Using combustion and direct gas introduction for the isotopic analysis of mercury

Material Information

Using combustion and direct gas introduction for the isotopic analysis of mercury
Rutherford, Danny Lee ( author )
Place of Publication:
Denver, CO
University of Colorado Denver
Publication Date:
Physical Description:
1 electronic file (92 pages). : ;


Subjects / Keywords:
Mercury -- Analysis ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Alternative techniques for the preparation and introduction of samples for mercury (Hg) isotopic analysis by multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) were investigated. An experimental method was developed in an attempt to improve the precision of Hg isotope measurements in high organic matrices, streamline these techniques, and utilize instrumentation available onsite. A temporary storage and syringe transfer method was developed and is characterized by five major parts: (1) sample Hg release by combustion, (2) collection and release of Hg on a gold trap, (3) collection of Hg in a temporary storage vessel, (4) sampling and transport of Hg by gas syringe, and (5) direct gas introduction to the plasma. Three standard reference materials (NIST-SRM 2711, SARM-20, and TORT-1) were prepared and analyzed using this experimental method and compared to the traditional acid digestion approaches with cold vapor introduction commonly used for Hg isotope measurements. This new experimental method improved upon the traditional approaches by decreasing preparation times and the use of hazardous chemicals (hydrochloric acid, nitric acid, and stannous chloride), but was unable to improve the precision. The new experimental method standard deviations for 8²⁰²Hg values resulted in approximately double the standard deviation values for bench top acid digestion preparation. A number of future experiments are suggested for further refinement and assessment of the experimental method.
Thesis (M.S.)--University of Colorado Denver. Department of Chemistry
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Danny Lee Rutherford, Jr.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
898035295 ( OCLC )


This item is only available as the following downloads:

Full Text


USING COMBUSTION AND DIRECT GAS INTRODUCTION FOR THE ISOTOPIC ANALYSIS OF MERCURY by DANNY LEE RUTHERFORD JR B.S., Colorado School of Mines, 2006 B.S., University of Colorado, Denver, 2008 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Chemistry Program 2014


ii This thesis for the Master of Science degree by Danny Lee Rutherford Jr has been approved for the Chemistry Program by Larry G. Anderson, Chair John Lanning Yong Liu April 18, 2014


iii Rutherford Jr, Danny Lee (M.S., Chemistry) Using combustion and direct gas introduction for the isotopic analysis of mercury Thesis directed by Professor Emeritus Larry G. Anderson. ABSTRACT Alternative techniques for the preparation and introduction of samples for mercury (Hg) isotopic analysis by multi collector inductively coupled plasma mass spectrometry (MC ICP MS) were investigated An experimental method was developed in an attempt to improve the precision of Hg isotope measurements in high organic matrices, streamline these techniques and utiliz e instrumentation available onsite. A temporary storage and syringe transfer method was developed and is characterized by five major parts: (1) sample Hg release by combustion, (2) collection and release of Hg on a gold trap, (3) collection of Hg in a tem porary storage vessel, (4) sampling and transport of Hg gas by syringe, and (5) direct gas introduction to the plasma. Three standard reference materials ( NIST SRM 2711, SARM 20, and TORT 1) were prepared and analyzed using this experimental method and co mpared to the traditional acid digestion approaches with cold vapor introduction commonly used for Hg isotope measurements Th is new experimental method improved upon the traditional approaches by decreasing preparation times and the use of ha zardous chem icals (hydrochloric acid, nitric acid, and stannous chloride) but was unable to improve the precision. The new e xperimental method standard deviations for 202 Hg values resulted in approximately double the standard deviation values for bench top acid digestion preparation. A number of future experiments are suggested for further refinement and assessment of the experimental method. The form and content of this abstract are approved. I recommend its publication. Approved: Larry G. Anderson


iv ACKNOWLEDGMENTS I would like to acknowledge Larry Anderson for guiding me through the thesis process, his willingness to provide feedback on experimentation and observations, and for referring me to the student position at the United States Ge ological Survey (USGS) to pursue this project I would also like to acknowledge Mike Pribil for the multitude of ways in which he has helped make this thesis project a reality. Not only did Mike recommend the project and initial suggestions to launch forw ard from, but he has also provided instruction on the operation of the MC ICP MS and DMA 80; introduction to the isotopic analysis of Hg; an extra set of hands for initial measurements; a sounding board to discuss approaches, results, and interpretations; and, constructive criticism for the manuscript herein. Finally, I would like to recognize the USGS as the provider for t he instrumentation, materials, and additional resources used in the completion of this thesis.


v TABLE OF CONTENTS CHAPTER I. INTRODUCTION AND BACKGROUND INFORMATION 1 Introduction 1 Mercury 1 Basics 1 Toxicology 3 Exposure 5 Fractionation 8 II. INSTRUMENT TECHNIQUES 14 Multi Collector Inductively Coupled Plasma Mass Spectrometry 14 Sample Introduction 14 Plasma and Ionization 16 Interface 17 Ion Optics 19 Electrostatic Sector 20 Magnetic Sector 21 Zoom Lens 22 Detectors 2 3 Entrance and Exit Slits 2 4 Direct Mercury Analyzer 2 5 Drying and Decomposition 2 6 Amalgamation 26 Detection 27 III. MERCURY SPECIFIC ANALYSIS 2 9 S ample Preparation Techniques 29


vi Bench Top and Micro wave Assisted Acid Digestion 29 Amalgamation 30 Solution Trapping 31 Reporting Protocol 3 1 Mass Bias Corre ctions and per mil Rep orting 32 Mass Independent Fractionation Reporting 3 3 Limited Usefulness of Raw Ratios 3 5 Sample Introduction Methods 3 6 Transient Signal Method 3 6 Cold Vapor Method 3 7 Syringe Method 38 Data Acquisition 40 Quality Analysis 41 IV. METHOD DEVELOPMENT 42 Purpose 42 Development 43 Dir ect and Indirect Interfacing 43 Samp le Transport and Interfacing 45 V. EXPERI MENTAL METHODS 54 Sample Preparation 54 Acid Digestions 54 Combustion 56 Isotope Analysis 58 Cold Vapor Analysis 59 Direct Gas Analysis 59


vii Results 60 Complications 60 Data Analysis 62 Discussion 64 Precision 64 Waste and Preparation Minimization 67 Application 68 Future Work 6 9 Conclusion 70 REFERENCES 72 APPENDIX A. Experimental Data 7 6


1 CHAPTER I INTRODUCTION AND BACKGROUND INFORMATION Introduction Elemental Mercury (Hg) and many of its compounds are toxic to humans and other animals (Clarkson and Magos, 2006; Schweiger, Stadler, and Bow es, 2006). A thorough understanding of its biogeochemical cycling in the environment is necessary to characterize its behavior on local and global scales. Mercury isotope ratios can be used to help build this understanding. Many believe isotope ratios w ill prove useful in trac ing Hg sources, sinks, and reactions once individual processes are better understood in the environment (Gray, et al. 2013; Jackson and Muir 2004 and 2012; Jackson, et al. 2008 ; Fen g et al. 2010; and, Sonke et al. 2010). Traditional sample preparation for Hg isotopic analysis can require significant time and the use of ha zardous chemicals ( Klaue and Blum, 1999; Foucher and Hintelmann, 2006). Alternative methods have been published that attempt ed to addresses these complic ations with varied success (Evans, Hintelmann, Dillon, 2001; Xie, et al., 2005; Sonke, Zambardi, and Toutain, 2008). A partially automated experimental method was developed to further explore and streamline these techniques. Standard reference materials were prepared and analyzed using this method and previously published methods in a direct comparison study. The focus of this study was to determine whether a combustion and direct gas introduction method for Hg isotope analyses in high organic and other sample matrices possesses similar or better precision than previously published methods. A second goal was to compare the accuracy and efficiency of these methods to identify preferred techniques based on sample matrix and study requirements. M ercury Basics Mercury is the eightieth element in the periodic table and possesses seven stable isotopes, Table 1.1. Mercury also possesses a [Xe]4f 14 5d 10 6s 2 electron configuration. These filled 4f, 5d


2 and 6s orbitals promote chemical stability. First, second, and third ionization energies are 1006 kJ/mol, 1809 kJ/mol, and 3300 kJ/mol as compared to those for carbon at 1086 kJ/mol, 2352 kJ/mol, and 4619 kJ/mol, respectively (CRC Handbook of Chemistry and Physics, 1972). M ercury engage s in oxidation redu ction chemistry regardless of its stable configuration The two most common ions of Hg are the mercuric ion (Hg 2+ ) and the mercurous ion (Hg + ), the latter being most commonly found in a diatomic form (Hg + Hg + state and other properties classify it as a soft acid, suggesting strong inte ractions with soft bases, as observed with various sulfur species. Table 1.1: Natural Mercury Isotope Abundance Isotope Mass (amu) 1 % Natural Abundance 1 196 Hg 195.9658 0.146 198 Hg 197.9668 10.02 199 Hg 198.9683 16.84 200 Hg 199.9683 23.13 201 Hg 200.9703 13.22 202 Hg 201.9706 29.80 204 Hg 203.9735 6.85 1 CRC Handbook of Chemistry and Physics, 1972 Mercury interacts with other metals to form alloys called amalgams. Of particular interest are the mercury silver and mercury gold amalgams. The mercury silver amalgam is used as a dental filling and is a potential route of human exposure to Hg (WHO, 1991). The mercury gold a malgam is generated in the mining and extraction of gold. Mercury gold amalgams are also generated though atmospheric analysis of Hg T he amalgam is formed by sampling air containing Hg through or across gold metal often in the form of beads or crumpled foil. The gold trap can be heated at a later time to release Hg vapor in a controlled setting for analytical analysis. Elemental Hg possesses a vapor pressure of 1.58 x 10 06 atm under typical atmospheric conditions (CRC Handbook of Chemistry and Physics, 1972) and creates a saturation concentration of 14 mg Hg/m 3 at room temperature (Clarkson and Magos, 2006). The combination of a stable monatomic form and high vapor pressure allow for atmospheric residence times around a year for Hg 0 and shorter times for oxidized species (Mason, 1994). Unreactive Hg


3 species can be transported far from sources of origin whereas more reactive species deposit nearby. Mercury contamination, particularly as a result of volcanic activity or energy generation from co al power plants, can result in long range transport This problem is only exacerbated by the behavior of Hg in natural systems. Toxicology The human body interacts with inorganic and organic Hg species differently leading to different lifetimes and locat ions of residence within the body and resultant deviations in the health. I nteraction with Hg compounds depend s largely on the compound itself and route of exposure, but complex building blocks (Clarkson and Magos, 2006) Gaseous Hg can enter the body through the lungs and disperse throughout the body due ross nonpolar barriers such as cellular membranes (Clarkson and Magos, 2006). Studies have shown that approximately 74% of inhaled Hg, from an atmospheric concentration of 0.1 mg Hg/m 3 can be retained in the body (Clarkson and Magos, 2006). Of that reta ined, 10% is found in the brain and 50% in the kidneys. M ercury can interact with catalase, an enzyme found throughout the body that is used to breakdown hydrogen peroxide molecules. Interaction with this pathway leads to in vivo Hg oxidation described in Equations 1.1 and 1.2 (Clarkson and Magos, 2006). 1.1 Cat OH + H 2 O 2 Cat OOH + H 2 O 1.2 Cat OOH + Hg (g) Cat OH + HgO Oxidized Hg can also be introduced to the body by direct exposure to compounds containing mercuric ions, mercurous ions, or other molecules that readily breakdown into these ions M ercurous ions may also dissociate into elemental Hg and a mercuric ion within the body (Equation 1.3). After oxidation or dissociation the resultant mercury ions are likely to exhibit


4 inh ibited mobility throughout the body due to their charged nature and remain largely deposited within the cell. 1.3 Hg 2 2+ Hg + Hg 2+ Mercury chemistry within the body is largely determined by interaction with sulfur containing molecules trong affinity to sulfur (Clarkson and Magos, 2006). According to Clarkson and Magos, t his is so much the case that only thiols have proven effective in removing Hg from the body (2006). This leads to interactions with the cysteine amino acid whose thiol group can play important roles in numerous proteins. O xidized Hg has also been found in liver bile as a complex with reduced glutathione, a tri peptide made up of L cysteine, L glutamic acid, and glycine (Clarkson and Magos, 2006). Dimethylmercury (CH 3 HgCH 3 ) is extremely toxic due to its ability to cross nonpolar barriers easily, but is unlikely to be encountered outside of laboratory synthesis. Methyl mercury (CH 3 Hg + ) and ethylmercury (CH 3 CH 2 Hg + ) are the primary organic Hg species of interest due to ea se of exposure (Clarkson and Magos, 2006). Other organic mercury compounds are found to readily metabolize into inorganic Hg (Clarkson and Magos, 2006). M ethyl mercury and ethylmercury are also capable of converting into inorganic forms, ethylmercury is c onsidered to undergo this transformation more readily (Clarkson and Magos, 2006). U nlike ethylmercury, methylmercury appears to harm the central nervous system exclusively (Clarkson and Magos, 2006). Methylmercury possesses relatively high mobility withi n the body by its interactio ns with proteins; t he ion appears to enter cells by complexing with cysteine and leave cells by complexing with reduced glutathione Mercury i s also found to bind strongly with Se 2 an interaction that may account for the long term residence of inorganic Hg in the brain (Clarkson and Magos, 2006). Selenium is a micronutrient used in complexes that remove reactive oxygen species, similar to catalase discussed above, particularly in the brain and endocrine systems (Ralston and Ra ymond, 2010; Carvalho, et al. 2008). Reaction with Hg is thought to irreversibly inhibit these function and may


5 explain the neurological symptoms associated with Hg toxicity (Ralston and Raymond, 2010; Carvalho, et al. 2008). Particular psychological con increasing shyness, loss of self confidence, anxiety, and a desire to remain unobserved and One study of the mercuric ion ex posed individuals to levels typical of the larger population by having them ingest 6 g of Hg 2+ (Clarkson and Magos, 2006). The amount of Hg absorbed was found to vary widely between individuals but reached as high as 16% of the dose. Two half lives were observed: 2 and 41 days for unabsorbed and absorbed Hg respectively. This study showed that Hg found in blood after exposure to the mercuric ion is one tenth the concentration when compared to Hg vapor and methylmercury exposure. The Hg content in human blood and hair is used to ascertain brain concentrations by the relationship described in Equation 1 .4 (Clarkson and Magos, 2006). T his is n ot to say that exposure to Hg 2+ is not dangerous; ingestion of as small a quantity as 1 g of Hg salts can lead to shock and death (Clarkson and Magos, 2006). 1.4 Exposure Natural processes, such as volcanic an d geological activity, as well as anthropogenic processes, such as mining and coal power plant energy production, are sources of Hg contamination in the environment. A tmospheric concentrations of Hg vapor are negligible when considering the overall inorga nic Hg exposure of the general population (Clarkson and Magos, 2006). The ingestion of liquid Hg is not directly problematic because it is poorly absorbed during passage through the digestive tract, but it does pose a threat by exposing the individual to Hg vapors (Clarkson and Magos, 2006). Though ingestion is one possible means of exposure, individuals can also be subjected to Hg vapor by spills of liquid Hg manufacturing processes, or dental amalgam. Gold mining that utilizes the formation of mercury gold amalgams can also be


6 of concern. In these processes the elements are separated by heating the amalgam until Hg is removed as a vapor. Similarly, Hg itself is commonly produced by heating a Hg ore, mercuric sulfide or cinnabar, and condensing the re sultant vapors. Mercury vapor can have a long atmospheric residence time and be transported great di stances from its point of origin, with eventual r emoval by oxidation and washout or deposition. Mercury that makes its way to aquatic systems finds itself controlled by a number of parameters including: pH, biological activity, nutrient and photo availability, temperature, and the presence of complexing agents (Ravichandran 2004). Bergquist and Blum conducted a study to elucidate the behavior of elemental and methylmercury in aquatic systems (Bergquist and Blum, 2007). The retention of Hg in aquatic systems was evaluated using photo availability, the presence of organic matter, and the specific form in which Hg was present. The amount of Hg undergoing pho toreduction in the presence of dissolved organic matter over the course of 300 minutes was observed for the mercuric ion and methylmercury: >90% and 20% respectively (Bergquist and Blum, 2007). Rates of reduction in the absence of light, also in the prese nce of dissolved organic matter, were observed. Under these conditions the mercuric ion underwent reduction at a reduced rate (20% in 300 min) and reduction of methylmercury was not detected (Bergquist and Blum, 2007). Although this experiment was conduc ted in a controlled laboratory setting, the observations indicate d that Hg is more than capable of escaping aquatic systems and reentering the atmosphere. Mercury can be captured through the passive and active uptake of microorganisms in aquatic systems Various forms of oxidized Hg can be reduced by microorganisms in addition to reduction by photolysis or interaction with dissolved organic matter. The presence of strong sulfur binding sites in dissolved organic matter, often found in higher concentration s than Hg in natural waters, can play a significant role in which form the Hg present. For these reasons the production of methylmercury is not necessarily indicative of total Hg initially present within an aquatic system, but the Hg available for methyla tion


7 Microorganisms found in aquatic sediments specifically sulfate reducing bacteria, methylate Hg These species find Hg 2+ to be more toxic than the organic forms of Hg, so they methylat e the element as a self protection safety measure (Clarkson and Magos, 2006). Overtime the microorganisms accumulate methylmercury and are consumed by larger organisms. This continues throughout the food chain, resulting in an increased organic Hg concentration in the larger predators as compared to those that are co nsumed. This can also lead to increasing concentrations of inorganic Hg as well as organic Hg compounds are broken down through biochemical pathways. Mercury is found to be present in nearly all aquatic species. This biomagnification of the element reac hes its maximum concentrations in those species at the highest trophic levels, such as long lived predatory fish. This process is sufficiently significant as it has been documented to take ocean water concentrations of approximately 1 ppb Hg up to shark m uscle tissues as high as 4 ppm Hg (Clarkson and Magos, 2006). M ercury is generally found as a CH 3 Hg cysteine complex in the proteins of fish muscle tissues (Clarkson and Magos, 2006). Some species of fish and aquatic mammals may have considerably larger portions of inorganic Hg by comparison (Clarkson and Magos, 2006). Of the methylmercury ingested by humans from fish consumption, about 95% is absorbed into the human body (Clarkson and Magos, 2006). Consumption of livestock can also lead to exposure to Hg particularly inorganic forms. This occurs by consuming livestock after it has converted the organic Hg in feed made with fish to inorganic Hg by metabolic processes (Clarkson and Magos, 2006). E xposure to organic Hg is primarily through fish consum ption, but other forms of exposure are possible such as chemical manufacturing and laboratory work. Extreme care should be taken when interacting with any quantity of these species in their pure chemical form as small amounts have proven to be lethal. In one case a scientist exposed to less than 0.5 mL of dimethylmercury in 1997 underwent rapid physical and mental degradation resulting in death after a short incubation period of a few months (Clarkson and Magos, 2006).


8 Fractionation Isotope f ractionation is the process by which a compound possessing two or more isotopes of the same element undergoes a physical or chemical change that results in the unequal distribution of the isotopes in the products and reactants. The preferential reaction o f one isotope over another most often occurs due to mass differences resulting in mass dependent fractionation (MDF), but may also occur by other processes resulting in mass independent fractionation (MIF). Mass Dependent Fractionation Mass dependent frac tionation originates from the differences in vibrational energy levels between molecules possessing different isotopes (Bigeleisen 1965). The vibrational energy levels of a bond between two atoms can be described by a harmonic oscillator as per Equation 1 .5. 1.5 where E v is the vibrational energy level k is a force constant characteristic of the bond 1 m 2 /(m 1 + m 2 ) v If the first atom, m 1 is kept the same while the second atom, m 2 is interchanged between two isotopes of the same element, a deviation is observed in vibrational energy levels. As m 2 increase s in mass the reduced mass becomes larger. Due to the reciprocal relationship between reduced mass and vibrational energy, the zero point energy decreases with increased mass. A molecule with a heavier isotope will possess a zero point vibrational energy level lower in energy than a molecule comprised of a lighter isotope. The harmonic oscillator description is incomplete as there is an element of anharmonicity present in vibrational bonds As bonded atoms move closer the potential energy grows steeper than might otherwise be described. As atoms move apart there is a definite point when the bond dissociates and the potential energy ceases to climb. Higher vibrational energy levels converge as they approach the point of dissociation (Carroll, 2011). Although the zero point energy levels


9 differ for each isotope, the point of dissociation does not. This creates the distinct energy difference that allows for mass dependent fractionation (Carroll, 2011). This differenc e can be seen in Figure 1.1, adapted from Carroll (2011). For two molecules of differing isotopic composition, each near their respective zero point vibrational energy level, the molecule with the heavier isotope requires a greater amount of energy to dis sociate than the molecule with the lighter isotope. H igher vibrational energy levels become increasingly occupied with higher temperatures Due to the nature of the convergence at higher vibrational energy levels, the gap between energy levels diminishes along with the mass driven differences described. The mass dependent fractionation of reactions at elevated temperatures is less extreme than it might otherwise be at cooler temperatures. Figure 1.1: Disparate Dissociation Energies between Isotopic Mol ecules


10 Reactions or transformations occurring in open systems allow products to leave upon formation and results in characteristic ratios of the concentration of one isotope to another. Unidirectional steady state kinetic fractionation, or mass dependent fractionation that occurs in ideal open systems, is sometimes called Rayleigh fractionation and is described by the Rayleigh equation, Equation 1.6 (Kendall and Caldwell, 1998). A full explanation of the Rayleigh equation is beyond the scope of this disc ussion, but the observations it describes will be explored. 1.6 where R = Ratio of heavy isotopes to light isotopes in the reactant R 0 = Initial ratio of isotopes in reactant X l = Concentration of lighter isotope in the reactan t = Initial concentration of lighter isotope in the reactant = Fractionation factor = R p /R r where R p and R r are the ratios of heavy isotopes to light isotopes in the products and reactants respectively Reactant materials become isotopically heavier as lighter isotopes more readily react and form products. This occurs because the dissociation energy for the lighter isotopic molecules is smaller and more readily overcome than the dissociation energy associated with heavier isotopes. The magn itude of the difference between isotope ratios in reactants and products depends on the reaction pathway, reaction rate, and relative bond energies (Kendall and Caldwell, 1998). In practice, the Rayleigh equation is also used to describe reactions and tra nsformations that occur in closed systems. These conditions allow for mixing between reactants and products, resulting in equilibria where available forward and backward reactions are at equal rates. Though lighter isotopes more readily transform into pr oducts initially, equilibrium allows heavier isotopes to eventually settle preferentially in more energetically stable states. This occurs because the associated dissociation energy in this state is larger and more difficult to overcome than all other sta tes under consideration. These states are often characterized by higher oxidation states or denser compounds (Kendall and Caldwell, 1998). Equilibrium conditions allow for smaller


11 differences between resulting isotope ratios in reactants and products (Ke ndall and Caldwell, 1998). It is important to note that equilibrium conditions allow for equal rates of forward and backward reactions, but does not necessitate equal isotope ratios for reactants and products. As alluded to above, these isotope ratios depend on a number of conditions such as temperature, chemical composition, crystal structure, and pressure (Kendall and Caldwell, 1998). Due to the relationship between mass dependent fractionation and the mass of the isotope, a relationship can be drawn between the various isotopes. For example, if the mass dependent fractionation of one isotope is measured, the expected mass dependent fractionation for the remaining isotopes can be calculated. In the case that only mass dependent fractionation occurs, observed isotope concentrations do not differ from the calculated values. Deviations from this trend indicate that mass independent fractionation also plays a factor. Mass Independent Fractionation Nuclear volume effect One process of mass independent fr actionation is the nuclear volume effect. The nucleus of a single element increases in size and decreases in charge density with increasing numbers of neutrons. The potential energy of electrons also increases with these factors, particularly for heavier elements (Schauble, 2007). This increase in potential energy promotes the loss of electrons close to the nucleus in heavier isotopes over similar losses for lighter isotopes. of the heavy isotope for the chemical species with the smallest number of s electrons in the bonding or valence 202 Hg is expected to preferentially accumulate in a Hg 2+ state over a Hg 0 state compared to 198 Hg. Upon first inspec tion the nuclear volume effect might be thought to be indistinguishable from mass dependent fractionation. Both the zero point energy and the potential energy of electrons change with isotopic mass; however, the change to nuclear shape and size with


12 incre asing neutrons is not consistent. The nuclear volume effect for odd isotopes produces a less drastic shift that more closely resembles the shift from the neighboring lighter isotope (Bigeleisen, 1996). Magnetic isotope effect The magnetic isotope effect (MIE) is an alternative process which results in mass independent fractionation. The nucleus of an odd isotope possesses an unpaired proton or neutron, resulting in non zero angular momentum, or spin. This nuclear magnetic moment can interact with elect rons as seen by the splitting in spectra produced by Electron Paramagnetic Resonance spectroscopy (Buchachenko, 2001). These interactions allow electrons to undergo spin conversions more readily than in atoms with non magnetic nuclei. Odd isotopes more r eadily undertake reactions requiring spin conversion. The MIE operates by altering the speed of a reaction, not the ground state energies of the products or reactants. The MIE is consequently unobservable in reactions that are at equilibrium. This obser vation can be used to distinguish between nuclear volume effects and magnetic isotope effects (Buchachenko 2001). UV self shielding Photochemical self shielding, also referred to as UV self shielding, is another process by which mass independent fractiona tion can occur. According to Lyons and Young, this process requires species of different isotopic composition to absorb light in narrow wavelengths unique to that isotope (2005). As light strikes a material, wavelengths associated with abundant isotopes are readily absorbed, leaving behind a proportionally large number of unexcited species as compared to less abundant isotopes. The less abundant an isotopic species the larger the proportion of its population will be that excites and reacts. Products gai n isotopic signatures with larger concentrations of isotopes that were proportionately less abundant within initial reactants. Example Application experiment examining Hg beha vior in aquatic systems. Mass dependent and mass independent


13 fractionation was observed in the photoreduction of Hg species in the presence of organic matter. Reduction in the presence of organic matter without light was found only to induce mass depende nt fractionation (Bergquist and Blum, 2007). Mass independent fractionation was observed with 201 Hg and 199 Hg, but not with 204 Hg. The nuclear volume effect predicts a deviation in 204 Hg. As this is not observed, mass independent fractionation is expected to occur by the MIE and photoreduction of Hg in aquatic systems occurs by radical intermediates that undergo spin conversion (Bergquist and Blum, 2007). This experiment also examine d the rates of the reduction processes. Bergquist and Blum suggest ed that the initial concentrations of methylmercury can be estimated by using the mass independent fractionation signature of the sample along with the estimated rate of photoreduction (200 7). Additional observations were also made when these results were compared to Hg isotopic data gathered from fish tissues. Further mass dependent fractionation was observed in fish tissue as compared to water content, but mass independent fraction signat ures remained the same (Bergquist and Blum, 2007). This suggests biological fractionation processes are mass dependent and mass independent fractionation signatures are characteristic of source waters and can be preserved in fish tissues Mercury isotope ratios within fish may yield information regarding the presence of methylmercury in natural waters throughout its lifetime. Methylmercury concentrations existing prior to losses by photoreduction may be calculated in the associated waters.


14 CHAPTER II INSTRUMENT TECHNIQUES Multi Collector Inductively Coupled Plasma Mass Spectrometer There are many different types of mass spectrometers available with eac h possessing its own strengths and weaknesses. Rather than considering the workings and effectivene ss of each instrument, the specific tools utilized for this project will be examined with references to alternative instrumentation and methods where appropriate. The multi collector inductively coupled plasma mass spectrometer (MC ICP MS) is a double fo cusing magnetic sector instrument that guides ions generated by a plasma source to an array of multiple detectors, in this case twelve detectors and three ion counters Figure 2.1 adapted from Nu Instruments (2003) The first sector includes an electrosta tic analyzer (ESA) which discriminates ions within the ion beam by an electrical field that only transfers ions with a specific energy. The second sector houses the magnetic sector which separates ions by mass and energy using angular dispersion. After t he magnetic sector, ions are directed towards individual detectors by enhanced variable zoom optics. The primary benefit of multi collector instruments is their ability to measure isotopes simultaneously rather than scanning or jumping the magnet to encomp ass the mass range of the isotopes of interest. Typical MC ICP MS instruments operate at high vacuum in the order of low in the 10 9 torr, using an Edwards E2M80 vacuum pump for the interface region, three turbo pumps backed by a single rough pump, and tw o ion pumps. Sample Introduction From laser ablation techniques to hydride generation, many different sample introduction methods exist for MC ICP MS analysis. Selection between the different methods depends on the nature of the sample and the specific i nformation required by the investigator. Two common modes for sample introduction on MC ICP MS instruments are wet aerosol and dry aerosol.


15 Additional sample introduction techniques employed in this project will be discussed in greater detail later. Figure 2.1: Multi Collector Schematic In the wet aerosol technique a sample solution is introduced through a pneumatic nebulizer into a spray chamber. The spray chamber and orientation of the nebulizer is designed to remove large vapor molecules while all owing the wet aerosol to travel in to the plasma of the instrument. Due to the introduction of larger amounts of solvent, a greater interference from the formation of oxides and hydrides can result. These form by interactions between the plasma, water mol dry aerosol, improving the sensitivity of MC ICP MS analyses (CETAC).


16 The desolvating nebulizer used in the United States Geological Survey ( USGS ) MC ICP MS high reso lution lab is an Aridus II TM The purpose of this introduction system is to transition solvated analytes into a dry aerosol by dispersion through a micro nebulizer and separation from the solvent by heating. First, a solution is aspirated through a capill ary tube and into a concentric tube pneumatic micro nebulizer (100 L/min) by a flow of argon gas (CETAC). The nebulizer is designed to convert a solution into a fine aerosol with high transport efficiency. Second, this aerosol is carried by argon gas in to a heated spray chamber where solvent molecules are volatilized. Third, a heated PTFE microporous membrane removes solvent molecules from the aerosol (CETAC). The solvent molecules move through the heated membrane and are carried away by a counter flow of argon gas. Finally, as the solvent vapors are removed, the nonvolatile analytes are carried into the plasma. Plasma and Ionization Plasmas are ionization sources with wide application due to their near universal ionization of species throughout the p eriodic table (Wieser and Schwieters, 2005). Montaser, et ionization potentials are ionized due to the extreme temperatures maintained by the plasma. Tempe ratures can range on the order of 6000 10,000 K. The argon plasma is generated by flowing argon gas through oscillating electric and magnetic fields. The fields are generated by supplying power from a radio frequency generator through a water cooled cop per load coil that circles the end of a torch (Montaser, et al. 1998) Figure 2.2 adapted from Thomas (200 1 Part IV) The torch is a quartz glass structure made up of three concentric tubes through which the argon gas and dry sample aerosol are introduce d. The torch is formed in such a way as to force the sample aerosol to maintain a course through the center of the plasma.


17 Figure 2.2: Plasma and Interface Region When a spark is discharged the argon is seeded with electrons that move from interaction with the magnetic field (Turner and Montaser, 1998). Collisions of the electrons and the argon atoms can convert the argon atoms into positively charged ions by the r elease of additional electrons. The oscillating magnetic field stimulates further movement in the charged argon ions and free electrons (Turner and Montaser, 1998). According to Turner and Montaser, steady state plasma is obtained when electrons are prod uced at the same rate they are lost through the recombination with positively charged argon ions (1998). This recombination is further characterized by the release of energy. Analyte atoms are ionized in the same manner as the argon gas, by collision wit h the energetic electrons produced by the plasma. Interface The interface region of the instrument consists of a series of two small metallic plates called cones Figure 2. 2 Each cone possesses a fu nnel shape to transfer ions from the plasma to the inst rument and also direct their flow. Between the cones is a chamber, referred to as the transfer region. According to the Nu Plasma instrument manual, t he transfer region is pumped by a large volume, two stage rotary pump that decreases the pressure from a tmospheric to a range of 5 to 10 mbar during normal operation (Nu Instruments, 2003) The internal vacuum of the


18 instrument is maintained at a pressure of 10 9 mbar. The step wise pressure reduction allows the intense vacuum conditions within the instrum ent to be maintained, important in minimizing collisions between analyte ions and internal gases thus increasing the transfer efficiency of the ions into the instrument and to the detectors. The cone openings are oriented towards the center of the plasma, allowing analyte ions to enter the instrument and be funneled into an ion beam. The first cone is called the skimmer and is broad with a wide angled shape; the second cone is the sampler which is narrow with a small angled shape. Ions enter the instrumen t by passing through the skimmer cone into the transfer region. Movement into this lower pressure chamber causes the sample to undergo supersonic expansion and spread radially outward. MS, the com position of the ion beam is maintained as ions travel from the plasma to the interface region (200 1 ). Space to charge interactions of ions can be considered small at the orifice of the skimmer cone due to the size of the orifice compared to the minimum di stance of interaction for the ions. Due to their low mass, electrons diffuse outward much more rapidly than the other ions. This allows the ion beam to develop a net positive charge and stronger interactions between the ions become available. Due to thes e repulsive interactions, heavy ions are more likely to dominate the center of the ion beam. This behavior results in fractionation within the instrument that must be corrected for. The specific corrections used for Hg analysis will be discussed later. A MS, 1 in every 1,000,000 ions generated in the plasma of a normal ICP 1 ). The two stage rotary pump results in lower pressures and improves this ion transfe r efficiency, but newer models outfitted with jet interfaces improve the ion transfer efficiency even further. Problems with transfer efficiency can stem from direct and indirect interactions between analyte species and matrix species. Matrix ions can do minate an ion beam and force the desired species out. Matrix species can also preferentially ionize within the plasma and decrease the ionization


19 efficiency of the analyte species. For isotopic work it is common practice to minimize concentrations of non analyte species through pre analysis preparation work. For many systems, like copper, iron, zinc, lead, and strontium, this includes column chromatography that yields Ion Optics The function of the ion optics is to pull positive ions out of the interface region and steer them into the mass analyzer of the instrument. Repulsive interactions must also be decreased in order to achieve high transfer efficiencies. This is accomplishe d by utilizing electric fields generated by applying voltages to metallic plates. From the transfer region positive ions are pulled into the remainder of the instrument by attraction to an extraction lens. The extraction lens has a negative voltage and accelerates ions into their flight path. Electrons and negative ions are repulsed by this initial electric field. This process grants positive ions with the kinetic energy profile described by Equation 2.1 (Nu Instruments, 2003). 2.1 KE = zeE = mv 2 whe re KE is the kinetic energy z is the number of charges on the ion e is the charge of the electron E is the electronic field of the extraction lens m is the mass of the ion v is the resulting velocity of the ion Ions of the same charge will possess the same kinetic energy, but the velocity of these ions will differ based on their mass. Although very low pressure is maintained in the instrument by multiple turbomolecular and ion pumps, collisions between the ions and gases within the instrument can occur. Due to collisions and repulsive interactions between ions within the ion beam, ions can stray from the desired path. The lens system resists these effects by steering stray ions back into the ion beam. This pro cess increases the transfer efficiency of the ions. The lens system is a series of metallic plates with various applied charges. It focuses the ion beam in both the vertical and horizontal planes as ions travel toward the ESA


20 Although electrons and nega tively charged ions are reflected by the electric field generated by the extraction lens, photons and neutral species may move through. To decrease the possibility of these species making it through the system to the detector, or from depositing on lens c omponents, the ion beam is normally steered off axis or a positively charged shadow stop is placed in the ion path. These elements are not present within the Nu Plasma MC ICP MS. Electrostatic Sector The magnetic sector that follows the ESA discriminates based on both mass and energy, but alone results in a suboptimal level of resolution. Interaction with the extraction lens results in ions that exhibit a distribution of accelerations. Ions of the same mass with unequal velocities ensue. These unequal velocities result in dissimilar paths for ions of the same isotope and the ion beam is observed to widen. To account for this widening the ESA is used to discriminate energy distributions prior to the magnetic sector. This double focusing characteristic of the instrument results in the greater resolution desired. After traveling through the entrance slit and initial ion optics, ions move through another series of lenses and into the ESA. These lenses determine the angle of the ion beam and serve to focus the beam in the vertical plane. The ESA consists of two charged, curved plates set parallel to one another The electric field created between the two plates is characterized by the equation V/d where V is the applied voltage and d is the distance betwe en the plates (Nu Instruments, 2003). Ions passing through interact with this electric field and experience a force that accelerates them to the center of the circular path. The force can be expressed as described in Equation 2.2 (Nu Instruments, 2003). 2.2 where F is the force exerted by the electric field z is the number of charges on the ion e is the charge of the electron V is the applied voltage of the ESA d is the distance between the parallel plates of the ESA


21 Equation 2. 3 expresses the centripetal acceleration experienced (Nu Instruments, 2003). 2.3 where a is the centripetal acceleration v is the velocity of the ion r e is the radius of the circular path the ion takes in the ESA The relationship between as in Equation 2.4 (Nu Instruments, 2003). 2.4 where m is the mass of the ion It appears as though the ESA separates the ions based on mass. Although this is not entirely untrue, it is not the full story. By substituting in the kinetic energy expression from above we observe a relationship between the initial acceleration of the ion and the interaction with the ESA. This relationship is described in Equation 2.5 (Nu Inst ruments, 2003). 2.5 The relationship between the centripetal force and radius the ions experience depends on their kinetic energy. Ions traveling too quickly for their mass will impact the outer plate of the ESA while ions traveling too slowly for their mass will impact the inner plate of the ESA. Those ions with the right kinetic energy pass through the ESA. Magnetic Sector The ion beam travels from the ESA, through a monitoring plate, and into the magnet sector. The magnetic sector behaves similarly to the ESA in that it curves the ions around a bend and transfer s only the ions desired. Two forces are balanced here, the force exerted by the magnetic field that pulls in towards the inner curve of the path and the centripetal force of the ion s that pushes outwards away from the inner curve of the path. The magnetic and centripetal forces are described by Equations 2.6 and 2.7 (Turner, et al., 1998). 2.6 where F M is the force exerted by the magnetic field B is the magnetic fi eld strength z is the number of charges on the ion e is the charge of an electron v is the velocity of the ion


22 2.7 where F C is the centripetal force m is the mass of the ion v is the velocity of the ion r m is the radius o f the circular path the ion takes in the magnet These forces are equal to one another for those ions that successfully traverse the magnetic sector (Equation 2.8). 2.8 By solving the kinetic energy equation for velocity and substituting in, the following relationships in Equation 2.9 are determined (Turner, et al., 1998). 2.9 interaction with the magnetic field, mass, and kinetic ene rgy. As such, the magnetic sector discriminates based on both mass and kinetic energy. The end effect is to separate each mass in to its own angular dispersion, resulting in a unique radius for each isotope ion from the ESA to the detectors. Zoom Lens Fro m the magnetic sector the ions move through a series of two quadratic fields created by a succession of two charged plates. The electric fields generated act in a similar manner as the optical lens systems discussed earlier. The ion beams, now separated into different angular dispersions, are focused or de focused such that they are directed towards the desired detectors, either faraday cups or ion counters. This focusing and de focusing allows the instrument to detect a larger or smaller array of masses, respectively, than would otherwise be available. The charged plates are used in place of a typical quadrupole system as the latter would cause further distortion of the ion beam (Nu Instruments, 2003). According to Wieser and Schwieters, zoom lens system s also filter out ions with less optimal kinetic energies and flight trajectories due to collisions (2005).


23 Detectors Although there are multiple detector options available for the MC ICP MS, the specific detectors used for Hg analysis are faraday cups. F araday cups work by receiving ions through the (Nu Instruments, 2003) The resulting current is mea sured and is proportional to the number of charges developed on the metallic cup and the concentration of the analyte. Large ohmic resistors on the order of 10 11 the signal. Resistors decrease the flow of electrons in the circuit by providing an increased resistance. This effectively reduces minor fluctuations in the current as a result of ion detector interactio ns in favor of a more stable signal. The entrances to the detectors are wide enough to receive ions across a mass range. While scanning the instrument over a range of masses the signal of the detector will climb sharply as the ion beam is moved towards t he detector entrance. Over the course of scanning the detector the signal will drop off sharply. This creates a distinct flat topped peak. Alterations in the peak size and shape can be used to identify interferences of similar masses. A pre amp gain calibration is performed every day by applying a known voltage to each detector. The detector output is then compared to the known value of the current to create a calibration factor that is automatically applied to subsequent signals. This allows the output from each detector to be directly compared, accounting for individual diffe rences between each detector. During this process the electronics of neighboring detectors can be linked to extend the range of any one single detector from 10 V to 20 V. This increases the flexibility of the instrument by allowing for the analysis of is otopic systems possessing an overabundance of a single isotope compared to others.


24 After the instrument has warmed up and the detectors have been calibrated, ion beam focusing and peak shape tuning is conducted. The edges of the flat top peaks should be s harp and coincide with the signals from the other detectors. This ensures fluctuations in ionization from the plasma, or other small disturbances throughout the instrument, are equally seen across all cups and each individual analyte. The focusing of the ion beam is tuned by adjusting the voltages applied to the metallic plates of the optical lens systems. The peak shape is tuned by adjusting the voltages of the zoom lens and by manual adjustment of the entrance and exit slits. Entrance and Exit Slits En trance and exit slits are used to decrease the space available for the ion beam to move through. The entrance slit appears after the initial ion optics system, but before the ESA The exit slit appears after the magnetic sector and zoom lens, but before the detectors. By decreasing this space the slits physically skim the ion beam. I nterferences possess a slightly different mass than the analyte and will occupy a different location within the respective ion beam. Both the entrance and exit slits are m anually adjustable and made from metallic strips. The entrance slit possesses a single narrow window that can be moved to cut off part of the ion beam entering the ESA. The exit slit possesses multiple windows that allow the passage of different ion beam s generated by the magnetic sector. This strip can be moved to cut off part of each ion beam entering the faraday cups. This tool reduce s polyatomic and isobaric interferences by increasing the mass resolution of the instrument Equation 2.10 (Nu Instru ments, 2003). T he width of the peak is traditionally (Nu Instruments, 2003). 2.10 where R = resolution M = mass of peak studied M = width of peak


25 Detector resolution, or the mass difference between detectors, is arbitrary in the sense that it can be altered by adjusting the zoom lens settings and changing the angles of the ion beams to different focal points guiding ions to differ ent detectors, based on the needs of the analysis performed. Although detector resolution can be helpful in resolving interferences, high resolution capabilities are primarily accessed by the use of entrance and exit slits. The range of mass resolution a vailable on the Nu Plasma is 300, 4000, and 8000 10,000. When the entrance and exit slits are used to resolve analyte peaks from their interferences in this manner, the MC ICP MS is said to be in high resolution mode. The resulting decrease in the signa l intensity requires higher concentrations of samples for analysis. The polyatomic oxygen interference for the 32 S isotope is one such example where high resolution mode is required. Only a resolution of approximately 200 is needed for the analysis of Hg isotopes T he full capabilities of the MC ICP MS are not required or utilized. The primary purpose for the use of the MC ICP MS in Hg isotopic analysis is the ability for simultaneous detection, allowing instrument fluctuations to be observed across all detectors and isotopes at the same time. Simultaneous measurement decreases error as compared to single collector systems that require jumping the magnetic sector or other mass to charge separating device, to different masses at different periods of time Direct Mercury Analyzer The direct mercury analyzer (DMA 80) available from Milestone Inc. measures the concentration of Hg within a sample using atomic absorption spectrophotometry as per the Beer Lambert Law, Equation 2.11 (Skoog, Holler, and Crouch, 2007) The absorbance of the 253.65 nm emission line signal generated by a known amount of sample is compared to a calibration curve generated by the analysis of numerous standard materials. The main advantage of the DMA 80 compared to other Hg analyzer s is the minimal sample preparation required for analysis.


26 2.11 where A = absorbance c = concentration l = path length Samples of up to 1.5 g or 1.5 mL are placed in tin or quartz sample boats. The sample boats are then loaded onto an auto sampler tray and introduced into the reaction tube at the appropriate time. Figure 2.3 describes the flow of operation for the DMA 80. Figure 2. 3 : DMA 80 Operational Schematic Drying and Decomposition During normal operating procedures the sample is dried by heating to 200 O C for 60 s followed by heating to 650 O C for 180 s for sample decomposition. Sample analysis is performed under a constant flow of oxygen at 200 mL/min. During this time the sample is dried and the organic and other combustible molecules are removed. The catalyst bed of the reaction tube contains potassium hydroxide and cobaltoxid, reactive species that trap interferences, oxidize organic species, and promote the reduction of non elemental Hg sp ecies (Milestone, Inc., 2008 and 2014). The remaining unabsorbed gases are flushed through the system prior to detection. Amalgamation A constant gas flow travels through a gold trap from the reaction tube. Mercury vapor that evolved during the drying an d decomposition cycle passes over the gold trap and forms an vented to waste, the gold trap is flash heated to 900 O C f or 12 s. As the gold trap is heated Hg is desorbed and the carrier gas pushes the Hg vapor through the detector cell.


27 Detection The DMA 80 has a single detector and three chamber detector housing, as described in Figure 2. 4 Light from a Hg lamp is directed through the first and third chambers a nd into a detector. The second chamber acts as a passage to prevent Hg vapors from occupying the first and third chambers simultaneously. Figure 2.4: Three Chamber Detector Housing Due to the disparate sizes of the two chambers, the first chamber posses ses a larger path length for light to travel than the third chamber. The increased path length of the first chamber allows for more accurate measurements of low Hg content on the order of 0 20 ng of Hg. The second chamber is used for larger Hg concentrat ions on the order of 20 1000 ng of Hg. From the three chamber detector housing the sample is directed through a carbon trap and vented through a fume hood to the atmosphere.


28 This set up allows for a larger range of detection otherwise unavailable to the instrument. A calibration is performed for each chamber separately and the instrument software automatically flips between the two based on a 30 ng of mercury threshold.


29 CHAPTER III MERCURY SPECIFIC ANALYSIS Sample Preparation Techniques Sample preparation for Hg isotopic analyses depends on the sample state. Solid samples require acid digestion methods or combustion and subsequent collection of released Hg vapors. Gas samples require trapping or collecting Hg from air. The three primary prepa ration methods for mercury analysis are: (1) acid digestion, (2) vapor amalgamation, and (3) solution trapping. Bench Top and Microwave Assisted Acid Digestion Typical bench top digestion is performed by subjecting a small aliquot on the order of 0.2 g of solid sample to a reverse aqua regia solution (1:9 HCl to HNO 3 ) with gentle heating. An alternative to this method is the use of microwave assisted digestion. Solid samples in the amount of 1 g are submitted to similar solutions as with bench top digest ion, but the mixture is retained within a microwave vessel and subjected to higher temperatures and pressures by means of microwave bombardment. top acid digestion preparation begins with submitting 0.2 g to 0.6 g of sample to 10 mL of a reverse aqua regia solution (2006). The acid solution is made up of doubly distilled concentrated HCl and HNO 3 Sulfuric acid and HNO 3 are mixed in a ratio of 3:7 respectively when H 2 SO 4 is used in place of HCl. The samp le acid mixture is heated to 100 120 O C for 4 6 hours. The solution is allowed to cool before filtering and/or centrifuging to remove remaining particulate matter. An aliquot of the supernatant is diluted with ultrapure water to a final concentration of approximately 2 3 ppb Hg in 2 5% HNO3 or 3% HCl. An oxidized form of Hg is maintained in the acidic solution. The low pH provides a proton rich environment that minimizes the reduction and escape of Hg as a vapor. According to EPA Method 3200, sample e xtracts for Hg analysis should be stored at 4 O C and analyzed within 846 prescribes additional protocol for the collection, digestion, and storage of Hg samples as listed in Table 3.1 (EPA, 2007).


30 Table 3.1: EPA Suggested Storage Protocol for Hg Samples 1 Sample Material Storage Conditions Collection Amount Digestion Amount Storage Time Aqueous HNO 3 to pH < 2 400 mL 100 mL 28 days Solid O C 200 g 0.2 g 28 days 1 Information reproduced from Table 3 2 of Chapter 3, SW 846 (EPA 2007) Despite the suggested storage requirements, an onsite Hg stock solution stored in a 1000 mL glass volumetric flask has been exposed to continuous light sources for multiple years without deviation from his torical data. This stock solution, prepared from a National Institute of Standards and Technology (NIST) Hg standard NIST 3133, was originally made up to the following specifications: 1000 mL of 1 ppm Hg in 3% HCl. The quality preservation of this solut ion may be attributable to its high concentration and large volume. Elemental mercury may evolve from st orage solution s and can exist in the head space of air tight containers but may be re solvated by multiple inversions. Microwave assisted acid digestion is another method used for the extraction of Hg. EPA Method 3200, as developed by Rahman and Kingston, begins by adding 10 mL of 4 M HNO 3 to 1.0 g of sample in an appropriately sized microwave vessel with a magnetic stir bar (2005). Appropriate ly sized microwave vessels are said to accommodate 1 10 g of sample an appropriate amount of solvent and apparatus specific to the microwave instrumentation (EPA, 2005). Vessels are sealed, inserted into the microwave device, brought to 100 O C over 2 mi n, and held at 100 O C for 10 min with magnetic stirring. Digestion solutions are filtered after they have cooled and are diluted with ultrapure water as necessary for analysis. Protocol is not provided within the method for pressure build up and headspac e gases within the vessels (EPA, 2005). Amalgamation Mercury gas released from direct emission sources and sample combustion can be collected for analysis. Sample gas is pumped across a gold surface and Hg vapor present forms an amalgam with the gold. Th e gold surface is often in a bead or foil form. The amalgam


31 produced is sealed and stored until analyses are performed As with the amalgam formed in the DMA 80, Hg is released by heating the gold trap. For their work with Hg emissions from volcanos, So nke, Zambardi, and Toutain pumped fumaro le gas es though a series of filters and two gold traps set in series (2008). The filters removed particulates acid vapor, and water vapor. Gold bead traps were set in series to monitor Hg that passed through the f irst trap, but Sonke, Zambardi, and Toutain reported that no such Hg break through was observed (2008). Solution Trapping A third preparation method is a combination of the previous two methods. Mercury vapor is released from a gold trap or by direct combustion of a sample and bubbled through an acidic solution. A strong oxidizing agent, such as KMnO 4 is included in the solution to help retain Hg (Biswas, et al., 2008). In their analysis of coal materials, Biswas, et al. decomposed and combusted 2 6 g of solid sample by heating it to 900 O C over the course of 6.2 hours (2008). A combined flow of Hg free oxygen and argon gases assisted the combustion and moved Hg vapors into a 25 mL solution of 1% KMnO 4 and 1.8 M H 2 SO 4 (Biswas, et al., 2008). Reporti ng Protocol Blum and Bergquist published recommendations for the reporting protocol of Hg isotopes to facilitate collaboration and data comparison between laboratories (2007). This protocol is based upon the previously developed methodologies used for lig ht stable isotope systems such as oxygen and nitrogen. Common practice for reporting light stable isotope ratios places the lightest isotope in the denominator. Ratios that are greater than zero and less than zero indicate samples with heavier or lighter isotopic contributions respectively. Due to a 0.15% na tural abundance of 196 Hg compared to a 10% natural abundance of 198 Hg, 198 Hg is the recommended Hg isotope to be used in the denominator for Hg analysis (Blum and Bergquist, 2007). The suggested reportable ratio for research interested in 196 Hg is 198 Hg/ 1 96 Hg.


32 Mass Bias Corrections and per mil Reporting Mass bias imparted by the method must be corrected. Two separate corrections are applied to account for fractionation imparted internally with respect to the MC ICP MS. Internal mass bias includes fluctua tions in plasma conditions and interactions within the ion beam Sufficient concentrations of non analyte species may hinder the ionization of Hg atoms or alter the impedance of species within the plasma. Heavy non analyte ions that enter the ion beam ca n dominate the beam and force lighter ions out by repulsive interactions. Mass bias within an analysis is corrected for by using a NIST thallium (Tl) isotope standard, NIST SRM 997. Thallium is the next element of higher mass after Hg and is assumed to be have similarly within the MC ICP MS. A 20 ppb Tl standard solution is prepared in 2% HNO 3 and introduced to a desolvating nebulizer system (Foucher and Hintelmann, 2006). The resulting Tl aerosol is combined with Hg gas in transit to the plasma. The 2 05 Tl/ 203 Tl ratio is corrected to its accepted value of 2.38714 and the correction factor applied to the Hg isotope data. Thallium correction is limited to instrument related mass bias fractionation. The NIST 997 standard solution is introduced to the ins trument by a different technique than that used for Hg. According to Blum and Bergquist, Tl correction is made further imperfect by the deviation in the mass bias of Tl and Hg that result s from the deviations in their mass (2007). Sample standard bracketing is adopted to account for the remainder of the correction and deviations in sample standard bracketing procedures. This correctio n manifests as part of the reporting protocol for Hg isotope concentration ratios. Mercury isotope ratios are reported 3.1 Where x = 204, 202, 201, 200, or 199


33 The ratio value used for the standard is an average of the bracketing standards analyzed before a nd after the sample. Del values are reported in per mil units, or parts per thousand, and Blum and Bergquist suggest ed the use of the elemental Hg standard NIST 3133 due to the fewer matrix effects observed compared to oth er standards available (2007). NIST 3133 comes as 9.954 ppm Hg in a 10% HNO 3 solution and is diluted to 2 5 ppb in 2 5% HNO 3 or 3% Table 3.3: Certified NIST 3133 H g Ratio Values Isotope Ratio Certified Value 204 Hg/ 198 Hg 0.68012 202 Hg/ 198 Hg 2.96141 201 Hg/ 198 Hg 1.31209 200 Hg/ 198 Hg 2.30468 199 Hg/ 198 Hg 1.68721 198 Hg/ 196 Hg 65.068 External mass bias includes matrix effects within the sample solution. According to Welz, the presence of species like silver and selenium can interfere with the reduction of Hg in the cold vapor technique discussed below (1985). External interferences i nclude species that compete for reduction and species that limit reduction by direct interaction with Hg atoms. These interferences are not corrected for by either technique discussed above and must be identified by sample screening or alternative methods Mass Independent Fractionation Reporting According to Blum and Bergquist, mass dependent fractionation should be reported by a 202 Hg (2007). This was chosen due to the larger abundance of 202 Hg compared to other eligible Hg isotopes a nd a lack of isobaric interferences. Mercury mass independent fractionation is reported using delta notation to remain consistent with the historical sulfur and oxygen mass independent fractionation reporting methods (Blum and Bergquist, 2007) Blum and Bergquist suggest ed determining mass dependent


34 fractionation (2007). They developed the calculations for Hg isotopes using the kinetic mass dependent fractionation law de 2007). These equations are referenced here under Equations 3.2, 3.3, 3.4, and 3.5. 3.2 3.3 3.4 3.5 Mass accurately de approximation can be made and Equations 3.6, 3.7, 3.8, and 3.9 used (Blum and Bergquist, 2007) 3.6 3.7 3.8 3.9 These equations take the following general form described in Equation 3.10. In this general equation y corresponds to a Hg isotope other than 202 Hg and m is a factor t hat describes the mass dependent fractionation of y Hg with respect to 202 Hg. This can be shown by assuming y Hg is equal to zero and solving for m (Equation 3.11). 3.10 3.11 This factor can also be described as a line in isotope space with a slope equal to m and intercept equal to zero. The presence of mass independent fractionation for values that fall within error of this line is questionable at best, such as with Point 1 i n Figure 3.1. Mass independent


35 fractionation is considered likely for values that fall outside error of this line, such as with Point 2 in the Figure 3.1. According to Bergquist and Blum, deviation of 204 Hg/ 198 Hg from expected mass dependent fractionati on values can indicate the extent of mass independent fraction from the nuclear volume effect, whereas for the magnetic isotope effect deviations from linearity are expected for the odd isotope ratios only: 199 Hg/ 198 Hg and 201 Hg/ 198 Hg (2007). The UV self shielding signature results in isotope ratios that reflect the individual abundance of each isotope (Lyons and Young, 2005). Isotopes with larger concentrations would preferentially remain in reactant or source materials whereas isotopes with smaller conc entrations would proceed to product. Limited Usefulness of Raw Ratios Del and delta notations are used to report data in published work for ease of reference and comparison between laboratories. Although NIST 3133 has the accepted values listed above, ana lyses of the standard will likely result in different values due to the nature of the instrumentation. The individual numerical values that are accepted for NIST 3133 are not in themselves important. It is only important that there are accepted values th at independent laboratories agree upon and correct to. y 202 Figure 3.1: Mass Dependent Fractionation in Isotope Space Point 1 Point 2


36 Raw isotope ratios are relatively useless when comparing data collected from other laboratories, or for data collected from the same laboratory at different points in time, without using standard sampl e bracketing. Data is reported to NIST 3133 under the assumption that isotopic fractionation applies equally to b oth the standard and the sample, and that instrumental drift behaves linearly between the opening and closing standard analyses. The raw ratio s do remain useful in evaluating performance characteristics of individual instruments. Large deviation in the raw ratios of NIST 3133 from accepted values is considered abnormal and requires troubleshooting to minimize mass bias. Raw ratios are also imp ortant in selecting the appropriate standard to bracket a given set of samples. Standards with ratio values near those of the sample are best due to the non linearity and imperfect mathematical modeling of mass dependent fractionation. Sample Introducti on Methods Transient Signal Method Initial forays into Hg isotope analysis began with transient signals (Figure 3.2) produced by Hg released from gold traps (Evans, Hintelmann, and Dillon, 2001). Gold traps were directly interfaced with a MC ICP MS system such that released Hg vapors were carried into the plasma with a flow of argon gas. The desolvated Tl internal standard aerosol was introduced to the Hg sample gas stream through a tee co nnection. Thermal release of Hg in this manner generates a short, transient signal as opposed to a long, steady signal. According to Krupp, the entirety of each signal is integrated and the resulting values compared between isotopes to obtain isotope rat ios (2005). Transient signal integration works for light stable isotopes such as hydrogen, carbon, oxygen, and sulfur, but is imprecise when applied to Hg. The ratio between signals is inconstant throughout analysis (Krupp, 2005). This leads to poor in ternal precision within an analysis. precise measurements because the beginning and ending extremes of the signal are lower in


37 voltage and affected more by nois e and instability (Krupp, 2005). The exact stretch for optimal integration is unclear if isotope ratios change throughout analysis. There are no consistent explanations agreed upon by investigators examining this behavior (Krupp, 2005). Cold Vapor Method The concentration of non analyte species introduced to the plasma should be minimized to avoid interferences in the plasma and ion beam. The analysis of most other isotope systems, such as iron and zinc, often require chromatography to provide matr ix free solutions. Filtering and centrifuging alone do not provide ready to analyze Hg sample solutions. Mercury gas production by reduction minimizes these post reduction interferences. Digested Hg samples in acidic solutions are not introduced to the MC ICP MS using the dry aerosol method discussed earlier. The se solutions are combined with a reducing agent to produce elemental Hg that evolves from the solution as a gas. The Tl internal standard is converted into a dry aerosol in a desolvating nebuli zer and combined with the Hg gas via a tee in transit to the plasma. The method, published by Foucher and Hintelmann, begins by digesting 0.2 0.5 g of sample in reverse aqua regia as discussed previously (2006). Tin (II) chloride (30% w/v in 10% HCl) is combined with sample solutions to reduce Hg to its ground state. All solutions are made Signal Intensity Time Figure 3.2: Example Transient Signal


38 up using ultrapure water and doubly distilled acids. The reduction employed is described in Equation 3.12. 3.12 HgCl 2(aq) + SnCl 2(aq) SnCl 4(aq) + Hg (g) Tin (II) ch loride and sample solutions are introduced simultaneously to a mixing block that tees into the main cold vapor generation vessel, Figure 3.3. The sample/reducing agent mixture is dispersed over a glass rod. The rod is frosted to increase its surface area and enhance the evolution of elemental Hg gas. Waste solution is pumped out the bottom of the main vessel while Hg gas is carried by a continuous stream of argon gas across a gas/liquid separator membrane. A dry aerosol of the NIST 997 thallium interna l standard (20 ppb Tl in 2% HNO 3 ) is generated by a desolvating nebulizer system and is introduced above the gas/liquid separator membrane. The resulting gaseous mixture of Hg, Tl, and argon carrier gas continues to the plasma of the MC ICP MS. The external reproducibility reported by Foucher and Hintelmann for the method ranges Syringe Method The cold vapor method possesses a lengthy preparation process requiring the use of hazardous che micals for sample digestion and reduction. Acidified Hg waste solutions and tin (II) chloride reducing solutions require special disposal. An alternative method requiring less in the way of preparation and waste avoids these drawbacks. Sonke, Zambardi, Hg sampling was performed using a gold trap (2008). Mercury gas is released from a gold trap heated to 500 O C and captured in a 50 mL syringe. The syringe is outfitted with a needle used to pierce the line carrying the NIST 997 internal Tl standard to the MC ICP MS, as described in Figure 3.4. A syringe pump is used to carefully introduce Hg gas to the gas stream at a rate of 10 mL/min.


39 Figure 3.3: Cold Vapor System Schematic


40 Fi gure 3.4: Syringe Injection Schematic for an in house Hg gas standard (2008). This method was developed for observing Hg isotopes in a degassing volcano, but has not been observed in additional published works. Data Acquisition Seven faraday cups are used during Hg analysis: H5, H4, H3, H2, H1, Ax, and L1. These labels correspond to their mass location with respect to one another and are each associated with a specific isotope as described in Table 3.2. Ax stands for axial, H for high mass and L for low mass A run file specific for Hg analysis is activated after a steady signal is obtained for the sample and Tl internal standard. This run file automates the collection of signal intensities, ratio calculations, and applies instrumental mass bias corrections derived from deviations in the 205 Tl/ 203 Tl ratio. Data collection begins by offsetting the magnet and recording an off peak background signal for 30 s before returning to the correct magnet setting. This background signal is subtracted from the sample si gnal during analysis. Sample signal intensities are collected continuously and reported in 10 s averaged intervals. Data collection continues for 5 minutes, allowing for 30 of these 10 s intervals to be collected. Signal intensities correspond to the nu mber of ions striking the detector and reflect the concentrations of the ions with respect to one another. The signal intensities of each faraday cup is measured simultaneous in time allowing fluctuation in the signal intensities caused by the instrument to be seen equally across all detectors.


41 Table 3.2: Detectors and Corresponding Isotopes Faraday Cup Isotope H5 205 Tl H4 203 Tl H3 202 Hg H2 201 Hg H1 200 Hg Ax 199 Hg L1 198 Hg Isotope ratios are obtained by dividing the signal intensity of one faraday cup by the signal intensity of another. For example, the ratio 202 Hg/ 198 Hg corresponds to the signal intensity of the H3 faraday cup divided by the signal intensity of the L1 fara day cup. Exact sample concentrations are unnecessary as isotope ratios are calculated directly from signal intensities. Quality Analysis Stable conditions within the instrument are expected to generate more accurate corrections. Instrument stability can be assessed by running NIST 3133 in sequence and calculating del values by self bracketing. In the case of perfect stability and no instrumental drift, to back analysis without in smaller These standard comparison values are expected to increase under standard sample bracketing protocol due to greater length of time in which the instrument can drift. Va lues on the order of smaller are ideal. Instrumental drift throughout the day is also observed by plotting 202 Hg/ 198 Hg ratios obtained for NIST 3133 analyses as they are acquired. A consistent trend in the values of one standard analysis to the next is preferred. Occasional deviations from a trend are expected and are corrected for by reanalysis of NIST 3133 before continuing with additional samples or standards. Values of samples ran in duplicate or triplicate as well as values for secondary reference materials like University of Michigan Almaden, can also be used to assess instrument stability and consistency.


42 CHAPTER IV METHOD DEVELOPMENT Purpose A review of the literature, as described in Figure 4.1, suggests external standard error varies more with larger quantities of organic material when samples are prepared by acid digestion and cold vapor production (Biswas, et al., 2008; Foucher and Hintelm ann, 2006; Jackson, et al. 2008; Carignan, et al., 2009; Stetson, et al., 2009). Coal data was prepared by combustion and solution trapping (Biswas, et al., 2008). Samples for the remaining data were prepared by acid digestion and cold vapor production ( Foucher and Hintelmann, 2006; Jackson, et al., 2008; Carignan, et al., 2009; Stetson, et al., 2009). Large variability is expected in coal samples due to large carbon content. The combustion and solution trapping preparation technique is observed to resu lt in smaller standard deviations than expected with values on the order of When Biswas, et al. employed the combustion and solution trapping method for sample analytical precision and simple reproducibility encountered during previous stud ies of coal Evans, et al. (2001) and a later adaptation of the same method by Xie, et al. (2005). Although the combustion and solution trapping technique has achieved greater precision, it does not avoid the use of hazardous chemicals or the production of waste associated with cold vapor techniques. Further advancements in the Hg isotope analytical technique need to address the analysis of biological material s and the efficiency of sample preparation and analysis. The goal of this project was to refine a method from those already developed that could accomplish three things: (1) attain high precision with high organic content samples, (2) avoid the use of haz ardous chemicals and limit waste production, and (3) minimize the time and sample amount required during preparation and handling.


43 The DMA 80 provides an automated, stream lined combustion system specific to Hg samples. Method development focused on in terfacing the DMA 80 with the MC ICP MS while avoiding problems associated with transient signal analysis. Nitrogen gas was used as a carrier within the DMA 80 in place of oxygen to avoid plasma instability. Utilization of oxygen gas was observed to exti nguish the plasma during isotope analysis. Development Direct and Indirect Interfacing Two different approaches were immediately available: direct and indirect interfacing. A direct interface approach between the DMA 80 and the MC ICP MS would be little d ifferent from transient signal methods previously discussed. Xie, et al. (2005) developed a direct 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 2SD of 202 Figure 4.1: Variability in 202 Hg Error Coal Sediment Aquatic Organism Lichen Calcine Minerals Cinnabar


44 interface method adapted from the initial transient signal work of Evans, et al. (2001). In this direct interface approach Hg gas that evolved from a gold trap was carried through a series of three 250 mL glass jars to extend and smooth the signal. A dry Tl internal standard aerosol was added to the gas stream before introduction to the plasma. Direct comparison to characteristics of modern techniques is d ifficult due to alternative reporting methods and lack of in depth data analysis. Xie, et al. claims an external reproducibility for repeated analysis of NIST 2225, a high 201 Hg/ 202 Hg and 199 Hg/ 202 Hg respectiv ely (2005). Method characteristics for more complex standards are described in Table 4.1Xie, et al. (2005). Table 4.1: Approximate Error Characteristics of a Direct Interface 1 Standard 200 Hg/ 202 199 Hg/ 202 Material DORM 2 0.4 1 Dogfish muscle CRM 422 0.325 0.5 Cod muscle DOLT 3 0.1 0.45 Dogfish liver NIST 2682b 0.175 0.2 Sub bituminous coal NIST 2709 0.075 0.225 Soil 1 Data obtained from Xie, et al., 2005 Higher precision was reported for Hg analysis in coal by the combustion and solution trapping technique (Biswas, et al. 2008). Any advancement in direct interfacing methods would require building upon but not directly duplicating this technique. Mercury gas could be temporarily collected in a storage vessel rather than introducing Hg gas to the plasma as it evolves from the gold trap. This temporary storage vessel would serve a similar purpose as the syringe in syringe inj ection method and would allow Hg isotopes to briefly mix before analysis (2008) A conceptual representation of such a direct interface set up is described in the Figure 4.2. Gas flow released from the DMA 80 could be guided by use of manual valves to a temporary storage vessel or waste ventilation. The temporary storage vessel is sealed off after the collection of Hg gases for a prescribed amount of time to allow for complete isotope mixing. The Hg gas can be carried downstream by a flow of argon, com bined with a dry aerosol of the Tl internal standard, and introduced to the plasma.


45 Figure 4.2: Conceptual Diagram of a Direct Interface Approach Method development did not proceed by a direct interface approach due to the bulk of the DMA 80 and limited space. Indirect interfacing necessitates transportation of Hg gas as with focused on three interrelated features: (1) collection, (2) transportation, and (3) sampl e introduction of Hg gas. Sample Transport and Interfacing Conceptual Approach Method development proceeded by splitting the interface into three segments: (1) a direct interface with the DMA 80, (2) a direct interface with the MC ICP MS, and (3) a Hg gas transport vessel. The Hg gas transport vessel must be capable of interacting with each direct interface for ease of gas collection and analysis The DMA 80 direct interface setup split gas flow in two directions by use of a tee post detector. Manu al valves were used to direct the gas flow in either of the two directions. One branch guided gas flow to a carbon trap and ventilation, and the other branch guided gas flow to the Hg gas transport vessel.


46 Two MC ICP MS direct interfaces were assembled to meet the specific needs of the particular Hg transport vessel used. In both cases the interface included a series of tee joints to combine Hg gas, a dry aerosol of the Tl internal standard, and an additional flow of argon carrier gas. The first interface utilized the cold vapor apparatus as conceptually described in Figure 4.3. The introduction of different gases mimicked the same setup as for normal cold vapor generation, but closed the waste line to minimize sample loss. The second interface, as descr ibed in Figure 4.4, combined Hg gas and the dry aerosol of the Tl internal standard directly before joining the additional argon gas in transit to the plasma. Figure 4.3: Diagram of an Interface with the MC ICP MS Using the CV Apparatus


47 Figure 4.4: Diagram of a Direct Gas Interface with the MC ICP MS Syringe Initial Hg gas transport vessels utilized were disposable plastic syringes outfitted with manual, gas The pres sure from the DMA 80 gas flow was insufficient to move the plunger of the se syringe s Manual assistance was required to overcome the friction between the plunger and barrel of the syringe. High quality 100 mL gas tight glass syringes with built in valves were obtained to address this inherent lack of technique consistency. Instrument gas flow was sufficient to smoothly displace the plunger for these syringes. Employment of a syringe pump allowed for consistent sample introduction. The small bulk, smoot h transitions, and gas tight properties of these high quality syringes were sufficient for vapor transportation. Two problems became evident over repeated evaluation: (1) ratio changes within a syringe over time and (2) insufficient volume. Standard solut ion in the amount of 0.3 mL of 1 ppm Hg from a NIST 3133 stock solution was prepared by application of the DMA 80 and collected with a 100 mL glass syringe. Multiple data collections from a single syringe were possible at this concentration syringe volum e, and syringe pump rate of 1.5 mL/min. Changes observed in the 202 Hg/ 198 Hg ratio, as described in


48 Figure 4.5, may be attributable to instrumental drift, interactions with the walls of the syringe, or an as yet unidentified fractionation source inherent to the technique employed at the time. Further investigations were performed later. The limited size of the 100 mL glass syringe was insufficient for complete Hg gas collection. Calculations indicated the maximum capacity of the syringe was attained before the entirety of Hg gas from the sample could be collected. A minimum flow of 200 mL/min of carrier gas is required for the operation of the DMA 80, and Hg gas collection over the full 60 s detection period is necessary to insure complete collection Another syringe was used to capture the remaining gas flow after the first syringe reached full capacity to verify these requirements. An average of 10 analyses from the first syringe resulted in a 202 Hg/ 198 Hg ratio of 2.9753 at a total Hg voltage of 2 .97. The second syringe possessed one third the Hg concentration from the first syringe. The syringe pump rate was increased to 5.5 mL/min to compensate. An average of the 2 available analyses from the second syringe resulted in a 202 Hg/ 198 Hg ratio of 2 .9694. These results indicated inadequate capacity for Hg gas collection and a fractionation mechanism within the DMA 80 sample preparation process. Lighter isotopes are expected to release preferentially from the amalgamator due to smaller energy require ments for breaking Au Hg interactions. These expectations were not observed. The first syringe possessed a heavier 2.9748 2.9750 2.9752 2.9754 2.9756 2.9758 2.9760 2.9762 0 1 2 3 4 5 6 7 8 9 10 202 Hg/ 198 Hg Analysis Figure 4.5: Ratio Changes Within a Syringe Over Time


49 isotopic ratio than the second syringe in this two syringe experiment. This may indicate a layering effect where lighter isotopes, prefere ntially released from samples prior to heavier isotopes, are trapped against the gold surface until later in the release. Glass Tube Multiple 250 mL glass sampling tubes were acquired to attain complete collection of Hg gas. These sampling tubes were out fitted with entry and exit points possessing gas tight valves and a centrally fixed septum. The structure of the tubes introduced an additional complication requiring resolution: absence of a built in mechanism to control storage space. The inability to attain a zero volume storage space required a method for adequate flushing of Hg gas and additional calculations to assess storage capacity. The tube was opened at both ends to flush t he tube of lingering sample gas es and minimize oxygen concentrations pr ior to sample preparation in the DMA 80, as described in Figure 4.6. For analysis, the second source of argon gas was the remainder of the MC ICP MS direct interface. Figure 4.6: Glass Tu be Gas Flow Schematic Storage capacity of Hg gas was further limited by the inability of the storage space to reach a zero volume. The storage space had to be shared between Hg sample gas and the nitrogen carrier gas at atmospheric pressure used to flush the tube. The 250 mL volume of a single tube appears sufficient by calculations, but a two tube combined volume of 500 mL was tested as well. Figure 4.7 shows a series of analyses from a single preparation and collection of 0.3 mL of 1 ppm Hg from a NIST 3133 stock solution. Mercury signals faded over time instead of the abrupt signal termination observed with syringe use. Adjustments were made on the Aridus and


50 the mass flow controller for the additional argon gas between analyses 8 and 9, and again bet ween 12 and 13. Temporary Storage and Syringe Transfer A third indirect interface setup was developed by combining a syringe transport vessel with a temporary storage tank, see Figure 4.8. Temporary storage tanks, described in Figure 4.9, for Hg gas wer e assembled by outfitting 1000 mL vacuum flasks with large rubber stoppers and manual valves. The shape and structure of the assemblies required flushing in the same manner outlined for the glass tubes. Mercury gas stored within a tank was manually sampl ed using the 100 mL glass syringes. Filled syringes were transported to the MC ICP MS direct interface and analyzed by the same methods outlined for the syringe method described earlier. This setup offered two advantages: (1) complete capture and mixing of sample isotope concentrations by use of a large volume storage tank and (2) succinct start and end times of isotope analysis by use of a syringe transport vessel. Additional experimentation was performed to examine the interactions between Hg gases and the syringe and vacuum flask. Preparing 0.3 mL of 1 ppm Hg from a NIST 3133 stock solution through the DMA 80 provided sufficient pressures and concentrations within the 0.00 0.50 1.00 1.50 2.00 2.50 3.00 1 2 3 4 5 6 7 8 9 10 11 12 13 Hg Total (V) Analysis Figure 4.7: Hg Total Voltage Over Time


51 temporary storage tank for 6 analyses. Each analysis required approximately 50 mL of gas at a syringe pump rate of 6.0 mL/min. Figure 4.8: Temporary Storage and Syringe Transport Five flasks were analyzed in succession throughout a single day to assess interactions between Hg gas and syringes. Flasks were analyzed in sets of three syringes. Each syringe held 100 mL of gas and allowed for two analyses each. Figure 4.9: Tempora ry Storage Tank The general trend observed is described in Figure 4.10 produced from the fourth vacuum flask syringe data. The 202 Hg/ 198 Hg ratio decreases between analyses from a single syringe. An anomalous result was observed for the second syringe of the second flask where the 202 Hg/ 198 Hg


52 was observed to increase. Heavier Hg isotopes are likely to preferentially hold on to the sides of the syringe barrel if interactions with the syringe barrel occur during extended analysis. This explanation provides for the general trend but does not explain the anomalous observation. Ratio shifts between syringes were inconsistent throughout this experiment. The method was altered such that only 50 mL of gas was pulled into a syringe at a time to avoid potential Hg gas syringe interactions. Figure 4.11 more aptly describes the behavior of Hg gases within a vacuum flask over time. The data shows the 202 Hg/ 198 Hg ratios observed by repeat analysis of the second vacuum flask prepared each day for three days. Most o f these analyses were spaced apart due to sample standard bracketing. The lack of a consistent pattern between flasks suggests little to no interaction between Hg gas and the walls of the vacuum flask used as a temporary storage tank. 2.9674 2.9675 2.9676 2.9677 2.9678 2.9679 2.9680 2.9681 2.9682 2.9683 1 2 202 Hg/ 198 Hg Analysis Figure 4.10: Hg Gas Syringe Interactions Syringe 1 Syringe 2 Syringe 3


53 2.9676 2.9678 2.9680 2.9682 2.9684 2.9686 2.9688 2.9690 2.9692 2.9694 1 2 3 4 5 6 202 Hg/ 198 Hg Analysis Figure 4.11: Hg Gas Flask Interactions 8/28/2013 8/29/2013 8/30/2013


54 CHAPTER V EXPERIMENTAL METHODS Three standard reference materials were prepared by bench top acid digestion, microwave assisted acid digestion, and combustion through the DMA 80. Acid digestion and combustion preparations were analyzed for Hg isotope ratios using c old vapor and direct gas introduction techniques respectively. These materials were independently prepared and analyzed a minimum of 8 times per preparation technique. These three standards were selected to provide a variety of sample types and are descr ibed in Table 5.1. Four additional reference materials were also analyzed by the combustion method for exploratory purposes: GXR 1 (jasperoid), GXR 2 (soil), NIST 1944 (New York/New Jersey waterway sediment), and NIST 1946 (Lake Superior fish tissue). Th ese additional reference materials were only prepared and analyzed once or twice. Table 5.1: Standard Materials Standard Composition [Hg] (ppm) 1 NIST SRM 2711 Montana soil 6.25 SARM 20 Coal (Sasolburg) 0.25 TORT 1 Lobster hepatopancreas 0.128 1 [Hg] for NIST 2711 and SARM 20 are certified [Hg] TOT values; TORT 1 [Hg] is the certified value for [CH 3 Hg] Sample Preparation Acid Digestions Both a cid digestion techniques used the same reagents and post digestion preparation process. Reagent grade concen trated HCl and concentrated HNO 3 were doubly distilled prior to use; ultra high purity 30 32% H 2 O 2 was used as is. Ultrapure DI H 2 O was obtained by running house distributed DI H 2 O through a MILLI Q water purification system until the water conductivity r ead 18 M Digested sample solutions were retained in pre cleaned 50 mL polypropylene conical tubes. These tubes were cleaned in two steps: (1) vigorous shaking with diluted MICRO 90 solution followed by (2) an interior wash with concentrated HNO 3 Tubes were rinsed a minimum of 5 times after each step with ultrapure DI H 2 O.


55 solutions retaining particulate matter were centrifuged at 3500 rpm for 5 minutes before addi tional filtering of the supernatant. Sample solutions were then diluted with ultrapure DI H 2 O to an ideal concentration between 2 3 ppb Hg, or as close a concentration as possible. Bench Top Digestion method (2006). Concentrated HCl in the amount of 1 mL was added to 0.200 g of sample in a 250 mL borosilicate Erlenmeyer flask. Concentrated HNO 3 and 30% H 2 O 2 were added in the amounts of 9 mL and 2 mL respectively. The solution was swirled between reagent additions to promote mixing and reaction. After initial reactions, solutions were heated to 80 O C for 4 to 6 hours in a fume hood. Digested solutions were allowed to co ol before filtration and dilution as described above. Microwave Assisted Digestion Concentrated HCl was added in the amount of 1 mL to approximately 0.200 g of sample in 50 mL PFA microwave vessels. Concentrated HNO 3 was added in the amount of 8 mL follow ed by 2 mL of H 2 O 2 in 0.5 mL increments. Hydrogen peroxide was added stepwise to avoid sample loss during vigorous reaction. Solutions were swirled between reagent additions to promote mixing and reaction. After initial reactions, digestion vessels wer e capped and placed in a Multiwave 3000 microwave apparatus. Samples were digested with microwave assistance under the settings described in Table 5.2. Digested samples were allowed to cool overnight before pressurized headspace gases were carefully vent ed by slow removal of digestion vessel caps in a fume hood. Digested solutions were then filtered and diluted as described above. Table 5.2: Microwave Digestion Instrument Parameters Phase Power (W) Ramp (min) Hold (min) Fan 1 1400 20:00 30:00 1 2 0 0 15:00 3


56 Combustion The temporary storage and syringe transfer method described in Chapter 4 was used for combustion preparation. The conceptual schematic for this method is described by Figure 4.8. Approximate masses used for the three primary standard reference materials are described in Table 5.3. Due to lower Hg concentrations in SARM 20 and TORT 1, Hg was concentrated on the gold trap over the course of two combustions before release and collection. Table 5.3: Sample Mass for Combustion Preparatio n Standard Total Mass (g) NIST 2711 0.05 SARM 20 0.825 TORT 1 0.30 Sample gases were guided from a direct mercury analyzer apparatus, DMA 80, using a combination of PVC and silicone tubing, plastic tees, and plastic one way stopcocks. Borosilicate vacuum flasks (1000 mL) and large rubber stoppers were also used in the co nstruction of temporary storage tanks. All connections with tubing were tightened with plastic zip ties. The carrier gas flow was set to the instrument standard of 9.9 L/hr. Two carrier gases were used, laboratory grade N 2 and ultra high purity O 2 Te mporary storage tanks were flushed with a flow of N 2 from the DMA 80 before and between sample collections for a minimum of 20 min. Two temporary storage vessels were used throughout the experiment: one was used only for NIST 3133 and the other for the re maining standard materials. The three primary standard materials were analyzed on separate days to allow for additional flush times to minimize the potential for cross contamination. Gas flow exiting the DMA 80 was manually controlled using a system of on e way stopcocks. Table 5.4 describes the schedule of events associated with the three stop cocks during sample preparation. The maximum starting temperature of the sample furnace was set to 250 O C. Additional parameters of the combustion method are desc ribed in Table 5.5.


57 Combustion schedules and carrier gases differed based on the sample capacity of the DMA 80, the Hg concentration of the standard material, and additional complications described later. The NIST 2711 and NIST 3133 standards were each co mbusted in a single pass of the DMA 80. The SARM 20 and TORT 1 standards required additional passes as described in Tables 5.6 and 5.7 respectively. During TORT 1 preparation, initial carrier gas was set to nitrogen to allow tank flushing. Prior to star ting sample combustion the temporary storage vessel was sealed off and the carrier gas was switched to oxygen The remainder of the schedule proceeded as described in Table 5.7. Table 5.4: Schedule of Gas Flow Events Step Event Stopcock 1 2 3 1 Tank Flush Before Combustion Closed Open Open 1 2 Prepare for Combustion Open Closed Closed 3 During Combustion Open Closed Closed 4 Hg Signal Recording Begins Closed Open Closed 5 Hg Signal Recording Ends Open Closed Closed 6 Hg Gas Sampling Open Closed Open 2 1 3 open and gas flow directed to fume hood ventilation 2 3 open to syringe Table 5.5: DMA 80 Combustion Method Process Step Time (sec) Temperature Range ( O C) 1 1 2 1 3 1 4 1 Drying 1 10 25 200 2 615 170 125 2 60 200 615 170 125 Decomposition 3 90 200 650 2 615 170 125 4 90 650 615 170 125 Purge 5 60 --3 615 170 125 Hg Release 6 20 --3 615 170 900 2 125 Signal Recording 7 60 --3 615 --3 125 1 Furnaces correspond to the following DMA 80 components: 1 Sample Housing of the Catalyst Tube, 2 Catalyst Housing of the Catalyst Tube, 3 Amalgamator, and 4 Cuvette 2 A temperature range indicates a ramp period where temperatures increase over the allotted period of time. 3 --to drop back to room temperature or a minimum base line Table 5.6: SARM 20 Combustion Preparation Schedule Pass Sample Carrier Gas Description 1 SARM 20 Nitrogen Hg Concentration 2 SARM 20 Nitrogen Hg Collection


58 Table 5.7: TORT 1 Combustion Preparation Schedule Pass Sample Carrier Gas Description 1 TORT 1 Oxygen Hg Concentration 2 Blank Oxygen System Flush 3 TORT 1 Oxygen Hg Concentration 4 Blank Oxygen System Flush 5 Blank Nitrogen Hg Collection Isotope Analysis Mercury isotope ratios were obtained for standard reference materials by analyzing preparations on a Nu Plasma, a double focusing magnetic sector MC ICP MS. Isotopes were directed towards individual detectors as described in Table 3.2. Signal intensities from faraday cups were gathered simultaneously in thirty 10 second averaged intervals. Instrument settings were tuned for peak shape and intensity at the beginning of each day before analyses of pre pared materials by using NIST 3133. Typical instrument settings used during analysis are described in Table 5.8. Table 5.8: MC ICP MS Instrument Settings Quads High Voltage Source Quad Voltage (V) Unit Voltage (V) Unit Voltage (V) Q1 35.1 HV1 3995 SV1 42.0 Q2 78.7 HV2 3129 SH1 24.0 HV3 2567 SH2 12.0 HV4 2043 TH1 11.0 HV5 1700 TV1 387.0 HV6 1744 TV2 88.0 Sample standard bracketing was followed for all analysis techniques using NIST 3133. For the cold vapor introduction technique described in Chapter 3, NIST 3133 solutions of 2 5 ppb Hg in 2% HNO 3 were prepared from a 1 ppm Hg in 3% HCl stock solution. For the combustion technique described in Chapter 4, 0.3 mL of 1 ppm Hg in 3% HCl was passed through the DMA 80 and coll ected in the NIST 3133 designated storage vessel. A single combustion provided sufficient gas for 6 analyses.


59 A dry aerosol of the NIST 997 Tl internal standard was prepared by passing 20 ppb Tl in 2% HNO 3 solution through an Aridus II TM desolvating nebu lizer system outfitted with a MicroMist 0.1 00 mL/min nebulizer. The Aridus II TM was set up to run argon for the sweep gas (approximately 5 7 L/min) and through the nitrogen line (1.0 mL/min). Cold Vapor Analysis Samples prepared by bench top and microwave assisted acid digestion were introduced to the plasma via cold vapor generation (Foucher and Hintelmann, 2006). A tin(II) chloride reducing solution was prepared by bringing 30 g of reagent grade SnCl 2 and 100 m L of doubly distilled concentrated HCl up to 1000 mL with ultrapure DI H 2 O. Solutions were prepared as needed, kept up to one week, and stored at 35 O F when not in use. by a peristaltic pump at a rate of 10 rpm. These solutions were combined in a mixing tee before introduced to an HGX 200 advanced membrane cold vapor and hydride generator, described in Chapter 3. The flow of argon carrier gas was controlled by a precision gas flow controller. settings. The Tl internal standard was introduced from the Aridus II TM into the cold vapor system after the gas liquid separator membrane. The co ld vapor system and sample introduction tubing was rinsed between analyses with ultrapure DI H 2 O followed by 3% HCl until background levels were achieved. The Aridus II TM was rinsed with ultrapure DI H 2 O followed by 2% HNO 3 Direct Gas Analysis Samples p repared by combustion through the DMA 80 were introduced to the plasma by the direct gas interface approach described in Chapter 4, see Figure 4.4. This direct interface was constructed from the same materials utilized for the DMA 80 interface. An NE 100 0 Multi Phaser TM programmable syringe pump was used to introduce the sample gas at a steady rate of 6.0 mL/min. Introduction of additional argon was controlled by a precision gas flow controller.


60 Sample gas (50 mL) was sampled from temporary storage vesse ls using 100 mL borosilicate glass syringes with PTFE tipped plunger s Two syringes were used: one for NIST 3133 and the other for the remaining standard reference materials. Each combustion preparation of NIST 3133 was utilized for 1 6 analyses. Each c ombustion preparation of NIST 2711, SARM 20, and TORT 1 was utilized for a single analysis each. Results Complications SARM 20 Microwave Samples Mercury signals could not be attained during the isotopic analysis of SARM 20 prepared by microwave assisted acid digestion. According to Welz, major interferences with Hg cold vapor production include silver, iodine, and selenium and o ther interferences include arsenic, bismuth, copper, and antimony (1985). To investigate the presence of possible interferences bench top and microwave assisted acid digestion sample solutions were analyzed by ICP MS. The concentrations of a number of elements w ere collected, though none of the certified literature values, or the values obtained by ICP MS, were sufficient to int erfere with Hg reduction. These observations are described in Table 5.9. Though not included in the list of interferences, aluminum concentrations have been added because this species displayed the greatest deviation between solutions. Aluminum was 17 times more concentrated in the microwave digestion sample than the bench top digestion sample. Aluminum does form an ama lgam with Hg (Chesworth, 1971). The literature value for aluminum content of 2711 is listed as 6.53% by mass, yet no discernable probl ems were encountered with the microwave assisted acid digestion of this standard. The extent this reaction may interfere with analysis is uncertain and would require additional experimentation. It is possible Hg entered the pressurized headspace of the microwave vessels during digestion. Mercury gas that did not re solvate by multiple inversions could have escaped when the headspace was vented into a fume hood. This possibility seems unlikely c onsidering


61 microwave assisted digestions of 2711 and TORT 1 analyzed without complication. This possibility could be tested by running the digested sample solutions through the combustion process and noting the resulting intensity of Hg signals. However, the low pH of these solutions would likely compromise the integrity of the DMA Table 5.9: Hg Cold Vapor Interference Analysis Species Interference Requirements (Weight %) Literature Values (ppm) Observed Values (ppb) Bench Top Microwave Al 1 ----2 24300 434000 Ag 0.005 --< 20 < 20 As 10 4.7 < 1000 < 1000 Bi 10 --< 100 < 100 Cu 10 18 709 < 500 I 0.003 ------Sb 1 0.4 3 ----Se 0.0005 0.8 < 500 < 500 1 Al was not reported as a potential interference by Welz, but was included here due to the observed differences between preparation methods (1985). 2 The certified Weight % of Al 2 O 3 is 11.27. 3 [Sb] was reported in the literature as a reference value only. Acid digestion solutions of SARM 20 were qualitatively compared post filtration and dilution. Microwave assisted preparations possessed a darker, more opaque brown color than bench top preparations. It may be possible that microwave preparation of coal matter allows the formation of materials that bind and retain Hg even in the presence of a reducing agent like SnCl 2 Further exploration of this possibility would require additional research, analysis, and repeated preparations. Combustion Samples Combustion preparations of TORT 1 were complicated by low Hg concentration and high organic biological materials. An unidentified slimy brown residue with a foul smell was observed to coat the inside of the DMA ing when only N 2 was used as a carrier gas. The Hg signal on the DMA 80 for these samples broadened and displayed difficulties returning to background levels during the 60 s signal collection time frame.


62 These observations called into question the comple te collection of sample Hg due to incomplete release within the DMA 80 and/or leaks by the greased joints in the direct interface. After repeated attempts and analyses, the system was dismantled and cleaned, a new temporary vessel constructed, and oxygen introduced as a carrier gas. The residue was not observed after oxygen schedule. Due to troubleshooting and lengthier preparation time, only five samples were analyzed for TORT 1 combus tion preparation. The lower Hg concentration of SARM 20 and TORT 1 presented difficulty in attaining a Hg total signal comparable to the typical Hg total signal obtained for the bracketing standard, NIST 3133. It was necessary to concentrate Hg on the ama lgamator over the course of multiple sample combustions. Sample analysis proceeded without voltage matching. Data Analysis Calculations Del and delta values were calculated using equations 3.1 and 3.8 as suggested for Hg isotope reporting protocol from Bl um and Bergquist (2007). Calculated values were averaged and the standard deviation obtained. Quality analysis generally proceeded as described in Chapter 3, allowing for NIST 3133 del bracketing values as large as 0.20 Values falling outside of this protocol were handled on a case by case basis. No observations were excluded as outliers. Initial analyses of TORT 1 prepared by combustion were omitted due to the complications described above. The first day of combust ion analysis was also omitted due to indications of decreased stability, primarily large fluctuations in 2711 and NIST 3133 values. A full propagation of error analysis has not been included in the reported values. Observations A summary of results are l isted in Table 5.10. A complete collection of raw data can be found in Appendix A Published values for standards have been included; values for similar standard materials were included where matching standards were unavailable.


63 Table 5.10: Summary Exper imental Data Standard Meth 1 n Tot Hg(V) SD(V) 2 202 2 2 201 2 2 271 1 b 8 2.376 0.270 0.199 0.095 0.148 0.023 271 1 m 8 2.196 0.222 0.163 0.091 0.158 0.043 2711 c 8 1.168 0.128 0.143 0.173 0.139 0.077 2711 3 lit 7 ----0.150 0.159 0.190 0.026 SARM 20 b 8 2.263 1.603 1.025 0.066 0.217 0.124 SAR M 20 c 8 1.139 0.507 0.705 0.485 0.244 0.068 TORT 1 b 8 1.457 0.139 0.815 0.174 0.350 0.035 TOR T 1 m 8 0.610 0.474 0.530 0.221 0.306 0.188 TORT 1 c 6 0.288 0.048 0.639 0.315 0.399 0.273 TORT 2 5 lit 3 ----0.330 0.080 0.670 0.060 GXR 1 c 1 0.919 --0.114 --0.116 --GXR 1 4 lit 3 ----0.260 0.130 0.070 0.035 GXR 2 c 1 1.085 --0.083 --0.042 --GXR 2 4 lit 2 ----0.110 0.106 0.010 0.028 NIS T 1944 c 2 1.153 0.181 0.120 0.070 0.040 0.174 NIST 1944 3 lit 10 ----0.420 0.221 0.020 0.032 NIS T 1946 c 1 1.229 --0.328 --1.905 --NIST 1947 5 lit 6 ----1.290 0.055 4.180 0.070 1 Method legend= b bench, m microwave, c combustion, lit literature 2 Values do not include a propagation of error analysis 3 Biswas, et al. 2008 (Combustion, solution trapping, and cold vapor introduction) 4 Estrade, et al. 2009 (Acid digestion and cold vapor introduction) 5 Masbou, Point, and Sonke 2013 (CH 3 Hg extraction technique and cold vapor introduction) The 202 201 Hg values are similar across all three sample preparation methods for 2711, including agreement with the published value. Standard error was larger for the combustion method, but comparable to the error reported in the literature. SARM 20 presents a special case due to the complications described previously and fractionation obse rved for combustion prepared materials. Three of the 8 combustion replicates were analyzed at a total Hg voltage greater than or equal to the voltage for the NIST 3133 bracketing standard, 1.2 1.5 V. The remaining 5 replicates were analyzed at a total Hg voltage 0.66 202 Hg value from the low voltage preparations agree more closely with the bench top digestion method. No observations could prescribe exclusion of the high voltage v alues.


64 Table 5.11: SARM 20 Combustion Fractionation Analysis Category n Tot Hg(V) SD(V) 1 202 1 1 201 1 1 Low Volt 5 0.832 0.115 1.025 0.067 0.251 0.087 High Volt 3 1.651 0.495 0.173 0.367 0.232 0.025 All 8 1.139 0.507 0.705 0.485 0.244 0.068 1 Values do not include a propagation of error analysis Standard deviations and errors for TORT 1 analyses were larger for all three preparation methods than for 2711 analyses. The combustion preparation technique possessed the highest variability, but also a central 202 Hg value compared to bench top and micr owave assisted acid 201 Hg values are similar to one another across all preparation methods. The standard deviation for TORT 2 listed in Table 5.10 is decidedly smaller than those observed during this experiment, but the preparation report ed was by an extraction technique for CH 3 Hg followed by cold vapor generation. The value from the literature was only included for general comparison. The additional reference materials were analyzed for exploratory purposes only. GXR 1, GXR 2, and NIST 1944 appear similar enough to literature values to warrant little discussion without the collection of additional observations from repeated analyses. The deviation of 202 Hg for NIST 1944 was unexpectedly high, but may be partially explained by the large r error associated with the literature value. Analysis of NIST 1946 proved difficult from a preparation stand point. The material is a frozen fish tissue that thaws quickly and is difficult to manipulate with warmer utensils. Deviations in calculated va lues from literature values were large, but not unexpected considering variation in source material (Lake Superior v ersus Lake Michigan fish tissue) and sample preparation (combustion vs CH 3 Hg extraction). Discussion Precision 202 Hg values were highest for the experimental combustion preparation method as compared to bench top and microwave assisted acid digestion methods.


65 The experimental method in its current incarnation does not appear to provide a more precise analysis than the acid digestion methods. Observed imprecision is likely inherent in the experimental method utilized and would require additional experimentation to isolate and explore. The underlying assumption that combustion preparation will improve analysis of high organic samples can still be examined despite the imprecision inherent to the experimental technique. In Table 5.12, standard deviations are listed for each preparation method for 2711 and TORT 1. These standard deviations are divided by one another to obtain a high organic to low organic ratio. If the combustion preparation method did improve analysis of high organic samples the ratio should be lower than for the other methods. Though the combustion method does not appear to improve analysis, it d oes not appear to be worse either. Table 5.12: High Organics and Combustion Standard 202 1 b m c 2711 0.095 0.091 0.173 TORT 1 0.174 0.221 0.315 TORT 1 / 2711 1.828 2.440 1.824 1 b bench, m microwave, c combustion The larger imprecision of the combustion preparation and direct gas introduction method may be addressable if its source can be identified. Compared to the higher voltage subset of SARM 202 Hg and associated standard deviation value s closer to the observed values for bench top acid digestion preparation. The lower voltage subset also possessed a tighter spread of total Hg voltages. It might be expected that larger 2 02 Hg as different quantities of analyte might behave differently in the plasma or ion beam; however, this is inconsistent with experimental observations, see Figure 5.1.


66 Initial identification of the source of imprecision may be possible by examining 31 33 bracketed against itself for both introduction methods: cold vapor and direct gas The standard solution for 3133 consists only of Hg analyte in a 3% HCl solution and is devoid of other matrix elements that may affect combustion preparation or plasma i onization. Deviation in precision between introduction methods is likely a consequence of the introduction method rather than preparation or instrumental analysis. 202 Hg for cold vapor was 0.188 202 Hg of 0.08 202 Hg for direct gas was 0.372 202 Hg of 0.105 (SD=0.090 ). The increased average 202 Hg value and respective standard deviation for direct gas introduction suggests that this introduction meth od may be the primary culprit of the increased imprecision and should be further refined. It is important to note that samples with low Hg concentrations, such as SARM 20 and TORT 1, must be handled differently than with 2711 or NIST 3133. Mercury must be concentrated on the amalgamator before release and increased sample loads are used in individual sample boats. The number of Hg concentration steps was not altered between high and low voltage subsets of SARM 20 and cannot explain the observed deviations The sample loading 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.000 0.250 0.500 0.750 1.000 1.250 1.500 1.750 202 Hg Total SD (V) Figure 5.1: Relationship Between Tot Hg Fluctuations and Precision Bench Top Microwave Combustion


67 technique may provide more adequate explanation. There may be a threshold past which sample mass and volume can start to interfere with complete Hg release from the sample. Fractionation from such an interaction is expected to leave heavier isotopes trapped in non combusted materials. This may explain the observations with SARM 20 and provide a source for some of the error observed in TORT 1, but the high voltage subset of SARM 20 analyses possesses a 202 Hg. Exploration of this possibility would require additional experimentation. 202 Hg value, is more difficult to assess. Values are similar between methods for 2711 analyses, but for SARM 20 and TORT 1 the differences are larger. These larger differences may or may not be due to the result of complications in the preparation process from higher organic content. For 2711 and 1944, non certified values are reported on the certificate for the mass fraction of carbon and total organic carbon content respectively. These values indicate 1944 ( 4.4% C by mass ) should possess nearly twice as much carbon content as that present in 2711 ( 2% C by mass ) though these values were listed as attained by differen t methods. The deviation of 202 Hg in 1944 from the reported literature value is larger than for 2711. 201 Hg values are similar across all methods for 2711, SARM 20, and TORT 1. Any fractionation inherent within the combus tion method is likely to be mass dependent, or similar to any fractionation imparted by the other methods. Waste and P reparation M inimization The experimental method did result in improvement over the bench top and microwave assisted acid digestion methods by minimizing the use of ha zardous chemicals, the production of hazardous waste materials, and the required time to prepare and analyze sample materials. H azardous chemicals were not utilized for the combustion method, and acidic waste products fo r sample digestion and reduction were not produced. The s ample mass required for analysis can be further minimized for the combustion method by utilizing a smaller temporary storage vessel. A vessel with a volume of 250 mL can


68 support liquid samples of 105 ppb Hg with a single pass through the DMA 80. For solid samples with densities of 1 g/mL or greater, 105 ppb Hg solids can also be prepared with a single pass. Required volumes and sample concentrations and masses might be further minimized by refini ng a precise collection period smaller than the current 60 s window. The ability to concentrate Hg on the amalgamator before release can also reduce required concentrations, but would increase the required amount of sample needed. This is in contrast to the acid digestion methods that require 0.2 g of 500 ppb Hg samples to attain 50 mL of 2 ppb Hg solution for analysis. Preparation time also favored the combustion method. All else being approximately equal, combustion preparation only requires the time n ecessary to flush the temporary storage vessel and the required number of passes through the DMA 80 to attain an adequate amount of Hg. This is in contrast to the time required for digestion, filtering, centrifuging, dilution, and long washout periods dur ing analysis. Though the meet precision goals requires further refinement on its development Utilization of the experimental method in its current state is likely to be minimal until its precision can be improved and the accuracy shown to be equal or superior to alternative methods. Application Despite complications with the analysis of SARM 20 prepared by microwave assisted acid digestion, both acid digestion methods with cold vapor introduction appear interchangeable. Additional standard preparations are necessary to produce a more robust comparison, but based on this assessment there is nothing to be gained by the greater time and material requirements of microwave assisted digestion. Both methods exhibited similar precision characteristics, though 202 Hg for TORT 1 deviated greatly. If the data for TORT 1 can be shown to be more accurate for samples prepared by microwave assisted digestion then this method may prove useful for samples with biological and/or high organic matrices.


69 The experimental method developed can be used as is without additional improvements for certain applications despite its higher levels of imprecision. The improved preparation ti me, analysis speed, and lack of hazardous chemicals allows the combustion and direct gas introduction method to be superior to the other methods examined for sample screening. An alternative method can later be used to examine individual samples of partic ular interest after initial screening analyses have been performed. The experimental method is also likely to be preferred when examining systems where del values between samples are large. In th e s e case s the increased imprecision does not inhibit interp retation of data and the same conclusions would be drawn had an acid digestion and cold vapor introduction method been utilized instead. Future Work Immediate future work should address the imprecision observed in the experimental method. Though initial i nterpretation of data suggests the direct gas introduction technique as the primary culprit of increased imprecision, further experimentation should be performed to explore this conclusion. This will require the method to be broken down where possible and the individual pieces explored for their overall contribution. An initial break down will separate the combustion process from direct gas introduction to the MC ICP MS. Mercury gas released from the DMA 80 can be trapped in an oxidizing solution of 1% KMnO 4 and 1.8 M H 2 SO 4 as utilized by Biswas and Blum (2008). Resulting solutions can be analyzed on the MC ICP MS by the cold vapor technique. Resulting observations can be directly contrasted to the acid digestion preparation methods as the only thing t hat differed was preparation technique, not introduction into the instrument. If the same level of imprecision is observed as in this experiment, additional steps may be taken to address the source of the error within the combustion technique. If the pre cision is observed to improve, additional steps can instead be taken to address error within the direct gas introduction technique. This experiment would also allow an assessment of fractionation inherent in the various methods. If del values appear simil ar for acid trapped solutions and acid digestions, the direct gas


70 202 Hg differences observed in this experiment. If the del values are differe nt, the combustion technique may be the culprit. Additional experimentation should also include the analysis of incomplete combustion as mentioned in the precision discussion above. The amount of sample loaded into a single sample boat can be varied, as well as the number of passes for collecting Hg on the amalgamator before 202 Hg and its associated standard deviation can be assessed for impact. By conducting the two major experiments discussed above, it will be easier to fully understand the mechanisms of the experime ntal method being developed. The resulting observations are expected to reveal sufficient information to further refine the method, suggest additional experiments to conduct, and further elucidate whether or not it will be possible to reach precision goal s. Conclusion The temporary storage and syringe transfer method was developed to explore and streamline alternate techniques for the preparation of samples for Hg isotopic analysis. Preparation time and the use of hazardous chemicals were improved over tr aditional approaches, but the precision of the experimental method was found to be less adequate. The standard 202 Hg values for 2711 and TORT 1 prepared by the experimental method were nearly twice those for bench top acid digestion with c old vapor introduction. Differences for the values of SARM 20 were even more dramatic, but the complications observed in its analyses make it difficult to interpret. Additional experiments will be necessary to further refine the experimental method and a ssess its applicability. The toxicity of Hg to humans and other animals warrants an understanding of its biogeochemical cycling throughout the environment. Mercury isotope ratios can prove useful in further building this understanding, but obtaining these ratios requires the use of relatively


71 inefficient methods. If these methods could be improved upon, observational data collected from environmental samples might be more forthcoming.


72 REFERENCES Bigeleisen J (1965) Chemistry of Isotopes. Science 147 463 471. Bigeleisen J (1 996) Nuclear Size and Shape Effects in Chemical Reactions. Isotope Chemistry of the Heavy Elements. J. Am. Chem. Soc. 118 3676 3680. Bergquist BA, Blum JD (2007) Mass Dependent and Independent Fractionation of Hg Isotopes by Photoreduction in Aquatic Systems. Science 318 417 420. Biswas A, Blum JD, Bergquist BA, Keeler GJ, Xie Z (2008) Natural mercury isotope variation in coal deposits and organic soils. Environ. Sci. Technol. 42 8303 8309. Blum JD, Bergquist BA (2007) Reporting of variations in the natural isotopic composition of mercury Anal. Bioanal. Chem. 388 353 359. Buchachenko AL (2001) Magnetic Isotope Effect: Nuclear Spin Control of Chemical Reactions. J. Phys. Chem. A 105 9995 10011. Carignan J, Estrade N, Sonke JE, Donard OFX (2009) Odd Isotope Deficits in Atmospheric Hg Measured in Lichens. Environ. Sci. Technol. 43 5660 5664. Carroll FA (2011) Chapter 6: Methods of Studying Organic Reactions. In Perspectives on Structure and Mechanism in Organic Chemistry Second Edition John Wiley & Sons, Inc., New Jersey. pp. 327 411 Carvalho CML, Chew EH, Hashemy SI, Lu J, Holmgren A, Holmgren A (2008) Inhibition of the human thioredoxin system: a molecular mechanism of mercury toxicity. J. Biol. Chem 283 11913 11923. CET AC Technologies Aridus II TM Chesworth, W (1971) Use of aluminum amalgam in mineral synthesis at low temperatures and 1 atmosphere total pressure. Clay Clay Miner. 19 337 339. Clarkson TW, Magos L (2006) The Toxicology of Mercury and Its Chemical Compounds. Crit. Rev. Toxicol. 36 609 662. CRC (1972) Handbook of Chemistry and Physics (eds Weast RC, et al.). The Chemical Rubber Co., Cleveland. pp. D147 D150, D171 D177, E56 EPA (2005) Method 3200 Mercur y Species Fractionation and Quantification by Microwave Assisted Extraction, Selective Solvent Extraction AND/OR Solid Phase Extraction. In SW 846 Test Methods for Evaluating Solid Waste, Physical/Chemical Methods pp. 12. EPA (2007) Chapter 3 Inorganic Analytes. In SW 846 Test Methods for Evaluating Solid Waste, Physical/Chemical Methods pp. 8. Estrade N, Carignan J, Sonke JE, Donard OFX (2009) Measuring Hg Isotopes in Bio Geo Environmental Reference Materials. Geostand. Geoanal. Res. 34 79 93.


73 Evans RD, Hintelmann H, Dillon PJ (2001) Measurement of high precision isotope ratios for mercury from coals using transient signals. J. Anal. At. Spectrom. 16 1064 1069. Feng X, Foucher D, Hintelmann H, Yan H, He T, Qiu G (2010) Tracing Mercury Contami nation Sources in Sediments Using Mercury Isotope Compositions. Environ. Sci. Technol 44 3363 3368. Foucher D, Hintelmann H (2006) High precision measurement of mercury isotope ratios in sediments using cold vapor generation multi collector inductivel y coupled plasma mass spectrometry. Anal. Bioanal. Chem. 384 1470 1478. Gray JE, Pribil MJ, Van Metre PC, Borrok DM, Thapalia A (2013) Identification of contamination in a lake sediment core using Hg and Pb isotopic compositions, Lake Ballinger, Washing ton, USA. Appl. Geochem. 29 1 12. Jackson TA, Muir DCG (2004) Historical Variations in the Stable Isotope Composition of Mercury in Arctic Lake Sediments. Environ. Sci. Technol. 38 2813 2821. Jackson TA, Muir DCG (2012) Mass dependent and mass indepen dent variations in the isotope composition of mercury in a sediment core from a lake polluted by emissions from the combustion of coal. Sci. Total Environ. 417 418 189 203. Jackson TA, Whittle DM, Evans MS, Muir DCG (2008) Evidence for mass independent and mass dependent fractionation of the stable isotopes of mercury by natural processes in aquatic ecosystems. Appl. Geochem. 23 547 571. Kendall C, Caldwell EA (1998) Chapt er 2 Fundamentals of Isotope Geochemistry. In Isotope Tracers in Catchment Hydrology (eds Kendall C, McDonnell JJ). Elsevier Science B.V., Amsterdam. pp. 51 86. Klaue B, Blum JD (1999) Trace Analysis of Arsenic in Drinking Water by Inductively Coupled P lasma Mass Spectrometry: High Resolution versus Hydride Generation. Anal Chem. 71 1408 1414. Krupp EM, Donard OFX (2005) Isotope ratios on transient signals with GC MC ICP MS. Int. J. Mass Spectrom. 242 233 242. Lyons JR, Young ED (2005) CO self shiel ding as the origin of oxygen isotope anomalies in the early solar nebula. Nature 435 317 320. Masbou J, Point D, Sonke JE (2013) Application of a selective extraction method for methylmercury compound specific stable isotope analysis (MeHg CSIA) in bio logical materials. J. Anal. At. Spectrom 28 1620 1628. Mason RP, Fitzgerald WF, Morel FMM (1994) The biogeochemical cycling of elemental mercury: Anthropogenic influences. Geochim. Cosmochim. Ac. 58 3191 3198. Milestone, Inc. (2008) Milestone DMA 80 U ser Manual Revision 7, pp. 135. Milestone, Inc. (2014) Principles of Operation. Available at mercury analyzer/dma80 principles.html


74 Montaser A, McLean JA, Liu H, Mermet JM (1998) An Introduction to ICP Spectrometries For Elemental Analysis. In Inductively Coupled Plasma Mass Spectrometry (eds Montaser A ). Wiley VCH, Inc., New York. pp. 1 31 Nu Instruments (2003) Nu Plasma Operating Manual Version 2. Rahman GM, Kingston HM (2005) Development of a microwave assisted ext raction method and isotopic validation of mercury species in soils and sediments. J. Anal. At. Spectrom. 20 183 191. toxicity. Toxicology 278 112 123. Ravichan dran M (2004) Interactions between mercury and dissolved organic matter a review. Chemosphere 55 319 331. Schauble EA (2007) Role of nuclear volume in driving equilibrium stable isotope fractionation of mercury, thallium, and other very heavy elements. Geochim. Cosmochim. Ac. 71 2170 2189. Schweiger L, Stadler F, Bowes C (2006) Poisoning Wildlife: The Reality of Mercury Pollution National Wildlife Federation, Virginia. Skoog DA, Holler FJ, Crouch SR (2007) Chapter 6 An Introduction to Spectrometric methods. Principles of Instrumental Analysis 6th Edition Thomson Brooks/Cole, California. pp. 132 163 Sonke JE, Zambardi T, Toutain J (2008) Indirect gold trap MC ICP MS coupling for Hg stable isotope analysis using a s yringe injection interface. J. Anal. At. Spectrom. 23 569 573. Sonke JE, Schfer J, Chmeleff J, Audry S, Blanc G, Dupr B (2010) Sedimentary mercury stable isotope records of atmospheric and riverine pollution from two major European heavy metal refineries. Chem. Geol. 279 90 100. Stetson SJ, Gray JE, Wanty RB, Macalady DL (2009) Isotopic Variability of Mercury in Ore, Mine Waste Calcine, and Leachates of Mine Wa ste Calcine from Areas Mines for Mercury. Environ. Sci. Technol. 43 7331 7336. Thomas R (200 1 MS, Part IV: The Interface Region. Spectroscopy 16 26 34. Thomas R (200 1 MS, Part V: The Ion Focusing S ystem. Spectroscopy 16 38 44. Turner PJ, Mills DJ, Schroder, E, Lapitajs G, Jung G, Iacone LA, Haydar DA, Montaser A (1998) Instrumentation for Low and High Resolution ICPMS In Inductively Coupled Plasma Mass Spectrometry (eds Montaser A ). Wiley VCH, Inc., New York. pp. 421 501


75 Turner PJ, Montaser A (1998) Plasma Generation in ICPMS In Inductively Coupled Plasma Mass Spectrometry (eds Montaser A ). Wiley VCH, Inc., New York. pp. 265 334 Welz B (1985) Atomic absorption spectrometry Weinheim, Florida. Wieser ME, Schwieters JB (2005) The development of multiple collector mass spectrometry for isotope ratio measurements. Int. J. Mass. Spectrom. 242 97 115. Wilcox J, Rupp E, Ying SC, Lim D, Negreira AS, Kirchofer A, Feng F, Lee K (2 012) Mercury adsorption and oxidation in coal combustion and gasification processes. Int. J. Coal Geol. 90 91 4 20. WHO (1991) Inorganic Mercury United Nations Environment Programme, the International Labour Organisation, and the World Health Organiza tion. First draft prepared by Friberg L. Geneva. Xie Q, Lu S, Evans D, Dillon P, Hintelmann H (2005) High precision Hg isotope analysis of environmental samples using gold trap MC ICP MS. J. Anal. Atom. Spectrom. 20 515 522.


76 APPENDIX A EXPERIMENTAL DATA Observational data collected for the isotopic analyses in the Chapter 5 experiment can be found in Table A1. NIST 3133 bracketing analyses are highlighted in green for ease of reference. The three SARM 20 analyses prepared by combustion and grouped highlighted in blue.


77 Table A1: Observational Data SE 1.30E 03 3.22E 03 1.60E 03 2.66E 03 1.38E 03 1.50E 03 1.89E 03 1.89E 03 2.06E 03 2.03E 03 2.23E 03 2.50E 03 2.25E 03 2.53E 03 3.06E 03 1.65E 03 3.83E 03 Tl tot (V) 3.06 3.10 3.07 3.07 2.99 2.97 2.93 2.90 2.87 2.95 3.08 3.05 2.98 2.99 3.00 2.97 2.93 SE 9.85E 04 8.28E 03 1.39E 03 2.43E 03 2.13E 03 3.32E 03 1.60E 03 3.64E 03 1.27E 03 1.18E 03 1.49E 03 1.06E 03 3.20E 03 2.16E 03 3.10E 03 1.08E 03 1.04E 03 Hg tot (V) 2.23 2.07 2.25 2.65 2.29 2.20 2.22 2.40 2.20 2.17 1.96 2.24 2.11 2.23 2.11 2.23 2.07 SE 3.66E 05 3.24E 05 3.05E 05 4.23E 05 3.10E 05 3.21E 05 3.37E 05 2.58E 05 3.41E 05 3.38E 05 2.54E 05 4.28E 05 3.41E 05 2.22E 05 3.57E 05 3.10E 05 4.22E 05 199/198 1.688177 1.687728 1.688226 1.687874 1.688126 1.687852 1.688158 1.687823 1.688195 1.688294 1.687975 1.688342 1.687891 1.688305 1.687878 1.688176 1.687888 SE 4.72E 05 4.01E 05 5.05E 05 5.85E 05 5.77E 05 4.54E 05 4.23E 05 4.09E 05 4.41E 05 6.34E 05 5.14E 05 6.32E 05 6.74E 05 4.25E 05 4.69E 05 4.34E 05 4.97E 05 200/198 2.306794 2.306468 2.307029 2.306893 2.306794 2.306714 2.307019 2.306862 2.307004 2.307006 2.307000 2.307355 2.306954 2.307192 2.306847 2.307080 2.306854 SE 3.32E 05 2.90E 05 3.83E 05 1.26E 04 4.03E 05 4.47E 05 3.45E 05 2.28E 05 3.43E 05 4.22E 05 3.97E 05 4.20E 05 4.74E 05 3.62E 05 3.31E 05 2.74E 05 3.53E 05 201/198 1.313820 1.313478 1.313924 1.313612 1.313764 1.313547 1.313875 1.313614 1.313867 1.314091 1.313806 1.314267 1.313738 1.314098 1.313660 1.314162 1.313629 SE 8.31E 05 7.67E 05 8.72E 05 1.02E 04 1.55E 04 1.39E 04 7.90E 05 7.46E 05 7.93E 05 1.03E 04 7.80E 05 1.07E 04 1.16E 04 8.25E 05 8.01E 05 7.40E 05 8.92E 05 202/198 2.966509 2.966101 2.966867 2.966742 2.966479 2.966325 2.967060 2.966484 2.966754 2.967159 2.966995 2.967525 2.966739 2.967262 2.966665 2.967482 2.966597 Date/Sample 7/25/2013 3133 2711 1 microwave 3133 2711 2 microwave 3133 2711 3 microwave 3133 2711 4 microwave 3133 3133 2711 5 microwave 3133 2711 6 microwave 3133 2711 7 microwave 3133 2711 8 microwave


78 Table A1: Observational Data 2.18E 03 1.72E 03 3.69E 03 2.14E 03 3.02E 03 2.65E 03 3.75E 03 2.20E 03 3.56E 03 2.29E 03 2.96E 03 3.16E 03 3.22E 03 2.15E 03 3.02E 03 2.65E 03 3.24E 03 1.96E 03 3.40E 03 2.98 2.95 1.89 2.93 1.86 2.91 1.65 2.87 1.39 2.82 2.30 2.82 1.42 2.74 1.75 2.75 1.52 2.75 3.12 1.64E 03 2.65E 03 1.31E 02 2.78E 03 1.71E 02 2.81E 03 6.29E 03 2.35E 03 7.17E 03 2.84E 03 2.60E 03 2.39E 03 1.74E 03 3.26E 03 2.58E 03 2.32E 03 2.21E 03 4.51E 03 2.44E 03 2.22 2.18 1.33 2.12 1.29 2.10 0.61 2.07 0.24 2.02 0.70 2.04 0.16 1.98 0.36 2.00 0.20 1.95 2.47 3.23E 05 3.91E 05 5.35E 05 3.55E 05 3.82E 05 3.37E 05 9.55E 05 3.46E 05 1.97E 04 4.76E 05 6.55E 05 3.24E 05 2.69E 04 3.34E 05 1.09E 04 4.35E 05 1.88E 04 4.46E 05 3.00E 05 1.688181 1.688200 1.688413 1.688348 1.688378 1.688295 1.687996 1.688350 1.689716 1.688301 1.688458 1.688444 1.687944 1.688396 1.688570 1.688422 1.688626 1.688371 1.688531 5.16E 05 4.08E 05 7.18E 05 4.92E 05 6.83E 05 5.27E 05 9.20E 05 4.83E 05 2.20E 04 6.70E 05 7.49E 05 5.96E 05 3.32E 04 4.94E 05 1.49E 04 5.77E 05 2.89E 04 7.49E 05 5.87E 05 2.307108 2.307001 2.306716 2.307213 2.306621 2.307164 2.305978 2.307208 2.306550 2.306916 2.306496 2.307020 2.306224 2.306761 2.306910 2.307068 2.305958 2.306876 2.307810 3.82E 05 4.85E 05 4.40E 05 3.41E 05 5.51E 05 3.52E 05 8.12E 05 2.85E 05 1.57E 04 5.16E 05 5.44E 05 4.75E 05 2.29E 04 3.16E 05 9.47E 05 5.57E 05 2.47E 04 5.90E 05 4.74E 05 1.314104 1.314066 1.314042 1.314099 1.313961 1.314091 1.313598 1.314155 1.314297 1.313953 1.314010 1.313961 1.313341 1.313863 1.314003 1.314073 1.313956 1.313908 1.314621 9.69E 05 8.86E 05 1.51E 04 7.70E 05 1.48E 04 1.03E 04 1.84E 04 8.69E 05 3.03E 04 1.26E 04 1.39E 04 1.13E 04 4.34E 04 9.72E 05 2.19E 04 1.08E 04 3.17E 04 1.18E 04 1.23E 04 2.967266 2.967167 2.965943 2.967395 2.965498 2.967392 2.964449 2.967338 2.965505 2.966693 2.966159 2.966872 2.965239 2.966378 2.965428 2.966995 2.965175 2.966644 2.968989 3133 3133 TORT 1 1 microwave 3133 TORT 1 2 microwave 3133 TORT 1 3 microwave 3133 TORT 1 4 microwave 3133 TORT 1 5 microwave 3133 TORT 1 6 microwave 3133 TORT 1 7 microwave 3133 TORT 1 8 microwave 3133 7/29/2013 3133


79 Table A1: Observational Data 5.29E 03 4.09E 03 2.64E 03 1.93E 03 2.91E 02 2.87E 03 3.57E 02 2.60E 03 2.53E 03 1.74E 03 1.57E 03 1.83E 03 2.11E 03 1.91E 03 1.96E 03 1.77E 03 2.01E 03 3.76E 03 5.81E 03 3.23 3.25 3.27 3.32 3.23 3.36 3.66 3.39 3.49 3.48 3.46 3.46 3.45 3.46 3.45 3.45 3.48 3.34 3.42 1.91E 03 1.96E 03 2.14E 03 1.76E 03 1.85E 02 1.83E 03 1.08E 03 2.25E 03 1.22E 03 4.83E 03 1.47E 03 7.92E 03 1.55E 03 9.84E 04 1.80E 03 1.56E 03 1.99E 03 1.28E 03 2.33E 03 2.26 2.52 2.50 2.54 2.13 2.56 2.29 2.58 2.63 2.81 2.66 2.71 2.65 2.25 2.63 2.06 2.61 2.74 1.40 4.12E 05 3.42E 05 2.30E 05 3.57E 05 4.11E 05 3.18E 05 2.73E 05 3.53E 05 3.10E 05 3.92E 05 3.20E 05 3.25E 05 3.43E 05 3.77E 05 2.61E 05 3.54E 05 3.69E 05 3.70E 05 4.50E 05 1.688034 1.688482 1.688044 1.688425 1.687885 1.688448 1.687978 1.688345 1.688509 1.688025 1.688403 1.687979 1.688386 1.688047 1.688431 1.688108 1.688433 1.688534 1.688896 4.99E 05 4.58E 05 5.69E 05 4.08E 05 5.83E 05 4.95E 05 6.01E 05 5.26E 05 5.46E 05 5.66E 05 4.46E 05 5.21E 05 4.07E 05 4.46E 05 2.51E 05 6.63E 05 4.28E 05 4.83E 05 5.70E 05 2.307387 2.307646 2.307411 2.307745 2.307171 2.307620 2.307273 2.307452 2.307787 2.307457 2.307555 2.307366 2.307446 2.307377 2.307511 2.307413 2.307621 2.307781 2.306616 3.68E 05 3.85E 05 3.88E 05 3.09E 05 4.16E 05 4.25E 05 4.02E 05 4.47E 05 4.09E 05 4.23E 05 4.48E 05 4.81E 05 3.28E 05 3.77E 05 2.61E 05 4.68E 05 3.72E 05 3.94E 05 3.20E 05 1.314075 1.314553 1.314126 1.314575 1.314025 1.314460 1.314036 1.314388 1.314639 1.314121 1.314445 1.314090 1.314293 1.314098 1.314408 1.314046 1.314382 1.314650 1.314031 7.11E 05 9.21E 05 1.08E 04 7.01E 05 1.98E 04 8.29E 05 1.13E 04 1.01E 04 1.14E 04 1.15E 04 9.41E 05 1.11E 04 8.16E 05 8.16E 05 6.38E 05 1.15E 04 6.72E 05 8.35E 05 6.60E 05 2.968084 2.968898 2.968053 2.968872 2.967783 2.968597 2.967919 2.968569 2.968987 2.968063 2.968282 2.967998 2.968215 2.967990 2.968251 2.967937 2.968340 2.968646 2.965385 2711 1 bench 3133 2711 2 bench 3133 2711 3 bench 3133 2711 4 bench 3133 3133 2711 5 bench 3133 2711 6 bench 3133 2711 7 bench 3133 2711 8 bench 3133 7/30/2013 3133 TORT 1 1 bench


80 Table A1: Observational Data 2.78E 03 1.96E 03 2.40E 03 3.20E 03 2.89E 03 5.03E 03 2.39E 03 3.20E 03 3.06E 03 3.28E 03 2.58E 03 2.38E 03 2.36E 03 1.83E 03 8.13E 03 6.19E 02 5.02E 03 9.37E 03 1.81E 02 3.45 3.40 3.50 3.48 3.46 3.39 3.51 3.72 3.67 3.75 3.75 3.72 3.75 3.80 3.77 3.57 3.91 3.96 3.64 7.13E 03 1.07E 03 6.57E 03 4.34E 03 6.74E 03 1.85E 03 4.24E 03 2.62E 03 2.38E 03 3.52E 03 2.48E 03 8.95E 03 2.99E 03 7.37E 03 6.25E 03 4.25E 02 2.31E 03 1.67E 03 8.26E 03 1.57 1.55 2.57 1.36 2.51 1.46 2.55 2.83 1.48 2.55 1.38 2.53 1.29 2.52 1.74 2.50 2.77 1.49 2.74 4.85 E 05 5.91E 05 3.37E 05 5.69E 05 4.45E 05 4.70E 05 4.36E 05 2.87E 05 5.51E 05 5.72E 05 3.01E 05 2.12E 05 5.05E 05 3.92E 05 6.24E 05 4.84E 05 3.22E 05 4.49E 05 2.51E 05 1.688374 1.688912 1.688439 1.688783 1.688479 1.689059 1.688603 1.688560 1.688785 1.688387 1.688974 1.688359 1.688798 1.688476 1.688892 1.688562 1.688168 1.687255 1.688065 6.58E 05 7.41E 05 6.08E 05 6.79E 05 3.99E 05 5.97E 05 5.34E 05 5.18E 05 5.13E+05 4.73E 05 5.96E 05 4.02E 05 7.22E 05 5.11E 05 7.71E 05 8.46E 05 3.74E 05 5.57E 05 4.55E 05 2.307385 2.306940 2.307559 2.306473 2.307568 2.306741 2.307722 2.307783 2.306641 2.307430 2.306794 2.307338 2.306547 2.307421 2.307026 2.307582 2.306968 2.305841 2.306866 4.61 E 05 4.92E 05 4.56E 05 4.11E 05 6.04E 05 3.65E 05 4.72E 05 2.89E 05 3.99E 05 3.37E 05 4.07E 05 2.72E 05 5.86E 05 3.66E 05 3.74E 05 5.43E 05 3.31E 05 3.21E 05 3.15E 05 1.314339 1.314169 1.314530 1.314028 1.314413 1.314196 1.314618 1.314799 1.313988 1.314364 1.314240 1.314371 1.314013 1.314451 1.314379 1.314602 1.314073 1.312549 1.313911 1.04E 04 1.29E 04 8.24E 05 1.11E 04 1.47E 04 8.17E 05 1.34E 04 8.05E 05 8.36E 05 7.80E 05 1.19E 04 1.19E 04 1.21E 04 8.20E 05 1.17E 04 1.73E 04 7.74E 05 6.05E 05 9.42E 05 2.967935 2.966087 2.968134 2.965200 2.967907 2.965689 2.968465 2.968829 2.965315 2.967951 2.965967 2.967687 2.965335 2.968021 2.966358 2.968158 2.967206 2.964034 2.967049 3133 TORT 1 2 bench 3133 TORT 1 3 bench 3133 TORT 1 4 bench 3133 3133 TORT 1 5 bench 3133 TORT 1 6 bench 3133 TORT 1 7 bench 3133 TORT 1 8 bench 3133 7/30/2013 3133 SARM 20 1 bench 3133


81 Tab le A1: Observational Data 1.62E 03 1.69E 03 1.69E 03 1.92E 03 1.43E 03 1.44E 03 2.17E 03 1.49E 03 1.63E 03 2.17E 03 4.82E 03 1.74E 03 2.04E 03 3.96E 03 1.56E 03 1.76E 03 1.78E 03 2.75E 03 2.36E 03 4.01 4.06 4.08 4.10 4.01 4.07 4.07 4.17 4.10 4.16 3.99 4.16 4.17 4.21 0.76 0.76 0.77 0.75 0.79 9.88E 04 5.79E 03 1.26E 03 7.80E 03 1.10E 03 1.10E 02 1.10E 03 9.64E 03 2.97E 03 7.07E 03 5.66E 03 2.41E 03 1.50E 03 1.22E 02 4.65E 03 7.81E 04 5.78E 03 1.96E 03 5.57E 03 1.71 2.76 1.46 2.81 1.62 2.78 1.80 2.71 2.09 2.61 6.20 2.90 1.73 2.79 1.34 0.93 1.36 1.11 1.35 5.61E 05 3.67E 05 4.25E 05 4.07E 05 4.93E 05 4.06E 05 4.44E 05 3.10E 05 4.39E 05 3.76E 05 3.26E 05 2.92E 05 4.11E 05 3.09E 05 6.40E 05 7.80E 05 5.45E 05 4.96E 05 1.14E 04 1.687331 1.688199 1.687383 1.688076 1.687280 1.688040 1.687337 1.688140 1.687257 1.688052 1.688102 1.688186 1.687441 1.688172 1.688525 1.688178 1.688378 1.688133 1.688689 3.84E 05 5.03E 05 6.49E 05 7.06E 05 5.95E 05 4.57E 05 6.13E 05 4.41E 05 4.16E 05 4.57E 05 4.17E 05 3.79E 05 5.89E 05 4.29E 05 9.77E 05 9.68E 05 7.31E 05 8.71E 05 2.36E 04 2.305517 2.306950 2.305725 2.306830 2.305557 2.306729 2.305764 2.306920 2.305712 2.306705 2.305594 2.306709 2.305803 2.306905 2.307749 2.307740 2.307700 2.307518 2.307997 3.22E 05 3.28E 05 3.75E 05 2.63E 05 3.57E 05 3.03E 05 4.25E 05 4.14E 05 3.07E 05 3.13E 05 2.89E 05 2.83E 05 3.41E 05 2.83E 05 7.18E 05 8.17E 05 5.53E 05 7.19E 05 2.10E 04 1.312403 1.313943 1.312511 1.313705 1.312421 1.313765 1.312486 1.313860 1.312550 1.313729 1.312867 1.313879 1.312575 1.313817 1.314419 1.314231 1.314466 1.314168 1.314604 7.39E 05 8.98E 05 9.46E 05 7.35E 05 1.01E 04 6.76E 05 6.87E 05 6.47E 05 6.80E 05 7.08E 05 7.16E 05 7.93E 05 1.01E 04 8.24E 05 1.67E 04 2.03E 04 1.28E 04 2.21E 04 1.37E 04 2.963592 2.967025 2.963938 2.966669 2.963597 2.966633 2.963742 2.966792 2.963728 2.966359 2.963293 2.966496 2.963774 2.966779 2.968859 2.968852 2.968910 2.968329 2.968754 SARM 20 2 bench 3133 SARM 20 3 bench 3133 SARM 20 4 bench 3133 SARM 20 5 bench 3133 SARM 20 6 bench 3133 SARM 20 7 bench 3133 SARM 20 8 bench 3133 8/29/2013 3133 2711 1 combustion 3133 2711 2 combustion 3133


82 Table A1: Observational Data 2.33E 03 1.88E 03 2.38E 03 2.54E 03 3.76E 03 3.99E 03 3.94E 03 3.89E 03 4.14E 03 3.76E 03 3.97E 03 4.36E 03 4.32E 03 4.32E 03 1.14E 03 9.87E 04 6.05E 04 8.69E 04 4.76E 04 0.77 0.78 0.77 0.79 0.86 0.88 0.88 0.87 0.88 0.88 0.88 0.96 0.93 0.95 0.90 0.85 0.91 0.85 0.91 2.16E 03 6.42E 03 4.03E 03 5.51E 03 5.14E 03 3.43E 03 3.46E 03 5.46E 03 3.48E 03 3.39E 03 3.79E 03 2.71E 03 2.18E 03 2.16E 03 3.69E 03 9.01E 04 6.56E 03 9.00E 03 2.29E 03 1.06 1.33 1.22 1.34 1.10 1.25 1.14 1.33 1.12 1.22 1.10 1.16 1.23 1.15 1.38 0.76 1.30 2.18 1.40 8.27E 05 9.83E 05 7.79E 05 4.07E 05 7.84E 05 6.64E 05 6.37E 05 7.61E 05 6.53E 05 1.05E 04 5.42E 05 5.74E 05 5.80E 05 5.24E 05 6.84E 05 7.37E 05 6.19E 05 3.83E 05 6.84E 05 1.688062 1.688746 1.687883 1.688635 1.688357 1.688061 1.688394 1.687941 1.688269 1.688181 1.688407 1.688450 1.687898 1.688348 1.688770 1.687433 1.688428 1.688015 1.688436 1.07E 04 5.69E 05 1.43E 04 7.52E 05 1.03E 04 8.96E 05 8.86E 05 1.32E 04 9.32E 05 1.44E 04 6.20E 05 6.94E 05 5.66E 05 6.29E 05 7.55E 05 1.07E 04 8.47E 05 6.62E 05 1.14E 04 2.307366 2.307803 2.306806 2.307666 2.307116 2.307409 2.307477 2.307421 2.307403 2.307513 2.307273 2.307610 2.307226 2.307411 2.307997 2.306202 2.307506 2.307523 2.307416 7.26E 05 4.40E 05 9.39E 05 5.41E 05 8.07E 05 7.54E 05 4.83E 05 9.46E 05 6.28E 05 8.45E 05 4.64E 05 6.30E 05 3.63E 05 4.95E 05 5.99E 05 5.21E 05 6.93E 05 6.44E 05 7.25E 05 1.314022 1.314360 1.313559 1.314264 1.313999 1.313994 1.313986 1.313844 1.314249 1.314004 1.314066 1.314357 1.313831 1.314044 1.314590 1.313261 1.314150 1.314061 1.314224 1.96E 04 9.71E 05 2.55E 04 1.45E 04 1.84E 04 1.70E 04 1.54E 04 2.16E 04 1.36E 04 3.09E 04 1.07E 04 1.27E 04 9.87E 05 1.28E 04 1.52E 04 1.57E 04 1.79E 04 1.22E 04 3.60E 04 2.968102 2.968459 2.966832 2.968395 2.967894 2.968114 2.968296 2.968119 2.968196 2.968020 2.967713 2.968476 2.967922 2.968033 2.968979 2.965834 2.968282 2.968494 2.967817 2711 3 combustion 3133 2711 4 combustion 3133 3133 2711 5 combustion 3133 2711 6 cvombustion 3133 2711 7 combustion 3133 3133 2711 8 combustion 3133 8/30/2013 3133 SARM 20 1 comb 3133 SARM 20 2 comb 3133


83 Table A1: Observational Data 7.82E 04 1.08E 03 8.35E 04 1.60E 03 4.66E 04 1.09E 03 1.00E 03 8.04E 04 2.22E 03 1.07E 03 5.90E 04 4.85E 04 1.44E 03 7.75E 04 2.40E 03 3.64E 03 7.56E 04 0.84 0.90 0.83 0.89 0.93 0.87 0.91 0.86 0.90 0.85 0.91 0.84 0.87 0.80 0.69 0.81 0.96 4.09E 03 4.17E 03 2.39E 03 5.45E 03 6.38E 03 2.20E 03 6.00E 03 7.21E 05 8.86E 03 1.39E 03 6.81E 03 1.20E 03 7.15E 03 5.07E 04 2.41E 03 4.02E 04 2.42E 03 1.56 1.28 1.21 1.21 1.46 0.66 1.41 0.90 1.32 0.94 1.36 0.89 1.23 0.60 0.27 0.52 1.10 5.23E 05 3.61E 05 5.27E 05 4.64E 05 1.25E 04 7.27E 05 3.71E 05 5.57E 05 5.32E 05 6.30E 05 5.70E 05 6.78E 05 5.66E 05 8.09E 05 1.33E 04 1.12E 04 6.76E 05 1.687927 1.688313 1.687539 1.688440 1.688812 1.687243 1.688474 1.687701 1.688524 1.687504 1.688622 1.687540 1.688558 1.688741 1.689189 1.688650 1.688446 8.55E 05 6.29E 05 5.77E 05 7.04E 05 2.84E 04 1.15E 04 5.90E 05 1.08E 04 8.12E 05 8.14E 05 6.74E 05 2.18E 04 8.37E 05 1.37E 04 1.90E 04 1.50E 04 7.10E 05 2.307263 2.307328 2.306729 2.307670 2.307948 2.306222 2.307580 2.306331 2.307700 2.306225 2.307672 2.305978 2.307307 2.307559 2.307396 2.307193 2.306908 6.29E 05 5.40E 05 4.64E 05 6.32E 05 2.63E 04 9.26E 05 7.07E 05 8.67E 05 6.19E 05 6.82E 05 4.89E 05 1.41E 04 5.79E 05 9.62E 05 1.37E 04 7.52E 05 6.66E 05 1.313678 1.314002 1.313262 1.314253 1.314471 1.312809 1.314277 1.312927 1.314304 1.313053 1.314348 1.312850 1.314161 1.314490 1.315384 1.314299 1.314028 2.22E 04 1.56E 04 1.09E 04 1.65E 04 7.27E 04 1.97E 04 1.68E 04 1.83E 04 1.51E 04 1.67E 04 1.61E 04 2.51E 04 1.51E 04 2.71E 04 3.10E 04 2.23E 04 1.42E 04 2.9 67529 2.967822 2.966387 2.968345 2.968808 2.965304 2.968484 2.965534 2.968543 2.965577 2.968787 2.965370 2.967963 2.968463 2.967781 2.967909 2.966965 SARM 20 3 comb 3133 SARM 20 4 comb 3133 3133 SARM 20 5 comb 3133 SARM 20 6 comb 3133 SARM 20 7 comb 3133 SARM 20 8 comb 3133 9/10/2013 3133 TORT 1 1 comb 3133 9/11/2013 3133


84 Table A1: Observational Data 1.63E 03 7.17E 04 9.02E 03 1.54E 02 2.31E 03 2.94E 03 2.98E 03 3.72E 03 2.42E 03 3.16E 03 2.53E 03 3.13E 03 4.30E 03 3.21E 03 1.97E 03 1.84E 03 1.85E 03 0.81 0.97 1.10 0.87 1.08 1.05 0.97 1.06 1.10 0.95 1.00 1.07 1.05 1.04 1.02 1.01 1.02 3.49E 03 3.43E 03 1.20E 03 6.77E 03 2.35E 03 1.23E 03 4.46E 03 7.38E 04 6.47E 04 4.92E 03 2.43E 03 1.53E 03 7.69E 03 3.74E 03 2.69E 03 2.42E 03 7.27E 04 0.24 1.05 1.07 0.33 0.99 1.03 0.28 0.76 1.06 0.24 1.32 1.58 0.36 1.70 1.12 1.02 0.92 2.14E 04 6.18E 05 6.88E 05 1.34E 04 8.11E 05 4.48E 05 1.59E 04 9.30E 05 7.10E 05 1.95E 04 5.74E 05 4.88E 05 1.52E 04 4.44E 05 5.52E 05 5.98E 05 5.75E 05 1.688972 1.688553 1.688361 1.688727 1.688284 1.688543 1.689626 1.688626 1.688772 1.689760 1.688576 1.688504 1.689437 1.688467 1.688395 1.688633 1.688846 2.76E 04 1.22E 04 1.49E 04 2.23E 04 1.19E 04 5.88E 05 2.06E 04 9.68E 05 1.31E 04 2.35E 04 9.98E 05 8.63E 05 1.75E 04 7.61E 05 7.80E 05 6.30E 05 6.64E 05 2.305973 2.307144 2.307012 2.306225 2.306626 2.306891 2.306482 2.307026 2.307235 2.306922 2.307014 2.306685 2.306306 2.306687 2.306547 2.306581 2.306690 1.39E 04 6.71E 05 7.99E 05 1.29E 04 9.49E 05 5.33E 05 1.54E 04 9.76E 05 7.49E 05 1.87E 04 7.27E 05 3.97E 05 1.19E 04 5.40E 05 5.98E 05 6.66E 05 5.38E 05 1.313499 1.313911 1.314023 1.313352 1.313725 1.313987 1.314127 1.314016 1.314068 1.313549 1.313929 1.313804 1.313422 1.313707 1.313381 1.313702 1.313597 3.81E 04 1.63E 04 2.02E 04 2.82E 04 1.62E 04 1.33E 04 2.68E 04 1.51E 04 1.74E 04 3.92E 04 1.62E 04 1.15E 04 1.82E 04 1.33E 04 1.54E 04 1.23E 04 1.32E 04 2.9 63987 2.966957 2.966710 2.964342 2.966133 2.966941 2.965524 2.966892 2.966981 2.965088 2.966731 2.966360 2.963467 2.966047 2.965835 2.965722 2.965537 TORT 1 2 comb 3133 9/12/2013 3133 TORT 1 3 comb 3133 3133 TORT 1 4 comb 3133 3133 TORT 1 5 comb 3133 3133 TORT 1 6 comb 3133 9/13/2013 3133 NIST 1944 GXR 1


85 Table A1: Observational Data 1.97E 03 2.09E 03 2.61E 03 2.24E 03 2.64E 03 4.96E 03 1.86E 03 1.53E 03 1.04 1.03 1.05 1.04 1.03 1.14 1.00 1.02 2.50E 03 4.99E 04 8.57E 04 1.87E 03 2.20E 03 2.18E 03 1.25E 03 4.01E 03 1.37 1.09 1.21 1.28 1.55 1.68 1.23 1.84 5.04E 05 4.55E 05 4.46E 05 6.45E 05 5.04E 05 6.31E 05 4.40E 05 3.74E 05 1.688415 1.688585 1.688463 1.688250 1.688390 1.688189 1.692499 1.688507 6.50E 05 7.13E 05 7.00E 05 8.14E 05 7.73E 05 6.72E 05 7.67E 05 6.21E 05 2.306584 2.306835 2.306705 2.306549 2.306644 2.306085 2.307115 2.306570 4.44E 05 6.46E 05 5.94E 05 5.91E 05 7.74E 05 5.63E 05 3.73E 05 3.95E 05 1.313735 1.313698 1.313608 1.313418 1.313782 1.313415 1.316396 1.313724 1.20E 04 1.24E 04 1.64E 04 1.80E 04 1.29E 04 1.65E 04 1.35E 04 1.21E 04 2.966029 2.966198 2.965874 2.965520 2.966174 2.965447 2.967011 2.966023 3133 GXR 2 3133 NIST 1944 3133 3133 NIST 1946 3133