Diss. ETH No. 14706

The Organic Chemistry of Volcanic Gases

at Vulcano (Aeolian Islands, Italy)

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZÜRICH

for the degree of

Doctor of Natural Sciences

presented by

Florian Maximilian Schwandner

Dipl. Geol-Paläontol., Freie Universität Berlin

born August 13th, 1970 citizen of the Federal Republic of Germany

accepted on the recommendation of

Prof. Dr. T.M. Seward Inst. of Mineralogy and Petrography, ETH Zürich examiner Prof. Dr. V.J. Dietrich Inst. of Mineralogy and Petrography, ETH Zürich co-examiner Dr. A. P. Gize Dept. of Earth Sciences, University of Manchester (UK) co-examiner

2002

To my family

i

Preface

Finally the printed “Pflichtexemplar” (mandatory copy) is done and printed, and life after the PhD can continue. In addition to the acknowledgements at the end of this thesis, a few remarks seem appropriate at this point. It has been a great pleasure and experience to conduct this work, with the professional, financial and personal support of Terry Seward, Volker “Wumme” Dietrich, Andy Gize, Jenny Cox, a variety of other colleagues as well as my family and friends. Christoph Wahrenberger preceeded me on the research topic and Alex Teague will continue on after me but I am sure there will be many more scientists “jumping on the train” in the nearest future.

There has been, still is and probably always will be great resistance to innovative ideas and approaches in science, especially by people who are so unfortunate to heavily depend on funding raised by and for mainstream “politically correct” research, or catastrophism. The idea to find permanent natural sources of natural halocarbons from volcanoes is not new but despite ist scientific foundation is still facing resistance. If a theory persists and dominates for a long time, then it eventually becomes a “fact” in many peoples eyes. This leads to publications starting with the phrase “It is well known that...”. To put the topic of this thesis in that context: many papers about halocarbons and ozone depletion for example contain the phrase “halocarbons are solely of anthropogenic origin” (or similar formulations).

In this PhD research, I had to learn that a strong scientific lobby exists that carries out research on anthropoghenic releases. To question one of their often stated basic assumptions (i.e. the phrase “halocarbons are solely of anthropogenic origin”) is in some ways threatening to “take the butter off their bread”, as an old German saying goes. The peer-reviewing of project funding proposals is therefore possibly sometimes hindering innovative ideas. However, an increasing number of research groups is since the nineties indeed focusing on natural sources of halocarbons.

It reminds me of a quote by Rabindranath Tagore (1861-1941; 1913 Nobel Laureate in Literature), “The false can never grow into truth by growing in power”("Stray Birds", 1916). The answer therefore has to be, to provide rigorous unbiased scientific evidence.

F.M. Schwandner, Zürich, August 28th, 2002

Preface ii

Table of contents

Preface i Abstract iv Zusammenfassung (German abstract) vi

Chapter I Introduction 1

Chapter II Halocarbons and other organic compounds in volcanic gases from Vulcano (Aeolian Islands, Italy) 7

Abstract 8 1. Introduction 9 2. Previous work 11 3. Volcanic impacts to the atmosphere 14 3.1. Global warming 14 3.2. Ozone depletion 17 4. Experimental 24 4.1. Sampling and preparative work 24 4.2. Analysis 26 4.3. Data treatment 27 4.4. Contamination, sample loss, degradation and blanks 30 5. Results 35 5.1. Hydrocarbons 35 5.2. Oxygenated compounds 42 5.3. Heteroatomic compounds excluding halocarbons and oxygenated hydrocarbons 43 5.4. Halocarbons 45 5.5. Other compounds 51 6. Flux estimates of volcanic halocarbon emissions 52 6.1. Total carbon fluxes and carbon recycling 52 6.2. Total halogen fluxes 52 6.3. Trace gas correlations 53 6.4. Single compound static flux estimate 56 6.5. Dynamic volcanogenic trace gas flux model 57 7. Discussion 63 7.1. Origin of source material 63 7.2. Variation with type of magmatic activity and/or tectonic setting 64 7.3. Processes of formation 65 8. Implications for further research and outlook 77 8.1. Atmospheric chemistry 77 8.2. Volcanic gas studies, exploration and volcanic surveillance 81 8.3. Metal transport in hydrothermal fluids 82 8.4. Origin of life and abiogenic hydrocarbons 84 9.Summary 85

Table of Contents iii

Chapter III Diffusive emission of organic trace gases from the flanks and crater of a quiescent active volcano (Vulcano, Aeolian Islands, Italy) 89

Abstract 90 1. Introduction 91 2. Experimental 93 2.1. Sample collection 93 2.2. Analytical methods 96 3. Results 99 3.1. Site parameters and CO2-flux 99 3.2. Chemical parameters 102 4. Summary and discussion 116 4.1. Vegetation effects and adsorption on soil particles 116 4.2. The crater region 118 4.3. Pyrolysis 119 4.4. Abiogenic formation 120

Chapter IV Conclusions 123

References 127

Appendix A1. A historic overview of organic chemical volcanic gas studies A1.1 A2. Additional GC-MS method information A2.1 A3. Chromatograms & mass spectra A3.1

Acknowledgements Curriculum vitae

Table of Contents iv

Abstract

In order to assess the source strength, composition and formation of trace organic constituents in volcanic gas emissions, we collected a total of 48 samples of sorbent-trapped volcanic gases, sublimates, condensates from active vents of the La Fossa crater on the island of Vulcano (Aeolian Islands, Italy) as well as ambient and industrial air, during three sampling campaigns in the years 1999, 2000 and 2001. Previously, trace organic gas analyses from volcanic sources have been limited by analytical problems such as the strong sulfur, halogen acid and water matrix that is deleterious to sensitive trace organic analysis by GC-MS. We developed a novel analytical and sampling approach that allows reliable, sensitive and quantitative analysis by Short-Path Thermal Desorption-Solid Phase Microextraction- Cryotrapping-Gas Chromatography/Mass Spectrometry (SPTD-SPME-CF-GC-MS.

In a total of 158 GC-MS analyses we detected and selectively quantified well over 100 compounds in each sample. We found alkanes, alkenes, arenes, phenols, aldehydes, carboxylic acids, esters, ketones, nitriles, furans, PAH's and their halogenated, methylated and sulfonated derivatives, as well as various heterocyclic compounds including thiophenes. Most compounds are present at concentrations well above laboratory, ambient air, adsorbent and field blank levels. For some analytes (e.g., CFC-11, CH2Cl2, CH3Br), concentrations are up to several orders of magnitude greater than even mid-latitudinal industrial urban air maxima. Air or laboratory contamination is negligible or absent on the basis of inert gas and noble gas isotopic measurements.

The organic compounds are interpreted as the product of inorganic, abiogenic gas-phase radical reactions. On the basis of isomer abundances, n-alkane distributions and the substitution pattern the compounds are considered to have been formed by high temperature (e.g., 900 °C) alkyl free radical reactions and halide electrophilic substitution on arenes, alkanes and alkenes.

Model global volcanic halocarbon fluxes from our own and published compositional data on a number of halocarbons have been obtained by scaling to published volcanic CO2 fluxes. This -9 -1 -4 -1 yields global halocarbon fluxes in the range from 1.0x 10 Tg y (CClF3) to 1.2 x 10 Tg y

(CH3Cl). Methyl bromide (CH3Br) and methyl iodide (CH3I) are estimated to have a volcanic model source strength of up to 1.3 x 10-6 Tg y-1, however, it is noted that other brominated

Abstract v

compounds were identified. The significance of organic emissions from quiescent volcanic degassing as opposed to explosive eruptions to stratospheric chemistry is that the negligible aqueous solubility of most emitted organic compounds does not lead to a quantitative tropospheric washout effect as postulated for explosive volcanic HCl and HF emissions.

Simple modeling suggests that episodic, catastrophically large, explosive events such as flood basalt eruptions do not impact the short-term ozone chemistry to the same magnitude as today’s anthropogenic emissions but may modify the natural atmospheric reactive balance over much longer (geological) time spans. In addition, the apparent abiogenic organic chemistry of volcanic gases may give insights into the possible hydrothermal origin of early life on Earth, as indicated by the presence of simple amino acids, nitriles, and alkanoic acids.

In order to clarify the source of these compounds we have sampled soil gas emissions of organic compounds and determined CO2 fluxes along a profile spanning from the vegetated base of the active La Fossa cone (Vulcano Island, Italy) over the unvegetated volcanic flank, the crater rim and base, into fumarolic areas.

The results indicate that a) the majority of volatile organic compounds in the soil gas show significant increases in concentration towards the crater and fumaroles, b) emissions of the 2 halocarbon CFC-11 (CCl3F) correlate well with soil CO2 fluxes measured on site (R =0.91 ±0.02) and both increase towards the crater and fumaroles, and c) diffusive emissions therefore contribute significantly to the volcanic halocarbon source strength. Other ozone depleting substances were found in concentrations significantly above those found in field and system blanks, including CH3Br, CH3Cl, CH3I, C2H5Br, and chlorinated benzenes. Abundances range from upper pptv to ppmv (e.g. CFC-11: max. 1200 pptv diffusive, 3700 pptv fumarolic/dry gas, dry air: 268 pptv). On the basis of these results, the natural volcanic source strength of halocarbon emissions to the atmosphere requires re-evaluation and correction to higher values.

Abstract vi

Zusammenfassung

Drei Beprobungskampagnen wurden in den Jahren 1999, 2000 und 2001 auf dem aktiven La Fossa Vulkan auf der Insel Vulkano (Aeolische inseln, Italien) durchgeführt, um die Quellstärken, Zusammensetzung und Formationsprozesse organischer Spurengase in vulkanischen Exhalationen zu charakterisieren. Weil der quantitative Erfolg früherer Studien an analytischen Problemen scheiterte, wurde ein neuer Ansatz zur Beprobung und Analyse entwickelten und erprobt.

Bei der Beprobung wurden Kondensat (Schwefel, Wasser, Halogensäuren) von Spurengas getrennt und auf Festphasen-Adsorbienten fixierten. Die Analyse erfolgte mittels der Technik “Kurzpfad-thermische Desorption mit Wiederauffangen mittels Festphasenmikroextraktion und nachfolgender gaschromatographisch-massernspektrometrischer Analyse” (Englisch: Short-Path Thermal Desorption-Solid Phase Microextraction-Cryotrapping-Gas Chromatography/Mass Spectrometry, SPTD-SPME-CF-GC-MS). Zur Entwicklung und Anwendung wurden insgesamt 158 GC-MS-Analysen durchgeführt, davon 48 an Proben von vulkanischem Material (Fumarolengase, Sublimate, Kondensate und Bodengase) sowie einige zusätzliche inorganische Gasanalysen).

Weit über 100 Komponenten wurden in jeder Probe durch einzelne Massenfragment-Analyse der Daten identifiziert, mit Hilfe von relativen Retentionsindizes bestätigt und ausgewählte Komponenten (vor allem halogenierte Kohlenwasserstoffe) quantifiziert. Blindprobenwerte waren durchwegs deutlich unterhalb der Probenwerte. Die Komponentengruppen beinhalteten Alkane, Alkene, Arene, Phenole, Aldehyde, Karboxylsäuren, Ester, Ketone, Nitrile, Furane, Polyaromatische Kohlenwasserstoffe (PAK) und deren halogenierte, methylierte und sulfonierte Derivate.

Einige Analyten zeigten Konzentrationen weit über atmosphärischen Hintergrundwerten. Kontamination wurde aufgrund sorgfältiger Blindprobenanalyse im Feld und im Labor ausgeschlossen. Die organischen Spurenchemie dieser vulkanischen Gase wird als das Resultat von Hochtemperatur-Radikal-Gasphasenreaktionen interpretiert. Auf Grundlage von Isomerverteilungen schliessen wir auf Alkylradikal-Reaktionen und eletrophile Substitution an Arenen und Alkanen.

Zusammenfassung vii

Die Quellstärken der vulkanischen Exhalationen wurde unter Zuhilfenahme von eigenen gemessenen und publizierten CO2-Entgasungsdaten modelliert. Globale jährliche -9 -1 -4 -1 Quellstärken erreichten dabei 1.0x 10 Tg y (CClF3) bis 1.2 x 10 Tg y (CH3Cl). -6 -1 Methylbromid (CH3Br) und Methyliodid (CH3I) werden auf bis zu 1.3 x 10 Tg y geschätzt.

Die Bedeutung dieser Werte zeigt, dass nicht-explosive Entgasungen ungehindert bis zur Tropopause aufsteigen können, weil sie aufgrund des Fehlens von Kondensationskernen (Aschenteilchen) nicht ausgewaschen werden, wie es bei explosiven Eruptionen der Fall ist. Diffuse Bodenentgasungen wurden am gleichen Vulkan beprobt, um die Quellen und

Bildungsprozesse zu beurteilen. Die spurenchemischen Daten sowie die CO2- Entgasungsdaten belegen, dass der Vulkan mit seinen Entgasungen die bei weitem stärkste Quelle darstellt. Natürliche Quelldaten von Ozonzerstörenden Substanzen müssen folglich global neu interpretiert werden.

Zusammenfassung viii

I. Introduction

2

Volcanoes discharge a variety of gases into the atmosphere, both during long lasting quiescent fumarolic activity and during episodic, catastrophic large explosive eruptions. Until recently, the major focus of the scientific community interested in volcanic emissions has focused on explosive eruptions only, however, few attempts to evaluate the source potential during intermittent, quiescent stages of activity have been done to date. Geochemists and volcanologists working with volcanic gases are usually not aware of trace constituents present in fumarolic volcanic gas discharges.

With the discovery of ozone depletion caused by anthropogenically released atmospheric halocarbon trace gases, research activity and funding for atmospheric trace gas research has increased dramatically. In accordance with this development, atmospheric budgets have been modeled with increasing accuracy but sources and sinks of some halocarbon compounds were found to not be sufficiently balanced. As a consequence, the search for additional natural sources has yielded many new results during the last decade, and attempts are being made to match known natural and anthropogenic sources to arrive at a coherent atmospheric halocarbon budget in order to more accurately predict future ozone and climate changes. On such predictions international treaties such as the Montral Protocol and the Kyoto Protocol are based.

The role of halocarbons in the atmosphere is both as strong greenhouse gases as well as ozone depleting substances. In order to understand long term trends in atmospheric budgets, long term acting sources have to be identified and characterized. Volcanoes have long been regarded as not emitting halocarbons because on the one hand geochemists are usually not concerned with atmospheric chemistry, and atmospheric chemists not with geochemistry. The few studies in the past that investigated volcanic halogen emissions were restricted to explosive eruptions. The eruption plumes of Mt. St. Helens have been extensively sampled by atmospheric chemists in 1980, but little attention has been paid to loss of compounds by adsorption on ash paricles, timely sampling in “undiluted” ash clouds, and fumarolic degassing prior to the eruption. After these studies were done, the conclusion became almost dogmatic that volcanoes don’t emit halocarbons. However, given the sulfur poor chemistry of the St. Helens eruption products, this eruption may not have been representative.

Chapter I. Introduction 3

Since quiescent degassing operates on a century to millennia scale and has similarly high major compound fluxes (i.e. SO2 and CO2), emissions of minor compounds may be significant but so far remained to be quantified. Some early studies gave indication of halocarbons present in volcanic gas discharges, however these results were either doubted by the authors themselves or regarded as air contamination. In addition to these dogmatic barriers, the sampling and analysis of organic trace gases in volcanic gas discharges is made difficult by the strong halogen acid, water and sulfur matrix which are deleterious to gas chromatographic mass spectrometric analysis. Halocarbon analyses from volcanic gases have therefore been to date severely hampered by analytical limitations, despite the great necessity to quantify natural sources.

In order to close this “scientific gap”, we developed novel sampling and analytical techniques to accurately and quantitatively determine halocarbon concentrations in >400°C hot volcanic gases. In chapter 2 the collection of volcanic gases, sublimates and condensates from active vents of the La Fossa crater on the island of Vulcano (Aeolian Islands, Italy) as well as ambient and industrial air by steam separation in impingers and subsequent sorbent trapping is described. The location of Vulcano Island, an active Island Arc volcano situated in the Thyrrenian Sea, is shown in Figure 1.

Figure 1 Location of the Island of Vulcano within the Aeolian Island Arc, north of Sicily (Italy). Vulcano is an active andesitic volcano capable of producing Plinian eruptions. The last eruptive (Vulcanian) activity was in the late nineteenth century and it is currently in the state of active quiescent fumarolic degassing, with temperatures reaching up to 600°C.

Chapter I. Introduction 4

We performed a total of 185 GC-MS analyses on adsorbent cartridge samples of ambient and industrial air samples, soil gases and fumarolic gas samples, as well as condensate, solid sublimate and Giggenbach bottle headspaces from thyree sampling campaigns, in the years 1999, 2000 and 2001. The adsorbed trace compounds were quantitatively analyzed by Short- Path Thermal Desorption-Solid Phase Microextraction-Cryotrapping-Gas Chromatography/Mass Spectrometry (SPTD-SPME-CF-GC-MS). In each sample we detected well over 100 compounds, many of which were in previous methods shielded by a large sulfur and water background. The analytes were identified by mass fragment analysis, confirmed by use of a relative retention index system and a selected range quantified by external dynamic standardization. The compound groups comprise alkanes, alkenes, arenes, phenols, aldehydes, carboxylic acids, esters, ketones, nitriles, furans, polyaromatic hydrocarbons (PAH's) and their halogenated, methylated and sulfonated derivatives, as well as various heterocyclic compounds including thiophenes. Most compounds are found at concentrations well above laboratory, ambient air, adsorbent and field blank levels. For some analytes (e.g., CFC-11,

CH2Cl2, CH3Br), concentrations are up to several orders of magnitude greater than even mid- latitudinal industrial urban air maxima. Air or laboratory contamination is negligible or absent on the basis of inert gas and noble gas isotopic measurements.

These results allow the quantitative determination of the volcanic halocarbon source strength of this individual volcano. Since air contamination during sampling can be ruled out after rigorous testing, the question remained if airborne halocarbons are the source of the halocarbons found in volcanic gases. In order to test this hypothesis, we performed a soil gas flux and trace chemical analytical survey along a profile spanning from the vegetated base of the volcano over its flanks into the crater and to the fumaroles. If the halocarbons were derived from atmospheric transport, the observed concentrations were expected to be similar along the profile. If the halocarbons were derived from plant activity or soil organic processes, the vegetated sites should show elevated halocarbon concentrations. The results are presented in chapter 3 and clearly show that the strongest local source is volcanic degassing and other sources are insignificant.

The implications of these findings are that the global halocarbon source potential of volcanoes has to be established in order to more accurately quantify the volcanic source strength. Even though the quasi-continuous emissions of some halocarbons represented by volcanic activity

Chapter I. Introduction 5

are dwarfed by current anthropogenic emissions, this process has operated on geological timescales, creating a natural background burden that has not yet been taken into account for instance in sensitive underground water tracing techniques. The new techniques offered by this study allow the relatively easy determination of volcanic gas organic trace constituents by other research groups using standard analytical equipment.

Other implications of these results are to the question of the possible hydrothermal origin of life, to ore formation in reducing hydrothermal fluid mixing environments, as well as to volcanic surveillance, possibly offering a new set of redox sensitive “natural indicators” of subsurface processes.

Chapter I. Introduction 6

Chapter I. Introduction II. Halocarbons and other organic compounds in volcanic gases from Vulcano (Aeolian Islands, Italy)

“Nous voulûmes recueillir de l’air, mais [il] avait oublié le flacon. Il fallut qu’il redescendit pour le chercher. Cela dura longtemps vu la difficulté de monter sur le cone et cette attente nous encouragea d’examiner le cratère en dedans et de descendre Presque à son fond. (…) Nous descendîmes cinq personnes, Bonpland, M Espelde, moi et les deux Indiens don’t l’un portait le baromètre, l’autre le flacon rempli d’eau pour cueillir d’air. (…) Nous recueillions l’air, c’était l’expérience la plus interessante et par laquelle je désirais surtout de descendre dans le cratère. Manque d’oxygène, abondance d’acide carbonique.”

A. v. Humboldt on Jorullo volcano (Mexico) 1803 (von Humboldt, 1803) 8

Abstract

Sorbent-trapped volcanic gases, sublimates and condensates from active vents of the La Fossa crater on the island of Vulcano (Aeolian Islands, Italy) as well as ambient and industrial air were quantitatively analyzed by Short-Path Thermal Desorption-Solid Phase Microextraction-Cryotrapping-Gas Chromatography/Mass Spectrometry (SPTD-SPME-CF- GC-MS). Among the well over 100 detected and quantified compounds are alkanes, alkenes, arenes, phenols, aldehydes, carboxylic acids, esters, ketones, nitriles, furans, PAH's and their halogenated, methylated and sulfonated derivatives, as well as various heterocyclic compounds including thiophenes. Most compounds are found at concentrations well above laboratory, ambient air, adsorbent and field blank levels. For some analytes (e.g., CFC-11,

CH2Cl2, CH3Br), concentrations are up to several orders of magnitude greater than even mid- latitudinal industrial urban air maxima. Air or laboratory contamination is negligible or absent on the basis of inert gas and noble gas isotopic measurements.

The organic compounds are interpreted as the product of inorganic, abiogenic gas-phase radical reactions. On the basis of isomer abundances, n-alkane distributions and the substitution pattern the compounds are considered to have been formed by high temperature (e.g., 900 °C) alkyl free radical reactions and halide electrophilic substitution on arenes, alkanes and alkenes.

Model global volcanic halocarbon fluxes from our own and published compositional data on a number of halocarbons have been obtained by scaling to published volcanic CO2 fluxes. This -9 -1 -4 -1 yields global halocarbon fluxes in the range from 1.0x 10 Tg y (CClF3) to 1.2 x 10 Tg y

(CH3Cl). Methyl bromide (CH3Br) and methyl iodide (CH3I) are estimated to have a volcanic model source strength of up to 1.3 x 10-6 Tg y-1, however, it is noted that other brominated compounds were identified. The significance of organic emissions from quiescent volcanic degassing as opposed to explosive eruptions to stratospheric chemistry is that the negligible aqueous solubility of most emitted organic compounds does not lead to a quantitative tropospheric washout effect as postulated for explosive volcanic HCl and HF emissions.

Simple modeling suggests that episodic, catastrophically large, explosive events such as flood basalt eruptions do not impact the short-term ozone chemistry to the same magnitude as

Chapter II. Halocarbons in volcanic gases 9 today’s anthropogenic emissions but may modify the natural atmospheric reactive balance over much longer (geological) time spans. In addition, the apparent abiogenic organic chemistry of volcanic gases may give insights into the possible hydrothermal origin of early life on Earth, as indicated by the presence of simple amino acids, nitriles, and alkanoic acids.

1. INTRODUCTION

UNDERSTANDING the occurrence, abiogenic formation and emission of organic compounds by volcanic gases has major implications for the interpretation of a great variety of natural phenomena, such as atmospheric organic chemistry and the “ozone hole” (Molina and Rowland, 1974), theories of a hydrothermal origin of life (e.g. Zolotov and Shock, 2000), the formation and alteration of ore deposits (Giże, 1999), and the availability and stability of organic ligands involved in metal transport in ore-forming hydrothermal fluids (Seward and Barnes, 1997). Irrespective of the ultimate source of the atomic constituents (i.e., C, H, N, O, and S), the conditions for abiogenic formation of organic molecules are potentially met in any active volcanic-hydrothermal system.

The origin of the carbon in hydrothermal systems is usually explained either by mantle degassing (primordial source, e.g. Mathez et al., 1984) or pyrolysis of subsurface organic matter (thermogenic source, Schoell, 1988). These natural volcanic “synthetic laboratories” not only produce a great variety of inorganic solid phases and volatile molecular species, but also provide reactive environments for extensive organic synthesis. Nearly the entire periodic table of elements is present at measurable concentrations and the major volatiles, carbon dioxide and water are present as liquid, vapor, gas or supercritical fluid. Furthermore, the availability of reactive and catalytic surfaces, extreme temperature gradients and variable redox potentials allow many simple reactions such as the Fischer-Tropsch synthesis to occur. Consequently, an understanding of abiogenic organic synthesis from simple inorganic carbon compounds (CO2, CO, CH4, CS2, OCS) together with hydrogen, water, halogen acids, ammonia and sulfur compounds under hydrothermal conditions, requires more than theoretical considerations, but first and foremost, reliable analytical data from active volcanic, hydrothermal and geothermal environments.

Chapter II. Halocarbons in volcanic gases 10

We present results of an ongoing study on the organic chemistry of volcanic gases using a newly developed quantitative methods for the collection and analysis of organic compounds from volcanic gases. A brief overview of the historic and relevant literature is given first, followed by our own data from Vulcano. A compilation of all significant literature data is used finally to derive a comprehensive quantitative global volcanogenic halocarbon emission inventory model. The global environmental significance of continuously quiescent volcanic degassing to stratospheric chemistry is that the negligible aqueous solubility of most emitted organic compounds does not lead to a quantitative tropospheric washout effect as has been postulated for volcanic HCl and HF emissions.

Episodic, large, explosive volcanic eruptions rapidly penetrate the otherwise relatively undisturbed tropopause, facilitating immediate and quantitative injection of contaminants into the stratosphere, including labile, short-lived compounds which normally never reach such altitudes. The significance of the emission of organic compounds from major eruptions and quiescent volcanic degassing has so far been neglected in environmental and atmospheric impact studies. New evidence now testifies to the importance of previously unrecognized natural sources of environmental pollutants to climate modeling, global warming, and ozone hole prediction.

Chapter II. Halocarbons in volcanic gases 11

2. PREVIOUS WORK

Since the earliest days of volcanic gas studies, simple organic compounds were recognized in volcanic gases. An involvement of organic matter in magmatic processes was already suspected by von Humboldt (1803) and Gay-Lussac (1823). Organic compounds in volcanic gases other than methane have been sought, inferred or reported by a number of scientists in the nineteenth century (Davy, 1828; Bunsen, 1847, 1857; Sainte-Claire Deville, 1856b, 1856c, 1856a, 1865; Fouqué, 1865b, 1865a, 1866a, 1866b, 1879; Gorceix, 1872a, 1872b, 1872c; Petschek, 1911; Hempel, 1915). Modern gas sampling and analytical techniques were introduced by the studies on Kilauea lava lake and Mauna Loa fumaroles (Hawaii), by Day, Jaggar, Shepherd, Perret and Brun (Day and Shepherd, 1913; Perret and Brun, 1915; Shepherd, 1919, 1920, 1921), who suspected hydrocarbons in the gases and found bituminous residues in the volatile fraction extracted from fresh lava. A similar phenomenon was found by Allen and Zies (1923) and Allen and Day (1927) for the Katmai (Alaska) and The Geysers (California) fumaroles.

From about 1935 to 1970, organic compounds in volcanic gases were not recognized (Noguchi, 1935; Chaigneau et al., 1970) but after that period, an increasing number of studies appeared. For low temperature subaerial hydrothermal fluids, a considerable number of studies have been published and subsequently reviewed by Darling (1998). For submarine hydrothermal systems, an increasing number of studies have appeared recently (Simoneit, 1993; Holloway and O'Day, 2000; Kelley and Früh-Green, 1999, 2000, 2001). In lower temperature volcanic hydrothermal systems, reports are available from Chaigneau et al. (1970), Pavlov and Karpov (1972), Giggenbach et al. (1990), Capaccioni et al. (1993), Capaccioni et al. (1995), Kiyoshu and Asada (1995), Higashihara et al. (1997) and Capaccioni et al. (2001). Kiyoshu and Asada (1995) distinguished the organic signatures in volcanic gases between low-temperature “volcanic geothermal thermogenic” and high-temperature “fumarolic type of inorganic origin” sources. Their low temperature type is equivalent to the emissions reviewed by Darling (1998). These low temperature sources are characterized by excess CH4 (similar to sedimentary basin gases), elevated normal/iso-ratios for butane and pentane, high amounts of aromatics and naphthenes (Capaccioni et al., 1993; Capaccioni et al., 1995; Darling, 1998), and ethylene depletion relative to n-butane (Kiyoshu and Asada, 1995). High temperature organic volcanic gas emissions are typically ethylene-enriched over

Chapter II. Halocarbons in volcanic gases 12 n-butane and are generally alkene-rich, becoming more saturated (hydrogen attraction on expense of double bonds) in post-eruptive emissions (Capaccioni et al., 1993; Capaccioni et al., 1995). In all cases, the redox potential of the fluid seems to control the amounts of hydrocarbons rather than temperature (Kiyoshu and Asada, 1995).

Organic compounds in gases emanating from lava flows and lakes were reported by Finlayson et al. (1968), Wilkniss et al., (1975), Markhinin et al., (1977), Graeber et al., (1979), Rasmussen et al., (1980), Markhinin et al., (1983) and Menyaylov et al., (1983). Experimental degassing of lava yielded organic compounds in the studies of Chaigneau (1975), Markhinin and Podkletnov (1977b, 1977a, 1978), Bondarev and Porshnev (1980), Podkletnov and Markhinin (1981) and Heide et al. (2000). During and shortly after explosive volcanic eruptions, organic compounds and sometimes halocarbons were reported from airborne plume sampling (Inn et al., 1981; Leifer et al., 1981; Cronn and Nutmagul, 1982a, 1982b; Leifer et al., 1984), and desorption of freshly fallen ash (Markhinin et al., 1975; Mun and Kuziakina, 1976; Markhinin and Podkletnov, 1977b, 1977a, 1978; Florovskaia et al., 1978; Kolesnikov and Egorov, 1979; Kolesnikov et al., 1979; Kravtsov et al., 1979; Ponomarev et al., 1979; Bondarev et al., 1980; Mun and Rogova, 1980; Pereira et al., 1980; Podkletnov and Markhinin, 1981; Rasmussen et al., 1981; Bondarev et al., 1982; Pereira et al., 1982; Rasmussen et al., 1982; Pereira and Rostad, 1983; Andreichikov et al., 1985; Porshnev and Bondarev, 1989).

Hydrocarbons, including in some cases halocarbons, were reported in high-temperature volcanic gases from Asama (Japan) by Noguchi (1935), Navado del Ruiz (Colombia) by Giggenbach et al. (1990), the Kamchatkan and Kurile volcanoes (Mendelev, Golovnin, Tyata, Mutnovskiy, Kikhpinych, Dolina Geyzerov, and Uzon) by Isidorov (1985, 1990, 1992), Isidorov et al. (1985), Isidorov and Zenkevich (1985), Isidorov and Ioffe (1986), and Isidorov et al. (1990); Kudriavy (Kuriles) by Wahrenberger et al. (1998), Etna (Italy) by Jordan et al. (2000), Galeras and Puerace in Colombia and Poas in Costa Rica (Capaccioni et al., 1993; Capaccioni et al., 1995), Mt. Erebus (Alaska) by Rasmussen and Rasmussen (1980) and Penkett et al. (1981), Masaya (Nicaragua) by Isidorov (1992), as well as at St. Augustine (Alaska), Teide (Canary Islands), and Krisuvik (Iceland) by Rasmussen and Rasmussen (1980). The fumaroles on Vulcano (Italy) have been analysed for organic compounds by Chaigneau et al. (1970), Capaccioni et al. (1993, 1995), Wahrenberger et al. (1996a), Wahrenberger et al. (1998), Schwandner et al. (2000a, b, 2001a, b).

Chapter II. Halocarbons in volcanic gases 13

Halocarbons were described for the first time in volcanic gases by Stoiber et al. (1971) who expressed doubts as to the validity of their data. Most of these studies have been interpreted to report contamination or mixing phenomena with ambient air, in some cases by the authors themselves (Stoiber et al., 1971; Wilkniss et al., 1975; Rasmussen et al., 1981; Rasmussen et al., 1982; Fink, 1983; Pereira and Rostad, 1983; Gaffney, 1995; Jordan et al., 2000). Equilibrium thermodynamic approaches to explain the presence of simple organic compounds in volcanic gases were presented by Symonds et al. (1988), Symonds, (1990), Symonds et al., (1990), Fegley, (1991), Wahrenberger, (1997), Shock and Schulte, (1998), Zolotov and Fegley, (1999), Fegley and Zolotov, (2000) and Zolotov and Shock, (2000).

Organic compounds of presumed magmatic origin have been reported from fluid inclusion studies on igneous complexes by Sobolev et al. (1974), Petersilie and Pavlova (1978) and Salvi and Williams-Jones (1997). Deep crustal rocks and borehole fluids were also shown to contain hydrocarbons (Isidorov, 1990; Mogk and Mathez, 2000). Carbon compounds have also been reported from mantle xenoliths and kimberlites (Deines, 1968; Chaigneau, 1975; Mathez and Delaney, 1981; Mathez et al., 1984; Mathez, 1987; Tingle et al., 1990; Tingle et al., 1991; Krot et al., 1993; Tingle and Hochella, 1993; Sugisaki and Mimura, 1994; Zubkov, 2001).

The aim of this study has therefore been to study the carbon-based (“organic”) chemistry of volcanic gas emissions at Vulcano (Aeolian Islands, Italy). We have developed novel sampling and gas chromatographic-mass spectrometric (GC-MS) analytical techniques in order to reliably identify many of the organic components present in the gases with a special emphasis on the halocarbon compounds. In addition, considerable care has been taken to either eliminate and/or quantify ambient air entrainment during sampling in order to obtain uncontaminated, pristine halocarbon concentrations of the gases emanating from the high- temperature crater fumaroles at Vulcano.

Chapter II. Halocarbons in volcanic gases 14

3. VOLCANIC IMPACTS TO THE ATMOSPHERE

Up to the present, reports on the volcanic impact on the Earth’s atmosphere are by large limited to the investigation of large-scale aerosol loading as particulates and SO2 into the lower stratosphere by large Plinian eruptions such as El Chichón (Mexico) in 1982 (reaching 25 km; D'Altorio and Visconti, 1983), and Pinatubo (Philippines) in 1991. El Chichón injected 8 Tg of sulfate directly into the stratosphere (in total 12 Tg aerosol) and Pinatubo emitted 20 - 30 Tg (McCormick et al., 1995). However, trace metal and trace organic emissions during explosive as well as quiescent fumarolic degassing activity have direct consequences on environmental health, the Earth’s climate and the protective ozone layer. The major atmospheric effects of volcanic activity are global warming and ozone depletion which are both facilitated by aerosol loading as well as organic compound emissions. In this section, we give a brief overview of the principles and opinions that constitute current scientific views on the impact of volcanic activity on the Earth’s ozone layer.

3.1. Global warming

The principal mechanism of global warming of the Earth’s atmosphere is the greenhouse effect caused by infrared (IR) radiation emitted from the Earth’s surface being absorbed by gaseous species and occlusion by clouds. This effect ensures an average global surface temperature of +14°C, instead of –18°C, as would be theoretically the case if the atmospheric heat budget was only governed by solar irradiance. The difference of 32°C is caused by naturally occurring greenhouse gases, such as H2O, CO2, CH4, O3, and O2 (Isidorov, 1990). Water in the form of clouds has a dual effect in enhancing the greenhouse effect by dispersion and backscattering of radiation at their lower surfaces but also by reduction of the warming by efficiently dissipating thermal energy on their upper surface through convection. One natural external modification to the atmospheric thermal energy balance is the intensity of solar UV- radiation that the planet receives, a value which is not constant and which has apparently increased significantly over the last century by a doubling of the Sun’s magnetic field (Lockwood et al., 1999), coupled with a doubled sunspot activity since 1870 (Foukal and Lean, 1996).

Chapter II. Halocarbons in volcanic gases 15

Figure 1 In this schematic, the atmospheric effects of active volcanoes are summarized as a model: a) quasi-steady quiescent degassing explosive volcanic eruptions (modified after Schwandner et al., 2001a); halocarbon compounds emitted from quiescent fumarolic and diffuse volcanic degassing activity partition according to their solubilities into the condensing steam clouds. Much alike industrial stack gas emissions, these compounds get dispersed in atmospheric circulation leading to tropospheric homogenization. Eventually transported to equatorial regions, Hadley convection carries them to the raised tropical tropopause, where they can get mixed into the lower stratosphere. At first mostly carried by the jet stream, they may soon reach the polar vortex where they can be a threat to the ozone layer. This transport process takes from weeks to up to over a year, mostly depending on latitude of the emission source; b) large explosive volcanic eruptions (modified after McCormick et al., 1995) are transient events that can penetrate the tropopause to directly inject material into the stratosphere, from where it is globally dispersed within just a few days. It can persist in the stratosphere for several years, initially as water vapor, ultrafine salt crystals, ash, SO2 and HCl/HF. Most HCl and HF dissolves in supercooled water droplets condensing on rapidly nucleating aerosols from sulfuric acid vapor (e.g., Tabazadeh and Turco, 1993), and rains out. SO2 is converted into H2SO4 under consumption of OH-radicals, the lack of which results into a negative feedback on CO and CH4 breakdown, causing an additional increase in abundance of these two greenhouse gases. H2SO4 condenses to an aerosol, decreasing atmospheric optical depth and thus increasing the earth’s albedo by reflecting incident solar radiation back into space – leading to tropospheric cooling. In addition, infrared radiation emitting from the troposphere and the earth’s surface is absorbed leading to a warming of the stratosphere. By the increased availability of aerosol particle surfaces in the stratosphere, heterogenous reactions lead to a release of halogens that are capable of entering the cyclic radical reactions of ozone destruction similar to the Antarctic ozone hole, where polar stratospheric clouds supply the necessary surfaces. Eventually, the aerosol particles sink into the troposphere where they modify cirrus cloud optical properties and continue to alter the radiative balance as described above.

The effect of a volcanic disturbance on the natural aerosol budget was recognized by Arrhenius (1896), Chamberlin (1897), and indirectly by Benjamin Franklin in 1783 (Franklin, 1784; Tilling, 1982). Physical and chemical atmospheric impact studies to date incorporate volcanic phenomena in climate models predicting global warming and uniformly conclude that volcanic eruptions counteract global warming. They suggest this counteraction to be due to an aerosol-loading induced increase in the planetary albedo and therefore occluding and scattering of incident solar UV radiation (e.g. Vogt, 1972; Sassen, 1992; Kondratev and Varotsos, 1995b, a). This effect has been discussed in detail since the beginning of the

Chapter II. Halocarbons in volcanic gases 16 twentieth century, following the Katmai 1912, Pelee 1902 and Krakatau 1883 eruptions (Arctowski, 1915).

Volcanic aerosol is held responsible for interrupting global warming trends in a similar fashion as predicted in “nuclear winter” scenarios, leading to a decreased global mean surface temperature of over 1°C for several years. This was observed after the Tambora 1815 eruption which was followed by the “year without summer” in 1816, with global crop failure and decreased optical depths for several years (Kerr, 1989; Vupputuri, 1992). Volcanic aerosol- induced tropospheric cooling will lead to a significant short-term disturbance of the carbon cycle by reduced plant and soil respiration and increased tropical productivity, resulting into notable carbon fixation by the terrestrial biosphere for several years post-eruption (Jones and Cox, 2001).

The main greenhouse gases are CO2, H2O, CH4, CO as well as other trace organic compounds, including halocarbons. The total anthropogenic emission of the major greenhouse -1 gas, CO2, is estimated to be 29,200 Tgy (IPCC, 2001). Large short term perturbations are known to have occurred in the past (Kuypers et al., 1999). In comparison, ca. 66 Tgy-1 is estimated to be released from subaerial volcanism (Gerlach, 1991), excluding hydrothermal emissions (Kerrick et al., 1995). The pre-industrial abundance was about 280 ppmv but is currently (1990) 369 ppmv (IPCC, 2001). Regulation of anthropogenic emissions of the strong greenhouse gas CO2 are currently in the state of international negotiation, in an attempt to modify (USA) or ratify (EU) the Kyoto Protocol.

The radiative forcing potential of a greenhouse gas is usually expressed as the Global

Warming Potential (GWP) relative to CO2 and is specific to a defined time horizon during non-steady state conditions (i.e. during industrial emission). Representative GWP values are listed in Table 4. Prominent greenhouse gases emitted by subaerial quiescent and eruptive volcanic activity are methane (GWP100=11), water vapor, and carbon monoxide. Their annual -1 volcanic mass fluxes correspond to 0.34 and 0.02 Tgy for CH4 and CO, respectively (Cadle, 1980). For water vapor, incorporation of the effect of volcanic aerosol clouds in troposphere- stratosphere transfer models show that tropospheric heating substantially increases the vertical transport of this strong greenhouse gas (Considine et al., 2001). After the Pinatubo 1991 eruption, an increase in CH4 and CO in the atmosphere was observed and attributed to a depletion effect in atmospheric OH caused by the conversion of sulfur dioxide to sulfuric acid

Chapter II. Halocarbons in volcanic gases 17

(Cadle et al., 1979), a process that represents over 80% of the sinks for atmospheric CH4 and CO (Law, 1999).

Many of the trace constituents of volcanic gases have an appreciably greater GWP than the major species H2O, SO2 and CO2. In some models, these trace gases account for up to 65% of the cumulative surface warming effect if extrapolated to the year 2080 (Vupputuri et al., 1995). The relatively long-lived greenhouse gases, CFC-11 and CFC-12 (chlorofluorocarbons

= CFC’s), are the second largest non-H2O contributors to global annual mean temperature increases after CO2 (Schönwiese, 1993). The halocarbons, methyl bromide and methyl chloride are emitted from volcanoes (Rasmussen and Rasmussen, 1980; Schwandner et al.,

2001) with CH3Cl being the major halocarbon compound (Gerlach, 1980). Both are considered to play only a minor role as greenhouse gases even though their GWP’s are similar to CH4, because their anthropogenic emission rates are too low compared to CH4 (Grossman et al., 1997).

3.2. Ozone depletion

Halogen emissions from volcanoes are a potential threat to the Earth’s protective stratospheric ozone (O3) layer which shields the surface from intense ultraviolet radiation. Their ability to reach the stratosphere, however, has for some time been a subject of controversial discussion (e.g., Cadle, 1975; Johnston, 1980; Russel III et al., 1996). Crutzen (1970) had originally postulated that nitrogen oxides (NOx) derived from atmospheric N2O react catalytically with

O3, leading to an acceleration of ozone destruction. The first discussion on the potential of volcanic eruptions and quiescent volcanic degassing to emit quantities sufficient to cause ozone destruction resulting in a localized ozone hole was by Stolarski and Cicerone (1974). Shortly after and on the basis of remote air analyses, experimental work and computational modeling it was suggested that chlorofluorocarbons were responsible for catalytic ozone destruction by chlorine radicals in the stratospheric ozone layer (Lovelock, 1971; Molina and Rowland, 1974). For their pioneering work, Rowland, Molina and Crutzen received the Nobel Price in Chemistry in 1995.

In 1984, a clear decline of ozone over Halley Bay, Antarctica, had been observed to reach 40%, a phenomenon that became known as the Antarctic Ozone Hole (Farman et al., 1985). International agreements have since then been successfully implemented, requiring the

Chapter II. Halocarbons in volcanic gases 18 participating nations to phase out the use, production and sale of ozone-depleting substances (ODS) in accordance with the Montreal Protocol (UNEP, 2000). The identification of an increasing number of ODS since, including earlier replacements for CFC’s has led to several amendments to the Montreal Protocol (London 1990, Copenhagen 1992, Vienna 1995, Montreal 1997, Beijing 1999), which are successively and at different rates per nation being ratified.

In parallel, a great number of non-anthropogenic natural sources of chlorine in the stratosphere have also been identified (see Gribble, 1994c; Gribble, 1998; Khalil et al., 1999; for recent reviews). Rigaud and Leroy (1990) provided evidence that the “ozone hole” already existed over Antarctica but was not centered at the polar vortex in 1958, when anthropogenic emissions of ODS were minor. These authors suggested a natural, perhaps volcanic, phenomenon to be responsible for the seasonably variable polar ozone densities (e.g., Angell and Korshover, 1964). Natural sources of ODS are a growing field of interest, since current budgets of several ODS are not balanced by anthropogenic emissions (Butler, 2000).

3.2.1. The volcanic contribution of halogens to the atmosphere

At the early stages of the CFC/ozone discussion stood the question, how much halogen in the stratosphere could be derived from volcanic eruptions (Ryan and Mukherjee, 1973, 1975; Stolarski and Cicerone, 1974; Cadle, 1975; Cicerone, 1975). These authors agreed that the threat to stratospheric ozone would, apart from rare large explosive volcanic eruptions, be largely from anthropogenic CFC’s, and that volcanic injections of halides were of a transient, short term nature lasting at most 2 years (Cadle, 1975). In reviewing the major perturbing influences on stratospheric ozone between 1955 and 1975, Bauer (1979) considered several major explosive eruptions, including Agung (Bali) in 1963, which represents the second largest injection of volcanic material into the stratosphere in the twentieth century, the first being Pinatubo in 1991 (McCormick et al., 1995).

After the El Chichón and the Mt. Pinatubo eruptions, a substantial ozone depletion by heterogeneous (aerosol particle surface) reactions involving sulfate particles and its dependence on the (ClONO2) - sulfate reaction rate have been suggested (Vupputuri, 1985; Hofmann and Solomon, 1989; Hofmann et al., 1992; Prather, 1992). Blanchette and McConnell (1994) demonstrated the dependence of ozone depletion on the reaction

Chapter II. Halocarbons in volcanic gases 19 probability of heterogeneous chemical reactions in the stratosphere and on the dramatic increases in reaction probability during times of volcanic aerosol loading. These authors interpreted this effect to be leading to an actual “volcanogenic” ozone depletion, not so much because of volcanogenic halogen loading but because of NOx emission (cf. Roscoe, 2001).

Ozone depletion was globally measured after the 1963 Agung (Bali), 1982 El Chichón and 1991 Pinatubo eruptions (Angell and Korshover, 1973; Angell et al., 1985; Solomon et al., 1993; Tie et al., 1994, and references therein). The increased chlorine dioxide after the Pinatubo eruption was not thought to be of volcanic origin but derived from pre-eruption stratospheric HCl and ClONO2. Ozone losses amounted to up to 25% between 40°-50°N at 20 km altitude in winter and spring in the lower stratosphere (Hofmann et al., 1994; McGee et al., 1994). On the basis of model simulations for Pinatubo and El Chichon, Tie and Brasseur (1995) concluded that the deviation from the expected effect of an actual ozone increase is only possible under the present conditions of increased anthropogenic chlorine loading and is directly linked to the amount of volcanic SO2 in the stratosphere. In contrast, emissions from high-altitude aircraft are believed to actually increase stratospheric ozone, especially after large volcanic eruptions that produce stratospheric aerosol (Tie, 1994). Volcanic bromine emissions have been suggested as an alternative reason for ozone depletion following eruptions, since Pinatubo may have emitted an amount equal to or larger than 0.011-0.025 Tg of Br (Bureau et al., 2000). A thorough discussion on the volcanic role on stratospheric chemistry is given in a recent review (Solomon, 1999).

In this context, it is interesting to remember that volcanoes also emit hydrocarbons (Stoiber et al., 1971; Muenow, 1973; Markhinin et al., 1983; Isidorov et al., 1990; Schwandner et al., 2000a, b). The effects of volcanic activity on the atmosphere are summarized in Figure 1. For an introduction to the chemical mechanisms of ozone depletion, see the reviews by Crutzen and Lielieveld (2001) and Staehelin et al. (2001). The global atmospheric halogen budgets were reviewed by a number of research groups, governmental agencies, the United Nations, and various interest groups (e.g., Cicerone, 1981; CMDL, 2001; Keene et al., 1999; WMO, 1998). Only a brief overview with respect to volcanic emissions is given here. Halogens may be emitted from volcanoes as volatile salts (ultrafine particles), aerosols, halides (including acids), molecular gases (i.e., Cl2 and F2), or as organic molecules (halocarbons). The volcanic source strength of gaseous molecular chlorine has been assessed by Ryan and Mukherjee (1975).

Chapter II. Halocarbons in volcanic gases 20

3.2.2. Halides

The term halide is used here for inorganic halogenides. The halide emission from volcanoes, though significant compared with the anthropogenic output, has been considered negligible due to the tropospheric wash-out effect by which HCl molecules in volcanic ash plumes dissolve in supercooled water droplets condensing on rapidly nucleating aerosols from sulfuric acid vapor (e.g., Cadle, 1975; Tabazadeh and Turco, 1993) and subsequently rain out before reaching the stratosphere. In addition, HCl is not photolyzed in the ultraviolet range. Exceptions to this removal mechanism may be large explosive Plinian volcanic eruptions that penetrate the tropopause. Such episodic injections into the stratosphere lead to a large but transient, short-term aerosol loading that in the case of the Mt. Pinatubo eruption in 1991 has been observed to decay within a few years (1995) to pre-eruptive levels (SAGE II experiment, WMO, 1994). HCl and HF reach the stratosphere in significant amounts as has been shown by remote Fourier transform spectrophotometric analyses, documenting that short-lived pronounced maxima from volcanic ash clouds are superimposed on the anthropogenic increase in stratospheric HCl and HF loading (Wallace and Livingston, 1991, 1992).

HCl and HF emissions from the active volcano Mt. Erebus have been quantified by Zreda- Gostynska et al. (1993, 1997). Emission rates between 1983 and 1991 vary from 2-30.5 Ggy-1 and 4-60.9 Ggy-1 for HF and HCl, respectively. St. Augustine volcano (Alaska) alone emitted 2920 Ggy-1 HCl in 1986 (Symonds et al., 1988), comparable to 3000 Ggy-1 anthropogenic HCl in the late 1980’s (WMO, 1990). Observed volcanic trace metal anomalies in ice cores can be explained by metal transport as volatile salts in the aerosol fraction of a volcanic eruption plume, as is indicated by experimental and field studies (Cadle et al., 1979; Lazrus, 1979; Smith et al., 1982; Phelan, 1983; Ammann, 1992; Zreda-Gostynska et al., 1997; Fischer et al., 1998).

Greenland ice core layers have elevated syndepositional fluorine levels originating from quasi-continuous fumarolic volcanic HF emissions at mid-latitudes (Preunkert et al., 2001) in contrast to high altitude alpine ice cores which block gas exchange with the atmosphere much faster than great inland ice sheet deposits (i.e. Greenland or Antarctica) because of higher precipitation rates. Sturges et al. (1996) showed that vertical diffusion of halocarbons occurs in the uppermost 0.5 m in alpine firn, leading to blurred signals but preservation of long-term trends. The detection limits of 1-5 pptv for halocarbons in air bubbles entrapped in ice may

Chapter II. Halocarbons in volcanic gases 21 still preclude identification, confirmation and quantification of natural sources, as stressed by Khalil and Rasmussen (1989), who conclude that “the possibility that small amounts of these chlorocarbons were in the pre-industrial atmosphere cannot be completely ruled out based on ice core analyses”.

3.2.3. Halocarbons and their emission sources

Halocarbons derive from a variety of anthropogenic and natural sources. Many important atmospheric trace gases that are emitted by volcanism (explosive eruptions, lava degassing, quiescent fumarolic activity) as well as by anthropogenic release have limited aqueous solubilities and are hence less affected by tropospheric wash-out effects (e.g., Johnston, 1980). Anthropogenic use of halocarbons is primarily as industrial solvents and in the chemical industry; as refrigerants (freons), as fumigants in agriculture (especially methyl bromide), as foaming agents and insulators, in air conditioners (home, industry and vehicles), and as fire retardants in fire extinguishers (halons).

Figure 2 This figure (modified after McLinden and Didion, 1987) displays the dilemma of choice of appropriate replacement alternatives for controlled substances. Chlorine substitution tends to increase a compound’s ODP (ozone depletion potential, see text), while fluorination increases its GWP (global warming potential, see text). HCFCs may be better alternatives in terms of a compromise between ODP and GWP, even though they are currently considered the “lesser evil”. To not fully substitute hydrogen for a halogen moves a compound (e.g., HCFC’s) closer into the field of increased flammability, which is a hazard in refrigation and an impossibility in fire retardants. Halons are not displayed, but any brominated compounds will have a much larger ODP than CFC’s, HCFC’s and PHC’s. Alternative refrigants and propellants currently promoted are volatile hydrocarbons (butane, propane), however they are flammable. The open circles display compositions of halocarbons found in volcanic gases (from this study only). Shaded fields after (McLinden and Didion, 1987), F=flammable, t = toxic, fh = fully halogenated (thus a long atmospheric lifetime). GWP and ODP trends are after (Calm and Didion, 1998).

Chapter II. Halocarbons in volcanic gases 22

To understand the problem of halocarbons in industrialized society, it is neccessary to understand the properties of this class of compounds that makes them so attractive for use in a variety of applications. For use in refrigeration and fire extinguishing, few compounds have the appropriate stability, expansivity, molecular mass and are inflammable in order to serve their industrial purpose (Figure 2). The requirements of low flammability, low Ozone Depletion Potential (ODP) and low GWP are mutually exclusive (McLinden and Didion, 1987).

3.2.4. Atmospheric fate of halocarbons

The longer the atmospheric lifetime and the more halogens contained in the molecule, the greater its Ozone Depletion Potential (ODP), reported relative to the reference compound CFC-11 (see Table 4). The ODP of a compound decreases by substitution of the different halogens from I>Br>Cl>F with iodinated compounds having the largest ODP. Ultraviolet active compounds can be photolyzed in the stratosphere to release halogen radicals that can enter the catalytic radical reaction cycle held responsible for the destruction of up to 10,000 ozone molecules per chlorine atom. Among these compounds are many halocarbons, including CFC’s (chlorofluorocarbons), HCFC’s (hydrochlorofluorocarbons, which are common CFC replacements), and halons (bromine containing mixed halocarbons), perhalocarbons (all hydrogens substituted by halogens of one element) and sulfur halides (e.g., sulfur hexafluoride).

In order to understand the reactive halogen mass balance of a given compound, several factors must be taken into account. These include its atmospheric lifetime (dependent on sink and source strengths, and photolysis rate), its source strengths (emission rates), sink strengths (removal rates, e.g. ozone destruction, rain-out, fixation as salts) and atmospheric burden (total mass in the atmosphere). In the Earth’s atmosphere, the lowermost layer (troposphere) is relatively well mixed and compound concentrations do not vary much in vertical profiles. Transport into the stratosphere is possible mainly in the tropics in uprising air masses of the Hadley convection. Once in the stratosphere, compounds are rapidly transported polewards, initially by the jet stream, and dispersed. Aerosol surfaces of polar stratospheric clouds (PSC) serve as reaction surfaces for ozone destruction leading to the observed ozone depletion at the polar vortex (e.g., Webster et al., 1993; see also Figure 1).

Chapter II. Halocarbons in volcanic gases 23

Ice nuclei forming at the extremely cold antarctic temperatures in PSC inhibit fixation of reactive chlorine as soluble HCl, thus increasing the availability of it for ozone destruction. Tabazadeh et al. (2002) showed, that during times of increased greenhouse gas loading in the troposphere that lead to a lowering of lower stratospheric temperatures, volcanic aerosols from major future explosive eruptions will increase the current levels of Arctic springtime ozone loss by up to 70%, independent of NOx chemistry. These effects will presumably be even larger for large explosive eruptions occurring at higher northern latitudes, such as in Alaska, Kamchatka, or in the Mediterranean (Santorini or Nisyros, Greece; or Vesuvius, Italy) with their potentially high halogen output from involvement of Messinian evaporites, .

A decrease in the atmospheric burden of some species has been observed since 1994 and interpreted as a result of regulation (i.e., the Montreal Protocol), and is predominantly due to decreasing trichloroethane (C2H3Cl3) emissions (Montzka et al., 1999). The sum of natural and anthropogenic emissions of other halocarbons has not shown such a dramatic decrease but the trend is already detectable for CFC-11, CFC-113 (C2Cl3F3) and CCl4 (Walker et al.,

2000), all of which started leveling off around 1991. Of those, only CCl4 showed a significant decrease. The natural occurrences of chlorinated compounds have been reviewed by Gribble (1992a, 1994d, 1996a, b, 1998), Khalil et al. (1999), Butler (2000) and Winterton (2000).

Chapter II. Halocarbons in volcanic gases 24

4. EXPERIMENTAL

Two major reasons for the lack of studies on volcanic organohalogen emissions have been the difficult sampling conditions (loss and reaction of compounds by photolysis, catalysis, oxidation, and reaction with condensate), and the complex analytical problems due to severe, adverse matrix effects (e.g. water, sulfur, and acids) during chromatographic separation of trace organic compounds, and their identification by gas chromatography-mass spectrometry (high sulfur, water and mineral acid backgrounds shielding and eliminating organic signals). We have developed a versatile method for the sampling and analysis of hydrocarbon and halocarbon compounds from volcanic-hydrothermal systems: i.e. Short-Path Thermal Desorption-Solid Phase MicroExtraction-CryoFocussing-High Resolution Gas Chromatography-Mass Spectrometry (SPTD-SPME-CF-HRGC-MS).

During the course of this study, a variety of adsorbents, condensation and sampling devices, as well as different sample preparation, gas chromatographic separation, and different mass spectrometric analytical techniques have been tested. As a result, trapping efficiency, peak shapes, and background linearity have been greatly improved. The analytical developments in our study may be summarized as follows: 1) separation of gases from condensates during sampling, and 2), the application of Solid Phase Microextraction (SPME) prior to high- resolution GC-MS analyses. The use of SPME markedly reduces the high sulfur background, which previously prevented detection of many trace organic compounds, and minimizes the introduction of water and inorganic acids, which are deleterious to high resolution GC-MS. The principles of the method have been discussed briefly by Schwandner et al. (2000a; b).

4.1. Sampling and preparative work

Sampling was done by focusing fumarolic gas flow into a 0.5 m long quartz glass tube with perforations on the lower to enhance ingress of gases. The quartz glass tubes were inserted into the high gas flow fumaroles, allowed to equilibrate within the high-temperature gas stream overnight, and if necessary cemented with sulfur melt naturally forming at the fumaroles. The temperature profile through the focusing tube was measured before and after sampling, to ensure that no cooling by entrainment or mixing with air had occurred.

Chapter II. Halocarbons in volcanic gases 25

The sampling tube was then connected to an all-glass sampling train. Preparation of the glasware involved several steps of acid washing or –boiling, and thermal desorption. Transport of glassware was in desorbed (300°C for several hours) aluminum foil with Kapak® endcaps (SKC Inc., PA, USA) to seal the tube ends. The all-glass sampling train components were connected head-to-head by support of brass fittings with M4-ferrules (Supelco, Inc.), such that no graphite or metal surfaces were in contact with the gas stream. In addition to organic adsorption samples, bulk gas samples (Tedlar bags and evacuated glass bulbs), absorbent samples with 4 N NaOH (“Giggenbach-bottles”) and cold-sealing copper tubes for noble gas isotopic analyses were taken.

The acid, sulfur and water matrix was removed from the gas stream by means of a 2-stage ice- cooled condenser array that could be sealed off with vacuum-stopcocks (acid washed and thermally desorbed), to preserve the condensate with its headspace for later analysis. The sampling train was wrapped with aluminium foil to avoid possible photolysis of compounds. The final adsorption temperature was 4°C. The natural gas flow was supported by a low-flow battery-driven air sampling pump with inert wetted surfaces (SKC Inc., PA, USA) downstream from the adsorbent. The pump was operated at a calibrated rate in the range of the natural flow, which also provided information on the sampled gas volume. See appendix 2 for further detail.

Adsorbent tubes were stored frozen and sealed with Kapak® endcaps, and condensates at ~4°C in the dark prior to analysis. A variety and combination of adsorbents were employed. Most successful proved to be Florisil (BDH, UK), Anasorb CSC (SKC Inc., PA, USA), Carboxen 569, and Carbotrap C (both Supleco Inc., USA). Tenax™ TA (Supleco Inc., USA) was tested as well, but was too sensitive for trace water and halogen acids, resulting into strong column bleed, to be taken into further consideration. Conditioning of the tubes was done under a stream of purified helium 5 (with in-line oxygen and hydrocarbon traps), at 300°C for the molecular sieve and carbon-based adsorbents, and at a lower temperature for the dry Florisil Mg-Si-gel. Sufficient blanks were run to ensure purity of the process, including field blanks of identical adsorbent tubes analyzed under identical conditions.

Chapter II. Halocarbons in volcanic gases 26

1.2. Analysis

1.2.1. Desorption

The adsorbent tubes were thermally desorbed at 300°C (120°C for Florisil), and re-adsorbed onto a SPME fiber at 25°C. The SPME needle was thermally desorbed in the splitless inlet of a GC and cryogenically focused on the column head using liquid nitrogen in order to ensure a low baseline and maximize chromatographic separation. This technique significantly lowers interferences from water and sulfur-containing compounds which comprise up to 95 mol % of the total volcanic gas, the remainder being CO2, HCl, HF, N2, NH3,noble gases and organic analytes in the pptv to ppbv range. The desorbed compounds were separated on the analytical column and analyzed by full-scan single quadrupole MS. To release the adsorbates from the adsorbents, we used a simple custom-built direct thermal desorption purge and trap system, optimized to minimize sample loss and fractionation of analytes (Hall Analytical Ltd., UK), with external temperature control, flow control and all-glass contact surfaces (cf. Appendix 2, Figure 2). The same high-purity helium supply was used as in the conditioning of the adsorbent tubes. We found Solid Phase Micro-Extraction (SPME) to be the most efficient technique to preserve the original composition (recovery ratios) in samples eluated directly by short-path thermal desorption, and by headspace SPME of eluates cryotrapped in vials.

1.2.2. Separation and detection

Separation was achieved using a 100% crosslinked low-polarity-polydimethylsiloxane (PDMS) fused silica Hall-GC1 analytical column (equivalent to the standard PDMS phase OV-1) with the dimensions 60 m x 0.32 mm x 0.5 µm (Hall Analytical Ltd., UK). Filtered helium was the carrier gas, the column head pressure was 8 psi (flow rate of 0.54 mL/min, average linear velocity of 18.8 cm/s) yielding a theoretical helium retention time of 5:32. The SPME fiber was PDMS mixed with Carboxen 569 (Supelco Inc., USA). The oven temperature program commenced with an isothermal stage (40°C for 8 min), followed by a heating ramp of 4°C/min to 290°C, and an isothermal stage for 17 min at the final temperature. We used two different GC-MS systems for the analyses, either a Thermoquest Voyager GC-MS system coupling a Fisons GC 8060 with a Thermoquest Voyager MD 800

Chapter II. Halocarbons in volcanic gases 27 single quadrupole mass spectrometer, or alternatively a HP GC 5890A coupled with a Micromass Trio-2000 single quadrupole mass spectrometer. The acquisition software used was MassLab 1.3 and VG MassLynx 3.0 for NT, respectively. Both MS systems were operated in positive electron ionization mode (EI+) at 70 eV electron current, and a scan rate of 1 scan per second. Mass ranges were 15-600 a.m.u (atomic mass units or daltons) for the MD 800, and 20-549 a.m.u. for the Trio 2000. The transfer line temperature was 300°C.

1.3. Data treatment

1.3.1. Mass fragment analysis

GC-MS analysis combines two independent analytical methods, that of chromatographic separation together with mass and molecular fragmentation detection. Identification of compounds is by independent determination from both methods. The mass spectrum yields the fragmentation pattern with the resulting structural and functional discrimination and in most cases the molecular ion (M•+) is preserved, yielding the approximate molecular mass of the compound. The chromatographic retention time is expressed either relative to a ubiquitous early eluting peak (CO2 in this study), or as a retention index relative to a homologous series of compounds (e.g. n-alkanes: Kovat’s Index, see below). Retention indices are specific to a compound in all its isomers that the stationary phase separates but cannot be distinguished by their mass spectra and. The overall GC-MS method employed thus provides a more accurate identification of the compounds in a sample than one of the combined methods used alone. Figure 3 shows the total ion chromatogram (TIC) of sample FL2 (Vulcano crater, fumarole F5AT, 362.3°C, June 7, 1999), a system blank, an inset of an enlarged portion of the TIC trace with specific ion fragments of compounds, and their observed mass spectra.

1.3.2. Chromatographic retention indices

The chromatographic retention relates logarithmically to the number of carbon atoms for n- alkanes under isothermal separation conditions. The Kovat’s Retention Index system is used to linearize this relationship , making it relatively easy to target specific compounds of interest (Kovats, 1958). In linear temperature programmed separation, an approximate linear relationship is observed (Davankov, 2001) that can be described by the linear temperature

Chapter II. Halocarbons in volcanic gases 28 programmed retention index (LTPRI) system, which only differs from the Kovat’s index system by not being logarithmic. There is a difference between accurate experimental indices to observed ("apparent") retention indices in any environmental sample. A single analysis of a sample does not reproduce experimental retention indices for the compounds with the same accuracy, but they can be compared to a database of more accurate indices. The database gives an retention index window for each compound (the average of all literature values and its standard deviation) that can be used to aid identification of unknowns, or as secondary confirmation.

Our data was treated in a multi-step process. First, all retention times were corrected relative to the CO2 peak, and subsequently n-alkane and/or n-aldehyde homologous series were identified to set up an observed retention index system. Compound-by-compound identification is then performed using background-subtracted mass spectra, using base, isotope, and molecular ion peaks. Reference library spectra were used as structural group suggestions only. The mass spectral interpretations were checked against expected retention times derived from our own literature based database (currently ~5000 entries). Selected results are presented in Table 1, arranged by structural group, substituent group and carbon number.

1.3.3. Quantification

Quantification of compounds of interest was done for the TO-14 series of volatile organic compounds (VOC), comprising mostly halocarbons (US EPA method “Determination of volatile organic compounds (VOCs) in air”), using a certified TO-14 reference standard gas (Scotty, Inc., NL). Calibration was achieved by charging the same adsorbent cartridges used in sampling with different known amounts of standard gas at controlled conditions, with subsequent desorption and analysis under conditions identical to those employed for the volcanic gas samples. Recoveries were determined by separate direct splitless injection of a volumetric gas syringe of standard gas. The calibration results are tabulated in Table 2, as residuals of the calibration curves in % of the response, and as R2 correlation coefficients. For

28 of 38 compounds (except CH3Cl, CFC-12, CFC-114, C2H5Cl, DCM, benzene, chlorobenzene, styrene, and m- and o-dichlorobenzene), precisions were better than 10 % and in most cases better than 5% (determined in the range from 12.5 pptv to 12.5 ppbv). Statistical procedures were adapted to IUPAC guidelines (Danzer and Currie, 1998).

Chapter II. Halocarbons in volcanic gases 29

Figure 3 Example of data quality obtained by the method employed in this study. The total ion chromatogram (TIC) of a volcanic gas sample (Vulcano Island, 1999) is compared to the TIC of a pre-run system blank, displaying the extremely low levels of contamination encountered in the procedure. Relative peak hight of the blank is scaled to the respective quantified peak areas of benzene as a reference compound in the sample and the blank, so that the blank response scale is identical to that of the sample (response in % of largest peak of the sample). In the blank, only traces of benzene, toluene, acetone, ethanol, dichloromethane and butane are detectable, in the worst case making up 12 % of the analytical response of a real sample. Compound identification was by mass fragment analysis, relative isotope peaks, base peak, and molecular ion where preserved, with optional confirmation and/or isomer distinction aided by retention indices relative to native n-alkane and n-aldehyde homologues. Separation was by a HP 5890 Series II gas chromatograph in splitless mode with a Hall GC-1 analytical column (60 m x 0.25 mm x 0.50 µm; equivalent to DB-1), GC/MS transfer line open split 200°C. Analysis used a Trio 2000 in electron impact mode with 70 eV current. The enlarged portion of the sample TIC on the right shows separated ion peaks of 1,2,4-trimethylbenzene, n-octanal and p-dichlorobenzene, as well as overlapping peaks of n-octanal and m-dichlorobenzene. Using reconstructed specific ion chromatograms, background subtraction leads to clear mass pectra that allow identification of the compounds. The three mass spectra displayed on top are extracted from the enlarged TIC on the right, yielding a clear spectrum with correct isotope peaks even for small amounts of m-dichlorobenzene (<100 pptv).

Chapter II. Halocarbons in volcanic gases 30

1.4. Contamination, sample loss, degradation, artifacts and blanks

1.4.1. Contamination

Laboratory contamination was evaluated by analysis of sufficient system blanks taken before and after each desorption, and individual equipment blanks (SPME needle, column blanks, carrier gas blanks). Laboratory contaminant concentrations were at least two orders of magnitude below sample concentrations, with the exception of dichloromethane, a common laboratory extraction solvent.

1.4.2. Air entrainment

Air contamination during storage was addressed by field blanks (unused sampling cartridge) that gave negligible readings for all compounds, with contamination levels being comparable to laboratory blanks (except a few higher n-alkanes that reached a maximum of 10% of the lowest sample peak area readings). Air contamination during analysis was determined by air blanks, comprising blanks of laboratory and Manchester urban air, on-site air blanks at Vulcano, and the system/equipment blanks. Special care was taken to avoid and quantify air contamination during sampling.

Entrainment of air through the shallow soil during sampling was minimized by choosing fumaroles of highest gas flow as well as by careful implacement of silica glass sampling tubes deeply within each fumarole. A second potential source of air entrainment is through leakage of the fittings connecting the individual parts of the sampling train. This source is negligible in this study as the flow and higher density of the gas in the sampling train created a small overpressure with respect to outside air. In addition, air residing inside the sampling train was purged by volcanic gas prior to sampling. A third potential contamination source was during the de-capping, installing, uninstalling and capping of the adsorbent tubes before and after sampling. This was minimized by an exposure time of a few seconds.

The condenser bottles (impingers) were flushed several times with the volcanic gas prior to installation of the sampling cartridge. The sum of all remaining atmospheric contamination

Chapter II. Halocarbons in volcanic gases 31 sources during sampling was quantified using two approaches. Gas samples from the same site were collected using evacuated flasks containing 5 m NaOH (“Giggenbach bottles”), and cold-welding copper-tubes used for noble gas sampling.

Giggenbach bottles were analyzed (Giggenbach and Goguel, 1989) using an HP GC 6890 Series II model gas chromatograph with an in-house custom-built expansion line, which was tested for its leak rate (2.17 x 10-4, 8.91 x 10-5 and 6.20 x 10-6 Pa m3 s-1 for non-linear initial, linear continuous 20 min and 8 weeks equilibration). Separation was performed on a 30 m Chrompack® Molsieve 5A porous layer open tubular (PLOT) fused silica (FS) capillary column using dual filtered 6.0 N2 and 5.0 He carrier gases. Detection was by a micro thermal conductivity (TCD). Peak areas were quantified by calibration with gas standards (Scotty Inc, NL).

Oxygen was either not detectable or below calibration limit (< 0.5 ppm), and the N2/Ar ratio was far from air values (see Table 2). The noble gas samples were analyzed using the methods described in Kipfer (1991) and Beyerle et al. (2000) for helium and neon isotopes. The helium 4 3 isotopic ratio, expressed as He/ He, was 5.6 and 5.7 (equivalent to R/RA = 4.01 and 3.97; R = 4He/3He, where A denotes the atmospheric value) for both sampling sites (F0 and F11b, respectively) used for organic gas sampling. These ratios fall well within the range of magmatic island arc-type end-member signatures found for Vulcano crater fumaroles by various authors (R/RA max. = 6.1, Tedesco et al., 1995). The samples are strongly enriched in mantle helium, indicating a magmatic origin of the gases.

The He/Ne ratios of 42.07 and 45.32 for fumaroles F11b and F0 are distinct from that of air (0.29). He and Ne concentrations as well as 20Ne/4He ratios indicate a strong radiogenic crustal component diluting the magmatic gas. In addition, the neon isotopic data indicate small degrees of kinetic fractionation, probably by expansion of the gas during ascent, or entrainment of minor amounts of atmospheric 20Ne, by admixing of a small component either of air or of meteoric or marine water. We evaluated the results by three-component analysis of three mixing endmember reservoirs, i.e. magmatic (MORB-type), air-saturated water (ASW), and a crustal radiogenic type in 4He/3He - 20Ne/4He space. As a result, the proportion of gas that can be traced to air represents a mole fraction of 3.3 x 10-6, and 3.4 x 10-6 for samples HP2 and HP4, respectively (Vulcano fumaroles F11b and F0, cf. Table 2).

Chapter II. Halocarbons in volcanic gases 32

These very small amounts of possible contamination most probably are characteristic of the gas before it reaches the surface rather than being introduced by the sampling method. This small non-juvenile component could not account for the observed halocarbons being derived from air entrainment. Air entrainment may have occurred either directly into the shallow gas stream or into the hot hydrothermal reservoir via dissolution into meteoric water that after infiltration subsequently mixes with juvenile fluid in the hydrothermal envelope. The hydrothermal aquifer was drilled and found to be at ~400°C at Vulcano (Giocanda and Sbrana, 1991), whereas the crater fumaroles are in several cases at a higher emission and temperature, reflecting a strong juvenile component. CFC’s are of very low solubilities (cf. Table 3) that are even lower in strong electrolytes such as hydrothermal brines. To balance the gas flux and observed concentrations in the volcanic gas, the amount of air required would be unrealistic, and in the case of the calculated small entrained air proportion would necessitate an impropable high halocarbon concentration in the incorporated air in the higher per cent range.

Furthermore, a simple mass-balance calculation also emphasizes the impossibility of ambient air entrainment being responsible for the organic compounds in our samples. Ambient air samples taken near Gelso in the South of the Island of Vulcano and on board of a ferry leaving from Vulcano showed significantly lower contaminant values than urban air samples (Manchester, UK). Using typical maximal urban air concentrations and CFC-11 (F11) as an F11 example, the following can be computed (C is concentration): C max. urban air = 0.6 ppbV, F11 C volcanic gas (range) = 15 – 80 ppbV. In order to create this concentration of CFC-11 in 500 mL sampled volcanic gas, 13 to 67 L of urban air would have to be pumped through the adsorbent tube in addition to the volcanic gas. These high concentrations are not consistent with the typical volumes (~ 0.5 L) sampled with the low-flow pump directly connected to the downstream side of the adsorbent tube (10 mL/min dry gas flow, in accordance to the natural fumarolic flow measured on the post-condensation dry gas). We conclude that ambient air as a source of the halocarbons can safely be excluded.

1.4.3. Loss of analytes to condensate or adsorption to particles

To test for possible loss of analytes by dissolution in the volcanic gas condensate, we analyzed thermal/sparging extraction analyses of the stopcock-sealed condensation flasks. Maximum dry gas concentrations of CFC-11 in the samples HP4 and HP2 were 3.74 and 0.34

Chapter II. Halocarbons in volcanic gases 33 ppbv, respectively, analyzed from the adsorption cartridges (dry ambient air: 0.2-0.3 ppbv) and below detection limit (ca.5 pptv) in the corresponding condensates. In the condensates collected with sample HP4 (fumarole F0), 0.33pptv were estimated from the m/z 101+103 fragments at the appropriate retentionm time, quantified by linear extrapolation of the lowest concentration of the CFC-11 calibration curve. It follows that solubility of CFC-11 in the condensate is negligible in our method, reflecting its’ extremely low Henry constant for CFC- -1 -1 11 (KH = 0.009 ± 0.0076 mol kg bar , average of 8 sources, in Sander, 2001) and its relatively low high boiling point (296.92 K).

We consider dissolution to play only a minor role, because of a) kinetic inhibition (fast bubbling and low KH), and b) the initially dramatically increasing ionic strength of the condensate further enhancing salting-out effects and inhibiting its solvent capacity by simultaneous condensation of steam and dissolution of strong mineral acids (HCl, HBr,

H2SO4), and more soluble, major gas species (CO2, NH3). To test if organic analytes were adsorbed onto solid sulfur sublimates, sulfur crystals around a fumarole were collected in a chromatography-grade cleaned borosilicate vial which was then capped immediately. Mild thermal desorption with the headspace-SPME technique yielded only minor amounts of adsorbed arenes (i.e. benzene and toluene), but no halocarbon compounds at detectable levels.

The thermal stability of hydrocarbons at hydrothermal conditions has been reviewed by Price (1993), and the thermal decomposition of chlorinated alkenes at temperatures over 400°C has been experimentally investigated by Yasuhara (1993). Thermal degradation will not play a role during sampling or analysis, unless temperatures above 400 °C are reached. Reactions and decomposition on the adsorbent during storage are possible but considered minimal if kept dry and frozen (e.g., Gan et al., 2001). Keeping the sampling cartridges in the dark avoided photolysis of compounds, by wrapping the sampling train and the cartridges with aluminium foil during sampling, and by cooling (ice pack) with subsequent storage in the dark in the field and at <0° in the laboratory prior to desorption. The photolysis rates of environmentally important compounds were reviewed by Atkinson (1994).

Aerobic oxidation of halocarbons by organisms may play a role via dehalogenation reactions if storage temperatures are not kept low enough for before analysis. Microbial remediation of halocarbons has been shown for a variety of bacterial genera. Olaniran et al. (2001) investigated bacterial dehalogenation potentials for mono- and trichloroacetic acid,

Chapter II. Halocarbons in volcanic gases 34 chloroform and carbon tetrachloride. Anaerobic (i.e. reductive) dehalogenation is another process affecting the stability of adsorbed halocarbon compounds. In this context, Zhao et al. (2001) determined the rate constants of reductive dehalogenation of a variety of halocarbon compounds including CFC-11 (trichlorofluoromethane).

Chapter II. Halocarbons in volcanic gases 35

5. RESULTS

A large range of organic compounds has been detected above blank concentrations in a variety of gas samples obtained from Vulcano crater (Italy). Selected results are reported in Table 1 (qualitative data) and Table 2 (selected quantitative data). We always found benzene and toluene to be the species of greatest concentrations (lower ppmv), followed by oxygenated and sulfonated compounds. Aliphatic hydrocarbons are of low concentration and often masked by co-eluting oxygenated compounds. Alkenes seem to dominate over alkanes consistently. The highest concentrations are detected for simple substituted benzenes, among them toluene, C2-C4 alkylated arenes, chlorinated benzenes (mono- to trichlorobenzenes), and oxygenated arenes, notably phenols.

5.1. Hydrocarbons

The hydrocarbons detected in fumarolic emissions include C4-C15 normal and branched alkanes, and C4-C12 mono-unsaturated hydrocarbons (alkenes). In the alkenes, a greater degree of branching is observed compared to the alkanes. The highest carbon number alkanes

(up to C21, not listed in Table 1) are observed in the 1999 sample of fumarole F8/9, which is also the one of highest gas discharge temperature. The narrowest range appears in the 2000 sample of fumarole F0 (the most reducing one with excess H2) at a discharge temperature of 394.8°C. From many of these compounds, substituted derivatives have been found, mainly with halogens as the substituents. The abundance of substituted halogens are in the order

Cl>Br>F>I. Halogenation is only observed for C1-C2 aliphatic hydrocarbons, even though halogenated C3-compounds were specifically targeted with the help of the TO-14 standard (cf. Table 2). This observed abundance pattern is different from typical polluted urban atmosphere, where halogenated propanes are among the common anthropogenic contaminants (US EPA, method TO-14). Unsaturated hydrocarbons can thermally be converted to naphthenes or arenes (aromatic compounds), many of which were found in our study. The naphthenes occur with varying degrees of unsaturation and substitution. No terpenes and terpenoids were detected, although they were specifically sought as the lower volcano flanks

Chapter II. Halocarbons in volcanic gases 36 are vegetated. Only an isolongifolene isomer was found, but as it is also used in adhesives, it is not possible to unambiguously assign its’ source.

In the arenes, several degrees of methylation and halogenation are observed (Table 1). C4- and higher substituted benzenes were found but are not listed yet in Table 1, because of uncertain structural assignment. Polycyclic aromatic hydrocarbons (PAH’s) were found consistently in all samples, including their methylated derivatives. Hydrogenated PAH’s were found but not unambigously. Naphthalene was found in all samples and phenanthrene was identified in a sample not listed in Table 1. Halogenated PAH’s were not detected. Oxygenated arenes are discussed below. All C0-C3 arenes occur at similar concentrations in all four samples, however trichlorobenzenes were not identified in fumarole F0. McCollom et al. (2001) found experimentally that toluene readily degrades catalytically to benzene under hydrothermal conditions (300-330°C, 350 bar), with further interactions involving benzaldehyde and phenol. All four compounds are present at appreciable amounts in our samples, with toluene being the most abundant organic compound in all analyses. The presence of chlorobenzene suggests that fixation of monocycles plays a role, as well as further substitution reactions indicated by small amounts of dichlorobenzene isomers and alkylated arenes. McCollom et al. (2001) suggested that abundant sulfur compounds may facilitate the formation of oxygenated derivatives.

Hydrocarbons (alkanes, alkenes, alkynes, naphthenes, arenes) dominate the chemistry of post- emplacement gas emissions from lava flows and ash deposits in which vegetation was pyrolyzed. These pyrolysates are either released through vents, or adsorbed onto the surfaces of ash particles (Finlayson et al., 1968; Muenow, 1973; Markhinin and Podkletnov, 1977b, 1977a, 1978; Graeber et al., 1979; Pereira et al., 1980; Podkletnov and Markhinin, 1981; Pereira and Rostad, 1983). In low-temperature hydrothermal manifestations, the same dominance is observed (Capaccioni et al., 1995; Darling, 1998; Capaccioni et al., 2001), as well as in high-temperature gas emissions from lava lakes or flows (Markhinin et al., 1977; Markhinin et al., 1983). In high-temperature fumarolic gas emissions, however, hydrocarbons are less abundant relative to oxygenated, sulfonated, and halogenated compounds of smaller molecular weight and greater volatility.

Chapter II. Halocarbons in volcanic gases 37

Table 1 Qualitative results of four volcanic gas analyses from Vulcano (Italy) crater fumaroles. Further details are given in the following captions.

Chapter II. Halocarbons in volcanic gases 38

Table 1 (continued)

Chapter II. Halocarbons in volcanic gases 39

Table 1 (continued)

Chapter II. Halocarbons in volcanic gases 40

r n n m ) from our ow ) from our lit R ), its maximu with rs, rounded molecula rounded rs, AL R & I R tural type and ascending carbo and ascending type tural ) retention indices, following IUPAC- following indices, ) retention AL R ) and separate runs (I runs separate ) and lit -aldehyde (I -aldehyde R n ) are given as DB-1 literature values (I R -alkane and -alkane and n sources of information (I n . Retention indices served as confirmatory information to mass spectral fragment analysis, n

or as an aid to distinguish isomers of a compound (noted “IR” if diagnostic). diagnostic). if “IR” (noted of a compound isomers distinguish to an aid or as Table 1 (continued) Qualitative results of four volcanic analyses gas from Vulcano (Italy) crater fumaroles, tabulated by struc numbe the CAS and synonyms, with acronyms the IUPAC-names formulae, are the Given or complexity. of substitution degree number, (I indices retention programmed temperature Linear information. index (M), and retention masses for runs, from 6 separate samples of actual values and measured entries), 5000 (ca. compilation values, of considered number ±”) and the mean (“err the from deviation recommended procedures. Each retention index value is shown as an average of of an average as shown is value index retention Each procedures. recommended

Chapter II. Halocarbons in volcanic gases 41 Table 1 (continued) Table

Chapter II. Halocarbons in volcanic gases 42

5.2. Oxygenated compounds

5.2.1. Alcohols & Ethers

Few alcohols were found. The highest concentrations were ethanol and phenol. Substituted phenols and a chloroethanol isomer were also found (Table 1). The only ether identified was diethyl ether. Oxidation at elevated temperatures may take place at shallow depth or during emission of a volcanic gas to form alcohols by an addition reaction on alkenes, and subsequent oxidation of alcohols to aldehydes (Muenow, 1973). The latter reaction may potentially be indicative of the redox state of a volcanic gas if the alcohol / aldehyde ratio for a given carbon number is measured. All three chlorophenol isomers were found in one sample (fumarole F8/9), of which only o-chlorophenol was identified in fumaroles F5AT and F11b, but not in F0. Di- and trichlorophenol isomers were found as well in samples from fumaroles F8/9 and F5AT.

5.2.2. Aldehydes

Normal aldehydes are present in most samples in slightly greater abundances (response ratios) than n-alkanes, making them a useful alternative retention index variable at the beginning of the chromatograms, where n-alkanes are often masked by other compounds. Aldehydic arenes are present in the form of benzaldehyde and its’ derivatives. Substitution by alkyl groups is less prominent as for the alkanes and alkenes. As the only halogenated aldehyde a fluorodihydrohy benzaldehyde isomer was identified in three samples.

5.2.3. Ketones and acids

Several C3-C8 ketones were identified. Acetone is also a common volatile laboratory contaminant, however, and is present at low concentrations in blanks. A hydrated furanone was found as well. C2-C7 carboxylic acids are always present at appreciable quantities in samples, especially acetic acid, but were absent in procedural blanks. Formic acid was identified in a sample not listed in Table 1. Since water is only a trace compound in the dry

Chapter II. Halocarbons in volcanic gases 43 gas fraction, acetic acid is unlikely to have been formed in-situ on the adsorbent after or during sampling by dissociation of carbon monoxide to formate with subsequent methylation.

5.2.4. Esters

Two groups of esters were identified: C4-C7 esters of carboxylic acids also identified in the ® samples, and C5-C8 methacrylates maybe derived from trace contamination (Kapak caps, SPME adhesive, septa). No halogenated derivatives were identified.

5.3. Heteroatomic compounds excluding halocarbons and oxygenated hydrocarbons

Several not halogenated cyclic and acyclic heteroatomic compounds were found in all samples. Among the heterocyclic compounds, oxygen and sulfur heteroatoms dominate. Dioxanes and substituted dioxanes were found in only one fumarole (F11b), whereas thiophenes with varying degrees of substitution of C1-C2 alkyl groups were found in all samples, and up to C5 in condensates. A benzothiophene isomer was identified in fumarole F5AT. A dioxin and partially saturated and methylated furans were found in several samples. Furaldehydes (2- and 3-furaldehyde), as well as nitrogenated heterocycles (pyrazine, pyridine and pyrrole) were identified as well in samples of fumaroles F11b, F0 and F8/9.

5.3.1. Nitrogen-containing compounds

NH3 and N2 (g) are major compounds in most volcanic gases. Methylated imidazole was found in one sample (fumarole F11b), as well as several nitriles (C2-C5) and amines. Only N- ethyl-N-nitrosoethenamine (C4H10ON2) was so far identified with a reasonable level of confidence as an amine. Two simple amino acids (aspartic and glutamic acid), which by themselfes can not be attributed to bacterial decomposition have been found in our condensates by ion chromatography but not in field and procedural blanks (Schwandner, Cox & Seward, unpublished results). These amino acids found are probably not pyrolisates of organic matter, since they are relatively unstable in aqueous pyrolysis experiments at temperatures comparable to the sampled volcanic gases (Douda and Basiuk, 2000).

Chapter II. Halocarbons in volcanic gases 44

Nitriles are compounds with a terminal carbon - nitrogen triple bond (a nitrile functional group, R-C≡N). Acetonitrile is present in all samples, 2-propenenitrile is confined to fumaroles F0 and F8/9, and 2-methyl-butanenitrile to F11b. Differing from nitriles, cyano- compounds contain a nitrile functional group in lower priority position. From all samples, we only identified isocyanoethane with confidence in fumarole F11b (sample HP2).

Nitriles are found in the lower troposphere in low ppbV concentrations (Becker and Ionescu, 1982; Isidorov, 1990). In the case of acetonitrile, a high tropospheric stability is assumed, as shown by the gas phase OH-radical reaction rate constant of 2-3 x 10-14 cm3molecule-1s-1 (Zetzsch, 1980). The environmental significance of nitrile emissions is mainly its influence on tropospheric OH-radical reactions which have a direct significance to heterogenous reactions promoting the release of reactive chlorine to destroy stratospheric ozone. Nitriles are also of interest with respect to the abiogenic synthesis of prebiotic molecules, as exemplified in NASA’s research efforts in finding nitriles in extraterrestrial atmospheric environments (Navale et al., 1998; Navarro-Gonzalez et al., 2001). The significance of nitrogenated compounds to the origin of life has been recently summarized (Navarro-Gonzalez et al., 2001).

The most simplest nitrile, HCN, has too simple a mass spectrum to be unambiguously detected in the volcanic gas dry gas fraction by our method, but it has been detected by other GC methods by Seward (1991), who reported 8 µg kg-1 HCN in the deep reservoir fluid of the Ohaaki-Broadlands geothermal system (New Zealand), and by Mukhin (1974) and Mukhin et al. (1974) in hydrothermal discharges in Kamchatka and in magmatic gas escaping from Alaid lavas. In addition, Pavlov and Karpov (1972) reported ferrocyanide complexes in hydrothermal discharges of Uzon Caldera (Kamchatka). Acetonitrile and acrylonitrile were reported from solfataric gas analyses of air of volcanoes on Kunashir Island (Kuriles/Kamchatka) by Isidorov et al. (1990). They were found on Suswa volcano, Kenya, and were interpreted as pyrolysis products of bitumen from ancient caldera lake vegetation by Giże and Macdonald (1993) who analyzed bitumen seepages but not high temperature volcanic gases as in our current study. A hot gas sample derived from pyrolized bitumen would yield more significant n-alkane (alkane) concentrations than observed in our volcanic gas analyses. The occurrence of nitriles in our Vulcano gas samples can therefore not be easily explained by pyrolysis (cf. Figure 6).

Chapter II. Halocarbons in volcanic gases 45

1.1.2. Sulfonated compounds

Apart from thiophenes, a small range of organic sulfides were detected. Among them are carbon disulfide (CS2) in fumaroles F8/9, F5AT and F11b as well as dimethyl sulfide in fumaroles F11b and F8/9, which also occurs in condensates of fumarole F0. Dimethyl disulfide was found in all fumaroles but F5AT, and dimethyl trisulfide only in fumarole F11b.

The methylated derivatives of CS2 represent one group that is distinct from the oxidation product of dimethyl sulfide (DMS), dimethylsulfone, (CH3)2SO2, that was found in all samples. No thiols or thionyls were found. CS2 has been previously identified in volcanic gases (e.g., Rasmussen et al., 1982). It is one of the compounds of interest to atmospheric science, that together with COS, is injected into the atmosphere by explosive volcanic eruptions.

1.4. Halocarbons

We detected a range of halocarbons in the volcanic gases, summarized in Tables 1 and 2. The mono- and polychlorobenzenes have probably not been found in previous studies, as they elute later than most CFC’s and may have been missed. Because of the greater interdisciplinary and political interest in natural sources of halocarbons, we specifically calibrated the mass spectrometer response for some prominent halocarbons of the TO-14 series, including the short-lived chlorobenzenes. As well as several CFC’s (CFC-11, CFC-12, CFC-113), we found methyl bromide and methyl iodide, both of which are currently topical in terms of their great ozone depletion potential and natural sources.

Andesitic volcanic fumaroles and rocks typically contain Cl and F in concentrations within an order of magitude of total sulfur, with halogen concentrations decreasing from Cl and F to Br and I. The source of halogens has been discussed in terms of the involvement of evaporites in the magmatic system in the case of the Aeolian and Aegean Arcs (Gill, 1981). However, other arc volcanoes without regional evaporites emit large quantities of halogens to the atmosphere (Symonds et al., 1988). Halocarbons are emitted during quiescent fumarolic as well as eruptive activity, and may potentially also be found in increased concentration directly before eruptions, when HF and HBr fluxes increase (e.g., Udisti et al., 1999). Chlorine isotopes have been used to trace origins of chlorine in volcanic gases (Eggenkamp and Kreulen, 1984; Wahrenberger et al., 1996b) and submarine hydrothermal fluids (Lüders et al., 2001).

Chapter II. Halocarbons in volcanic gases 46

Table 2 Quantitative results of TO-14 compounds present in two volcanic gas samples. Types of information given are the TO-14 compound name, compound code according to ASHRAE nomenclature (refrigant code, as Halon=H, and Freon=F), formula, CAS-number, exact molecular mass, and ions used for quantification and calibration. Calibration parameters are given, with R2 values and residual errors of the curve fit in % at the 12.5 ppbv level. Sample information is given, with analytical information of noble gases, noble gas isotopes, nitrogen and molar “concentration” of representative condensate that gets separated before adsorption during sampling. Analyses of condensates by mild heating and sparging with helium yielded negligible concentrations of the analytes presented here.

Chapter II. Halocarbons in volcanic gases 47

1.4.1. Chlorinated halocarbons

Methyl chloride (CH3Cl), dichloromethane (DCM, CH2Cl2), chloroform (CHCl3), and carbon tetrachloride (CCl4) in volcanic gases have been previously qualitatively reported (Rasmussen et al., 1980; Isidorov, 1990, 1992; Wahrenberger, 1997; Jordan et al., 2000), ranging in concentration from 0.8-930 ppbv (air: 0.6), 0.04-69 (air: 0.02), 0.01-820.8 (air:0.01), and 0.1-

19.1 (air:0.1) for CH3Cl, CH2Cl2, CHCl3 and CCl4, respectively. Their presence above ambient air is confirmed by our data, relative to air blank concentrations.

CH2Cl2 and CHCl3 have concentrations of 199-307 and 20-30 ppbv, respectively, in dry volcanic gas (corresponding to 12.9-25.3 and 1.4-2.6 ppbv in fumarolic vapor). The average tropospheric dry air concentrations are 0.02 and 0.01 ppbv (McCulloch and Midgley, 1996; CMDL, 2001; Hoekstra et al., 2001). Other chlorinated aliphatic hydrocarbons are restricted to C2 alkanes and alkenes, comprising chloroethane and 1,2-dichloroethane for the alkanes.

For the alkenes, chloroethylene was found in fumarole F5AT, all three isomers of C2H2Cl2 in fumaroles F5AT and F8/9, as well as tri- and tetrachloroethylene in fumaroles F5AT and F11b.

1.4.2. Brominated and fluorinated halocarbons

Several brominated compounds are found in volcanic gases, as reported by Rasmussen et al. (1982), Jordan et al. (2000) and Schwandner et al. (2000b). In our study, we find mainly methyl bromide, bromoethane, dibromoethane and tribromoethylene at minor abundances (up to 7.8 ppbv CH3Br in dry volcanic gas, dry air: 0.01). A few other bromocarbons were not unambiguously identified. As opposed to fluorine, bromine readily partitions into an exsolving H2O-rich vapor phase upon cooling of a magmatic intrusion (Bureau et al., 2000; Bureau and Metrich, 2001).

Fluorinated organic compounds in our analyses are dominated by CFC-11 (CCl3F), with minor CFC-12, CFC-113 and fluorodihydroxy benzaldehyde. 2,2-difluoropropane (HFC-272) is suspected in a sample not listed in Table 1 and fluorobenzene was not unambiguously identified in fumarole F11b. Fluorinated organic compounds are strongly underrepresented

Chapter II. Halocarbons in volcanic gases 48 relative to chlorinated compounds, even though HF generally exceeds HBr concentrations. There are several reasons for this discrepancy. The F/Cl ratio for Vulcano crater emissions is about 0.2 (Wahrenberger, 1997), while magmatic melt inclusions in Vulcano dacites and andesites have F/Cl of ca. 0.38 (Zanon et al., 2001).

Fluorine concentrations in all volcanic rocks on Vulcano range from 280–1285 ppm, the highest values representing the more highly differentiated magmas (Martini, 1984). Fluorine has a greater solubility in andesitic melts than Cl (Burnham, 1979; Giggenbach, 1996, 1997), and the exsolved magmatic gas encounters several sinks for fluorine in volcanic-hydrothermal systems en route to the surface. At elevated hydrothermal temperatures (600°C, 1000 bar), fluorine is incorporated preferentially into trioctahedral sites of potassic micas present either in the magmatic host rock, or as secondary alteration minerals in the hydrothermal envelope (Robert et al., 1993). The relatively low concentrations of fluorinated compounds in volcanic gases relative to chlorinated compounds and to the fluorine content of the host magma may reflected preferential fluorine loss during these processes.

A strong enrichment of F relative to Cl is shown by advanced and intermediate argillic alteration on Vulcano with F/Cl mass ratios between 9.26 and 41.82, and silicic alterites ranging 0.78-4.19 (Fulignati et al., 1998). Crystallization as fluorite in veins is a similarly efficient sink for magmatic fluorine. Strongly fluorescent hydrothermal fluorite (“Stinkspat”) has been shown to contain up to 0.6 ppm of organic fluorinated compounds, including CFC’s (Kranz, 1966; Harnisch et al., 1996; Harnisch and Eisenhauer, 1998; Harnisch et al., 2000). Owing to the great solubility of both HF and HCl, they are easily removed from the gas by vapor condensation. The other major non-HF fluorinated species is the highly soluble SiF4, which has been found by remote sensing techniques (Francis et al., 1995) to be emanating from Vulcano and Etna, in fumarolic gases (Symonds et al., 1992) and in fumarole incrustations (Brun, 1910).

In addition, Vulcano crater fumaroles of intermediate temperature sublime ammonium borofluoride crystals (Barberiite). Our own investigations indicate possibly as many as nine independent fluorinated species in the volcanic gas condensates from Vulcano (Italy), based on analyses of condensates by 19F-NMR (Schwandner, Ellis, Cox, Mabury & Seward, unpublished results). Four CFC’s and one HCFC have been identified by GC-MS (CFC-11, - 12, -113, -114; HCFC-142b) in this and previous studies (Wahrenberger et al., 1996a;

Chapter II. Halocarbons in volcanic gases 49

Wahrenberger, 1997; Wahrenberger et al., 1998). Jordan et al. (2000) lists 5 fluorinated compounds (CFC-11, C2H3F, C6H5F, C6H4ClF, C6H2F4), of which we identified only CFC-11, and suspect fluorobenzene in one sample (fumarole F11b).

1.4.3. Iodinated halocarbons

Methyl iodide was identified in our gas samples. Its very high photolysis rate (cf. Table 4) makes it labile during sampling and storage, so that the concenrations determined are probably an underestimation (not calibrated, by analogy to CH3Br ca. 11-17 pptv; air: 3.4 pptv, cf. Table 3). Other iodinated compounds have not been identified in this study, reflecting the generally low iodine concentration in volcanic gases. Other iodinated compounds (C2H5I and CH2ICl) were reported by Jordan et al. (2000), but these results are questioned because they used NaOH solution during sampling that remained in contact with the sampling bottle headspace during storage and transport and may have resulted into dissolution and reaction of iodinated compounds. Yu et al. (1996) demonstrated that NaOH solutions efficiently dissolve iodide adsorbed onto volcanic ash.

Iodide may also have a sink in the hydrothermally altered volcanic edifice, as described for fluorine. Gaseous molecular iodine in volcanic gases may dissolve into water as soon as vapour condenses on surfaces during ascent and emission of the gas. The dominating species will then be iodide if the conditions are sufficiently reducing, as is the case for the high- temperature crater fumaroles on Vulcano, and will get retained by the volcanic soil rapidly (Yu et al., 1996). Other compounds may have never reached the surface, since clay minerals and silica within the volcanic edifice are effective sorptive surfaces for organic compounds, including halocarbons (Goss, 1992, 1993). Some new evidence on the sources of iodine comes by 129I isotope tracing of volcanic fluids (Snyder and Fehn, 2002). The available quantitative literature data on volcanic halocarbon trace gas concentrations are summarized in Table 3.

Chapter II. Halocarbons in volcanic gases 50 . on ati m r o f n e i v titati n a u q

d n a

p u o r g

d elly et al., 1993; S. this study. this S. 1993; al., et elly n t al. 2000; K. IPCC 2001; L. McCulloch 2001; IPCC K. 2000; al. t u o gle reported value, with or without reported without or with value, reported gle nkett, et al. 1981; E. Isidorov et al. 1985b; F. 1985b; al. et E. Isidorov 1981; al. et nkett, p m o , c y atilit l vo ess r p x e

to

° p , respectively). Statistical information are given as the standard deviation (95% deviation standard the as are given information Statistical , respectively). max c at T°, H ), number of literature valuesliteratureof (in some number of cases averages several ranges of repeats), ), ° k and and avg. ) and literature quality discussed (data the in the text). Given are analyte name, molecular avg. trop S * i c c . * Average is one of three data types: a) average of reported range, error is 1/2 range, some ranges error based range, 1/2 is on are some range, of a) average reported types: data of three one is . * Average avg * c ass, heteroatoms, CAS number, boiling point at p° and Henry’s law-constants p° law-constants and at point Henry’s boiling number, CAS heteroatoms, ass, eferences to data sources, and average and maximum concentrations reported ( reported concentrations maximum and average and sources, data to eferences several samples/measurements; b) average of deviating multiple reports of the same analysis in consecutive publications; c) sin publications; in consecutive analysis same of the multiple reports deviating of b) average samples/measurements; several Pe D. Gerlach, 1980; in: pers. commun., Rasmussen, et al. 1980; C. Rasmusssen et al. 1975; B. A. Wilkniss, error. Data sources: e I. Jordan 1997; & Wahrenberger 1996a al., et Wahrenberger H. 1992; al. et Isidorov, G. 1990; Isidorov & 1990a al. et Isidorov and Midgley, 1996; M. CMDL 2001; 1990; N. Isidorov O. Bassford1999; P. estimate after Rasmussen et al., and Khalil, 1983; Q. K Table 3 List of halocarbon and arene analyses of volcanic gases from this study ( study from this gases volcanic of analyses arene and halocarbon of 3 List Table m r confidence interval) in concentration units and in % of and in units concentration interval) in confidence Concentration information comprise average global tropospheric abundances ( abundances tropospheric global average comprise information Concentration

Chapter II. Halocarbons in volcanic gases 51

1.5. Other compounds

Few inorganic compounds are detectable by GC-MS analysis at the conditions used. Among them are CO2, SO2, water, native sulfur and silanes. The silanes are decomposition products of the stationary phase of the analytical column (PDMS) following exposure to acids. Difluorodimethylsilane and fluorotrimethylsilane are listed in Table 1, but may also be a decomposition product. In addition to the approach discussed in this study, GC-TCD analyses of giggenbach-bottle headspaces were analysed for noble gases, hydrogen, oxygen and nitrogen. Some of the results are listed in Table 2 and the results and techniques were discussed above in the context of air contamination.

Chapter II. Halocarbons in volcanic gases 52

6. FLUX ESTIMATES OF VOLCANIC HALOCARBON EMISSIONS

6.1. Total carbon fluxes and carbon recycling

The total elemental and molecular fluxes of CH4, C, halogens, and total organic fluxes have also been related to CO2 fluxes. The mantle carbon flux has been assessed by Javoy et al. (1982), and a more recent model estimates the global elemental carbon fluxes into the mantle by subduction to be 1.2 ±0.3 x1010 kgy-1 of sediment plus 3 ±0.5 1010 kgy-1 from hydrothermally altered basalt, summing up to a total of ca. 4.2 ±0.5 1010 kgy-1 for global subduction recycling. The arc flux out of the mantle is estimated as 2.0 ±01.7 1010 kgy-1 (Coltice et al., 2001).

An additional pathway of carbon through the crust would be via earthquake degassing of CO2 13 and CH4, which shows an enrichment of C , suggesting a deep origin (Voitov et al., 1986;

Voitov and Yunusov, 1996; Voitov, 2000a, 2000b). The volcanogenic CH4 flux was estimated as 0.2 Tg y-1 by Ehalt (1974), for subaerial volcanoes alone; another estimated stated 0.34 Tg -1 y (Cadle, 1980). The anthropogenic CH4 emissions were estimated to be on the order of ~600 Tg y-1 for late 1990’s (IPCC, 2001). The total organic flux from volcanoes to the atmosphere was estimated as 2 to 5 Tg y-1 (Markhinin, 1985; Isidorov, 1990). The earth’s crust as a hydrocarbon reservoir was reported to hold about 5.8 x 107 Tg of organic matter (Isidorov, 1990).

6.2. Total halogen fluxes

Total halogen fluxes were estimated for several volcanoes and on a global scale by various authors. Iwasaki et al. (1968) tried to estimate the inorganic chlorine flux from the mantle to the surface on the basis of >500 samples from different tectonic settings and of different magmatic type, but did not calculate a global mass balance. Vulcano (Italy) has been estimated to emit 13t HCl per day (Francis et al., 1995). The global atmospheric concentration of organic chlorine is estimated as 3.7 ± 0.1 ppbv (WMO, 1998). Two major sinks for atmospheric non-sea salt chlorine from long-lived chlorinated compounds such as CFC’s are dry deposition (30%) and acid rainfall (70%), accounting for a sink strength of ~ 3 Tgy-1 that

Chapter II. Halocarbons in volcanic gases 53 is well balanced by the atmospheric deposition estimate of 1.85 Tgy-1 (Sanhueza, 2001). The anthropogenic and natural, non-volcanic bromine flux is 100 kty-1 (Bureau et al., 2000). Mt. St. Helens emitted about 2.4-5.6 kt in 1980 and Tambora released 790-1820 kt in 1815 (Bureau et al., 2000). The fluorine flux is not quantified for volcanic sources yet, but the global natural fluorine flux is estimated as 0.1-10 Mgy-1 (Harnisch et al., 1996) and supported 3 -1 by ice core data. The anthropogenic CF4 flux is estimated to be 28 x 10 y (Fabian et al.,

1987), amounting to ca. 35% of global CF4 emissions into the atmosphere. Fluorinated organics are common in fluid inclusions (Kranz, 1966; Harnisch et al., 1996; Harnisch et al., 2000). The volcanic iodine flux to the atmosphere will have a negligible effect, since its enriched hydro- and atmospheric concentration is several orders of magnitude higher (~1800 ppb) than its magmatic and mantle concentration of ~ 8 ppb, with a rare maximum of 260 ppb in some basaltic melts (Jambon, 1995).

6.3. Trace gas correlations

To quantify the volcanic source strength of halocarbons, the organic composition may be scaled to SO2-fluxes determined independently by correlation spectrometer (COSPEC) measurements. However, there are discriminatory shortcomings associated with SO2-fluxes determined by this method (sulfate aerosol and sulfate-covered ash particles). Sulfur speciation can be variable in a volcanic gas, making it difficult to correlate SO2-fluxes and

SO2-concentrations in the pure fumarolic gas, in that in some cases H2S dominates over SO2, or vice-cersa. CO2 concentrations in a fumarolic gas are not as variable, even though volcanic

C-O-H fluids speciate into CO and CH4, depending on temperature and redox potential. CO2 is in most cases the predominant species. In addition, CO2-fluxes determined by COSPEC have the advantage that confirmatory data are available for an increasing number of volcanoes by grid mapping of diffusive soil degassing (e.g., Koepenick, 1995; Chiodini et al., 1996; Chiodini et al., 1998; Giammanco et al., 1998; Wardell et al., 1999; Brombach et al., 2001; Chiodini et al., 2001).

6.3.1. Correlation with diffusive CO2 degassing

To test the applicability of CO2-degassing data to organic trace gas emissions, we also performed a survey on Vulcano’s La Fossa (Italy) flank and crater diffusive emissions to find

Chapter II. Halocarbons in volcanic gases 54 suitable correlation parameters (Schwandner et al., 2002a). Chiodini et al. (1996) determined the La Fossa crater diffusive gas discharge by soil gas surveying to be on the order of 0.073 Tgy-1, in addition to the fumarolic maximum emissions of 0.037 Tgy-1 (max. 0.402 Tgy- 1 during the seismic-volcanic 1988-1992 crisis). Baubron et al. (1990) showed clear evidence that the Fossa crater gas escaping by diffusive emission is of the same, magmatic, source as 13 12 the fumarolic emissions, based on He/CO2 and C/ C data of the soil gases.

Figure 4 Correlation plot of CFC-11 (CCl3F) concentration of soil gas samples obtained in the crater and on the flanks of Vulcano (Italy) at soil depths of ~60 cm, with soil gas fluxes determined by the dynamic method (Chiodini et al., 1998). Inner crater samples are significantly above tropospheric air concentrations in terms of CFC-11 concentration. The good correlation suggests a genetically coherent origin of the observed CFC-11 with diffusively degassing magmatic CO2. Precision is better than 2%, residual errors of the CO2-flux calibration are <0.1%. Symbol sizes are larger than error ranges.

Soil gas CO2-emissions were measured using the dynamic chamber method (Chiodini et al.,

1998). CO2 was detected by a portable IR-Sensor calibrated to CO2 (Multiwarn II, Dräger GmbH, Kiel, Germany). The corresponding soil gas emissions of organic trace gases were measured by a modification of US EPA method SOP 2042: a steel probe is lowered into the ground to an appropriate depth (0.5-0.8 m), a sampling tube (focusing tube) is inserted and sealed in, and after slow purging, a known volume of soil gas is passed through a mixed-bed solid adsorbent tube (SKC Inc., PA, USA) using a battery driven calibrated low-flow sampling pump. Of the sampling sites, only one had to be discarded because of sealing problems which resulted in significant air contamination. All other results are displayed in the

Chapter II. Halocarbons in volcanic gases 55

CO2-flux vs. CFC-11 correlation diagram in Figure 4. On the basis of the good correlation and the above discussed advantages, we have scaled the natural halocarbon emissions to the CO2- fluxes instead of SO2. Analyzing SO2 in soil gases would be complicated since it adsorbs very quickly onto soil matter. We tried analyzing sulfur species fluxes in the field, but the data showed variations controlled by soil type and vegetation. CO2 degassing is directly related to subsurface CO2 flow and structures only.

6.3.2. Mass flux calculations

Molar fluxes of individual compounds from one source of interest (e.g. a specific volcano) can be calculated using: F G F X G  CO2 i Fi V , F X CO2 where F = flux; G = global, X = source of interest, i = component of interest. Using global averages or representative concentrations of trace gases as an approximation, this formulation becomes:

G G F C V G  CO2 i Fi V C G CO2 where C = molar concentration. The global atmospheric burden of a compound is computed 21 using the constant atmospheric mass matm = 4.879 x 10 Tg and the average molecular mass of air Mair = 28.97 g/mol, i.e. C M m  i i atm Bi 21 , M air 10

where B = atmospheric mass, Ci is in units of ppbv, Mi is in g/mol and Bi in Tg. To estimate global emissions, we used initial averages and maximum concentrations from the literature values. These were global subaerial volcanic, non-hydrothermal CO2 emissions of 66 and 79.2 -1 7 Tgy (Gerlach, 1991), for average and maximum CO2 concentrations (8.092 x 10 and 27.911 x 107 ppbv, respectively), and the average and maximum analyte concentrations in Table 3. The results of the calculations are presented in Table 4.

Chapter II. Halocarbons in volcanic gases 56

6.4. Single compound static flux estimate

The flux of single compounds present in volcanic gases can be roughly estimated for a particular volcanic eruption or fumarolic degassing if the flux of a major compound of sufficient compositional stability (assumption of static emission) is known (e.g., CO2 or SO2), and their molar ratio. As an example of a single volcanic eruption calculation, we evaluated the source potential of the Kilauea (Hawaii) 1977 East Rift eruption. Volcanic gas halocarbon analyses from this eruption are from Rasmussen et al. (1980) where methyl chloride (CH3Cl) was reported for Sept. 20, 1977 and bulk gas analyses were taken from Graeber et al. (1979) for the same eruption, a day later and in the following week. The eruptive phase ended the morning of September 20 after 2 days of intermittent fountaining (Eaton, 1977; Moore et al., 1980) before new activity began on the 23rd and 25th. The of degassing of “fresh” magmatic vapor was therefore probably not possible to observe. The bulk gas analyses are given without air correction and had to be corrected. CO2 concentrations after air removal ranged 0.0187 - 0.1656 mol%.

The Kilauea CH3Cl values are reported as 0.8-2.3 ppbv in the air as close to the eruption as could be done, and 20-200 ppbv sampled directly from gas emanating from the cooling lava flow during the following two years (Rasmussen et al., 1980). A conservative time average of 100 ±50 ppbv is adopted here. For the November 16, 1979 eruption, which was sampled on a day of intense fountaining (Banks and Klein, 1979), as much as 930 ppbv were reported

(Rasmussen et al., 1980). Gerlach and Graeber (1985) estimated the annual Kilauea CO2 mass -1 -3 -1 - flux to be 1.022 Tgy , corresponding to a CH3Cl flux of 5.527 ±5.404 x 10 Tgy (26.9 x 10 3 Tgy-1 for the 1979 eruption).

Scaling the first mentioned methyl chloride value to the current global subaerial volcanic -1 CO2-outgassing of 66 ± 13 Tgy (Gerlach, 1991), a global annual volcanogenic methyl chloride emission of 0.357 ±0.071 Tgy-1 is derived, which is about one third of current anthropogenic emissions (1.074 Tgy-1, including human induced biomass burning) and about half of oceanic emissions (0.812 Tgy-1). The 1979 value corresponds to an annual emission of 1.74 Tgy-1, roughly equivalent to the sum of all other considered sources together. Other estimates of global trace gas fluxes from volcanoes have been attempted on the basis of SO2- emissions (Symonds et al., 1988; Wahrenberger, 1997; Jordan et al., 2000).

Chapter II. Halocarbons in volcanic gases 57

6.5. Dynamic volcanogenic trace gas flux model

Tables 3 and 4 summarize emission data for a variety of halocarbons found in volcanic gases.

We scale the emissions to CO2 rather than SO2 emissions for several reasons:

1. The anthropogenic fluxes of CO2 are better known than SO2 fluxes, and to a similar quality as for volcanic emissions. The resulting values are directly comparable.

2. Sufficient CO2-flux data are available for the volcanoes where halocarbon data are available, especially at Vulcano (Italy). 3. We observe a good positive correlation (R2 = 0.923, n=8, see Figure 4) of CFC-11

concentration with CO2-flux for volcanic soil gas emanations on the La Fossa di Vulcano cone and especially within the crater, where the gases from fumaroles and diffusively escaping soil gases were shown to be of the same origin (Baubron et al., 1990).

6.5.1. Model assumptions

The model atmospheric concentrations and burdens (Table 4) were calculated for a steady- state pre-industrial emission scenario at quasi-equilibrium conditions assuming sufficient mixing with respect to the atmospheric lifetime of a given compound. In reality, the volcanic component of natural emissions operates on two different modi: a) quasi-continuous degassing of non-erupting volcanoes that have a cooling, shallow magma reservoir emitting gases to the surface by exsolution of a vapor phase.

This type of activity operates on time scales several orders of magnitude above the atmospheric lifetimes of the emitted compounds and in our model, is therefore taken to be continuous; and b) explosive volcanism: this type of activity is episodic and creates pulses of emissions into the atmosphere that equilibrate on the order of the lifetime of its compounds, or even longer. Over geologic time, this type of activity is approximately continuous but further modeling and a sufficient database will be required to assess the impact of such eruptions on the atmospheric halogen chemistry. A schematic of these two types is displayed in Figure 1.

Chapter II. Halocarbons in volcanic gases 58

6.5.2. Model calculation

The simple box model used in this calculation assumes a direct relationship between a continuous emission and a fixed lifetime of the compound (IPCC, 2001, chapter 4.1.3). Other approaches are reviewed elsewhere (SORG, 1990). For a given compound i in Tg over a time -1 interval t in years, and assuming zero-order (linear) input rates into the system (rate k1, tg y ; -1 emission fluxes Fi in Table 4), and first-order removal sinks (rate k2, y ), the following differential equation can be used to describe the overall behavior of a compound i in the model system: di  k k [i] dt 1 2

This equation was solved numerically using the Matlab R12 software package as an ordinary differential equation initial value problem using Matlab's "ode45" solver. Input rate values (global emission estimates) were based on published fluxes as specified in Table 4, and own flux estimates in the case of emissions from volcanoes (Table 4). Output rate values were based on published compound half-lives, τ½, and life times, r (Table 4). The initial value of i was either based on known annual additions to the atmospheric burden or was arbitrarily chosen. An arbitrary choice was justified since the initial value of i only affects the time required to reach a steady state, and not the final amount of i at steady state conditions. Once the steady-state atmospheric burden was derived, it was used to compute a realistic input value.

For globally averaged emission scenarios, available literature data of volcanic halocarbon emissions (Table 3) were used to compute the three global model parameters compound flux, atmospheric mixing ratio (molar concentration) and global burden (total mass of a compound in the atmosphere). The results are shown in Table 4. For the model calculation, two scenarios were considered, in which the average concentration of a compound in volcanic gases was scaled to global CO2 fluxes and concentrations. In the first, conservative, scenario (model 1), average global CO2 fluxes and a maximum CO2 gas mole fraction were used and in the second, progressive, scenario (model 2) maximum global CO2 fluxes and average CO2 concentrations were used as input values. Maximum reported compound concentrations are reported in Table 3 as well but were not used in our model calculations.

Chapter II. Halocarbons in volcanic gases 59

(figure caption overleaf)

Chapter II. Halocarbons in volcanic gases 60

Table 4 (previous page) Compilation of natural and anthropogenic source and sink strengths for selected halocarbons found in volcanic gases (for compound details, see Table 3). The sinks are listed as the sink strength: the rate constant k of atmospheric removal, defined as k ≡ ln2 / τ½ with τ½ being a compound’s atmospheric half-life (τ½ = 0.693r, where r is the atmospheric lifetime, the time needed to decrease by 1/e, i.e. to 37% of the starting amount); if k>1, the compound is unlikely to reach the stratosphere, if k<1, it islikely to do trop so. The global average tropospheric mixing ratio is given for reference (ci avg), as well as literature values of G G known global anthropogenic (Fi A) and natural, non-volcanic source strengths (Fi N-V). The literature source data were recalculated to compound molecular mass and include biomass burning classified as an anthropogenic source, since >90% of which are induced by humans (Rowland, 2002). In addition, model source strengths are listed as estimated from halocarbon concentrations in volcanic gases (Table 3), for average and maximum abundance scenarios, also considering average and maximum global subaerial volcanic non-hydrothermal CO2- fluxes after (Gerlach, 1991). The resulting model burdens for pre-industrial volcanic as well as total natural emission conditions are computed (see text), as well as the resulting model mixing ratios in pptv and the percentage of the global pre-industrial burden represented by the volcanic contribution. To understand the significance of such emissions better, the global effects of these trace gas species are given: the global ozone depletion potential (ODP, for a 100-year time horizon, relative to CFC-11), the global warming potential (GWP, 100 year time horizon), as well as the current state of regulation of the compounds by the Montreal Protocol (MP). Abbreviations: c = molar concentration (mixing ratio); trop = tropospheric; k = rate constant; F=flux; G=global; N=natural; A = anthropogenic; V = volcanic; avg. = average; capital letters: data source code (see below). Data sources: K. IPCC 2001; L. McCulloch & Midgley, 1996; M. CMDL 2001; N. Isidorov 1990; O. Bassford, et al. 1999; P. estimate after Rasmussen & Khalil, 1983; Q. Kelly, et al. 1993; R. Atkinson 1994 & 1995; T. UNEP 2000 (Montreal Protocol, incl. amendments until Beijing 1999); U. Khalil, 1999; V. Solomon, 1996; W. Solomon, et al. 1992; various for non-volcanic natural and anthropogenic emissions. Note: non-volcanic source strengths as of 1998, except L (northern hemisphere only, 1992), N (1983), O (1996), and P (1980-83).

1.1.3. Results for steady-state quiescent fumarolic emissions

The model results are displayed in Table 4. The calculated pre-industrial atmospheric model burden shows a discrepancy to the actual observed burden (derived from the measured current global abundance). This is either a reflection of the simplicity of the model, or of an underestimate of the volcanic CO2-fluxes, or of the underestimation of surface sinks. There are still many unknown sources not accounted for in the global methyl chloride budget (Khalil et al., 1999). In addition, the global subaerial volcanic CO2 emission estimate does not include basaltic eruptions from OIB sources alone. In our model calculation (Table 4), global methyl chloride emissions vary “only” between 0.287 and 1.19 x 10-4 Tg y-1. The rate limited model presented here is more likely to produce precise results than the single eruption calculation presented in the previous section. However, the input values are averages of the few published analyses and there is still too little data available on volcanic halocarbon emissions. The individual volcano approach may yield more accurate results, since its input values are specific for its source, however less representative on a global scale. The model discrepancies illustrate the great necessity of starting to set up a database of measurements of halocarbon emissions from volcanoes on a global scale, measurements of which are yet to be performed.

Chapter II. Halocarbons in volcanic gases 61

1.1.4. Results for single catastrophic eruption events

So far we showed calculations for quasi-steady-state emissions on a global scale. We now turn to large-scale perturbations caused by single events. Extremely large eruptions occurred throughout Earth’s history such as the Deccan Trap event 65-69 m.y. ago which may have 3 15 emitted as much as 30 x 10 Tg CO2 (30 x 10 g, ≈ 0.00058 % of the total atmospheric mass) 3 in one eruption episode, while emitting 3.66 Tg of CO2 / km basalt (the eruption released roughly 8.2 x 106 km3 of basalt) (Courtillot, 1990).

Scaling this emission to the CFC-11/CO2 ratio of a volcanic gas, 0.000353 Tg of CFC-11 would have been released over a very short time, creating a tropospheric concentration equivalent to ca. 14.5 pptv, a value that was reached in modern times in the years 1961-62 (Walker et al., 2000), when a measurable ozone depletion was already evident, but not recognized (Rigaud and Leroy, 1990). If Etna is used as an example of a basaltic volcanic gas with a CO2 concentration of 86.7% (D'Allessandro et al., 1997) and CFC-11 of ca. 0.45 ppbv (Jordan et al., 2000), the Deccan event would have put out 15.6 t (1 .56 x 10-5 Tg) at once. On top of the estimated volcanic background loading of the atmosphere of 0.00056 - 0.13608 Tg (Table 4), this injection is not large, creating a positive perturbation of only between 0.01 to 2.76 % of the steady-state background volcanic CFC-11 burden.

Figure 5 displays the CFC-11 atmospheric burden starting with an injection of 1.56 x 10-5 Tg of CFC-11 by the Deccan event on top of the persisting, undisturbed steady-state volcanic background burden. The model results show that depending on which model pre-eruption background burden is used (cf. Table 4), it will take 610 - 760 years to attain the pre-injection conditions again in model 1 and 2, respectively. The simple model calculation used does not represent a non-equilibrium approach applicable to series of discrete injection events. The above figures may thereofore have quite a large error. The simple box model does not take feedback, circulation, layering nor other atmospheric species into account.

Chapter II. Halocarbons in volcanic gases 62

Figure 5 Model results of the exponential decay of a single perturbation event to the pre-industrial volcanic background atmospheric model burden of CFC-11 (cf. Table 4). The perturbation adds to the steady-state model burden, and with time the sinks remove the disequilibrium loading to re-attain the equilibrium background burden from before the perturbation within a time-scale on the order of 610-750 years, depending on the source strength estimated from average and maximum volcanic-gas CFC-11 concentrations (cf. Table 3). The perturbation event is the Deccan trap eruption, and the CFC-11 loading by the eruption is scaled to its estimated CO2-output on the basis of the correlation presented in Figure 4 and the CO2-concentration range of typical basaltic eruptions (Etna, Kilauea).

The actual tropospheric mixing ratios resulting from a single large perturbation event will probably be larger, since the Deccan eruption probably did not perturb the tropopause. Short- term concentrations of current polluted atmospheric conditions cannot be expected from such an event. The very long recovery time from such a perturbation that changes the reactive regimes in the lower stratosphere significantly may lead to long-lasting feedbacks on heterogenous reactions, similar to the current problems encountered with ozone-depletion facilitated by volcanic aerosol loading during times of increased anthropogenic stratospheric halocarbon concentrations (Blanchette and McConnell, 1994; Tie and Brasseur, 1995; Tabazadeh et al., 2002).

Chapter II. Halocarbons in volcanic gases 63

7. DISCUSSION

As a result of the increased sensitivity of our method, halocarbon concentrations higher than previously reported have been detected because these are no longer masked by high water and sulfur backgrounds. In addition, we have been able to obtain quantitative information on a variety of compounds such as simple alkanes, alkenes, and arenes, polyhalogenated, oxygenated and sulfonated compounds, HCFC's, CFC's, and methyl halogenides. The question now arises as to the chemistry of formation of these compounds.

7.1. Origin of source material

Contamination as a source for organic compounds, especially halocarbons, in volcanic gases has been held responsible in several cases (Rasmussen et al., 1982; Gaffney, 1995) but can be ruled out in this study for reasons described above. The origin of the carbon may be from 3 13 pyrolysates or from CO2-mantle degassing. On the basis of CO2/ He and δ C correlations, it was suggested that up to 80% of the carbon in volcanic arc CO2 is recycled by subduction (e.g., Varekamp et al., 1992). The isotopic composition of carbon in magmatic rocks is light (Hoefs, 1973), as also observed for abiogenic hydrocarbons from MOR-fluids at the East 13 Pacific Rise which have δ CCH4 of approximately –9 to –15 ‰ (Welhan and Craig, 1982). Irrespective of the carbon source, inorganic reactions have experimentally and thermodynamically been shown under volcanic gas conditions to result in abiotic hydrocarbons (see above). In this context, the origin of natural oil and gas deposits has been discussed controversially (Robinson, 1966; Porfir'ev, 1974; Gold, 1979; Gold and Soter, 1980, 1982; North, 1982; Welhan and Craig, 1982; Morency et al., 1986; Welhan and Lupton, 1987; Welhan, 1988; Price, 1994; Seewald, 1994; Laughrey, 1998). Several authors noted that an apparent correlation of “fossil fuel” deposits with subduction zone environments coul be explained by arc magmatism not only supplying an adequate heat source and porous reservoir rocks, but also as a potential source of abiogenic hydrocarbons (Hedberg, 1964; Ranneft, 1979; Gold and Soter, 1980).

Previous studies report a hydrocarbon dominated organic compounds in volcanic products while our study shows a heteroatomic (N-, O-, S- and halogen substitution and inclusion),

Chapter II. Halocarbons in volcanic gases 64 aromatic and undersaturated dominance in volcanic gases. This is not surprising considering the methodology of most previous studies (with the notable exception of Isidorov’s work), which did not allow sampling and analysis of very soluble and reactive compounds such as alcohols and halocarbons. In most cases these studies did not separate condensates from analytes during sampling.

Pyrolysate of plant material buried in the volcanic edifice is not evident in our samples, as it would produce a notable odd-even carbon number abundance alternation and a dominance of straight-chain hydrocarbons over more complex molecules. In addition, no biomarkers such as terpenes or terpenoids were identified in our samples, as opposed to the findings by Isidorov et al. (1985) and Isidorov (1990). In addition, Schwandner et al. (2002a) have shown a direct relationship of soil gas halocarbon concentration to the proximity of sampling sites to crater fumaroles. Zolotov and Shock (2000b) required carbon isotopic measurements to prove pyrolysis, however, organic geochemical patterns are independently diagnostic tools to show if pyrolysis plays a role. Carbon isotopic measurements have been published. Kravtsov et al. 13 (1979) found C18-C39 n-alkanes in chloroform extracts (δ C = -2..63‰ vs. PDB), and high- boiling heteroatomic and halogenated and oxygenated PAHs in acetone and alcohol-benzene extracts (δ13C = -2..62‰ vs. PDB). In addition, Beskrony and Lobkov (1977) analyzed

Kamchatkan volcanic ash extracts and volcanic bombs, finding that C1-C3 alkanes and CO2 showed 13C enrichment, thus distinguishing volcanogenic from biogenic hydrocarbons in sedimentary rocks. They observed correlations between the hydrogen content vs. mantle 3 13 3 helium portion ( He) and δ CH4. The mantle derived He increases with hydrogen content 13 13 and with δ CH4 of the hydrothermal sources (see also Isidorov, 1990). Enriched heavy δ C signatures in CO2 and CH4 degassing during earthquakes also point to deep primordial sources (Voitov et al., 1986; Voitov and Yunusov, 1996; Voitov, 2000a, 2000b), however this trend reversed if incomplete Fischer-Tropsch reaction was assumed (Des Marais et al., 1981). Water vapor sampled within eruption ash clouds showed a significant deuterium-depletion, typical for magmatic H2 (Rose et al., 1980).

1.2. Variation with type of magmatic activity and/or tectonic setting

There is an apparent difference between organic compounds found in magmatic gases of “Hot Spot” volcanoes, subduction-related volcanic gases, and intracontinental volcanism (Isidorov, 1990). Ocean island- as well as subduction-related basaltic melts are relatively depleted in

Chapter II. Halocarbons in volcanic gases 65

halogens and release predominantly alkanes, alkenes and PAH’s, and CH3Cl, only small quantities of heterocyclic, other halogenated and oxygenated compounds. Subduction-related volcanic gases are often rich in HCl and HF, and contain a very heterogenous mixture of the organic compound groups, with oxygenated compounds (alcohols and aldehydes) dominating over hydrocarbons, and simple arenes (benzene, toluene) over alkanes, while halocarbons show significant abundances. Intracontinental volcanoes are not well studied in terms of their organic content in volcanic gases but their low abundance of halogens (with the exception of the fluorine-rich lavas of Oldoinyo Lengai) may not give rise to a large emission of halocarbons.

1.3. Processes of formation

A number of processes have been held responsible for the occurrence of organic compounds in volcanic gases and eruption products, including “primary” mantle hydrocarbons (Sugisaki and Mimura, 1994). These include reactions within melts (Heide et al., 2000), catalytic reduction of CO or CO2 (i.e., Fischer-Tropsch reaction), contamination en route to the surface, and pyrolysis of sedimentary organic matter with subsequent cracking and distillation processes (e.g., Capaccioni et al., 1995; Simoneit, 1995; Darling, 1998).

The occurrence of “primary” mantle hydrocarbons has been suggested by findings organic compounds on crack surfaces of peridotitic olivine, in fluid inclusions of mantle minerals, and as continuous meshworks on grain boundaries in mantle xenoliths and undifferentiated mantle melts (Mathez et al., 1984; Mathez, 1987; Sugisaki and Mimura, 1994). They are considered to have been formed by catalytic reactions, although it has also been suggested that they have been exogenically delivered during early meteorite bombardement, or recycled in subduction zones (Sugisaki and Mimura, 1994). Heide et al. (2000) discuss hydrocarbon formation in melts in terms of the reaction:

- - C + 4OH ↔ CH4 + 2O + O2 , (1) and in addition, melt-dependent pathways including Si-C bond formation, network reduction, and polymerization.

Chapter II. Halocarbons in volcanic gases 66

1.1.1. Fischer-Tropsch Synthesis

Two similar processes are potentially involved in the abiogenic hydrothermal synthesis of hydrocarbons: CO2-reduction and CO-reduction (Fischer-Tropsch synthesis). The latter reaction, originally proposed by Brodie (1873), was first used by Fischer (1935) to synthesize hydrocarbon fuel from coal. First, coke is treated with steam to generate carbon monoxide and hydrogen; i.e.

C + H2O ↔ CO + H2 .(2)

These products are then reacted over a metal catalyst to produce light hydrocarbon mixtures, according to the net reaction,

CO + 3 H2 ↔ H2O+ CH4 . (3)

Hydrocarbons derived from Fischer-Tropsch synthesis characteristically exhibit a pattern of decreaseing concentration with increasing carbon number (Schulz-Flory distribution) and have a limited range of carbon numbers (Satterfield and Huff, 1982; Salvi and Williams- Jones, 1997; Giże, 1999). All isomers of a given compound can be expected in the synthetic product, in contrast to a few isomers selected through enzymatic control in biologically mediated processes. Catalytic Fischer-Tropsch synthesis has been suggested in order to account for abiogenic formation of hydrocarbons in a range of chemical environments in the Earth’s crust (Muenow, 1973; Markhinin, 1985; Giggenbach and Goguel, 1989; Isidorov, 1990; Basiuk and Navarro-Gonzalez, 1996; Salvi and Williams-Jones, 1997; Giże, 1999; Zolotov and Shock, 2000). Suitable conditions occur in the mantle (Ballhaus and Frost, 1994; Matveev et al., 1997; Zubkov, 2001), within the crust (Mogk and Mathez, 2000) and in near- surface hydrothermal systems (Seward, 1974; Kelley and Früh-Green, 2001). Several authors have experimentally demonstrated, that serpentinization by thermal hydration of olivine yields excess hydrogen that can catalytically reduce CO or CO2 to hydrocarbons (Morency et al., 1986; Berndt et al., 1996). Alternatively, the formation of lower-carbon chain length hydrocarbons from n-alkanes has been proposed by transition metal catalysis on the hydrophobic domains of kerogen, a process similar to both cracking and Fischer-Tropsch synthesis, to explain the formation of natural gas from pre-existing organic sedimentary matter (Mango, 1992, 1996).

Chapter II. Halocarbons in volcanic gases 67

Under the conditions of an emitting volcanic gas during fumarolic activity as well as during eruptive phases, rapid quenching, fast flow, phase separation and dilution may lead to the occurrence of metastable compounds such as alkynes, acyclic C5-C8 unsaturated hydrocarbons, and radicals (Muenow, 1973; Zolotov and Shock, 2000). In volcanic gases of Vulcano, Chaigneau et al. (1970) previously interpreted aliphatic compounds to have been formed through either a secondary source (contamination) or by reaction of carbon monoxide and hydrogen. An additional gas-phase redox reaction involves carbon dioxide reduction with hydrogen (Giggenbach et al., 2001), controlled by a rock- (usually assigned to FeO/FeO1.5 ) or gas buffer (e.g., H2S/SO2):

CO2 + 4 H2 ↔ 2 H2O + CH4 . (4)

This reaction is also interpreted to be responsible for hydrocarbon formation during serpentinization, whereby hydration reactions lead to transformation of olivine and pyroxenes to serpentinite, brucite and magnetite which yields excess hydrogen (Morency et al., 1986; Berndt et al., 1996; McCollom and Seewald, 2001). Other redox reactions involved in controlling the redox state of a volcanic/hydrothermal fluid/gas (Seward, 1974; Giggenbach and Le Guern, 1976; Giggenbach, 1980) include,

SO2 + 3 H2 ↔ H2S + 2 H2O (5)

H2S ↔ (S) + H2 , (6)

CO2 + H2 ↔ CO + H2O, (7) and

2 NH3 ↔ N2 + 3 H2 , (8)

as well as a variety of possible metal sulfide – water reactions involving H2S, CO2, CH4 and

H2 (Giggenbach, 1980). The formation of carbonyl sulfide, COS, has also been considered and it can be reduced to CS2 to serve as a reaction partner in halogenation reactions (reactions 12, 13) (Giggenbach and Le Guern, 1976):

Chapter II. Halocarbons in volcanic gases 68

CO2 + 3 H2 + SO2 ↔ COS + 3 H2O . (9)

Zolotov and Shock (2000b) calculated the thermodynamic likelihood of abiogenic formation of n-alkanes and PAH’s in a cooling, trapped volcanic gas by Fischer-Tropsch reaction. They observed from their model the well known fact that reducing conditions favor synthesis of purely C-H and C-C bonded hydrocarbons, consistent with practice in industrial organic synthesis. The synthesis of hydrocarbons from CO2 and H2 is energetically less favored than from CO and H2 at temperatures at or below 250°C in the stability field of graphite. At elevated temperatures, PAH’s are stable at slightly higher temperatures than long chain n- alkanes. Chamberlin (1908) suggested hydration reactions of metal carbides as alternative routes to form ethylene (C2H2) and methane in magmatic gases, but metal carbides have to the best of our knowledge not yet been found in fumarolic environments. They have however been reported in group IIICD iron meteorites. Noguchi (1935) explained the abiogenic formation of ethylene in volcanic gases via reaction of two methane molecules with release of hydrogen gas. Lancet and Anders (1970) found experimentally that incomplete Fischer- 13 Tropsch reaction produced isotopically heavy methane and C-depleted C2+ hydrocarbons. This gives rise to a “fingerprint” of abiogenic hydrocarbons that is distinct from biogenic or thermogenic signatures, for which this trend is much less pronounced or completely reversed (Des Marais et al., 1981).

1.1.2. Pyrolysis and Cracking

Pyrolysis and cracking of living or buried organic matter has also been invoked to account for the formation of organic compounds found in high-temperature volcanic gases and eruption products by most of the authors of the 19th century (see above), as well as by modern studies (Gerlach, 1979; Pereira et al., 1980; Pereira et al., 1982; Pereira and Rostad, 1983; Isidorov, 1990, 1992; Giże and Macdonald, 1993; Gribble, 1994b; Capaccioni et al., 1995; Simoneit, 1995; Darling, 1998). For example, oil in the Uzon Caldera contains biomarkers with an aliphatic predominance and contains 14C-matter, documenting recent plant material involvement in the oil generation process, even though its alkenes are largely hydrothermal (Bazhenova et al., 1998). Non-pyrolytic volcanic gas hydrocarbons consequently should be devoid of such characteristics.

Chapter II. Halocarbons in volcanic gases 69

1.1.3. Thermal cracking of pre-existing biologically synthesized molecules

Cracking of pre-existing (biological) organic matter is a process often invoked to account for light hydrocarbons found in low-temperature geothermal sources such as the Porto di Levante (Vulcano, Italy) fumaroles (Mangani et al., 1991; Capaccioni et al., 1993; Capaccioni et al., 1995; Darling, 1998; Capaccioni et al., 2001). Organic compounds derived from this process are termed thermogenic hydrocarbons. It is possible to explain isotopically light hydrocarbons in volcanic gases by this process, since 12C-12C bonds are more easily broken than 12C-13C bonds. In contrast, polymerized carbon chains become isotopically heavier with increasing carbon number (Des Marais et al., 1981). For example, gas seeps in the Green Tuff basin (Japan) have been investigated to elucidate their origin and have high 3He/4He ratios, suggesting a magmatic or mantle origin. They nevertheless have δ13C values which become heavier with increasing molecular weight, indicating cracking of biological organic matter

(Sakata, 1991). Rasmussen et al. (1980) argue that magmatic CH3Cl/CO2 ratios are an order of magnitude lower than those obtained experimentally for high-temperature wood incineration. Some evidence points to a preservation of natural halocarbons in fossils, sedimentary rocks and fossil groundwater (see Gribble, 1994d for review), supported by CFC analyses of deep borehole fluids (cf. Isidorov, 1990). However, it appears more likely that the isotopically lighter hydrocarbons of higher-molecular mass on Vulcano are abiogenic (Gherardi and Panichi, 2001). As an alternative explanation, Fischer-Tropsch synthesis produces isotopically heavier hydrocarbons of lower molecular weight and isotopically lighter hydrocarbons of higher molecular mass.

1.1.4. Pyrolysis of buried biological matter

Enzymatic polymerization via an acetic acid cycle leads predominantly to molecules having an even number of carbon atoms and having distinct isomeric characteristics, unlike abiogenic molecules formed by Fischer-Tropsch synthesis. Since this is not observed in our samples, and due to the volatile rich and highly hydrocarbon-depleted character of our samples, pyrolysis is unlikely to be responsible for the compounds. Muenow (1973) considered pyrolysis reactions were not responsible for the short-chain hydrocarbons found in inclusions of Hawaiian lava, because of the absence of high molecular weight compounds typically associated with pyrolyzed vegetation.

Chapter II. Halocarbons in volcanic gases 70

1.1.5. Serpentinization

Serpentinization of olivine by hydration at 300°C and 500 bar has been shown experimentally to be capable of producing aliphatic and aromatic compounds via reaction 4 (Berndt et al., 1996). In submarine serpentinite seamount fluid seeps, aromatic, parraffinic and carboxylic organic compounds have been found in chimney fluid inclusions and interstitial pore waters (Haggerty, 1991). However, McCollom and Seewald (2001) have demonstrated experimentally that the formation of hydrocarbons by reaction of mantle-derived CO2 with H2 derived from serpentinization of olivine under hydrothermal conditions is less likely but may in fact be generated from more reduced carbon compounds already present within the olivine. The possibility of organic compounds at Vulcano being produced by crustal processes such as serpentinization can not be ruled out. A crustal involvement in the Vulcano magmatic system is indicated by non-magmatic quartz xenoliths and melt inclusions (Zanon et al., 2001). There is additional noble gas isotopic evidence for a crustal component in the Vulcano gases (Tedesco and Nagao, 1996). However, the relatively unsaturated character of the dominant aliphatic hydrocarbons observed in our samples is characteristic of thermal processes but atypical for organic matter associated with serpentinization and we therefore render this possibility as a source of the hydrocarbons unlikely.

1.1.6. Halogenation

Some authors (Stoiber et al., 1971; Cadle et al., 1979) have proposed HCl and HF from volcanic gases reacted with organic compounds to produce organohalogens. A similar type of reaction is the hydrous thermal halogenation reaction studied experimentally by Kanters and Louw (1996), which may be formulated as,

R-H + HCl + 1/2 O2 ↔ R-Cl + H2O. (10a)

For gas-phase reactions, Isidorov et al. (1990) have speculated on the formation of carbon tetrachloride by the free radical substitution of gaseous elemental chlorine for hydrogen in methane, i.e.

CH4 + 4 Cl2 ↔ CCl4 + 4 HCl . (10b)

Chapter II. Halocarbons in volcanic gases 71

Furthermore, Isidorov (1990, 1992) and Isidorov et al. (1990) assumed a fluorination reaction dependent on HF partial pressure in the volcanic gas and involving the substitution of chlorine by fluorine on CCl4 molecules to form CFC’s according to the reaction,

2 CCl4 + 3 HF ↔ CCl3F + CCl2F2 + 3 HCl . (11)

Their proposed reaction remains to be proven experimentally in conditions of saturated water vapor. As an alternative reaction route, Isidorov et al. (1990) suggest the Kolbe-Hoffman process, which involves reacting CS2 with chlorine gas to yield sulfuryl chloride in the presence of a metal catalyst, i.e.

CS2 + 3 Cl2 ↔ S2Cl2 + CCl4, (12) and

CS2 + 2 S2Cl2 ↔ 6 S + CCl4 . (13)

The formation of CHCl2F (HCFC-21) by reaction of chloroform with hydrogen fluoride (a common constituent in volcanic gases) is explained by the same authors as:

CHCl3 + HF ↔ CHCl2F + HCl . (14)

Direct fluorination of alkenes has been discussed by Arimura et al. (1994), and many other synthetic pathways are known. The role of sulfur in the industrial hydrothermal synthesis of fluorocarbons is crucial and may be applicable to volcanic gases, as for example with the classic reaction of Riche (1854):

C2H3S + 8 Cl ↔ C2Cl2 + ClS + 3 HCl (15)

Harnisch et al. (2000) have discussed several processes acting to form fluorocarbons within magmatic environments, including high temperature-high pressure gas phase reactions involving HF and water, catalytic fluorination, heterogeneous fluorination of halogenated precursors and radiochemical production of 19F to randomly fluorinate compounds. Jordan et al. (2000) have suggested that the formation of halocarbons in a volcanic gas may occur via

Chapter II. Halocarbons in volcanic gases 72

the reaction from CH4 to acetylene with subsequent halogenation to form precursor molecules of halogenated hydrocarbons. However, if that were the case, only even carbon numbered would be present, which is not the case in their list of compounds.

1.1.7. Heteroatomic reactions excluding halogenation

Sulfonated, oxygenated and nitrogenated compounds are given in volcanic gas analyses of several authors (see above), including this study. Oxygenation of compounds by formation of a functional group may occur by oxidation upon first mixing with surrounding air at fumarole emission points. Under such conditions of drastic temperature decrease accompanied by oxidation of CO to CO2 (or of H2S to SO2) and in attend condensation of water from gas to subcritical vapor, oxidized radicals readily dissociate into the newly formed vapor droplets. A well known reaction of this type is the reaction of carbon monoxide in water to form formate (Giggenbach and Matsuo, 1991), i.e.

- + CO + n H2O ↔ HCOO + (n-2) H2O + H3O (16)

Oxygenation to heterocyclic oxocompounds may occur by ring closure of oxygenated alkenes on the site of the double bond. Sulfonation in the gas phase is known to proceed by several redox reactions. In high-temperature volcanic gases (>600°C), reduction of SO2 to CS2 with oxidation of CH4 to CO2 may be the dominant process, as given by the equilibrium,

2 SO2 + 3 CH4 ↔ CS2 + 6 H2 + 2 CO2 , (17)

or, by reduction of CO2 by SO2, as is probable in eruption clouds, which may be expressed by the reaction,

CO2 + 2 SO2 ↔ CS2 + 3 O2 . (18)

In a medium- to low-temperature volcanic gas (<600°C), H2S dominates over SO2, and thus reactions of the type,

CH4 + 2 H2S ↔ CS2 + 4 H2 , (19)

Chapter II. Halocarbons in volcanic gases 73 take place (e.g., Petherbridge et al., 2002) and may involve hydrocarbons larger than methane. As a possible product, thiophene can be formed in the gas-phase by the reaction,

- + C4H3S + H ↔ C4H4S (20) as noted by DePuy et al. (1988). The formation of nitriles in volcanic gases has been described by Navarro-Gonzalez et al. (2001) and experimentally demonstrated by

Pietrogrande et al. (2001) to be readily derived from catalytic reactions with CH4 and N2 .

1.1.8. Implications for high-temperature volcanic gases

Because all of the above reactions can occur in combination or succession, three approaches to understanding the formation of organic compounds in volcanic gases are experimental, thermodynamic modeling, and analysis in natural systems. In this study, we focus on analysis, which can later be explained by thermodynamic modeling and supporting experimental data. Figure 6 shows four chromatograms, a) a volcanic gas sample, b) ambient polluted air (Vulcano harbor during ferry docking and vehicle traffic), c) Mt. St. Helens 1980 volcanic ash extracts with plant pyrolysates (Pereira et al., 1982), and d) a synthetic Fischer-Tropsch product (Davis et al., 1999).

The chromatograms a-c are limited to the hydrocarbon fraction by displaying a reconstructed ion chromatogram (RIC) of m/z = 57, chromatogram d is by the nature of its experimental conditions already limited to hydrocarbons. The Fischer-Tropsch product (chromatogram d) was synthesized by reacting a CO-H2 gas mixture over an iron-based catalyst (Davis et al., 1999). The numbered peaks are n-alkanes, the next prominent peak series are their corresponding n-alkenes that closely elute before the n-alkanes of the same carbon number. The next prominent peaks are mono-unsaturated isomers. The Schulz-Flory distribution (Satterfield and Huff, 1982) is not immediately obvious but can be seen for the n-alkenes. Sets of same isomers of a given carbon number hydrocarbon are repeated in multiplets.

Chapter II. Halocarbons in volcanic gases 74

(figure caption overleaf)

Chapter II. Halocarbons in volcanic gases 75

Figure 6 (previous page) Hydrocarbon distributions from four different sample types: a) RIC m/z 57 (reconstructed ion chromatogram) of volcanic gas analysed in this study, b) RIC m/z 57 of ambient polluted air (contamination) in Vulcano harbour during ferry loading with trucks; c) RIC m/z 57 of volcanic ash extract from Mt. St. Helens 1980 sampled directly after the eruption by Pereira et al. (1982), the organic source was interpreted as pyrolysed vegetation on the basis of abundant biomarkers and terpenoids; d) FID-TIC of an experimental product from Fischer-Tropsch-Reaction synthesized in a CO-H2 gas stream over an iron-based catalyst (Davis et al., 1999). Discussion in text.

The only other chromatogram of a-c that shows a hint of this pattern is b) where a predominance of aliphatic hydrocarbons leads to clearly distinguishable, large peaks of homologues. In this chromatogram, a step in carbon number distribution can be seen between the C13 and C14 isomer peaks. From C14 to C17, the alkane concentrations decrease and are not detectable over C18. By comparison, diesel fuel has its strongest concentrations in the C14-C17 alkane range whereas unleaded gasoline contains concentrated C4-C6 alkanes. In the harbor traffic sample taken on the dock, residuals of incomplete diesel combustion can be seen, reflecting ferry loading with truck traffic. Experimentally, the thermal stability of C15+ hydrocarbons is limited if temperatures >350°C are encountered (Price, 1993).

Pyrolysed vegetation in volcanic ash fall samples from Mt. St. Helens (Figure 6 c) show, however, that even at temperatures superceding these experimental temperatures for a short time, alkanes up to C29 are preserved. A notable odd/even predominance occurs in these samples, reflecting that in the original plant matter (Pereira et al., 1982). As these authors note, the unimodal distribution with a maximum at C25 indicates a single source of the compounds. Aqueous hydrolysis experiments performed on organic matter in sulfur-rich marls show carbon-heteroatom cleavage occurring at temperatures as low as 220°C within three days (Putschew et al., 1998), leading to a strong depletion in heteroatomic molecules. Pyrolysis of bitumen leads to alkane carbon number distributions that still have their strongest peaks at C16/C17, in both anhydrous and hydrous conditions (Schlepp et al., 2001).

In light of all these observations, the hydrocarbon distribution found in the volcanic gas sample can neither be contamination by fuel exhaust (chromatogram b), nor pyrolyzed vegetation (chromatogram c). Neither are any biomarkers present, nor is any odd/even predominance observable. The pyrolyzed vegetation in chromatogram (c) has experienced much greater temperatures than the Vulcano volcanic gas (chromatogram c) near the surface. In the case of Vulcano, little if any vegetation could have been incorporated in the volcanic edifice during the last eruption (1880’s), if present. In addition, halocarbon concentrations

Chapter II. Halocarbons in volcanic gases 76 increase in volcanic flank soil gas from the more distal, vegetated sampling sites towards the unvegetated crater and fumaroles (Schwandner et al., 2002a). Pyrolysis and cracking have been postulated for the Vulcano beach fumaroles (Capaccioni et al., 1995; Capaccioni et al., 2001) but this pattern cannot be established for the crater fumaroles on the basis of above discussed observations.

Fischer-Tropsch synthesis may play a role in the fumarole gases, especially since there is no isomeric preference or absence of specific isomers in the volcanic gas sample. The normal alkanes from C9 to C12 approximate a Schulz-Flory distribution, which may indicate Fischer- Tropsch synthesis. Based on the range of isomers present, we deduce these compounds to have been synthesized by gas-phase reactions in an ascending gas column together with subsequent abiogenic alkylation, halogenation, and sulfidation. Although the roles of surface catalysis by mineral surfaces at high temperatures are still poorly known, the organic distribution is interpreted as produced by high temperature (e.g., 900 °C) alkyl free radical reactions, and halide electrophilic substitution on arenes (e.g., Weigert and Davis, 1993). In contrast to volcanic gases, hydrocarbon distributions in submarine vent fluids and in associated sediments have a biogenic signature (Leif and Simoneit, 1995) not detected in high-temperature volcanic gases.

Chapter II. Halocarbons in volcanic gases 77

8. IMPLICATIONS FOR FURTHER RESEARCH AND OUTLOOK

8.1. Atmospheric chemistry

8.1.1. Natural halocarbon emissions

In certain largely government funded scientific disciplines, the statement that halocarbons are solely anthropogenic seems to persist in an almost dogmatic manner. Recently however, an increasing number of publications have admitted the possibility of natural, including volcanic, sources of halocarbons (e.g., Golitsyn et al., 1991; Gribble, 1992b, a, 1994a, b, c, d, e, 1996a, b, 1998, 1999, 2000, 2001; Gribble and Worman, 1998; Butler et al., 1999; Khalil et al., 1999; Hoekstra et al., 2001). Strong natural sources such as volcanism have already been discussed in the mid-70’s (Gassmann, 1974; Lovelock, 1975; Rowland and Molina, 1975; Wilkniss et al., 1975; Rowland and Molina, 1976). In 1999, the first acknowledgement of Isidorov’s work on volcanoes (Isidorov et al., 1990) as natural sources of chlorinated and fluorinated halocarbons (i.e. CFC’s & HCFC’s) came from the side of the dominating scientific community of atmospheric chemists (Aucott et al., 1999). They cited Isidorov’s work on Kamchatkan volcanoes, with CFC-11 and -12 as being naturally emitted from volcanic vents; however, they explicitly mention HCFC-22 as not being produced, although it had already been found in volcanic gases in the early 1970’s (Stoiber et al., 1971). A number of non- volcanic natural halocarbon sources were identified in the past, as reviewed by Gribble (1992b, a, 1994c, b, 1995a, b, 1998, 1999, 2000), Khalil et al. (1999), and Butler (2000). The role of anthropogenic and natural factors modifying the Earth’s ozone layer has been a matter of controversial public debate (Maduro and Schauerhammer, 1992; Hanson, 1993; Gribble, 1994c, 1995c, 2001; Gaffney, 1995; Isidorov, 1995).

In our study, we have found a great variety of halocarbons in volcanic gases that are characterized by small Henry’s Law constants (<< 1.0, see Table 3). Even though halocarbons are present in volcanic gases at much lower abundances than the principal volcanic gas chlorine species, HCl, their potential atmospheric impact is significant due to their lesser solubility similarly to anthropogenic emissions. They should therefore be taken into consideration in atmospheric chemical trace gas models. In Table 4, several atmospheric

Chapter II. Halocarbons in volcanic gases 78 parameters of the photochemically active compounds found in this study are indicated for clarification, where available. These are the international status of the compound in terms of its status under the Montreal Protocol (MP) and amendments (banned or controlled substances are marked in Table 4); the short term ozone creation potential (POCP) relative to ethylene, affecting mainly ground level tropospheric ozone increases; the ozone depletion potential (ODP) relative to CFC-11; the atmospheric half-life (t½), and the global warming potential (GWP, relative to CO2). In addition, each compounds’ volatility is indicated in the table by its’ Henry’s law coefficient (KH) and the boiling point (b.p.).

8.1.2. Past air archives

Polar ice cores offer a convenient access to natural past air archives. The finding of CFC’s in air bubbles entrapped in Antarctic ice that was deposited in pre-industrial times was attributed to vertical diffusion of modern-day atmospheric contaminants, but in some cases (Tunu,

Antarctica), increasing concentrations of CH3Br with depth could not be sufficiently explained by diffusion (Butler et al., 1999). In some polar firn air archives “locked” to vertical diffusion, CCl4, CH3Br and CH3I even show “anomalously” high concentration in pre- industrial air that cannot be explained by anthropogenic emissions (Butler et al., 1999; Sturges et al., 2001). In the same ice cores, other halocarbons trend to zero concentration (below detection limit) towards the base of the profile (ca. the year 1880 in the study of Butler et al., 1999). The same was observed for CH3Br data at Siple Dome (Antarctica) and CH3I data at Tunu (CMDL, 1999) but remained unexplained.

In an earlier study (Penkett et al., 1981), anthropogenic CH3Br emissions had been estimated to be only 10% of the natural input. Additional data from three polar stations yielded similar results that were interpreted as being derived from in-situ formation of CH3Br and CH3I (Sturges et al., 2001). However the observed up-section trends of Antarctic pre-industrial halocarbons are neither readily explained by the section being open to diffusion, nor by in-situ production and a sufficient explanation remains missing. Butler et al. (1999) admitted that

“based upon current understanding, there is a large source of this gas [CH3Br] that has not been identified”. They concluded, that “CH3Br and CCl4 are, for unknown reasons, not always conserved in firn air, and other reactive halocarbons could conceivably behave similarly”.

Chapter II. Halocarbons in volcanic gases 79

In parallel, volcanic ash-derived trace metal and halogen records were reported from similar profiles that document historic eruptions, and at the prime measurement station for polar ozone research, McMurdo Station downwind of Mt. Erebus, sulfate and especially PM10 aerosols (PM10 = particles Ø < 10 µm) have recently been attributed to proximal Mt. Erebus emissions, accounting for up to 24% of the total PM10 precipitation (Mazzera et al., 2001). Since volcanogenic halocarbons have been desorbed from volcanic ash before (Stoiber et al., 1971; Wilkniss et al., 1975; Markhinin and Podkletnov, 1977b, 1977a; Kravtsov et al., 1979; Pereira et al., 1980; Rasmussen and Rasmussen, 1980; Rasmussen et al., 1980; Rasmussen et al., 1981; Rasmussen et al., 1982), a volcanic source cannot be ruled out as a contributor of the observed halocarbon abundances to Antarctic snow and ice core profiles.

The dominant halocarbon in the atmosphere is methyl chloride (CH3Cl), which is known to have strong natural sources including bacterial and algal activity, biomass burning, sea-salt spray and volcanic gases (e.g., Lovelock, 1975; Rasmussen and Rasmussen, 1980). Brominated and iodinated halocarbons are of special interest becauase of the by far higher ozone depletion potentials. Our finding of several brominated halocarbons and methyl iodide necessitate the consideration of their atmospheric impact in the context of this study.

1.1.3. Brominated halocarbons

Bromine is estimated to be at least 40 times as efficient in stratospheric ozone destruction as chlorine in polar regions, and possibly 100 times more efficient in mid-latitude regions

(Solomon et al., 1992; Butler et al., 1998; Daniel et al., 1999). Methyl bromide (CH3Br) is the single largest carrier of bromine to the stratosphere (Butler, 2000). Currently, 32% of the

CH3Br sources are anthropogenic (incl. biomass burning), 38.8% are natural (including oceans), but 29.2% are unaccounted for (WMO, 1998; Butler, 2000). Naturally occurring brominated compounds have been reviewed recently (Gribble, 1992a, 1999, 2000; Reeves and Penkett, 1993; Varner et al., 1999; Bureau et al., 2000; Butler, 2000) and entered public scientific debate (Gribble, 1992b, 1994e; Butler, 2000). In order to have volcanoes contribute significantly to the natural pre-industrial atmospheric CH3Br budget, at least 2 ppm would -1 have to be present in any volcanic gas, assuming a global CO2-flux of 66 Tgy (a flux of -1 CH3Br of 0.00044 Tgy , thus 0.2% of the global flux from all sources). However, CH3Br may not be the only volcanic organic Br carrier.

Chapter II. Halocarbons in volcanic gases 80

CH3Br, CH2Br2 and the Halons H-1301 (CBrF3), H-1211 (CBrClF2), H-2402 (C2Br2F4) constitute the bulk upper tropospheric organic bromine and thus are significant carriers of Br to the lower stratosphere (Butler et al., 1998), where CH3Br makes up approximately 55% of the total organic bromine burden (Lee-Taylor et al., 1998, and references therein). The pre- 1986 Halon concentrations are poorly constrained and their atmospheric abundance has been steadily increasing until 1994 (Butler et al., 1998). Reeves and Penkett (1993) estimated the anthropogenic contribution to the atmospheric bromine burden by methyl bromide to be 54%, and conclude that “our understanding of the way in which CH3Br is produced naturally must be increased in order to quantify its source strength”. New discoveries of sources are sought in order to match observed mixing ratios (molar concentrations) and to balance source-sink relationships (e.g., Varner et al., 1999).

Halons were mainly used as fire retardants due to their low flammability. They are characterized by very long atmospheric lifetimes, making them an efficient halogen carrier to the stratosphere. Halons are phased out by the Montreal Protocol for production and sale in developed nations as of January 1st, 1994. They are generally considered solely of anthropogenic origin and their atmospheric abundance is expected to remain growing (e.g., Butler et al., 1998). They may be formed in volcanic gases as well, as indicated by findings of Halons 1011, 1012, 1021, 1031, 2011 and 2031 by (Jordan et al., 2000), and analyses from several volcanic eruptive emissions by Rasmussen and Rasmussen (1980). Wamsley et al. (1998) estimated a relatively high photolysis rate of Halons in the lower stratosphere. The combined non-volcanic anthropogenic and natural source strength has been estimated as 0.1 Tgy-1, and the volcanic source was estimated to be within the same order of magnitude (Bureau et al., 1999; Bureau et al., 2000 and references therein). The same authors considered the observed ozone depletion observed after the Pinatubo eruption (see above) to be likely due to the release of bromine, estimated to be on the order of 0.011 to 0.025 Tg.

1.1.4. Iodinated halocarbons

Iodine is weakly bonded to carbon in iodocarbons, giving them too short of an atmospheric half-life (i.e. 4 days for iodomethane, Solomon et al., 1994) to be of significance for the stratosphere from tropospheric sources. This statement excludes large Plinian eruption plumes that can carry matter from the surface into the stratosphere within minutes. Transport models and observations of tropical convection provide insights into the possible transport of

Chapter II. Halocarbons in volcanic gases 81 iodinated gases to the tropopause (Solomon et al., 1994). These authors concluded a strong natural source for iodomethane and a relatively insignificant anthropogenic input in order to balance observed atmospheric concentrations of 1-10 pptv.

Few tenths of 1 pptv in the stratosphere are already considered significant for ozone loss. This figure is two orders of magnitude below that of CFC-11. It was estimated that iodine released in the stratosphere is >1000 times more efficient in ozone destruction than chlorine (Solomon et al., 1994). CH3I has been consistently found in volcanic emissions during eruptive periods, on Hawaii (Mauna Loa and Kilauea), St. Augustine (Alaska), and Mt. Erebus, Antarctica (Rasmussen and Rasmussen, 1980).

1.1.5. Atmospheric transport and fate of halocarbons

Halocarbon concentrations below the tropopause are always higher by several orders of magnitude than in the stratosphere, which contains the seasonally and locally variable ozone layer responsible for absorption of incident solar UV radiation. Penetration of the tropopause would allow the ascent of long-lived halocarbons and occurs primarily as a result of rising tropical air masses, rare turnover events, or large Plinian volcanic eruptions. In contrast, permanent slow emissions are carried south within the troposphere, rising in the Hadley convection of tropical latitudes, where they may ascend into the stratosphere. They are transported to the polar regions within the stratosphere, largely by the lower stratospheric jet stream. Volcanic gases are a relatively weak but continuous background source of halocarbon emissions to the atmosphere in industrialized times - but almost certainly a strong but yet poorly understood contributor and modulator to the natural halogen chemistry and ozone balance of the pre-industrial stratosphere (Golitsyn et al., 1991). This fact has been neglected in all atmospheric models predicting future ozone loss. We thus foresee a great necessity for further research of volcanic trace gas discharges in both atmospheric and geochemistry.

1.2. Volcanic gas studies, exploration and volcanic surveillance

The F/Cl ratio of volcanic gases is used as a tool for monitoring active volcanoes, reflecting the different solubilities of the two halogens in andesitic melts (Giggenbach, 1996; Giggenbach, 1997). Apart from the halogen systematics that may be altered by conditions

Chapter II. Halocarbons in volcanic gases 82 favouring fixation or release by substitution onto hydrocarbons, the organic chemistry of a volcanic gas discharge is redox sensitive and may therefore provide insight into changing proportions of meteoric vs. magmatic signals in the mixing zone of the hydrothermal envelope surrounding an intrusive body. Because of the time-consuming analytical technique, organic chemical information is currently not useful for short-term monitoring of “dangerous” volcanoes, unless operator-intensive field gas-chromatographic methods were employed.

1.3. Metal transport in hydrothermal fluids

Fluorine and chlorine are competing ligands in volcanic gases, acting in principal by the equilibrium reaction of a gaseous halogen acid with a solid metal oxide to form a MeX (X=F or Cl) gaseous complex and water (Schrön, 1989). The importance of halogenated species in metal transport processes in lunar volcanism was discussed by Fegley (1991). The gas-phase transport of trace metals by inorganic as well as organic ligands and resulting deposition of metals and solid halides as sublimates in a cooling volcanic gas has been assessed by Symonds et al. (1992) for Augustine Volcano (Alaska), and by Wahrenberger et al. (2002) for Kudriavy (Kuriles), who emphasize the importance of oxoacids, sulfides, Br, I, and dominantly chlorine for gas-phase metal transport. Organic matter is abundant as a major compound group in certain ore deposits (Giże, 1987; Giże and Manning, 1993; Landais and Giże, 1997; Giże, 1999). Organic compounds play a crucial role in the formation and alteration of some ore deposits (Giże, 1999), and for the availability and stability of organic ligands involved in metal transport in ore-forming hydrothermal fluids (Manning and Giże, 1993; Seward and Barnes, 1997).

In low-temperature oil field and basin brines, the dominating complexes are after chloride probably, mono- and bidentate carboxylic anions (acetate, propanoic, and malonic), by concentration. In petroleum liquids, metalloporphyrins have the greatest transport potential. Petroleum is thought to be an important control for metallic mineralization of such large economic hydrothermal mineralizations as the El Soldado Cu deposit (Chile), located in andesites (Wilson and Zentilli, 1999). The role of organic complex formation at the high temperatures of a volcanic gas is to date poorly constrained, but our quantitative data may give a first insight into compounds that may play a role, by abundance. Our finding of considerable quantities of short-chain oxygenated molecules of high volatility, especially carboxylic acids (notably acetic acid), ketones (acetone and homologues), as well as esters

Chapter II. Halocarbons in volcanic gases 83 and aldehydes, as opposed to heavier, more complex molecules of higher molecular weight and carbon number, is interesting in that context.

Since toluene is the by far most abundant organic molecule, its transporting capabilities under hydrothermal conditions should be further investigated. Benzenes and substituted derivatives including chlorobenzene, are known to associate with trace metals, oxygen, Ar, Ne and halogens in aqueous and vapor phase (Emmenegger, 1987; Casero and Joens, 1997; Froudakis et al., 2001; Kelly, 2001; Korolev et al., 2001; Pandey et al., 2001; Bakker et al., 2002). The strong surface complexation properties of arenes with Pt (Kryachko et al., 2001) necessitates avoiding Pt tubing in sampling of a volcanic gas, if aromatic compounds are sought. Some of the compounds found in our samples are condensed versions of their probably radical counterparts in the gas phase, and formed during sampling.

Fluorinated hydrocarbons have recently been used as active (transient) tracers in some geothermal sytems, to monitor re-injected fluids in geothermally exploited areas (e.g., Adams et al., 2001). Such methods will have to be used with caution to prevent costly misinterpretations, if naturally formed fluorocarbons mix with tracers. Results from such studies will have to be limited to well-characterized, steady-state natural effluent sources or wells. The upper thermal stability limits for halocarbons are poorly known for volcanic gas- or hydrothermal fluid matrices.

Experimental data for methanol, propanol, SF6, C2H2F4 (HCFC-134a), CHF3 (HFC-23) and

CClF3 (CFC-13) in water-steam-nitrogen mixtures with and without rock chips at up to 310°C, varying pH and salinity in quartz-glass ampules for up to two weeks indicate a very slow decomposition at hydrothermal conditions compared to the time needed for ascent of fluids (Adams and Kilbourn, 2000; Adams et al., 2000), making them a suitable active tracer in geothermal injection wells if no natural contribution is assumed (Adams et al., 2001). Halocarbons are also used as passive tracers in hydrology, assuming the source to be only anthropogenic and the atmospheric growth rates to be known (Hofer and Imboden, 1998). If the latter assumption is valid, a groundwater model age can still be established with a natural halocarbon background source, but lower trace abundances would have to be interpreted with caution.

Chapter II. Halocarbons in volcanic gases 84

1.4. Origin of life and abiogenic hydrocarbons

Alkyl thiols such as methyl sulfide, dimethylsulfide, and ethylsulfide in volcanic gases may react with carbon monoxide at the presence of iron sulfide to form carbonylated organometallic intermediates that are capable of producing pyruvate (Cody et al., 2000). Formic acid can be a source of carbon monoxide as well as the result of carbon monoxide dissociation in a condensed aqueous phase and is in this study found in volcanic gases and their condensates. Nitriles are thought to play a crucial role in prebiotic chemistry as well, as emphasized by Navarro-Gonzalez et al. (2001). At last, the finding of apparently de-novo synthesized amino acids may support the theory of a volcanic association to the origin of life.

Chapter II. Halocarbons in volcanic gases 85

9. SUMMARY

Organic trace gas emissions of organic compounds including halocarbons from volcanoes are of interest to volcanology, high-temperature fluid geochemistry, economic geology, astrobiology, and atmospheric chemistry. In volcanic hazard assessment and monitoring, the observation, quantification and detailed chemistry of diffusive as well as fumarolic vent emissions is of central importance for early warning systems.

On the island of Vulcano (Aeolian Islands, Italy) we sampled volcanic gases, sublimates and condensates from active vents of the La Fossa crater as well as ambient and industrial air, in 1999 and 2000. Among the well over 100 detected and quantified compounds are several alkanes, alkenes, arenes, phenols, aldehydes, carboxylic acids, esters, ketones, nitriles, furans, PAH's and their halogenated, methylated and sulfonated derivatives, and various heterocyclic compounds including thiophenes. A large number of the reported compounds are found well above laboratory, ambient air, adsorbent and field blank concentrations. For some analytes

(e.g., CFC-11, CH2Cl2, CH3Br), concentrations are up to several orders of magnitude greater than even midlatitudional industrial urban air maxima. We took special care to avoid and quantify air and laboratory contamination by the collection and analysis of field, equipment, ambient air and laboratory air blanks. In addition, we quantified the amount of air entrained during sampling by analyzing noble gas isotopes contents of the volcanic gases from focused through the sampling train.

In this study we present a method to sample and analyze high-temperature volcanic gas emissions from fumaroles for organic trace gases, with special emphasis on halocarbons. The sampling technique is based on the separation of water, halogen acids and sulfur from organic compounds present at trace levels in volcanic gases by flow-through steam condensation and subsequent sorbent trapping of organic compounds on mixed or single bed adsorbent cartridges, and preservation of the condensate headspace. For analysis, we developed and applied Short-Path Thermal Desorption-Solid Phase Microextraction-Cryotrapping-Gas Chromatography/Mass Spectrometry (SPTD-SPME-CF-GC-MS) as a reliable standard protocol, with an external calibration procedure allowing quantification of trace organic compounds in the range of 0.12-12 ppbv. These techniques are applicable to any hot (>350°C) steam-saturated, corrosive sulfurous emissions.

Chapter II. Halocarbons in volcanic gases 86

We observed a very good correlation of CO2 fluxes from diffusive soil degassing to halocarbon concentrations. This correlation provides a safe trace gas source strength quantification method that is more reliable than SO2 as a reference gas since it does not adsorb or react as easily and preserves subsurface flux ratios. CO2 degassing data are independently available from soil gas mapping as well as remote COSPEC observations, whereas SO2 degassing can only be quantified by remote methods. Apart from this first suggestion to use CO2 rather than SO2 in trace compound flux calculations, it indicates a geogenic source or process giving rise to emissions of halocarbons. This observation is supported by organic geochemical evidence that shows that organic compounds are not of ambient air or pyrolytic origin but rather the product of inorganic abiogenic gas-phase radical reactions. Linearly falling concentrations of n-alkanes heavier than n-nonane (Schulz-Flory distribution) suggest catalytic Fischer Tropsch reaction to be responsible for the formation of simple hydrocarbons. On the basis of isomer abundances, n-alkane distributions and the substitution pattern of the more complex compounds we interpret them to have been formed by high temperature (e.g., 900 °C) alkyl free radical reactions, and halide electrophilic substitution on arenes, alkanes and alkenes.

The significance of organic emissions from quiescent volcanic degassing as opposed to explosive eruptions to stratospheric chemistry is that the negligible aqueous solubility of most emitted organic compounds does not lead to a quantitative tropospheric washout effect as postulated for explosive volcanic HCl and HF emissions. The halocarbons emitted by this quasi-continuous type of volcanic activity are therefore exposed to the same atmospheric transport and removal mechanisms as anthropogenic emissions and their source strength is directly comparable. Using own and literature data on halocarbon emissions from volcanoes, we calculated a simple atmospheric box model to estimate source strengths of selected compounds found in volcanic gases, and to assess the potential impact of a large explosive eruption.

Source strengths of several methyl halides are (ranges are from conservative and progressive -6 -1 -6 -1 model parameters, respectively): CH3Br 0.3-1.25 x 10 Tgy , CH3I 0.49-2.01 x 10 Tgy , -4 -1 -9 -4 CH3Cl 0.29-1.19 x 10 Tgy , and other halocarbons and hydrocarbons range from 10 - 10 Tgy-1. Even though the fluxes are minor relative to current anthropogenic releases, they have operated throughout geologic history and created a low but constant background burden to the Earth’s atmosphere and thus participated in stratospheric heterogenous reactions and catalytic

Chapter II. Halocarbons in volcanic gases 87 ozone destruction. Rare episodic catastrophically large explosive eruptions such as the Deccan flood basalt eruption did not impact the short-term ozone chemistry to the same magnitude as today’s anthropogenic emissions, but modified the natural atmospheric balance for centuries until this chemical perturbation is consumed by the natural sinks and reaches the pre-eruption steady-state background conditions again.

Compounds of ring geometry and/or polar functional groups are capable of complexing metal ions in solution and in the vapor phase, acting as metal transporting media. The respective presence of halocarbons, gaseous halides, metal-organic complexes and other organic compounds may in conjunction give rise to a better understanding of metal transport mechanisms in volcanic gases. The presence of such polar molecules also alter the dielectric properties of fluids in the shallow subsurface, and the knowledge of their presence and properties at high temperatures aids interpretation of magnetotelluric exploration results in active hydrothermal systems. In addition, the apparent abiogenic organic chemistry of volcanic gases may give insights into the possible hydrothermal origin of early life on Earth. The presence of simple amino acids, nitriles, and alkanoic acids, metal catalysts and other compounds supports such theories.

Chapter II. Halocarbons in volcanic gases 88 III. Diffusive emission of organic trace gases from the flank and crater of a quiescent active volcano (Vulcano, Aeolian Islands, Italy)

“I would love to know more facts about halogen compounds that come out of volcanoes. If I went to a federal agency and asked to get funded for such a study, they would probably throw me out. Why? Because looking at this question assumes the need to look further into issues related to possible sources of presumed depletion of ozone. It would also suggest that the concept that man-made CFCs are the greatest threat to the ozone layer may not tell the whole story.”

Prof. Dr. Frederick Seitz, former President of the National Academy of Sciences, past President of the American Physical Society, President Emeritus of Rockefeller University, in an interview with the George C. Marshall Institute on September 3, 1997. 90

Abstract

Volcanoes discharge a range of major gas species (i.e. H2O, CO2, SO2, CO) not only during eruptions but also in quiescent phases, through fumarolic as well as diffusive degassing in their craters and on their flanks. The emission of trace gases such as natural halocarbon compounds from volcanoes is in the same regard not expected to be restricted to discrete fumarolic gas discharges alone. To test this hypothesis, we have sampled soil gas emissions of organic compounds and determined CO2 fluxes along a profile spanning from the vegetated base of the active La Fossa cone (Vulcano Island, Italy) over the unvegetated volcanic flank, the crater rim and base, into fumarolic areas.

The results indicate that a) the majority of volatile organic compounds in the soil gas show significant increases in concentration towards the crater and fumaroles, b) emissions of the 2 halocarbon CFC-11 (CCl3F) correlate well with soil CO2 fluxes measured on site (R =0.91 ±0.02) and both increase towards the crater and fumaroles, and c) diffusive emissions therefore contribute significantly to the volcanic halocarbon source strength. Other ozone depleting substances were found in concentrations significantly above those found in field and system blanks, including CH3Br, CH3Cl, CH3I, C2H5Br, and chlorinated benzenes. Abundances range from upper pptv to ppmv (e.g. CFC-11: max. 1200 pptv diffusive, 3700 pptv fumarolic/dry gas, dry air: 268 pptv). On the basis of these results, the natural volcanic source strength of halocarbon emissions to the atmosphere requires re-evaluation and correction to higher values.

Chapter III. Diffusive emission of organic trace gases from volcanoes 91

1. INTRODUCTION

SINCE THE dimunition of the Earth’s ozone layer by the increased emission of anthropogenic ozone-depleting substances (ODS) such as chlorofluorocarbons (CFC), halons (brominated halocarbons) and perhalocarbons became evident in theory and observation (Molina and Rowland, 1974; Farman et al., 1985), a large number of studies have addressed the question of a natural contribution to global ODS emissions that may be responsible for a pre-industrial constant background burden previously unaccounted for in atmospheric growth models (Isidorov, 1990; Gribble, 1992a, 1994d, 1996b, 1998, 1999, 2000; Butler et al., 1998; Khalil et al., 1999; see previous chapter). A range of inorganic and organic natural sources have been identified. Among the most prominent inorganic natural sources are sea-salt aerosol (Keene et al., 1999), volcanism (Wilkniss et al., 1975; Rasmussen et al., 1980; Isidorov et al., 1990; Wahrenberger et al., 1996a; Jordan et al., 2000; Schwandner et al., 2000b; Schwandner et al., 2001a), the oxidation of soil organic matter (Keppler et al., 2000) and crustal degassing (Harnisch et al., 2000). Organic sources include biomass burning (Lobert et al., 1999), fungal metabolic activity (Khalil et al., 1999) and processes in soils (Khalil et al., 1999). Biomass burning is today considered to be approximately 90% anthropogenically caused, on a lower scale however it may have been a natural contributor in the past (Levine, 1998).

For brominated compounds, methyl bromide (CH3Br) is the major carrier of atmospheric bromine, with natural sources accounting for up to 46% (Reeves and Penkett, 1993) of the total. Natural sources of CH3Br are still to a large extent unknown (Butler, 2000), although new sources have recently been identified which include emissions from coastal salt-marshes (Rhew et al., 2000), inorganic production in soils (Keppler et al., 2000, 2001), and non- eruptive quiescent volcanic emissions (Wahrenberger et al., 1996a; Jordan et al., 2000; Schwandner et al., 2000b). The natural sources of brominated compounds have been discussed recently by Gribble (1999, 2000) and Butler (2000). Volcanic gaseous emissions of halocarbons have been interpreted by most authors to date as the product of pyrolysis of adjacent vegetation (Gerlach, 1980; Lobert et al., 1999) or air contamination (Gaffney, 1995; Jordan et al., 2000). In contrast, organic geochemical evidence indicates a pristine abiogenic origin by high-temperature gas-phase radical reactions (see previous chapter). The formation of methyl chloride in volcanic emissions of Kilauea (Hawaii) by pyrolysis has been in some

Chapter III. Diffusive emission of organic trace gases from volcanoes 92

doubt on the basis of a comparison of CH3Cl/CO2 ratios found in volcanic gases, in the products of pyrolysis experiments and in biomass burning emissions (Rasmussen et al., 1980).

In this study, we have carried out a field survey to address the question of the origin of halocarbons emitted from the crater and partially vegetated flanks of a quiescently degassing active volcano. As a test site, we chose the La Fossa cone of Vulcano Island (Italy), an andesitic volcanic complex that had its last explosive eruption in the late 1880’s (Mercalli and Silvestri, 1891), and a rejuvenation of unrest a hundred years later (e.g., Tedesco et al., 1995). Vulcano Island is probably one of the best and longest studied volcanoes in the world

(Giggenbach et al., 2001), and has a long record of diffusive CO2 flank degassing measurements (e.g., Chiodini et al., 1998), making it the ideal test site to investigate the origin of halocarbons observed in its fumarolic emissions (Wahrenberger et al., 1996a; Jordan et al., 2000; Schwandner et al., 2000a, 2000b, 2001b, see previous chapter).

Figure 1 Map display of the La Fossa volcanic cone on Vulcano (Aeolian Islands, Italy), after Keller (1970), modified, showing the sampling sites. The stippled field displays the approximate location of the area of fumarolic activity, the solid black areas are lava flows. Thick lines indicate crater outlines, and the open circles show the locations of the fumaroles described previously (Schwandner et al., 2002), of which site 14 (fumarole F0) is included in this study. Point “ref” is a geodetic reference point.

Chapter III. Diffusive emission of organic trace gases from volcanoes 93

2. EXPERIMENTAL

2.1. Sample collection

The site profile for soil degassing measurements extended from the lower flanks of the volcanic edifice up to and into the crater and was chosen by the following criteria: i) the profile intersected areas that are vegetated on the lower flanks of the volcanic cone, ii) it is representative of the entire cone from a geological point of view in terms of known subsurface structures and the soil substrate, iii) it does not intersect with active fumarolic vents, iv) supporting geophysical data are available (Gex, 1992), and v) previous diffusive CO2- measurements of specific parts of the area are available (Chiodini et al., 1996). A map of the sites is shown in Figure 1, indicating the sampling sites and the two fumarolic emission points included in this study. Note that the term ‘soil’ may be misleading because the unvegetated surface of the volcanic cone is by and large made up of unstable, coarse, granular rock material (lapilli and ash) of the 1880 eruption that has never supported any visible vegetation.

Soil gas CO2-emissions were determined using a dynamic chamber method (e.g., Chiodini et al., 1998) with CO2 being measured by a portable IR-Sensor calibrated to CO2 (Multiwarn ® II , Dräger GmbH, Kiel, Germany). The untreated raw data from the CO2-flux measurements are shown in Figure 3. All sampling sites except number 13 yielded data in the linear range of calibration for the CO2-sensor. In order to calculate diffusive soil mass fluxes by the dynamic chamber method, the detector responses have to be corrected for air pressure and temperature, assuming ideal gas behavior. Temperatures were measured using an Inconel sheathed K-type thermocouple that was inserted along the glass tube into the ground which was carefully sealed around the glass tube where it intersected the surface. The air pressure [kPa] was measured using a simple barometric altimeter (Casio 1160 DW-6500-1V) which was daily calibrated against mean local sea level (precision 1%). The site temperatures were also measured in air (10 cm above ground, shaded) and at the soil surface 0.5 cm beneath the surface as well as in the shallow soil at 1 cm intervals to varying depth. These measurements yielded soil temperature gradients which were a function of the local geothermal gradient and a solar surface heating effect for the degassing profile (see Figure 4).

Chapter III. Diffusive emission of organic trace gases from volcanoes 94

The organic trace gases were sampled using an active adsorbent tube technique which was a customized version of US EPA method SOP 2042. A 2-piece stainless steel probe was inserted into the ground to about 0.6 m depth, its solid central rod removed and replaced by a glass tube (1 m length, 4 cm diameter, lower end perforated) and then finally, the outer steel jacket tube pulled out. The top of the glass tube was connected to a custom made conical- convex glass reducing union to 6 mm OD. This connection was further stabilized with a piece of wide-bore silicone rubber hosing that did not come into contact with the sampled gas. A Swagelock® brass union with Supelco M4 ceramic/graphite composite ferrules (Supelco Inc., PA, USA) was used to connect the sampling tube (adsorbent cartridge) to the reducing union. Figure 2 displays a schematic of the sampling setup.

Figure 2 Schematic diagram of the sampling setup for diffusive soil gas sampling (left) and CO2-flux measurements.

A battery-driven, low-flow pump was connected to the sampling cartridge on its downstream end by an identical Swagelock® union, a piece of 6 mm OD glass tubing, and a piece of 5 mm ID silicone hose. The pump’s wetted surfaces were Teflon coated and the pump exhaust was collected in Tedlar® bags to avoid any backflow of air. The pump was a calibrated SKC 222 Series Low-Flow Sample Pump (SKC Inc., PA, USA), set to a flow rate of approximately 80

Chapter III. Diffusive emission of organic trace gases from volcanoes 95

mL/min. This yielded approximately 1.5 L of gas per sample. Prior to sampling, the air contained in the sampling train was carefully purged to allow slow admixing of soil gas but to avoid further air entrainment. The pumped volume was calculated from a) the collected volume in the Tedlar® bag, b) the flow rate and sampling time, and c) the number of strokes and the calibrated stroke volume against a similar resistance.

® Figure 3 Raw data of the CO2-flux measurements obtained from the CO2-infrared sensor (Multiwarn ). All site measurements except site 13 are in the linear range of the sensor.

The adsorbent cartridges (SKC Inc., PA, USA) contained three adsorbent beds consisting of Carbotrap B, Carboxen 569, and Anasorb CSC. Pre-sampling conditioning of the cartridges was done under a stream of purified helium (grade 5) at 300°C with in-line oxygen and hydrocarbon traps. During sampling, they were cooled by ice packs and wrapped in aluminium foil to avoid possible photolysis of analytes by sunlight. After collection, the cartridge ends were immediately capped by the same type of standard endcaps (Kapak®) with which they were supplied and designed specifically for that purpose in air monitoring (SKC Inc., PA, USA). Storage prior to analysis was at –18°C. Sampling of the fumarolic volcanic gases follows a different protocol that is detailed elsewhere (Schwandner et al., 2001a; see previous chapter and appendix).

Chapter III. Diffusive emission of organic trace gases from volcanoes 96

Couch et al. (2000) used an alternative method which involved the sampling of a portion of soil into containers. The advantage of our method is that transport and storage of sealed adsorbent cartridges is much easier and avoids contamination problems compared to soil samples. Furthermore, the decomposition and reaction of compounds during transport and storage of soil material may be a problem given the complex substrate of an actual soil sample, including the microorganisms contained. It also precludes a meaningful soil gas analysis due to significant air entrainment, as opposed to in situ methods such as the one employed in this study.

2.2. Analytical methods

The adsorbent cartriges were thermally desorbed at 300°C under a stream of purified helium, re-adsorbed onto a PDMS/Carboxen (PDMS = polydimethylsiloxane) coated SPME fiber (Solid Phase Microextraction, Supelco Inc., USA) at constant temperature (25°C) and then thermally desorbed in the inlet system (splitless) of the gas chromatograph (GC) and the separated eluents analyzed by full-scan single quadrupole mass spectrometry (MS). To release the adsorbates from the adsorbents, we used a simple custom-built short-path thermal desorption purge-and trap system, optimized to minimize sample loss and fractionation of analytes (Hall Analytical Ltd., UK), with external temperature control, flow control and all- glass contact surfaces (Schwandner et al., 2001a). The same high-purity helium supply was used for desorption as in the conditioning of the adsorbent tubes. A simple custom-built cryocooling device concentrated the analytes at the head of the analytical column in order to ensure a low baseline. The analytical column was a dedicated 100% crosslinked low-polarity PDMS fused-silica Hall-GC1 analytical column (equivalent to the stationary phase OV-1) with the dimensions 60m x 0.32mm x 0.5 µm (Hall Analytical Ltd., UK). Using in-line filtered helium 5.0 as a carrier gas, the column head pressure was 8 psi (flow rate of 0.54 mL/min, average linear velocity of 18.8 cm/s) with a theoretical helium holdup time of 5:32. The phase ration β was determined as 160. The oven was run in a temperature program of 8 min isothermal (40°C), 4°C/min ramp to 290°C, and isothermal again for 17 min at the final temperature.

Chapter III. Diffusive emission of organic trace gases from volcanoes 97

The analyses were performed on two different GC-MS systems: a Thermoquest Voyager GC- MS system coupling a Fisons GC 8060 with a Thermoquest Voyager MD 800 single quadrupole mass spectrometer and an HP GC 5890A coupled with a Micromass Trio-2000 single quadrupole mass spectrometer. The latter was used on one sample from a vegetated site at the base of the volcanic cone to determine target compounds and retention indices at higher resolution than in the routine analyses of all sampling sites as well as for analysis of a field blank (unused cartrige). The acquisition software used was MassLab 1.3 and VG MassLynx 3.0 for NT, respectively. Both MS systems were operated in electron impact mode (EI+) at 70 eV electron current, and a scan rate of 1 scan per second. Mass ranges were 15-600 a.m.u for the MD 800, and 20-549 a.m.u. for the Trio 2000. The transfer line was kept at 300°C, connected in both cases to the GC column end by a capillary open split connection. Sufficient system, SPME needle and column blanks were analyzed in between sample runs to ensure purity of the process, including also a field blank of identical adsorbent tubes analyzed under identical conditions, and ambient and laboratory air blanks.

The raw data were analyzed by indexing all compounds relative to n-alkane and separately to n-aldehyde homologues, and performing a scan-by-scan mass fragment analysis of the compounds. Natural n-aldehyde retention indices are easier to target in samples of soil gases influenced by vegetation than n-alkanes and were used as secondary confirmation. For a number of compounds, molecular ion peaks were preserved and indicated approximate molecular masses of compounds, library spectra were used as suggestions and were considered indicative of a compound’s identity. Retention indices served as targeting help for specific compounds, as isomer identification and in some cases, as confirmation for mass spectral identifications.

Standardization was performed by charging several adsorbent tubes with different known amounts of a certified standard gas (TO-14 series compounds, Scotty Inc., NL). Calibration results and repeat analyses gave a precision of better or equal to 2%; R2 values of the calibration curve from 0.125 ppbv to 12.5 ppbv were better than 0.99 for all compounds except CH3Cl and benzene (not quantified), CFC-114 (0.90), DCM (0.88), toluene (0.97), chlorobenzene (0.82), and m- and o-dichlorobenzene (both 0.96). The greatest source of error (i.e. contamination) appears to be from compounds present on the desorbed, uncharged, adsorbent cartridges or from laboratory handling, as indicated by the field blank (reported as

Chapter III. Diffusive emission of organic trace gases from volcanoes 98

BLK). In all cases, the contamination from these sources was less than 5% except in the case of benzene and some n-alkanes) for which the contribution to the sample analyte peak area readings was up to 10%. The quantitative results of the fumarolic gases are reported in more detail in the previous chapter. Here we report quantitative and relative values in order to be able to compare a larger range of compounds and groups with each other and with measured parameters at the various sampling sites.

Chapter III. Diffusive emission of organic trace gases from volcanoes 99

3. RESULTS

3.1. Site parameters and CO2-flux

To identify the sources and boundary conditions for organic trace gases emanating from an active volcano, an understanding of subsurface structures as well as site conditions is required. The site parameters considered in this study are the precise location and altitude, air pressure, thermal properties of the ground, the diffusive CO2 flux through the soil, vegetation distribution as well as the conductive heat flow as calculated from the temperature measurements. From the available literature, the apparent ground resistivity (Gex, 1992) as well as geological surface and subsurface information was included (Keller, 1970, 1980; Chiodini et al., 1996). The available geological, geochemical and geophysical literature on Vulcano is large and will not be reviewed here. For an introduction, see Fulignati et al. (1998). The results of the physical measurements are displayed in Figure 5.

In the air pressure profile diagram (Figure 5), it can be seen that expected variations with altitude of over 3 kPa occur systematically throughout the profile, but that the actual topography is not accurately reproduced. The air pressure was measured as a correction factor entering in the calculation of the CO2-flux from the sensor measurements (Brombach, 2000), as specified in the previous section.

The temperature measurements yielded several different trends. Figure 4 shows two representative temperature measurements at two sites in the profile. At both sites, the temperature changes linearly with depth but this trend is in both cases clearly separated into two discrete sections with different slopes. The upper section trend is due to solar heating of the black soil surface (the “shallow soil insolation (=solar heating) gradient” in Figure 5). The second, deeper, section at each site is the steady-state geothermal gradient (displayed as its directly measured inverse value dz/dt in Figure 5). In the upper section of site 1 (open squares), the solar heating produces an expected signal with a positive slope whereas site 10 displays a negative slope, indicating anomalous warming to depth or cooling at the surface. In the crater at site 10, the heat released from depth is expected to be greater than at site 1 (at the base of the flank of the volcanic edifice). At site 10 the subsurface heat flow dominates the

Chapter III. Diffusive emission of organic trace gases from volcanoes 100

temperature profile and solar heating is expected to cause only a minor disturbance. The upper section of site 10 is most likely explained by a relative surface cooling of the shallow soil that is continuously heated by a strong subsurface heat flow despite extensive solar heating after a warm sunny day (measured at 6 pm). Surface cooling merely alters the strong thermal signal of the crater region slightly.

Figure 4 Temperature / depth graph of two sampling sites at the base of the cone flank (site 1) and within the crater (site 10). Both measurements show a clear distinction between the linear insolation and the geothermal gradients.

In Figure 5 (column a, bottom), the depth of intersection of the shallow soil solar heating gradient and the geothermal gradient are plotted, together with the air temperature. The two curves show no relation to each other, indicating that the depth of intersection is not controlled by solar heating but by surface advective cooling and the geothermal gradient. The depth of intersection is also a measure of how strongly the subsurface thermal release penetrates through the soil. Finally, the conductive heat flow, H, was calculated from the thermal data (Brombach et al., 2001) using Fourier’s law,

dt H  O dz

Chapter III. Diffusive emission of organic trace gases from volcanoes 101

where λ is the thermal conductivity constant having a value of 1.25 ± 0.25 W m-2 °C-1 in unconsolidated volcanic sediments (see Brombach, 2000), t is the temperature in degrees Celsius and z is depth in meters. The resulting graph in Figure 5 shows that as expected, the heat flow increases with proximity to areas of fumarolic activity.

Figure 5 Parameters measured in the field and derived quantities. The abscissa refers to the distance from the first sampling site (i.e. site 1) at the base of the flank of the volcano. In diagrams (a-c), thermal parameters are displayed, in diagrams (d-g), altitude, pressure-corrected CO2-flux, and atmospheric pressure are shown together with basic botanical, structural and surface geological data. For comparison, literature values of the apparent resistivity along the same profile are also shown (Gex, 1992).

Chapter III. Diffusive emission of organic trace gases from volcanoes 102

The measured CO2 soil gas flux (Figure 5) shows a general increase towards the crater, and is largest at sites where faults are inferred from surface geology and surface degassing structures. At the crater bottom, it approaches zero because of decreased permeability due to the underlying magmatic plug and the consolidated character of the accumulated fine epiclastic alluvial volcanic debris that fills the crater bottom.

Similar CO2-measurements in 1995 resulted in qualitatively similar flux profiles (Chiodini et al., 1996), indicating that the structural and thermal regime has been relatively unchanged over the last 6 years. The variations of the fluxes along the profile are similar to variations in ground resistivity data obtained along the same profile (Figure 5) 10 years earlier (Gex,

1992). The CO2 and resistivity profiles suggest active faults and/or active degassing at shallow depth. An increased gas pressure dilates pores and the gas is in itself a poor conductor, leading to an increased resistivity over subsurface degassing features, such as actively degassing subvolcanic magmatic intrusions and/or deep reaching faults.

3.2. Chemical parameters

A large range of compounds was targeted and identified and only a representative selection is given here. Figures 6 to 11 display the chemical signals along the measured profile ordered by compound group (Figures 6 to 8), as the C7-C20 n-alkane distribution (Figure 8), and as correlations between selected compounds and physical parameters (Figure 11 and 11). The sites along the profile can be subdivided into four major zones: a) the vegetated lower flank of the volcano, b) the barren upper flanks of the volcano facing Vulcano town, c) the barren crater area with obvious diffusive degassing structures, and d) the fumaroles and fumarolic fields with ground temperatures > 50°C and high diffusive CO2 fluxes. Site 1 is excluded in the diagrams of Figures 6 to 8 since it was analyzed by higher resolution GC-MS system in order to derive enrichment factors of sample vs. blank readings (Table 1). Table 1 shows that all analytes are significantly enriched, indicating that contamination is negligible even under lengthy, hot and corrosive field sampling conditions where the field blank (i.e. an unused adsorbent cartridge) was carried along at all times.

Chapter III. Diffusive emission of organic trace gases from volcanoes 103

Table 1 Comparison of two analyses, that of site 1, located on the vegetated lower flank of the volcano, with a field blank. For all compounds shown, the enrichment factor of the sample reading over the blank reading is shown. This value is obtained by dividing the absolute detector response of the compound in the sample by the response in the blank. Note that the field blank analysis incorporates all sources of contamination except the actual sampling process (i.e., preparation, transport, storage, analysis). Halocarbons are highlighted in bold face and sums of several compounds or isomers are indicated by italics.

Chapter III. Diffusive emission of organic trace gases from volcanoes 104

The least enrichment is shown for thiophene (f = 1.74), reflecting the extremely low abundance in the vegetated zone of the profile (cf. Figure 9), where it was only detected in the site 1 sample because of the use of a higher resolution MS in analysis. Consequently, enrichment factors are up to an order of magnitude higher in samples that have a higher concentration, i.e. in the crater area. The highest enrichment factors are found for halocarbons, nitriles and arenes. In a number of case, compounds were not detected in the blank at all. This shows that contamination can be safely ruled out as a source for these compounds. Air contamination during sample collection was previously shown to be minimal using the techniques employed as indicated by noble gas isotopic measurements and inorganic atmospheric gases (see previous chapter).

Before discussing the individual compound groups, the Figures are introduced as follows. In Figures 6 to 8, a range of representative compounds are displayed in abundance vs. “profile distance” graphs, ordered by compound groups. The profile distance is measured relative to the starting point at the base of the profile on the vegetated lower flank of the volcano (site 1). In these diagrams, enrichments can be directly read for individual compounds at certain sites or zones along the profile. For clarity of comparison, the detector responses are displayed relative to the highest response found in the profile in per cent, after correcting for sampling volume.

In Figure 8, n-alkane carbon number distribution patterns are shown for a representative selection of sites that are characterized by varying influences of volatiles from fuel exhaust, vegetation and volcanic gas. These diagrams help in distinguishing the sources and processes giving rise to the organic chemistry found at the various sites. In Figure 11 and 11, a selection of well correlating compounds and CO2 flux are displayed. We calculated and ranked a correlation matrix of over 100 compounds and physical parameters, in order to gain insight into processes by which physical and chemical signals seem to be related. Only a representative selection can be presented here.

Chapter III. Diffusive emission of organic trace gases from volcanoes 105

Figure 6 Correlation diagrams of compounds and CO2-flux (cf. Figure 11). In all diagrams, the abscissa refers to absolute detector responses (sampling volume corrected) of compounds present at various locations through the profile, excluding the fumarole F0 (site 14). In diagrams (a-d), the ordinate displays the detector responses of other compounds in the same samples. In diagrams (e) and (f), the ordinate is the measured site CO2 flow, F (cf. Figure 11). Correlation parameters are shown for each diagram, as the correlation function derived from linear least-squares regression of the data, and the R2 correlation coefficient. Note that in diagram (f), two data sets are shown: the sum of all C2-substituted thiophenes, and the first eluting dimethyl thiophene isomer. The summed data include the data points represented by the dimethyl thiophene isomer.

3.2.1. Hydrocarbons

The first row of Figure 7 shows the results for n-alkanes and styrene and the remaining diagrams display the variations for aromatic hydrocarbons (arenes) and oxygenated compounds along the profile. Normal alkanes (n-alkanes) are straight-chain aliphatic saturated hydrocarbons of the general formula CnH2n+2. The n-hexane concentration is markedly increasing towards the crater. The concentrations of n-nonane are more variable from one site

Chapter III. Diffusive emission of organic trace gases from volcanoes 106

to the other but generally increase towards the crater as well, with pronounced maxima over sites of active subsurface degassing (crater ring fault at site 9). Hexane is a prominent hydrocarbon found in a variety of low-temperature volcanic and hydrothermal degassing sites around the world (Capaccioni et al., 1993; Kiyoshu and Asada, 1995), whereas n-nonane dominates other hydrocarbons in air near vegetation (cf. Figure 8a, ambient air sample). Despite the relative dominance of n-nonane over other n-alkanes at sites influenced by vegetation, the absolute concentrations nevertheless notably increase towards the crater, indicating that the relatively weak signal from the vegetation is quickly overwhelmed by subsurface degassing. Interestingly, n-butane shows an apparent variation with atmospheric pressure of the sampling site in all samples which is considered to be an artifact.

The even/odd predominance of biological organic matter is a common phenomenon used to assess maturation of fossil fuels, however, this signature is destroyed/altered by hydrous pyrolysis (Putschew et al., 1998). In addition, no even/odd predominance is observed for abiogenically synthesized compounds (e.g. via Fischer-Tropsch reaction). In the uppermost right diagram in Figure 7, the abundance of odd numbered n-alkanes decreases towards the crater, however, it is somewhat lower on the vegetated zone itself as well. The highest abundances are found on the upper flank of the volcano but they decrease towards the crater. In Figure 8, a strong even/odd predominance can be seen to be present in the samples from the vegetated zone, even though the total abundance of odd alkanes is not elevated (Figure 7). The strong signal of odd hydrocarbons must therefore have a second source in addition to plant volatiles. In Figure 8, samples from the upper slopes as well as within the crater resemble a mixture of fuel combustion exhaust volatiles (represented by the Vulcano harbor air sample) and plant derived volatiles with a strong C12 peak similar to sites 3 and 4 (Figure 8).

Further information on the n-alkanes can be extracted from Figure 8, displaying the C7-C20 n- alkane distribution of selected individual sampling sites, a high-temperature fumarole at site 14 (394.8°C fumarole F0), as well as the field blank and a remote Vulcano air sample (taken at the shore near vegetation at Gelso, on the south coast of the island of Vulcano). A polluted air blank taken in Vulcano harbour during ferry loading an truck traffic is also shown. For comparison, a synthetic Fischer-Tropsch product is shown as well (Davis et al., 1999). The samples from the vegetated zone display a systematic abundance alteration of odd and even

Chapter III. Diffusive emission of organic trace gases from volcanoes 107

hydrocarbons (Figure 8, column a), indicating a dominance by biogenic organic matter. The hydrocarbon curve has its maximum at about C12, shifting towards heavier molecular weight hydrocarbons at site 6 (not shown). The remote air sample (Gelso air) taken near vegetation also has two maxima at C12 and and C9. Sites 7 through 12 are characterized by a rising/falling hydrocarbon distribution, centered at C14, with a prominent additional peak at

C12.

Figure 7 Hydrocarbons and oxygenated compounds present in diffusive emanations (soil gases) on the different sections/zones along the profile. The ordinate in each diagram displays the measured detector response of the GC-MS system, in per cent of the largest signal obtained. The abscissa refers to the distance from site 1, where the profile was started at the base of the flank of the volcano. The extent of the different zones crossed by the profile are indicated on the top of the Figure: the vegetated lower flank of the volcano, the barren upper flanks (facing away from the crater and hot gas emissions), the barren crater, and the areas of active fumarolic degassing. Site 13 is grouped with the fumaroles because it is situated within the outer area of a fumarolic field, with ground temperatures approaching 100°C and high CO2-fluxes (cf. Figure 3). Site 14 is a fumarole (F0) and is slightly off the profile (cf. Figure 1), details are given in the text and in (Schwandner et al., 2002). Responses were corrected for sampled volumes for each site. The dashed lines connect sites between which datapoints are not given. These “missing” data points (indicated by “?”) are detected compounds that could not be quantified with a reasonable degree of confidence due to overlaps with coeluting peaks with similar mass fragments.

Chapter III. Diffusive emission of organic trace gases from volcanoes 108

The harbour air sample exhibits a pronounced maximum at C14 as well, suggesting an anthropogenic source of a significant portion of the hydrocarbons in soil gases on the upper flanks of the La Fossa cone (sites 7-12). The harbour air with its relatively heavy n-alkane distribution is dominated by the products of incomplete combustion of diesel fuel from the commercial truck traffic and the ferry engines.

Figure 8 Normal (straight-chain) alkane molecular carbon number distributions for selected representative samples from the profile (i.e. three different sources). The abscissa refers to the number of carbon atoms in the normal alkanes (n-alkane) found in each sample (only 1 compound per n-alkane carbon number is possible). The ordinates represent detector responses (peak areas) relative to the strongest peak in the carbon number range C7- C20, in per cent. The uppermost graph in column (c) displays peak heights obtained from published data.

The fumarolic gas sample (site 14, fumarole F0) is very reducing in nature with excess, detectable, free hydrogen gas. Its hydrocarbon distribution has its modus at C8, with abundances linearly falling with increasing molecular weight. This type of pattern is known as the Schulz-Flory distribution (Satterfield and Huff, 1982) and is regarded as a strong indicator of the abiogenic catalytic Fischer-Tropsch synthesis (Salvi and Williams-Jones, 1997). For comparison, a product of this process is shown in Figure 8 (column c, uppermost diagram)

Chapter III. Diffusive emission of organic trace gases from volcanoes 109

displaying the Schulz-Flory distribution and a number of synthetic polymers of lower molecular weight. In the example shown, only CO and H2 were used to synthesize these compounds over an iron-based catalyst (Davis et al., 1999). We consider that this fumarolic gas sample (site 14) represents the result of abiogenic formation of hydrocarbons in the gas phase.

3.2.2. Aromatic compounds (arenes)

A number of aromatic compounds, including polycyclic aromatic hydrocarbons (PAH), have been reported from volcanic gases (see previous chapter) but they also may typically derive from anthropogenic/biogenic sources. Benzene, toluene, the xylenes (BTX), styrene, substituted arenes, naphthalene and higher PAH’s and their hydrated derivatives are found in our samples. All of these compounds increase in abundance towards the crater (Figure 7, second row), but are relatively low in volcanic gases. Blank readings are very low and diffusive flank degassing appears to be a significant source of these compounds. In contrast to this pattern, styrene is consistently at low concentration throughout the profile with a prounounced single peak at site 13 (a site of diffusive but >80°C degassing).

3.2.3. Oxygenated and nitrogenated compounds

Among the oxygenated compounds (rows 2 and 3 in Figure 7), ketones and alcohols show a strong enrichment towards the crater area. Acetophenone (C8H8O) is, in addition, enriched on the upper flanks of the cone. All trends of oxygenated compounds except ethyl methacrylate show a clear enrichment towards the crater. Acetone correlates reasonably well with the measured CO2-flux, indicating a volcanic origin (Figure 6). Ethyl methacrylate (C6H10O2) does not show any similarity to any other physical or chemical signal and is generally of uniformly low abundance (Figure 7). It is a common contaminant derived from adhesives and plastics, such as in the SPME fiber and the analytical column. It most probably is a procedural contaminant. Acrolein (2-propenal) correlates well with ethanol (Figure 6), which is strongly biased towards the crater area (Figure 7). Both compounds may thus have a strong volcanic contribution as well. Acetaldehyde (ethanal) shows a recognizable correlation with dimethyl sulfone (Figure 6), some of the scatter possibly being attributable to oxidation of dimethyl monosulfide to yield excess dimethyl sulfone (see below).

Chapter III. Diffusive emission of organic trace gases from volcanoes 110

The second row of Figure 9 demonstrates that the nitrogen containing heterocycles pyridine

(C5H5N), pyrazine (C4H5N2) and pyrrole (C4H5N) are enriched on the upper flanks of the volcano but decrease from the crater rim towards the fumaroles which contain lower concentrations, with a pronounced increase at the crater bottom. Pyridine and pyrazine exhibit this pattern more strongly than pyrrole. Acetonitrile (C2H3N) is plotted as a representative of several nitriles found in the samples and is enriched in the volcanic gases. Acetonitrile and methyl bromide appear to correlate reasonably well (Figure 6), however, too few samples contained enough methyl bromide to produce a statistically relevant regression.

Figure 9 Heteroatomic (S, N) organic compounds present in diffusive emanations (soil gases) on the different sections/zones along the profile. Abscissa and ordinate as in Fig. 6; note that the upper two diagrams have a logarithmic scale on the ordinate. Further details are given in the text and in Figure 7.

3.2.4. Sulfur compounds

Volcanic gases contain sulfur as a major element in various forms and native sulfur and sulfide phases sublime around fumaroles. Sulfur in volcanic gases occurs predominantly as

SO2 and H2S, however, we have also identified organosulfur compounds. Row 1 of Figure 9 shows a notable increase of all organosulfur compounds towards the crater with a strong

Chapter III. Diffusive emission of organic trace gases from volcanoes 111

enrichment in the volcanic gases. The organosulfides additionally show an enrichment over inferred or known faults and in areas of fumarolic activity.

The oxygenated compound dimethyl sulfone (C2H6O2S), was also identified on the upper flanks of the cone but is diminished in the vegetated area at the base of the profile. Dimethyl monosulfide (C2H6S) was found in several fumarolic gas samples (see previous chapter) but not detected along the profile. Trace dimethyl monosulfide was also detected in the site 1 gas sample which was analyzed at higher resolution. The dimethyl monosulfide concentrations are probably slightly higher in the volcanic gas inside the volcanic edifice since dimethyl monosulfide is readily oxidized to dimethyl sulfone.

All sulfides (as well as some arenes and halocarbons) show a small peak at site 4 at the onset of visible vegetation on the slope (note: logarithmic scale). CO2 fluxes and resistivities are elevated at this site as well (Figure 7), indicating a local subsurface degassing feature such as an active fault buried by slope debris. The heterocyclic thiophenes were not detected in samples from the vegetated lower flanks of the cone but are relatively prominent in the volcanic gases and on the upper volcanic flanks. Thiophene (C4H4S) and its C2- and C3- substituted derivatives were not only found enriched in the crater area and the volcanic gases, but they also correlate well with the CO2-flux, indicating a volcanic origin (Figure 8).

3.2.5. Halocarbons

The halocarbons (Figure 10) uniformly increase in abundance from the flank towards the crater, with in some cases dramatic increases in the volcanic gas samples (note the logarithmic scale of some of the diagrams in Figure 10). All three methyl halides, CH3Br, CH3I and

CH3Cl, show strong enrichment towards the fumaroles and CH3I is already notably enriched in crater soil gases. CH3I and CH3Cl abundances correlate well, supporting a volcanic origin consistent with the apparent strong enrichment in the fumarolic area. Bromoethane, C2H5Br, is most prominent at site 13, in an area where strong diffusive gas emanation (hissing sound) was observed with temperatures reaching 100°C in the range from 80 to 100°C and near the surface sampling site. This suggests a possible lower temperature of formation for bromoethane in this environment than for methyl bromide which is more pronounced at the much higher temperatures of the fumarolic volcanic gases.

Chapter III. Diffusive emission of organic trace gases from volcanoes 112

Figure 10 Halocarbon compounds present in diffusive emanations (soil gases) on the different sections/zones along the profile. Abscissa and ordinate as in Fig. 6, note that the upper two and the lower left diagram have a logarithmic scale on the ordinate. Further details are given in the text and in Figure 7. “Missing” data points indicate that the compound was not detected.

Increased methylene chloride, CH2Cl2, concentrations are associated/correlated with high CO2 fluxes and elevated ground resistivities (cf. Figures 5 and 8b), indicating a volcanic origin. Methylene chloride reaches diffusive gas concentrations as high as 800 ppbv but is not any more enriched in the volcanic gas of site 14. CFC-11 (CFCl3) also correlates well with CO2 degassing (Figure 8), but in contrast to methylene chloride is also present at high concentrations in the volcanic gases (Figure 7). At the bottom of Figure 10, data for chlorinated benzenes (mono- and dichlorobenzenes) are shown. The sum of all three isomers of dichlorobenzene, C6H4Cl2, as well as chlorobenzene, C6H5Cl, show a clear increase towards the crater but a smaller abundance in the volcanic gases. The chlorobenzenes appear to have a much stronger source in diffusive rather than fumarolic degassing.

Tetrachloromethane, CCl4, (Figure 10 lower left) was not detected at the vegetated base of the profile but in the crater area and in fumarolic discharges. Pronounced peaks are evident at site 11 (inner crater rim) and in the fumarolic gas (site 14, fumarole F0), indicating a strong

Chapter III. Diffusive emission of organic trace gases from volcanoes 113

subsurface source. The other crater sites (i.e. sites 9, 10, 12, 13) are all at similar concentration, indicating a deep rather than shallow formation of this compound. If shallow formation would give rise to the observed CCl4 concentrations in the crater area, site specific parameters such as degree of substrate cementation and temperature would lead to stronger concentration differences. The same applies to site 7, where despite the very impermeable argillaceous “Fossa Rossa” substrate, small scale sublimation phenomena can be observed along cracks.

Figure 11 Quantitative correlation diagram of volume-corrected CFC-11 concentrations of all sites along the profile except the fumarolic gas from site 14 (fumarole F0). The ordinate shows the measured CO2 flux at each site, the abscissa displays CFC-11 concentrations in adsorbent cartridge samples taken at the same location and time. The hatched line indicates the average tropospheric CFC-11 concentration (cf. Figure 12). The correlation function and R2 correlation coefficient were obtained from linear least-squares regression of the data. Symbol diameters are larger than analytical uncertainties (ca. 2.4% for CFC-11, <2% for CO2 flux) and the symbols are sized accordingly.

Figure 12 shows quantitative data of concentration variations along the profile, for methylene chloride and CFC-11. This diagram is based on the same data as the data points in Figure 10.

Methylene chloride, CH2Cl2, shows a strong enrichment towards the crater, but no pronounced enrichment in fumarolic gas, whereas CFC-11 is by an order of magnitude enriched over background air and even average crater soil gas concentrations in the fumarolic discharges. The dashed line represents the average tropospheric CFC-11 concentration, however, the soil gas samples and ambient air blanks not included in this diagram indicate that the actual local concentration may be only around 200 pptv.

In both cases (average tropospheric air or local air with 200 pptv), the volcanic gas as well as a number of samples of soil gases from the crater area are significantly enriched in CFC-11

Chapter III. Diffusive emission of organic trace gases from volcanoes 114

and cannot be derived from air contamination. The air component for the fumarolic gas considered in this study was found to be less than 0.001% based on He and Ne isotopic analyses and He, Ne, Ar and N2 atomic ratios of headspace analyses of volcanic gases sampled by conventional methods (“Giggenbach bottle”) from the same sampling train (see previous chapter).

Figure 12 Quantitative variations of halocarbon concentrations (methylene chloride and CFC-11) present in diffusive emanations (soil gases) on the different sections/zones along the profile. Abscissa and ordinate as in Fig. 6, note that the upper two and the lower left diagram have a logarithmic scale on the ordinate, and that two different ordinates refer to concentrations in ppbv (for methylene chloride) and pptv (for CFC-11). The dashed line represents the average tropospheric CFC-11 concentration, however, the soil gas samples and ambient air blanks not included in this diagram indicate that the actual local concentration may be only around 200 pptv. Further details are given in the text and in Figure 7.

Figure 11 shows an excellent positive correlation of CFC-11 concentration to CO2 fluxes measured at the sampling site. This observation as well as the above mentioned strong enrichments in the crater area, especially in the fumarolic gases, documents a deep, abiogenic and non-atmospheric origin of this compound. In addition, this correlation allows quantitation of fluxes for individual volcanoes as well as on a global scale (see previous chapter).

3.2.6. Botanical observations

In order to asses the influence of the vegetation on the volatile organic chemistry of the samples, we carried out a plant collection in autumn 2001. We found the visible vegetation not only to be limited to the first four sampling sites but is also strongly zoned despite the apparently similar substrate conditions (soil material, slope, etc.). At the base of the profile, a great variety of grasses and few thistles dominate which are quickly replaced by Leguminosae, dominated by woody Genista sp. brooms (including Spartium junceum) as one

Chapter III. Diffusive emission of organic trace gases from volcanoes 115

moves upslope. Since Leguminosae are not nitrogen-limited and appear to dwell under elevated CO2-atmospheres (Lüscher and Nösberger, 1997; Gloser et al., 2000); hence, actual local advantage can be postulated. Thus, vegetation zonation on volcanic slopes may indicate subsurface degassing structures. In the organic trace chemistry of the soil gases, several terpenes and terpenoids were identified, some of which are characteristic for Leguminosae. Among the terpenes found in sites 1-4 soil gases are alpha-copaene and ylangene (isomers, tricyclic sesquiterpenes), camphene, limonene, pinene, alpha-terpinene, isolongifolene, alpha- gurjunene, alpha-phellandrene, beta-myrcene, tricyclene, and several others. However, some were also found in minor quantities at the upper slopes of the profile, perhaps having been adsorbed onto the volcanic ash particles. Some typical “plant volatile” compounds, such as cumene (isopropylbenzene), found on the upper slopes may also be of either pristine volcanic or anthropogenic origin.

Chapter III. Diffusive emission of organic trace gases from volcanoes 116

4. SUMMARY AND DISCUSSION

The physical and chemical data presented in this study show that volcanic gases and on-site vegetation are not the only contributors to organic trace gases found in soil gas. The investigated soil gas profile can be subdivided into four major zones: the vegetated base of the profile with a granular loose lapilli and ash substrate, the unvegetated upper flanks of the volcanic cone with different substrates of volcanic material, the vegetation-free crater area with cemented pyroclastic and loose epiclastic substrates and strong diffusive degassing features, and fumarolic areas including discrete volcanic gas emanations from fumaroles.

4.1. Vegetation effects and adsorption on soil particles

Vegetation-derived volatile organic trace gases dominate the soil gas chemistry of the vegetated base of the volcanic cone of La Fossa. Among them are terpenes and terpenoids, nitrogenated compounds, and C9-dominated hydrocarbons, with an apparent even/odd predominance, as well as dimethyl monosulfide and carbon disulfide. The organic content of soil gases from the upper flanks of the volcanic ash slopes are dominated by what appears to be adsorbed organic matter derived from a variety of sources, including anthropogenic atmospheric background contaminants, airborne proximal plant organic volatiles from the cone flank base, anthropogenic emissions from fuel combustion, and a component attributable to volcanic emanations.

The vegetation derived compounds comprises n-alkanes, aldehydes, ketones (except acetone), nitrogenated heterocycles, terpenes, and dimethyl sulfone as an oxidation product of dimethyl monosulfide. The anthropogenic component is best characterized by the n-alkane distribution in the harbour air sample (Figure 8, “air 2”). This sample represents a mixture of vegetation- derived volatiles (i.e., strong C9 ) and an alkane normal distribution with a maximum at C14 (Figure 8). Traces of its n-alkane distribution are partial recognizable at sites 7 and 11 as well

(Figure 8), mixed with a small vegetation-derived component indicated by a C12 peak. The volcanic gas component comprises arenes, acetone, halocarbons, ethanol, and thiophenes. No overprint of volcanogenic n-alkane distributions (site 14) can be identified on the upper slopes.

Chapter III. Diffusive emission of organic trace gases from volcanoes 117

The airborne “atmospheric” component responsible for compounds found on the upper slopes and to much lesser extent in the crater, manifests itself in compounds that are probably adsorbed onto soil particles. The sources of this airborne component are all of the above mentioned individual sources, including volcanic gas. If one theoretically considers the volcano would be temporarily “switched off”, concentrations of some compounds would probably after equilibration of the local atmosphere be decreasing towards the crater. The background signal of halocarbons derived from the atmosphere (and thus by large anthropogenic sources) would encounter a soil substrate that has a drastically reduced surface availability for adsorption due to its previously cemented character. In our study, we see no evidence for atmospheric halocarbons at significant concentrations relative to the other local strong sources. In contrast, we observe that the halocarbons derived from subsurface degassing dominate the organic chemistry of the volcanic soil gases.

For example, the tetrachloromethane (CCl4) concentrations would be expected to be of the same concentration throughout the entire profile if one attempted to explain the observed

CCl4 concentrations by a non-volcanic, remote atmospheric, source. This is obviously not the case since strong peaks are observed at sites of known structures of subsurface degassing (in this study confirmed by CO2 flux and resistivity data) and the base of the profile is free of detectable CCl4. In addition to the uniform concentration profile expected of the remote atmospheric interpretation, CCl4 concentrations should actually be elevated at sites dominated by vegetation (sites 1-4), because carbon tetrachloride retention in soils is strongly controlled by its organic carbon content (Duffy et al., 1997) and the substrate of sites 5-14 is extremely poor in organic carbon. However no CCl4 was detected in the samples from the vegetated zone.

Any supposed adsorbed halocarbons on soil particles are clearly overwhelmed by a strong volcanic signal, even though the dry climate at the Aeolian Islands actually facilitates enrichment of halocarbons in the shallow soil. Volatile organic compound adsorption was experimentally found to increase the total volatile organic carbon sorption capacity to soil minerals at low water content by several orders of magnitude for some compounds including trichloroethylene (Poulsen et al., 2000). The dry climate of the Aeolian Islands during the summer requires a separate control on adsorption in organic carbon-poor volcanic ash. One

Chapter III. Diffusive emission of organic trace gases from volcanoes 118

such facilitating process could be intraparticle mass transfer which affects the adsorption process at sites with mostly coarse-grained substrate character, such as the upper flank regions (Chen and Wu, 1995). The only remaining possible explanation other than subsurface degassing would be a local anthropogenic source of CCl4 but this hypothesis is not verified by the harbor air sample.

Nitrogenated compounds show enrichment patterns favoring the upper flanks (heterocyclic compounds) or the crater and fumaroles (nitriles). In this local case, the local volcanic source is by far dominant. The presence of acetonitrile in soils facilitates pyridine adsorption on soil particles (Borisover et al., 2001), but this coupled effect does not explain the observed enrichment patterns since acetonitrile increases towards the crater and pyridine in contrast decreases towards the crater. A local strong volcanic source of acetonitrile would therefore be the only explanation for the observed enrichment of this nitrile.

4.2. The crater region

The crater region is characterized by a high diffusive gas flow. Individual sites of elevated gas discharge (site 9 and the fumarole, F0, at site 14) produce volcanic peaks that are strongest for

CFC-11, n-hexane, CCl4, thiophene, and the chlorobenzenes. Hexane increasingly dominates the n-alkanes discharge with proximity of fumaroles. Strong bimodal patterns in n-alkane distribution are not noted and the vegetation indicators n-nonane (C9) which dominates the soil gas alkane chemistry at the vegetated sites at the base of the profile looses this dominance relative to other alkanes in the crater area. Its own absolute concentration however increases towards the crater as well, indicating a strong volcanic source. The n-alkane distribution is dominated by C12-C14.

All halocarbons show elevated values in the crater. The increase in CCl4 (see above) and methyl chloride is less pronounced but quite strong for methyl iodide and –bromide, methylene chloride, CFC-11 and the chlorobenzenes. Strong general increases in the concentration of compounds are observed for halocarbons, acetone, acetonitrile, acetaldehyde and ethanol among others. All the halocarbons except the haloarenes and bromoethane show their strongest enrichment in the volcanic gas dominated emanations (fumarole fields and fumaroles). The same applies to thiophene and its substituted derivatives, dimethyl disulfide

Chapter III. Diffusive emission of organic trace gases from volcanoes 119

and acetonitrile. In areas of active fumarolic degassing, styrene, C2-substituted arenes, acetophenone, benzaldehyde, methyl bromide and bromoethane are strongly enriched where intense, diffusive warm (~80-100°C) degassing occurs (i.e., site 13). The very reducing volcanic gas sampled at fumarole F0 (site 14) bears no resemblance to any of the non- volcanic sources, and its halocarbon concentrations are the highest found in all samples. The n-alkane distribution is centered at a lesser molecular weight with linearly decreasing abundances with increasing carbon number (Schulz-Flory distribution), suggesting inorganic catalytic Fischer-Tropsch synthesis.

4.3. Pyrolysis

In the investigation of diffusive organic trace gas emissions from volcanic flanks and craters, the question of a possible pyrolytic origin of the organic matter arises. For natural thermal sources of halogenated, volatile organic compounds, biomass burning is responsible for a significant fraction of global emissions of reactive chlorine (Keene et al., 1999). Since about 90% of vegetation fires are currently induced by human activity, it is actually considered an anthropogenic source (Keene et al., 1999). In biomass burning, the two cases of flaming and smoldering stages of vegetation fires differ greatly in their emissions of pollutants and in the associated processes acting during combustion (Lobert and Warnatz, 1993).

In subaerial volcanic emissions, a similar distinction can be made between different types of degassing activity: violent episodic explosive eruptions, lasting hours to days, and being capable of penetrating the tropopause, and continuous quiescent degassing periods of fumarolic activity, lasting 100’s to 10’000’s of years between major eruptions (see previous chapter). In the cases of flaming and smoldering vegetation fires, the processes and emissions differ as do the cases of explosive and fumarolic volcanic activity. However, true volcanic flames are a rare phenomenon and are either due to burning of vegetation buried by lava flows (Bunsen, 1847, 1852), self-ignition of sulfur in high-temperature fumarolic fields (Harris et al., 2000), or hydrogen gas burning (Cruikshank et al., 1973; Naughton, 1973). Volcanic hot degassing activity should not be confused with combustion processes. There is no indication for a pyrolytic origin of any of the samples from the sites investigated in this study

Chapter III. Diffusive emission of organic trace gases from volcanoes 120

4.4. Abiogenic formation

Some of the organic compounds having the degree of complexity of those found in volcanic gases may not be stable (i.e. in equilibrium) at the discharge temperatures of the gases themselves but form during cooling of the volcanic gas to exit temperatures. Other compounds may later form during dilution by air after exit to the atmosphere (Zolotov and Shock, 2000; see previous chapter). However, the thermal stability of many of the relevant organic compounds is poorly known at temperatures >400°C. In our case, the gas is essentially quenched during sampling so that the compounds trapped on the adsorbent cartridge probably represent an equilibrium mixture of higher temperature conditions (Symonds et al., 1994). The compounds most likely form from simple organic precursors by gas-phase radical reactions which may be mediated by available catalytic surfaces. Organic precursor molecules may form via catalytic Fischer-Tropsch reaction, as indicated by a number of isomers found (see previous chapter) as well as by the apparent Schulz-Flory distribution in reducing volcanic gas samples as demonstrated by fumarole F0 (site 14). Subsequent gas-phase or catalytic surface aided polymerization may form higher molecular weight compounds condensing on surfaces or being emitted to the atmosphere. Gas phase electrophillic substitution and alkyl free radical reactions may form and modify halocarbons, heterocycles and arenes in the hot volcanic gas.

Many of the arguments presented in this study are supported by strong positive correlations of concentration with CO2-flux. The subsurface degassing operates quasi-continuously on a century to millennia time scale giving rise to very high daily CO2 discharges at Vulcano (180- 230 t d-1, Chiodini et al., 1996). From the data presented in this study, we conclude that in terms of a volcanic source of halocarbons in the atmosphere, diffusive volcanic emission probably plays a role about equal to that of discrete fumarolic emissions (cf. previous chapter).

To estimate the volcanic source strength of halocarbons, the organic composition may be scaled to SO2-fluxes determined independently by COSPEC-measurements. However, there are shortcomings associated with SO2-fluxes determined by that method (sulphate aerosol and sulphate-covered ash particles). Sulphur speciation can be quite variable in a volcanic gas,

Chapter III. Diffusive emission of organic trace gases from volcanoes 121

making it difficult to correlate SO2 fluxes and SO2 abundances in the pure fumarolic gas, in that H2S may in some cases dominate over SO2. In other cases, the reverse can be observed.

CO2 concentrations in a fumarolic gas are not as variable and even though volcanic C-O-H fluids speciate into CO and CH4 as well, depending on temperature and redox potential, CO2 is in most cases the dominating species. In addition, CO2 fluxes determined by COSPEC have the advantage that confirmatory data are available for an increasing number of volcanoes by grid mapping of diffusive soil degassing (e.g., Koepenick, 1995; Chiodini et al., 1996; Chiodini et al., 1998; Giammanco et al., 1998; Wardell et al., 1999; Brombach et al., 2001; Chiodini et al., 2001).

We previously estimated a global volcanic CFC-11 source strength from quasi-continuous, discrete fumarolic degassing to have a global annual mass flux from volcanoes to the atmosphere of 0.3-1.25 x 10-5 Tg y-1 (data from previous chapter). If diffusive degassing contributes the approximately same amount as appears to be the case at Vulcano, the above figure will double. This estimate does not take the phenomenon described in this study into account. Some brominated halocarbons are apparently emitted from hot but strongly diffusive emissions. In the case of methyl iodide, the diffusive emissions might actually superceed the discrete fumarolic discharges, provided the case of Vulcano can be extrapolated to other volcanoes. In order to quantify the volcanic source strength of halocarbons to the atmosphere, diffusive and discrete degassing features will have to be investigated on a more representative number of volcanoes to arrive at a meaningful global estimate.

Chapter III. Diffusive emission of organic trace gases from volcanoes 122 IV. Conclusions 124

This study was concerned with the sampling, analysis, data treatment, analytical quality and interpretation of natural halocarbon and other organic compound discharges from a quiescent active volcano (Vulcano, Aeolian Islands, Italy), as well as the extrapolation and modeling of the results to arrive at a reasonable estimate of global volcanogenic halocarbon emissions (chapter II), and the testing of the interpretation as a natural source by means of a source evaluation (chapter III).

The aim of the study can be divided into several key approaches: N to develop new sampling and analytical techniques for the task of determining trace organic compounds in a complex matrix of major compounds that is adverse to successful collection, conservation, separation and detection of analytes of interest; N to quantitatively determine concentrations, fluxes and contamination sources of selected halocarbons that are known or suspect to be causing stratospheric ozone depletion; N to evaluate these quantitative results in terms of their volcanic source strength as a constant pre-industrial natural contributor to the atmospheric halogen chemistry; N to test the hypothesis of a natural contribution of volcanic halocarbon emissions by identifying possible abiogenic natural processes and other non-volcanic sources interfering with a postulated volcanic source.

In order to meet these goals, we developed, tested and applied novel sampling and analytical techniques to efficiently separate the major matrix compounds, water, carbon dioxide, inorganic sulfur, metals and halogen acids and at the same time preserve the analytes of interest from losses, degradation, dilution or contamination overprint. We used a custom-built gas focusing and steam separation apparatus for sampling and tested a variety of solid adsorbent materials, including Florisil®, Tenax TA®, Carboxen®569, and Carbotrap, for use of adsorption of trace organic compounds during sampling. For analysis, we used short-path thermal desorption-solid phase microextraction-cryotrapping-gas chromatography-mass spectrometry to analyze the volcanic gas samples fully quantitatively. In addition to these techniques, we carefully evaluated the possible role of air or laboratory contamination by use of separate sampling and analysis of non-condensible gases and noble gas isotopes and a complex blank system.

Chapter IV: Conclusions 125

In addition, we performed a field survey along a profile spanning from the vegetated base of the volcano, over the flanks and crater, into the fumarole field. We measured a variety of parameters in the field, including thermal gradients in the ground and CO2-fluxes, and took soil gas samples for analysis of trace organic compounds such as in the first part of the study.

The results comprise N a new versatile method of sampling and quantitative analysis of trace organic compounds, including halocarbons, in high-temperature volcanic gas discharges, N qualitative and semiquantitative information of a large range (>100) of organic compounds found significantly above blank and ambient remote and polluted air samples, N fully quantitative information of a selected range of halocarbons, including several CFC’s (e.g. CFC-11), brominated hydrocarbons, as well as methyl iodide, N a precise quantitation of the amount of air entrainment that showed to be insignificant (factor 10-6) for our samples N a large number of relative compound concentration profiles along a traverse that includes the vegetation on onwe end and the pure volcanic gas on the other end, N correlation parameters suitable for scaling compound fluxes independently of

COSPEC-measurements, using CO2-flux measurements and CFC-11 concentrations, N quantitative static as well as dynamic source strengths of a selected number of compounds.

Well over 100 detected and quantified compounds are alkanes, alkenes, arenes, phenols, aldehydes, carboxylic acids, esters, ketones, nitriles, furans, PAH's and their halogenated, methylated and sulfonated derivatives, as well as various heterocyclic compounds including thiophenes. Most compounds are found at concentrations well above laboratory, ambient air, adsorbent and field blank levels. For some analytes (e.g., CFC-11, CH2Cl2, CH3Br), concentrations are up to several orders of magnitude greater than even mid-latitudinal industrial urban air maxima. Ozone depleting substances were found in concentrations significantly above those found in field and system blanks, including CH3Br, CH3Cl, CH3I,

C2H5Br, and chlorinated benzenes. Abundances range from upper pptv to ppmv (e.g. CFC-11: max. 1200 pptv diffusive, 3700 pptv fumarolic/dry gas, dry air: 268 pptv).

Chapter IV: Conclusions 126

The organic compounds are interpreted as the product of inorganic, abiogenic gas-phase radical reactions. On the basis of isomer abundances, n-alkane distributions and the substitution pattern the compounds are considered to have been formed by high temperature (e.g., 900 °C) alkyl free radical reactions, Fischer-Tropsch synthesis, and halide electrophilic substitution on arenes, alkanes and alkenes.

From the measurements along the profile and from blank, as well as remote and polluted local air analyses, we found vegetation to play a very minor role in subsurface degassing on the flanks and crater of the volcano, and no role in the volcanic gases. The same was found for anthropogenically derived sources, such as fuel combustion exhaust volatiles. Concentrations in the fumarolic gases are up to an order of magnitude higher in concentration compared to ambient air and local soil gases at vegetation-rich sites.

The implications of these results are that N organic chemical information can be easily acquired in volcanic gas discharges using our method,on the basis of these results, the natural volcanic source strength of halocarbon emissions to the atmosphere requires re-evaluation and correction to higher values. N The volcanic source strength of halocarbons and its impact on the Earth’s atmosphere was significant during pre-industrial times, even though current anthropogenic emissions are stronger, N single large explosive volcanic events such as the Deccan trap eruption may have influenced the Earth’s atmospheric chemistry for centuries, despite the relatively small increase of concentration of volcanic halocarbons post-eruption, N organic chemical information will offer new tools in the evaluation of volcanic gases in volcanic surveillance, and N these data offer support for the theory of a hydrothermal origin of life on Earth by abiogenic formation of important precursor molecules such as for instance nitriles and carboxylic acids.

Chapter IV: Conclusions 127

References

Adams M. C., Beall J. J., Enedy S. L., Hirtz P. N., Kilbourn P., Koenig B. A., Kunzman R., and Smith J. L. B. (2001) Hydrofluorocarbons as geothermal vapour-phase tracers. Geothermics 30 (6), 747-775. Adams M. C. and Kilbourn P. M. (2000) Thermal stability of the vapor-phase tracer R-134A. Proc. Twenty-Fifth Workshop on Geothermal Reservoir Engineering, Stanford University, Stanford, California (USA); 5 p. Adams M. C., Yamada Y., Yagi M., Kondo T., and Wada T. (2000) Stability of methanol, propanol, and SF6 as high-temperature tracers. Proc. World Geothermal Congress (WGC 2000), Beppu and Morioka (JP); 3015-3019. Adriani A., Fiocco G., Gobbi G. P., and Congeduti F. (1987) Correlated behavior of aerosol and ozone contents of the stratosphere after the El Chichon eruption. J. Geophys. Res.-Atmos. 92 (D7), 8365-8372. Allen E. T. and Day A. L. (1927) Steam wells and other thermal activity at "The Geysers" California. Carnegie Institution of Washington Publication Series Vol. 378, Carnegie Institution of Washington, Washington, D.C., 106 p. Allen E. T. and Zies E. G. (1923) A chemical study of the fumaroles of the Katmai region. Contributed Technical Papers: Katmai Series Vol. 1(2), National Geographic Society, Washington D.C. (USA), 155 p. Ammann M. (1992) Ultrafeine Teilchen in vulkanischen Emissionen [ultrafine particles in volcanic emissions]. PhD, ETH Zürich, 101 p. Andreichikov B. M., Porshnev N. V., and Bondarev V. B. (1985) The interrelation between vapor-gas mixture components in the high-temperature hydrotherms. Dokl. Akad. Nauk SSSR 280 (1), 206-211. Andronova N. G., Rozanov E. V., Yang F., Schlesinger M. E., and Stenchikov G. L. (1999) Radiative forcing by volcanic aerosols from 1850 to 1994. J. Geophys. Res.-Atmos. 104 (D14), 16807-16826. Angell J. K. and Korshover J. (1964) Quasi-biennial variations in temperature, total ozone, and tropopause height. J. Atmos. Sci. 21, 479-492. Angell J. K. and Korshover J. (1973) Quasi-biennial and long-term fluctuations in ozone. Mon. Weather Rev. 101, 426. Angell J. K., Korshover J., and Planet W. G. (1985) Ground-based and satellite evidence for a pronounced total- ozone minimum in early 1983 and responsible atmospheric layers. Mon. Weather Rev. 113 (641-646). Arctowski H. (1915) A study of the influence of volcanic dust veils on climatic variations. Science XLI (1050), 252-255. Arimura T., Shibakami M., Tamura M., Kurosawa S., and Sekiya A. (1994) Perfluorination of allene derivatives by direct fluorination. J. Chem. Res. 89, 89. Arrhenius S. (1896) On the influence of carbonic acid in the air upon the temperature of the ground. Philos. Mag. (Ser. 5) 41, 237-276. Atkinson R. (1994) Gas-phase tropospheric chemistry of organic compounds. J. Phys. Chem. Ref. Data (Monograph) 29 (2), 1-216. Aucott M. L., McCulloch A., Graedel T. E., Kleiman G., Midgley P. M., and Li Y.-F. (1999) Anthropogenic emissions of trichloromethane (chloroform, CHCl3) and chlorodifluoromethane (HCFC-22): reactive chlorine emissions inventory. J. Geophys. Res. 104 (D7), 8405-8415. Bakker J. M., Satink R. G., von Helden G., and Meijer G. (2002) Infrared photodissociation spectroscopy of benzene-Ne, Ar complex cations. Phys. Chem. Chem. Phys. 4 (1), 24-33. Ballhaus C. and Frost B. R. (1994) The generation of oxidized CO2-bearing basaltic melts from reduced CH4- bearing upper mantle sources. Geochim. Cosmochim. Acta 58 (22), 4931-1940. Banks N. and Klein F. (1979) Kilauea (Hawaii, USA); brief eruption from upper east rift zone. Scientific Event Alert Network Bulletin 4 (11). Basiuk V. A. and Navarro-Gonzalez R. (1996) Possible role of volcanic ash-gas clouds in the earth's prebiotic chemistry. Orig. Life Evol. Biosph. 26 (2), 173-194. Bassford M. R., Nickless G., Simmonds P. G., Lewis A. C., Pilling M. J., and Evans M. J. (1999) The concurrent observation of methyl iodide and dimethyl sulphide in marine air; implications for sources of atmospheric methyl iodide. Atmos. Environ. 33 (15), 2373-2383. Baubron J. C., Allard P., and Toutain J. P. (1990) Diffuse volcanic emissions of carbon dioxide from Vulcano island, Italy. Nature 344 (6261), 51-53. Bauer E. (1979) A catalogue of perturbing influences on stratospheric ozone. J. Geophys. Res. 84 (C11), 6929- 6940. Bazhenova O. K., Arefiev O. A., and Frolov E. B. (1998) Oil of the volcano Uzon caldera, Kamchatka. Org. Geochem. 29 (1-3), 421-428.

References 128

Becker K. H. and Ionescu A. (1982) Acetonitrile in the lower troposphere. Geophys. Res. Lett. 9 (12), 1349- 1351. Berndt M. E., Allen D. E., and Seyfried W. E., Jr. (1996) Reduction of CO2 during serpentinization of olivine at 300°C and 500 bar. Geology 24, 351-354. Beskrony N. S. and Lobkov V. A. (1977) Relationships of spreading of hydrocarbon gases in modern hydrothermal systems of Kamchatka. In: Geothermal Process in the Regions of Tectonic-Magmatic Activity. Nauka, Moscow. Beyerle U., Aeschbach-Hertig W., Imboden D. M., Baur H., Graf T., and Kipfer R. (2000) A mass spectrometric system for the analysis of noble gases and tritium from water samples. Environ. Sci. Technol. 34 (10), 2042-2050. Blanchette C. and McConnell J. C. (1994) Modulations of stratospheric ozone by volcanic eruptions. Proc. Quadrennial Ozone Symposium: Ozone in the Troposphere and Stratosphere, Charlottesville, VA (USA); 619-622 (abstr.). Bluth G. J. S., Rose W. I., Sprod I. E., and Krueger A. J. (1997) Stratospheric loading of sulfur from explosive volcanic eruptions. J. Geol. 105, 671-683. Bondarev V. B. and Porshnev N. V. (1980) Quantitative-Analysis of Products of the Thermal Synthesis, Modeling Volcanic-Eruption, by Gasochromatographic Method (the System CH4-H2O). Dokl. Akad. Nauk SSSR 252 (2), 455-458. Bondarev V. B., Porshnev N. V., and Nenarokov D. F. (1980) Analysis of the Composition of Vapor-Gas Shows of Kamchatka Thermal Fields by the Gasochromatographic Method. Dokl. Akad. Nauk SSSR 253 (5), 1204-1207. Bondarev V. B., Porshnev N. V., and Nenarokov D. F. (1982) Gas-chromatographic analysis of the vapors and gases discharged from the thermal fields of Kamchatka. J. Chromatogr. 247 (2), 347-351. Borisover M., Reddy M., and Graber E. (2001) Solvation effect on organic compound interactions in soil organic matter. Proc. EUG XI, Strasbourg (F); 57 (abstr.). Briffa K. R., Jones P. D., Schweingruber F. H., and Osborn T. J. (1998) Influence of volcanic eruptions on northern hemisphere summer temperature over the past 600 years. Nature 393 (6684), 450-455. Brodie B. C., Sir. (1873) Note on the synthesis of marsh-gas and formic acid, and on the electric decomposition of carbonic oxide. Proc. Royal Soc. London 21 (7), 245-247. Brombach T. (2000) Fluid geochemistry of hydrothermal systems in volcanic island arcs: Guadaloupe (Lesser Antilles) and Nisyros (Greece). PhD, Universite de Lausanne, 114 p. Brombach T., Hunziker J. C., Chiodini G., Cardellini C., and Marini L. (2001) Soil diffuse degassing and thermal energy fluxes from the southern Lakki plain, Nisyros (Greece). Geophys. Res. Lett. 28 (1), 69- 72. Brun A. (1910) Der Kohlenstoff in der Lava. Chem. Ztg. 34 (37). Bryson R. A. and Goodman B. M. (1980) Volcanic activity and climatic changes. Science 207 (4435), 1041- 1044. Bunsen R. W. (1847) Ueber den inneren Zusammenhang der pseudovulkanischen Erscheinungen Islands. Ann. Chem. 62 (1), 1-59. Bunsen R. W. (1852) Ueber vulkanische Exhalationen. Jahresberichte der Schlesischen Gesellschaft für Vaterländische Kultur 30, 29-30. Bunsen R. W. (1857) Gasometrische Methoden Vol., Friedrich Vieweg, Braunschweig (D), 305 p. Bureau H., Keppler F., and Metrich N. (1999) An experimental study of the partitioning of halogens (Cl, Br, I) between hydrous fluid and silicate melt. Terra Nova 11, 694. Bureau H., Keppler H., and Métrich N. (2000) Volcanic degassing of bromine and iodine: experimental fluid/melt partitioning data and applications to stratospheric chemistry. Earth Planet. Sci. Lett. 183 (1- 2), 51-60. Bureau H. and Metrich N. (2001) Bromine behaviour in synthetic and natural silicic melts. Proc. EUG XI, Strasbourg (FR); 786 (abstr.). Burnham W. (1979) Magmas and hydrothermal fluids. In: The Geochemistry of hydrothermal ore deposits, (ed. H. L. Barnes), pp. 71-136. John Wiley & Sons, Inc., New York. Butler J. H. (2000) Better budgets for methyl halides? Nature 403 (6767), 260-261. Butler J. H., Battle M., Bender M. L., Montzka S. A., Clarke A. D., Saltzman E. S., Sucher C. M., Severinghaus J. P., and Elkins J. W. (1999) A record of atmospheric halocarbons during the twentieth century from polar firn air. Nature 399 (6738), 749-755. Butler J. H., Montzka S. A., Clarke A. D., Lobert J. M., and Elkins J. W. (1998) Growth and distribution of halons in the atmosphere. J. Geophys. Res. 103 (D1), 1503-1511. Cadle R. D. (1975) Volcanic emissions of halides and sulfur compounds to the troposphere and stratosphere. J. Geophys. Res. 80 (12), 1650-1652.

References 129

Cadle R. D. (1980) A comparison of volcanic with other fluxes of atmospheric trace gas constituents. Rev. Geophys. 18 (4), 746-752. Cadle R. D., Lazrus A. L., Huebert B. J., Heidt L. E., Rose W. I., Woods D. C., Chuan R. L., Stoiber R. E., Smith D. B., and Zielinski R. A. (1979) Atmospheric implications of studies of central American volcanic eruption clouds. J. Geophys Res.-Oceans Atmos. 84 (NC11), 6961-6968. Calm J. M. and Didion D. A. (1998) Trade-offs in refrigant selections: past, present, and future. Int. J. Refrig. 21 (4), 308-321. Capaccioni B., Martini M., and Mangani F. (1995) Light hydrocarbons in hydrothermal and magmatic fumaroles - hints of catalytic and thermal reactions. Bull. Volcanol. 56 (8), 593-600. Capaccioni B., Martini M., Mangani F., Giannini L., Nappi G., and Prati F. (1993) Light hydrocarbons in gas emissions from volcanic areas and geothermal fields. Geochem. J. 27 (1), 7-17. Capaccioni B., Tassi F., and Vaselli O. (2001) Organic and inorganic geochemistry of low temperature gas discharges at the Baia di Levante beach, Vulcano Island, Italy. J. Volcanol. Geotherm. Res. 108 (1-4), 173-185. Casero J. J. and Joens J. A. (1997) Thermochemistry of the gas phase molecular complex of benzene with oxygen. J. Phys. Chem. A 101 (14), 2607-2609. Chaigneau M. M. (1975) Essai sur la composition des gaz magmatiques profonds. Bull. Volcanol. 39 (3), 407- 419. Chaigneau M. M., Conrad K., and Wyart M. J. (1970) Sur les gaz volcaniques de Vulcano (Iles Eoliennes). CR Acad. Sci. Paris D 271 (2), 165-167. Chamberlin R. T. (1908) The Gases in Rocks. Contributions to cosmology and the fundamental problems of geology Vol. 106, Carnegie Institution of Washington, Washington, DC, USA, 80 p. Chamberlin T. C. (1897) A group of hypotheses bearing on climatic changes. J. Geol. 5, 653-683. Chen C.-y. and Wu S.-c. (1995) The adsorption of benzene, toluene and ethylbenzene on soils near infinite dilution. Chemosphere 31 (10), 4225-4235. Chiodini G., Cioni R., Guidi M., Raco B., and Marini L. (1998) Soil CO2 flux measurements in volcanic and geothermal areas. Appl. Geochem. 13 (5), 543-552. Chiodini G., Frondini F., Cardellini C., Granieri D., Marini L., and Ventura G. (2001) CO2 degassing and energy release at Solfatara volcano, Campi Flegrei, Italy. J. Geophys. Res.-Solid Earth 106 (B8), 16213-16221. Chiodini G., Frondini F., and Raco B. (1996) Diffuse emission of CO2 from the Fossa crater, Vulcano Island (Italy). Bull. Volcanol. 58 (1), 41-50. Cicerone R. J. (1975) Comment on 'Volcanic emissions of halides and sulfur compounds to the troposphere and stratosphere' by R. D. Cadle. J. Geophys. Res. 80 (27), 3911-3912. Cicerone R. J. (1981) Halogens in the atmosphere. Rev. Geophys. Space Phys. 19 (1), 123-139. CMDL. (2001) Climate monitoring and diagnostics laboratory summary. Rep.No. 25 (1998-1999), NOAA, Boulder, CO, USA. Cody G. D., Boctor N. Z., Filley T. R., Hazen R. M., Scott J. H., Sharma A., and Yoder H. S., Jr. (2000) Primordial carbonylated iron-sulfur compounds and the synthesis of pyruvate. Science 289 (5483), 1337-1340. Coltice N., Simon L., Gillet P., and Lecuyer C. (2001) Carbon recycling and the structure of the mantle. J. Conf. Abs. 6 (1), 422. Considine D. B., Rosenfield J. E., and Fleming E. L. (2001) An interactive model study of the influence of the Mount Pinatubo aerosol on stratospheric methane and water trends. J. Geophys. Res.-Atmos. 106 (D21), 27711-27727. Couch M. W., Schmidt C. J., and Wasdo S. C. (2000) A comparison of sampling techniques for VOCs in soil. Adv. Environ. Res. 4 (1), 97-102. Corti S., Molteni F., and Palmer T. N. (1999) Signature of recent climate change in frequencies of natural atmospheric circulation regimes. Nature 398 (6730), 799-802. Courtillot V. E. (1990) A volcanic eruption. Sci. Am. 263 (4), 53-60. Cronn D. R. and Nutmagul W. (1982a) Characterization of trace gases in 1980 volcanic plumes of Mt. St. Helens. J. Geophys Res.-Oceans Atmos. 87 (NC13), 1153-1160. Cronn D. R. and Nutmagul W. (1982b) Volcanic gases in the April 1979 Soufriere eruption. Science 216 (4550), 1121-1123. Cruikshank D. P., Morrison D., and Lennon K. (1973) Volcanic gases: hydrogen burning at Kilauea volcano, Hawaii. Science 182 (4109), 277-279. Crutzen P. J. (1970) The influence of nitrogen oxides on the atmospheric ozone content. Q. J. Royal Meteor. Soc. 96, 320-325. Crutzen P. J. and Lielieveld J. (2001) Human impacts on atmospheric chemistry. Annual Reviews of Earth and Planetary Sciences 29, 17-45.

References 130

D'Allessandro W., De Gregorio s., Dongarrà G., Gurrieri S., Parello F., and Parisi B. (1997) Chemical and isotopic characterization of the gases of Mount Etna (Italy). J. Volcanol. Geotherm. Res. 78, 65-76. D'Altorio A. and Visconti G. (1983) Lidar observations of dust layers transience in the stratosphere following the El Chichon volcanic eruption. Geophys. Res. Lett. 10 (27-30). Daniel J. S., Solomon S., Portmann R. W., and Garcia R. R. (1999) Stratospheric ozone destruction: the importance of bromine relative to chlorine. J. Geophys. Res.-Atmos. D 104, 23871-23880. Danzer K. and Currie L. A. (1998) Guidelines for calibration in analytical chemistry. Pure Appl. Chem. 70 (4), 993-1014. Darling W. G. (1998) Hydrothermal hydrocarbon gases: 1, genesis and geothermometry. Appl. Geochem. 13 (7), 815-824. Davankov V. A. (2001) Retention parameters in chromatography - part B: retention parameters in gas chromatography (IUPAC recommendations 2001). Pure Appl. Chem. 73 (6), 969-992. Davis B. H., Iglesia E., Li A., Tu M., Li S., and Li W. (1999) Technology development for iron and cobalt Fischer-Tropsch catalysts. Rep.No. Quarterly Report (January 1, 1999 to March 31, 1999 - 1999), University of California, Berkeley & University of Kentucky (USA), Berkeley, CA (USA). Davy H., Sir. (1828) On the phenomena of volcanoes. Abstracts of the Papers Printed in the Philosophical Transactions of the Royal Society of London 2 (1815-1830), 338-339. Day A. L. and Shepherd E. S. (1913) L'eau et les gaz magmatiques. CR Acad. Sci. Paris 157 (20), 958-961, erratum 964. de Silva S. L. and Zielinski G. A. (1998) Global influence of the AD1600 eruption of Huaynaputina, Peru. Nature 393 (6684), 455-458. Deines P. (1968) Carbon oxygen isotopic composition of carbonates from a mica peridotite dike near Dixonville Pennsylvania. Geochim. Cosmochim. Acta 32 (6), 613-625. Delmas R. J., Kirchner S., Palais J. M., and Petit J.-R. (1992) 1000 years of explosive volcanism recorded at the South Pole. Tellus 44 B, 335-350. DePuy C. H., Kass S. R., and Bean G. P. (1988) Formation and Reactions of Heteroaromatic Anions in the Gas- Phase. J. Org. Chem. 53 (19), 4427-4433. Des Marais D. J., Donchin J. H., Nehring N. L., and Truesdell A. H. (1981) Molecular carbon isotopic evidence for the origin of geothermal hydrocarbons. Nature 292 (5826), 826-828. Dessler A. E., Stimpfle R. M., Daube B. C., Salawitch R. J., Weinstock E. M., Judah D. M., Burley J. D., Munger J. W., Wofsy S. C., Anderson J. G., McCormick M. P., and Chu W. P. (1993) Balloon-borne measurements of ClO, NO, and O3 in a volcanic cloud: an analysis of heterogenous chemistry between 20 and 30 km. Geophys. Res. Lett. 20 (22), 2527-2530. Dlugokencky E. J., Dutton E. G., Novelli P. C., Tans P. P., Masarie K. A., Lantz K. O., and Madronich S. (1996) Changes in CH4 and CO growth rates after the eruption of Mt. Pinatubo and their link with changes in tropical tropospheric UV flux. Geophys. Res. Lett. 23 (20), 2761-2764. Douda J. and Basiuk V. A. (2000) Pyrolysis of amino acids: recovery of starting materials and yields of condensation products. J. Anal. Appl. Pyrol. 56, 113-121. Duffy C. C., McCallister D. L., and Renken R. R. (1997) Carbon tetrachloride retention by modern and buried soil horizons. J. Environ. Qual. 26 (4), 1123-1127. Dütsch H. U. (1984) Total ozone trend in the light of ozone soundings, the impact of El Chichon. In: Atmospheric Ozone, (ed. C. S. Zerefos and A. Ghazi). D. Reidel Publishing, Dordrecht, NL. Eaton G. (1977) Kilauea (Hawaii, USA); eruption near the Kalalua cone ends on 10 October. Scientific Event Alert Network Bulletin 2 (10). Eggenkamp H. G. M. and Kreulen R. (1984) Chlorine stable isotopes in volcanic gases, volcanic springs and crater lakes from Indonesia. Geologica Ultraiectina Vol. 93, 102 p. Ehalt D. H. (1974) The atmospheric cycle of methane. Tellus 26 (1-2), 58-70. Emmenegger F. P. (1987) Stability of complex metal bromides in the gas-phase and in toluene. Z. Anorg. Allg. Chem. 545 (2), 56-68. Fabian P., Borchers R., Kruger B. C., and Lal S. (1987) CF4 and C2F6 in the atmosphere. J. Geophys. Res.- Atmos. 92 (D8), 9831-9835. Farman J. C., Gardiner B. G., and Shanklin J. D. (1985) Large losses of ozone in Antarctica reveal seasonal ClOx/NOx interaction. Nature 315 (6016), 207-210. Fegley B. (1991) Thermodynamic models of the chemistry of lunar volcanic gases. Geophys. Res. Lett. 18 (11), 2073-2076. Fegley B. and Zolotov M. Y. (2000) Carbon chemistry of volcanic gases on Io. Meteorit. Planet. Sci. 35, A52- A52. Fink L. E. (1983) Characterization of organic contaminants in environmental samples associated with the Mount St. Helens 1980 volcanic eruption - Comment. Environ. Sci. Technol. 17 (5), 312-313.

References 131

Finlayson J. B., Barnes I. L., and Naughton J. J. (1968) Developments in volcanic gas research in Hawaii. In: The crust and upper mantle of the Pacific Basin, AGU Geophysical Monograph Vol. 12 (ed. L. Knopoff, C. L. Drake, and P. J. Hart), pp. 428-438. AGU Geophysical Monograph, Washington, DC (USA). Fischer F. (1935) Die Synthese der Treibstoffe (Kogasin) und Schmieröle aus Kohlendioxyd und Wasserstoff bei gewöhnlichem Druck. Brennst. Chem. 16, 1-11. Fischer T. P., Giggenbach W. F., Sano Y., and Williams S. N. (1998) Fluxes and sources of volatiles discharged from Kudryavy, a subduction zone volcano, Kurile Islands. Earth Planet. Sci. Lett. 160 (1-2), 81-96. Florovskaia V. N., Teplitskaia T. A., Markhinin E. K., and Podkletnov N. E. (1978) Polynuclear arenes in products of volcanic eruptions. Dokl. Akad. Nauk SSSR 240 (1), 221-222. Fulignati P., Gioncada A., and Sbrana A. (1998) Geologic model of the magmatic-hydrothermal system of Vulcano (Aeolian Islands, Italy). Miner. Petrol. 62, 195-222. Foukal, P. and Lean J. (1996) An empirical model of total solar irradiance variation between 1874 and 1988. Science 247 (4942), 556-558. Fouqué F. (1865a) Sur l'éruption de l'Etna en 1865 et sur les phénomènes eruptifs de l'Italie méridionale (5eme lettre à Ch. Sainte-Claire Deville). CR Acad. Sci. Paris LXI (10), 421-424. Fouqué F. (1865b) Sur les phénomènes éruptifs de l'Italie méridionale (7eme lettre à Ch. Sainte-Claire Deville). CR Acad. Sci. Paris LXI (18), 734-737. Fouqué F. (1866a) Recherches sur les phénomènes chimiques des volcans. Compt. Rend. hebd. S. Acad. Sci. Paris 62 (26), 1366-1377. Fouqué F. (1866b) Recherches sur les phénomènes chimiques des volcans. CR Acad. Sci. Paris LXII (11), 616- 617. Fouqué F. (1879) Santorin et ses éruptions Vol., Masson, Paris, 433 p. Francis P., Maciejewski A., Oppenheimer C., Chaffin C., and Caltabiano T. (1995) SO2:HCl ratios in the plumes from Mt Etna and Vulcano determined by Fourier Transform Spectroscopy. Geophys. Res. Lett. 22 (13), 1717-1720. Franklin B. (1784) Meteorological imaginations and conjectures. Memoirs of the Literary and Philosophical Society of Manchester 2, 373-377. Froudakis G. E., Andriotis A. N., and Menon M. (2001) Structural properties of metal-benzene, M- n(benzene)(m), M = Ni, V complexes: an ab initio study. Chem. Phys. Lett. 350 (5-6), 393-398. Fulignati P., Gioncada A., and Sbrana A. (1998) Geologic model of the magmatic-hydrothermal system of Vulcano (Aeolian Islands, Italy). Miner. Petrol. 62, 195-222. Gaffney J. S. (1995) Volcanic CFCs. Environ. Sci. Technol. 29 (1), A8-A8. Gan J. Y., Megonnell N. E., and Yates S. R. (2001) Adsorption and catalytic decomposition of methyl bromide and methyl iodode on activated carbons. Atmos. Environ. 35 (5), 941-947. Gaffney J. S. (1995) Volcanic CFCs. Environ. Sci. Technol. 29 (1), A8-A8. Gassmann M. (1974) Freon 14 im "Krypton reinst" und in der Atmosphäre. Naturwissenschaften 61 (2), 127. Gay-Lussac L. J. (1823) Réflexions sur les volcans. Ann. Chim. Phys. tome XXII, 415-429. Gerlach T. M. (1979) Evaluation and restoration of the 1970 volcanic gas analyses from Mount Etna, Sicily. J. Volcanol. Geotherm. Res. 6, 165-178. Gerlach T. M. (1980) Evaluation of volcanic gas analyses from Kilauea volcano. J. Volcanol. Geotherm. Res. 7, 295-317. Gerlach T. M. (1991) Present-day CO2 emissions from volcanoes. EOS Transact. AGU 72, 249-256. Gerlach T. M. and Graeber E. (1985) Volatile budget of Kilauea volcano. Nature 313 (6000), 273-277. Gex P. (1992) Etude géophysique des environs du cratère de Vulcano, Italie. B. Soc. Vaudoise Sci. Nat. 82 (2), 157-172. Gherardi F. and Panichi C. (2001) Chemical and isotopic composition of light hydrocarbon gases in hydrothermal systems: preliminary results from three Italian thermal areas. In 10th International Symposium on Water-Rock Interaction, WRI-10 (ed. R. Cidu), pp. 835-838. Swets & Zeitlinger, Lisse: Balkema (NL). Giammanco S., Inguaggiato S., and Valenza M. (1998) Soil and fumarole gases of Mount Etna: geochemistry and relations with volcanic activity. J. Volcanol. Geotherm. Res. 81 (3-4), 297-310. Giggenbach W. F. (1980) Geothermal gas equilibria. Geochim. Cosmochim. Acta 44, 2021-2032. Giggenbach W. F. (1996) Chemical composition of volcanic gases. In: Monitoring and mitigation of volcanic hazards, (ed. Scarpa and R. I. Tilling), pp. 221-256. Springer Verlag, Berlin. Giggenbach W. F. (1997) The origin and evolution of fluids in magmatic-hydrothermal systems. In: Geochemistry of Hydrothermal Ore Deposits, (ed. H. L. Barnes), pp. 737-796. John Wiley and Sons, Inc., New York.

References 132

Giggenbach W. F., Garcia N., Londono A., Rodriguez L., Rojas N., and Calvache M. L. (1990) The chemistry of fumarolic vapor and thermal-spring discharges from the Nevado-Del-Ruiz volcanic-magmatic- hydrothermal system, Colombia. J. Volcanol. Geotherm. Res. 42 (1-2), 13-39. Giggenbach W. F. and Goguel R. L. (1989) Collection and analysis of geothermal and volcanic water and gas discharges (4th ed.). Rep.No. CD 2401, Chemistry Division, Department of Scientific and Industrial Research (DSIR), Petone, New Zealand. Giggenbach W. F. and Le Guern F. N. (1976) Chemistry of magmatic gases from Erta'ale, Ethiopia. Geochim. Cosmochim. Acta 40 (1), 25-30. Giggenbach W. F. and Matsuo S. (1991) Evaluation of results from second and third IAVCEI field workshops on volcanic gases, Mt Usu, Japan, and White Island, New Zealand. Appl. Geochem. 6, 125-141. Giggenbach W. F., Tedesco D., Sulistiyo Y., Caprai A., Cioni R., Favara R., Fischer T. P., Hirabayashi J.-I., Martini M., Menyailov I., and Shinohara H. (2001) Evaluation of results from the fourth and fifth IAVCEI field workshops on volcanic gases, Vulcano island, Italy and Java, Indonesia. J. Volcanol. Geotherm. Res. 108, 157-172. Gill J. B. (1981) Orogenic andesites and plate tectonics. Minerals and rocks Vol. 16, Springer Verlag, Berlin (D), 390 p. Giocanda A. and Sbrana A. (1991) "La Fossa Caldera", Vulcano, inferences from deep drillings. Acta Vulcanol. 1, 115-125. Giże A. and Macdonald R. (1993) Generation of compositionally atypical hydrocarbons in CO2-rich geologic environments. Geology 21 (2), 129-132. Giże A. P. (1987) The organic geochemistry of two Mississippi Valley-type lead-zinc deposits. Econ. Geol. 82, 457-470. Giże A. P. (1999) Organic alteration in hydrothermal sulfide ore deposits. Econ. Geol. 94, 967-980. Giże A. P. and Manning D. A. C. (1993) Aspects of the organic geochemistry and petrology of metalliferous ores. In: Organic Geochemistry, (ed. M. H. Engel and S. A. Macko), pp. 565-580. Plenum Press, New York (USA). Gleason J. F., Bhartia P. K., Herman J. R., McPeters R., Newman P., Stolarski R. S., Flynn L., Labow G., Larko D., Seftor C., Wellemeyer C., Komhyr W. D., Miller A. J., and Planet W. (1993) Record low global ozone in 1992. Science 260 (5107), 523-526. Gloser V., Jezikova M., Lüscher A., Frehner M., Blum H., Nösberger J., and Hartwig U. A. (2000) Soil mineral nitrogen availability was unaffected by elevated atmospheric pCO2 in a four years old field experiment (Swiss FACE). Plant Soil 227, 291-299. Gold T. (1979) Terrestrial sources of carbon and earthquake outgassing. J. Petrol. Geol. 1 (3), 3-19. Gold T. and Soter S. (1980) The Deep-Earth-Gas hypothesis. Sci. Am. 242, 130-137. Gold T. and Soter S. (1982) Abiogenic methane and the origin of petroleum. Energ. Explor. Exploit. 1 (2), 89- 104. Golitsyn G. S., Grechko E. I., Elansky N. F., and Pugachev N. S. (1991) Some Soviet measurements of trace gases. Tellus A 43 (4), 164-175. Goodmann B. M. (1984) The climatic impact of volcanic acitvity. PhD, The University of Wisconsin, 264 p. Gorceix H. (1872a) Etat du Vésuve et des dégagements gazeux des Champs Phlégréens, au mois de juin 1869. CR Acad. Sci. Paris LXXV (3), 154-156. Gorceix H. (1872b) Etude des dégagements gazeux de Santorin pendant la fin de l'éruption de 1866. CR Acad. Sci. Paris LXXV (5), 270-272. Gorceix H. (1872c) Résumé des phénomènes dont le volcan de Santorin a été le siége, à la fin de l'éruption de 1866 (de décembre 1869 au mois d'octobre 1871). CR Acad. Sci. Paris LXXV (6), 372-374. Goss K.-U. (1992) Effects of temperature and relative humidity on the sorption of organic vapours on quartz sand. Environ. Sci. Technol. 26, 2287-2294. Goss K.-U. (1993) Effects of temperature and relative humidity on the sorption of organic vapours on clay minerals. Environ. Sci. Technol. 27, 2127-2132. Graeber E. J., Modreski P. J., and Gerlach T. M. (1979) Composition of the gases collected during the 1977 East Rift eruption, Kilauea, Hawaii. J. Volcanol. Geotherm. Res. 5, 337-344. Gribble G. W. (1992a) Naturally occurring organohalogen compounds - a survey. J. Nat. Prod. 55 (10), 1353- 1395. Gribble G. W. (1992b) Sources of halogenated alkanes. Chem. Eng. News 70 (45), 3-3. Gribble G. W. (1994a) The abundant natural sources and uses of chlorinated chemicals. Am. J. Public Health 84 (7), 1183-1183. Gribble G. W. (1994b) Natural chlorine? naturally! Today's Chemist at Work 3 (7), 62-64. Gribble G. W. (1994c) Natural organohalogens - many more than you think. J. Chem. Educ. 71 (11), 907-911. Gribble G. W. (1994d) The natural production of chlorinated compounds. Environ. Sci. Technol. 28 (7), A310- A319.

References 133

Gribble G. W. (1994e) Volcanic emissions. Chem. Ind. 94 (11), 390. Gribble G. W. (1995a) The diversity of natural organohalogen compounds. Abstr. Pap. Am. Chem. Soc. 209 (CHED), 482. Gribble G. W. (1995b) The natural production of organochlorine compounds. Abstr. Pap. Am. Chem. Soc. 210 (ENV), 65. Gribble G. W. (1995c) Volcanic CFCs - Response. Environ. Sci. Technol. 29 (1), A8-A8. Gribble G. W. (1996a) The diversity of natural organochlorines in living organisms. Pure Appl. Chem. 68 (9), 1699-1712. Gribble G. W. (1996b) Naturally occurring organohalogen compounds - a comprehensive survey. Prog. Org. Nat. Prod. 68, 1-423. Gribble G. W. (1998) Naturally occurring organohalogen compounds. Accounts Chem. Res. 31 (3), 141-152. Gribble G. W. (1999) The diversity of naturally occurring organobromine compounds. Chem. Soc. Rev. 28 (5), 335-346. Gribble G. W. (2000) The natural production of organobromine compounds. Environ. Sci. Pollut. Res. 7 (1), 37- 49. Gribble G. W. (2001) Kamchatka and Montreal. Am. Sci. 89 (2), 107-107. Gribble G. W. and Worman J. J. (1998) Natural organohalogens - an update. Abstr. Pap. Am. Chem. Soc. 216 (CHED), 426. Grossman A. S., Grant K. E., Blass W. E., and Wuebbles D. J. (1997) Radiative forcing calculations for CH3Cl and CH3Br. J. Geophys. Res. 102 (D12), 13651-13656. Haggerty J. A. (1991) Evidence from fluid seeps atop serpentine seamounts in the Mariana fore-arc - clues for emplacement of the seamounts and their relationship to fore-arc tectonics. Mar. Geol. 102 (1-4), 293- 309. Hansen A. D. A. (1997) Measurement of combustion effluent aerosols from Amundsen-Scott South Pole Station. Antarct. J. US (reviews) 32 (5), 163-168. Hanson B. M. (1993) What happened to science? AAPG Bulletin 77 (9), 1573. Harnisch J., Borchers R., Fabian P., Gaggeler H. W., and Schotterer U. (1996) Effect of natural tetrafluoromethane. Nature 384 (6604), 32-32. Harnisch J. and Eisenhauer A. (1998) Natural CF4 and SF6 on Earth. Geophys. Res. Lett. 25 (13), 2401-2404. Harnisch J., Frische M., Borchers R., Eisenhauer A., and Jordan A. (2000) Natural fluorinated organics in fluorite and rocks. Geophys. Res. Lett. 27 (13), 1883-1886. Harris A. J. L., Sherman S. B., and Wright R. (2000) Discovery of self-combusting volcanic sulfur flows. Geology 28 (5), 415–418.

Hedberg H. D. (1964) Geologic aspects of the origin of petroleum. AAPG Bull. 48 (11), 1755-1803. Heide K., Gerth K., and Hartmann E. (2000) The detection of an inorganic hydrocarbon formation in silicate melts by means of a direct-coupled-evolved-gas-analysis-system (DEGAS). Thermochim. Acta 354 (1- 2), 165-172. Hempel W. (1915) Über die vulkanische Tätigkeit. Z. Vulkanol. 1 (4), 150-167. Higashihara M., Igari S., Maekawa T., Noda T., Sakata S., Asada N., and MiyatakE A. (1997) C5+ hydrocarbons in fumarolic gases from the Kakkonda (Takinoue) geothermal area, Japan. Geochem. J. 31 (2), 63-73. Hoefs J. (1973) Ein Beitrag zur Isotopengeochemie des Kohlenstoffs in magmatischen Gesteinen. Contrib. Min. Petrol. 41, 277-300. Hoekstra E. J., Duyzer J. H., de Leer E. W. B., and Brinkmann U. A. T. (2001) Chloroform - concentration gradients in soil air and atmospheric air, and emission fluxes from soil. Atmos. Environ. 35, 61-70. Hofer M. and Imboden D. M. (1998) Simultaneous determination of CFC-11, CFC-12, N2, and Ar in water. Anal. Chem. 70 (4), 724-729. Hofmann D. J., Oltmans S. J., Harris J. M., Solomon S., Deshler T., and Johnson B. J. (1992) Observation and possible causes of new ozone depletion in Antarctica in 1991. Nature 359 (6393), 283-287. Hofmann D. J., Oltmans S. J., Komhyr W. D., Harris J. M., Lathrop J. A., Langford A. O., Deshler T., Johnson B. J., Torres A., and Matthews W. A. (1994) Ozone loss in the lower stratosphere over the United States in 1992-1993 - evidence for heterogeneous chemistry on the Pinatubo aerosol. Geophys. Res. Lett. 21 (1), 65-68. Hofmann D. J. and Solomon S. (1989) Ozone destruction through heterogeneous chemistry following the eruption of El Chichon. J. Geophys. Res.-Atmos. 94 (D4), 5029-5041. Holloway J. R. and O'Day P. A. (2000) Production of CO2 and H2 by diking-eruptive events at mid-ocean ridges: Implications for abiotic organic synthesis and global geochemical cycling. Int. Geol. Rev. 42 (8), 673- 683. Inn E. C. Y., Vedder J. F., Condon E. P., and Ohara D. (1981) Gaseous constituents in the plume from eruptions of Mount St. Helens. Science 211 (4484), 821-823.

References 134

IPCC. (2001) Climate change: the scientific basis. Rep.No. Third Assessment Report, UNEP/WMO, Nairobi, Kenya. Isidorov V. A. (1985) Organicheskaya khimiya atmosfery (The Organic Chemistry of the Atmosphere) [in Russian] Vol., Khimiya, Leningrad, 215 p. Isidorov V. A. (1990) Organic Chemistry of the Earth's Atmosphere Vol., Springer, Berlin, 215 p. Isidorov V. A. (1992) Volatile organic compounds in steam-gas outflows of several volcanoes and hydrothermal systems in Kamchatka. Volcanol. Seismol. 13 (3), 287-293. Isidorov V. A. (1995) Isidorovs response. Environ. Sci. Technol. 29 (1), A8. Isidorov V. A. and Ioffe B. V. (1986) Non-antropogenic sources of earth atmosphere halogenocarbons. Dokl. Akad. Nauk SSSR 287 (1), 86-90. Isidorov V. A., Titov V. S., and Strukova T. P. (1985) Halogen hydrocarbons in the composition of volcanic gases. Meteorologiya i Gidrologiya 10, 110-113 (translation); 229. Isidorov V. A. and Zenkevich I. G. (1985) The atmosphere organic components in the regions of volcanic activity. Dokl. Akad. Nauk SSSR 280 (1), 223-227. Isidorov V. A., Zenkevich I. G., and Ioffe B. V. (1990) Volatile organic compounds in solfataric gases. J. Atmos. Chem. 10 (3), 329-340. Iwasaki I., Katsura T., Ozawa T., Yoshida M., and Iwasaki B. (1968) Chlorine content of volcanic rocks and migration of chlorine from the mantle to the surface of the Earth. In: The crust and upper mantle of the Pacific Basin, AGU Geophysical Monograph Vol. 12 (ed. L. Knopoff, C. L. Drake, and P. J. Hart), pp. 428-439. AGU Geophysical Monograph, Washington, DC (USA). Jäger H., Mücke R., and Harris J. (2000) The stratospheric aerosol layer after the decay of the Pinatubo perturbation. Proc. 5th European Symposium: Stratospheric Ozone 1999 (Air pollution research report 73), Saint Jean de Luz, FR; 87-90 (abstr.). Jambon A. (1995) Earth degassing and large-scale geochemical cycling of volatile elements. In: Volatiles in Magmas, Reviews in Mineralogy Vol. 30 (ed. M. R. Caroll and J. R. Holloway), pp. 479-517, Washington, DC (USA). Javoy M., Pineau F., and Allègre C. J. (1982) Carbon geodynamic cycle. Nature 300, 171-173. Johnston D. A. (1980) Volcanic contribution of chlorine to the stratosphere: more significant to ozone than previously estimated? Science 209 (4455), 491-493. Jones C. D. and Cox P. M. (2001) Modeling the volcanic signal in the atmospheric CO2 record. Global Biogeochem. Cy. 15 (2), 453-465. Jordan A., Harnisch J., Borchers R., Le Guern F. N., and Shinohara H. (2000) Volcanogenic halocarbons. Environ. Sci. Technol. 34 (6), 1122-1124. Kanters J. and Louw R. (1996) Thermal and catalysed halogenation in combustion reactions. Chemosphere 1, 89-97. Keene W. C., Khalil M. A. K., Erickson III D. J., McCulloch A., Graedel T. E., Lobert J. M., Aucott M. L., Gong S. L., Harper D. B., Kleiman G., Midgley P. M., Moore R. M., Seuzaret C., Sturges W. T., Benkovitz C. M., Koropalov V., Barrie L. A., and Li Y.-F. (1999) Composite global emissions of reactive chlorine from anthropogenic and natural sources: reactive chlorine emissions inventory. J. Geophys. Res. 104 (D7), 8429-8440. Kelley D. S. and Früh-Green G. L. (1999) Abiogenic methane in deep-seated mid-ocean ridge environments: insights from stable isotope analyses. J. Geophys. Res.-Solid Earth 104 (B5), 10439-10460. Kelley D. S. and Früh-Green G. L. (2000) Volatiles in mid-ocean ridge environments. In: Ophiolites and Oceanic Crust: New Insights from Field Studies and the Ocean Drilling Program, Geological Society of America Special Paper Vol. 349 (ed. Y. Dilek, E. M. Moores, D. Elthon, and A. Nicolas), pp. 237- 260. Geological Society of America, Boulder, Colorado. Kelley D. S. and Früh-Green G. L. (2001) Volatile lines of descent in submarine plutonic environments: insights from stable isotope and fluid inclusion analyses. Geochim. Cosmochim. Acta 65 (19), 3325-3346. Keller J. (1970) Carta geologica dell'Isola di Vulcano (Isole Eolie). 1:10,000. Keller J. (1980) The island of Vulcano. Rend. Soc. Ital. Min. Petrol. 36 (1), 369-414. Kelly T. J., Ramamurthi M., Pollack A. J., Spicer C. W., and Cupitt L. T. (1993) Ambient concentration summaries for Clean Air Act Title III hazardous air pollutants. Rep.No. U.S. EPA Contract No. 68- D80082, United States Environmental Protection Agency (U.S. EPA). Kelly W. J. (2001) Computational reevaluation of benzene-chlorine atom complexation. Abstr. Pap. Am. Chem. Soc. 221 (COMP), 135. Keppler F., Eiden R., Niedan V., Pracht J., and Scholer H. F. (2000) Halocarbons produced by natural oxidation processes during degradation of organic matter. Nature 403 (6767), 298-301. Keppler F., Eiden R., Niedan V., Pracht J., and Scholer H. F. (2001) Correction on: halocarbons produced by natural oxidation processes during degradation of organic matter (vol 403, pg 298, 2000). Nature 409 (6818), 382.

References 135

Kerr R. A. (1989) Volcanoes can muddle the greenhouse. Science 245 (4914), 127-128. Kerrick D. M., McKibben M. A., Seward T. M., and Caldeira K. (1995) Convective hydrothermal CO2 emission from high heat-flow regions. Chem. Geol. 121 (1-4), 285-293. Khalil M. A. K., Moore R. M., Harper D. B., Lobert J. M., Erickson D. J., Koropalov V., Sturges W. T., and Keene W. C. (1999) Natural emissions of chlorine-containing gases: reactive chlorine emissions inventory. J. Geophys. Res. 104 (D7), 833-8346. Khalil M. A. K. and Rasmussen R. A. (1989) Temporal variations of trace gases in ice cores. In: The environmental record in glaciers and ice sheets (Dahlem Konferenz, Berlin), (ed. H. Oeschger and C. C. Langway), pp. 193-206. John Wiley & Sons Ltd., Kipfer R. (1991) Primordiale Edelgase als Tracer für Fluide aus dem Erdmantel. PhD, ETH Zürich, 223 p. Kiyoshu Y. and Asada N. (1995) Light hydrocarbons in volcanic gases from the Japanese island-arc. Geochem. J. 29 (4), 231-242. Koepenick K. W. (1995) Volatile emissions from Oldoinyo Lengai volcano, Tanzania. Masters, Pennsylvania State University, 75 p. Kolesnikov M. P. and Egorov I. A. (1979) Metalloporphyrins and molecular complexes of amino acids with porphyrins in juvenile volcanic ash. Orig. Life Evol. Biosph. 9 (4), 267-277. Kolesnikov M. P., Voronova N. I., and Egorov I. A. (1979) Metalloporphyrins and perylene in the products of volcano eruption. Dokl. Akad. Nauk SSSR 247 (3), 749-752. Komhyr W. H., Oltmans S. J., Chopra A. N., Leonard R. K., Garcia T. E., and McFee C. (1984) Results of Umkehr, Ozonesonde, and sulfur dioxide observations in Hawaii following the eruption of El Chichon volcano in 1982. In: Atmospheric ozone, (ed. C. S. Zerefos and A. Ghazi). D. Reidel Publishing, Dordrecht, NL. Kondratev K. Y. and Varotsos C. A. (1995a) Atmospheric ozone variability in the context of global change. Int. J. Remote Sensing 16 (10), 1851-1881. Kondratev K. Y. and Varotsos C. A. (1995b) Volcanic eruptions and global ozone dynamics. Int. J. Remote Sensing 16 (10), 1887-1895. Korolev A. V., Delpech F., Dagorne S., Guzei I. A., and Jordan R. F. (2001) Main-group-metal chlorobenzene complexes. Organometallics 20 (16), 3367-3369. Kovats E. (1958) Gas-Chromatographische Charakterisierung Organischer Verbindungen .1. Retentionsindices Aliphatischer Halogenide, Alkohole, Aldehyde Und Ketone. Helv. Chim. Acta 41 (7), 1915-1932. Kranz R. (1966) Organische Fluorverbindungen in den Gaseinschlüssen der Wölsendorfer Flußspäte. Naturwissenschaften 53, 593-600. Kravtsov A. I., Bobrov V. A., Kropotova O. I., Markhinin E. K., and Podkletnov N. E. (1979) Carbon isotopy of gases and organic compounds in juvenile material of volcanos of the Kuril-Kamchatka volcanic Arc. Geokhimiya (10), 1554-1559. Krot A. N., Posukhova T. V., Guseva E. V., Galimov E. M., Botkunov A. I., Ramenskaya M. E., and Oglobina A. I. (1993) Genesis of the garnets containing hydrocarbon inclusions (Mir kimberlite pipe). Geokhimiya (6), 891-899. Kryachko E. S., Arbuznikov A. V., and Hendrickx M. F. A. (2001) Modeling chemisorption of benzene and its hydrogenation on platinum surfaces. 1. Complexes of benzene with Pt and Pt2. J. Phys. Chem. B 105 (17), 3557-3566. Kuypers M. M. M., Pancost R. D., and Sinninghe Damsté J. S. (1999) A large and abrupt fall in atmospheric CO2 concentration during Cretaceous times. Nature 399 (6734), 342-345. Labitzke K., Naujokat B., and McCormick M. P. (1983) Temperature effects on the stratosphere of April 4, 1982 eruption of El Chichon. Geophys. Res. Lett. 10 (24-26). Lancet M. S. and Anders E. (1970) Carbon-isotope fractionation in the Fischer-Tropsch synthesis and in meteorites. Science 170 (3961), 980-982. Landais P. and Giże A. P. (1997) Organic matter in hydrothermal ore deposits. In: The Geochemistry of hydrothermal ore deposits, (ed. H. L. Barnes), pp. 613-655. John Wiley & Sons, New York (USA). Laughrey C. D. (1998) The origin of oil. Pennsylvania Geology 29 (1), 9-14. Law K. S. (1999) Theoretical studies of carbon monoxide distributions, budgets and trends. Chemosphere: Glob. Change Sci. 1, 19-31. Lazrus A. L. (1979) Sulfur and halogen chemistry of the stratosphere and of volcanic eruption plumes. J. Geophys. Res. 84 (C12), 7869-7875. Lee-Taylor J. M., Doney S. C., and Brasseur G. P. (1998) A global three-dimensional atmosphere-ocean model of methyl bromide distributions. J. Geophys. Res.-Atmos. 103 (D13), 16039-16057. Legrand M. R., Delmas R. J., and Charlson R. J. (1988) Climate forcing implications from Vostok ice-core sulphate data. Nature 334 (6181), 418-420. Leif R. N. and Simoneit B. R. T. (1995) Ketones in hydrothermal petroleums and sediment extracts from Guyamas Basin, Gulf of California. Org. Geochem. 23 (10), 889-904.

References 136

Leifer R., Juzdan Z. R., and Larsen R. (1984) COS in the stratosphere - El Chichon observations. Geophys. Res. Lett. 11 (5), 549-552. Leifer R., Sommers K., and Guggenheim S. F. (1981) Atmospheric trace gas measurements with a new clean air sampling system. Geophys. Res. Lett. 8 (10), 1079-1081. Levine J. S. (1998), pers.comm. In: Lobert et al., 1999 Lobert J. M., Keene W. C., Logan J. A., and Yevich R. (1999) Global chlorine emissions from biomass burning: Reactive Chlorine Emissions Inventory. J. Geophys. Res. 104 (D7), 8373-8389. Lobert J. M. and Warnatz J. (1993) Emissions from the combustion processs in vegetation. In: Fire in the environment: the ecological, atmospheric, and climate importance of vegetation fires, (ed. P. J. Crutzen and J. G. Goldammer), pp. 15-37. John Wiley & Sons Ltd., Chichester (UK). Lockwood M., Stamper R., and Wild M. N. (1999) A doubling of the Sun's coronal magnetic field during the past 100 years. Nature 399 (6735), 437 - 439). Lovelock J. E. (1971) Atmospheric fluorine compounds as indicators of air movements. Nature 230 (5293), 379. Lovelock J. E. (1975) Natural halocarbons in the air and in the sea. Nature 256 (5514), 193-194. Lüders V., Banks D., and Halbach P. (2001) Cl/Br and d 37Cl isotope fractionation in fluids of modern submarine hydrothermal systems. Eur. J. Mineral. 13, 113. Lüscher A. and Nösberger J. (1997) Interspecific and intraspecific variation in the response of grasses and legumes to free air CO2 enrichment. Acta Oecol. 18, 269-276. Maduro R. A. and Schauerhammer R. (1992) The holes in the ozone scare - the scientific evidence that the sky isn't falling Vol., 21st Century Science Associates, Washington D.C., USA, 357 p. Mangani F., Cappiello A., Capaccioni B., and Martini M. (1991) Sampling and analysis of light hydrocarbons in volcanic gases. Chromatographia 32 (9-10), 441-444. Mango F. D. (1992) Transition metal catalysis in the generation of petroleum and natural gas. Geochim. Cosmochim. Acta 56, 553-555. Mango F. D. (1996) Transition metal catalysis in the generation of natural gas. Org. Geochem. 24 (10-11), 977- 984. Mann M. M., Bradley R. S., and Hughes M. K. (1998) Global-scale temperature patterns and climate forcing over the past six centuries. Nature 392 (6678), 779-787. Manning D. A. C. and Giże A. P. (1993) The role of organic matter in ore transport processes. In: Organic Geochemistry, (ed. M. H. Engel and S. A. Macko), pp. 547-563. Plenum Press, New York (USA). Markhinin E. K. (1985) Vulkanizm (Volcanism) [in Russian] Vol., Nedra, Moscow, RU, p. Markhinin E. K. and Podkletnov N. E. (1977a) The phenomenon of formation of pre-biological compounds in volcanic processes. Orig. Life Evol. Biosph. 8 (3), 225-235. Markhinin E. K. and Podkletnov N. E. (1977b) Phenomenon of pre-biological compound formation observed at eruption of volcano Tolbachik. Dokl. Akad. Nauk SSSR 235 (5), 1203-1206. Markhinin E. K. and Podkletnov N. E. (1978) Hydrocarbons and other complex organic compounds in volcanic products. Geolog. Geofiz. 19 (12), 21-32. Markhinin E. K., Podkletnov N. E., and Zbrueva A. I. (1975) Amino acids, hydrocarbons and other organic compounds in juvenile jolcanic ash. Dokl. Akad. Nauk SSSR 222 (6), 1438-1440. Markhinin E. K., Urakov V. A., and Podkletnov N. E. (1977) Hydrocarbons in gases of basalt lava streams of Tolbachik Fissure Eruption of 1975-1976. Dokl. Akad. Nauk SSSR 236 (5), 1214-1217. Markhinin E. K., Urakov V. A., Podkletnov N. E., and Ponomarev V. V. (1983) Gases from the basalt flows of the Tolbachik fissure eruption 1975-1976. In: The Great Tolbachik Fissure Eruption; geological and geophysical data 1975-1976, (ed. Fedotov-SA and Markhinin-E.K.), pp. 156-163. Cambridge University Press, Cambridge (UK). Martini M. (1984) On the behaviour of fluorine in volcanic processes. Bull. Volcanol. 47 (3), 483-489. Mathez E. A. (1987) Carbonaceous matter in mantle xenoliths: composition and relevance to isotopes. Geochim. Cosmochim. Acta 51, 2339-2347. Mathez E. A. and Delaney J. R. (1981) The nature and distribution of carbon in submarine basalts and peridotite nodules. Earth Planet. Sci. Lett. 56, 217-232. Mathez E. A., Dietrich V. J., and Irving A. J. (1984) The geochemistry of carbon in mantle peridotites. Geochim. Cosmochim. Acta 48, 1849-1859. Matveev S., Ballhaus C., Fricke K., Truckenbrodt J., and Ziegenbein D. (1997) Volatiles in the Earth's mantle: I. synthesis of CHO fluids at 1273 K and 2.4 GPa. Geochim. Cosmochim. Acta 61 (15), 3081-3088. Mazzera D. M., Lowenthal D. H., Chow J. C., and Watson J. G. (2001) Sources of PM10 and sulfate aerosol at McMurdo station, Antarctica. Chemosphere 45 (3), 347-356. McCollom T. M. and Seewald J. S. (2001) A reassessment of the potential for reduction of dissolved CO2 to hydrocarbons during serpentinization of olivine. Geochim. Cosmochim. Acta 65 (21), 3769-3778. McCollom T. M., Seewald J. S., and Simoneit B. R. T. (2001) Reactivity of monocyclic aromatic compounds under hydrothermal conditions. Geochim. Cosmochim. Acta 65 (3), 455-468.

References 137

McCormick M. P., Thomason L. W., and Trepte C. R. (1995) Atmospheric effects of the Mt. Pinatubo eruption [review]. Nature 373 (6513), 399-404. McCulloch A. and Midgley P. M. (1996) The production and global distribution of emissions of trichloroethylene, tetrachloroethene and dichloromethane over the period 1988-1992. Atmos. Environ. 30 (4), 601-608. McGee T. J., Newman P., Gross M., Singh U., Godin S., Lacoste A. M., and Megie G. (1994) Correlation of ozone loss with the presence of volcanic aerosols. Geophys. Res. Lett. 21 (25), 2801-2804. McInturff R. M., Miller A. J., Angell J. K., and Korshover J. (1971a) Possible effects on stratosphere of 1963 Mt Agung volcanic eruption. EOS Transact. AGU 52 (4), 219. McInturff R. M., Miller A. J., Angell J. K., and Korshover J. (1971b) Possible effects on stratosphere of 1963 Mt Agung volcanic eruption. J. Atmos. Sci. 28 (7), 1304-1307. McLinden M. O. and Didion D. A. (1987) Quest for alternatives. ASHRAE J. 29 (12), 32-36. Menyaylov I. A., Nikitina L. P., Vargasova L. P., Guseva R. V., Shapar' V. N., and Rozhkov A. M. (1983) Chemistry, metal content and reaction products of gases from the New Tolbachik Volcanoes of 1975. In: The Great Tolbachik fissure eruption; geological and geophysical data 1975-1976, (ed. Fedotov-SA and Markhinin-E.K.). Cambridge University Press, Cambridge (UK). Mercalli G. and Silvestri O. (1891) Le eruzioni dell'isola di Vulcano, incominciate il 3 Augusto 1888 e terminate il 22 Marzo 1880. Ann. Uff. Cent. Meteor. Geodin. 10 (4), 213. Mogk D. W. and Mathez E. A. (2000) Carbonaceous films in midcrustal rocks from the KTB borehole, Germany, as characterized by time-of-flight secondary ion mass spectrometry. Geochem. Geophy. Geosy. 1, 29 pp. Molina M. J. and Rowland F. S. (1974) Stratospheric sink for chlorofluoromethanes: chlorine atom catalyzed destruction of ozone. Nature 249 (5460), 810-814. Montzka S. A., Butler J. H., Elkins J. W., Thompson T. M., Clarke A. D., and Lock L. T. (1999) Present and future trends in the atmospheric burden of ozone-depleting halogens. Nature 398 (6729), 690-664. Moore R. B., Helz R. T., Dzurisin D., Eaton G. P., Koyanagi R. Y., Lipman P. W., Lockwood J. P., and Puniwai G. S. (1980) The 1977 eruption of Kilauea volcano, Hawaii. J. Volcanol. Geotherm. Res. 7 (3-4), 189- 210. Morency M., Mineau R., Zeller E., and Dreschhoff G. (1986) Are all fossil fuels really fossils? Oil Gas J. 84, 92- 95. Muenow D. W. (1973) High temperature mass-spectrometric gas-release studies of Hawaiian volcanic glass: Pele's Tears. Geochim. Cosmochim. Acta 37, 1551-1561. Mukhin L. (1974) Evolution of organic compounds in volcanic regions. Nature 251 (5470), 50-51. Mukhin L. M., Vasilev V. N., and Ponomarev V. V. (1974) Detection of hydrogen cyanide and its derivatives in regions of active volcanism. Dokl. Akad. Nauk SSSR 215 (5), 1253-1254. Mun T. K. and Kuziakina T. I. (1976) Proportion of microflora organic compounds in volcanic ashes. Dokl. Akad. Nauk SSSR 231 (1), 232-235. Mun T. K. and Rogova O. A. (1980) Microbic colonization of volcanic ashes. Izv. Akad. Nauk Biol. 2, 261-266. Myhre G., Myhre A., and Stordal F. (2001) Historical evolution of radiative forcing of climate. Atmos. Environ. 35, 2361-2373. Naughton J. J. (1973) Volcanic flame - source of fuel and relation to volcanic gas-lava equilibrium. Geochim. Cosmochim. Acta 37 (5), 1163-1169. Navale V., Harpold D., and Vertes A. (1998) Development and characterization of gas chromatographic columns for the analysis of prebiological molecules in Titan's atmosphere. Anal. Chem. 70 (4), 689-697. Navarro-Gonzalez R., Ramirez S. I., de la Rosa J. G., Coll P., and Raulin F. (2001) Production of hydrocarbons and nitriles by electrical processes in Titan's atmosphere. In: Space Life Sciences: Life in the Solar System: Prebiotic Chemistry, Chirality and Space Biology, Advances in Space Research Vol. 27, pp. 271-282, Noguchi K. (1935) Geochemical investigation of volcanoes in Japan, I. Studies on the gases and the spring waters of volcano Asama I. Nippon Kagaku Kaishi 56 (12), 1495-1510. North F. K. (1982) Review of Thomas Gold's deep-earth-gas hypothesis. Energ. Explor. Exploit. 1 (2), 105-110. Olaniran A. O., Babalola G. O., and Okoh A. I. (2001) Aerobic dehalogenation potentials of four bacterial species isolated from soil and sewage sludge. Chemosphere 45 (1), 45-50. Palais J. M. (1985) Tephra layers and ice chemistry in the Byrd station ice core, Antarctica. PhD, The Ohio State University, 545 p. Palmer A. S., van Ommen T. D., Curran M. A. J., Morgan V., Souney J. M., and Mayewski P. A. (2001) High- precision dating of volcanic events (A.D. 1301-1995) using ice cores from Law Dome, Antarctica. J. Geophys. Res.-Atmos. 106 (D22), 28089-28095. Pandey R., Rao B. K., Jena P., and Blanco M. A. (2001) Electronic structure and properties of transition metal- benzene complexes. J. Am. Chem. Soc. 123 (31), 7744.

References 138

Pavlov A. L. and Karpov G. A. (1972) Physicochemical features of recent ore formation in Uzon Caldera (Kamchatka). Dokl. Akad. Nauk SSSR 206 (3), 719-722. Peel D. A. (1989) Trace metals and organic compounds in ice cores. In: The environmental record in glaciers and ice sheets (Dahlem Konferenz, Berlin), (ed. H. Oeschger and C. C. Langway), pp. 207-223. John Wiley & Sons Ltd., Penkett S. A., Prosser N. J. D., Rasmussen R. A., and Khalil M. A. K. (1981) Atmospheric measurements of CF4 and other fluorocarbons containing the CF3 grouping. J. Geophys. Res.-Atmos. 86 (C6), 5172-5178. Pereira W. E. and Rostad C. E. (1983) Terpenoid marker compounds derived from biogenic precursors in volcanic ash from Mount St-Helens, Washington. Geochim. Cosmochim. Acta 47 (12), 2287-2291. Pereira W. E., Rostad C. E., and Taylor H. E. (1980) Mount St-Helens, Washington, 1980 volcanic eruption - characterization of organic compounds in ash samples. Geophys. Res. Lett. 7 (11), 953-954. Pereira W. E., Rostad C. E., Taylor H. E., and Klein J. M. (1982) Characterization of organic contaminants in environmental samples associated with Mount St-Helens 1980 volcanic eruption. Environ. Sci. Technol. 16 (7), 387-396. Perret F. A. and Brun A. (1915) Volcanic research at Kilauea in the summer of 1911 with a report by Dr. Albert Brun on the material taken directly from "Old Faithful". Am. J. Sci. 36, 475-488. Petersilie I. A. and Pavlova M. A. (1978) Organic compounds in volcanic and metamorphic rocks. Int. Geol. Rev. 20, 339-344. Petherbridge J. R., May P. W., Fuge G. M., Rosser K. N., and Ashfold M. N. R. (2002) In situ plasma diagnostics of the chemistry behind sulfur doping of CVD diamond films. Diam. Relat. Mat. 11, (in press). Petit J. R., Jouzel J., Raynaud D., Barkov N. I., Barnola J.-M., Basile I., Bender M., Chappellaz J., Davis M., Delaygue G., Delmotte M., Kotlyakov V. M., Legrand M., Lipenkov V. Y., Lorius C., Pépin L., Ritz C., Saltzman E., and Stievenard M. (1999) Climate and atmospheric history of the past 420'000 years from the Vostok ice core, Antarctica. Nature 399 (6735), 429-436. Petschek P. (1911) Studien über Gase. PhD, Technische Hochschule zu Dresden, p. Phelan J. M. (1983) Volcanoes as a source of volatile trace elements in the atmosphere (aircraft sampling, plume chemistry). PhD, University of Maryland College Park, 175 p. Pietrogrande M. C., Coll P., Sternberg R., Szopa C., Navarro-Gonzalez R., Vidal-Madjar C., and Dondi F. (2001) Analysis of complex mixtures recovered from space missions - Statistical approach to the study of Titan atmosphere analogues (tholins). J. Chromatogr. A 939 (1-2), 69-77. Podkletnov N. E. and Markhinin E. K. (1981) New data on abiogenic synthesis of prebiological compounds in volcanic processes. Orig. Life Evol. Biosph. 11 (4), 303-315. Ponomarev V. V., Burkova V. N., Serebrennikova O. V., and Titov V. I. (1979) Method of extraction of volcano-derived organic matter. Geokhimiya (5), 792-794. Poulsen T. G., Yamaguchi T., Moldrup P., de Jonge L. W., and Rolston D. E. (2000) Predicting volatile organic vapor sorption from soil specific surface area and texture. J. Environ. Qual. 29 (5), 1642-1649. Porfir'ev V. B. (1974) Inorganic origin of petroleum. AAPG Bull. 58 (1), 3-33. Porshnev N. V. and Bondarev V. B. (1989) Hydrocarbon of the C1-C11 range in high-temperature geotherms - quantitative variations and pathways of formation. Dokl. Akad. Nauk SSSR 306 (3), 708-712. Prather M. (1992) Catastrophic loss of stratospheric ozone in dense volcanic clouds. J. Geophys. Res. 97 (D9), 10187-10191. Preunkert S., Legrand M., and Wagenbach D. (2001) Causes of enhanced fluoride levels in Alpine ice cores over the last 75 years: Implications for the atmospheric fluoride budget. J. Geophys. Res.-Atmos. 106 (D12), 12619-12632. Price L. C. (1993) Thermal stability of hydrocarbons in nature - limits, evidence, characteristics, and possible controls. Geochim. Cosmochim. Acta 57 (14), 3261-3280. Price L. C. (1994) Petroleum geochemistry - metamorphic free-for-all. Nature 370 (6487), 253-254. Putschew A., Schaeffer-Reiss C., Schaeffer P., Koopmans M. P., De Leeuw J., Lewan M. D., Sinninghe Damsté J. S., and Maxwell J. R. (1998) Release of sulfur- and oxygen-bound components from a sulfur-rich kerogen during simulated maturation by hydrous pyrolysis. Org. Geochem. 29 (8), 1875-1890. Rampino M. R. and Self S. (1982) historic eruptions of Tambora (1815), Krakatau, and Agung (1963), their stratospheric aerosols, and climate impact. Quaternary Res. 18, 127-143. Ranneft T. S. M. (1979) Segmentation of island arcs and application to petroleum geology. J. Petrol. Geol. 1 (3), 35-53. Rasmussen R. A. and Khalil M. A. K. (1983) Atmospheric benzene and toluene. Geophys. Res. Lett. 10 (11), 1096-1099. Rasmussen R. A., Khalil M. A. K., Dalluge R. W., Penkett S. A., and Jones B. (1982) Carbonyl sulfide and carbon disulfide from the eruptions of Mount St. Helens. Science 215 (4533), 665-667.

References 139

Rasmussen R. A., Khalil M. A. K., Penkett S. A., and Casadevall T. J. (1981) Analysis of trace gases from Mt. St. Helens volcanic ash. EOS Transact. AGU 62 (8), 81. Rasmussen R. A. and Rasmussen L. E. (1980) Trace gases of volcanic origin. EOS Transact. AGU 61 (6), 67. Rasmussen R. A., Rasmussen L. E., Khalil M. A. K., and Dalluge R. W. (1980) Concentration distribution of methyl chloride in the atmosphere. J. Geophys Res.-Oceans Atmos. 85 (NC12), 7350-7356. Reeves C. E. and Penkett S. A. (1993) An estimate of the anthropogenic contribution to atmospheric methyl bromide. Geophys. Res. Lett. 20 (15), 1563-1566. Rhew R. C., Miller B. R., and Weiss R. F. (2000) Natural methyl bromide and methyl chloride emissions from coastal salt marshes. Nature 403, 292-295. Riche M. A. (1854) Recherches sur des combinaisons chlorées dérivées des sulfures de méthyle et d'éthyle. CR Acad. Sci. Paris XXXIX (19), 910-914. Rigaud P. and Leroy B. (1990) Presumptive evidence for a low value of the total ozone content above Antarctica in September, 1958. Ann. Geophys. 8 (11), 791-794. Robert J.-L., Beny J.-M., Della Ventura G., and Hardy M. (1993) Fluorine in micas: crystal chemical control of the OH-F distribution between trioctahedral and dioctahedral sites. Eur. J. Mineral. 5, 7-18. Robinson R., Sir. (1966) The origins of petroleum. Nature 212 (5066), 1291-1295. Robock A. (1981) The Mount St. Helens volcanic eruption of 18 May 1980: minimal climatic effect. Science 212 (4501), 1383-1384. Robock A. and Mass C. (1982) The Mount St. Helens volcanic eruption of 18 May 1980: large short-term surface temperature effects. Science 216 (4546), 628-630. Roscoe H. K. (2001) The risk of large volcanic eruptions and the impact of this risk on future ozone depletion. Nat. Hazards 23 (2-3), 231-246. Rose W. I., Jr., Cadle R. D., Heidt L. E., Friedman I., Lazrus A. L., and Huebert B. J. (1980) Gas and hydrogen isotopic analyses of volcanic eruption clouds in Guatemala sampled by aircraft. J. Volcanol. Geotherm. Res. 7, 1-10. Rowland F. S. (2002), pers.comm. In: Rowland F. S. and Molina M. J. (1975) Chlorofluoromethanes in environment. Rev. Geophys. 13 (1), 1-35. Rowland F. S. and Molina M. J. (1976) Estimated future atmospheric concentrations of CCl3F (fluorocarbon-11) for various hypothetical tropospheric removal rates. J. Phys. Chem. 80 (19), 2049-2056. Russel III J. M., Louo M., Cicerone R. J., and Deaver L. E. (1996) Sattelite confirmation of the dominance of chlorofluorocarbons in the global stratospheric chlorine budget. Nature 379 (6565), 526-529. Ryan J. A. and Mukherjee N. R. (1973) Estimate of HCl tropospheric cycle. EOS Transact. AGU 54 (11), 1098- 1098. Ryan J. A. and Mukherjee N. R. (1975) Sources of stratospheric gaseous chlorine. Rev. Geophys. 13 (5), 650- 658. Sainte-Claire Deville C. (1856a) 2eme lettre a M. Dumas sur quelques produits d'émenation de la Sicilie. CR Acad. Sci. Paris LXIII (7), 359-370. Sainte-Claire Deville C. (1856b) 9eme lettre a M. Elie de Beaumont sur les phénomènes éruptifs de l'Italie méridionale. CR Acad. Sci. Paris XLIII (14), 681-686. Sainte-Claire Deville C. (1856c) 10eme lettre a M. Elie de Beaumont sur les phénomènes éruptifs de l'Italie méridionale. CR Acad. Sci. Paris LXIII (16), 745-752. Sainte-Claire Deville C. (1865) Sur les émanations volcaniques des Champs Phlégréens (3eme lettre). CR Acad. Sci. Paris LXI (19), 760-764. Sakata S. (1991) Carbon isotope geochemistry of natural gases from the Green Tuff Basin, Japan. Geochim. Cosmochim. Acta 55 (5), 1395-1405. Salvi S. and Williams-Jones A. E. (1997) Fischer-Tropsch synthesis of hydrocarbons during sub-solidus alteration of the Strange Lake peralkaline granite, Quebec/Labrador, Canada. Geochim. Cosmochim. Acta 61 (1), 83-99. Sander R. (2001) Henry's law constants. In: NIST Chemistry WebBook, NIST Standard Reference Database Number 69, (ed. P. J. Linstrom and W. G. Mallard). National Institute of Standards and Technology, 20899 (http://webbook.nist.gov), Gaithersburg, MD (USA). Sanhueza E. (2001) Hydrochloric acid from chlorocarbons: a significant global source of background rain acidity. Tellus B 53 (2), 122-132. Sassen K. (1992) Evidence for liquid-phase cirrus cloud formation from volcanic aerosols: climatic implications. Science 257 (5069), 516-519. Satterfield c. N. and Huff G. A. (1982) Carbon number distribution of Fischer-Tropsch products formed on an iron catalyst in a slurry reactor. J. Catal. 73, 187-197. Schlepp L., Elie M., Landais P., and Romero M. A. (2001) Pyrolysis of asphalt in the presence and absence of water. Fuel Process. Technol. 74, 107-123.

References 140

Schneider S. H. and Mass C. (1975) Volcanic dust, sunspots, and temperature trends. Science 190 (4216), 741- 746. Schoell M. (1988) Multiple origins of methane in the Earth. Chem. Geol. 71, 1-10. Schönwiese C.-D. (1993) Treibhauseffekt und natürliche Klimaschwankungen. Atomwirtschaft 38 (7), 522-529. Schrön W. (1989) Solid-gas-equilibria in geo- and cosmochemistry: I. geochemistry. Eur. J. Mineral. 1, 739- 763. Schwandner F. M., Giże A. P., Hall K., Seward T. M., and Dietrich V. J. (2001a) Short Path Thermal Desorption - SPME-CT-GC/MS: analysis of trace organic compounds present in hot volcanic gases with high sulfur, water and acid matrix. Chimia 55 (7-8), 590. Schwandner F. M., Giże A. P., Hall P. A., Seward T. M., and Dietrich V. J. (2002) Diffusive emission of organic trace gases from volcanic flanks and craters. J. Geophys. Res.-Atmos., (submitted). Schwandner F. M., Giże A. P., Seward T. M., Hall K., and Dietrich V. J. (2000a) Natural halocarbon compounds in volcanic gases. Proc. Geochemistry of Crustal Fluids, EURESCO/ESF/INTAS Conference, Grenada (Spain); 91 (abstr.). Schwandner F. M., Giże A. P., Seward T. M., Hall K., and Dietrich V. J. (2000b) Natural halocarbon compounds in volcanic gases. J. Conf. Abs. 5 (2), 898. Schwandner F. M., Giże A. P., Seward T. M., Hall K., and Dietrich V. J. (2001b) Organic compounds in volcanic gases. Proc. EUG XI, Strasbourg (France); 255 (abstr.). Schwandner F. M., Gize A. P., Seward T. M., Hall K., and Dietrich V. J. (2002) Halocarbons and other organic compounds in volcanic gases from Vulcano (Aeolian Islands, Italy). In: The Geochemistry of Fluids and Ore Deposits: Rulers and Witnesses of Processes within the Earth - a Tribute to Werner F. Giggenbach, Vol. 8 (ed. S. F. Simmons and I. Graham), pp. submitted. Geochemical Society, St. Louis, MO (USA). Seewald J. S. (1994) Evidence for metastable equilibrium between hydrocarbons under hydrothermal conditions. Nature 370 (6487), 285-287. Seward T. M. (1974) Equilibrium and oxidation potential in geothermal waters at Broadlands, New Zealand. Am. J. Sci. 274 (2), 190-192. Seward T. M. (1991) The hydrothermal chemistry of gold. In: Gold metallogeny and exploration, (ed. R. P. Foster), pp. 432. Blackie and Son Ltd., London (UK). Seward T. M. and Barnes H. L. (1997) Metal transport in hydrothermal ore fluids. In: The Geochemistry of Hydrothermal Ore Deposits, (ed. H. L. Barnes), pp. 435-486. John Wiley & Sons Ltd., New York (USA). Shepherd E. S. (1919) The composition of the gases of Kilauea. Bull. Hawaiian Volcano Obs. 7 (7), 94-97. Shepherd E. S. (1920) Two gas collections from Mauna Loa. Bull. Hawaiian Volcano Obs. 8 (5), 63-64. Shepherd E. S. (1921) Kilauea Gases, 1919. Bull. Hawaiian Volcano Obs. 9 (5), 83-88. Shock E. L. and Schulte M. D. (1998) Organic synthesis during fluid mixing in hydrothermal systems. J. Geophys. Res.-Planets 103 (E12), 28513-28527. Simoneit B. R. T. (1993) Aqueous high-temperature and high-pressure organic geochemistry of hydrothermal vent systems. Geochim. Cosmochim. Acta 57, 3231-3243. Simoneit B. R. T. (1995) Evidence for organic synthesis in high-temperature aqueous media - facts and prognosis. Orig. Life Evol. Biosph. 25 (1-3), 119-140. Smith D. B., Zielinski R. A., Rose W. I., Jr., and Huebert B. J. (1982) Water-soluble material on aerosols collected within volcanic eruption clouds. J. Geophys. Res. 87 (C7), 4963-4972. Snyder G. T. and Fehn U. (2002) The origin of iodine in volcanic fluids: 129I results from the Central American Volcanic Arc. Geochim. Cosmochim. Acta 66, (in press). Sobolev V. S., Bazarova T. Y., and Kostyuk V. P. (1974) Inclusions in the minerals of some types of alkaline rocks. In: The Alkaline Rocks, (ed. H. Sorensen), pp. 389-401. John Wiley & Sons, London (UK). Solomon S. (1999) Stratospheric ozone depletion: a review of concepts and history. Rev. Geophys. 37 (3), 275- 316. Solomon S., Garcia R. R., and Ravishankara A. R. (1994) On the role of iodine in ozone depletion. J. Geophys. Res.-Atmos. 99 (D10), 20491-20499. Solomon S., Mills M., Heidt L. E., Pollock W. H., and Tuck A. F. (1992) On the evaluation of ozone depletion potentials. J. Geophys. Res.-Atmos. 97, 825-842. Solomon S., Sanders R. W., Garcia R. R., and Keys J. G. (1993) Increased chlorine dioxide over Antarctica caused by volcanic aerosols from Mount Pinatubo. Nature 363 (6426), 245-248. SORG. (1990) Stratospheric ozone 1990; Third report of the UK SORG (Stratospheric Ozone Review Group). Rep.No. 3, Department of the Environment, London (UK), London, UK. Staehelin J., Harris N. R. P., Appenzeller C., and Eberhard J. (2001) Ozone trends: A review. Rev. Geophys. 39 (2), 231-290.

References 141

Stoiber R. E., Leggett D. C., Jenkins T. F., Murrmann R. P., and Rose W. I. (1971) Organic Compounds in Volcanic Gas from Santiaguito Volcano, Guatemala. GSA Bull. 82 (8), 2299-2302. Stolarski R. S. and Cicerone R. J. (1974) Stratospheric chlorine: a possible sink for ozone. Can. J. Chemistry 52 (8), 1610-1615. Sturges W. T., McIntyre H. P., Penkett S. A., Chappellaz J., Barnola J.-M., Mulvaney R., Atlas E., and Stroud V. (2001) Methyl bromide, other brominated methanes, and methyl iodide in polar firn air. J. Geophys. Res.-Atmos. 106 (D2), 1595-1606. Sturges W. T., Penkett S. A., Barnola J.-M., and Chappellaz J. A. (1996) Fluorocarbon tracers of the age of air in alpine firn. In: Chemical exchange between the atmosphere and polar snow, NATO ASI Series Vol. 143 (ed. E. W. Wolff and R. C. Bales). Springer Verlag, Berlin. Sugisaki R. and Mimura K. (1994) Mantle hydrocarbons - abiotic or biotic. Geochim. Cosmochim. Acta 58 (11), 2527-2542. Symonds R. B. (1990) Application of multicomponent chemical equilibria to volcanic gases at Augustine Volcano, volcanic halogen emissions, and volcanological studies of gas-phase transport. Ph.D., Michigan Tech. (USA), p. Symonds R. B., Reed M. H., and Rose W. I. (1992) Origin, speciation, and fluxes of trace-element gases at Augustine Volcano, Alaska - insights into magma degassing and fumarolic processes. Geochim. Cosmochim. Acta 56 (2), 633-657. Symonds R. B., Rose W. I., Bluth G. J. S., and Gerlach T. M. (1994) Volcanic gas studies - methods, results, and applications. In: Volatiles in Magmas, REVIEWS IN MINERALOGY Vol. 30, pp. 1-66. Mineralogical Society of America, Washington, D.C. (USA). Symonds R. B., Rose W. I., Gerlach T. M., Briggs P. H., and Harmon R. S. (1990) Evaluation of gases, condensates, and SO2 emissions from Augustine Volcano, Alaska - the degassing of a Cl-rich volcanic system. Bull. Volcanol. 52 (5), 355-374. Symonds R. B., Rose W. I., and Reed M. H. (1988) Contribution of Cl-bearing and F-bearing gases to the atmosphere by volcanos. Nature 334 (6181), 415-418. Tabazadeh A., Drdla K., Schoeberl M. R., Hamill P., and Toon O. B. (2002) Arctic "ozone hole" in a cold volcanic stratosphere. P. Natl . Acad. Sci. USA 99 (5), 2609-2612. Tabazadeh A. and Turco R. P. (1993) Stratospheric chlorine injection by volcanic eruptions: HCl scavenging and implications for ozone. Science 260 (5111), 1082-1086. Tedesco D., Miele G., Sano Y., and Toutain J. P. (1995) Helium isotopic ratio in Vulcano island fumaroles: temporal variations in shallow level mixing and deep magmatic supply. J. Volcanol. Geotherm. Res. 64, 117-128. Tedesco D. and Nagao K. (1996) Radiogenic 4He, 21Ne and 40Ar in fumarolic gases on Vulcano: Implication for the presence of continental crust beneath the island. Earth Planet. Sci. Lett. 144 (3-4), 517-528. Tett S. F. B., Stott P. A., Allen M. R., Ingram W. J., and Mitchelll J. F. B. (1999) Causes of twentieth-century temperature change near the earth's surface. Nature 399 (6736), 569-572. Tie X. X. (1994) Potential impact on stratospheric ozone due to emission of hydrocarbons from high-altitude aircraft. Tellus B 46 (4), 286-293. Tie X. X. and Brasseur G. (1995) The response of stratospheric ozone to volcanic eruptions - sensitivity to atmospheric chlorine loading. Geophys. Res. Lett. 22 (22), 3035-3038. Tie X. X., Brasseur G. P., Briegleb B., and Granier C. (1994) Two-dimensional simulation of Pinatubo aerosol and its effect on stratospheric ozone. J. Geophys. Res.-Atmos. 99 (D 10), 20545-20562. Tilling R. I. (1982) Volcanic cloud may alter climate. Natl. Geogr. 162 (5), 672-675. Tingle T. N. and Hochella M. F. (1993) Formation of reduced carbonaceous matter in basalts and xenoliths - reaction of C-O-H gases on olivine crack surfaces. Geochim. Cosmochim. Acta 57 (14), 3245-3249. Tingle T. N., Hochella M. F., Becker C. H., and Malhotra R. (1990) Organic compounds on crack surfaces in olivine from San-Carlos, Arizona, and Hualalai volcano, Hawaii. Geochim. Cosmochim. Acta 54 (2), 477-485. Tingle T. N., Mathez E. A., and Hochella M. F. (1991) Carbonaceous matter in peridotites and basalts studied by XPS, SALI, and LEED. Geochim. Cosmochim. Acta 55 (5), 1345-1352. Udisti R., Barbante C., Castellano E., Vermigli S., Traversi R., Capodaglio G., and Piccardi G. (1999) Chemical characterisation of a volcanic event (about AD 1500) at Styx Glacier plateau, northern Victoria Land, Antarctica. In: Annals of Glaciology, Vol 29, 1999, ANNALS OF GLACIOLOGY Vol. 29, pp. 113- 120. Int Glaciological Soc, Cambridge (UK). UNEP. (2000) The Montreal Protocol on substances that deplete the ozone layer, including amendments London 1990, Copenhagen 1992, Vienna 1995, Montreal 1997, Beijing 1999 Vol., Ozone Secretariat, UNEP (United Nations Environment Programme), Nairobi (Kenya), p. Varekamp J. C., Kreulen R., Poorter R. P. E., and Van Bergen M. J. (1992) Carbon sources in arc volcanism, with impolications for the carbon cycle. Terra Nova 4, 363-373.

References 142

Varner R. K., Crill P. M., and Talbot R. W. (1999) Wetlands: a potentially sigificant source of atmospheric methyl bromide and methyl chloride. Geophys. Res. Lett. 26 (16), 2433-2435. Vogt H.-H. (1972) Vulkanausbrüche wirken abkühlend [volcanic eruptions result in cooling]. Kosmos 68 (Suppl.), 109. Voitov G. I. (2000a) Geochemical and carbon isotopic instabilities of spontaneous gases in seismically active regions. Geochem. Int. 38 (11), 1084-1106. Voitov G. I. (2000b) Variability of methane fluxes during cold earth degassing. Geochem. Int. 38 (3), 274-280. Voitov G. I., Milkes M. R., Kucher M. I., Belikov V. M., and Krivomazova N. G. (1986) The anomalous variations of CH4 and CO2 carbon isotopes in free gases of the seismoactive region of Prekopet-Dag depression. Dokl. Akad. Nauk SSSR 290 (6), 1335-1338. Voitov G. I. and Yunusov S. S. (1996) Isotope-carbon precursor of strong tectonic earthquakes. Dokl. Akad. Nauk 346 (3), 392-395. von Humboldt A. (1803) Diaries: Rio Magdalena. Vupputuri R. K. R. (1985) The effect of ozone photochemistry on atmospheric and surface temperature changes due to increased CO2, N2O, CH4 and volcanic aerosol in the atmosphere. Atmos. Ocean 23 (4), 359-374. Vupputuri R. K. R. (1992) The Tambora eruption in 1815 provides a test on possible global climatic and chemical perturbations in the past. Nat. Hazards 5, 1-16. Vupputuri R. K. R., Higuchi K., and Hengeveld H. G. (1995) A 1-D modelling of climatic and chemical effects of greenhouse gases. Theor. Appl. Climatol. 52 (3-4), 151-167. Wahrenberger C. (1997) Some Aspects of the Chemistry of Volcanic Gases. Ph.D., ETH Zürich, DISS- ETH.No.12252, 233 p. Wahrenberger C., Dietrich V., Seward T. M., Dolezal I. S., Farrenkothen I., and Mutter K. (1996a) CFC's; natural background emission by volcanoes? EOS Transact. AGU 77 (46), F804. Wahrenberger C., Eastoe C. J., Seward T. M., and Dietrich V. (1996b) Stable chlorine isotopic composition of volcanic gas condensates. Proc. 7th annual V.M. Goldschmidt conference, Tucson, Arizona (abstr.). Wahrenberger C., Seward T. M., and Dietrich V. (1998) Halocarbon compounds in high temperature volcanic gases. Proc. IAVCEI meeting, Cape Town (South Africa) (abstr.). Wahrenberger C., Seward T. M., and Dietrich V. (2002) Volatile trace element transport in high temperature gases from Kudriavy volcano (Iturup, Kurile Islands, Russia). In: Water-rock Interaction, Ore Deposits, and Environmental Geochemisty: A tribute to David A. Crerar, Geochemical Society Special Publications Vol. 7 (ed. R. Hellmann and S. A. Wood), pp. 307-327., Walker S. J., Weiss R. F., and Salameh P. K. (2000) Reconstructed histories of the annual mean atmospheric mole fractions for the halocarbons CFC-11, CFC-12, CFC-113 and carbon tetrachloride. J. Geophys Res.-Oceans Atmos. 105 (C6), 14285-14296. Wallace L. and Livingston W. (1991) Spectroscopic observations of atmospheric trace gases over Kitt Peak . 3. long-term trends of hydrogen chloride and hydrogen fluoride from 1978 to 1990. J. Geophys. Res.- Atmos. 96 (D8), 15513-15521. Wallace L. and Livingston W. (1992) The effect of the Pinatubo cloud on hydrogen chloride and hydrogen fluoride. Geophys. Res. Lett. 19 (12), 1209-1211. Wamsley P. R., Elkins J. W., Fahey D. W., Dutton G. S., Volk C. M., Myers R. C., Montzka S. A., Butler J. H., Clarke A. D., Fraser P. J., Steele L. P., Lucarelli M. P., Atlas E. L., Schauffler S. M., Blake D. R., Rowland F. S., Sturges W. T., Lee J. M., Penkett S. A., Engel A., Stimpfle R. M., Chan K. R., Weisenstein D. K., Ko M. K. W., and Salawitch R. J. (1998) Distribution of Halon-1211 in the upper troposphere and lower stratosphere and the 1994 total bromine budget. J. Geophys. Res. 103 (D1), 1513-1526. Wardell L. J., Kyle P. R., and Chaffin C. (1999) CO2 emissions from Mount Erebus volcano, Antarctica. Proc. 8th International Symposium on Antarctic Earth Science, Wellington, New Zealand (abstr.). Webster C. R., May R. D., Toohey D. W., Avallone L. M., Anderson J. G., and Solomon S. (1993) In-situ measurements of the ClO/HCl ratio - heterogenous processing on sulfate aerosols and polar stratospheric clouds. Geophys. Res. Lett. 20 (22), 2523-2526. Weigert F. J. and Davis R. F. (1993) Synthesis and aromatization of 2+2 cycloadducts of butadienes and tetrafluoroethylene. J. Fluorine Chem. 63, 69-84. Welhan J. A. (1988) Origins of methane in hydrothermal systems. Chem. Geol. 71, 183-198. Welhan J. A. and Craig H. (1982) Abiogenic methane in mid-ocean ridge hydrothermal fluids. In Deep source gas workshop, Vol. (US Dept of Energy) DOE-METC-82-50 (ed. W. J. Gwilliam), pp. 122-128. Welhan J. A. and Lupton J. E. (1987) Light hydrocarbon gases in Guaymas Basin hydrothermal fluids: thermogenic versus abiogenic origin. AAPG Bull. 71 (2), 215-223. Wilkniss P. E., Swinnerton J. W., Lamontagne R. A., and Bressan D. J. (1975) Trichlorofluoromethane in troposphere, distribution and increase, 1971 to 1974. Science 187 (4179), 832-834.

References 143

Wilson N. S. F. and Zentilli M. (1999) The role of organic matter in the genesis of the El Soldado volcanic- hosted manto-type Cu deposit, Chile. Econ. Geol. Bull. Soc. Econ. Geol. 94 (7), 1115-1135. Winterton N. (2000) Chlorine: the only green element - towards a wider acceptance of its role in natural cycles. Green Chem. 2 (5), 173-225. WMO. (1990) Scientific assessment of stratospheric ozone. Rep.No. 1989, WMO report No. 20 (2 vols.), World Meteorological Organisation (WMO), Global Ozone Research and Monitoring Project, Geneva (CH). WMO. (1994) Scientific assessment of ozone depletion. Rep.No. WMO Rep. No. 37, World Meteorological Organisation (WMO), Global Ozone Research and Monitoring Project, Geneva (CH). WMO. (1998) Scientific assessment of ozone depletion. Rep.No. WMO Rep. No. 44, World Meteorological Organisation (WMO), Global Ozone Research and Monitoring Project, Geneva (CH). Wuebbles D. J. (1987) Natural and anthropogenic perturbations to the stratosphere. Rev. Geophys. 25 (3), 487- 493. Yasuhara A. (1993) Thermal decomposition of tetrachloroethylene. Chemosphere 26 (8), 1507-1512. Yu Z. S., Warner J. A., Dahlgren R. A., and Casey W. H. (1996) Reactivity of iodide in volcanic soils and noncrystalline soil constituents. Geochim. Cosmochim. Acta 60 (24), 4945-4956. Yurganov L. N., Grechko E. I., and Dzhola A. V. (2000) An estimate of changes in the tropospheric hydroxyl concentration with long-term total column CO measurements. Proc. 2nd International Symposium on non-CO2 greenhouse gases (NCGG-2); Scientific Understanding, Control and Implementation, Noordwijkerhout, NL; 660p. (abstr.). Zanon V., Frezzotti M. L., and Angelo P. (2001) Melt and fluid inclusions study on quartz xenoliths from the Aeolian Island Arc (Southern Italy): evidence for crustal contamination. Proc. EUG XI, Strasbourg (FR); 816 (abstr.). Zetzsch C. (1980) Gepulste Vakuum-UV-Photolyse mit Resonanzfluoreszenznachweis von OH als Testmethode troposphärischen Reaktionsverhaltens von Chemikalien. Proc. Internationale Arbeitstagung über Prüfungsmethoden und Bewertungsverfahren zur Bestimmung des photochemischen Abbauverhaltens von chemischen Stoffen, Berlin (abstr.). Zhao H., Chen J., Quan X., Yang F., and Peijnenburg W. J. G. M. (2001) Quantitative structure-property relationship study on reductive dehalognation of selected halogenated aliphatic hydrocarbons in sediment slurries. Chemosphere 44 (7), 1557-1563. Zielinski G. A. (1995) Stratospheric loading and optical depth estimates of explosive volcanism over the last 2100 years derived from the Greenland Ice Sheet Project 2 ice core. J. Geophys. Res.-Atmos. 100 (D10), 20937-20955. Zielinski G. A., Mayewski P. A., Meeker L. D., Whitlow S., and Twickler M. S. (1996) A 110'000-yr record of explosive volcanism from the GISP2 (Greenland) ice core. Quaternary Res. 45 (2), 109-118. Zielinski G. A., Mayewski P. A., Meeker L. D., Whitlow S., Twickler M. S., Morrison M., Meese D., Alley R. B., and Gow A. J. (1994) Record of volcanism since 7000 B. C. from the GISP2 Greenland ice core and implications for the volcano-climate system. Science 264 (5161), 948-952. Zolotov M. Y. and Fegley B. (1999) Oxidation state of volcanic gases and the interior of Io. Icarus 141 (1), 40- 52. Zolotov M. Y. and Shock E. L. (2000) A thermodynamic assessment of the potential synthesis of condensed hydrocarbons during cooling and dilution of volcanic gases. J. Geophys. Res.-Solid Earth 105 (B1), 539-559. Zreda-Gostynska G., Kyle P., and Finnegan D. (1993) Chlorine, fluorine, and sulfur emissions from Mount Erebus, Antarctica and estimated contributions to the Antarctic atmosphere. Geophys. Res. Lett. 20 (18), 1959-1962. Zreda-Gostynska G., Kyle P., Finnegan D., and Prestbo K. M. (1997) Volcanic gas emissions from Mount Erebus and their impact on the Antarctic environment. J. Geophys. Res.-Solid Earth 102 (B7), 15039- 15055. Zubkov V. S. (2001) Composition and speciation of fluid in the system C-H-N-O-S at P-T conditions of the upper mantle. Geochem. Int. 39 (2), 109-122.

References 144 Appendices

A.1.1

A1. A historic overview of organic chemical volcanic gas studies

Abstract

Volcanic gas studies during the last one hundred years have been carried out by scientists working in volcanic surveillance, ore deposit exploration, geothermal energy exploration and exploitation, and pure scientific research. Prior to that time, volcanic gas research was deeply interwoven with natural gas exploration. In this chapter, we a) compile the early volcanic gas research in so far as it is concerned with organic chemistry, from the earliest attempts of A. von Humboldt, Sir H. Davy, Gay-Laussac, R. W. Bunsen, and F. Fouqué to A. Day and S. Shepherd, and b) critically review the post-war research efforts to elucidate the sources and composition of organic compounds in high-temperature volcanic gases. Historical data are used in modern theoretical, deterministic and experimental studies of volcanic surveillance, exploration, exobiology, medical sciences, atmospheric chemistry, and in hydrothermal studies concerned with the question of the origin of life on earth.

1. INTRODUCTION

UNDERSTANDING the occurrence, abiogenic formation and emission of organic compounds in volcanic gases has major implications for the interpretation of a great variety of natural phenomena, such as formation and alteration of ore deposits (Gize, 1999), availability and stability of organic ligands involved in metal transport in ore-forming hydrothermal fluids (Seward and Barnes, 1997), atmospheric organic chemistry and the “ozone hole” (Molina and Rowland, 1974), and theories of a hydrothermal origin of life (e.g. Zolotov and Shock, 2000). Irrespective of the ultimate source of the atomic constituents (i.e., C, H, N, O, and S), the conditions for inorganic abiogenic formation of organic molecules are potentially met in any active volcanic-hydrothermal system. The origin of the carbon in hydrothermal systems is usually explained either by mantle degassing (primordial source, e.g. Mathez et al., 1984) or pyrolysis of subsurface organic matter (thermogenic source, Schoell, 1988).

Appendix I A.1.2

These natural volcanic “synthesis laboratories” not only produce a great variety of inorganic solid phases and solutes, but also offer suitable conditions for organic synthesis. Nearly the whole periodic table of elements is abundant in appreciable amounts; CO2 and water are present as liquid, vapor, gas or supercritical fluid. Furthermore, a high availability of reactive and catalytic surfaces, extreme temperature gradients and variable redox potentials potentially allow many simple reactions such as the Fischer-Tropsch synthesis to occur. Consequently, understanding abiogenic organic synthesis from simple inorganic carbon compounds (CO2,

CO, CH4, CS2, COS) together with H2, H2O, mineral acids, NH4 and S compounds under hydrothermal conditions, requires more than theoretical considerations, but first and foremost, solid analytical data from active geological sites. Since the earliest days of volcanic gas studies, simple organic compounds have been detected in volcanic gases. For historic interest, a summary of the earliest work is given first followed by an overview of the previous relevant research.

2. HISTORIC STUDIES

Perhaps the most interesting historical attempt to sample volcanic gases was by Alexander von Humboldt, who climbed Jorullo volcano in Mexico on September 19th, 1803 together with two native guides and two French gentlemen. He detected several gas species, as such water vapor, hydrogen sulfide, sulfuric acid and carbon dioxide. He also describes taking “air samples”, inside the crater, but to our knowledge, no analyses are available (von Humboldt, 1803). He had experimented already on air sampling on the Pico de Tenerife (Teide, Tenerife) on June 19-25, 1799, before departing on his ‘American Journey’ (de Terra, 1958). For his field analyses, he used a portable field analytical kit designed in 1783 by Goyton de Morveau (McCosh, 1984; Smeaton, 1966 p.47). Together with his college friend Leopold von Buch, and his roommate, friend and scientific collaborator on gas analyses at that time, Gay-Lussac (e.g., Gay-Lussac and von Humboldt, 1805), he climbed Vesuvius in the summer of 1805 to investigate its lava flows and gas emanations. Gay-Lussac later investigated volcanic gases on Vesuvius, during an eruption (McCosh, 1984) but did not report much on his studies (Gay- Lussac, 1823). One of his renowned results, however, is the important observation that the composition of fumarolic gases change with temperature. von Humboldt and Gay-Lussac believed in the origin of petroleum as a result of impregnation of marine sediments by sub- aqueous volcanic hot springs (e.g., Laughrey, 1998).

Appendix I A.1.3

Sir Humphry Davy had since 1812 speculated on the role of air and water on volcanic activity (Davy, 1828). He examined the few days old eruption products of Vesuvius in December 1819-February 1820, trying to catch escaping vapors from the lava flows in bottles either empty, or filled with sand or water. On the same occasion, he performed combustion experiments by throwing various chemicals onto the lava. His observations led him to conclude that pyrolysis of organic matter or combustion of coal is not the heat source of the volcanic “fire”, and that there was no indication of carbonaceous matter in the volcanic “smoke”. In 1831, Bossingault, who had attended Gay-Lussac’s lectures, experimented on fumarolic vapors on Purace (Colombia), also collecting samples for later analysis (McCosh, 1984). Together with von Buch, Elie de Beaumont climbed Etna in 1834 to perform similar studies on fumarolic gas chemistry and speculated on the origin of flames emanating from lava flows (de Beaumont, 1836). His early research on volcanic gases led to a milestone publication on volcanic gases (de Beaumont, 1847), in which he ascribes the origin of CO2 in the atmosphere to volcanic outgassing (p. 1293).

The first accessible and still reliable literature mention of the association of organic matter with volcanic gases comes from the German chemist Robert Wilhelm Bunsen, who investigated the gaseous syn-eruptive emissions of Vesuvius in 1841 and sampled the Hekla (Iceland) fumaroles, geysers and thermal springs in July 1846, just months after an eruption.

He speculated on the pyrolitic origin (“dry distillation”) of the nitrogen and NH3 in volcanic gases emanating from lava flows that were emplaced over vegetation (Bunsen, 1847, p.9) and estimated the ammonium chloride yield from pyrolysis to be 23g per m2 of grassland overrun by lava (Bunsen, 1852). Little of Bunsen’s volcanic gas analyses are accessible today, except an analysis of natural-gas hydrocarbons in mud volcano gas from Titarowka (Taman peninsula) which is tabulated in a later compilation (Tschirwinsky, 1915), recording identification of unsaturated aliphatic hydrocarbons and ethane. In those days, the origin of the gas species in natural emissions was not known well, but a frequent matter of discussion. Nitrogen was often seen as evidence of quantitative air entrainment and recycling.

Since the “gasometric methods” (Bunsen, 1857) of the time were mainly targeted to the early petroleum gas industry, hydrocarbon analyses were standard procedure in any “natural” gas analyses, including volcanic and hydrothermal gases. Bunsen’s landmark book

Appendix I A.1.4

“Gasometrische Methoden” (Bunsen, 1857) extensively describes the difficulties and techniques for sampling volcanic gases, and the analytical methods for qualitative and quantitative determination of the gas composition. His methods were later used in the petroleum industry as standard procedures. The methods of sampling were chiefly the replacement of a liquid (mercury) by the sample gas and flow-through sampling in arrays of glass containers, aided by a hand-pump if necessary, that could be clamped off or torch- sealed. His analytical methods comprise volumetric, gravimetric, and titrimetric analyses after separations aided by selective absorption, diffusion, dissolution and combustion of individual or groups of compounds. Following the discovery of oil in the Drake well in Pennsylvania (USA) in 1859, the “black gold” rush also swept over to Europe, inspiring increased exploration activity for natural gas seeps. Such work was performed by the same scientists that also worked on volcanic gases (Vélain, 1884), often sampling volcanic gases within the framework of a natural gas exploration campaign and vice versa, using Bunsen’s methods (Fouqué and Gorceix, 1869). Unfortunately, little is preserved in the literature until newer research was documented at the beginning of the twentieth century, when the principles of Bunsen’s methods were still commonly applied.

Bunsen had several students who profited from his gas sampling and analytical techniques developed for his visit to Hekla, among them the German Alfons Stübel (1859/60). After the naturalists Alexander von Humboldt, Charles Lyell and Leopold von Buch, the pioneering volcanological community in the mid-1800’s (in terms of gas chemistry) was dominated initially by C. Daubeny (1795-1867), Elie de Beaumont (1798-1874) and J. B. Boussingault (1802-1887), and later by R. W. Bunsen (1811-1899) and Beaumont’s student and successor, Ch. Sainte-Claire Deville (1814-1876), as well as certain Italian volcanologists, mostly studying Vesuvius and Etna. There is a direct lineage of student-mentor relationships as well as intellectual influence and challenge recognizable, that proceeds from von Humboldt and Gay-Laussac through the above mentioned generation, to Deville and Bunsen, to Bunsen’s student A. Stübel (1835-1904), to Deville’s student Fouqué (1828-1904) and Fouqué’s student C.-H. Gorceix (1842-1819), to Fouqué’s student and son-in-law A. Lacroix (1863- 1949). Lacroix was challenged by and himself critical of A. L. Day (1869-1960) and E. S. Shepherd (1879-1949) in the early 1900’s, who are regarded as the originators of modern volcanic gas sampling and analytical techniques (e.g. Gerlach, 1980).

Appendix I A.1.5

New eruptions occurred in Italy in the 1850’s - 1860’s, and with the new techniques available, the young science of the study of volcanic emanations accelerated. Sainte-Claire Deville introduced simple hydrocarbon analytical techniques developed for Italian natural gas seeps and the methane emanations of Solfatara de Pozzuoli (Sainte-Claire Deville, 1856c, 1856a). This was follwed by the discovery of “hydrocarbons” in volcanic gases of the Eolian Islands in Italy (Sainte-Claire Deville, 1856b) and on Etna (Fouqué, 1865a). Deville and his student Ferdinand Fouqué then discovered alkenes in the gases emanating at the 1861 lava flows at Torre del Greco, Santa-Lucia and other locations at the foot of Vesuvius (Fouqué, 1865b; Sainte-Claire Deville, 1865). They then started to study the gases emanating from the actual active craters. Sainte-Claire Deville later described the alkenes from Torre del Greco in more detail, noting a change in gas composition from 1862 to 1865 from hydrogen-containing to hydrogen-free with “a small quantity of a gas more enriched in carbon than marsh gas, an alkeneic gas” (Sainte-Claire Deville, 1865). Fouqué confirmed Deville’s theory, stating finally, that hydrocarbons are a common product of volcanic activity (Fouqué, 1866b) and summarized his findings on Vesuvius in 1861, and Etna in 1865 (Fouqué, 1866a). The Italian Silvestri described the emission of hydrocarbons in reducing gases, with a notable petroleum odor, after an earthquake preceding an eruption of Etna in 1866, at the Salse de Paterno (Silvestri, 1866, 1869).

With Bunsen’s and Deville’s former students, a new generation received their chance when the island of Santorini (Greece) showed signs of a beginning eruptive phase that climaxed with the emergence of two new islets in the dome area in 1866. This led to probably the best volcanological work of its times in terms of gas chemistry by Fouqué (1879), who in his gas analyses from Santorini also clearly found and documented ethene in several samples collected between 1866 and 1875. Two young German volcanologists and geographers Wilhelm Reiss and Alfons Stübel (a student of Bunsen) arrived on Santorini in 1866 first, during the time of the new eruptive phase, and stayed simultaneously with Fouqué, M. Gorceix, and Ch. Sainte-Claire Deville. Gorceix (1872a) had already reported on alkenes in gases from the Phlegrean fields the same year. Fouqué (1872) studied the freshly erupted lava in detail and occupied himself with early excavations at the Minoan settlement of Akrotiri on the island, while his student Gorceix (1872b, c) continued to study the gases. Fouqué (1879) returned to Santorini in 1867 and 1875 to collect more gases.

Appendix I A.1.6

During 1866, the Germans departed early from Santorin while the other scientists remained. News of imminent war had reached the Germans, making it impossible to complete their investigations, and they sent their gas sampling equipment home from Santorini (Reiss and Stübel, 1868). While staying in Athens during their return (June 12-20, 1866), they had a short time to investigate the nearby Greek islands of Aegina and Methana, as well as the fumaroles of Sousaki near Corinth (Reiss and Stübel, 1867). In retrospect, it is unfortunate that they were unable to sample the fumaroles of Sousaki, which may have been promising for analyzing hydrocarbons (see below), since excess hydrogen is abundant and they are closely associated with serpentinites; an association discussed in some detail by Fouqué, and Reiss and Stübel (1867). The quite different Greek volcanic islands of Aegina and Methana had been visited just months before them by Fouqué on his way to Santorini (and shortly after by Saint-Claire Deville), who had first identified the site of the historic Kameni eruption on Methana (Strabon, 7 b.c.). Since they already had personal information about this from Fouqué, they must have had contact while being on Santorini, where Fouqué (1879, p.232) also tested some improvements he had made to Bunsen’s gas sampling techniques. In January 1868, Reiss and Stübel departed on a journey to Hawaii which they never reached, because what was originally planned as a short visit to South America, turned into probably the most significant and thorough scientific expedition through South and Central America since that of Alexander von Humboldt, and which lasted until August 1877, almost 10 years (Rast, 1985).

These historical facts are significant to the early work on organic compounds in volcanic gases because Reiss and Stübel not only investigated, sampled and mapped in great detail more Andean volcanoes than von Humboldt, but they also brought back a large number of torch-sealed glass tube volcanic gas samples. Through unknown circumstances, it was many decades until they were finally analyzed under the supervision of Walter Hempel by Paul Petschek in Dresden, Germany (Petschek, 1911; Hempel, 1915). Several of the samples yielded simple hydrocarbons, but no use was made of the published data. The mud volcanoes of Turbaco near Cartagena (Columbia), sampled March 20, 1868 (Dietzel, 1921, p. 34 & 36) contained up to 7.8 % (of dry gas) ethane in two samples; those of Volcan del Tigre 1.8 %. In the same publication, Hempel assumed hydrocarbons to be present in lava (Hempel, 1915, p.165). The growing interest in fossil fuel as a result of the mass production of the automobile and war efforts encouraged increased exploration for natural gas and more analyses of mud

Appendix I A.1.7

volcanoes thus appeared in the literature (e.g. Charitschkoff, 1911; Tschirwinsky, 1915). In 1908, Rollin Thomas Chamberlin published an important book, “The gases in rocks”, where he ascribes hydrocarbon formation to a metal carbide hydration reaction to form ethylene, and to the Brodie reaction (Fischer-Tropsch reaction) to form alkanes (Chamberlin, 1908).

In a series of volcanic gas studies on Hawaii that started in May 1912 (Day and Shepherd, 1913b, 1913a; Gautier, 1913; Lacroix, 1913; Brun, 1914; Day and Shepherd, 1915; Perret and Brun, 1915; Shepherd, 1919, 1920, 1921), Shepherd analyzed vacuum tube samples of the Kilauea lava lake and Mauna Loa fumaroles, also analyzing for hydrocarbons, but with only little success (Day and Shepherd, 1913b; Shepherd, 1919, p.94 & 960; Shepherd, 1920, p.66; Shepherd, 1921, p.88). The 1919 samples from Mauna Loa (Shepherd, 1920, p.66), collected by T.A. Jagger, Jr., were taken from vents, where “an odor like pistachio” had been repeatedly recognized. The typical pistachio odor is today known to be of a complex organic compound mixture (Palazoglu and Balaban, 1998), and inorganic analogues are yet unidentified (e.g., organosulfurous or arsenosulfurous compounds may be likely candidates). The Swiss pharmacist Brun analyzed Kilauea samples taken by Perret and Shepherd, and performed a fusion experiment to remove remaining volatiles. His analyses include the finding of “very plentiful of bitumen” (Perret and Brun, 1915). The analyses from Hawaii led to three important publications of their time, one very detailed on the analytical methods (Shepherd, 1925), a summary and conclusion on the long years of gas research on Hawaii (Shepherd, 1938), and a general work on magmatic gases by Jaggar (Jaggar, 1940). The original Day, Shepherd & Jaggar analyses were much later restored and re-evaluated by Gerlach (1980) and interestingly, they appeared, at least theoretically, to contain hydrocarbons.

In 1919, Allen and Zies (1923) studied the fumaroles in the Katmai area (Valley of Ten Thousand Smokes, Alaska). They noticed an alkane-like odour that suggests light alkanes emanating from the fumaroles but did not find any hydrocarbons in their analyses even though they specifically looked for them (Allen and Zies, 1923), p.127). However, in their sublimate samples, “a heavy oil resembling crude petroleum was (…) distilled from it”. In Italy in 1923, the Istituto Chimico Farmaceutico of the University of Napoli, together with the Volcanological Institute Friedländer, also analyzed the solfataric gases of the Solfatara di Pozzuoli and the easterly Agnano complex for hydrocarbons but did not find any (Olita, 1924;

Appendix I A.1.8

Salvatore, 1924). Later, Allen and Day (1927, p. 70) found trace ethane in the hydrothermal well gas analyses of The Geysers (California).

The last pre-war report on hydrocarbons in volcanic gases was an extensive study on the gases of Asama volcano in Japan by Noguchi (1935), published in Japanese and later referenced by the volcanic gas review of White and Waring (1963). Noguchi reported on “higher hydrocarbons” (consisting mainly of ethane) up to 3.15 % (dry mass). He explained the formation of trace ethylene via a reaction pathway of two methane molecules under release of hydrogen gas. No other reports of volcanic gas analyses followed after this publication until the early 1970’s (Chaigneau et al., 1970). The two most complete reviews of volcanic gas analyses prior are by Allen (1922) and White and Waring (1963), both of which deserve due mention even though they are incomplete, especially in terms of organic compounds, as summarized above. The 1920’s and 1930’s were a time without great explosive eruptions and because of the world political situation in that time, no significant studies were published.

Appendix I A.1.9

3. ENVIRONMENTS OF OCCURRENCES

Considering the question of the possible natural abiogenic organic synthesis and the occurrence of natural geogenic organic compounds reported in the literature, it is most useful to distinguish the environments of extraterrestrial occurrences, as well as low-temperature hydrothermal fluids from high-temperature volcanic gases, and experimental and theoretical studies simulating simple and natural systems.

3.1. Extraterrestrial organic compounds

Since the early organic chemical research on meteorites by Berzelius (1834), the growing interest in exo- and astrobiology in extraterrestrial studies has shown ample theoretical and analytical evidence of the existence and abiotic formation of extraterrestrial organic matter in interstellar space, the solar nebula and planetary volcanism. This has been through thermodynamic models (e.g., Fegley, 1991; Fegley and Zolotov, 2000; Zolotov and Shock, 2000a, 2001), remote earth- and space-born spectroscopic observation (e.g. Vuitton et al., 2001) and the study of meteorites (especially the Tagish Lake and Murchison meteorites; for an introduction, see Pizzarello et al., 2001). The evolution and modification of the early Earth’s atmosphere has been explained by various processes, including volcanic outgassing and meteoritic imports. In the case of carbon compounds, carbonaceous chondrites are considered significant (e.g., Hunten, 1993; Kasting, 1993).

3.2. Terrestrial abiogenic organic occurrences

3.2.1. Magmatic fluid inclusions

Magmatic and hydrothermal vapour-rich fluid inclusions contain hydrocarbons. This simple statement is supported by the vast literature on the subject, a review of which is beyond the scope of this chapter. For introductions, see the special issue of Geochimica et Cosmochimica Acta (Vol. 5, January 1992), Roedder (1984), and Salvi and Williams-Jones (1997). Gases trapped in late hydrothermal inclusions in fluorite (“Stinkspat”) from Wölsendorf in the

Oberpfalz region (Bayern, Germany), were shown to contain C1-C3 alkanes and alkenes,

Appendix I A.1.10

alkylated benzenes, thionylfluorides (SOFx), sulfurylfluorides (SO2Fx), a fluoroborane (~BF3) compound, CH[1+(4-x)]Fx, and amines (Kranz, 1966). Harnisch et al. (1996), Harnisch and Eisenhauer (1998), and Harnisch et al. (2000) re-evaluated and continued the investigation of Kranz (1966). In a study of 26 fluorite samples from a variety of source rock materials, they found that SF6 and CF4 are present in accessory fluorite in granites to a much higher degree than in hydrothermal fluorite gangue, and concluded that Earth degassing is sufficient to account for the observed natural background concentration of these two fluorinated compounds. On the basis of their own and cited work, they state that emissions from volcanoes are only a very minor contribution (Harnisch and Eisenhauer, 1998). In addition, they find CF3Cl (CFC-13), CHF3 (R-23), and NF3 at 0.8, 3.7 and 600 ppbm, respectively, in “Stinkspat” (purple fluorite) from the same locality in Wölsendorf as analyzed by Kranz (1966). In addition, they reported analyses of 33 natural rock samples yielding 1.8 – 61000 pptm of CF4, 0.9 – 1200 pptm for CFC-12, 0.4 – 460 ppt for CFC-11, and 0.6 – 30000 pptm for SF6. By lithology, the Wölsendorf fluorite yields the highest contents, and the 13 volcanic rocks the lowest.

3.2.2. Igneous complexes

A number of accounts of hydrocarbon-rich fluid inclusions in alkaline igneous intrusions have been reported (Wahler, 1956; Petersilie, 1958,1964; Goguel, 1963; Gorstka et al., 1965; Petersilie et al., 1965; Galimov and Petersilie, 1967, 1968; Ikorskii, 1967; Petersilie et al., 1967; Petersilie and Sorensen, 1970; Petersilie and Pavlova, 1978; Salvi and Williams-Jones, 1997). For a review, see Sobolev et al. (1974). In addition, the potential of hydrocarbon formation within an exsolving gas phase of an alkaline melt has been thermodynamically modeled (Gerlach, 1980) using gas analyses of the Nyiragongo lava lake. It was concluded that later-stage meteoric circulation would not significantly alter magmatic bituminous substances and hydrocarbons.

3.2.3. Deep crustal sources

V. A. Isidorov & Y. N. Federov (Isidorov, 1990) reported 2-2.6 km deep borehole gases to contain volatile hydrocarbons (alkanes and alkenes) and halocarbons, including CHCl3, CCl4, -3 and CFCl3 (CFC-11) at concentrations ranging between 2 and 3 mg m . Assuming standard

Appendix I A.1.11

pressure and temperature at the point of emission, this equals 415-620 ppbv CHCl3, 320-480 ppbv CCl4, and 360-540 ppbv CFCl3 (CFC-11), making it a relatively strong source. Mogk and Mathez (2000) investigated samples from the KTB borehole (Germany) using time-of- flight – secondary ion mass spectrometry (ToF-SIMS), finding a meshwork of carbonaceous matter (mostly graphite) to be present along grain boundaries. Some of their mass spectra with ion fragments at m/z (mass/charge) = 134-138, 151-155, 290-292, and 306-309 show heavy organic compounds to be present as well.

3.2.4. The Earth’s upper mantle

Independent of the route to the surface (basaltic volcanism/xenoliths, kimberlites, serpentinite domes), the presence of hydrocarbons in the Earth’s mantle is likely, considering a) the dominance of CO2 in mantle degassing features and their correlation with primordial helium (e.g., Giggenbach, 1997), b) the apparent reduced oxidation state of the upper mantle (e.g., Ballhaus and Frost, 1994), c) the availability of metal oxides being able to act as catalytic redox buffers in reactions such as the Fischer-Tropsch reaction, and d) the abundance of elemental carbon and hydrocarbons in mantle rocks (Mathez et al., 1984; Sugisaki and Mimura, 1994). Some authors have assumed abiotic, juvenile, hydrocarbon gases to be present in the Earth’s mantle, as an alternative source of oil deposits (i.e. “Deep Earth Gas Hypothesis” of Gold and Soter, 1980, 1982), whereas others see mantle carbon species to be the product of subduction recycling processes, with hydrocarbons derived from serpentinization, the incorporation of sedimentary biogenic matter, and degassing reactions (Javoy et al., 1982; Javoy et al., 1986; Giggenbach, 1992; Zhang and Zindler, 1993; Jambon, 1995; Nishio et al., 1998; Snyder et al., 2001). Sano and Marty (1995) estimate up to 20% of the carbon in high-temperature volcanic gases to be derived from a MORB-type source, and the remainder largely to be derived from CO2 produced from subducted carbonates. The latter hypothesis currently prevails, but is not necessarily opposite to the juvenile hypothesis, which has a strong thermodynamic and experimental basis (Ballhaus and Frost, 1994; Matveev et al., 1997; Zubkov, 2001). The juvenile hypothesis suggests C-H-N-O-S mantle fluids to be generating hydrocarbons at depths from the transitional zone between the lower and upper mantle upwards as well as in the shallow upper mantle.

Appendix I A.1.12

The occurrence of carbon in mantle peridotites in alkalic basalts and kimberlites has been detailed from a large variety of occurrences and was observed not only to be dominantly present as CO2 in fluid inclusions (e.g., Roedder, 1965), but also as condensed carbonaceous matter in the form of discrete platy grains and on grain boundaries and cracks (Mathez and Delaney, 1981; Mathez et al., 1984; Mathez, 1987; Tingle et al., 1990; Tingle et al., 1991). Carbonaceous matter may have formed on olivine in-situ on the sites of deposition (Tingle and Hochella, 1993). Basiuk and Navarro-Gonzalez (1996a) in a review on terrestrial volcanic dust assign great importance to kimberlites as a source of abiogenic, prebiotic hydrocarbons on the early Earth. The Earth’s upper mantle is thought to have had the same oxidation state 3.9 Ma ago as today, resulting into the same type of volatile degassing as seen in today’s basaltic volcanism with a dominance of SO2, H2O, and CO2 (Delano, 2001). Partial reduction of these species to H2, CO and H2S is thought to happen in the hydrothermal envelope of a magmatic intrusion facilitated by water-rock interaction processes (Giggenbach, 1987), as long as a condensed meteoric aqueous phase was available. These mixing zones of volcanic- hydrothermal aquifers are thought to play an important role in prebiotic chemistry (Washington, 2000). Hydrocarbons in kimberlites have also been reported (e.g., Krot et al., 1993; Zubkov, 2001).

Chaigneau (1975) thermally degassed basalts and basaltic olivine nodules from the oceanic hot-spot volcano Réunion (Indian Ocean), finding C2-C4 alkanes and alkenes, acetylene, and benzene. Sugisaki and Mimura (1994) analyzed 227 mantle-derived rocks from fifty localities worldwide for hydrocarbons by gas chromatography-mass spectrometry (GC-MS), finding less evolved magmatic rocks and mantle xenoliths to contain heavier n-alkanes on grain boundaries and in fluid inclusions, whereas more evolved igneous rocks as well as mafic cumulates lack them. They stated that “mantle hydrocarbons” resemble the aliphatic signatures of meteorites and petroleum and are isotopically relatively light (~ -27‰). Their sample chromatograms show typical petroleum pyrolysate distributions, but differ significantly from thermogenic and hydrothermal types (Gize, 1999). Diamond carbon isotopic signatures are in contrast quite heavy (i.e. ~ -2 to –5 ‰; Craig, 1953) but are superceded by carbonate associated with kimberlites, where heavy values up to +33.6 ‰ are reached (Mamchur, 1980) as well as by some carbonaceous matter in serpentinites where +12 to +25 ‰ have been reported (Deines, 1968).

Appendix I A.1.13

3.2.5. Low-Temperature hydrothermal fluids and geothermal steam discharges

Many studies have addressed the occurrence of organic compounds in aqueous low- temperature hydrothermal fluids from different environments such as deep borehole fluids, continental basin brines, submarine accretionary wedge fluids, submarine mid-ocean ridge vent fluids, and low-temperature volcanic-hydrothermal fluids. These observations have been supplemented by a range of theoretical and experimental studies (see below). In submarine mid-ocean ridge hydrothermal fluids, abiogenic methane synthesis has been recently discussed and reviewed by Kelley and Früh-Green (1999, 2000, 2001) and Holloway and O'Day (2000). Temperatures of vent discharges at which organic compounds have been found reach ~ 400°C (Simoneit, 1993).

For subaerial hydrothermal fluids, a great number of studies have been published (see review by Darling, 1998, and special issue of Chemical Geology 71, 1988). Data for a number of volcanic-hydrothermal systems have also been reported by Chaigneau et al. (1970), Pavlov and Karpov (1972), Giggenbach et al. (1990) and Capaccioni et al. (1993, 1995, 2001). Chaigneau et al. (1970) analyzed low-temperature (i.e. t<350°C ) “volcanic gas” samples from Vulcano (Italy) and reported hydrocarbons (alkanes and alkenes) at concentrations of ≤

1000 ppm. They found thionyl fluoride (SOF2) in gases that they desiccated over P2O5 during sampling. Giggenbach et al. (1990) reported on hydrocarbons from fumarolic vapor and thermal spring discharges from the Nevado del Ruiz volcanic-magmatic-hydrothermal system in Colombia and suggested that the unsaturated hydrocarbons (alkenes and arenes) were derived from high temperature interaction of the hydrothermal fluid with sedimentary organic matter at depth.

Capaccioni et al. (1993, 1995) analyzed low temperature volcanic gases from fumarolic and hydrothermal emissions from a number of volcanoes and hydrothermal systems for organic compounds (see discussion on high-temperature volcanic gases below). Higashihara et al.

(1997) found an unusual C10-C11 predominance in gases of the Kakkonda (Takinoue) geothermal area (Japan), with alkanes, arenes and naphthenes dominating the organic chemistry of the geothermal fluid discharge. Capaccioni et al. (2001) reported alkanes, alkenes and aromatics from the Baia di Levante (Vulcano, Italy) low-temperature steam discharges, assuming these compounds to be derived from degrading organic matter by

Appendix I A.1.14

hydrolysis, catalytic cracking, and recombination to naphthenes and arenes at higher temperatures and fugacities of CO2.

Darling (1998) has reviewed the literature pertaining to hydrocarbon analyses from low- temperature sources of hydrothermal systems of different tectonic settings worldwide, including data from sedimentary basin gas fields, crystalline sources, hydrothermal sources of subduction-related fluids, spreading environments and continental systems. He observed that hydrothermal hydrocarbons differ from sedimentary hydrocarbons by a relative “excess CH4”

(higher C1/C6 ratio), higher normal/iso ratios (normal alkanes = straight-chain alkanes, iso- alkanes = branched alkanes) for butane and pentane, and high amounts of C6 volatiles. It has 3 4 to be noted that despite the high R/RA (the ratio R of He/ He, relative to that of air, A) values of some systems observed, he included only those which are characterized by reservoir temperatures <320 °C, thus excluding volcanic gases in the supercritical range. Darling (1998) states, that volcanic gases are “generally devoid of higher hydrocarbons because of their extremely high temperatures”, and concludes that “hydrothermal hydrocarbon gases are generated in the shallow subsurface by the action of heat on organic precursors”.

A significant part of his argument supporting these statements assumes that “the hotter the hydrothermal system (or part thereof), the greater the input of mantle helium”, which is incorrect, since mantle helium reaches shallow hydrothermal convection cells via several routes including deep faults, magmatic activity, and tapping of temporary crustal storage reservoirs. The first and second routes may show significant differences within the same subduction zone. For example, in the Aegean Volcanic Arc (Greece), some relatively cold hydrothermal sources (e.g., Aegina Island) with no magmatic activity in the last 1 million years (Dietrich et al., 1991) that are situated on deep active faults, have elevated R/RA and “excess 3He” concentrations (i.e. excess concentration of 3He with respect to its concentration in solubility equilibrium with air). These testify to mantle helium emissions similar to “warm” hydrothermal systems within the same volcanic arc that display evidence of shallow crustal magmatic reservoirs (e.g. Nisyros Island, Schwandner et al., 1998, Dietrich et al., 1999, Schwandner et al., 1999).

The comprehensive study of light hydrocarbons in fumarolic discharges from most of the Japanese Island Arc has been carried out by Kiyoshu and Asada (1995) in order to distinguish

Appendix I A.1.15

low-temperature “volcanic geothermal thermogenic” from high-temperature “fumarolic type of inorganic origin” sources of hydrocarbons, with a possible magmatic source of hydrocarbons via the Fischer-Tropsch reaction. The low temperature type is approximately equivalent to that reviewed by Darling (1998). They conclude that for both types, the amounts of hydrocarbons are dependent on the oxygen fugacity rather than temperature or other factors

(i.e. fCO2). In the case of the low-temperature hydrothermal discharges of Porto di Levante (Vulcano, Italy), the gases were classified as being of the volcanic-hydrothermal type, in contrast to gases from high-enthalpy geothermal sources explained by thermogenic origins/cracking. High-temperature manifestations are typically ethylene enriched relative to n-butane, whereas low-temperature hydrothermal hydrocarbons show the reverse (e.g., Kiyoshu and Asada, 1995).

Several experimental studies have attempted to confirm the hypothesis of an abiogenic organic synthesis origin of life (Darwin, 1871; Oparin, 1924). One of the most famous was the pioneering study of Miller (1953), known today as “Miller’s experiment”. A great number of articles have been published on the matter since, which are beyond the scope of this review (for an introduction, see Amend and Shock, 1998; Huber and Wächtershauser, 1998; Orgel, 1998). Extraterrestrial delivery has also been discussed repeatedly in theoretical studies based on findings of organic matter in meteorites as well as from relevant experimental data (e.g., Douda and Basiuk, 2000, and references therein). Russian research of the 1960’s to 70’s was also concerned with the question of the origin of life in association with volcanic activity, some of which has been summarized by Mukhin (1976, 1979). Of direct concern to volcanic systems, however, are a few selected experimental, theoretical and field studies, which are briefly summarized below.

3.3. High temperature volcanic gases

Organic compounds in high-temperature volcanic gases are more difficult to sample and hence, they have seldom been reported in the literature. For the case of high-temperature volcanic gases, three cases have to be considered (excluding diffusive emanations): N Explosive volcanic eruptions, N Degassing of lava flows and –lakes (“magmatic gas”), N Active quiescent fumarolic degassing (“fumarolic gases”).

Appendix I A.1.16

3.3.1. Explosive volcanic eruptions

For the case of explosive volcanic eruptions, the organic gas chemistry has been documented by very few data, mainly on halocarbons and thiocarbon compounds in Plinian eruption plumes and compounds adsorbed onto ash particles. Markhinin et al. (1975) found amino acids, hydrocarbons and other organic compounds in volcanic ash of the July 1973 Tyatya (Kuriles) eruption, estimated to have yielded a total of 2 x 108 t of pyroclastic material consisting mainly of ash. Several hundreds of samples of ash were collected just after they fell and while still warm (50-70°C) over a 2 x 2 km grid. Soxhlet extractions were performed with chloroform, acetone and an ethanol-benzene azeotrope and the authors obtained a total estimate of 1.5 x 105 t (0.072 wt%) of organic material erupted during the Tyatya eruption. They found 14 optically inactive amino acids, alkanes, naphthenes, and more complex organic compounds (mainly mono- and polycyclic arenes), most of which with had high boiling points (>300°C), as well as oxygen- and nitrogen-containing compounds. On Alaid volcano (Kuriles), they calculated an organic content of 0.242 wt% in the eruption products. In a later study, Markhinin and Podkletnov (1977a, b, 1978; Podkletnov and Markhinin, 1981) detected over 150 compounds in the volcanic ash of seven volcanoes in Kamchatka (Shiveluch, Bezymyanny, Klynchevskoy, Klyuchevskaya Sopka, and Tolbachik) as well as on the Kurile Islands (Alaid and Tyatya; mainly on ash and a bomb of Tyatya volcano/Kunashir Island, as they were falling during the 1973 eruption) and Indonesia (Agung, Bali). They list saturated, aromatic and polycyclic hydrocarbons, amino acids and sugars, porphyrins, and N, O, S and Cl heterocompounds. The total porphyrin yield in the Tyatya ashes was 5 x 10-8 % (Podkletnov and Markhinin, 1981), and 1 x 10-8 % in a bomb of the Tolbachik eruption (Kolesnikov and Egorov, 1979; Kolesnikov et al., 1979).

These studies employed a rigorous variety of sampling (sterilized glass containers with analytical blanks) and analytical techniques, including wet chemical extraction techniques, Soxhlet-extraction, Biotronic® Amino Acid Analyzer, thin layer chromatography (TLC), GC, and infrared (IR)-spectrophotometry. The gas chromatograms of the n-hexane soluble fraction of chloroform extracts resemble pyrolysates of complex organic matter with a predominance of higher normal (straight chain) alkanes. Whether this is pristine or pyrolysed biogenic organic matter hosted in the ejected pre-eruptive older volcanic products previously forming the crater filling remains unresolved, although the question is raised in their paper. They also

Appendix I A.1.17

sampled and cultivated airborne microorganisms directly adjacent to the sampling sites and the erupting crater. They calculated the mass yield of the maximum possible pyrolysate from such microbial-derived organic matter in the eruption products to be significantly below the observed integrated mass yield over their sampling grid. The possible microbial influence on volcanic ash analyses has been described by Mun and Kuziakina (1976) and Mun and Rogova (1980). Some of the methods used by Russian scientists were described by Ponomarev et al. (1979). Kravtsov et al. (1979) also analyzed Tyatya volcano ash to find that the acetone, alcohol-benzene, and chloroform extracts generically contain n-alkanes, halogenated high- boiling heteroatomic compounds and polycyclic aromatic hydrocarbons (PAH’s). PAH’s have also been found in solid volcanic eruption products by Florovskaia et al. (1978). Other organic analyses of steam discharges on Kamchatkan geothermal fields exist, and have been modeled experimentally (Bondarev and Porshnev, 1980; Bondarev et al., 1980, 1982; Porshnev et al., 1982; Porshnev et al., 1983; Andreichikov et al., 1985; Porshnev and Bondarev, 1989).

In Markhinin's (1985) book “Volcanism”, the Russian work on organic compounds in pyroclastic material is summarized (Isidorov, 1990). Over 150 compounds have been reported in the fresh solid eruption products of seven volcanoes on Kamchatka, the Kuriles and Indonesia (see also Markhinin and Podkletnov, 1977a, b; 1978), comprising C15-C36 n- alkanes, C18-C36 i-alkanes, PAH’s (e.g. pyrene, fluoranthene, benzo-[a]pyrene, benzo[ghi]perylene, coronene), as well as oxygenated and chlorinated compounds, with a predominance of C24-C28 hydrocarbons showing a notable enriched abundance of odd over even homologues.

Several studies by atmospheric scientists investigated the possibility if volcanogenic halogenated hydrocarbons being released by explosive volcanic eruptions. These studies focused on airborne eruption plume sampling of the Soufriere and Mt. St. Helens eruptions. Cronn and Nutmagul (1982b) analyzed airborne volcanic ash-gas cloud samples of the 1979 Soufriere (Guadaloupe) eruption. Using an electron capture detector (ECD) gas chromatograph, they found CFC-11, CFC-12, methyl chloroform and carbon tetrachloride to be within expected background air concentrations, but COS and CS2 were slightly elevated. After the climactic eruption of Mt. St. Helens in May 18, 1980, detailed gas studies were persued on the eruption plume and the fallen ash (Science, 211:815-838, 1981). Three

Appendix I A.1.18

airborne gas studies of the eruption plume were performed by several groups on various days (February 28, March 28, May 13, May 18, May 25, May 28, June 13, July 24) at altitudes up to 20.4 km (by U-2), and at various locations (Inn et al., 1981; Leifer et al., 1981; Cronn and Nutmagul, 1982a).

Leifer et al. (1981) sampled tropospheric and lower stratospheric air along the northern west coast of the USA in February 28, May 13 and July 24, 1980, the later of which flights coincided with eruptive events on July 22nd (sub-Plinian ash eruption, dome formed in June destroyed) and minor steam and ash eruptions on May 13th (Hoblitt, 1986). The samples were analyzed (methodology not available) for CFC and hydrochlorofluorocarbon (HCFC) compounds (# 11, 12, 113, 22), chlorocarbons (CH3Cl, C2H3Cl3, CCl4), sulfur hexafluoride

(SF6), carbonyl sulfide (COS), and other inorganic gas species (CH4, CO2, N2O, CO). On the July 24 flight, they intersected the July 22 eruption cloud at 19.2 km altitude (lower stratosphere), but no significant enrichment relative to ambient air was found except for COS. North-south vertical profiles however do show slight enrichment at lower stratospheric altitudes for several compounds other than COS (i.e., CFC-11, CH4, N2O, CH3Cl, CO, HCFC-

22, CO2, CCl4). It is probable that at the time of sampling two days after the eruption, all non- adsorbed compounds were sufficiently diluted by entrained air to be at background concentration. The elevated concentration of COS most likely reflects a dynamic equilibrium via heterogenous interaction with the sulfate aerosol of the eruption cloud and is therefore a secondary product. The analyses are hence not representative for eruption clouds. It is noteworthy however, that the abundances of certain halocarbons were elevated in the lower troposphere over the San Francisco Bay area.

222 Inn et al. (1981) measured CH3Cl, Rn, CFC-11, CFC-12, COS, CO2, N2O and SO2 in samples obtained during cryogenic aircraft sampling of the climactic eruption cloud of May 18, and of plumes of the May 25 and June 13 eruptions, by GC with photoionization detection

(PID) and flame ionization detection (FID). They found the freons, CO2, COS and N2O to be at background air concentrations, but SO2 and CH3Cl to be consistently elevated. Since they note possible sample degradation problems, their data have to be interpreted with caution. Casadevall and Greenland (1981) measured a variety of fumarolic emissions on Mount St.

Helens, unsuccessfully seeking to analyze C2 hydrocarbons.

Appendix I A.1.19

Cronn and Nutmagul (1982a) analyzed a large number of samples from the Mt. St. Helens eruption which were taken in a similar fashion to their Soufrière plume samples. On March

28, 1980, they detected COS, CS2, CFC-12 and methyl chloroform enriched in the eruption cloud relative to background air at several altitudes and distances to the vent, CFC-11 slightly enriched and CCl4 at background air concentration. They also detected ethane (C2H6), ethylene (C2H4) and acetylene (C2H2) during the same flight, but did not discuss their results in detail: all three compounds were slightly above air background (~1 ppb), with acetylene several tens of ppb above air background. Their data show apparent correlations that were not mentioned in the original publication but may have genetic implications and these are discussed below.

Acetylene and ethane correlate positively (R2=0.89), but neither shows any correlation with ethylene. This may be explained either by a similar origin of the first two species (e.g. pyrolysis, cf. Pereira et al., 1982), or by faster reaction rates of ethylene, which seems unlikely since acetylene with its triple bond is the most labile of the three compounds.

Ethylene also correlates well with all the halocarbons except CCl4 but this merely reflects the small scatter and constancy of the three halogenated species. CFC-11 shows a good positive correlation to methyl chloroform (R2=0.82), suggesting a similar source or process of formation. CFC-11 correlates negatively with CCl4, indicating either different sources or coupled formation and destruction. Because all three C2-hydrocarbons, CFC-11, and methyl chloroform are present in quantities above background air concentrations, and taking the observed correlations into account, a volcanogenic source seems likely. Interpreting time- series measurements, Cronn and Nutmagul (1982a) observed increased concentrations of th CFC-12 and CCl4 on the day of the climactic eruption (May 18 ), and significant stratospheric CFC-11 and CFC-12 peaks after that event. The eruption plume samples of Cronn and Nutmagul (1982a) are mixtures of particulate eruption plume material, volcanic gas, and mainly air. The above correlations show that some but not all of the enriched compounds may be in portions derived from the same source or process:

N volcanic gas as a source, mixing with background concentrations in air; N localised enrichment of entirely air derived portions of a particular compound, by adsorption on ash particles; N in-situ formation of a portion of a particular compound within the ash-cloud.

Appendix I A.1.20

Since the correllations indicate the same origin (i.e., a, b or c) for some of the compounds but not for all compounds, more than one out of the three sources/processes (a-c) must be active. Therefore air entrainment alone can not explain the halocarbon abundances data in the study by Cronn and Nutmagul (1982a).

Rasmussen et al. (1981, 1982) desorbed freshly fallen ash of the 18 May 1980 Mt. St. Helens eruption which released COS and CS2. They also analyzed air samples of the eruption plume that were sampled by aircraft, in SUMMA®-canisters. Their data show above tropospheric background values for methyl chloride (CH3Cl) in the ppmv range, and lower ppb-range for the other methyl halogens (CH3Br and CH3I). In addition, they detected various halocarbons (ethyl chloride, dibromochloromethane, dichlorobromomethane) and more complex organic compounds such as acetone, naphthalene and anthracene. They interpreted their observations as reflecting extreme air dilution of the volcanic gas prior to sampling, and simultaneous in- situ breakdown of more complex organics. Pereira et al. (1980) analyzed freshly fallen ash of the 18 May 1980 eruption and found chloro-aromatics and three previously unknown pentachlorobiphenyls (PCB’s) isomers that would structurally not be expected from industrial emissions. Later, Pereira et al. (1982) showed new data on the same material as well as of surface waters and of “bottom material” samples (pre-eruptive soil surface). They found a large range of organic compound classes, including resin acids and insect hormones, of which many were similar to typical paper and pulp industry effluents. They concluded that eruption- pyrolyzed pre-eruption plant and soil organic matter was the most likely source of the organics. It was doubted that the compounds were the product of de novo synthesis (Fink, 1983).

Using data from Inn et al. (1981) and Cronn and Nutmagul (1982a), Turco et al. (1983) discussed and reviewed evidence for, and the effects of global temperature changes and ozone depletion after the 1980 Mt. St. Helens eruption. A plume simulation model was used to shed light on possible causes and atmospheric effects related to Plinian eruptions. They admitted volcanogenic methyl chloride to be an indigenous contributor to their long-term volcanic cloud model, but considered it to be of minor importance on the assumption that chlorine gases from volcanoes are an insignificant source compared to anthropogenic chlorine emissions (~106 t/a), and because of the photochemical ozone response being equal to or greater than the cloud dispersion times. It was implied that volcanogenic halogens would not

Appendix I A.1.21

reach the polar regions and hence play a role in Antarctic/Arctic ozone depletion, an assumption in conflict with observations of volcanic ash archives established later in Antarctic ice profiles (Palais, 1985; Petit et al., 1999). The first assumption is invalid if global volcanic chlorine fluxes are considered (Johnston, 1980; Symonds et al., 1988; Symonds et al., 1992; WMO, 1998). Turco et al. (1983) concluded that the observed changes (<0.2% maximum hemispherical ozone depletion and with a northern hemispheric surface cooling of

≤0.05 K ) were mainly due to SO2, water vapor and ash loading rather than “exotic” compounds. They considered (incorrectly) that the Mt. St. Helens eruption was of large magnitude even though it is dwarfed by large eruptions such as Mt Pinatubo 1991, El Chichón 1982, Krakatau 1883, and Tambora 1815. The Tambora eruption caused a significant global cooling that resulted in global crop failures (Kerr, 1989). Its impact on the ozone layer has not been studied (to the best of our knowledge), but the growing polar ice-core data will allow quantitative models and correlations in the near future. The review by Basiuk and Navarro- Gonzalez (1996b) summarizes the evidence and theories that indicate the importance of terrestrial volcanic ash cloud phenomena to “prebiotic” abiogenic organic compounds, assuming formation by catalytic Fischer-Tropsch reaction.

The results of the Mt. St. Helens studies led atmospheric chemists to believe that volcanoes may not emit large quantities of halocarbons. No further efforts were made to investigate the halocarbon content of ash clouds of other eruptions and later airborne sampling studies focused on emissions of sulfate aerosol. The St. Helens eruption was unfortunately not representative of most other Plinian volcanic eruptions, since its sulfur output was very limited and the halogen chemistry shifted accordingly (Symonds et al., 1988; Kerr, 1989).

3.3.2. Degassing of lava flows and -lakes

The pioneering studies of Day, Jaggar and Shepherd on magmatic gases from the Kilauea lava lake have been summarized above. Finlayson et al. (1968) analyzed lava lake gases at Kilauea Iki, Alae and Makaopuhi and considered pyrolysis reactions of adjacent vegetation to be responsible for the unspecified non-methane hydrocarbons. From magmatic gas and hot springs on Kurile Islands volcanoes (Mendeleev and Golovnin on Kunashir Island; Alaid on Atlasov Islands) and Kamchatkan volcanoes (Uzon, Malyi Semyachek and Mutnovsky), Mukhin (1974) and Mukhin et al. (1974) reported cyanides (thio- and ferrocyanates) from

Appendix I A.1.22

several thermal springs and ‘pristine’ hydrocyanic acid (metastable, dissociated HCN) from 900°C gas in lava crevices of a 1972 Alaid lava flow emitted from a cinder cone (~0.01mgl-1). Wilkniss et al. (1975) sampled air near a Kilauea eruption and the gas escaping from cracks in lava flows on March 18, 1974, in order to analyze the volcanic gas for CClF3 (CFC-13), following the findings of Stoiber et al. (1971). Concentrations were identical to ambient air of the time (75-86 pptv), which they separately sampled a day after the eruption at the Mauna Loa Observatory.

Mukhin et al. (1978) report cyanide, aldehyde, amino acids, amines, and hydrocarbon concentrations in gases (Alaid, Atlasov Island/Kuriles, same site), thermal water samples (Golovnin volcanic springs and bore hole fluid; Uzon Caldera thermal waters and Koshelev volcano bore hole fluids, on Kamchatka) and sublimate samples from Mutnovsky. They confirmed their earlier results (Mukhin, 1974; Mukhin et al., 1974) on HCN from the Alaid high-temperature cinder cone gases (900-1000°C, 0.01 mg l-1), and additionally documented a rich aromatic, aliphatic, alkenoic and aldehydic chemistry, especially in bore hole fluids from the different volcanic sites although contamination by drilling fluids was not discussed. Other research was concerned with experimental simulation of eruption clouds (Mukhin et al., 1976a), organic compounds in volcanic-hydrothermal systems (Mukhin et al., 1976b), and the role of volcanic processes in the origin of life (Mukhin, 1976; Mukhin et al., 1979). During the Tolbachik fissure eruption of 1975, Menyaylov et al. (1983) sampled volcanic gases from over 30 effusive vents from flows and cones, resulting in ca. 100 gas samples taken by methods based on replacement of a solution (introduced by Bunsen, 1857) and later analyzed by GC-MS. It has to be noted that they used fluoropolymer tubing in their sampling train. They found ethane and propane in up to 9 vol % of the dry gas after steam separation.

Markhinin et al. (1977), and Markhinin et al. (1983) took 36 gas samples from flowing basaltic lava at 1000 ±65°C, during the 1975 Tolbachik eruption, by direct insertion of ceramic tubes into the flowing lava and into gas bubbles before they broke. Using GC, a range of hydrocarbons was identified, including normal (n-) and 2-methyl (iso, i-) alkanes up to hexane, and alkenes (mainly C2-C4). Some of the gases were extremely rich in hydrogen gas, in some cases up to 35% of the dry content. They estimated the overall amount of adsorbed hydrocarbons to be on the order of 0.08 to 0.3 cm3/kg of lava. A direct relationship was noted between the H2 content and hydrocarbon abundance in the magmatic gases, indicating the

Appendix I A.1.23

importance of the redox state in the abiogenic formation of organic compounds. The authors also referred to unpublished spectroscopic investigations of A. N. Senkov & A. A. Trom, who obtained spectra of self-combustion flash flames over bursting gas bubbles at the source of the lava flows, which also indicated the presence of hydrocarbons (CH4, C3H8, C2H2 were identified). At Kilauea, Hawaii, evidence for hydrogen combustion/burning has been reported in spectroscopic observations of volcanic flames (Cruikshank et al., 1973; Naughton, 1973). Graeber et al. (1979) analyzed gases of the 1977 Kilaeuea East Rift eruption, finding simple hydrocarbons that they assigned to pyrolysis of organic matter observed to grow near the eruption.

Rasmussen et al. (1980) analyzed the methyl chloride contents of magmatic gas–air mixtures from the eruption of Kilauea Kalapanain 1977 and 1979, and from degassing lava of the 1977 eruption 8 weeks and ½ year after the eruption. They found CH3Cl concentrations to increase with cooling and degassing after the eruption, as measured by CH3Cl/CO2 ratios. They also found that the observed ratios were far from those expected from prolysis of vegetation during the eruption. Their study was later used by Gerlach (1980) to assess the equilibrium thermodynamic probability of the occurrence of these compounds (see below).

Several studies dealt experimentally with degassing solid products of volcanic eruptions. Muenow (1973) noted “organic constituents, previously unobserved in volcanic gas analyses” in a study on gas release from glassy lava spheres (Pele’s Tears) of Hawaiian tholeiites. He thermally opened enclosed gas bubbles and analyzed the released volatiles by on-line high- temperature (HT) mass spectrometry. The observed mass fragments were assigned to unsaturated C2-C4 hydrocarbons and alcohols. Light alkanes, alkenes and simple arenes have been found in the solid products of the 1971 Teneguia (Canary Islands) alkalibasaltic eruption (Chaigneau and Füster, 1973). Reduced “organic” carbon in magmatic rocks is apparently isotopically light (Hoefs, 1973). Podkletnov and Markhinin (1981) extracted bombs from the Tolbachik 1975 and Tyatya 1972 eruptions and identified PAH’s (including pyrene, fluoranthene, 3,4-benzopyrene and 1,12-benzoperylene) in the Tolbachik bomb, and C15-C36 n- paraffins, C18-C26 i-paraffins with ca. 146 unidentified additional saturated hydrocarbons in the Tyatya bomb.

Appendix I A.1.24

3.3.3. Active quiescent fumarolic degassing

The term volcanic gas is here used as gaseous or steam discharges emanating from volcanically active areas which are characterized by a dominance of juvenile components directly exsolved from a magmatic body, as opposed to hydrothermal steam discharges which are aquifers heated by magmatic thermal anomalies. This distinction is not always readily defined for individual systems, but several chemical lines of evidence have led different authors to be able to differentiate between these two cases. Most successful has been the use of rare gases (i.e., He, Ne, Ar, in addition to CO2 and N2) as well as noble gas isotopes (Giggenbach, 1997) and to a lesser extent sulfur, oxygen and hydrogen isotopes (Giggenbach, 1992). The noble gas isotopic signatures are possibly influenced by non-magmatic, tectonic deep crustal and/or mantle outgassing. Giggenbach (1987) postulated a two-phase vapor-brine envelope surrounding magmatic systems, providing the environment for fluid-rock interaction. The associated processes result in conversion of initially acid, oxidizing magmatic fluids to neutral, reducing hydrothermal fluids. Fluid mixing provides the reactive interface on which organic synthesis may occur. An active field example of the processes is the low- temperature reduced fumarole field which outcrops on the beach at Porto di Levante, Vulcano (Italy) and which is considered to represent a heated aquifer with minor influences from seawater and a juvenile condensation component. The Vulcano crater fumaroles in contrast, are high-temperature reduced acidic SO2-rich gas discharges with a magmatic dominance. The highest temperature island arc static (non-lava flow or spatter cone) fumarolic volcanic gas discharge which has been tapped for analysis of organic compounds is a 900°C gas sample from Kudriavy, Kurile Islands by Wahrenberger et al. (1998).

Rasmussen and Rasmussen (1980) have reported CH3Cl, CH3Br, and CH3I in volcanic emissions of Mt Erebus (Antarctica), St. Augustine (Alaska), Teide (Canary islands), Krisuvik (Iceland) in some cases during eruptive periods. Penkett et al. (1981) analyzed background air and aluminum plant emissions from a variety of locations by GC/MS, including the crater of Mt. Erebus in Antarctica. They found CF4 and C2F6 concentrations at

Erebus to be at background air level (55 and 74 pptv, and 4.3 pptv, respectively), but CF3Cl to be elevated (3.0 and 4.6 pptv) relative to all air and aluminum plant emission plumes (max. 4.3 pptv). It is noteworthy that Sugisaki and Mimura (1994) found Mt. Erebus ultramafic xenoliths to contain up to 1µg/g alkanes (C14-C33).

Appendix I A.1.25

Isidorov and co-workers sampled air taken near fumarolic gas exhausts in “solfatara fields” of the three Kamtchatkan volcanoes Mendelev, Golovnin and Tyatya (Kunashir Island) by GC- ECD after collection of dry gas on activated carbon and Tenax® adsorbent tubes (Isidorov, 1985; Isidorov et al., 1985; Isidorov and Zenkevich, 1985; Isidorov and Ioffe, 1986; Isidorov et al., 1990). They found that CFC-11 and -12, CHCl3, CCl4 and C2Cl4 in their samples exceeded background air concentrations by up to an order of magnitude. They speculated that high-temperature reactions of hydrocarbons with HCl and HF or with other Cl- or F-bearing inorganic compounds and metal halide catalysts took place to form halocarbons in volcanic gases. Considerable differences in the abundance of halocarbon species from different volcanic sites was attributed to varying F- and Cl-concentrations of the volcanic gases. The authors noted by way of confirmation of their results, the work of Khalil and Rasmussen (1982) on Antarctic ice core data, where CFC-11 found in >1100 years old entrapped air bubbles (17 ppt) was interpreted as air contamination during sampling. Isidorov et al. (1985) suggested that if that was the case CFC-12 should also be present in much higher concentration, which is not the case in the ice core samples. Isidorov's (1990) excellent textbook on the organic chemistry of the atmosphere presents a number of organic analyses of volcanic gases. He affirms, that “apart from methane, volcanic gases always contain its C2-C6 homologues and light unsaturated hydrocarbons of the ethylene series”.

With a modified sampling technique, Isidorov (1992) directly sampled fumarolic volcanic gases, hydrothermal gases, and solfataric air from several volcanoes on Kamchatka (Kikhpinych, Dolina Geyzerov, Uzon Caldera, Mutnovskiy) and from Masaya Volcano, Nicaragua for GC-MS analysis of organic compounds, especially halocarbons. They list 73 compounds, 29 of which they had already reported from volcanic gases on Kunashir Island (Isidorov et al., 1985). Alkanes and alkenes as well as oxygenated and sulfonated compounds (including O, S, N heterocycles) and halocarbons were reported. They quantitatively determined volcanic gas halocarbon concentrations to be significantly above polluted industrial urban air maxima in the cases of CFC-11 (CFCl3), CFC-12 (CF2Cl2), carbon tetrachloride (CCl4), and methyl chloride (CH3Cl). These findings led to a debate about the contribution of volcanoes to the atmosphere’s halocarbon budget. Gribble (1994d) in his review article on “The Natural Production of Chlorinated Compounds” mentions volcanoes as a natural source of CFC’s and aromatics, citing Stoiber et al. (1971) and Isidorov et al. (1990). Gaffney (1995) commented that since the samples by Isidorov et al. (1990) were

Appendix I A.1.26

taken “downwind in volcanic gas plumes”, volcanoes may not be a ‘real’ source of CFCs. He considered CFCs in volcanic plumes were rather due to a mixing phenomenon with the surrounding troposphere. Gribble (1995c) replied that “gas samples were taken directly from solfataric vents where mixing with tropospheric air is expected to be minimal or nonexistent”. Isidorov (1995) clarified the misinterpretation of Gaffney (1995), stating that the CFC concentration in his measurements of volcanic gases was “3-4 orders of magnitude higher than background concentrations and exceeds by 1-3 orders of magnitude the value of typical urban air”, making it impossible to consider tropospheric air contamination as the only source of halocarbons in his samples.

Capaccioni et al. (1993, 1995, 2001) analyzed the headspace gases of NaOH absorption samples (commonly referred to as “Giggenbach-bottles”) from fumarolic and hydrothermal gas emissions from a number of volcanoes and hydrothermal systems around the world by

GC-FID and GC-MS and identified light hydrocarbons (C1-C5 n- & i- alkanes and –alkenes; naphthenes, aromatics). They observed three characteristic compositional abundances: alkane- rich, low-temperature gas emissions from recently active volcanic areas; alkene-rich magmatic fumaroles on active volcanoes; and aromatic- and naphthene-rich hydrothermal manifestations. A number of processes were considered responsible for the observed results. In the case of alkane and alkene-rich emanations, thermal cracking of hydrocarbon-enriched hydrothermal fluids occurred upon mixing with magmatic gas, and in the case of aromatic- rich hydrothermal emanations, smectite or zeolite catalyzed rearrangement processes had taken place. The only high-temperature (>350°C) fumaroles studied were on Galeras volcano (642°C), which showed distinct characteristics but were broadly similar to some lower- temperature (118°-313°C) magmatic fumaroles which they also studied at Vulcano crater (Italy), Purace (Colombia) and Poas (Costa Rica). Further improvements to the sampling approach and novel analytical techniques were employed by the ETH group since 1993, using the method of Isidorov et al. (1990) as a starting point (Wahrenberger et al., 1996a; Wahrenberger, 1997; Wahrenberger et al., 1998; Schwandner et al., 2000a, b; Schwandner et al., 2001a; Schwandner et al., 2001b; this work). The studies have comprised fumarole samples from a number of volcanoes having discharge temperatures in the range from 100° to 900°C. No summary on organic compounds in volcanic gases is complete without reference to the two Russian texts “Volcanism” by Markhinin (1985), and “Volcanogenic Organic Matter” by Podkletnov (1984).

Appendix I A.1.27

Jordan et al. (2000) analyzed volcanic gas samples from Etna and Vulcano (Italy) as well as

Kuju and Iwojima (Japan) by absorption of acid gases, CO2 and water in partially NaOH- filled evacuated glass bottles (i.e. Giggenbach bottles), allowing partitioning and enrichment of less soluble and/or reactive compounds into the associated headspace. They claimed identification of numerous halocarbons, including several iodinated compounds, all of which were estimated to be at background air concentrations. These data have to be interpreted with caution for several reasons. First, the analytical technique was not specified in sufficient detail (no mass fragment spectra or chromatograms were published), making independent interpretation impossible. The identification of iodated compounds by ion trap detector mass spectrometry (ITD-MS) is questionable as these compounds tend to show atypical fragmentation behavior by ITD-MS, by over-facile fragmentation. Furthermore, by starting mass spectrometric scans at 48 amu, identification of low-molecular weight fluorinated hydrocarbons is vague by eliminating diagnostic ions such as m/z 31 (CF●) and m/z 33 ● (CH2F ). Second, there is ample reason to believe that quantitative analyses of NaOH- headspace leads to ion dissociation, dissolution and formation of new compounds by reaction and interaction with the liquid. We consider their analyses to be not quantifiable because of the possible loss of analytes into solution, and because the solubilities of most of these compounds are poorly known for strong aqueous alkaline electrolytes.

As an example of potential problems, it should be noted that artifacts may arise from dissolution of carbon monoxide in the strongly alkaline solution to form formate (Giggenbach and Matsuo, 1991), or formic acid at the solution-gas phase boundary (e.g., Brodie, 1873), which may react further to form more complex molecules. No corrections were applied for such effects. In addition to spelling and identification errors (e.g., chloroselenophene has only two isomers), a third major shortcoming is associated with the Jordan et al. (2000) paper: they speculated on the formation of halocarbons in a volcanic gas via the reaction from CH4 to acetylene with subsequent halogenation to form precursor molecules of halogenated hydrocarbons. If acetylene (C2H2) was the intermediary, even carbon numbers would predominate in the products, which is not evident in their list of compounds. For example, the report to have identified tetrachloropropene (C3HCl4) which has at least one α-chlorine, making it unstable during transport, storage and analysis. Their data and interpretation thereof must therefore be treated with caution.

Appendix I A.1.28

3.4. Non-volcanogenic associated occurrences

Biogenic organic matter originating from different sources but discharged by volcanic systems is clearly identifiable using organic geochemical and biomarker techniques, although the materials may show anomalous geochemical signatures. Such anomalies may be explained by processes acting within the volcanic edifice (i.e. fractionation/distillation, reactions, and chromatographic separation) in response to the magmatic heat supply and fluid transport properties and the resulting geochemical signatures are in contrast to organic matter found in basin brines (Igari and Sakata, 1992; Matsumoto and Watanuki, 1992; Gize and Macdonald, 1993). For example, Matsumoto and Watanuki (1992) found a rich biogenic organic chemistry in the crater lake water of Kusatsu-Shirane volcano (Japan) which they attributed to sulfur reducing bacteria on the basis of biomarker compounds. In contrast, unusual bituminous material was found within a trachyte flow on Suswa volcano (Kenya Rift valley, Kenya) that is characterized by an enrichment of O and N heteroatomic organic compounds and a depletion in aliphatics (Gize and Macdonald, 1993). The authors explained this phenomenon by liquid or supercritical CO2-transport of crater lake sediment organic matter. Gize and Macdonald (1993) attributed the depletion in aliphatics to transport processes whereas Matsumoto and Watanuki (1992) documented a non-depleted aliphatic signature in the Kusatsu-Shirane crater lake chemistry.

The crater lake sample from Kusatsu-Shirane serves as an example of the precursor to the transported product occurring at Suswa. Igari and Sakata (1992) have also shown that low temperature clay mineral and zeolite assemblages may cause some degree of fractionation of aliphatic compounds, especially at lower temperatures. Such conditions occur on many volcanoes, where low-temperature volcanic gases (>100 to ca. 350 °C) migrate through altered volcanic material. A different range of volatile organic compounds is emitted during the interaction of lava with shallow seawater. The resulting vapor is a mixture of magmatic gas, evaporated seawater, and the reaction and decomposition products of organic matter already present in the seawater. The products of this process are dominated by biogenic organic matter and have been investigated to assess possible health effects on humans in terms of volatile metals and organic compounds (Kullman et al., 1994). Coal deposits are known to contain bromine and chlorine (Vassilev et al., 2000), but no study has been done to date to evaluate their halocarbon source potential.

Appendix I A.1.29

3.5. Experimental and theoretical studies

Gulbransen (1966) experimentally reacted graphite with deuterated hydrogen gas to determine if any hydrocarbon reaction products formed via C-D bonds. He found hydrocarbons other than methane to have formed in greater amounts compared to non-deuterated hydrogen gas experiments at the same conditions and suggested the formation of hydrocarbons in the early Earth’s atmosphere and oceans by volcanic outgassing. Shimizu (1975) calculated the temperature dependence (1500°–5500°C, ~100 bar) of the partial pressures of important prebiological molecules including HCN and HCOH in volcanic gases under magmatic conditions of the early molten Earth. He concluded sufficient concentrations of such compounds were present in the early Earth’s atmosphere to have also dissolved in the newly forming oceans. Bondarev and Porshnev (1980) exposed a water vapor-methane mixture to 1000°C lava and produced low molecular weight unsaturated hydrocarbons, acetaldehyde, acetic acid, benzene and homologues, and PAH’s including naphthalene, diphenyl (biphenyl), and pyrene. They attempted to mimic the reaction conditions at the base of volcanic ash plumes during an eruption, an environment where erupting gas/ash jets quickly remove new compounds from the zone of formation to cooler zones with greater air dilution (Isidorov, 1990).

Gerlach (1980) re-evaluated the pioneering 1913-1922 Kilauea gas analyses by Day, Jaggar and Shepherd (see above) theoretically. After restoration of the analyses, he calculated the probability of formation of organohalogens and simple organics using a simple equilibrium thermodynamic approach. He considered methyl chloride to be potentially abundant at 1 bar with 0.0004 – 1.0 ppbv, depending on which real input gas analysis was used. The maximum calculated concentrations of methyl chloride at near atmospheric pressure occurred in the temperature range from 100-500°C. At higher pressures, equilibrium concentrations reached 56 ppbv (900°, 100 bar). However, the actual measured concentrations observed by Rasmussen et al. (1980) in Kilauea samples reach 930 ppbv which possibly indicates non- equilibrium processes, which Gerlach (1980) assumed to be pyrolysis, with lava serving as the heat source only. However, the CH3Cl/CO2 ratios obtained by Rasmussen et al. (1980) on the same samples are not consistent with typical pyrolysis products. An important observation is the greater amount of methyl chloride found at higher temperatures, corresponding to the increased gas phase dissociation of HCl. Gerlach (1980) calculated the methyl halogens

Appendix I A.1.30

CH3Br and CH3I to be abundant at concentrations one to two orders of magnitude above that of CH3Cl, corresponding to their respective bond energies as HX acids (X=any halogen).

However, the abundances of the fluorinated halocarbons (i.e., CH3F) were reported to always occur at several orders of magnitude lower concentration than CH3Cl. Using common cellulose (C6H10O5) as a theoretical carbon source, calculated equilibrium concentrations of about 250 ppbv were obtained. On the basis of these observations, Gerlach (1980) interpreted any hydrocarbons, including halocarbons, in volcanic gases to be derived from pyrolysis of plant matter that grows near the eruption sites.

Thermodynamic equilibrium modeling has been applied to evaluate and predict field analyses. Symonds et al. (1988) and Symonds (1990) performed a thermodynamic study of the stability of individual halogenated compounds in cooling volcanic gases, to assess the different transport mechanisms of halogens in a gas dominated hydrothermal volcanic environment and their emission to the atmosphere. The study was based on gas data from St. Augustine volcano (Alaska), Mt. St. Helens (WA, USA), Merapi (Indonesia), White Island (New Zealand), and Etna (Italy). Their equilibrium thermodynamic calculations suggested extremely low halocarbon concentrations and they explained the higher concentrations in field samples as resulting from pyrolysis of vegetation. Equilibrium models do not necessarily mimic reality in a cooling volcanic gas, and metastable disequilibrium compositions may be expected in cases where the magmatic gas has been modified by wall rock and hydrothermal fluid interaction (Symonds et al., 1994).

Fegley (1991) thermodynamically modeled the lunar volcanic gas stabilities of various halogenated gas species, including CF4 and CCl4, concluding that they may have been even more abundant than HCl and HF. Based on a theoretical thermodynamic approach, Shock and

Schulte (1998) discussed the reduction of CO2 in 50°-250°C hydrothermal vent fluids upon mixing with cold seawater to produce a mixture of carboxylic acids, alcohols, and ketones in the presence of excess hydrothermal H2 (controlled by the host rock redox buffers). Bureau et al. (1999, 2000) and Bureau and Metrich (2001) experimentally determined partition coefficients (Df/m) for Cl, Br and I between albite melt and an associated fluid phase at 2 kbar and 900°C under conditions approximating volcanic systems. Their data indicated that all three halogens partition into the fluid phase and allow one to estimate the amount of Cl, Br and I emitted by large explosive volcanic eruptions on the basis of mass balance calculations

Appendix I A.1.31

and known average concentrations of the pre-eruption magmatic educts. Their calculations show the global annual volcanogenic bromine flux of large explosive eruptions to be comparable to that of current anthropogenic emissions. Zolotov and Shock (2000b) discussed the possible abiogenic synthesis of condensed hydrocarbons in volcanic gases upon cooling and dilution on the basis of theoretical thermodynamic calculations assuming metastable chemical equilibria involving CO, CO2, H2 and H2O in the gas phase. They found n-alkanes and PAH’s of preferentially high H/C ratios to form metastably from CO and H2 at temperatures below 250°C in a cooling volcanic gas, within the stability field of graphite. Heide et al. (2000) developed an on-line gas-measuring device for experimental silica melt charges. They found molecular fragments of hydrocarbons in the experimental range of 800 to 1200°C. It has to be noted though, that their input components were organic materials (saccharides and Na-acetate).

A recurring problem in experimental studies is the potential for direct and indirect reactions between the metal / transition metal surfaces of reaction vessels and sampling equipment and the compounds of analytical interest. Direct surface effects involve processes such as leaching by complex formation, and can influence the speciation and availability of the analytes. Reactions may also occur between the surfaces and the matrix which can secondarily influence the analytes. Indirect reactions involve surface adsorption / surface complexation, with the potential for catalytic effects and selection of certain species or products of the analytes. Typical transition metal alloy materials used in experiments are commonly based on iron/chromium/molybdenum/nickel (stainless steels and Hastelloys), titanium and titanium alloys, platinum, or gold. For example, Voglesonger et al. (2001) presented experimental evidence for the quartz and quartz-magnetite catalysed abiogenic formation of methanol from a CO2-H2-H2O mixture. The reaction vessel was made of stainless steel, which is not an inert surface. Stainless steel is deactivated for ambient temperature storage of gaseous samples of volatile organic compounds by electropolishing (e.g., Summa®-process) or deactivation using phenyl group chemisorption. Using Ti- or Pt-alloys instead is disadvantagous for several reasons. For example, at the maximum temperatures of the experiments by Voglesonger et al.

(2001), Ti can produce H2 if only minor amounts of HCl are present (>0.1%, Giggenbach and

Matsuo, 1991) by the oxidation (corrosion) reaction Tis + 2 H2O ↔ TiO2, s+ 2H2, representing a direct reactive effect. Metallic Pt is a surface on which benzyl groups readily form stable surface complexes up to elevated temperatures (Kryachko et al., 2001), representing a

Appendix I A.1.32

secondary sorptive effect. The only other appropriate inert coating apart from Au may be Silcosteel® treated surfaces, as used in high-temperature applications for NASA’s space shuttle and in gas chromatographic equipment (Thompson and Wise, 1998).

3.6. Medical sciences

In the medical science literature, an awareness of volcanoes as possible sources of organic toxins has increased in the last three decades. Shabad et al. (1971) recorded carcinogenic benzo[a]pyrene concentrations in soil and considered the possibility of volcanic eruptions as a source (p.1189). Kung et al. (1981) speculated on the possible intake of an unknown carcinogenic agent present in lava through consumption of fish products. Spitz et al. (1988) suspected a carcinogenic agent in volcanic lava following a regional survey of thyroid cancer rates which found populations inhabiting volcanic areas (Hawaii, Colombia, Philippines, Polynesia and New Zealand) to have significantly increased incident rates, irrespective of the local resident ethnicity. In their speculation of the carcinogen source, they cite Ilnitsky et al. (1976) postulating a theory of possible hydrocarbon releases from the pyrolysis of complex organic material during volcanic events. Ilnitsky et al. (1976) analyzed 1974 soil, vegetation and juvenile volcanic ash samples in the vicinity of Kamtchatkan volcanoes (Burlyashiy, Uzon, Klutshevskaya Sopka and Valley of Geysers in Kronotsky National Park) and Tyatya volcano (Kunashir Island, Kuriles). Having specifically analyzed for the carcinogen, benzo[a]pyrene, they estimated a global annual emission of 120 kg/a as well as a global annual volcanic ash-derived soil loading of 2.35 x 10-4 µg/m2 and speculated on an even higher prebiotic endogenic production on the early Earth.

Ilnitsky et al. (1977a,b) supported their results with new data from Tobalchik and Tyatya volcanoes, correcting their global annual benzo[a]pyrene emission estimate to 12-14 tons/a. They also speculated on the formation on nitroso compounds during electrical discharges in eruption ash clouds following detection of amino acids, hydrocarbons and other organic compounds in juvenile volcanic ash by Markhinin et al. (1975). Ilnitsky et al. (1979) showed ample evidence of pre-industrial, volcanically derived background concentrations of benzo[a]pyrene in soils, with data from up to 10,000 years old epigenetically frozen permafrost soil and several glacially denuded surfaces, glacial ice and snow layers intercalated with ash of known historic eruptions of nearby volcanoes in the Kamchatka and

Appendix I A.1.33

Chukotsk regions. Inorganic volcanic products may represent a greater health hazard during volcanically quiescent times on the local population through trace metals, sulfate aerosols, acid rain collected for drinking water (Baxter et al., 1982), As and Hg emission and deposition (Signorelli, 1997; Signorelli et al., 1998; Pirrone et al., 2001), or as sudden outburst of CO2 from crater lakes (e.g., Lake Nyos , Cameroun, in 1986: Baxter et al., 1989; Ze et al., 1996).

4. CONCLUSIONS

In this review, we presented a) a compilation of the historic research on the organic chemistry of volcanic gases from the earliest attempts of A. von Humboldt, Sir H. Davy, Gay-Laussac, R. W. Bunsen, and F. Fouqué to A. Day and S. Shepherd, and b) a critical review of the post- war research on the sources and composition of high-temperature volcanic gases. Historical data are used in modern theoretical, field and experimental studies of volcanic surveillance, exploration, exobiology, medical sciences, atmospheric chemistry, and in hydrothermal studies concerned with the question of the origin of life on Earth. Several theoretical studies have suggested de novo abiogenic formation of organic compounds to occur in volcanic gases by catalytic Fischer-Tropsch synthesis. The results are broadly consistent with the findings of the few existing experimental studies. Field studies on natural systems have claimed to have found evidence for a variety of processes to be responsible for organic compounds in volcanic gases, among them contamination, pyrolysis and cracking of biological matter (i.e. fossil fuel), and pyrolysis of vegetation in the case of explosive volcanic eruptions and lava flows emplaced over vegetated areas. Transport processes by hydrothermal activity may form, alter and destroy organic compounds. We find the atmospheric impact of volcanic emissions of volatile organic carbon compounds (VOC) to be a neglected field of research.

Appendix I A.1.34 A.2.1

A2. Additional GC-MS method information

Figure 1 Schematic of the sampling method employed in this study. The strength of this method is that organic trace analytesin the pptv to ppbv concentration range get quantitatively trapped after the strong matrix components water, sulfur and halogen acids get separated off during sampling. Headspace/sparging experiments with the condensates show a great separation efficiency, leading to an up to over 600 times enrichments in the adsorbent cartridge versus the condensate. Sampling was done by focusing fumarolic gas flow into a 0.5 m long quartz glass tube with perforations on the lower to enhance ingress of gases. The quartz glass tubes were inserted into the high gas flow fumaroles, allowed to equilibrate within the high-temperature gas stream overnight, and if necessary cemented with sulphur melt naturally forming at the fumaroles. The temperature profile through the focusing tube was measured before and after sampling, to ensure that no cooling by entrainment or mixing with air has occurred. The sampling tube was then connected to an all-glass sampling train. The all-glass sampling train components were connected head-to- head by support of brass fittings with M4-ferrules (Supelco, Inc.), such that no graphite or metal surfaces were in contact with the gas stream. The acid, sulphur and water matrix was removed from the gas stream by means of a 2-stage ice- cooled condenser array that could be sealed off with vacuum-stopcocks (acid washed and thermally desorbed), to preserve the condensate with its headspace for later analysis. The sampling train was wrapped with aluminium foil to avoid possible photolysis of compounds. The final adsorption temperature was 4°C. The natural gas flow was supported by a low-flow battery-driven air sampling pump with inert wetted surfaces (SKC Inc., PA, USA) downstream from the adsorbent. The pump was operated at a calibrated rate in the range of the natural flow, which also provided information on the sampled gas volume. Adsorbent tubes were stored frozen and sealed with Kapak® endcaps, and condensates at ~4°C in the dark prior to analysis. Sufficient blanks were run to ensure purity of the process, including field blanks of identical adsorbent tubes analyzed under identical conditions.

Appendix II A.2.2

Figure 2 The custom-built short-path thermal desorption system employed in the development of this method (a) is shown together with modifications of the system for analysis of condensation bottle headspaces (b) and solid sublimate samples (c), with external temperature control, flow control and all-glass contact surfaces. The major advantage of this combined method is that the analytes get quantitatively transferred while remnant water and sulfur background gets reduced dramatically. This leads to a strong enrichment and sensitive signals of analytes with a low baseline, allowing analyses into the pptv range despite the severe adverse matrix constituents of volcanic gases. The adsorbent tubes were thermally desorbed at 300°C (120°C for Florisil), re-adsorbed onto a SPME fiber at 25°C, which was thermally desorbed in the splitless inlet of the GC, cryogenically focused on the column head with liquid nitrogen, separated by the analytical column and the eluents then analyzed by full-scan single quadrupole MS. The same high-purity helium supply was used in the conditioning of the adsorbent tubes for desorption and as a carrier gas of the gas chromatographic separation. The detailed conditions are listed in Table 1.

Appendix II A.2.3

Figure 3 Reconstructed single-ion chromatograms of the volatile end of analyses of volcanic gases from Vulcano (Vulcano Island, Aeolian Islands, Italy) crater fumaroles taken in 1999 and 2000 at temperatures between 390 and 440°C are shown. The three types of adsorbent cartridges show differing trapping efficiencies. The Carboxen 569 tube (sample HP4) is despite co-eluting water highly sensitive at the very volatile beginning of the chromatogram, where the simple but important CFC’s (e.g., CFC-12, CFC-11), as well as methyl bromide and methyl iodide elute, among others. Aldehydes and sulfonated compounds are resolved well, with strong peaks. The volcanic gases contain benzene and toluene as the two compounds of highest concentration, and their peak areas are slightly surpressed in the first sample. The Tenax® cartrige is affected by strong alcohol and acid peaks, and the very volatile beginning of the chromatogram is depleted. This reflects the low breakthrough volume for water that is characteristic for Tenax®. The Florisil® cartrige yields strong signals for aliphatic hydrocarbons as well as arenes, but weak signals at the very volatile beginning of the chromatogram. Both the Florisil® and the Tenax® tubes emphasize aliphatic hydrocarbons but show poorer resolution for the very volatile spectrum. Similar tests with Anasorb CSC and Carbotrap yielded good results, leading to the use of composite (multibed) adsorbent tubes of all three adsorbent materials in this method.

Appendix II A.2.4

Figure 4 Total ion chromatogram (TIC) trace of a standard run using 12.5 ppbv of a TO-14 compound standard mixture (Scotty Inc., PA, USA). The inset shows the complete resolution of the highly volatile fraction, even m- and p-xylene resolve reasonably well. Spectrum obtained on the MD 800 GC-MS system (see Table 1).

Appendix II A.2.5

Figure 5 Diagram showing the relationship of the relative retention time to the carbon number of two homologous series, of n-alkanes and n-aldehydes, under the changing run conditions. Apparent relative retention indices are derived by interpolation and mathematical treatment. The retention time, either expressed as retention time relative to a common early eluting peak (in our case, CO2 proved most convenient), or as a time- independent resolution-limited retention index relative to a homologous series of compounds (e.g. n-alkanes or n-aldehydes). These observed (“apparent”) retention indices can be compared to a retention time window of a specific compound given by the average and standard deviation of a set of reported literature values obtained in experimental studies under controlled conditions on the same stationary phase. Most isomers of the compounds can be distinguished by this method, provided that a good chromatographic separation is achieved, and that the literature reference values are accurate. Further details are given in chapter 2.

Appendix II A.2.6

Figure 6 Example of calibration curves obtained from separate and repeat analyses of certified reference gas (cf. Figure 4). The data shown are those used in the calibration of methyl bromide in chapter 2.

Appendix II A.2.7

Figure 7 Enrichment ratios of sample detector responses showing strong enrichment of analytes in the adsorbent cartrige over the corresponding sampled condensate (sampling volume corrected), for most compounds. The poor enrichment of some very volatile compounds is mainly due to the above described depletion caused by using Tenax® TA as an adsorbent and is individual to the displayed sample only. These values are strong underestimates because of insufficient trapping of the very volatile fraction by Tenax®. A clear trend towards stronger enrichment of later eluting compounds reflects their higher octanol-water partition coefficients and lower volatilities. See also Table 2.

Appendix II A.2.8

Table 1 Detailed analytical conditions of the final method used in analyses described in chapters 2 and 3.

Appendix II A.2.9

Table 2 Detailed list of apparent retention indices of a selection of organic compounds found in this study. Enrichment factors for sample over condensate are given as well, for sample TX 1, taken from fumarole F8/9 on Vulcano island (Italy). Two consecutive desorptions using different durations show desorption efficiencies being a function of desorption time. The enrichment efficiency of the analytes on the adsorbent tube can be read from the response ratio of the sample analysis over the corresponding condensate analysis (spl/cond.). Symbols: CAS # = Chemical Abstracts Service identity code; I = Intensity (detector response); s = standard deviation; n = number of analyses; spl/cond = response ratio of sample over condensate analysis (cf. Figure 7).

Appendix II A.2.10

Appendix II A.3.1

A3. Chromatograms and mass spectra

On the following pages, several RIC and TIC chromatograms as well as 102 selected erpresentative and exemplaric mass spectra from analyses performed in this study are reproduced. For each of the four fumarolic samples discussed in the text, three chromatograms are given. First, a reconstructed ion chromatogram (RIC) of m/z 43+57 (typical hydrocarbons ion mass fragments), then a RIC of m/z 35 (a typical chlorocarbon ion mass fragment), and finally the total ion current (TIC) chromatogram, representing the sum of all analysed ions. The mass spectra are taken from samples discussed in the text. Note the clarity of the spectra despite the use of a benchtop GC-MS system, the trace abundances, complex matrix and high number of compounds in each sample.

A3. Chromatograms & Mass Spectra A3.2

Sample TX1, fumarole F8/9, t=438.3°C, June 4 1999 (scan EI+ mode)

m/z = 43+57 3.33e6 100 1901 n-tetradecane 1504 n-dodecane 1708 n-tridecane

% 1994 not identified 1286 n-undecane 488 acetoacetone 578 butyl acetate 1132 acetophenone 1471 n-decanal 474 toluene 281 chloroethanol 1247 n-nonanal 237 acetic acid 158 acetone 1055 n-decane 175 DCM 1615 n'nonanoic acid 135 acetaldehyde 815 n-nonane 0

100 m/z = 35 1.61e6 y-scale x 150 281 chloroethanol 133 methyl chloride

% 1025 1,4-dichlorobenzene 632 chlorobenzene 569 not identified 1401 trichlorobenzene (isomer) 236 chloroform 474 toluene 263 1,2-dichloroethane (ethylene chloride) 347 trichloroethylene (TCE) 1234 not identified 1969 not identified

0

100 TIC 4.12e7 474 toluene 175 DCM 175 DCM

% 1132 acetophenone 705 & 707 m- p-xylene 1025 1,4-dichlorobenzene 1901 n-tetradecane 129 sulfur dioxide 632 chlorobenzene 880 benzaldehyde 1620 impurity 1415 naphthalene 1076 1,2-dichlorobenzene 1003 1,2,4-trimethylbenzene 281 chloroethanol 684 ethylbenzene 757 o-xylene 1708 n-tridecane 1504 n-dodecane 1222 not identified 66 nitrogen 578 butyl acetate 1682 n-undecanal 1247 n-nonanal 1286 n-undecane 966 2-ethyltoluene 1471 n-decanal 292 benzene 488 acetoacetone 237 acetic acid & chloroform 746 styrene 0 200 400 600 800 1000 1200 1400 1600 1800 2000 scan number

A3. Chromatograms & Mass Spectra A3.3

Sample FL2, fumarole F5AT, t=362.3°C, June 7 1999 (scan EI+ mode)

100 m/z = 43+57 2.05e6 y-scale x 150 1571 n-dodecane 1354 n-undecane 643 acetic acid butyl ester 1123 n-decane

% 1315 n-nonanal 555 2,4-pentanedione 541 toluene - benzene isomer 4 1983 not identified 223 acetone 1539 n-decanal 1914 not identified 1969 n-tetradecane 1670 not identified 1776 n-tridecane 1147 C 881 n-nonane 1510 not identified 240 DCM 1158 2,3-dihydro-1H-indene (indan) 1078 n-octanal 200 not identified 304 ethyl acetate 438 n-heptane 0

100 m/z = 35 1.74e6 y-scale x 150 240 DCM 699 chlorobenzene

% (193 methyl chloride) 1093 1,4-dichlorobenzene 301 chloroform 1469 not identified 541 toluene 413 trichloroethylene (TCE) 1080 1,3-dichlorobenzene 1380 not identified 1716 not identified 1805 not identified 1995 not identified 0 TIC 100 6.35e7 542 toluene - benzene isomer 4 1148 C 240 DCM isomer

12

% H 194 sulfur dioxide 10 772 & 774 m- p-xylene 1291 C 699 chlorobenzene 1071 1,2,4-trimethylbenzene 824 o-xylene 746 styrene 994 3-ethyltoluene 1093 1,4-dichlorobenzene 1315 n-nonanal 1354 n-undecane 975 n-propylbenzene 1571 n-dodecane 643 acetic acid butyl ester 1539 n-decanal 357 benzene 555 acetoacetone 1983 not identified 1123 n-decane 1078 n-octanal 1776 n-tridecane 0 200 400 600 800 1000 1200 1400 1600 1800 2000 scan number

A3. Chromatograms & Mass Spectra A3.4

Sample HP2, fumarole F11b, t=403.2°C, June 2 2000 (scan EI+ mode)

x 50 m/z = 43+57 5.75e6 100 (y-scale x 50) 1255 an ethyl alcohol 910 n-nonane 218 acetic acid 183 not identified 1344 not determined

% 1020 not determined 572 furfural 1156 n-octanal 678 bleed 528 dimethyl-1,4-dioxane (isomer) 1210 n-decane 1634 bleed 132 not identified 616 not determined 1207 not determined 742 3-methylbutyric acid 1432 n-nonanal 1677 n-decanal 1773 not determined 1083 1H-pyrrole-2,5-dione 231 thiophene 296 n-heptane 314 2,4,4-trimethyl-1-pentene 165 2-methylpentane 844 n-heptanal 1964 bleed 1474 n-undecane 143 DCM 419 toluene 116 acetaldehyde

0

100 m/z = 35 4.55e5 F-12 y-scale x 100 143 DCM 185 chloroform 124 ethyl chloride % 130 Br-Cl cmpd. 319 not determined 232 carbon tetrachloride b.d. 275 trichloroethylene (TCE) 1091 not determined 1011 not determined 553 tetrachloroethylene 612 742 3-methylbutyric acid 1165 1,4-dichlorobenzene 419 toluene 792 not determined 1149 1,3-dichlorobenzene 932 i-propylbenzene 1376 not determined 1491 impurity 1609 naphthalene 1766 not determined 2009 not determined 0

100 2 TIC 7.43e7 112 COS 148 CS 419 toluene 184 THMF* 116 acetaldehyde

% 227 benzene 218 acetic acid 678 bleed 109 carbon dioxide 231 thiophene 1243 bleed 1643 benzoic acid 1646 bleed 165 2-methylpentane 1116 phenol 433 not identified (not acetoacetone) 1634 bleed 572 furfural 987 benzaldehyde 1964 bleed 0 2027 not determined 200 400 600 800 1000 1200 1400 1600 1800 2000 *THMF = tetrahydro-3-methylfuran scan number

A3. Chromatograms & Mass Spectra A3.5

Sample HP4, fumarole F0, t=394.8°C, June 4 2000 (scan EI+ mode)

m/z = 43+57 x 50 151 dimethyl sulfone 8.45e6 100 674 bleed 409 toluene 181 n-hexane isomer 24 H 1248 2-ethyl-1-hexanol 12 1238 bleed % 904 n-nonane 567 n-octane 1219 C 812 not determined 1013 2-ethylhexanal 464 2-hexanone 444 not determined 432 2-methyl thiophene (isomer) 130 acetone 490 n-hexanal 116 acetaldehyde 524 not determined 183 chloroform 798 3-heptanone 125 ethanol 338 dimethyl disulfide 163 2-methyl-pentane 591 not determined 311 not determined 1640 bleed 839 n-heptanal 358 4-pentenal ? 866 not identified 1153 n-octanal 142 DCM 151 171 3-methyl-pentane 1515 bleed 1681 not determined 2021 bleed 0

132 CFC-11 m/z = 35 100 5.38e6

147 CFC-113 y-scale x 800 142 DCM 183 chloroform 111-112 methyl chloride

% 229 tetrachloromethane 124 ethyl chloride 209 not determined 162 bleed 271 trichloroethylene (TCE) 406 notdetermined 338 dimethyl disulfide 467 not determined 626 bleed 606 not determined 723 ethyl thiophene isomer (2-, 3-) 946 not determined 1150 1,3-dichlorobenzene 1164 1,4-dichlorobenzene 772 not determined 986 benzaldehyde 1086 not determined 0

130 acetone TIC 100 6.61e7

182 tetrahydro-3-methyl furan 142 DCM 125 ethanol 128 674 bleed 1-butene 146 carbon disulphide 409 toluene % 111-112 methyl chloride 1238 bleed 224 benzene 109 carbon dioxide 564 2-furancarboxaldehyde (furfural) 1640 bleed 163 2-methyl-pentane 1496 bleed 202 211 428 not identified 470 impurity 151 dimethyl sulfone 1515 bleed 2020 bleed 0 805 styrene 200 400 600 800 1000 1200 1400 1600 1800 2000 202 methyl-cyclopentane 211 3-methyl-butanal scan number

A3. Chromatograms & Mass Spectra A3.6

aliphatic hydrocarbons - paraffins (alkanes)

Florian HP2 B130201D 132 (2.336) Cm (131:133-(133:134+130:131)) Scan EI+ 43 100 6.08e5

% 58

29

27 28 42

15 41 57

39

26 38 30 37 40 44 59 16 19 20 22 24 25 34 36 47 49 50 52 53 60 61 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62

n-butane

FL2 SPME FL2 299 (5.507) Rf (7,3.000) Scan EI+ 57 100 6.17e4

28

43

56 % 41

42

86

29 83 27 39 55 44 71 58 15 17 77 26 37 38 48 50 53 64 6869 70 73 79 80 21 24 30 3334 46 51 59 62 65 82 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88

n-hexane

profile tube 1, desorbed 10 mins @ 300C SITE01 837 (15.371) Cm (836:838-(838:846+832:836)) Scan EI+ 85 100 5.03e5

43

2-furancarboxaldehyde

57 71

%

41

95 96

56 70

29 39 84 114 55 42 27

86 28 37 38 40 67 69 94 32 44 53 58 115 25 50 51 63 65 72 76 92 97 104 24 26 31 3336 47 49 62 66 68 75 77 79 81 83 87 89 93 99 106 110 0 Da/e 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115

n-octane (coeluting with 2-furancarboxaldehyde)

A3. Mass Spectra A3.7

aliphatic hydrocarbons - paraffins (alkanes)

Florian HP2 B130201D 138 (2.441) Cm (137:138-(139+136:137)) Scan EI+ 43 100 3.36e4 42

41

%

27

57

72 55 15

29 56 44 70 21 46 64 22 23 35 73 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 n-pentane

FL2 SPME FL2 438 (8.055) Rf (7,3.000) Scan EI+ 43 100 5.65e4

71

57

41 %

56 70

42 100

29 55 27

39

15 72 40 58 85 34 53 61 63 64 65 67 78 81 87 92 101 102 20 30 33 36 38 51 73 77 84 86 8891 93 96 104 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 n-heptane

FL2 SPME FL2 881 (16.177) Cm (880:883-(883:885+878:880)) Scan EI+ 57 100 7.70e4

43

% 85

41 71

56 70

55 84 42 29 99 39 27 32 44 58 69 98 128 18 40 86 25 50 53 62 72 78 82 90 100 115 122 129 17 22 26 30 34 36 38 45 47 51 6163 65 67 77 79 83 87 91 93 96 107 108113 120 126 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 n-nonane

A3. Mass Spectra A3.8

aliphatic hydrocarbons - paraffins (alkanes)

FL2 SPME FL2 1123 (20.614) Cm (1121:1123-(1123:1126+1116:1121)) Scan EI+ 2.80e5 100 57

43

% 71

85

41

56 70 55 84 28 42 29 9899 58 69 27 39 44 142 72 86 113 25 32 50 53 64 73 8283 87 92 94 100 112 128 18 21 26 30 35 38 45 51 59 6567 7779 93 97 102 114 121 126133 136 140 143 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145

n-decane

FL2 SPME FL2 1571 (28.828) Rf (7,3.000) Scan EI+ 57 100 1.07e6

71

43

%

85

41 56

55 70 84 69 99 98 2829 58 72 83 86 112 27 39 44 97 113 170 38 51 53 59 65 67 73 100 126 127 165 17 26 32 7782 91 111 114 133 140 141 147 148 168 171 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170

n-dodecane

FL2 SPME FL2 1969 (36.125) Rf (7,3.000) Scan EI+ 57 100 2.69e5

71

%

43 85

56 41 84 99 55 98 29 58 70 72 44 86 112 141 156 54 113 198 27 30 39 48 76 100 126 127 142 155 157 16 1726 36 60 64 68 92 96 106 114 128 140 169 180 191 199 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

n-tetradecane

A3. Mass Spectra A3.9

aliphatic hydrocarbons - paraffins (alkanes)

FL2 SPME FL2 1353 (24.831) Rf (7,3.000) Scan EI+ 57 100 1.27e6

43 71

%

85

41

56 70 55 84

69 9899 29 58 27 39 44 72 83 86 112 113 156 17 26 33 38 5253 59 65 67 77 97 100 127 137 18 22 32 45 81 87 93 108 114 126 128 140 146 152 157 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 n-undecane

FL2 SPME FL2 1776 (32.586) Rf (7,3.000) Scan EI+ 57 100 2.36e5

71

43

% 85

41 55

28 84 99 29 70 82 98 32 58 72 67 86 27 39 54 112 113 44 51 81 100 126 127 135 157 169 184 20 38 60 64 79 92 93 96 106 114 133 141 149 154 155 166 177 185 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 n-tridecane

FL2 SPME FL2 2151 (39.461) Cm (2148:2151-(2151:2152+2146:2147)) Scan EI+ 5.96e4 100 57

71

85

%

43

55 41 99

69 111 197 72 28 58 83 86 113 141 155 29 97 44 100109 127 140 154 198 18 27 39 48 53 59 79 81 87 114 126 156 191 196 208 212 21 35 6467 96 128 151 163 170 179 181 214 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 n-pentadecane

A3. Mass Spectra A3.10

aliphatic hydrocarbons - paraffins (alkanes)

HARBOUR AIR, SEPT 10, 2001 HARBOUR 2185 (40.085) Cm (2182:2186-(2186:2192+2175:2181)) Scan EI+ 1.63e4 100 43

28

57

41

29

% 71

55 85

27 69

39 16 84 97 99 127 44 58 72 17 53 86 126 211 26 35 47 67 82 111 113 141 169 183 33 60 74 125 131140 151 163 165 182 210 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230

n-hexadecane

Florian HP2 HP2 2679 (46.912) Cm (2677:2681-(2681:2686+2674:2677)) Scan EI+ 2.10e3 100 57

43

71

41 %

85

55 69

29 32 40 58 39 84 99 15 44 82 98 72 101 192 36 64 86 96 113 20 24 52 53 125 127 161 209 212 78 79 129 153 159 162 224 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260

n-octadecane

HARBOUR AIR, SEPT 10, 2001 HARBOUR 2861 (52.478) Cm (2860:2862-(2862:2865+2856:2860)) Scan EI+ 2.55e4 100 28

%

43

57

41 71 16 55 85

27 69 83 99 22 3739 54 5867 73 86 97 45 93 111112 119 134 163191 194 207 209 265 267 281282 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290

n-heneicosan

A3. Mass Spectra A3.11

aliphatic hydrocarbons - paraffins (alkanes)

HARBOUR AIR, SEPT 10, 2001 HARBOUR 2265 (41.551) Cm (2263:2265-(2265:2268+2260:2263)) Scan EI+ 8.02e3 100 32

57

43 %

18

41 29 55

71

40 69 85 99

98 83 97 211 76 26 27 58 64 86 113 127 141 37 48 81 111 123 207 281 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 n-heptadecane

Florian HP4 (carboxen) HP4 163 (2.878) Rf (7,3.000) Scan EI+ 43 100 1.38e6

%

42

27 41

55

71 15 70 86 29 39

26 28 56 57 37 24 34 38 44 49 51 53 88 16 25 30 32 36 46 50 58 62 65 67 81 87 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88

2-methylpentane

Florian HP4 (carboxen) HP4 171 (3.019) Rf (7,3.000) Scan EI+ 56 100 4.03e5

57

41

29

%

27

39

28

71 26 53 50 58 85 86 30 34 51 70 84 16 20 22 24 35 46 62 64 65 67 69 74 76 78 80 82 87 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88

3-methylpentane

A3. Mass Spectra A3.12

aliphatic hydrocarbons - haloalkanes

Florian HP2 B130201D 141 (2.493) Rf (7,3.000) Scan EI+ x 100 142 1.97e5 100 28 31

43

79+81 6.32e3

108 3.37e4

27

127+142 2.97e3 127 %

TIC 1.59e7

135 140 145 26 45

147 15 18 47 74 56 121 73 139 25 55 62 67 39 48 79 157 162 20 61 81 35 65 82 91 95 96 105 111 184 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 methyl iodide (coeluting with bromoethane) inset: RIC and TIC of CH3I (m/z 127+142) and C2H5Br (m/z 79+81 and 108)

TX1 SPME TX1 133 (2.463) Cm (133:134-(134:135+131:132)) Scan EI+ 50 100 6.64e4

%

52

49

51 35 24 33 36 37 55 0 Da/e 15 20 25 30 35 40 45 50 55 methyl chloride

FL2 SPME FL2 300 (5.525) Cm (300:302-(302:304+299:300)) Scan EI+ 83 100 3.46e4

85

%

47

87

48 49 35 17 18 82 84 23 37 38 50 51 78 20 23 25 34 54 6364 66 68 72 74 77 89 92 0 Da/e 20 30 40 50 60 70 80 90 trichloromethane (chloroform)

A3. Mass Spectra A3.13

aliphatic hydrocarbons - haloalkanes

TX1 SPME TX1 143 (2.647) Cm (143:144-(144:145+141:142)) Scan EI+ 1.74e4 100 28

%

94

96

32

15 40 93 25 36 70 81 91 95 20 21 23 35 38 46 50 5557 62 69 76 83 85 88 100 100 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 methyl bromide

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 279 (5.140) Cm (278:280-(280:282+273:277)) Scan EI+ 49 100 1.23e5

84

%

86

51

47

35 79

48 88 37 83 20 22 36 50 64 65 75 41 54 60 63 67 69 71 73 77 81 85 90 9293 0 Da/e 20 30 40 50 60 70 80 90 methylene chloride (dichloromethane, DCM)

Florian HP4 (carboxen) HP4 228 (4.016) Cm (228:231-(231:232+227:229)) Scan EI+ 84 100 x 50 4.08e4 56 1.54e5

69+84 7.12e5

117 117+119 1.04e4 119

31 TIC 1.90e7

58

225 230 235 240

% 45 27

56 tetrachloromethane (carbon tetrachloride, m/z 117+119),

41 43 coeluting with thiophene (m/z 69+84); 39

92 box: m/z 56 at scan 234 is cyclohexane, 121

38 107 26 scan 229 is CCl4, 98 155 29 33 81 85 86 151153 47 49 50 69 94 103 112 18 25 60 93 127 134 140 145 157 19 36 80 123 150 scan 228 is thiophene 54 67 73 88 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

A3. Mass Spectra A3.14

aliphatic hydrocarbons - haloalkanes

Florian HP4 (carboxen) B130201A 125 (2.213) Cm (123:125-(125:126+123)) Scan EI+ 44 100 x 100 8.76e5

64

%

66

35

51 60 34 37 48 61 36 32 62 73 76 80 88 23 24 30 56 69 80 0 Da/e 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88

ethyl chloride (chloroethane)

FL2 SPME FL2 589 (10.824) Rf (7,3.000) Scan EI+ 107 100 109 6.78e3

%

97

27

64

96 112 23 63 68 83 20 103 69 79 81 91 95 117 41 98 146 153 186188 26 30 46 51 7477 127 139 170 33 115 137 173 34 86 89 159 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

1,2-dibromoethane

FL2 SPME FL2 652 (11.979) Rf (7,3.000) Scan EI+ 55 100 5.16e4

70

56

% 41

112 69

42 83 39 57

29 84 27 54 53 67 43 68 71 40 81 50 64 65 91 113 15 37 46 48 51 76 87 92 98 103 21 23 26 31 36 58 60 62 77 79 82 86 99 109 110 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115

octene (isomer 1)

A3. Mass Spectra A3.15

aliphatic hydrocarbons - haloalkanes

Florian HP2 B130201D 140 (2.476) Cm (140:141-(141:143+138:140)) 108 1.02e4 100 110 79+81 6.32e3

108 3.37e4

127+142 2.97e3

TIC 1.59e7

% 135 140 145

methyl iodide

79 142 81 93 80 95 111 127 0 83 88 90 91 104 122 m/z 92 98 101 106 112 115 121 125 128 134 138 141 147 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 ethyl bromide (bromoethane), coeluting with CH3I (m/z 127+142) aliphatic hydrocarbons - olefins (alkenes)

Florian HP2 B130201D 118 (2.091) Cm (118:120-(120+117:118)) Scan EI+ 39 100 1.58e4

41

%

56

54

53 55

51 50 37 52 49 57 23 36 40 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 butene (isomer)

FL2 SPME FL2 670 (12.309) Rf (7,3.000) Scan EI+ 55 100 1.35e4

70 56

% 41

69

112

42

39 83 29 44 57 94 64 53 17 84 27 43 67 16 37 51 65 79 103 20 22 38 50 62 68 71 81 87 25 35 46 48 88 97 98 108 115 26 33 58 78 92 106 118 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 octene (isomer 2)

A3. Mass Spectra A3.16

aliphatic hydrocarbons - alkenes & haloalkenes

Florian HP2 B130201D 350 (6.151) Rf (7,3.000) Scan EI+ 97 100 2.01e4

55

%

41 112

69

57 29 43 84 56 98 42 84 67 72 77 20 60 74 113 19 24 33 68 70 91 95 109 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115

trimethyl-2-pentene isomer

FL2 SPME FL2 327 (6.020) Cm (327:329-(329:330+326:327)) Scan EI+ 62 100 1.12e3

64 27

%

29 40

16 34 63 17 49 25 47 50 52 51 61 26 35 24 45 60 56 58 37 38 42 20 55 67 22 48 68 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68

chloroethylene

Florian HP4 (carboxen) B130201A 272 (4.786) Rf (7,3.000) Scan EI+ 95 100 4.83e3

130 132

97

%

60

47 20 35 48 84 99 133 24 25 33 73 75 83 134 64 62 76 80 46 78 136 23 77 90 122 128 88 137 0 Da/e 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140

trichloroethylene (TCE)

A3. Mass Spectra A3.17

aliphatic hydrocarbons - alkenes & haloalkenes

FL2 SPME FL2 992 (18.212) Cm (989:992-(992:994+986:988)) Scan EI+ 185 100 8.48e3

266 264

183 187 %

262 268 81 44 43

56 69 25 60 70 132 24 143 160 182 191 213 228231 30 36 48 96 137 151 161 240 248 258 269 17 97 112 126 128 196 203 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 tribromoethene

FL2 SPME FL2 262 (4.829) Rf (7,3.000) Scan EI+ 61 100 1.11e4

96

%

98

63

18

31 44

60

70 17 39 26 41 100 62 47 59 30 40 55 57 66 80 95 97 25 37 46 48 65 68 21 23 35 49 52 83 87 104 58 73 89 92 107 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 1,2-dichloroethene (isomer)

Florian HP2 B130201D 553 (9.705) Cm (551:554-(554:558+548:551)) Scan EI+ 28 100 3.04e3

43

% 44

87

56 85 18 32 42 129131 166 164 84

67 168 57 94 96 4748 64 7681 82 59 89 117 120 133 170 20 36 93 103 113 125 143 153 159 24 65 72 80 98102 109 134 145 175 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 tetrachloroethylene

A3. Mass Spectra A3.18

cyclic hydrocarbons - naphthenes & arenes

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 335 (6.167) Rf (7,3.000) Scan EI+ 56 100 2.97e5

41

%

39 69

27 55 42

28

29

67 40 53 84 26 38 50 51 57 68 65 30 37 52 62 70 73 77 79 82 19 22 25 46 49 59 63 74 85 93 96 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 cyclopentane

Florian HP2 B130201D 205 (3.614) Rf (7,3.000) Scan EI+ 56 100 6.75e4

39

70

69

%

41

55

38 28 40 84 37 74

46 62 18 26 53 30 47 50 57 67 71 33 36 51 64 22 66 73 79 80 98 99 21 83 8690 93 95 102105 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105

1,1-dimethylcyclopentane

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 365 (6.717) Rf (7,3.000) Scan EI+ 78 100 1.75e5

%

50 51 77 52

39

32 27 38 79 26 37 63 74 16 57 76 49 61 36 53 62 73 75 22 25 3435 4748 64 66 70 80 82 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82

benzene

A3. Mass Spectra A3.19

cyclic hydrocarbons - naphthenes & arenes

Florian HP4 (carboxen) HP4 234 (4.121) Rf (7,3.000) Scan EI+ 56 100 1.51e5

84

41

redo, no portrait %

HP427 234 39 69 55

42

29 40 53 54 26 67 85 15 44 50 51 34 62 65 73 81 21 24 35 47 61 63 7577 82 87 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 cyclohexane

Florian HP2 B130201D 981 (17.196) Rf (7,3.000) Scan EI+ 100 1.03e4 32

83

55

82

%

41

39 28 29 67 18 43

126

56 69 16 81 125 26 99 51 6566 70 84 109 34 4244 25 30 38 72 91 92 100 19 34 4648 5052 78 87 95 127 61 103 120 105 119 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 propylcyclohexane

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 505 (9.284) Rf (7,3.000) Scan EI+ 91 100 3.71e6

92

%

39

65 63 51 50 28 27 38 62 89 41 45 46 93 15 26 32 37 40 52 53 61 18 25 60 64 66 67 69 73 74 77 78 85 86 87 95 20 24 30 34 36 44 49 56 79 84 90 94 96 100102 104 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 toluene

A3. Mass Spectra A3.20

cyclic hydrocarbons - arenes

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 684 (12.565) Rf (7,3.000) Scan EI+ 91 100 1.82e5

%

32

106

51 39 65 77 78

50 63 92 41 43 52 79 89 103 38 62 74 75 105 37 48 53 56 58 61 67 71 73 81 86 94 98 107 34 42 44 49 57 64 66 80 87 90 93 97 102 0 Da/e 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 ethylbenzene

HARBOUR AIR, SEPT 10, 2001 HARBOUR 732 (13.445) Rf (7,3.000) Scan EI+ 104 100 8.74e3

78

17 %

16 51

30 103

40

50 27 29 77

33 63 82 102 26 74 38 52 76 87 89 37 45 55 68 34 48 61 6466 67 84 98 53 69 83 92 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105

styrene (cinnamene)

FL2 SPME FL2 992 (18.212) Cm (989:992-(992:994+986:988)) Scan EI+ 185 100 8.48e3

266 264

183 187 % choose one

262 268 81 44 43

56 69 25 60 70 132 24 143 160 182 191 213 228231 30 36 48 96 137 151 161 240 248 258 269 17 97 112 126 128 196 203 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270

4-ethyltoluene

A3. Mass Spectra A3.21

cyclic hydrocarbons - arenes

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 702 (12.895) Rf (7,3.000) Scan EI+ 91 100 3.30e5

%

106

39 105 51 77

27 63 65 78 79 50 52 92 32 103 15 38 53 62 89 17 74 75 102 107 25 30 36 37 41 55 60 64 67 81 85 98 26 44 46 49 54 61 66 73 80 86 90 9395 97 110 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 m-xylene (1,3-dimethylbenzene)

DF8/5 condensate DF85A 921 (16.911) Cm (921:922-(922:924+918:920)) Scan EI+ 91 100 2.13e4

%

120 56 52 38 65 48 92 105 36 40 47 66 50 63 68 78 82 114 42 70 81 101 118 54 55 88 109 115 49 64 76 93 94 100 123 0 Da/e 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 n-propylbenzene

Florian HP4 (carboxen) B130201A 1212 (21.239) Cm (1210:1213-(1213:1217+1200:1209)) Scan EI+ 105 100 2.79e3

%

120

119

16 63 77 103 106 91 51 96 44 50 62 76 78 79 94 117 121 20 2425 46 59 8687 100 110 115 127 31 52 61 66 73 75 125 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 1,2,3-trimethylbenzene

A3. Mass Spectra A3.22

cyclic hydrocarbons - arenes & haloarenes

Florian HP4 (carboxen) B130201A 1066 (18.683) Rf (7,3.000) Scan EI+ 105 100 6.98e3

% 120

HP4 1066

77 44 39 119 51 42 54 79 106 55 91 73 103 37 48 52 63 65 75 15 50 56 60 78 117 102 115 121 49 74 81 89 93 97 24 59 61 71 99 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 1,3,5-trimethylbenzene

Florian HP2 B130201D 653 (11.455) Rf (7,3.000) Scan EI+ 112 100 9.50e4

77

%

114

50 51

38 74 75 37 49 18 32 73 39 78 81 84 85 86 115 15 25 26 29 30 34 35 5556 60 63 64 69 71 88 96 103 20 42 44 47 52 72 97 101 108 130 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 chlorobenzene

TX1 SPME TX1 1025 (18.817) Rf (7,3.000) Scan EI+ 146 100 2.95e6

FL2 1079 148

redo 111 %

75

74 113 50

73 150

51 55 84 110 38 76 85 37 49 56 6162 87 114 151 22 26 28 36 45 54 63 72 77 86 96 100 108 131 135 145 25 32 39 47 68 83 88 9799 120 121 137 153 0 Da/e 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 p-DCB (1,4-dichlorobenzene)

A3. Mass Spectra A3.23

cyclic hydrocarbons - arenes & haloarenes

DF8/5 condensate DF85A 1094 (20.082) Cm (1093:1094-(1094:1096+1091:1093)) Scan EI+ 119 100 9.42e4

%

91 134 DF85A 1093

77 117 65 120 58 115 63 79 103 105 16 3941 57 78 89 93 133 135 26 34 36 50 56 100 19 23 35 44 5153 67 74 75 88 97 108 110 17 21 2931 49 61 69 72 818385 90 124127 128 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 methylisopropylbenzene isomer (cymene)

Florian HP2 HP2 240 (4.226) Cm (240:241-(242:243+238:239)) Scan EI+ 96 100 1.24e4

28

HP230 240

% redo

29

86 70 67

79 58 76 82 50 51 77 26 37 38 68 74 81 97 19 94 24 25 34 49 52 62 63 65 99 21 47 54 78 88 93 100 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 fluorobenzene

TX1 SPME TX1 1401 (25.711) Rf (7,3.000) Scan EI+ 180 100 3.15e5

182

% tx1 1401

109 74 145 184

105 75 147 73

84 72 111 90 28 85 148 49 50 55 92 144 61 62 112 120 37 52 186 18 38 44 47 56 79 96 97 103 118 121 150154 178 22 25 30 36 64 69 82 131 132 143 164 172 187 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 trichlorobenzene isomer

A3. Mass Spectra A3.24

cyclic hydrocarbons - PAH & others

HARBOUR AIR, SEPT 10, 2001 HARBOUR 1422 (26.096) Rf (7,3.000) Scan EI+ 128 100 6.33e3

% Harbor 1422

40

44

51 63 50 64 102 121 74 87 45 129 73 7576 101 127 77 31 62 103 48 81 122 3839 60 78 106 66 34 93 26 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 naphthalene

Florian HP4 (carboxen) B130201A 1410 (24.704) Rf (7,3.000) Scan EI+ 67 100 1.32e4

95

79 94 39 % 41 6668 121

136 93 HP4 1410 77 80

65 53 107 91 32 108 52 29 40 81 51 55 63 42 69 78 82 96 16 26 38 50 71 73 137 58 59 105 117 122 133 21 37 49 56 73 8586 120 22 25 33 46 99 102 112 141 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 octahydro-4,7-methano-1H-indene others: inorganic compounds

Florian HP2 B130201D 114 (2.021) Cm (114-(115+113)) Scan EI+ 81 100 5.28e6

HP2 114

%

47

96

77 82 27 83 15 66 67 18 26 41 43 49 54 61 69 74 94 97 98 23 40 42 51 53 55 58 62 68 72 75 78 84 86 89 91 93 99 101 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 difluorodimethylsilane

A3. Mass Spectra A3.25

cyclic hydrocarbons - PAH & others

FL2 SPME FL2 1748 (32.073) Rf (7,3.000) Scan EI+ 142 100 7.45e4

141

% FL228 1748 115 70

69 71

95 143 131 139 126 39 63 89 96 116 51 58 62 77 91 50 53 6566 74 8687 109 145 40 92 110 124 130 17 94 99 102 123 133 138 146 16 26 30 34 38 4648 61 7880 103 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 methyl naphthalene (1/2 isomers)

FL2 SPME FL2 1159 (21.274) Rf (7,3.000) Scan EI+ 117 100 5.38e5

118

%

FL2 1159 115 57

91

58

43 55 63 70 89 39 41 65 83 119 51 77 78 95 103 53 62 69 74 84 2526 27 29 37 42 45 50 52 64 67 71 75 81 87 90 92 93 98 99 18 20 31 34 38 40 44 48 59 96 104 112 113 123 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 2,3-dihydro-1H-indene (indan) others: inorganic compounds

Florian HP4 (carboxen) HP4 161 (2.843) Rf (7,3.000) Scan EI+ 75 100 1.29e6

% select one

45

47

42 43 76 77 59 61 69 79 20 32 49 56 66 71 81 9293 30 34 48 53 57 63 65 73 8082 85 88 9091 0 Da/e 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 trimethylsilanol

A3. Mass Spectra A3.26

oxygenated compounds - alcohols & ethers

Florian HP4 (carboxen) HP4 125 (2.213) Rf (7,3.000) Scan EI+ 18 100 1.05e7

31

45 %

16 29 27 46 43

26 28 42

25 41 20 21 23 24 32 33 3435 36 37 47 49 50 0 Da/e 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 1-ethanol (coeluting with water, m/z 18)

Florian HP2 HP2 1116 (19.558) Rf (7,3.000) Scan EI+ 94 100 1.36e6

%

66 39 65

38 40 55 63 95 27 37 50 51 26 29 47 61 62 15 20 31 32 43 53 67 68 71 73 74 75 88 97 17 19 25 33 36 42 44 49 57 60 77 79 82 84 86 89 91 96 99 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 phenol

Florian HP2 HP2 2311 (40.472) Rf (7,3.000) Scan EI+ 205 100 7.53e3

%

41 206 220

105 145 91 177 40 55 77 161 50 58 65 81 81 95 102 115 53 72 73 119 34 39 87 129 135 149 159 163 84 97 178 193 194 208 61 112 141 175 204 221 154 182 0 Da/e 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 2,6-bis(1,1-dimethyl)-4-methylphenol

A3. Mass Spectra A3.27

oxygenated compounds - alcohols & ethers

HARBOUR AIR, SEPT 10, 2001 HARBOUR 1857 (34.071) Cm (1856:1859-(1861:1863+1853:1855)) Scan EI+ 41 100 9.58e3

29 32

43

55 57

%

27 44

69 82 97 39

67 70 83 71 81 96 54 45 84 95 53 98 111 58 26 66 86 112 125 137 104 110 140 141 153 156 35 48 73 80 94 20 37 49 61 76 106 124 134 149 159 167 0 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 1-dodecanol

Florian HP2 HP2 1119 (19.611) Cm (1119:1122-(1123:1125+1115:1118)) Scan EI+ 28 100 8.56e3

% 64

128

92 81 130

18 98 83 97 140 102 2022 46 78 100 111 16 19 23 35 104 126 135138 144 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 chlorophenol (isomer) m/z 28 is contaminant

TX1 SPME TX1 168 (3.105) Cm (167:169-(169:171+165:167)) Scan EI+ 31 100 8.51e4

59

74

45

% 29

27

43

41 18 73 32 44 15 26 28 30 42 60 75 19 46 58 17 21 23 2425 33 35 36 3940 57 6567 69 72 76 78 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 diethyl ether

A3. Mass Spectra A3.28

oxygenated compounds - aldehydes

Florian HP2 B130201D 116 (2.056) Cm (115:117-(117:118+114:115)) Scan EI+ 29 100 6.75e5

30 %

17

28 44 18 43

19 16 42 25 31 41 20 21 22 23 24 34 35 36 40 45 46 0 Da/e 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 acetaldehyde (ethanal)

Florian HP4 (carboxen) B130201A 1154 (20.224) Cm (1152:1158-(1159:1160+1149:1150)) Scan EI+ 28 100 4.26e3

43 41

% 29 18

5657 55 84

42 39 69 67

44 81 82 40 95 26 85 15 72 51 70 100 17 20 25 54 91 110 30 36 50 65 75 76 83 89 92 101 133 139 34 4648 59 62 86 98 109 115 119123 128 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135

n-octanal

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 912 (16.746) Rf (7,3.000) Scan EI+ 77 100 1.24e5 105

28 106 51

%

50

32

78 52 29 39 74

38 37 17 49 63 75 91 26 62 79 107 53 61 73 103 22 24 30 44 58 89 96 108 111 19 25 34 36 48 66 68 84 86 95 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 benzaldehyde

A3. Mass Spectra A3.29

oxygenated compounds - aldehydes

Florian HP4 (carboxen) HP4 1013 (17.756) Rf (7,3.000) Scan EI+ 72 100 4.49e4

57

%

41 28 43

29

39

32 55 42 97 53 15 40 54 69 70 73 82 18 34 37 38 58 59 66 67 77 81 85 86 100 22 25 30 34 49 51 74 90 99 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105

2-ethylhexanal

FL2 SPME FL2 1539 (28.241) Rf (7,3.000) Scan EI+ 57 100 3.16e5

55

43

82 70

% 41 71 68 83

44 81 67

95 96 112 84

29 97 54 72 110 39 45 109 85 27 53 58 66 79 86 99 113 128 59 65 73 105 38 46 51 77 91 133 16 2023 26 31 36 115 119 123 138 145148 153 154 0 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 n-decanal

FL2 SPME FL2 649 (11.924) Rf (7,3.000) Scan EI+ 95 100 96 6.02e4

%

39

38

29 97 36 37 42 67 17 51 54 58 60 68 46 48 53 59 66 72 77 79 85 92 94 98 100 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 furaldehyde (furfural)

A3. Mass Spectra A3.30

oxygenated compounds - ketones

FL2 SPME FL2 223 (4.114) Cm (222:224-(224:226+221:222)) Scan EI+ 43 100 1.81e5

58 %

28

42 29 27 15 39 26 32 41 57 17 25 37 38 44 59 16 19 20 22 23 25 30 31 33 34 36 40 45 46 48 50 51 52 53 54 55 60 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 acetone (propanone)

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 729 (13.390) Cm (727:731-(731:736+723:727)) Scan EI+ 57 100 4.41e4

29

%

27

41

85

39 72 43 42 26 114 15 40 53 55 56 37 44 58 68 70 96 24 38 49 50 65 73 76 81 86 87 92 98 111 115 19 25 31 36 45 52 59 61 66 77 79 83 91 95 97 99 104 109 119 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 3-heptanone

Florian HP4 (carboxen) HP4 464 (8.147) Cm (462:465-(466:467+457:462)) Scan EI+ 43 100 2.02e4

%

58

29 87 28 41 57

32 59 39 85 100 15 18 71 72 26 40 20 34 38 50 5153 61 80 17 19 24 30 34 36 54 56 60 64 66 68 73 77 81 83 88 90 101 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 2-hexanone

A3. Mass Spectra A3.31

oxygenated compounds - ketones

FL2 SPME FL2 555 (10.200) Rf (7,3.000) Scan EI+ 43 100 7.35e5

85

100

%

41 42 39 72 29 58 86 15 27 44 55 69 101 17 19 26 31 37 38 45 49 53 62 66 84 25 36 40 50 57 61 63 67 71 73 81 87 89 102 105 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105

2,4-pentanedione

HARBOUR AIR, SEPT 10, 2001 HARBOUR 1138 (20.889) Cm (1138:1140-(1140:1147+1133:1138)) Scan EI+ 77 100 1.00e4

105

51 %

43

50 120

15 29

39 78 74 106 27 33 63 26 30 57 99 38 69 73 75 16 41 52 61 62 82 100 121 19 37 49 58 65 70 84 97 109 20 25 34 45 47 53 67 72 86 89 93 111 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 acetophenone (1-phenyl-ethanone)

Florian HP4 (carboxen) HP4 769 (13.485) Cm (768:771-(771:775+762:768)) Scan EI+ 55 100 8.67e3

42

28 98 %

39 41

69

70

32 40 26 29 67 83 56 73 17 38 43 52 74 80 44 51 95 15 18 37 49 78 79 81 96 99 25 36 45 48 62 64 65 68 71 100 19 22 34 58 61 75 84 86 8991 94 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 cyclohexanone

A3. Mass Spectra A3.32

oxygenated compounds - acids & esters

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 347 (6.387) Rf (7,3.000) Scan EI+ 43 100 9.64e5

45

%

60 15

29 42 28 44

16 18 31 41 17 25 26 30 46 19 20 21 22 24 34 36 40 47 48 49 52 55 56 5758 59 61 62 64 65 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 acetic acid (ethanoic acid)

GELSO LIGHTHOUSE AIR, SEPT 10, 2001 LIGHTHOU 317 (5.837) Rf (7,3.000) Scan EI+ 29 100 4.26e5

% 46

45

44

32

16 17 30 22 23 27 31 35 36 41 47 48 52 0 Da/e 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 formic acid (methanoic acid)

FL2 SPME FL2 643 (11.814) Rf (7,3.000) Scan EI+ 43 100 1.21e6

56

%

73

41 61

55 57 85 28 29 71 42 27 39 44 58 69 70 87 15 37 40 45 50 75 77 84 86 91 96 98 109 112 18 26 31 34 38 51 53 60 62 64 68 72 74 81 88 92 97 101 102 114 115 118 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 acetic acid butyl ester

A3. Mass Spectra A3.33

oxygenated compounds - acids & esters

HARBOUR AIR, SEPT 10, 2001 HARBOUR 1225 (22.484) Cm (1223:1227-(1227:1232+1220:1223)) Scan EI+ 73 100 1.92e4

88

41

29

27

%

39 32 55 57

43

45 87

101

116 17 69 26 53 18 74 115 59 70 89 38 51 83 19 25 50 60 67 77 79 939798 102 119 91 109 126 134 135 147 0 Da/e 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 2-ethylhexanoic acid

Florian HP2 B130201D 742 (13.013) Rf (7,3.000) Scan EI+ 60 100 4.07e4

%

41

43

28

45

39 87 42

29 32

15 38 61 17 57 55 95 19 23 26 35 37 48 49 50 54 63 70 74 79 82 86 16 22 40 69 75 80 83 88 91 93 100 101 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 isovaleric acid (3-methylbutyric acid)

Florian HP2 B130201D 294 (5.171) Rf (7,3.000) Scan EI+ 41 100 1.16e5

39 69

%

100 28 99

40 38 15 59 85 32 37 98 53 56 81 82 44 45 17 22 33 64 70 77 79 86 101 107 21 25 36 47 60 62 80 83 90 93 105 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 methyl methacrylate (2-methyl-2-propenoic acid methyl ester)

A3. Mass Spectra A3.34

oxygenated heterocyclic compounds

Florian HP2 B130201D 136 (2.406) Cm (136:138-(138:139+136)) 42 1.13e4 100

31

55

% 29 45

39 18

27

70

64 68

32 58 30 46 44 15 28 41 40 48 56 19 34 38 72 17 62 65 66 69 16 20 21 23 35 61 71 75 0 m/z 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76

furan

Florian HP2 B130201D 194 (3.421) Rf (7,3.000) Scan EI+ 55 100 2.43e4

84

%

39

69 56

18 53

54 61 50 83 51 85 33 66 35 68 70 21 22 37 65 72 77 80 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 dihydromethylfuran

HARBOUR AIR, SEPT 10, 2001 HARBOUR 401 (7.377) Cm (399:401-(401:409+394:399)) Scan EI+ 28 100 1.09e4 29

88

58

% 27

31

15 43 30

26 57

44

42 45 55 87 33 97 17 19 20 41 46 47 59 66 72 73 80 84 89 93 98 24 36 38 50 53 61 63 65 69 76 79 81 85 86 94 103 104 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 1,4-dioxane

A3. Mass Spectra A3.35

oxygenated heterocyclic compounds

Florian HP2 B130201D 198 (3.491) Cm (196:198-(198:200+194:196)) Scan EI+ 42 100 2.27e4

71

72 %

41

28

39 40 29 26 38 57 1718 30 22 37 22 25 34 36 58 76 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 tetrahydrofuran

Florian HP2 B130201D 293 (5.154) Cm (291:293-(293:295+289:291)) Scan EI+ 95 100 96 6.71e3

28

%

53

81 18 86

45 50 51 97 65 44 73 77 78 47 48 58 76 84 21 23 68 80 87 89 93 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 2,5-dimethylfuran

Florian HP2 HP2 239 (4.209) Cm (238:240-(240:242+236:238)) 4.15e4 Scan EI+ 29 56+102 100 2.41e4

70+96 1.83e4 86

31 86 5.13e4 30

TIC TIC 1.30e6

27 28 235 240 245 250

%

56

41

58

43 26

42

57

33 15 39 83 87 38 55 18 19 24 25 44 49 52 72 73 20 36 46 48 50 59 66 74 88 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 2,3-dihydro-1,4-dioxin (m/z 56 is from coeluting compound)

A3. Mass Spectra A3.36

sulfonated compounds - thiophenes, sulfides & sulfones

FL2 SPME FL2 362 (6.662) Rf (7,3.000) Scan EI+ 28 100 5.22e4

84

%

58

45

44 57 85 86 16 17 83 26 40 49 51 64 69 81 27 29 37 38 42 50 60 63 66 71 72 19 21 23 33 36 47 53 56 68 74 76 80 88 89 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 thiophene

FL2 SPME FL2 245 (4.517) Rf (7,3.000) Scan EI+ 76 100 6.12e5

%

78 28 32 44 77 38 39 42 6264 6667 83 17 1920 23 24 26 31 34 36 43 46 48 5360 61 65 6970 72 74 80 82 85 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 carbon disulfide

Florian HP2 B130201D 343 (6.029) Cm (342:344-(344:346+338:342)) Scan EI+ 94 100 4.10e4

% 79

45

46

47

28 61 64 96 48 84 29 43 54 55 81 95 32 38 39 44 50 49 53 66 82 103 33 35 40 41 56 58 60 63 65 68 71 76 80 85 97 100102 0 Da/e 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 dimethyldisulfide

A3. Mass Spectra A3.37

sulfonated compounds - thiophenes, sulfides & sulfones

Florian HP2 B130201D 748 (13.118) Rf (7,3.000) Scan EI+ 100 111 4.67e3

112

32

97

%

94

59

49

94 71 113 34 36 30 58 46 33 47 98 20 70 35 100 101 114 21 25 119 0 Da/e 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 dimethylthiophene isomer (1/4)

TX1 SPME TX1 164 (3.032) Cm (164-(164:165+162:163)) Scan EI+ 28 100 1.02e4

62

%

32

47

27 29 45 61 18 46 57 25 34 40 59 23 26 33 51 58 71 74 16 19 22 35 5253 60 73 7578 80 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 dimethyl sulfide

FL2 SPME FL2 256 (4.719) Cm (255:257-(257+254:255)) Scan EI+ 79 100 3.66e5

%

47 80 45 94 64 77 78 81 15 24 27 28 29 31 37 43 54 59 61 65 75 83 92 1620 26 34 38 42 46 49 51 53 56 63 67 69 71 73 82 85 87 89 91 95 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 dimethylsulfone

A3. Mass Spectra A3.38

nitrogenated compounds - heterocycles, nitriles & amines

Florian HP4 (carboxen) HP4 368 (6.467) Cm (368:373-(373:378+364:367)) Scan EI+ 67 100 1.05e4

% 39

40 41

28 38

37 32 17 18 34 68 16 36 52 63 51 23 25 26 34 43 46 4748 54 56 59 64 65 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 1H-pyrrole

site 10 SITE10 121 (2.143) Cm (120:122-(122+119:120)) Scan EI+ 53 100 2.01e5

52

26

%

51

28 44 17

50

54 24 23 36 0 Da/e 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 acrylonitrile (2-propenenitrile)

Florian HP2 HP2 288 (5.066) Rf (7,3.000) Scan EI+ 55 100 5.71e4

54

%

29

28 18

52 17 26 38 40 56 37 51 16 21 25 32 48 0 Da/e 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 isocyanoethane (ethylisonitrile)

A3. Mass Spectra A3.39

nitrogenated compounds - heterocycles, nitriles & amines site 13 SITE13 377 (6.624) Cm (377:380-(387:391+374:376)) Scan EI+ 79 100 5.35e4

52

%

51 50

43

57 39 41 54 69 72 49 80 31 53 58 62 66 75 76 82 30 34 46 48 60 62 64 65 71 77 81 0 Da/e 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 pyridine

site 11 SITE11 309 (5.434) Cm (309:312-(313:315+307:308)) Scan EI+ 80 100 9.48e4

%

26

53

31

52 2829 4142 51 39 59 56 40 25 38 45 65 81 30 54 66 72 73 17 24 36 50 55 61 71 18 2021 24 34 37 44 47 60 62 69 77 78 82 84 0 Da/e 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 pyrazine

Florian HP2 B130201D 382 (6.712) Rf (7,3.000) Scan EI+ 102 100 28 3.68e4

27

29

58

42 87

%

44 41

43 39 32 26

57 30

18 40 72 59 38 70

19 25 75 82 36 61 81 88 91 92 99100 22 34 47 60 66 69 78 97 104 0 Da/e 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 N-ethyl-N-nitroso-ethenamine

A3. Mass Spectra

Acknowledgements This project was initiated by T.M. Seward and V.J. Dietrich who had already examined Christoph Wahrenberger’s thesis before mine. The two of them were always a great source of support and guidance, even if the duties as an assistant at IMP didn’t always allow full-time occupation with the project. Over the years we became good colleagues and friends and I will always lok back at the time at IMP and the geochemistry group as that of a family life.

Andy Gize joined the project later and we always had a great time in ice cream competitions in Italy, Zürich and Manchester. Having him as a co-advisor and friend resulted in a lot of quite humorous situations which made frustrating moments of machine failure bearable. Our trips to Wales and through Italy were a pleasure and “almost a holiday”, and we laughed a lot - especially during the re-ocurring situations of ferry embarakation with a British vehicle on the continent or when Italian highway patrol saw no driver on the left seat of the british rental truck (with flashing lights). He taught me a lot.

Terry Seward became my mentor and friend whose endless experience, unbreakable optimism and always abundant knowledge of minute details was a vital resource when I got stuck in my interpretations. He introduced me to the art of dining in southern Italy and the extremely exhausting daily work schedule on Vulcano. On one memorable occasion we ended up in a seemingly Mafia-run posh restaurant on Sicily, with the greatest fresh seafood display I have ever seen. His selfless ability to carry water and oranges up a volcano was never matched by any of us and I enjoyed every drop of it! I thank him also for the freedom and resources to spend extra time no getting the ETH-IMP GC-lab up and running. I am very much loocking forward to further culinary/field experiences.

Wumme (Volker Dietrich) has been my advisor and friend for many years even predating my time at ETH-IMP. I always enjoyed his unconventional approach to any obstacles and his down-to-earth view on many issues. Spending time in the field, sailing, eating out or just trying to get field equipment re-stocked, prepared and packed within one day were typical experiences with him. My first impression of field work with Wumme was: getting up at the crack of dawn, running up and down the mountains in a hurry, and having a nice cold beer at lunch, with a side dish of garlic-fried shrimps.

Jenny showed great patience during the last half year of the thesis and without her selfless support (...lots of dal with lots of love) none of this would have ever been bearable.

A great number of people at ETH and elsewhere have been an influence or help during the last 4 years and I want to thank everyone for patience, friendship and support. Among them are the Hall family and co-workers at Hall Analytical Inc. (Keith, Margaret, Tony, Mike, Vic and Sue), whom I want to thank for fruitful discussions, good coffee, and logistic help developing and performing our method. Don Maurizio’s and Don Piruccudu’s help on Vulcano is greatly appreciated in supplying us daily with fresh ice for work. I want to thank N. Keller, A. Stefansson, C. Wahrenberger and A. Widler (all ETH-IMP) for help during sampling campaigns, Rolf Kipfer (EAWAG, CH) and Zhou Zheng (ETH IGMR, CH) are thanked for analyzing some volcanic gases for noble gas isotopes. Joanne M. Kauffman (MIT) provided valuable insights into the politics of environmental sciences during discussions. Prof. Dr. Josef Nösberger and Dr. Matthias Baltisberger (both ETH) are thanked for identification of some collected plant species and botanical advice. Dr. Pierre Gex (Univ. Lausanne, CH) kindly shared his original geophysical data along our soil degassing profile.

Curriculum Vitae

Name Florian Maximilian Schwandner Date of birth 13.08.1970, München (Germany) Adress Florastr. 21 8008 Zürich, Schweiz Citizenship German Marital status single Parents • Ms Erika M. Tappe (artist, Berlin, Germany) • Mr Prof. Dr. Dipl.-Ing. E.-L. Schwandner (archaeologist, Berlin) Siblings • Ms M.A. Stephanie Schwandner-Sievers (ethnologist, London, UK) • Ms Dipl.-Ing. Angela Schwandner (architect, Berlin) • Mr Julian Thaddeus Schwandner (student, Berlin) • Ms Imogen Adele Schwandner (student, Berlin)

Education 1977-81 Johannes-Tews-Grundschule (elementary school), 14129 Berlin, Germany 1981-83 Nord-Grundschule (elementary school), 14163 Berlin 1983-91 Schadow-Oberschule (highschool), 14163 Berlin 1991-97 Freie Universität Berlin, Berlin; undergraduate & graduate program “Geologie-Paläontologie” 1995-1996 University of Washington, USA (Dept. of Geological Sciences, Visiting Graduate Scholar) 1/98-7/02 ETH Zürich, D-ERDW, PhD student

Degrees (pre-PhD) 04.06.91 Abitur (university qualifying exam), Schadow OG, Berlin 09.11.93 Vordiplom Geologie-Paläontologie, FU Berlin 04.11.97 Diplom (MSc equiv.) Geologie-Paläontologie, FU Berlin, thesis: “Die Geologie der Insel Poros (Griechenland)” (The geology of Poros Island, Greece)

Stipends & Awards 1995-96 Exchange scholarship, University of Washington for 3 quarters (9 months) 04/97 Travel stipend (research) of FU Berlin 11/97 Diplom “mit Auszeichnung” (with distinction), FU Berlin 09/01 Travel stipend (research) of SANW (Schweizerischen Akademie der Naturwissenschaften) 10/01-3/02 PhD student scholarship of ETH Zürich

Professional experience 11/89-11/90 Assistant to manager, sales & store, City Music GmbH, Berlin (Steglitz branch) 08/1992 Drilling assistant, archaeological excavation of D.A.I. Berlin (Deutsches Archäologisches Institut) in Stratos/ Greece 10/94-09/97 Teaching assistant (“Studentische Hilfskraft mit Unterrichts- aufgaben, Tutor”) Geologie-Paläontologie, FU Berlin 10/95-06/96 Research assistant (Geochemistry), University of Washington 01/98-09/01 Research & teaching assistant (SNF), ETH Zürich, D-ERDW; set-up, operation and maintenance of gas chromatography labs, XRF and ICP preparation lab, webmaster of IMP analytical labs 04-06/02 Research assistant (EU grant), ETH Zürich, D-ERDW & D-BAUG since 07/02 Research scientist (EU grant), ETH Zürich, D-ERDW & D-BAUG

Teaching experience 1994-present field work (mapping/sampling), geology, geochemistry, optical mineralogy & petrography, sample preparation techniques (ICP-MS & XRF), petrology and phase equilibria, aqueous geochemistry, thermodynamics, undergraduate thesis co-advisor, voluntary undergraduate course guidance/councelling.