On the reliability of methods for the speciation of mercury based on

chromatographic separation coupled to atomic spectrometric detection

by

Johanna Qvarnström

AKADEMISK AVHANDLING

Som med vederbörligt tillstånd av Rektorsämbetet vid Umeå Universitet för erhållande av filosofie doktorsexamen framlägges till offentlig granskning vid Kemiska Institutionen, Sal KB3A9 (Lilla hörsalen) KBC-huset, fredagen den 23

maj 2003, Klockan 13.00.

Fakultetsopponent: Dr. Jörg Feldmann, Department of Chemistry, University of Aberdeen, Old Aberdeen, Scotland, UK

Title: On the reliability of methods for the speciation of mercury based on chromatographic separation coupled to atomic spectrometric detection

Author: Johanna Qvarnström Department of Chemistry, Analytical Chemistry, Umeå University, S-90187 Umeå, Sweden

Abstract: This thesis deals with the reliability of methods for the speciation of mercury in

environmental and biological samples. Problems with speciation methods that couple chromatography to atomic spectrometric detection and how to overcome the problems are discussed. Analytical techniques primarily studied and evaluated are high performance liquid chromatography-cold vapour-atomic absorption spectrometry (HPLC-CV-AAS), HPLC-inductively coupled plasma-mass spectrometry (HPLC-ICP-MS), capillary electrophoresis-ICP-MS (CE-ICP-MS) and gas chromatography-ICP-MS (GC-ICP-MS). Applying a multi-capillary approach increased the analyte amount injected into a CE-ICP-MS system and improved the overall sensitivity. A microconcentric nebulizer with a cyclone spray chamber was shown to improve the detection limits for mercury species 3-13 times in HPLC-ICP-MS and 11-19 times in CE-ICP-MS compared to a cross-flow nebulizer with a Scott spray chamber. To decrease the interference of water vapour in HPLC-CV-AAS a Nafion dryer tube was inserted between the CV-generation and the detector. Methyl mercury was however lost in the Nafion unless it was reduced to elemental mercury prior transport through the dryer tube.

During sample pre-treatment, incomplete extraction, losses and transformation (alkylation, dealkylation, oxidation and reduction) of mercury species can lead to significant errors (underestimation and overestimation) in the determination of the concentrations. Methods to detect and determine the degree of transformation as well as correct for errors caused by transformation are presented in the thesis. The preferable method use species-specific enriched stable isotope standards in combination with MS detection and a matrix based calculation scheme. This approach is very powerful as both the concentrations of the species as well as the degrees of transformation can be determined within each individual sample.

Keywords: Mercury, speciation, hyphenated techniques, HPLC, CE, GC, CV-AAS, ICP-MS, species-specific enriched stable isotope

ISBN: 91-7305-429-1

Copyright, Johanna Qvarnström, 2003. All rights reserved. Printed by Solfjädern Offset AB

-ii-

This thesis includes the following papers, which are referred to in the text by their roman numerals.

I V. Majidi, J. Qvarnström, Q. Tu, W. Frech and Y. Thomassen

Improving sensitivity for CE-ICP-MS using multicapillary parallel separation

Journal of Analytical Atomic Spectrometry, 14(12): 1933-1935, 1999

II Q. Tu, J. Qvarnström and W. Frech

Determination of mercury species by capillary zone electrophoresis-inductively coupled plasma mass spectrometry: a comparison of two spray chamber–nebulizer combinations

Analyst, 125(4): 705-710, 2000

III J. Qvarnström, Q. Tu, W. Frech and C. Lüdke

Flow injection-liquid chromatography-cold vapour atomic absorption spectrometry for rapid determination of methyl- and inorganic mercury

Analyst, 125(6): 1193-1197, 2000

IV J. Qvarnström and W. Frech

Mercury species transformations during sample pre-treatment of biological tissues studied by HPLC-ICP-MS

Journal of Analytical Atomic Spectrometry, 17(11): 1486-1491, 2002

V J. Qvarnström, L. Lambertsson, S. Havarinasab, P. Hultman and W. Frech

Determination of methyl-, ethyl- and inorganic mercury in mouse tissues, following administraion of Thimerosal, by species-specific isotope dilution GC-ICP-MS

Manuscript submitted to Analytical chemistry

Papers I-IV are reproduced by permission of The Royal Society of Chemistry.

-iii-

Also by the author:

O. Al-Dirbashi, J. Qvarnström, K. Irgum and K. Nakashima

Simple and sensitive high-performance liquid chromatographic determination of methamphetamines in human urine as derivatives of 4-(4,5-diphenyl-1H-imidazol-2-yl) benzoyl chloride, a new fluorescence derivatization reagent

Journal of Chromatography B, 712(1-2): 105-112, 1998

-iv-

Table of Contents

1 Introduction.............................................................................................................................................1

1.1 Mercury ..............................................................................................................................................1

1.1.1 Physical and chemical properties of mercury.........................................................................................1

1.2 Speciation of mercury...................................................................................................................2

1.3 Chromatographic separation of mercury species ..............................................................3

1.3.1 HPLC............................................................................................................................................................................3

1.3.2 CE...................................................................................................................................................................................4

1.3.3 GC..................................................................................................................................................................................5

1.4 Spectrometric detection ...............................................................................................................7

1.4.1 CV-AAS........................................................................................................................................................................7

1.4.2 ICP-MS ........................................................................................................................................................................7

1.4.3 Other detectors used for the determination of mercury....................................................................9

1.5 Sample preparation ..................................................................................................................... 10

1.5.1 Alkaline dissolution...........................................................................................................................................10

1.5.2 Acid leaching ........................................................................................................................................................10

1.5.3 Microwave digestion .........................................................................................................................................11

1.5.4 Aqueous distillation...........................................................................................................................................11

1.5.5 Solvent extraction ...............................................................................................................................................11 2.0 Problems and possibilities......................................................................................................... 12

2.1 Chromatography coupled to atomic spectrometric detection................................... 12

2.1.1 HPLC-CV-AAS......................................................................................................................................................12

2.1.2 HPLC-ICP-MS ......................................................................................................................................................13

2.1.3 CE-ICP-MS .............................................................................................................................................................16

2.1.4 GC-ICP-MS ............................................................................................................................................................18

2.2 Transformation of species during analysis........................................................................ 18

2.2.1 Sources of transformation ..............................................................................................................................19

2.2.2 Determining the degree of transformation and correcting for errors caused by transformation .................................................................................................................................................................20

-v-

2.3 Using enriched stable isotopes............................................................................................... 22

2.3.1 Calculation of corrected signals .................................................................................................................23 3.0 Concluding remarks and future perspectives .................................................................. 25 References................................................................................................................................................. 27 Glossary...................................................................................................................................................... 33 Acknowledgements .............................................................................................................................. 35 Appendix Paper I-V

-vi-

1 Introduction

1.1 Mercury

For over 3500 years humans have used mercury for various applications. For example in the form of cinnabar, HgS, mercury was used as a red pigment in cosmetics and paint. Calomel, Hg2Cl2, was one of the ingredients in “teething powder” that prior 1960 was used to prevent ache in the mouth of children having their first teeth.

Metallic mercury has been used in a variety of laboratory equipment such as thermometers, barometers and diffusion pumps. Elemental mercury forms alloys, called amalgam, with for example gold and silver and this has been used in tooth fillings since the early nineteenth century. In the beginning of the last century alkyl mercury compounds were introduced as agricultural fungicides. This eventually led to numerous mercury poison disasters, the most serious disaster being in Iraq 1971-1972 where maybe as many as 40,000 people were poisoned of which 500 died.1

Even though more knowledge of its toxicity is now available, mercury in different forms is today still used as constituent in tooth filling material, pesticides, preservatives, catalysts, lamps and batteries. Lately there have been discussions about the potential neurotoxic effect of ethyl mercury thiosalicylic acid (see figure 1), a preservative used in vaccines, medicines and cosmetic products.2 But there is today not enough evidence for a clear relationship between exposure of this preservative and increased risk of neurodevelopmental disorders, and therefore it has not been banned.3

SHg

O

OH Figure 1. Ethyl mercury thiosalicylic acid, common names arethiomersal or thimerosal.

1.1.1 Physical and chemical properties of mercury

The melting point for metallic mercury is –38.8°C and it is hence the only common metal that is in a liquid state at room temperature. Air saturated with mercury vapour

at +20.0°C contains 13.18 mg mercury m-3 which many times exceed the toxic limit. (The concentration limit for metallic mercury in air in a working environment in Sweden is 0.03 mg m-3.4)

Because mercury in some forms is very volatile it can easily be transported in the atmosphere and it is therefore classified as a global pollutant. The solubility for elemental mercury in water is only 0.025 mg l-1 but as it can be transformed into other forms that are more soluble, mercury can be found almost everywhere in the

-1-

environment; for example in water, sediment, soil and fish. Mercury in its different forms is poisonous and can be inhaled, ingested and even absorbed through unbroken skin; it is bio-accumulated and bio-magnified and some organometallic mercury species, such as monomethyl- and dimethyl mercury, are known to be extremely toxic.

1.2 Speciation of mercury

Most elements are found in different forms, species, in the environment. They can have different oxidations states and they can be covalently or ionic bound to other atoms or ions. Because different species of the same element show differences in properties and toxicity, the European Union legislation calls for not only total determinations, but also species-specific analysis of mercury, tin, arsenic and selenium.

Mercury is found in the environment as elemental mercury, Hg0, inorganic mercury, Hg2

2+, Hg2+ and organo-mercury, Hg+II covalently bound to alkyl-groups. By biological activity in the environment, methyl mercury and dimethyl mercury are synthesised from inorganic mercury. Inorganic ionic mercury forms complexes with for example chloride, sulphur, humic- and fulvic acids. The diversity of existing mercury species and the inter-conversion between them makes it complicated to understand and provide a picture of the mercury pathways in nature. The concentration of mercury normally found in non-contaminated environmental and biological samples is close to detection limits for most methods, which further complicates the determination of mercury species and studies on its pathways in the environment.

A method for mercury speciation can include many steps and it is important to know that the original species composition in the sample is preserved throughout the analysis. If species are transformed or lost during storage, pre-treatment and analysis the determined concentration will not represent that of the original in the sample. Accurate results in the speciation analysis can be obtained if the integrity of mercury species is retained or if the degree of transformation, for example alkylation and dealkylation, can be quantified and corrected for. By understanding mechanisms and conditions for various possible reactions, which might take place during the analysis analytical errors might be avoided by choosing suitable conditions during sample work-up or compensated for, by taking into account redistribution, transformation or losses of mercury species.

In most situations empirical procedures are used for quality assurance. By comparing results obtained from several methods based on different sample preparation procedures and detection techniques it is often possible to evaluate the performance of the methods with respect to accuracy and precision. If several independent

-2-

methods give agreeable results the data have a larger probability to be accurate than if only one method has been used. But if results differ, which is a common problem for speciation analysis, a “true” value cannot be given. For speciation of trace elements, most laboratories have only one method available and they therefore have to resort to certified reference materials (CRMs). Often CRMs with an appropriate matrix composition and concentration of the species of interest are not available and in this situation recovery tests have to be performed for quality assurance. Analytical recovery is defined6 as the fraction or percentage of a species residue recoverable following extraction and analysis of a matrix containing the species. In practice the recovery is determined from analyte amounts added to the sample and this might not reflect the recovery of the incipient analyte of the sample.

The aim with this thesis was to develop and improve methods for the speciation of mercury in biological and environmental samples. The aim was also to study and improve the reliability of the methods by gaining more knowledge about processes that occur during the analysis and develop procedures to accurately correct for species transformation reactions.

1.3 Chromatographic separation of mercury species

At present a large number of methods utilising different pre-treatment procedures and separation/detection techniques for speciation of mercury in environmental and biological samples are described in the literature.7 Most methods for the determination of mercury species couple a technique for species separation with a suitable detection system. Separation by gas chromatography8,9 (GC), capillary electrophoresis 10,11 (CE), or high performance liquid chromatography12 (HPLC), are often applied. There are also so called operationally defined methods for the “fractionation” of mercury species, but they will not be discussed in this thesis.

1.3.1 HPLC

Difference in polarity of analytes is utilized for the separation of molecules in HPLC and it has become the most popular chromatographic separation technique.13 There are different types of columns available for the separation of analytes. Most columns

are packed with silica-based particles (3-10 µm in diameter) and to the silica surface a stationary phase is chemically attached. The most common stationary phase is octadecylsilane, ODS, also called C18. The ODS is non-polar and it is used in combination with a relatively polar mobile phase. For each analyte injected into the column there will be an equilibrium between the stationary phase and the mobile phase. A polar analyte has a short retention time as the equilibrium is shifted towards the polar mobile phase and for a non-polar analyte the equilibrium is shifted towards the stationary phase and consequently it has longer retention time. This type of HPLC

-3-

is called reversed phase. The opposite is normal phase HPLC where the stationary phase is polar and the mobile phase is non-polar. By changing the composition and polarity of the mobile phase the separation between analytes can be optimised. Ions can also be separated with HPLC by including counter ions, with an opposite charge

to the analyte ion, in the mobile phase. Derivatisation of species prior to separation is often obsolete in HPLC, consequently there are less steps involved that can lead to errors. Compared to GC sample preparation in HPLC is simplified, which is advantageous.

For mercury speciation with HPLC mercury species are usually mixed with some kind of sulphur containing reagent to form a complex, which improves the chromatographic separation properties of the species. For example mercaptoethanol,14-

16 mercaptobenzothiazole17 and pyrrolidine dithiocarbamate salts18,19,20,III are successfully used as complexing agents for speciation of mercury in HPLC. The most common approach is employing reversed phase chromatography with a mobile phase consisting of mainly acetonitrile/water or methanol/water. In order to lower the detection limits species can be on-line pre-concentrated on a “pre-column”19,20,III.

The type of material used in the HPLC system should be carefully chosen. Since mercury is easily adsorbed on steel, memory effects can arise from injectors, capillaries and columns made of steel. PEEK or glass should therefore be used instead of steel to minimize memory effects.21

1.3.2 CE

CE is one of the most powerful techniques for separating a large variety of analytes; from elemental ions up to large bio-molecules. The separation of analyte species in CE is based on species having different migration rates in an electrical field. Typical

capillaries used are made from fused-silica tubing, usually 25-100 µm in inner diameter and 50-100 cm long. A potential of 20-30 kV is applied over the capillary to obtain the migration and separation of analytes. When a voltage is applied over the capillary, cations migrate towards the cathode and as the ions are solvated in a solution they will “carry” the solution towards the negative end of the capillary. This solution flow is called the electro-osmotic flow, EOF, and it has a flat flow profile. If a

Figure 2. Flow profiles in a tube, flat (left) and laminar (right) flow.

Derivatisation is defined by IUPAC as “chemical reaction of analytes to yield derivatives that are

easier to chromatograph, easier to separate or easier to detect than the original analytes”

-4-

solution is forced through a capillary by pressure the flow profile is parabolic, called a laminar flow. A laminar flow causes much more broadening of a solvent zone compared to a flat flow, see figure 2. As there is ultimately no laminar flow in CE, broadening of the peaks within the column is minimised and as a result better resolution is obtained.

In CE highly charged cations have highest migration rate and arrive first at the cathode, neutral analytes are carried with the EOF and are not separated from each other and negatively charged ions reach the end of the capillary last. The anions are transported to the negative cathode because of the EOF. The flow rate of the EOF affects the chromatographic resolution and the ions will not have time to separate if the flow rate is too high. Very low flows result in long retention- and analysis times. To increase the EOF the electric field can be increased or the buffer composition in the capillary can be altered to have a higher pH, lower ionic strength or lower viscosity.

Samples can be injected electrokinetically by applying a voltage over the capillary when its injection end is immersed in the sample solution. With this injection method the analyte amount injected is dependent on its migration rate in the electrical field and for different samples and analytes this may lead to severe sampling bias. This is not the case when samples are instead injected hydrodynamically. The same sample volume will be injected by altering the pressure at the injection or elution end of the capillary. Hydrodynamic injections are preferred in quantitative analysis even if electrokinetc injections may be more efficient.

Mercury species do not need to be derivatised but should be mixed with a complexing agent, supplied with the electrolyte, to form charged complexes. Cysteine is the complexing agent mainly used in CE10,22-24 but also dithizone sulphonate25 and CaCl2

II has proven to be useful. A drawback when using CE is the limited amount of sample, only a few nl, utilized during analysis resulting in, compared to HPLC and GC, poor relative detection limits.26

1.3.3 GC

In GC species are separated on the basis of volatility and the interaction between the analyte and a stationary phase. There are both packed and open/capillary columns for GC, the latter being the most efficient and nowadays therefore the mainly used of the two. Capillary columns are typically between 10 and 100 meters long, having an

internal diameter of 0.2-0.7 mm and inner walls coated with a 0.2-5 µm thick film of stationary phase. The mobile phase is a gas, typically helium that is transporting the vaporized analytes towards the detector. The GC column is situated in an oven that is heated to required temperatures. The type of column (stationary phase), oven

-5-

temperature and mobile phase gas-flow rate are variables that can be changed to improve separation between analytes. The oven can be held at a constant temperature but better separation, more symmetric peak profiles and shorter analysis time are usually obtained if the oven is temperature programmed. The injector is set

on a high temperature so that the whole sample, 1-2 µl, is immediately volatilised when injected.

For pre-eminent performance in GC most mercury species need to be derivatised to become volatile and thermally stable. Butylation with Grignard reagent in an organic solvent is a common derivatisation scheme.27 As it requires extraction of the mercury species into an organic solvent before addition of the Grignard reagent, difference in the extraction efficiency between various species and various samples will lead to uncertainties in the determination. Within 5 minutes the reaction yield between butyl magnesium chloride and the mercury species has reached a maximum.28 (The actual derivatisation efficiency can hardly ever be determined as there are usually no salts or standards of the derivative available.) If the excess Grignard reagent is not deactivated, methyl butyl mercury can be transformed to dibutyl mercury and the same transformation has also been seen for dimethyl mercury.28 The presence of halogens such as iodine in a sample has also been shown to decompose alkyl mercury.29 Iodide and bromide can react with alkyl mercury to form mercury halides or alkyl mercury halides. Butylated mercury standards containing no halides were even shown to react with iodide and bromide in a GC that was contaminated with these halides.28

Sodium tetraethyl borate, NaBEt4, is used for derivatisation of mercury species in aqueous solution,30 and the produced volatile ethyl mercury species can be purged out of the solution with an inert gas onto a Tenax-packed column.31,32 The enriched mercury species are then desorbed from the Tenax by heating and with a gas stream the species are injected into the GC.32,33 The adsorption efficiency can however differ between the different Tenax-packed tubes resulting in large uncertainties in the analysis. Ethylated mercury species can also be extracted into an organic solvent, for example isooctane, nonane or hexane34 and injected traditionally with a syringe onto a GC-column. For samples containing high concentrations of mercury, when enrichment on Tenax is not necessary, this can be an easier approach than using Tenax-packed tubes. Varying purity of the NaBEt4 salts can be a problem as traces of methyl- replacing ethyl groups in the salt may cause overestimation of methyl mercury concentrations. Another drawback using NaBEt4 is that inorganic mercury, mono ethyl mercury and diethyl mercury cannot be separated as these species will be derivatised to the same chemical form. It has also been shown that if the sample

contains high concentrations of halides, (>50 µg ml-1) methyl mercury is reduced to

-6-

Hg0 during derivatisation.32 When tetrapropyl borate was used in the same way as NaBEt4, there was no reduction of mercury and also monoethyl-, diethyl- and inorganic mercury can be separated.32

1.4 Spectrometric detection

In this chapter the most common detection techniques used in mercury analysis are briefly described focusing on the types of detection used in paper I-V.

1.4.1 CV-AAS

Cold vapour-atomic absorption spectrometry, CV-AAS, is traditionally the most widely used technique for mercury determination. Mercury in a liquid sample is reduced, normally with SnCl2 or NaBH4, to Hg0. With an inert gas the mercury vapour is purged out of the solution and transported to an AAS where the absorption at 254 or 185 nm is measured. A typical absorption cell for mercury determination with CV-AAS is made of quartz, is 25 cm long and has an inner diameter of about 0.5 cm. An amalgamation step can be incorporated for obtaining lower relative detection limits. In an amalgamation step the gaseous mercury is transported to a quartz tube filled with for example platinum-gold wool. Mercury will be trapped as amalgam in the wool and it is in this way pre-concentrated. It is then released from the amalgam by heating the “wool” and Hg0 is transported to the detector. With an amalgamation trap, and by using the more sensitive absorption wavelength (185 nm) concentrations in the low ng l-1 range can be measured.

1.4.2 ICP-MS

Recently mass spectrometers (MS) have become more commonly used as detectors in analytical chemistry. In the mass spectrometer instrument the sample introduced is ionised and components are separated on a mass to charge ratio basis in the mass analyser. For elemental analysis samples are usually introduced as liquids into an inductively coupled plasma-MS (ICP-MS). The liquid passes through a nebulizer to form an aerosol that is introduced into the hot (3000-10000 K) plasma. The solvent (usually water) evaporates and molecules are ultimately decomposed to single atomic ions in the plasma. As the ICP is operated at atmospheric pressure and high temperature and the MS requires high vacuum and ambient temperatures it is necessary with an interface that reduces pressure and temperature. Aligned with the axes of the plasma the interface is divided into sections with two cones, sampler and skimmer, having an orifice diameter of about 1 mm. The pressure and temperature is reduced between the cones by means of a vacuum pump, see figure 3. Two turbo pumps are further reducing the pressure in the MS-part.

-7-

The ability to §distinguish between masses is dependent on the resolution of the

mass analyser. The resolution, R, is usually defined as R=m/∆m where m is the mean

value of two masses that are separated and ∆m is the mass difference between them. A criterion for considering two peaks to be separated is usually that the valley between the peaks is lower than 10% of the peak heights. The quadropole mass analysers are considered to offer low-resolution and they normally have unit resolution, meaning that they can separate only ions with different nominal masses. Considering the definition described above, these mass analysers have varying resolution depending on actual mass. High-resolution mass analysers, for example sector field analysers, have the same resolution in the whole mass range, for commercial ICP-sector field-MS up to approximately 10 000.

The main advantage with ICP-MS is that it can be used for detecting most elements at a low concentration; detection limits are generally in the ng l-1 range or lower. By scanning the MS, signals for many different elements, or isotope to charge ratios, can be detected within a few seconds. For elements with a mass above 80 u, there are generally only minor problems with spectral interferences even with a low resolution MS. Of the seven stable naturally occurring isotopes of Hg the isotope most likely to

TABLE 1. Abundance (%) of mercury and potential spectral interferences. Data from ELAN ICP-MS Instrument Control Software, version 2.3.2.

Nominal mass

Abundance of Hg

Potential spectral interferences, elemental and oxides, natural abundance of elements within brackets

196 0.15 Pt (25.24), TaO, HfO, WO 198 9.97 Pt (7.16), TaO, HfO, WO 199 16.87 WO, TaO 200 23.10 WO 201 13.18 WO 202 29.86 WO 204 6.87 Pb (1.40), WO

Figure 3. Schematic view of an ICP and the interface between ICP and MS.

Mass analyser Plasma

Sampler Skimmer

Torch

Injector tube

-8-

be affected by spectral interference is 204, see Table 1. Non-spectral interferences can however be a problem in ICP-MS when samples with complicated matrices are introduced.35 If non-spectral interference effects are present matrix-matched standards, the method of standard addition or internal standards can be used to obtain (more) accurate results.

The ICP-MS performance is readily affected by introduction of solutions containing high amounts of salts or organic solvents, as these will cause deposits of salt and carbon in the interface and MS. If the orifices of the cones are partly covered with salt or carbon the ion transport efficiency into the MS is decreased. Introduction of organic solvents give rise to a high organic vapour pressure that will cause a high carbon load of the plasma. High amounts of carbon and salt in the plasma absorb energy and therefore decrease the ionisation efficiency of the plasma leading to deterioration in sensitivity. Since carbon absorbs energy, a high organic vapour pressure can even give rise to difficulties in sustaining the plasma.35 The change in sensitivity observed for samples containing different amounts of salt is an example of a non-spectroscopic interference. It should however be mentioned that the sensitivity may also increase if a small amount of salt or organic solvent is added to an aqueous sample.

1.4.3 Other detectors used for the determination of mercury

Electron capture detector, ECD, is not a spectrometric detector but it should be mentioned as it was traditionally used in combination with GC for the speciation of mercury.36 ECD’s are however not selective towards mercury but instead selective towards molecules that contain electronegative functional groups, such as halides and the mercury species have therefore to be derivatised to form for example chloride derivatives prior detection.

Atomic fluorescence spectrometry is the most sensitive detection technique for elemental mercury. It is used in combination with CV and it can provide detection limits at the pg l-1 level.30

Absolute detection limits for mercury in the low pg range is achieved with microwave induced plasma-atomic emission spectrometry, MIP-AES. It has been successfully used as a detector for mercury species eluting from a GC.8,34,37 The disadvantage with MIP is the low energy of the plasma that prevents it from being able to atomise large amounts of sample, for example a solvent peak from a GC entering the MIP might extinguish the plasma.

UV detectors have been used in the determination of mercury after HPLC separation14,38. These detectors are however not selective, the sensitivity is poor and

-9-

therefore the concentrations of mercury in natural non-contaminated samples cannot be determined without pre-concentration of the samples.

MS has been directly coupled to GC39 and LC16 for mercury speciation analysis. By chemical or electrospray ionisation the molecules are ionised but they are not totally decomposed to atomic ions. The advantage with direct coupling to a MS is therefore the positive identification of a compound and the major drawback is poor methodological sensitivity.

1.5 Sample preparation

Solid samples cannot be directly injected into a chromatographic system necessitating the extraction of mercury species from the samples. The mercury species are either leached out from the solid sample or the whole sample is digested or dissolved in acidic or alkaline media. Various methods have been described in the literature40 of which some are outlined below. As samples from non-polluted areas contain low mercury levels the risk of contaminating samples during sampling, storage and pre-treatment is large. Because of this, extreme precaution is necessary to avoid contamination from sample containers and reagents. As each step in an analysis may contribute to errors it is generally advantageous to include as few steps as possible.

1.5.1 Alkaline dissolution

Most biological materials are easily dissolved in an aqueous41 or organic42 alkaline solution. The dissolution can be accelerated by heating43,44 or ultrasonic irradiation45. Potassium hydroxide42,43,45 and tetramethyl ammonium hydroxide, TMAH41, are the most commonly used alkaline reagents for dissolution of biological materials. A portion of the sample is simply mixed with a solution of the reagent, agitated on a mechanical shaker, placed in an ultrasonic bath or heated in a water bath. The whole sample is dissolved and the solution may be analysed directly or it might need further clean-up before derivatisation and injection into a chromatographic column.

1.5.2 Acid leaching

In 1966 Westöö published a pre-treatment method for methyl mercury determination in foodstuff based on acid leaching and enrichment of mercury in organic solvent.36 Hydrochloric acid was added for releasing mercury species from proteins followed by extraction into benzene or toluene. The leaching time should be optimised and controlled for the method to have similar extraction efficiencies for all samples analysed. The acid leaching approach has been further investigated and developed8,27,34,44,46 to be applicable for various other sample types.

-10-

1.5.3 Microwave digestion

Acidic microwave digestion has often been used for total mercury determination, as most samples can be totally decomposed in a closed vessel high-pressure system. Methyl mercury is however easily decomposed to inorganic mercury in the microwave field. This system has anyway, with careful optimisation, been used for mercury speciation analysis.47 A different type of microwave digestion system operating at atmospheric pressure and with individual microwave magnetrons for each vessel, is called “open focused microwave”. It has shown great potential of being a fast and suitable pre-treatment tool in speciation analysis,48 but it is still critical to find operating conditions that do not decompose the different mercury species. Evaluation of different extraction media in the microwave field showed, for methanolic KOH or aqueous TMAH, recoveries between 95 and 105% for both methyl mercury and inorganic mercury.48 But with HCl and HNO3 methyl mercury was dealkylated to inorganic mercury.48 (The open focused microwave system is to my knowledge no longer manufactured.)

1.5.4 Aqueous distillation

Distillation has been used for extracting mercury species, especially from sediment samples.43 The sample, KCl, H2SO4 and water are added to a distillation vessel and the

vessel is heated to about 150°C. The distillation is performed in a nitrogen atmosphere and the distillate containing the mercury species is collected in a vessel kept on ice.

1.5.5 Solvent extraction

For speciation of mercury in soil and sediment samples a solvent extraction method has been applied. Mercury species are extracted from a freeze dried sample into an acidic solution of KBr and Cu(NO3)

2. The extraction efficiency has been thoroughly investigated by Qian et al49 and methyl mercury formation has not been observed with this method.50

-11-

2.0 Problems and possibilities

2.1 Chromatography coupled to atomic spectrometric detection

The advantage with coupling together different analytical techniques is that the analytes can be separated with respect to more than one chemical and/or physical property. For example in the GC-ICP-MS case: analytes are first separated by means of their boiling point in the GC and in the ICP-MS analytes are separated and detected on behalf of their isotopic mass. Spectroscopic and non-spectroscopic interferences can be diminished or even avoided by a chromatographic separation. To couple chromatographic systems to atomic spectroscopic detection techniques are however not always straightforward. The form in which samples are eluted from a chromatographic system cannot always be directly injected into the detector; therefore the chromatographic methods often needs to be modified and/or special couplings and interfaces has to be implemented between the instruments.

2.1.1 HPLC-CV-AAS

For CV-AAS, the reduction and gas-liquid separation that can be obtained in an on-line flow-system makes it suitable as a detector for HPLC. The HPLC eluate is online mixed with a reducing agent such as NaBH4 to form volatile mercury species. Inert gas is added through a second T-connection and in a gas-liquid separator the volatile mercury species are purged out from the liquid following the gas stream into the AAS. In paper III it was shown that the most critical step is the formation of volatile mercury species and the gas-liquid separation. About 80% of the mercury is lost in these stagesIII and it is therefore a possibility to largely improve detection limits if the volatile mercury formation and transport efficiency from the gas-liquid separator could be increased. With a higher argon flow the volatile mercury is more efficiently transported to the AAS but the improved mercury transport efficiency is then counteracted by a decrease in sensitivity, caused by the higher gas flow through the absorption cell.

Drying gas

H2O (g)

H2O (g) Gas containing water vapour Dried gas

Outer tube

Nafion tube

Figure 4. The concept of drying gas with a Nafion dryer tube.

-12-

There is often a problem with baseline drift in the AAS because of water vapour that enters the absorption cell and absorbs light.51 By implementing some kind of water trap between the gas-liquid separator and the AAS this problem can be diminished. The use of Nafion dryer tubes is one of the most efficient and convenient ways for drying gas.51,III It also causes less band broadening compared to physical moisture traps with for example anhydrous salt.51 The gas to be dried is transported in a Nafion tube inserted in an outer tube. In the outer tube dry gas is flown in the opposite direction. Water is absorbed on the sulfonic acid surface of the Nafion and thereafter transported through the Teflon membrane of the Nafion, see figure 4. In paper III, when a Nafion dryer tube was first inserted in the system the methyl mercury peak almost completely disappeared. Methyl mercury was probably lost in the Nafion as

methyl mercury hydride. Therefore it is necessary to heat up (to about +800°C in the presence of MgO) the gas stream coming from the liquid-gas separator so that all volatile mercury species are reduced to Hg0 before being transported through the Nafion dryer tube.

The HPLC-CV-AAS approach is inexpensive to operate but it demands operator skills, as there are no fully automatic systems available. The large void volume caused by on-line reduction and coupling between the techniques gives rise to band broadening and it can lead to overlapping signals when the resolution between chromatographic peaks is small.

2.1.2 HPLC-ICP-MS

There should not be any practical problems to couple HPLC to ICP-MS as the ICP-MS is usually operated with a liquid flow rate between 0.1 and 1 ml min-1, which is in agreement with most HPLC applications. But the problem is that the HPLC-eluents generally contain considerable amounts (5-100%) of methanol and/or acetonitrile. As discussed in chapter 1.4.2 introduction of organic solvent and vapour into the ICP-MS will give rise to non-spectroscopic interference and carbon deposits in the interface and MS. Adding oxygen to the plasma and using a cooled spray-chamber that lowers the vapour pressure can diminish these problems. Carbon reacts with the added oxygen and produces carbon oxides. But the excess oxygen will also oxidise and destroy the sampler and skimmer cones (figure 3) that are usually made of nickel. For applications when oxygen is added to the plasma there are cones with platinum coating that should preferably be used, as they are less easily oxidised than nickel.

To omit cooled spray chambers and oxygen addition, HPLC methods avoiding or minimizing organic solvent could be applied when coupled to ICP-MS. Bloxham et al52 used an HPLC-eluent containing only 1% acetonitrile, which was reported to cause no interference in the ICP-MS. Wan et al53 determined mercury species in water by a HPLC-CV-ICP-MS method with an aqueous eluent containing 0.5% cysteine. The

-13-

HPLC-eluate was mixed with a 0.1% NaBH4 solution and in this way more volatile mercury species are formed. The CV generation was reported to increase mercury ion signals 8-36 times compared to signals when NaBH4 was not added to the eluent. The difference in signal enhancement was due to the variation in vapour generation efficiency for different mercury species, methyl mercury being the most efficient and inorganic mercury being the least efficient. A similar HPLC approach but without CV was applied in paper IV.

Development of the HPLC-ICP-MS method used in paper IV. As a starting point the AAS detector in the flow injection-HPLC-CV-AAS method described in paper III was replaced by ICP-MS. After cold vapour generation, elemental mercury and other volatile mercury species were introduced directly into the ICP-MS via a Teflon capillary, inserted into the injector tube (figure 3) of the instrument. The whole coupling device was made from Teflon and to accomplish optimal gas flow settings of the ICP-MS it was constructed with the ability to introduce an additional make-up gas flow. Organic vapours from the HPLC eluent and the NaBH4 reducing agent caused however salt and carbon deposits on the interface and high carbon load of the plasma. This could not be avoided completely by adding oxygen to the nebulizer gas, and by carrying the gas through a Nafion drying tube after the cold vapour generation. (Note that the gas has to be heated prior transport through a Nafion dryer tube, see chapter 2.1.1.) Although the Nafion dryer removed part of the methanol as was evident from reduced ArC+ signals, residual carbon was still substantial. Increasing the length of the Nafion drying tube would probably further reduce carbon introduction to the ICP-MS but it would also give rise to band broadening of the eluted peaks. Therefore, an alternative HPLC separation procedure based on an ammonium acetate-cysteine-water eluent and direct HPLC-ICP-MS coupling without enrichment or CV generation was developed. Instead of pyrrolidine dithiocarbamate, cysteine was chosen to form more polar mercury complexes, which could be

separated and eluted from the ODS-column (150x3mm, 5µm, α-chrom, Upchurch, USA) without the presence of organic solvent. A primary investigation for eluent optimisation was performed using the HPLC-CV-AAS system.III A two level full factorial chemometric design54 was applied for the eluent composition optimisation; the concentration of acetonitrile was varied between 0 and 1% (v/v), ammonium acetate between 0 and 60 mM and L-cysteine between 0 and 3 mM and the resolution

Figure 5. Ammonium pyrrolidinedithiocarbamate (left) and cysteine (right).

NS

SNH4

+ NH2 CH OH

O

SH

-14-

between methyl- and inorganic mercury was determined for the different eluents. Results were evaluated in the chemometric software Modde 4.0 (UMETRICS AB, Umeå, Sweden) and this showed that cysteine was the most important eluent constituent for separating methyl- and inorganic mercury. In agreement with Bloxham et al52 increased acetonitrile concentration was found to decrease resolution between the two mercury species. Increasing the concentrations of ammonium acetate did not significantly affect the resolution. From these results an eluent composition of 10.4 mM (0.08%, w/v) ammonium acetate and 1.65 mM (0.02%, w/v) L-cysteine was chosen. The total salt concentration was 0.1% and the reason for including ammonium acetate was to increase the buffer capacity of the eluent. The low salt concentration of the eluent does not give rise to any detectable non-spectroscopic interference effects for mercury. This was found by using a simple flow

injection system to inject water standards of methyl- and inorganic mercury with either Milli-Q water or the HPLC eluent as carrier solution.

Figure 6. Chromatogram for alkaline dissolved oyster tissue, the bottom trace shows the ArC+ signal.

0

500

1000

1500

2000

0 100 200 300 400

Time / Seconds

Cou

nts,

202 H

g+ /

s-1

0.0E+00

2.0E+04

4.0E+04

6.0E+04

Cou

nts,

53A

rC+ /

s-1

ArC+

CH3Hg+

Hg2+

With this HPLC-system there was no significant difference between sensitivities of inorganic mercury and methyl mercury when analysing aqueous standards. But for a digested oyster tissue sample spiked with the two species of mercury the sensitivity for inorganic mercury at a radio frequency power of 1100 W in the ICP was 50% lower than for methyl mercury. When the power was increased to 1500 W the inorganic mercury signal was about 90% of the methyl mercury signal. Therefore the method of standard addition or species-specific isotope dilution calibration should here be applied for more accurate determination of inorganic mercury in biological or environmental samples. When the ArC+ signal was monitored for the digested oyster sample it was seen that a carbon peak is co-eluting with mainly the inorganic mercury, see figure 6. As also can be seen from the figure, the baseline between the two peaks is slightly negative probably as a result of carbon load and this baseline

-15-

shift might deteriorate precision of peak area evaluations. The above zero baseline originates mainly from mercury ions, indicated by the isotope ratios, and was only reduced to 70% in the absence of eluent introduced.

Absolute detection limits for the HPLC-ICP-MS method were 13 and 16 pg for methyl- and inorganic mercury with a cross-flow nebulizer and a Scott spray chamber. With a micro-concentric nebulizer and a cyclone spray chamber the absolute detection limits were improved 3-13 times to 1 and 5 pg for methyl- and inorganic mercury, respectively. Detection limits were based on 3 times the standard deviation of integrated peak area signals for six blank injections. For further details on the experimental part, see paper IV.

HPLC-ICP-MS is a much more costly approach than HPLC-CV-AAS because of the high cost for running the ICP-MS, which requires about 17 L of Argon per minute. It can also be more difficult finding HPLC-conditions that are applicable with the ICP-MS. The advantage is that species-specific enriched stable isotope standards can be used, this is thoroughly discussed in chapter 2.3.

2.1.3 CE-ICP-MS

In CE nl volumes of samples are injected and the liquid flow rate is in the low µL min-

1 range. To successfully couple CE to ICP-MS a liquid make-up buffer flow is added to obtain a stable aerosol from the nebulizer. The make-up buffer flow is normally utilized for closing the electric circuit in the CE. In figure 7, the ground of the power

Figure 7. Schematic drawing of the type of coupling55 between CE and ICP-MS used in paper I and II.

Argon nebulizer gas

Cross flow nebulizer

ICP-MS CE-capillary

Make-up buffer

Run buffer

Peristaltic pump

30 kV power supply

-16-

supply is connected to a thin platinum wire that, in the T-connection, is in contact with the buffer solution.55 To inject a sample, the run buffer is replaced by a cup containing the sample and hydrodynamic or electrokinetic injection is applied, see chapter 1.3.2.

The main disadvantage with CE is high relative detection limits caused by the small sample volume, which can be injected onto the column. To be able to use this technique for trace element speciation analysis the detection limits should be improved. This has been achieved in paper I by the application of multiple parallel capillaries in CE-ICP-MS. With this approach the total amount of analyte introduced can be increased without using electrokinetic injection or sacrificing the separation efficiency by injecting a larger sample amount in one capillary. However for obtaining identical retention times, conditioning and adjusting of several capillaries is a critical and time-consuming task.I

The type of nebulizer used in the ICP-MS will also greatly effect the detection limit. Using a cross-flow nebulizer with a liquid make-up flow between 0.4 and 1 ml min-1 will cause an extensive dilution of the CE eluate and consequently worsen the detection limit. A micro-nebulizer having an optimal liquid flow rate below 0.1 ml min-1 would be a better choice for CE-ICP-MS couplings. An improved detection limit with a microconcentric- compared to a conventional cross-flow nebulizer was also seen in paper II where a method for mercury speciation with CE-ICP-MS is described.

When using a microconcentric nebulizer with a 96 µl min-1 liquid make-up flow the detection limit for methyl- and inorganic mercury was 14 and 6 ng g-1, respectively.

The cross-flow nebulizer was operated with a liquid make-up flow of 170 µl min-1 and the detection limits were, compared to the microconcentric nebulizer, about 10 times higher for methyl mercury and almost 20 times higher for inorganic mercury.II

The make-up buffer flow and nebulizer gas flow rate of the ICP-MS will both affect the chromatographic resolution in the CE. This is because certain settings might create a low pressure at the end of the capillary that will give rise to an unwanted laminar flow within the CE-capillary. This can be seen in paper I when comparing the CE chromatograms for a cross-flow nebulizer and a direct injection high efficiency nebulizer, DIHEN. The resolution is higher and the retention times longer when using the DIHEN, this is probably because of less laminar flow or a high pressure at the cathode causing a negative flow through the capillary. The optimisation of nebulizer gas flow and make-up liquid flow for CE-ICP-MS applications can be difficult as the chromatographic resolution is affected by these parameters. Both flows should be optimised by monitoring a steady state signal for mercury or another element not included in the make-up buffer. This element should be introduced from the run buffer, see figure 7. By monitoring the element signal both with and without

-17-

turning on the power it is possible to detect if there is a laminar flow within the capillary. If there is a laminar flow that deteriorates the chromatography the optimised flow rate of the nebulizer gas and/or make-up buffer might need to be changed or the ends of the analytical capillary could perhaps also be positioned at different heights to equalize the pressure in the whole capillary. A laminar flow can also be utilized to decrease retention times for a faster analysis.

2.1.4 GC-ICP-MS

When a GC is coupled to an ICP there is no need to have a nebulizer as the sample (GC eluate) is in gaseous form. The transfer line from the GC to the ICP-MS should preferably be evenly heated to prevent condensation of the eluate. It is also necessary to add a make-up gas, as the gas flow rate in the GC is low compared to optimal nebulizer gas flow settings in the ICP-MS. To avoid problems with carbon when injecting an organic solvent into the GC-ICP-MS, oxygen should be added to the

make-up or plasma gas. But as only about 1-2 µl of sample (in organic solvent) is usually injected into a GC, the carbon deposit problems are much less severe than those described above for HPLC-eluents. If the solvent peak has different retention time from the analytes of interest it is also possible to avoid carbon deposits by implementing a valve that can be automatically switched to vent the solvent peak.56

Various types of GC-ICP interfaces have been constructed in-house by many research groups57,58 and there are now commercially available interfaces.59

2.2 Transformation of species during analysis

In the end of the last decade it was found that inorganic mercury was methylated to form methyl mercury during aqueous distillation of sediment samples.60,61 As the inorganic mercury concentration in theses samples are normally more than 100 times higher than the natural methyl mercury concentration, the methyl mercury concentration will be significantly overestimated even if only a small fraction of the inorganic mercury (<0.1%) is methylated. In the same articles60 methylation of inorganic mercury in other type of samples and with other extraction methods was also observed. After these findings the certified methyl mercury concentrations in some sediment reference materials were questioned.62 With this new knowledge the sediment reference materials should be carefully re-evaluated regarding the methyl mercury content. The solvent extraction method (chapter 1.5.5) could be applied, as no transformations have been observed to occur with this method.50

In biological samples the concentration of methyl mercury is usually higher than that of inorganic mercury thereby methylation might not cause any significant overestimation of the methyl mercury concentration. Demethylation of methyl mercury during sample preparation might however cause overestimation of inorganic

-18-

mercury if the methyl mercury content is considerably higher than the inorganic mercury. Therefore the degree of both “dealkylation” and “alkylation” of mercury species should be examined for a sample preparation method, especially if, within a sample, there is a great difference in concentrations for various mercury species.

Standards of mercury should always be carefully prepared. Water standards with

concentrations >1µg g-1 are usually stable when kept in glass bottles. Inorganic solutions however should be acidified and organo-mercury standards should be stored in darkness. Snell et al63 have investigated the stability of mercury species in organic solutions. Stock standards containing both inorganic and alkyl mercury should be avoided or the stability should be carefully checked as the distribution of the species might easily be altered.

2.2.1 Sources of transformation

As mentioned earlier (chapter 1.3.3), transformation of species has been reported to occur in the GC and during derivatisation with magnesium butylchloride and sodium tetraethylborate. During sample preparation with microwave digestion (chapter 1.5.3) transformations have also been seen and investigated. It can however be difficult to find direct causes to the degree of transformation. Bloom et al60 did show that a larger carbon content in a water sample resulted in higher degree of methyl mercury produced during aqueous distillation of the sample. Also Prestbo et al64 found a relationship between degree of methylation and the amounts of sulphur and iron in combustion flue gas samples when acetic acid was used for pH-adjustment. When they replaced the acetic acid with citric acid the transformation disappeared.

In paper IV the degree of methylation in a pike sample was lower in a TMAH-ultrasonicated-digested sample that had been stored in the refrigerator for 24 hours compared to a sample that was pH-adjusted and analysed directly after digestion. This suggests that methylation was mainly taking place during pH-adjustment and not during the digestion. In this case however the acetic acid, which was used for pH-adjustment, was not the main methyl-donor as methylation occurred even when citric acid was used for pH-adjustment, see molecular structure in figure 8.

Figure 8. Acetic and citric acid.

O

OH

O

OH

OH O

OHOOH

-19-

Larsson et al65 described a method for the determination of gaseous Hg0, (CH3)2Hg and CH3Hg+ in air. During the analytical process 20-40% of the organomercury species were transformed, (methylated, demethylated and/or reduced,) and unless corrected for this would cause great errors in the determination of concentrations.

Decomposition of alkyl mercury is perhaps more easily accepted as a phenomenon. Depending on the stability of the covalent bond between mercury and carbon in an alkyl group the degree of dealkylation differs. The carbon-mercury bond in ethyl mercury is more stretched than that in methyl mercury66 and ethyl mercury has been shown to more easily dealkylate during sample preparation than methyl mercury.V,56 The covalent bonds may be destroyed in strong heat and in UV-light.

From the examples mentioned above it is evident that there is still no general simple explanation to the methylation of mercury. Therefore it is impossible to determine the degree of methylation in a sample from just knowing the main components of the sample. The degrees of transformations need to be determined within the sample and determined species concentrations should be compensated for any effects caused by transformation.

2.2.2 Determining the degree of transformation and correcting for errors caused by transformation

Bloom et al60 added known amounts of inorganic mercury to portions of the same sample. They showed that the determined methyl mercury concentration increased linearly with an increasing inorganic mercury concentration and could so determine the degree of methylation. Hintelmann et al61 used species with isotopically enriched mercury to determine the degree of methylation in a similar way as Bloom. The degree of methylation, TM, is calculated as:

Total

MM nT =

n(1)

In this equation nTotal represents the total amount of Hg2+ that has been added to the sample and nM is the amount of the added mercury that has been methylated. The amount of Hg2+ that has been methylated (nM) can be determined from comparing the determined amount of methyl mercury in the sample without additional Hg2+ with the sample where Hg2+ has been added. In paper IV a method is described that applies species-specific isotope dilution to calculate the amount of species transformed within each individual sample.

To correct for the transformations there are two different ways to go: (i) if you know or can justify an assumption in which order the transformations occur the concentrations should be corrected in several subsequent steps, (ii) the transformations can be assumed to occur simultaneously and then the

-20-

concentrations have to be corrected “simultaneously”. In the first case (i) the equations are easily solved but in case two (ii) the equations are most conveniently solved by matrix calculations especially if there are more than two species involved and all species transform into each other. Below is an example how to solve this when internal standards are used so that all losses, including those caused by transformation, are compensated for. In this example there are three species of the same element, A, B and C, and all species can be transformed into each other, see figure 9. For example A transforms into B and TAB is the degree of that transformation. In the example A is lost as B and C but the concentration of A could also be overestimated as B and C are transformed into A. Zi represents the incipient concentration of species Z in the sample and Zm,IS is the measured concentration that has been determined with internal standard and therefore all types of losses are compensated for using the equation system written below:

Figure 9. Transformation pathways of three species; A, B and C. TZW is the degree of Z that is transformed to W.

A

B

C TAC

TAB

B

A

CTBC

TBA

C

A

B TCB

TCA

iBCiACiISm CTBTAC +⋅+⋅=,

CBiiABiISm TCBTAB ⋅++⋅=,

CAiBAiiISm TCTBAA ⋅+⋅+=,(2a)

(2b)

(2c)

The corresponding matrix system (A) will then be:

XM1

11

Y,

,

,

⋅

=

i

i

i

BCAC

CBAB

CABA

ISm

ISm

ISm

CBA

TTTTTT

CBA

(A)

The unknown matrix X is solved by multiplying the inverse matrix of M with Y. If the apparent amounts are determined without internal standard the loss of a species has also to be corrected for. This is a bit more complicated and a matrix system (B) for determining the incipient concentrations is shown below for a case when only two species, A and B, are determined. In this case the measured concentration is Zm.

-21-

The values for Zi will be closer to the “real” values the higher k is set, but normally close enough values are obtained with k=2 or k=3.

It should be noted that it is not always necessary to use these corrections to obtain sufficiently accurate results, as transformations do not always give rise to significant determination errors.

2.3 Using enriched stable isotopes

As discussed before, different species of the same element have specific chemical and physical properties and they might show different extraction and derivatisation yields during sample preparation. To be able to use an internal standard of the same chemical form as the analyte species investigated is therefore a major advantage.

Enriched stable isotopes as internal standards for total determination of elements have been applied in MS for more than 40 years.67 In trace element speciation analysis it is a rather new approach but with increasing popularity. There are still no species-specific enriched stable isotope standards commercially available and therefore synthesis of the species has to be made in-house.58,68,69,V Besides using the species-specific standards with enriched stable isotopes as internal standards they can be added to samples to study transformation reactions during sample preparation and to study biotic reactions, for example methylation by sulphur reducing bacteria in a sediment.33 But whenever a chemical is added to a sample it may effect the natural equilibrium. To accurately study transformations and reactions taking place in a natural sample it is therefore important to disturb the equilibrium as little as possible, for example by adding as small amounts as possible.

The enriched stable isotope dilution calibration method is used in combination with MS and it is applicable for elements that have at least two natural stable isotopes. By

monitoring the signals for both the mass of an enriched isotope added to the sample and the mass of a reference isotope of the same element that is natural abundant, the difference of this isotope ratio from the natural isotope ratio is used to determine the amount of analyte present in the original sample. A simplified example of how stable isotope dilution calibration works is outlined below.

)(1)(

)()(1

11

1

1

1

1

⋅

⋅−⋅+⋅−⋅

⋅−⋅⋅−⋅+

=

∑∑

∑∑

−−

=

−

=

−i

knnnn

knnnn

k

n

nAB

nBA

nAB

nBA

k

n

nBA

nAB

nBA

nABm

B

A

TTTTTTTT

TTTTTTTT

B

A(B)

XM

Y 11

==

in

ABBAABBAn

BAABBAABm

-22-

It is actually the signal for a mass to charge ratio, m/z, that is monitored in the mass spectrometer.

An element has the natural isotopes x and y. Rx/y,mix is the isotope ratio for isotope x over isotope y in a sample to which a stable isotope standard, I, enriched with isotope x, has been added. The abundance of isotopes x and y in the isotope standard is AI,x and AI,Y respectively. Rx/y,natural, is the x/y ratio in a sample with only natural abundant analyte and it is known to be the natural fractions or abundances of x, AW,x, over y, AW,y: Rx/y,natural=AW,x/AW,y. If the natural unknown concentration of the element in the sample is W molar and the amount (molar) of enriched stable isotope standard added is I, the following equation 3a is true, giving the unknown W in equation 3b after rearrangements:

IAWAIAWA

RyIyW

xIxWmixyx ⋅+⋅

⋅+⋅=

,,

,,,/

yImixyxxI

AARIARIA

W ,,/,

−⋅

⋅⋅−⋅=

xWyWmixyx ,,,/

(3a)

(3b)

The experimental and certified values for abundance are usually not the same because of mass bias in the instrument. This should therefore be compensated for in the calculation, either by a correction

factor or by using experimental abundance values.70

It is possible to see differences in the isotope ratio even if the addition of the enriched isotope is small. The additions are usually smaller than, or in the same concentration range as that of the analyte naturally present in the sample. The paradox is that it would be best to know the concentration of a species in a sample before analysing the sample, as an optimal amount of enriched isotope can then be added. Because, if the addition is too small the difference from the natural ratio is small and the uncertainty in the ratio will be larger, if the addition is too big the ratio will be large which also leads to uncertainty in the ratio determination.71

The importance of having a species-specific internal standard is particularly observed in paper V where the loss of ethyl mercury, caused by dealkylation, is large but there are no detectable losses of methyl mercury and inorganic mercury. When more than 2 species are determined65 and there are transformation of the species it is advantageous to use species-specific isotope dilution calibration for easier corrections (chapter 2.2.2). If the transformation differs a lot between samples it is time-consuming to determine the degree of transformation in each sample without the use of species-specific enriched stable isotope standards.

2.3.1 Calculation of corrected signals

There are no enriched isotopes that do not have any impurities from other isotopes of the same element. When more than one enriched isotope standard is added to a sample these impurities have to be compensated for in order to obtain more accurate results.IV For example one standard, s1, of species A, is enriched with the stable isotope y and it has 5% impurities of isotope x. If 10% of the impurity is transformed into B, there will be a contribution to the signal of isotope x for species B and this

-23-

may result in an overestimation of the B-concentration. Below is an example where 2 standards, s1 enriched with isotope y and s2 enriched with isotope z, has been added to a sample. The signals for the three isotopes x, y and z for one species can be corrected for species and isotope impurities in s1 and s2. The signal of each isotope is a sum of the natural abundant (n) species signal and signal from s1 and s2.

Accordingly ΣSi is the total measured signal of isotope i, Si,q is the contribution from solution q to the isotope signal of i giving the following equations 4a-4c:

2,1,, sxsxnxx SSSS ++=∑

2,1,, sysynyy SSSS ++=∑2,1,, szsznzz SSSS ++=∑

(4a)

(4b)

(4c)

To be able to solve the three equations above, the equation system should be modified to include only three unknown values. Therefore the Si,q values are rewritten as a function of isotope ratios in each solution. One example is given in equation 5 below where Rx/y,s1 is the isotope ratio of x over y in standard s1.

(5) 1,1,/1, sysyxsx SRS ⋅=

Similar rewritings for the other signals in equation 4a-4c will give the equations 6a-6c with only three unknown values (Sx,n, Sy,s1 and Sz,s2) as shown below.IV

2,2, szsz S⋅

2,2,/1,1,/,,/ szszzsysyznxnxzz S⋅

2,2,/1,1,/,,/ szszxsysyxnxnxxx SRSRSRS ⋅+⋅+⋅=∑/1,1,/,,/ ysysyynxnxyy RSRSRS +⋅+⋅=∑

RSRSRS +⋅+⋅=∑

(6a)

(6b)

(6c)

By using the same type of matrix calculations as for matrix A, chapter 2.2.2, signals corrected for impurities are obtained. These signals are then used in an isotope dilution equation derived in paper IV. See example below:

= ⋅ ⋅

nM s sM 1 C⋅ 1

n , xA s , 1 ⋅ s 1, yA

yS n , xS (7)

nC

Cn and Cs1 are the concentrations (mol g-1) of a species in the sample and in the standard s1. Ms1 and Mn are the respectively amounts (g) of s1 and sample in the mixture. Ay,s1 and Ax,n is the abundance of enriched isotope y in s1 and abundance of isotope x in the natural sample, respectively. Example of how the corrected signals may be used for determining the degree of transformation is found in paper IV.

-24-

3.0 Concluding remarks and future perspectives

Species of mercury have been shown to transform into various forms during pre-treatment. The degree of transformation is different for different species and samples and it may give rise to significant determination errors unless corrected for. Therefore internal standards should be used, preferably species-specific enriched stable isotope standards. With these standards, in combination with a mass spectrometric detector, it is possible to both determine the concentration of mercury species and the degree of transformation within each individual sample. Hence the errors caused by transformation can be compensated for and a more accurate analysis is obtained.

Incomplete extraction and losses during sample pre-treatment will also be compensated for after addition of the species-specific internal standards to the sample. In CE-ICP-MS these internal standards would allow for the use of the more efficient electrokinetic injection.

There is however a need for commercially available species-specific enriched stable isotope standards before they can be used in routine analysis. There is also a need for more research concerning the uncertainties in the isotope ratio measurements for transient signals. The uncertainty in isotope ratio measurements has so far only been thoroughly studied for steady state signals.72

Among the mercury speciation methods used in paper II-V the HPLC-CV-AAS instrumentation is the least expensive both to purchase and use. If transformation of species is known to be negligible the HPLC-CV-AAS could be automatized and used in routine analysis. HPLC is easily coupled to ICP-MS but it is more difficult to optimise the HPLC performance when the eluent has to be compatible with the ICP-MS. The coupling between CE and ICP-MS is also easily accomplished but the optimisation of species separation as well as the high relative detection limits and poor robustness can be problematic. The GC-ICP-MS methods have shown superior chromatographic resolution compared to HPLC- and CE- ICP-MS, but drawbacks with GC is that the mercury species need to be derivatised and the coupling to the ICP-MS is difficult to construct or expensive to purchase. The GC is however robust, easy to optimise and operate and therefore the disadvantages are often tolerated. With more available species-specific enriched stable isotope standards GC-ICP-MS methods will be the first choice in mercury “multi-speciation” analysis.

For transformation corrections the species-specific enriched stable isotope approach is however limited such that, to include n species of the same element in the correction calculation, n+1 stable isotopes are required. Among the elements that are of primary interest in speciation analysis tin and selenium have 9 and 5 natural stable isotopes, respectively, whereas arsenic has only one stable isotope.

-25-

-26-

References

1. F. Bakir, S. F. Damluji, L. Amin-Zaki, A. Khalidi, N. Y. al-Rawi, S. Tikriti, H. I. Dahahir, T. W. Clarkson, J. C. Smith, and R. A. Doherty, Methylmercury poisoning in Iraq. Science 181, 230-241 (1973).

2. T. W. Clarkson, The three modern faces of mercury. Environmental Health Perspectives 110, 11-23 (2002).

3. H. Frankish, Report finds no link between thimerosal and neurodevelopmental disorders. Lancet 358, 1163 (2001).

4. Hygiensiska gränsvärden och åtgärder mot luftföroreningar. AFS 2000:3. (2000).

5. Om metaller, Statens Naturvårdsverk, (1976).

6. P. T. Holland, IUPAC Pesticides report .36. Glossary of terms relating to pesticides (IUPAC Recommendations 1996). Pure and Applied Chemistry 68, 1167-1193 (1996).

7. M. Morita, J. Yoshinaga, and J. S. Edmonds, The determination of mercury species in environmental and biological samples. Pure and Applied Chemistry 70, 1585-1615 (1998).

8. E. Bulska, D. C. Baxter, and W. Frech, Capillary column gas-chromatography for mercury speciation. Analytica Chimica Acta 249, 545-554 (1991).

9. L. Liang, N. S. Bloom, and M. Horvat, Simultaneous determination of mercury speciation in biological- materials by GC/CVAFS after ethylation and room-temperature precollection. Clinical Chemistry 40, 602-607 (1994).

10. I. Medina, E. Rubi, M. C. Mejito, and R. Cela, Speciation of organomercurials in marine samples using capillary electrophoresis. Talanta 40, 1631-1636 (1993).

11. A. M. CarroDiaz, R. A. LorenzoFerreira, and R. CelaTorrijos, Capillary electrophoresis of methylmercury with injection by sample stacking. Journal of Chromatography A 730, 345-351 (1996).

12. C. F. Harrington, The speciation of mercury and organomercury compounds by using high-performance liquid chromatography. Trac-Trends in Analytical Chemistry 19, 167-179 (2000).

13. D. A. Skoog and J. J. Leary. Principles of Instrumental Analysis, 4th edition. 26 High Performance Liquid Chromatography. (1992).

14. M. Hempel, H. Hintelmann, and R. D. Wilken, Determination of organic mercury species in soils by high- performance liquid-chromatography with ultraviolet detection. Analyst 117, 669-672 (1992).

15. H. Hintelmann and R. D. Wilken, The analysis of organic mercury-compounds using liquid-chromatography with online atomic fluorescence spectrometric detection. Applied Organometallic Chemistry 7, 173-180 (1993).

16. C. F. Harrington, J. Romeril, and T. Catterick, The speciation of mercury and organomercury compounds by high performance liquid chromatography atmospheric pressure ionization mass spectrometry. Rapid Communications in Mass Spectrometry 12, 911-916 (1998).

-27-

17. Y. C. Wang and C. W. Whang, High-performance liquid-chromatography of inorganic mercury and organomercury with 2-mercaptobenzthiazole. Journal of Chromatography 628, 133-137 (1993).

18. C. Sarzanini, G. Sacchero, M. Aceto, O. Abollino, and E. Mentasti, Simultanous determination of methyl-mercury, ethyl-mercury, phenyl-mercury and inorganic mercury by cold vapor atmic-absorption spectrometry with online chromatographic-separation. Journal of Chromatography 626, 151-157 (1992).

19. R. Eiden, R. Falter, B. AugustinCastro, and H. F. Scholer, Distillation, on-line RP C18 preconcentration and HPLC-UV-PCO- CVAAS as a new combination for the determination of methylmercury in sediments and fish tissue. Fresenius Journal of Analytical Chemistry 357, 439-441 (1997).

20. X. F. Yin, W. Frech, E. Hoffmann, C. Ludke, and J. Skole, Mercury speciation by coupling cold vapour atomic absorption spectrometry with flow injection on-line preconcentration and liquid chromatographic separation. Fresenius Journal of Analytical Chemistry 361, 761-766 (1998).

21. C. F. Harrington and T. Catterick, Problems encountered during the development of a method for the speciation of mercury and methylmercury by high-performance liquid chromatography coupled to inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 12, 1053-1056 (1997).

22. A. M. CarroDiaz, R. A. LorenzoFerreira, and R. CelaTorrijos, Capillary electrophoresis of methylmercury with injection by sample stacking. Journal of Chromatography A 730, 345-351 (1996).

23. T. H. Lee and S. J. Jiang, Determination of mercury compounds by capillary electrophoresis inductively coupled plasma mass spectrometry with microconcentric nebulization. Analytica Chimica Acta 413, 197-205 (2000).

24. M. S. da Rocha, A. B. Soldado, E. Blanco-Gonzalez, and A. Sanz-Medel, Speciation of mercury compounds by capillary electrophoresis coupled on-line with quadrupole and double-focusing inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 15, 513-518 (2000).

25. S. Hardy and P. Jones, Capillary electrophoresis determination of methylmercury in fish and crab meat after extraction as the dithizone sulphonate complex. Journal of Chromatography A 791, 333-338 (1997).

26. B. G. Sun, M. Macka, and P. R. Haddad, Speciation of tin, lead, mercury, arsenic and selenium compounds by capillary electrophoresis. International Journal of Environmental Analytical Chemistry 81, 161-205 (2001).

27. H. Emteborg, N. Hadgu, and D. C. Baxter, Quality-control of a recently developed analytical method for the simultaneous determination of methylmercury and inorganic mercury in environmental and biological samples. Journal of Analytical Atomic Spectrometry 9, 297-302 (1994).

28. H. Emteborg, J. Snell, J. Qian, and W. Frech, Sources of systematic errors in mercury speciation using Grignard reagents and capillary gas chromatography coupled to atomic spectrometry. Chemosphere 39, 1137-1152 (1999).

29. Y. Talmi and R. E. Mesmer. Studies on vaporization and halogen decomposition of methylmercury compounds by GC [gas chromatography] with a microwave detector. Water research 9[5-6], 547-552. (1975).

-28-

30. N. Bloom, Determination of picigram levels of methylmercury by aquous phase ethylation, followed by cryogenic gas-chromatography with cold vapor atomic fluorescence detection. Canadian Journal of Fisheries and Aquatic Sciences 46, 1131-1140 (1989).

31. L. Liang, M. Horvat, and N. S. Bloom, An improved speciation method for mercury by GC CVAFS after aquous-phase ethylation and room-temperature precollection. Talanta 41, 371-379 (1994).

32. N. Demuth and K. G. Heumann, Validation of methylmercury determinations in aquatic systems by alkyl derivatization methods for GC analysis using ICP-IDMS. Analytical Chemistry 73, 4020-4027 (2001).

33. Lambertsson L., Lundberg E., Nlsson M.B., and W. Frech. Application of enriched stable isotope tracers in combination with istope dilution GC-ICP-MS to study mercury species transformation in sea sediments during in situ ethylation and determination. Journal of Analytical Atomic Spectrometry 16, 1296-1301 (2001).

34. Q. Tu, J. Qian, and W. Frech, Rapid determination of methylmercury in biological materials by GC-MIP-AES or GC-ICP-MS following simultaneous ultrasonic- assisted in situ ethylation and solvent extraction. Journal of Analytical Atomic Spectrometry 15, 1583-1588 (2000).

35. A. Montaser, M. G. Minnich, H. Liu, A. G. T. Gustavsson, and R. F. Browner, Fundamental aspects of sample introduction in ICP spectrometry. In Inductively Coupled Plasma Mass Spectrometry, 1 Edn, ed. by A. Montaser, Wiley-VCH, New York (1998).

36. Westoo G, Determination of methylmercury compounds in foodstuffs. I. Methylmercury compounds in fish, identification and determination. Acta Chemica Scandinavia 20, 2131-2137 (1966).

37. B. D. Quimby and J. J. Sullivan, Evaluation of a microwave cavity, discharge-tube, and gas-flow system for combined gas-chromatography atomic emission detection. Analytical Chemistry 62, 1027-1034 (1990).

38. M. Foltin, S. Megova, T. Prochackova, and M. Steklac, Speciation of mercury by ion chromatography with post-column derivatization. Journal of Radioanalytical and Nuclear Chemistry-Articles 208, 295-307 (1996).

39. C. M. Barshick, S. A. Barshick, P. F. Britt, D. A. Lake, M. A. Vance, and E. B. Walsh, Development of a technique for the analysis of inorganic mercury salts in soils by gas chromatography mass spectrometry. International Journal of Mass Spectrometry 178, 31-41 (1998).

40. J. E. S. Uria and A. Sanz-Medel, Inorganic and methylmercury speciation in environmental samples. Talanta 47, 509-524 (1998).

41. M. S. Jimenez and R. E. Sturgeon, Speciation of methyl- and inorganic mercury in biological tissues using ethylation and gas chromatography with furnace atomization plasma emission spectrometric detection. Journal of Analytical Atomic Spectrometry 12, 597-601 (1997).

42. L. Liang, M. Horvat, E. Cernichiari, B. Gelein, and S. Balogh, Simple solvent extraction technique for elimination of matrix interferences in the determination of methylmercury in environmental and biological samples by ethylation gas chromatography cold vapor atomic fluorescence spectrometry. Talanta 43, 1883-1888 (1996).

43. M. Horvat, N. S. Bloom, and L. Liang, Comparison of distillation with other current isolation methods for the determination of methyl mercury-compunds in low-level environmental-samples .1. Sediments. Analytica Chimica Acta 281, 135-152 (1993).

-29-

44. M. Logar, M. Horvat, I. Falnoga, and V. Stibilj, A methodological study of mercury speciation using Dogfish liver CRM (DOLT-2). Fresenius Journal of Analytical Chemistry 366, 453-460 (2000).

45. A. M. Caricchia, G. Minervini, P. Soldati, S. Chiavarini, C. Ubaldi, and R. Morabito, GC-ECD determination of methylmercury in sediment samples using a SPB-608 capillary column after alkaline digestion. Microchemical Journal 55, 44-55 (1997).

46. C. R. Hammerschmidt and W. F. Fitzgerald, Formation of artifact methylmercury during extraction from a sediment reference material. Analytical Chemistry 73, 5930-5936 (2001).

47. M. J. Vazquez, A. M. Carro, R. A. Lorenzo, and R. Cela, Optimization of methylmercury microwave-assisted extraction from aquatic sediments. Analytical Chemistry 69, 221-225 (1997).

48. C. M. Tseng, A. DeDiego, F. M. Martin, D. Amouroux, and O. F. X. Donard, Rapid determination of inorganic mercury and methylmercury in biological reference materials by hydride generation, cryofocusing, atomic absorption spectrometry after open focused microwave-assisted alkaline digestion. Journal of Analytical Atomic Spectrometry 12, 743-750 (1997).

49. J. Qian, U. Skyllberg, Q. Tu, W. F. Bleam, and W. Frech, Efficiency of solvent extraction methods for the determination of methyl mercury in forest soils. Fresenius Journal of Analytical Chemistry 367, 467-473 (2000).

50. H. Hintelmann, Comparison of different extraction techniques used for methylmercury analysis with respect to accidental formation of methylmercury during sample preparation. Chemosphere 39, 1093-1105 (1999).

51. W. T. Corns, L. Ebdon, S. J. Hill, and P. B. Stockwell, Effects of moisture on the cold vapor determination of mercury and its removal by use of membrane dryer tubes. Analyst 117, 717-719 (1992).

52. M. J. Bloxham, A. Gachanja, S. J. Hill, and P. J. Worsfold, Determination of mercury species in sea-water by liquid chromatography with inductively coupled plasma mass spectrometric detection. Journal of Analytical Atomic Spectrometry 11, 145-148 (1996).

53. C. C. Wan, C. S. Chen, and S. J. Jiang, Determination of mercury compounds in water samples by liquid chromatography inductively coupled plasma mass spectrometry with an in situ nebulizer/vapor generator. Journal of Analytical Atomic Spectrometry 12, 683-687 (1997).

54. S. Wold, M. Sjöström, R. Carlsson, T. Lundstedt, S. Hellberg, B. Skagerberg, C. Wikström, and J. Öhman, Multivariate design. Analytica Chimica Acta 191, 17-32 (1986).

55. V. Majidi and N. J. Miller-Ihli, Two simple interface designs for capillary electrophoresis inductively coupled plasma mass spectrometry. Analyst 123, 803-808 (1998).

56. E. Bulska, H. Emteborg, D. C. Baxter, W. Frech, D. Ellingsen, and Y. Thomassen, Speciation of mercury in human whole-blood by capillary gas-chromatography with a microwave-induced plasma emission detector system following complexometric extraction and butylation. Analyst 117, 657-663 (1992).

57. A. Prange and E. Jantzen, Determination of organometallic species by gas-chromatography inductively-coupled plasma-mass spectrometry. Journal of Analytical Atomic Spectrometry 10, 105-109 (1995).

-30-

58. J. P. Snell, I. I. Stewart, R. E. Sturgeon, and W. Frech, Species specific isotope dilution calibration for determination of mercury species by gas chromatography coupled to inductively coupled plasma- or furnace atomisation plasma ionisation-mass spectrometry. Journal of Analytical Atomic Spectrometry 15, 1540-1545 (2000).

59. GC-ICP-MS Interface. 5988-3071EN. 2001. Agilent Technical Note.

60. N. S. Bloom, J. A. Colman, and L. Barber, Artifact formation of methyl mercury during aqueous distillation and alternative techniques for the extraction of methyl mercury from environmental samples. Fresenius Journal of Analytical Chemistry 358 , 371-377 (1997).

61. H. Hintelmann, R. Falter, G. Ilgen, and R. D. Evans, Determination of artifactual formation of monomethylmercury (CH3Hg+) in environmental samples using stable Hg2+ isotopes with ICP-MS detection: Calculation of contents applying species specific isotope addition. Fresenius Journal of Analytical Chemistry 358, 363-370 (1997).

62. N. S. Bloom, D. Evans, H. Hintelmann, and R. D. Wilken, Methyl mercury revisited. Analytical Chemistry 71, 575A (1999).

63. J. Snell, J. Qian, M. Johansson, K. Smit, and W. Frech, Stability and reactions of mercury species in organic solution. Analyst 123, 905-909 (1998).

64. E. M. Prestbo and N. S. Bloom, Mercury speciation adsorption (MESA) method for combustion flue-gas-methodology, artifacts, intercomparison, and atmispheric implications. Water Air and Soil Pollution 80, 145-158 (1995).

65. T. Larsson and W. Frech. Species specific isotope dilution with permeation tubes for accurate determination of gaseous mercury species. Submitted to Analytical Chemistry, 2003.

66. M. Kaupp and O. L. Malkina, Density functional analysis of C-13 and H-1 chemical shifts and bonding in mercurimethanes and organomercury hydrides: The role of scalar relativistic, spin-orbit, and substituent effects. Journal of Chemical Physics 108, 3648-3659 (1998).

67. A. A. Smales and R. K. Webster, The determination of rubidium by the method of stable isotope dilution. Analytica Chimica Acta 18, 582-586 (1958).

68. R. C. R. Martin-Doimeadios, T. Stoichev, E. Krupp, D. Amouroux, M. Holeman, and O. F. X. Donard, Micro-scale preparation and characterization of isotopically enriched monomethylmercury. Applied Organometallic Chemistry 16, 610-615 (2002).

69. C. J. Cappon and J. C. Smith, Gas-chromatographic determination of inorganic mercury and organomercurials in biological materials. Analytical Chemistry 49, 365-369 (1977).

70. H. P. Longerich, The application of isotope-dilution to imductively coupled plasma-mass spectrometry. Atomic Spectroscopy 10, 112-115 (1989).

71. F. Vanhaecke, G. de Wannemacker, L. Moens, R. Dams, C. Latkoczy, T. Prohaska, and G. Stingeder, Dependence of detector dead time on analyte mass number in inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 13, 567-571 (1998).

72. P. K. Appelblad, I. Rodushkin, and D. C. Baxter, Sources of uncertainty in isotope ratio measurements by inductively coupled plasma mass spectrometry. Analytical Chemistry 73, 2911-2919 (2001).

-31-

-32-

Glossary

Mercury speciation Species-specific analysis of mercury

Alkylation Formation of a covalent bond between a (mercury) atom and a carbon in an alkyl group like for example methyl- or ethyl-.

Dealkylation Breakage of a covalent bond between an alkyl group and an element atom

IUPAC International Union of Pure and Applied Chemistry

HPLC High Performance Liquid Chromatography

GC (capillary) Gas Chromatography

CE Capillary Electrophoresis

CV-AAS Cold Vapour Atomic Absorption Spectrometry

ICP-MS Inductively Coupled Plasma Mass Spectrometry

TMAH Tetra methyl ammonium hydroxide

NaBEt4 Sodium tetraethyl borate

-33-

-34-

Acknowledgements

Först och främst vill jag tacka Erik och Lasse som en sen kväll i diskrummet (efter någon kursfest) övertygande mig om att jag skulle söka doktorandtjänst på analyten...

Jag vill tacka min handledare Wolfgang som lät mig börja som doktorand och som lärt mig otroligt mycket om kemi, kritiskt skrivande och livet. Du har varit ett bra stöd i både med- och motgång!

Ett stort tack till alla i AAS-gruppen som bidragit till många trevliga stunder med mat, vin, klota och kemi-diskussioner(!): Erik för ditt ovärdeliga stöd och din hjälp, att du varit ett toppen-bra “bollplank”, trevlig rumskompis och att du stod ut med mitt prat och mina tvära kast, Karin för all hjälp och alla trevliga pratstunder, Tom för alla idéer och roliga historier..., Lars för gott samarbete, James för hjälp med allt möjligt från salsa-recept till krånglande datorer och en massa kemi funderingar, Jin, Qiang, Daniel, Dong och så navet i det hela; Wolfgang för att du tagit hand om oss alla!

Tack till alla andra förre detta och nuvarande doktorander och personal på analytisk kemi som alla på sitt sätt bidragit till den här boken: Anders, Anita, Anna N, Anna S, Ann-Helene, Britta, Camilla, Einar, Erika, Josefina, Juliene, Knut (som genom att släppa iväg mig till Japan bidrog till att jag ville fortsätta med analytisk kemi, tack!), Lasse, Magnus, Michael, Patrik, Solomon, Svante J, Svante Å, Tobias, Ulla och Ulrika.

Sen finns det ju annat än jobb som gör livet roligt och som också bidrar till att man trivs på jobbet, tex: att dricka god skotsk (helst) malt-whisky tillsammans med medlemmarna i Uisge Beatha (se www.acc.umu.se/~gramse/Whisky1.html) tack för alla super-trevliga stunder att lära andra hur skoj det är att åka slalom, tack hela gänget i Umeå Nya Skidskola att spela innebandy, tack fysikal-gänget och de få tappra från analyten att spela golf (okej då, jag erkänner att jag inte alltid är sprudlande glad på golfbanan...) tack alla golfvänner ni gör golfrundan roligare

Jag vill även passa på att tacka gänget på biokemi som bidragit till många trevliga stunder: Magnus, Linda, Maria (och Anders), Anna, Peter (och Ulrika) Mikael O (och Pia)!

Mamma, Pappa, Malin, Fredrik och Fergus ska också ha tack!

Till sist vill jag tacka den person som betyder mest för mig, Mikael, du är bäst! (…fast det förstås, slalom, du är kanske inte så bra på att åka slalom) ...du vet!!

-35-