arsenic standards become available, the task of identifying the two compounds, which were not successfully identified in this study (Peaks #1 and #2), should be pursued. The identity of these compounds may allow us to establish the intermediates in the biosynthesis of arsenobetaine. (b) The biotransformations of other arsenic compounds which exist in the marine environment should also be investigated. For example, TMAs+ is a compound whose presence in marine animals is well established but whose origin and role is completely unknown. For example, is this compound further metabolized? (c) Also of great interest would be to identify the stage in the food chain at which the interconversion of the arsenicals occurs. This can only be accomplished by conducting well controlled feeding experiments, involving selected organisms at various trophic levels and investigating their capabilities to biotransform arsenic compounds. (d) The relationship between the biomineralization processes responsible for shell formation and the presence of arsenic in the shells should be established. This can be accomplished through similar experiments of longer time duration. Shell layers would have H-labeled to be removed and analyzed separately in order to investigate the presence of 3 arsenic in the newly formed layers.  163  CHAPTER 7  SUMMARY  In the previous chapters of this thesis, research concerning the development and application of analytical methods for arsenic speciation was described. During this work, a major emphasis was placed on the development and evaluation of mass spectrometric techniques for the analysis of arsenicals found in both the marine and terrestrial environment. In particular, DCI-MS, high resolution DCI-MS, MALDI-TOF-MS, dynamic-LS1MS-MS, thermospray-MS, LSIMS, and HG-GC-MS were investigated. Satisfactory DCI-MS spectra can be obtained with 100 ng arsenic placed on the DCI filament, making the technique eminently suitable for the investigation of environmental samples. The spectra of the arsenicals offered recognizable molecular ions as well as characteristic fragment ions. To extend the usefulness of DCI-MS, an analytical method was developed that permitted the use of mass deficient reference standards for calibration purposes in accurate mass measurements of positive ions under ammonia DCI conditions. This was accomplished by employing a mixture of ammonia and methane as reagent gases. In the high resolution accurate mass measurement experiment, this gas mixture allows for simultaneous detection of the mass spectrum of perfiuorokerosene (calibration substance) adequate for calibration purposes, and the spectrum of the analyte which contains molecular weight information. A needle valve system was used to control the composition of the gas mixture introduced into the ion source. For positive ion accurate mass measurements of higher masses (up to mlz=23 00), Fomblin 18/8 oil was used successfully as a reference standard under ammonia,  methane, and iso-butane desorption chemical ionization conditions.  164  Both low and high resolution DCI-MS were used to identif,’ arsenic compounds present in Mytilus cahfornianus. This was accomplished by first extracting the arsenicals from the mussel flesh with methanol, and then isolating and purifjing the compounds by means of conventional chromatography. Subsequently the mass spectrometric techniques described above were used to analyze the purified materials and provide spectra suitable for the structural characterization of two principle arsenic species present in the mussels; arsenobetaine and the tetramethylarsonium ion. MALDI-TOF-MS is another mass spectrometric technique which exhibits potential for use in the speciation of arsenic in the environment. The method is very sensitive and the minimum detectable amount was determined to be 0.3 ng of arsenic or 4 pmole of compound. This feature suggests that the method could be used for the identification of minor arsenic components in environmental samples. Other advantages which make it attractive for arsenic speciation are the following: a) the method is capable of providing molecular ions as well as fragment ions for a variety of arsenicals, b) compound fragmentation can be controlled to a certain extent by adjusting the N 2 laser power, and c) the analysis requires only very small sample volumes 1-1.5 1 .tL. HG-GC-MS was also developed in this work and was used to provide conclusive evidence of -CD 3 incorporation from L-methionine-methyl-d 3 into arsenic compounds produced from arsenate by alga cell cultures. The results from this work are of great importance because they contribute towards the elucidation of the mechanism of arsenic biomethylation. These findings strongly support the notion that the oxidation-reduction pathway involving carbonium ions, as originally suggested by Challenger for the alkylation of arsenic by microorganisms (Figure 1. i),44 applies to marine cellular alga and probably other marine organisms. The advantages of micro-HPLC were demonstrated by the work reported in Chapter 4 of this thesis. Micro-LC columns fabricated and packed in house can be conveniently  165  coupled on-line to a variety of mass spectrometric systems, mainly because of the extremely low flow rates they require. It was shown that a 99 % reduction in the volume of solvent waste is achieved by switching from conventional to micro-LC, and a 87.5 % reduction by switching from microbore to micro-LC. This reduction in solvent waste can translate into large savings, in terms of both solvent consumption, and solvent waste handling and disposal. Compared to previous work, lower detection limits for arsenic compounds were also obtained when micro-LC columns were employed in conjunction with UV and ETA.A 74 detectors. The micro-LC columns are able to efficiently separate arsenicals that are used as animal feed additives as well as their potential metabolites. This separation technique in conjunction with a variety of compatible detectors, can help in the evaluation of the physiological role of these compounds as well as to assess their impact on the environment. When considering the general use of these feed additives and their possible impact on the environment and public, this method is of prime importance because only a small number of analytical techniques had been developed for the analysis of this class of compounds. Most of these are for target analysis of a few specific compounds, which would probably be unsuitable for the detection of metabolites. In order to accomplish the accurate determination of arsenic in environmental samples it is essential that the optimization of the analytical method used be performed in conjunction with a standard reference material of similar matrix to that of the sample being investigated. In the present work Simplex optimization was used to efficiently delineate the optimum experimental conditions to be used for the electrothermal atomic absorption spectrometric analysis of total arsenic in a standard reference material of marine origin. Four experimental variables, were considered:  ashing temperature, atomization temperature, modifier  concentration, and atomization ramping time. This combination of methods and materials provides a powerful means of rapidly optimizing the experimental conditions used for the  166  analysis of arsenic in a wide variety of samples of environmental origin. Excellent recoveries of arsenic were obtained when using the optimum electrothermal atomic absorption spectrometry conditions to analyze standard solutions of arsenobetaine, arsenocholine and tetramethylarsonium iodide. This procedure also allowed for the accurate determination of arsenic in Californian mussels collected from a variety of locations along the B.C. coast. Finally, in Chapter 6 it was demonstrated that MAsA is a likely precursor to MB in the marine environment. Mytilus calfornianus exposed for 9 days to a seawater system containing [ H]-methylarsonic acid, was found to contain [ 3 H]-methylarsonic acid along 3 with [ H]-arsenobetaine and two unknown 3 3 H-labeled compounds in the tissue parts of this mussel. A linear increase with time in the specific activity present in the flesh of Mytilus calfornianus was also observed, indicating uptake of the labelled compound or its metabolites. The highest specific activity was found in the visceral mass and the gills of the mussel.  167  REFERENCES 1.  Cullen, W.R.; Reimer, K.J. Chem. Rev. 1989, 89, 713.  2.  National Research Council, Arsenic; National Academy of Sciences: Washington, DC, 1977.  3.  Tanaka, T. Appl. Organomet. Chem. 1988, 2, 283.  4.  National Research Council of Canada The Effects ofArsenic in the Canadian Environment; Ottawa, 1978.  5.  Ishiguro, S. Appi. Organomet. Chem. 1992, 6, 323.  6.  Chilvers, D.C.; Peterson, P.J. In Lead Mercury, Cadmium andArsenic in the Environment; Hutchinson, T.C., Heerna, KM. Eds.; Wiley: New York, 1987; p 279.  7.  Squibb, K.S.; Fowler, BA. In Biological and Environmental Effects ofArsenic; Fowler, B.A., Ed.; Elsevier: Amsterdam, 1983; p 233.  8.  Shibata, Y.; Morita, M.; Fuwa, K. Adv. Biophys. 1992, 28, 31.  9.  Bertrand, G. Comptes Rendus Hebdomadaires des Seances de 1’ Academie des Sciences 1902, 134, 1434.  10.  Jones, A.J. In Year Book ofPharmacy, Transactions of the British Pharmaceutical Conference; Hampshire, C.H. Ed.; London, 1922; pp 388-395.  11.  Cox, H.E. TheAnalyst 1925, 50, 3.  12.  Edmonds, J.S.; Francesconi, K.A. Nature 1977, 265, 436.  13.  Francesconi, K.A.; Edmonds, J.S. Oceanogr. Mar. Biol. Annu. Rev. 1993, 31, 111.  14.  Lawrence, J.F.; Michalik, P.; Tam, G.; Conacher, H.B.S. J Agric. Food Chem. 1986, 34, 315.  15.  Edmonds, J.S.; Shibata, Y.; Francesconi, K.A.; Yoshinaga, J.; Morita, M. Sci. Tot. Environ. 1992, 122, 321.  16.  Maher, W.A. Compar. Biochem. Physiol. 1985, 80C, 199.  17.  Luten, J.B.; Riekwel-Booy, G.; Greef J.v.d; Ten Noever De Brauw, M.C. Chemosphere 1983, 12, 131.  168  18.  Francesconi, K.A.; Edmonds, J.S.; Hatcher, B.G. Compar. Biochem. Physiol. 1988, 90C, 313.  19.  Shiomi, K.; Shinagawa, A.; Azuma, M.; Yamanaka, H.; Kikuchi, T. Compar. Biochem. Physiol. 1983, 74C, 393.  20.  Shiomi, K.; Shinagawa, A.; Azuma, M.; Yamanaka, H.; Kikuchi, T. Nippon Suisan Gakkaishi (Bull. Jpn. Soc. Sci. Fish.) 1983, 49, 79.  21.  Kaise, T.; Fukui, S. Appi. Organomet. Chem. 1992, 6, 155.  22.  Irvin, T. R.; Irgolic, K.J. Appi. Organomet. Chem. 1988, 2, 509.  23.  Norm, H.; Christakopoulos, A. Chemosphere 1982, 11, 287.  24.  Norm, H.; Ryhage, R.; Christakopoulos, A.; Sandsträm, M. Chemosphere 1983, 12, 299.  25.  Lunde, G. .1 SeE. FoodAgric. 1975, 26, 1257.  26.  Penrose, W.R.; Conacher, H.B.S.; Black, R.; Méranger, J.C.; Miles, W.; Cunningham, H.M.; Squires, W.R. Environ. Health Perspect. 1977, 19, 53.  27.  Morita, M.; Shibata, Y. Anal. Sd. 1987, 3, 575.  28.  Shiomi, K.; Kakehasi, Y.; Yamanaka, H.; Kikuchi, T. Appl. Organomet. Chem. 1987, 1, 177.  29.  Norm, H.; Christakopoulos, A.; Sandstrom, M.; Ryhage, R. Chemosphere 1985, 14, 313.  30.  Francesconi, K.A.; Edmonds, J.S.; Stick, R.V. J. Chem. Soc., Perkin Trans. 1 1992, 1349.  31.  3m, K.; Hayashi, T.; Shibata, Y.; Morita, M. Appi.  32.  Jin, K.; Shibata, Y.; Morita, M. AgrEe. Biolog. Chem. 1988, 52, 1965.  33.  Shibata, Y.; Morita, M.; Edmonds, J. S. AgrEe. Biol. Chem. 1987, 51, 391.  34.  Morita, M.; Shibata, Y. Chemosphere 1988, 17, 1147.  35.  Edmonds, J.S.; Francesconi, K.A. I Chem. Perkin Trans. 11982, 2989.  169  Organomet. Chem. 1988, 2, 365.  36.  Edmonds, J.S.; Francesconi, K.A. Nature 1981, 289, 602.  37.  Edmonds, J.S.; Francesconi, K.A. I Chem. Perkin Trans. I 1983, 2375.  38.  Edmonds, J.S.; Morita, M. I Chem. Perkin Trans. I 1987, 577.  39.  Shibata, Y.; Morita, M. Agric. Biol. Chem. 1988, 52, 1087.  40.  Francesconi, K.A.; Stick, R.V.; Edmonds, J.S. I Chem. Soc. Chem. Commun. 1991, 928.  41.  Irgolic, K.J.; Woolson, E.A.; Stockton, R.A.; Newman, R.D.; Bottino, N.R.; Zingaro, R.A.; Kearney, P.C.; Pyles, R.A.; Maeda, S.; McShane, W.J.; Cox, E.R. Environ. Health Perspectives 1977, 19, 61.  42.  Edmonds, J.S.; Francesconi, K.A. Experientia 1987, 43, 553.  43.  Thayer, J. S. In Organometallic Compounds and Living Organisms; Academic: New York, 1984; p 189.  44.  Challenger, F. Chem. Rev. 1945, 36, 315.  45.  Challenger, F.; Lisle, D.B.; Dransfield, P.B. .1 Chem. Soc. 1954, 1760.  46.  Cullen, W.R.; McBride, B.C.; Reglinski, J. I Inorg. Biochem. 1984, 21, 45.  47.  Cullen, W.R.; McBride, B.C.; Reglinski, J. I Inorg. Biochem. 1984, 21, 179.  48.  Beauchemin, D.; Bednas, M.E.; Berman, S.S.; McLaren, J.W.; Siu, K.W.M.; Sturgeon, R.E. Anal. Chem. 1988, 60, 2209.  49.  Larsen, E.H.; Pritzl, G.; Hansen, S.H. I Anal. At. Spectrom. 1993, 8, 557.  50.  Shibata, Y.; Morita, M. AppL Organomet. Chem. 1992, 6, 343.  51.  Anderson, R.K.; Thompson, M.; Culbard, E. Analyst 1986, 111, 1143.  52.  Anderson, R.K.; Thompson, M.; Culbard, E. Analyst 1986, 111, 1153.  53.  Aggett, J.; Aspell, A.C. Analyst 1976, 101, 341.  54.  Andreae, M.O. Anal. Chem. 1977, 49, 820.  55.  Braman, R.S.; Foreback, C.C. Science 1973, 182, 1247.  170  56.  Braman, R.S.; Johnson, D.L; Foreback, C.C.; Ammons, J.M.; Bricker, J.L. Anal. Chem. 1977, 49, 621.  57.  Rabbins, W.B.; Caruso, J.A. Anal. Chem. 1979, 51, 889A.  58.  Godden, R.G.; Thomerson, DR. Analyst 1980, 105, 1137.  59.  Nakahara, T. Frog. Analyt. At. Spectrosc. 1983, 6, 163.  60.  Welz, B. Chem. Brit. 1986, February, 130.  61.  Reimer, K.J. Appl. Organomet. Chem. 1989, 3, 475.  62.  Odanaka, Y.; Tsuchiya, N.; Matano, N.; Goto, S. Anal. Chem. 1983, 55, 929.  63.  Cullen, W.R.; Hao, L.; Pergantis, S.A.; Eigendorf G.K.; Harrison, L.G. Chemosphere 1994 (In press).  64.  Le, X-C.; Cullen, W.R.; Reimer, K.J. Talanta 1993, 40, 185.  65.  Atallah, R.J.; Kalman, D.A. Talanta 1991, 38, 167.  66.  Brinckman, F.E.; Jewett, K.L.; Iverson, W.P.; Irgolic, K.J.; Ehrhardt, K.C.; Stockton, R.A. I Chromatogr. 1980, 191, 31.  67.  Shiomi, K.; Orii, M.; Yamanaka, H.; Kikuchi, T. Nippon Suisan Gakkaishi (Bull. Jpn. Soc. Sci. Fish.) 1987, 53, 103.  68.  Hansen, S.H.; Larsen, E.H.; Pritzl, G.; Cornett, C. .1 Anal. At. Spectrom. 1992, 7, 629.  69.  Shibata, Y.; Morita, M. Anal. Sd. 1989, 5, 107.  70.  Stockton, R.A.; Irgolic, K.J. Intern. I Environ. Anal. Chem. 1979, 6, 313.  71.  Slavin, W.; Schmidt, G.J. I Chromatogr. 1979, 17, 610.  72.  Ebdon, L.; Hills, S.; Jones, P.1 Anal. At. Spectrom. 1987, 2, 205.  73.  Nygren, 0.; Nilsson, C.A.; Frech, W. Anal. Chem. 1988, 60, 2204.  74.  Cullen, W.R.; Dodd, M. Appl. Organomet. Chem. 1989, 3, 401.  75.  Gast, C.H.; Kraak, J.C.; Poppe, H.; Maessen, F.J.M.J. I Chromatogr. 1979, 185, 549.  171  76.  Florence, T.M. Talanta 1982, 29, 345.  77.  Van Loon, J.C.; Barefoot, R.R. Analyst 1992, 117, 563.  78.  Luten, J.B.; Riekwel-Booy, G. Chemosphere 1983, 12, 131.  79.  Luten, lB.; Riekwel-Booy, G.; Rauchbaar, A. Environ. Health Perspectives 1982, 45, 165.  80.  Lau, B.Y.; Michalik, P.; Porter, C.J.; Krolik, S. Biomed. Environ. Mass Spectrom. 1987, 14, 723.  81.  Siu, K.W.M.; Gardner, G.J.; Berman, S.S. Rapid Commun. Mass Spectrom. 1988,2, 69.  82.  Siu, K.W.M.; Guevremont, R.; Le Blanc, J.C.Y.; Gardner, G.J.; Berman, S.S. J Chromatogr. 1991, 554, 27.  83.  a) Pergantis, S.A.; Madilao, L.L.; Emond, C.A.; Eigendorf G.K. Proceedings of the 41st ASMS Conference on Mass Spectrometry and Allied Topics, San Fransisco, 1993, 994a. b) Pergantis, S.A.; Emond, C.A.; Madilao, L.L.; Eigendorf G.K. Org. Mass Spectrom. (In press).  84.  Thomson, J.J. Rays ofPositive Electricity and Their Application to Chemical Analyses, Green & Co., London, 1913.  85.  Aston, F.W. Mass Spectra and Isotopes, Arnold, London, 1942.  86.  Baldwin, M.A.; McLafferty, F.W. Org. Mass Spectrom. 1973, 7, 111.  87.  Caccamese, S.; Amico, V.; Hardy, M. Bull. Soc. Chim. Fr. 1988, 1, 91.  88.  McElvany, S.W.; Callahan, J.H. .1 Phys. Chem. 1991, 95, 6186.  89.  Monneret, C.; Sellier, N. Biomeci Environ. Mass Spectrom. 1986, 13, 319.  90.  Thimmaiah, K.N.; Thomas, M.J.; Sethi, V.S.; Made Gowda, N.M. Microchem. J 1990, 41, 183.  91.  Guglielmetti, G.; Dalcanale, E.; Bonsignore, S.; Voncenti, M. Rapid Commun. Mass Spectrom. 1989, 3, 106.  92.  Vincenti, M.; Pelizzetti, E.; Guarini, A.; Constanzi, S. AnaL Chem. 1992, 64, 1879.  172  93.  Hansen, G.; Munson, B. Anal. Chem. 1980, 52, 245.  93a. Beuhier, RI.; Flanigan, E.; Greene, L.J.; Friedman, L. I Am. Chem. Soc. 1974, 96, 3990. 93b. Daves, G.D.Jr. Ace. Chem. Res. 1979, 12, 359. 94.  Harrison, A.G. Chemical Ionization Mass Spectrometry; CRC Press, Inc., 1983.  95.  Allan, A.R.; Roboz, 3. Rapid Commun. Mass Spectrom. 1988, 2, 235.  96.  Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299.  97.  Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151.  98.  Hillenkamp, F.; Karas, M.; Beavis, R.C.; Chait, B.T. Anal. Chem. 1991, 63, 1 193A.  99.  Danis, P. 0.; Karr, D. E.; Holle, A.; Mayer-Posner F. Proceedings of the 4IstASMS Co,?ference on Mass Spectrometry andAllied Topics, San Fransisco, 1993, 1093a.  100. Bileci, T. M.; Stults, J. T. Anal. Chem. 1993, 65, 1709. 101. Nordhoff E.; Karas, M.; Hillenkamp, F.; Kirpekar, F.; Kristiansen K.; Roepstorfi P. Proceedings of the 41st ASMS Conference on Mass Spectrometry andAllied Topics, San Fransisco, 1993, 246a. 102. Mock, K.K.; Davey, M.; Cottrell, J.S. Biochem. Biophys. Res. Commun. 1991, 177, 644. 103. Cotter, R JAnal. Chem. 1992, 64, 1027A. 104. Edmonds, J.S.; Francesconi, K.A.; Cannon, J.R.; Raston, C.L.; Skelton, B.W.; White, A. Tetrahedron Lett. 1977, 18, 1543. 105. Irgolic, K.J.; Junk, T.; Kos, C.; Mcshane, W.S.; Parppalardo, C.C. Applied Organomet. Chem. 1987, 1, 403. 106. Kaise, K.; Hanaoka, K.; Tagawa, S. Chemosphere 1987, 16, 2551. 107. Kaise, T, Yamauchi, H, Hirayama, T and Fukui, S App!. Organomet. Chem. 1988, 2, 339. 108. Lohe, C.; Kohl, C.D. I Vac. SeE. TechnoL B 1989, 7, 217.  173  109. Dzidic, I.; Desiderio, D.M.; Wilson, M.S.; Cram, P.F.; McCloskey, J.A. Anal. Chem. 1971, 43, 1877. 110. Lawrence, D.L. Rapid Commun. Mass Spectrom. 1990, 4, 546. 111. Ballantine, J.A.; Barton, J.D.; Carter, J.F.; Fussell, B. Org. Mass Spectrom. 1987, 22, 316. 112. Langley, G.L. Org. Mass Spectrom. 1990, 25, 429. 113. Bruins, A.P. Biomedical Mass Spectroin. 1980, 7, 454. 114. Brilis, G.M.; Brumley, W.C. Anal. ChimicaActa 1990, 229, 163. 115. Heller, D.N.; Chen, T.S.; Hansen, G.; Fenseleau, C. In Proceedings of the 29th Conference on Mass Spectrometry and Allied Topics; Minneapolis, MN, 1981 p. 574. 116. Guarini, A.; Guglielmetti, G.; Vincenti, M.; Guarda, P.; Marchionni, G. AnaL Chem. 1993, 65, 970. 117. Maeder H.; Gunzelmann, K.H. Rapid Commun. Mass Spectrom., 1988, 2, 199. 118. McCamish, M.; Allan R.; Roboz, J. Rapid Commun. Mass Spectrom. 1987, 1, 124. 119. Smith, C.H.; Noyes, R.A.; Kelly, P.B.; Chang, D.P.Y.; and Jones, A.D. Proceedings of the 41st ASMS Conference on Mass Spectrometry and Allied Topics, San Fransisco, 1993, 116a. 1 19a. Li, L.; Wang, A.P.L.; Coulson, L.D. Anal. Chem. 1993, 65, 493. 1 19b.Murray, K.K.; Russell, D.H. Anal. Chem. 1993, 65, 2534. 119c.Murray, K.K.; Russell, D.H. Anal. Chem. 1994, 5, 1. 120. Proudfoot, F.G.; Jackson, E.D.; Hulan, H.W.; Salisbury, C.D.C. Can. J Anim. Sd. 1991, 71, 221. 121. Worden, A.N.; Wood, E.C. J Sd. FdAgric. 1973, 24, 35. 122. Nelson, R.W.; McLean, M.A.; Vestal, M.L. Proceedings of the 40th ASMS Conference on Mass Spectrometry andAllied Topics, Washington, DC ,1992, 1919. 123. Preston, L.M.; Murray, K.K.; Russell, D.H. Biol. Mass Spectrom. 1993, 22, 544.  174  124. Li, H. Ph.D. Thesis ; University of British Columbia, 1994 (In preparation). 125. Comber, S.D.W.; Howard, A.G. Anal. Proceedings 1989, 26, 20. 126. Maeda, S.; Wada, H.; Kumeda, K.; Onoue, M.; Ohki, A.; Higashi, S.; Takeshita, T. Appi. Organomet. Chem. 1987, 1, 465. 127. Odanaka, Y.; Tsuchiya, N.; Matano, N.; Goto, S. Anal. Chem. 1983, 55, 929. 128. Lussi-Schlatter, B.; Brandenberger, H. In Advances in Mass Spectrometiy in Biochemistry and Medicine ; Spectrum Publications; New York 1976, Vol.2 pp. 231 129. Lussi-Schlatter, B; Brandenberger, H. Z. Kim. Chem. Biochem. 1974, 12:224 130. Cullen, W.R.; Froese, C.L.; Lui, A.; McBride, B.C.; Patmore, D.J.; Reimer, M. I Oganomet. Chem. 1977, 139, 61. 131. Slavin, W. In Graphite Furnace AAS: A Source Boolç Perkin-Elmer; Ridgefield, 1984. 132. Subramanian, K.S. Can. I Spectrosc. 1988, 33, 6. 133. Matousek, P. Prog. Anal. At. Spectrosc. 1981, 4, 247-310. 134. Riley, W. At Spectrosc. 1982, 3, 120. 135. Brzezinska-Paudyn, A.; Van Loon, A.; Hancock, R. At. Spectrosc. 1986, 7, 72. 136. Ediger, R.D. At. Absorp. News!. 1975, 14, 127. 137. Newsteacl, R.A.; Price, W.J.; Whiteside, P.J. Prog. Analyt. At. Spectrosc. 1978, 1, 267. 138. Chakrabarti, D.; De Jonghe, W.; Adams, F. Anal. Chim. Ada 1980, 119, 331. 139. Betteridge, D.; Wade, A.P.; Howard, AG. Talanta 1985, 32, 709. 140. Betteridge, D.; Wade, A.P.; Howard, AG. Talanta 1985, 32, 723. 141. Walsh, A. Spectrochim. Acta 1955, 7, 108. 142. L’vov, B.V. Spectrochim. Acta 1961, 17, 761. 143. Rothery, E.; In Analytical Methods for Graphite Tube Atomizers; Varian Techtron, Victoria, Australia, 1982.  175  144. Fabec, J.L. Anal. Chem. 1982, 52, 2170. 145. Brooks, R.R.; Douglas; R,E.; Zhang, H. Anal. Chim. Acta 1981, 131, 1. 146. Lawrence, H.K. In: E.P.A. Lewis Publishers, Inc., 1990.  Sampling and Analysis Methods, Database Manual  147. Schmidt, F.J.; Royer, J.L. Anal. Lett. 1973, 6, 17. 148. Pollock, E.N.; West, S.J. Ibid 1973, 12, 6. 149. Fernandez, F.J. At. Absorpt. NewsL 1973, 12, 93. 150. Thompson, K.C.; Thomerson, D.R. Analyst 1974, 99, 595. 151. Hershey, J.W.; Keliher, P.N. Appl. Spectros. Rev. 1990, 25, 213. 152. Van Loon, J.C.; In: Selected methods of Trace Analysis: Biological and Environmental Samples, Wiley: New York, 1985. 153. Kingston, H.M.; Jassie, L.B. In: Introduction to Microwave Sample Preparation: Theory and Practice, American Chemical Society, Washington, DC, 1988. 154. Matusiewicz, H.; Sturgeon, R.E.; Berman, S.S. .1 Anal. At. Spectrom. 1989, 4, 323. 155. Schachter, M.M.; Boyer, K.W. Anal. Chem. 1980, 52, 360. 156. De Boer, J.L.M.; Maessen, F.J.M.J. Spectrochim. Acta 1983, 38B, 739. 157. Abu-Samra, A.; Morris, J.S.; Koirtyohann, S.R. Anal. Chem. 1975, 47, 1475. 158. Barrett, P.; Davidowski, Jr.L.J.; Penaro, K.W.; Copeland, T.R. Anal. Chem. 1978, 7, 1021. 159. Nadkarni, R.A. AnaL Chem. 1984, 56, 2233. 160. Matthes, S.A.; Farrell, R.F.; Mackie, A.J. Tech. Prog. Rep. 120.  -  US, Bur. Mines, 1983,  161. Lamothe, P.J.; Fires, T.L.; Consul, 3.3. Anal. Chem. 1986, 58, 1881. 162. Fernando, L.A.; Heavner, W.D.; Gabrielli, C.C. Anal. Chem. 1986, 58, 511.  176  163. Papp, C.S.E.; Fischer, L.B. Analyst 1987, 112, 337. 164. Kingston, H.M.; Jassie, L.B. Anal. Chem. 1986, 58, 2534. 165. Nakashima, S.; Sturgeon, R.; Willie, L.B.S.; Berman, S. Analyst 1988, 113, 159. 166. Spendley, W.; Hext, G.R.; Himsworth, F.R. Technometrics 1962, 4, 441. 167. Deming, S.N.; Morgan, S.L. Anal. Chem. 1973, 45, 278A. 168. Morgan, S.L.; Deming, S.N. Anal. Chem. 1974, 46, 1170. 169. Betteridge, D.; Sly, T.J.; Wade, A.P.; Tiliman, J.E.W. Anal. Chem. 1983, 55, 1292. 170. Bajo, S.; Suter, U.; Aeschliman, B. Anal. Chim. Acta 1983, 149, 321. 171. Wade, A.P.; Shiundu, P.M.; Wentzell, P.D. Anal. Chim. Acta 1990, 237, 361. 172. Krivan, V.; Arpadjan, S. Frezenius Z Anal. Chem. 1989, 335, 743. 173. Fernandez, F.J.; Giddings, R. At. Spectrosc. 1982, 3, 61. 174. Larsen, E.H. J Anal. At. Spectroc. 1991, 6, 375. 175. Holak, W. Anal. Chem. 1973, 41, 1712. 176. Goulden, P.D.; Brooksbank, P. Anal. Chem. 1974, 46, 1431. 177. Braman, R.S.; Justin, L.L.; Foreback, C.C. Anal. Chem. 1972, 44, 2195. 178. Dodd, M. Ph.D. Thesis, University of British Columbia, 1988. 179. Bye, R. Talanta 1990, 37, 1029. 180. Chen, H.; Brindle, I.D.; Le, X-C. Anal. Chem. 1992, 64, 667. 181. Takahashi, Y.; Ono, T.; Yokoyama, T.; Taruntami, T. Chinetsu 1987, 24, 383. 182. Matsuto, S.; Stockton, R.A.; Irgolic, K.J. Sd. Total Emviron. 1986, 48, 133. 183. Smith, L. Undergraduate Thesis, University of British Columbia, 1989. 184. Le, X-C Ph.D. Thesis, University of British Columbia, 1993.  177  185. Huang, J.; Goltz, D.; Smith, F. Talanta 1988, 35, 907. 186. Friel, J.K.; Skinner, C.S.; Jackson, S.E.; Longerich, H.P. Analyst 1990, 115, 269. 187. Nakao, M. I Osaka City Med. Center (Osaka Shiritsu Daigaku Igakubu Zasshz) 1960, 9, 541. 188. Ishizaki, M. Nippon Eiseigaki Zasshi. 1979, 34, 605. 189. World Health Organization, Trace Elements in Human Nutrition, WHO technical series No. 532. World Health Organization, Geneva, 1973. 190. Maher, W.A. Marine Biol. Lett. 1984, 5, 47. 191. Gilbert, P.R.; Wells, G.A.H.; Gunning, R.F. Vet. Rec. 1981, 109, 158. 192. Health and Welfare Canada, In: Drugs Directorate, Ottawa, ON 1981. 193. Agriculture Canada, In: Compedium ofMedicating Ingredients Bronchures, 5th ed. Ottawa, ON 1984. 194. Weston, R.E.; Wheals, B.B.; Kensett, M.J. Analyst, 1971, 96, 601. 195. Hoodless, R.A.; Tarrant, K.R. Analyst, 1973, 98, 502. 196. Morrison, J.L. I Agr. Food Chem., 1968, 16, 704. 197. Maruo, M.; Hirayama, N.; Wada, H.; Kuwamoto, T. I Chromatogr. 1989, 466, 379. 198. Hirayama, N.; Kuwamoto, T. I Chromatogr. 1988, 457, 415. 199. Ishii, D. In: Introduction to Microscale High-Performance Liquied Chromatography, VCH Publishers Inc., 1988. 200. Erlich, P.; Bertheim, D. Berichte 1907, 40, 3292. 201. Becker, E.R.; Morehouse, N.F. I Parasitol. 1936, 22, 60. 202. Morehouse, N.F.; Mayfield, O.J. I ParasitoL 1946, 32, 20. 203. Humayoun Akhtar, M.; Patterson, J.R.; Salisbury, C.D.C.; Ho, S.K.; Jui, P.Y.; Hartin, K.E., Can. I Anim. Sd., 1992, 72, 389.  178  204. Edmonds, M.S.; Baker, D.H. I Anim. SeE. 1986, 63, 553. 205. Edmonds, M.S.; Baker, D.H. I Anim. SeE. 1986, 63, 553. 206. Rice, D.A.; McMurray, C.H.; McCracken, R.M.; Bryson, D.G.; Maybin, R. Vet. Rec. 1980, 106, 312. 207. Horváth, C.G.; Preiss, B.A.; Lipsky, S.R. Anal. Chem. 1967, 39, 1422. 208. Ishii, D. Jasco Report, 1974, 11 (no. 6). 209. Ishii, D.; Asai, K.; Hibi, K.; Jonokuchi, T.; Nagaya, M. .1 Chromatogr. 1977, 144, 157. 210. Knox, J.H. I Chromatogr. SeE. 1980, 18, 453. 211. Scott, R.P.W.; Scott, C.J.; Munroe, M.; Hess, J. I Chromatogr. 1974, 99, 212. McFadden, W.H.; Schwartz, H.L.; Evans, A. I Chromatogr. 1976, 122, 389. 213. Tal’roze, V.L.; Skurat, V.E.; Gorodetskii, I.G.; Zolotai, N.B. Russ. I Phys. Chem. 1972, 46, 456. 214. Blakley, C.R.; McAdams, M.J.; Vetsal, M.L. I Chromatogr. 1978, 158, 261. 214a.van de Greef J.; Niessen, W.M.A. 857.  mt. I Mass Spectrom. Ion Proc.  1992, 118/119,  215. Blakley, C.R.; Vetsal, M.L. Anal. Chem. 1983, 55, 750. 216. Caprioli, R.M.; Suter, M.J.F. Intern. I Mass Spectrom. Ion Proc. 1992, 118/119, 449.  217. Caprioli, R.M.; Fan, T.; Cottrell, J.S. Anal. Chem. 1986, 58, 2949. 218. Fenn, J.B.; Mann, M.; Meng, C.K.; Wong, S.F.; Whitehouse, C.M. Science, 1989, 264, 64.  219. Fenn, J.B.; Mann, M.; Meng, C.K.; Wong, S.F.; Whitehouse, C.M. Mass Spectrom. Rev. 1990, 9, 37. 220. Fenn, J.B. I Am. Soc. Mass Spectrom. 1993, 4, 524. 221. Guevremont, R.; Siu, K.W.M.; Le Blanc, J.C.Y.; Berman, S.S. I Am. Soc. Mass Spectrom. 1992, 3, 216.  179  222. Loo, J.A.; Edmonds, C.D.; Udseth, H,R.; Smith, RD. Anal. Chem. 1990, 62, 693. 223. Arpino, P. Mass Spectrom. Rev. 1990, 9, 631. 224. Pacey, G.E.; Ford, J.A. Talanta 1981, 28, 935. 225. Woolson, E.A.; Aharowsow, N. J Assoc. Off Anal. Chem. 1980, 63, 523. 226. Barber, M.; Bordoli, R.S.; Sedwick, R.D.; Tyler, A.N. I Chem. Soc., Chem. Commun., 1981, 325. 227. Gower, J.L. Biomeci Mass Spectrom., 1985, 12, 191. 228. Visentini, J.; Gauthier, J.; Bertrand, M.J. Rapid Commun. Mass Spectrom., 1989, 3, 390. 228a.de Wit, J.S.M.; Deterding, L.J.; Moseley, M.A.; Tomer, K.B.; Jorgenson, J.W. Rapid Commun. Mass Spectrom., 1988, 2, 100. 228b. Pleasance, S.; Thibault, P.; Moseley, M.A.; Deterding, L.J.; Tomer, K.B.; Jorgenson, J.W. I Am. Soc. Mass Spectrom., 1990, 1, 312. 229. Francesconi, K.A. Ph.D. Thesis, The University of Western Australia, 1991. 230. Maugh II, T.H. Science 1979, 203, 637. 231. Cullen, W.R.; Dodd, M. Appl. Organomet. Chem. 1989, 3, 79. 232. Francesconi, K.A.; Edmonds, J.S.; Stick R.V. Appi. Organomet. Chem. 1992, 6, 247. 233. Francesconi, K.A.; Edmonds, J.S.; Stick R.V. Total Science of the Environm. 1989, 79, 59. 234. Francesconi, K.A.; Edmonds, J.S. Compar. Biochem. Physiol. 1987, 87C, 345. 235. Edmonds, J.S.; Francesconi, K.A.; Healy, P.C.; White, AC. I Chem Soc., Perkin Trans. 11982, 1989. 236. Cannon, J.R.; Saunders, J.B.; Toia, R.F. Sci. Total Environ. 1986, 31, 181. 237. Shiomi, K.; Shinagawa, A.; Igarashi, T.; Hirota, K.; Yamanaka, H.; Kikuchi, T. Nippon Suisan Gakkaishi (Bull. Jpn. Soc. Sci. Fish.) 1984, 50, 293.  180  238. Francesconi, K.A.; Micks, P.; Stockton, R.A.; Irgolic, K.J. Chemosphere 1985, 14, 1443. 239. Lunde, G. Acta Chem. Scandin. 1973, 27, 1586. 240. Lai, V. Undergraduate Thesis, University of British Columbia, 1994 (In preparation). 241. Cullen, W.R.; Dodd, M.; Nwata, B.U.; Reimer, D.; Reimer, K.J. Appi. Organomet. Chem. 1989, 3, 351. 242. Woolson, E.A. Topics in Environmental Health; Biological and Environmental Effects ofArsenic, Elsevier, Amsterdam, 1983, 6, 51. 243. Walsh, L.M.; Sumner, M.E; Keeney, DR. Environ. Health Perspect. 1977, 19, 67. 244. Anake, M.; Schneider, H.J.; Bruechner, C. Spurenelem Symposium, Arsenic, 3, Friedrich-Shiller-University, Jena, 1980. 245. Lederer, W.H.; Fensterheim, R.J. Arsenic: Industrial Biomedical Environmental Perspectives (Proc. Arsenic Symp. 1981), Van Nostrand, New York, 1983. 246. Gugger, P.; Willis, C.A.; Wild, S.B. I Chem. Soc., Chem. Commun. 1990, 1169. 247. Andreae, M.O. Deep Sea Res. 1978, 25, 391. 248. Andreae, M.O. Limnol. Oceanigr. 1979, 24, 440. 249. Reimer, K.J.; Thompson, J. A.J. unpublished results. 250. Francesconi, K.A.; Stick, R.V.; Edmonds, J.S. Experientia 1990, 46, 464. 251. UnlU, M.Y.; Fowler, S.W. Marine Biology 1979, 51, 209. 252. Klumpp, D.W. Marine Biology 1980, 58, 265. 253. Cullen, W.R.; Nelson, J.C. Appi. Organomet. Chem. 1993, 7, 319. 254. Cullen, W.R.; Mcbride, B.C.; Pickett, A.W.; Hasseini, M. Appl. Organomet. Chem., 1988, 3, 71. 255. Lisk, D.J. Agric. and Food Chem. 1960, 8, 2. 256. Ober, R.E.; Hansen, A.R.; Mourer, D.; Baukema, 3.; Gwynn, G.W. Inter. I of Appi. RacL and Isotopes 1969, 20, 703.  181  257. UnIU, M.Y. Chemosphere 1979, 8, 269. 258. Wrench, J.; Fowler, S.W.; UnlU, MY. Mar. Pollut. Bull. 1979, 10, 18. 259. Edmonds, J.S.; Francesconi, K.A. J Chem. Soc., Perkins Trans. 1 1983, 2383. 260. Pentreath, R.J. .1 Mar. Biol. Ass. UK 1973, 53, 127. 261. Fowler, S.W.; Heyraud, M.; Beasley, T.M. Vienna: International Atomic Energy Agency, 1975, 157. 262. George, S.G.; Pine, B.J.S.; Coombs, T.L. .1 Exp. Mar. Biol. Ecol. 1976, 23, 71.  182  APPENDIX A Thermospray mass spectra of arylarsenicals  —  C) Cu C)  :  o Cu -  5  t  —  H  -  ;  +  C., --C.,  +  -  —  -  I  I  I  I C)  p.  I  I —  £;!suuI A!WI)I Therm ospray mass spectrum of 4-nitrophenylarsonic acid. Temperature settings: Probe at 129 °C, vaporizer at 185 0 C, source block at 223 °C, and jet at 120 °C.  183  i  qoij  Do OZJ  :suus  V81  i poq amos ‘Do 8J i 1zuodA ‘Do 611 uosiupCuqdxoip- jo uin.ods ssw idsow.iij  i1iu ‘Do pp  Relativeintensity  I  I  I  I  I  I  I  I  I  I .  —  I °1= 1  8I  if i poq amos ‘j çj i izuodA ‘30 6Z1 w qoij ‘Do :suus aEfl11QdtUj. pp oruosi pCuqd&xoJpi(4-fr-011!U-E jo unuds ssw idsouuq  Relative Intensity  .1  + C’  + +  C’  ‘1—f Z  C  a.  —.  •  0—  —  100  16  150  154  *  150  I  214  214  200  1  248  +  +  i 250  r 300  289  300  t  289  (+) LSIMS  250  Ii  246  rrpr  176  200  199  *  1  329  tTh  350  {  350  L 1  1  347  L  r  *  i  1  502  Z  +  400  T LI  1 L  450  441  500  M!Z  M/Z  ii  518  473  500  516  383  *  450  1 T- , Fi T 4  442  +  45  400  395  z  I’  +  z  3  381  z+  LSIMS mass spectra of 4-nitrophenylarsonic acid in negative and positive detection mode.  107 *  *  100  107  *  168  306  *  153  (-) LSIMS  I  0  *  i  CD  Cl)  —  LST  pow uopmp ArnSOd pu AflU Ut  po njiutsi-d jo uxids sw siisi  Relative Intensity 0  I  I  I  -  -  -  II—  --  —  [M+H]  -  [M-H]  c12 —%  0z  -  *  r  —  [M+NBA,-H20,-H]  [M+NBA,-H2O,+H] w  “3  [M+2NBA,-2H2O,+H]  [M+2NBA,-2H 0 2 ,-H]  —  uoup  AtTsOd  pu  881  ui pi owosi  .pow uqdiCxoip(q- jo ids ssm swisi  Relative Intensity 0  0  I  —  —  0*  .—  -  E_ -I  L  [M+H]  +  -  [M-Hj  Cfi  C1 C17  0  :*  +  +  2  2 “I  C -  +  2 —*  —  [M+2NBA,2H2O,+H1  —‘.I  U’  -  100%  50%  0%  128  130  148  158  168  178  188  198 MJZ  Mass spectrum of 2,6-dimethyiphenol obtained by using DLI-MS; amount injected 500 ng; ; ion source pressure: 10-1 1 mobile phase: 80% methanoll 20% water; flow rate: 2 pL.min torr; masses scanned: 120-140 amu; temperature of copper sleave 180 °C.  190  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061715/manifest

Comment

Related Items